Lecture Notes On General Relativity and Cosmology Mhadrat Fy Nzryt Alnsbyt Alamt

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/347200758
Lecture Notes on General Relativity and Cosmology ‫ﻣﺤﺎﺿﺮات ﻓﻲ ﻧﻈﺮﻳﺔ اﻟﻨﺴﺒﻴﺔ‬

‫اﻟﻌﺎﻣﺔ‬
Book · February 2019
CITATIONS READS
0 1,010
1 author:
Mohammed Basil Altaie

Yarmouk University
37 PUBLICATIONS 220 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Dark Matter is it Necessary? View project
Back-reaction of Quantum Fields in curved spacetimes View project
All content following this page was uploaded by Mohammed Basil Altaie on 15 December 2020.
The user has requested enhancement of the downloaded file.

Lecture Notes on General Relativity and
Cosmology
Prof. M.B. Altaie

Department of Physics, Yarmouk University
Feb. 2019
2
Contents
1 Review of Special Relativity 9

Pre-relativistic Physics . . . . . . . . . . . . . . . . . . . . . 9
Basic De…nitions . . . . . . . . . . . . . . . . . . . . . 10
Galilean Transformations . . . . . . . . . . . . . . . . . . . . 11
Relative Velocity . . . . . . . . . . . . . . . . . . . . . 11
Acceleration and Newtonian Mechanics . . . . . . . . . 12
Laws of Electrodynamics . . . . . . . . . . . . . . . . . 12
Lorentz Transformations . . . . . . . . . . . . . . . . . . . . 13
Implications of Lorentz Transformations . . . . . . . . 14
The Postulates of Special Relativity . . . . . . . . . . . . . . 14
Postulate (1): The Principle of Special Relativity . . . 15
Postulate (2): The velocity of light . . . . . . . . . . . 15
The spacetime interval . . . . . . . . . . . . . . . . . . . . . 15
Derivation of the Lorentz Transformations . . . . . . . . . . 16
Lorentz Transformations and Rotation . . . . . . . . . 18
Consequences of Lorentz transformations . . . . . . . . 19
Relativity of Simultaneity . . . . . . . . . . . . . . . . . . . 21
The Proper time . . . . . . . . . . . . . . . . . . . . . 21
Mass and Energy . . . . . . . . . . . . . . . . . . . . . . . . 22
Energy and Momentum . . . . . . . . . . . . . . . . . . 22
Minkowski Spacetime . . . . . . . . . . . . . . . . . . . . . . 23
Spacetime diagrams . . . . . . . . . . . . . . . . . . . . 24
The Light-Cones . . . . . . . . . . . . . . . . . . . . . 25
4-Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Velocity 4-Vector . . . . . . . . . . . . . . . . . . . . . 27
The 4-momentum . . . . . . . . . . . . . . . . . . . . . 27
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2 The Principles of General Relativity 31

Why relativistic Gravity . . . . . . . . . . . . . . . . . . . . 31
Di¢ culties . . . . . . . . . . . . . . . . . . . . . . . . . 32
Inertial mass and Gravitational mass . . . . . . . . . . 32
The Principle of Equivalence . . . . . . . . . . . . . . . . . . 33
Gravity and Acceleration . . . . . . . . . . . . . . . . . 33
General Covariance . . . . . . . . . . . . . . . . . . . . . . 34
Physical implications . . . . . . . . . . . . . . . . . . . 35
The Line Element . . . . . . . . . . . . . . . . . . . . 35
3
4 CONTENTS
Examples of line elements . . . . . . . . . . . . . . . . 36

Geometrical realization . . . . . . . . . . . . . . . . . . 36
Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Accelerated Observers . . . . . . . . . . . . . . . . . . 38
Rotating Coordinates . . . . . . . . . . . . . . . . . . . 38
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3 Tensors and Tensor Calculus 41

Coordinate transformations . . . . . . . . . . . . . . . . . . 41
Scalars . . . . . . . . . . . . . . . . . . . . . . . . . . 42
Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . 42
Tensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Contravariant tensors . . . . . . . . . . . . . . . . . . . 45
Covariant tensor . . . . . . . . . . . . . . . . . . . . . 45
Mixed tensors . . . . . . . . . . . . . . . . . . . . . . . 45
The Metric Tensor . . . . . . . . . . . . . . . . . . . . 46
Raising and lowering indices . . . . . . . . . . . . . . . 48
Magnitude of a Vector . . . . . . . . . . . . . . . . . . 48
Symmetric and antisymmetric tensors . . . . . . . . . . 49
Levi-Civita tensor: . . . . . . . . . . . . . . . . . . . . 50
The Volume and surface elements . . . . . . . . . . . . 50
Parallel Transport . . . . . . . . . . . . . . . . . . . . . . . 50
Covariant di¤erentiation . . . . . . . . . . . . . . . . . 52
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4 Spacetime Curvature 55
Spacetime Geometry . . . . . . . . . . . . . . . . . . . . . . 55
The Gaussian curvature . . . . . . . . . . . . . . . . . . . . 56
Riemannian geometry . . . . . . . . . . . . . . . . . . . . . 57
The locally Inertial Coordinate system . . . . . . . . . . . . 57
Geodesics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
Unchanging Direction . . . . . . . . . . . . . . . . . . . 58
Equation of geodesics . . . . . . . . . . . . . . . . . . . 59
Shortest Distance . . . . . . . . . . . . . . . . . . . . . 60
Examples . . . . . . . . . . . . . . . . . . . . . . . . . 61
Riemann Curvature Tensor . . . . . . . . . . . . . . . . . . 64
i
Symmetry Properties of Rmlk . . . . . . . . . . . . . . 64
Number of components . . . . . . . . . . . . . . . . . . 65
Ricci tensor . . . . . . . . . . . . . . . . . . . . . . . . 65
Constant Curvature . . . . . . . . . . . . . . . . . . . . 65
The Einstein Tensor . . . . . . . . . . . . . . . . . . . . . . 66
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
5 The Field Equations 67

Heuristic derivation of Einstein Equation . . . . . . . . . . . 67
The Newtonian Approximation . . . . . . . . . . . . . . . . 68
Energy-Momentum Tensor . . . . . . . . . . . . . . . . . . . 70
CONTENTS 5
Special Cases . . . . . . . . . . . . . . . . . . . . . . . 71
The Einstein Equations for Empty Spacetime . . . . . . . . 72
6 The Schwarzschild Solution 75

Spherically Symmetric Spacetime . . . . . . . . . . . . . . . 76
The Schwarzschild Solution . . . . . . . . . . . . . . . . . . 76
Schwarzschild Black Hole . . . . . . . . . . . . . . . . . . . . 79
Orbits in Schwarzschild Spacetime . . . . . . . . . . . . . . 80
Precession of Planetary orbits . . . . . . . . . . . . . . 82
7 Experimental Test of General Relativity 85

The Gravitational redshift . . . . . . . . . . . . . . . . . . . 86
The Bending of Light . . . . . . . . . . . . . . . . . . . . . . 87
Gravitational Lensing . . . . . . . . . . . . . . . . . . . . . . 90
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
8 Cosmology 97
The Einstein Static Universe . . . . . . . . . . . . . . . . . . 98
Dust Universe . . . . . . . . . . . . . . . . . . . . . . . 99
The Cosmological Constant . . . . . . . . . . . . . . . 100
The de Sitter solution . . . . . . . . . . . . . . . . . . . . . 101
The Robertson-Walker metric . . . . . . . . . . . . . . . . . 102
Friedmann Models . . . . . . . . . . . . . . . . . . . . . . . 103
Cosmological Red shift . . . . . . . . . . . . . . . . . . 103
The Einstein …eld equations . . . . . . . . . . . . . . . 104
The Equation of Fluid . . . . . . . . . . . . . . . . . . 106
The Friedmann Equations . . . . . . . . . . . . . . . . 106
Solving the Friedmann Equations . . . . . . . . . . . . 107
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
9 The Evolution of the Universe 115

6 CONTENTS
Preface
This is a brief introduction to the theory of general relativity and mod-

ern cosmology intended for a postgraduate course at Yarmouk Univer-
sity. The students are assumed to have little information about rela-
tivity, for this reason a short review of special relativity is provided in
Chapter 1. This will make the note more self-contained.
When teaching relativity theory it is important to stress the basic
concept of integrated spacetime, otherwise the real conceptual value of
relativity might not be properly valued. On the other hand the true
meaning of curved spacetime will not be realized unless the integrity
of the spacetime is well understood; it is important to understand how
time get bent through getting dilated by the gravitational redshift.
Throughout the book minimum use of tensors is made, only the nec-
essary relations concerning Riemann and Ricci tensors are presented,
however the implications of spacetime curvature by the geometrical
presentation is emphasized.
The last chapter of these notes is an introduction of modern cos-
mology. This part is presented mainly as an application of general
relativity and the solutions of the Einstein …eld equations, for this rea-
son no much details of the big bang theory are given.
There are a number of useful references that can be of real help in
this course, these are [1], [2], [3], [5], [6], [7].
Acknowledgement 1 I would like to thank my ex-postgraduate stu-

dent Usama Al-Ahmad for his help in preparing some of the home works
and part of the material of the last chapter.
7
8 PREFACE
Chapter 1
Review of Special Relativity
In this chapter I will present the basics of the special theory of relativ-
ity proposed by Albert Einstein in 1905 and the years that followed,
emphasizing the basic formulations and projecting the main aspects
related to the concept of spacetime continuum. It should be pointed
out here that although much of the basic ingredients of the theory of
special relativity were available in the literature at the end of the nine-
teenth century, but it was Albert Einstein who collected all the beans
of the chain to produce a meaningful and consistent framework that
can be used to understand the world. It was Einstein who consistently
uni…ed the laws of Mechanics and the laws of electromagnetism under
one scheme which was called the theory of Special relativity. There are
some useful videos on the you tube some of them are my own. The
reader can view these at
http://www.youtube.com/watch?v=p59E7jPZsG0
http://www.youtube.com/watch?v=czY8mDXmOVw
http://www.youtube.com/watch?v=l4x7Bd1fYrw
http://www.youtube.com/watch?v=l4x7Bd1fYrw
http://www.youtube.com/watch?v=8sYUvN88fnE
Pre-relativistic Physics
Classical Physics was based on three major subjects
1. Newton’s mechanics and his theory of Gravity
2. Maxwell’s electromagnetic theory and classical optics.
9
10 Review of Special Relativity
Figure 1-1 Refrence Frames
3. The kinetic theory of gases, thermodynamics and the Boltzmann

statistics.
The basic assumption in Newtonian mechanics is that space and

time are independent entities and are absolute; meaning that these
entities exist independent of any other physical e¤ect. Motion, accord-
ingly can be de…ned be de…ned con…dently as the change of position
during passage of time. However this de…nition would be incomplete
unless we de…ne certain frame of reference with respect to which the
position coordinates at each moment are de…ned.
Basic De…nitions
Before discussing the di¤erent transformation laws of motion let me
introduce some basic de…nitions which are necessary for this topic.
Frame of Reference
Any point which is at rest or moving can be considered a reference
for the motion. A frame of reference is any coordinate system that is
at rest or moving with any speed. A moving ( or even at rest ) car or
a ship or a plane or a lift can be considered as a frame of reference.
Inertial Frame of Reference
Is any coordinate system which is at rest or moving with constant
velocity with respect to any coordinate system.
Non-Inertial frame of reference
Is any coordinate system moving with arbitrary velocity. Accord-
ingly a rotating frame of reference is non-inertial.
Review of Special Relativity 11
Galilean Transformations
According to the classical doctrine space and time were considered as
two separate entities. In Newtonian Mechanics an absolute reference
for space is needed in order to de…ne motion properly. Time was taken
as a measure for change and motion1 . Time has no need to have any
reference except for its de…nition in terms of natural units (year, day
...etc.) or conventional units (hour, minute, seconds), otherwise time
was considered to be absolute as if it is a continuously ‡owing river.
In terms of this classical description of space and time, the relative
0 0
position of an object in a reference frame S moving along the xx axis
with constant velocity v with respect to another frame S which might
be considered at rest is
0
x =x t
0
y =y
0
z =z (1.1)
0
t =t
These relations were called Galileo Transformations. It expresses the

relativity of spacial positions and velocities with respect to moving
frames of reference.
Relative Velocity
0 0
According to (1.1) the velocity u of the object with respect to S in
terms of its velocity u with respect to S is
0
dx dx 0
= )u =u+
dt dt
This is the law of the velocity addition according to Galileo. The
velocity of an object depends on the frame of reference of the observer.
For example if two persons are standing on the back of a truck …ring
bullets in opposite directions then the bullets have di¤erent velocities
depending on the frame of the observer. In both directions the velocity
of the bullets is 1000 km/hr with respect to the men on the truck,
but these velocities are di¤erent with respect to the person in the rest
frame at the road. See the …gure below
The shape of the trajectory of a moving particle may also look
di¤erently when viewed from di¤erent frames of reference. For example
a ball thrown upward by an observer in a moving vehicle will have a
straight trajectory up and down but it will have a parabolic trajectory
when observed by an observer at rest on the road side.
1
However some Muslim thinkers declared that space and time are to be consid-
ered as two integral entities that cannot be separated from one another. (see: M.B.
Altaie, Daqiq al-Kalam: the Islamic Approach to the Philosophy of Nature).
Vel
2 :jpg
Figure 1-2 Relative velocity
Figure 1-3 Di¤erent trajectories from di¤erent frames of reference
Acceleration and Newtonian Mechanics

The relative acceleration observed by any inertial observer will be the
same for all since 0
d2 x d2 x
= 2
dt2 dt
The laws of Newtonian Mechanics fully satisfy Galilean transforma-
tions; the force in Newton’s second law of motion is invariant under
these transformations. If the velocity of light is taken to be c with
respect to one frame of reference, then the speed of the of light with
respect to any other frame moving with velocity v with respect to the
…rst frame is given by
u=c v (1.2)
where the sign depends on the direction of motion.
According to Galilean transformations the length of a vector de-
scribed in three dimensional space is invariant. An example of this is
the square of the distance between two points such that
0 0 0 0
L2 = x2+ y2+ z2= x2 + y2 + z 2 = L2
Laws of Electrodynamics
James Clark Maxwell formulated the basic laws of electromagnetism
in one package to become laws of classical electrodynamics. These can
be written in several forms di¤erent units. In vacuum these equations

take the form
r:E = 0
r:B = 0 (1.3)
@B
r E=
@t
1 @E
r B= 2
c @t
where E is the electric …eld, and B is the magnetic …eld. The above
set of equations are called Maxwell Equations.
Anton Lorentz found that Maxwell’s equations are invariant un-
der Lorentz transformations while Newton’s laws are invariant under
Galilean transformations. Accordingly, and in order to resolve the
problem we have either to change the laws of electrodynamics and get
them to satisfy Galilean transformations or modify Newton’s laws to
make them satisfy Lorentz transformations. This caused a problem for
physicists studying the electrodynamics of moving bodies. The force
expressed in 3 components is not invariant under Lorentz transforma-
tions. Therefore, a new theory was needed to reformulate Newtonian
mechanics in terms of the new transformations so that it will be in
harmony with Maxwell’s electrodynamics.
Lorentz Transformations
Anton Henderik Lorentz was an active theoretician who tried to explain
the behavior of electrically charged moving bodies using the notion of
the aether. His investigations led him to discover that Maxwell’s equa-
tions are not invariant under Galilean transformations but are invari-
ant under new transformations that involves space and time. These
transformations, which are now called Lorentz transformations. In one
dimension these are given by
x t
x0 = q = (x vt)
2
1 c2
y0 = y
z0 = z (1.4)
x
t c2 xv
t0 = q = t
1
2 c2
c2
p
where c = 1="0p0 is the velocity of light (or EM radiations) in vac-
uum, and = 1= 1 v 2 =c2 . From (1.4) it is clear that space is tightly
related with time and time is tightly related with space, whereas in the
Galilean transformations given by (1.1) space and time are separate
and independent.
Implications of Lorentz Transformations

From (1.4) it is clear that space and time are inter-related. Speci…cally
the 4th relation in Lorentz transformation indicates that the measure
of time depends on the frame of reference, and that determination of
the time in any inertial frame of reference is related to the position of
the point in space too and vise-versa. Beside this it is important to
notice that the determination of both space and time depends on the
relative velocity of the frame of reference. Moreover, the appearance of
the velocity of light c in these transformations is meant to be a constant
of nature, a point which was not well recognized before Einstein. It is
quite clear that as v becomes much less that c L.T. will give nearly the
same results obtained from G.T. Therefore one can say that G.T. is
the limiting case of L.T.
The Postulates of Special Relativity

This discrepancy between the Galilean transformations and the Lorentz
transformations, i.e., between Newtons laws of mechanics and the elec-
trodynamics required a resolution. Comparing (1.1) with (1.4) we can
easily see that Lorentz Transformations ! Galilean Transformation
when vc ! 0; which means that the discrepancy between mechani-
cal and electromagnetic phenomena may go unnoticed when vc << 1.
Despite this, it is important to know that, on the conceptual level,
Newton’s mechanics is only approximate and does not give the accu-
rate description of nature. In order to remedy this problem we have
to set up a new mathematical construct that would reformulate the
Newtonian mechanics in such a way that it would be compatible with
electrodynamics which means that it should satisfy the Lorentz trans-
formations. This task was actually done by Albert Einstein.
In 1905 Albert Einstein published his paper "On the Electrodynam-
ics of Moving Bodies" in which he suggested, what was later called,
the Theory of Special Relativity. Einstein viewed the time on the same
footing as space and for this reason he considered time as a 4th di-
mension, accordingly the Einsteinian space is a 4d spacetime in which
3 dimensions are for space and one for time. But to build his theory
Einstein found it necessary to assume two basic postulates, these are:
Postulate (1): The Principle of Special Relativity

Physics should apply to every point in space and time. Einstein recog-
nized that it should be necessary to say that:
The laws of physics must be invariant under inertial coordinates
transformations.
This principle was called The Principle of Special Relativity. There
are other expression of the above postulate but the statement chosen
above is perhaps the most accurate one. For example one can say also
that: the laws of physics must be expressed in the same form in all in-
ertial coordinate systems or say that: all inertial frames are equivalent.
In order to implement this principle physical vector quantities should
be expressed as 4-Vectors (vectors with 4 components) not 3-vectors:
a time-component is added in order to comply with the requirements
of invariance under Lorentz transformation.
Postulate (2): The velocity of light

Einstein correctly recognized that the velocity of light in vacuum c must
be a universal constant in order that the Lorentz transformations be
universal; c in Lorentz laws must be a universal constant. Therefore,
Einstein had to postulate that:
The velocity of light in vacuum is same for all inertial observers and is
independent of the state of motion of the observers or the source.
In fact this postulate was implemented already by Maxwell’s theory of
electromagnetism where c2 = 1= 0 0 where both are universal constant
which are independent of the state of motion of the observer or the
source.
Michelson-Morely experiment can now be explained by saying since
the light is not to be a¤ected by the relative motion of Earth through
the assumed ether therefore it would not be surprising to have a neg-
ative result and see no interference fringes. Note that this explanation
does not eliminate the possible presence of ether.
We now need to reformulate the laws of physics so that it becomes
invariant under Lorentz transformations not Galilean transformations.
This requires that the physical quantities be expressed in terms of
scalars or 4components vectors (3 for space and 1 for time).
The spacetime interval

The implementation of the principle of special relativity of Einstein,
postulate (1), requires that the physical quantities be described in 4
dimensions (3 spacial and 1-time) instead of being described in 3 di-
mensions. All physical laws are written in terms of space and time
components. Space and time is described by (x; y; z; t). In 4 dimen-
sions the in…nitesimal spacetime interval s is de…ned as
q
s = ( x1 )2 + ( x2 )2 + ( x3 )2 + ( x4 )2
the spacial coordinates x1 ; x2 ; x3 stands for x; y; z and x4 should contain
the time. The easiest way to construct the time in harmony with spacial
coordinates is to take as ct. but since we are looking for an axis which
is to be perpendicular to the other 3 axes, we should consider setting
the new axis to be imaginary so that x4 = ict: As such the spacetime
interval can be expressed as
q
s = ( x)2 + ( y)2 + ( z)2 c2 ( t)2
ds2 = dx21 + dx22 + dx23 + dx24 : (1.5)

where
x1 = x; x2 = y; x3 = z; x4 = ict
This should be invariant under Lorentz transformations given by (1.4).
Derivation of the Lorentz Transfor-

mations
Einstein realized that the space-time interval de…ned by (1.2) is invari-
ant under Lorentz transformations and accordingly he re-derived the
Lorentz transformations implementing his postulates mentioned above.
If we consider a two dimensional case only (that is to say that the mo-
0 0
tion of S with respect to S is along the xx axes only we will have
0 0
x12 + x42 = x21 + x24 (1.6)
this correspond to a rotation in the x1 x4 -plane which has the solutions

0
x1 = x1 cos + x4 sin
0
x4 = x1 sin + x4 cos (1.7)
In matrix form this can be written as

0
x1 cos sin x1
0 = (1.8)
x4 sin cos x4
where is the angle of rotation in the plane which would be a function

0
of the velocity v of S with respect to S. The relation between and
v can be determined as follows.
0
An object at rest in the system S must have a velocity v with respect
0
dx1 dx1 iv
to system S; hence, for 0
dx4
= 0; dx4
= c
. Now Eqn. (1.7) give
dx1
dx1
0
dx4
cos + sin
0 = dx1
=0 (1.9)
dx4 sin + cos
dx4
it then follows that

iv
tan = =i ; (1.10)
c
where = v=c: Accordingly
1 iv=c
cos = =q ; sin =i =q (1.11)
v2 2
1 c2
1 vc2
Substituting in (1.7) we get

x t
x0 = q = (x t)
2
1 c2
and x
t c2
t0 = q = (t x=c)
2
1 c2
which are the …rst and the fourth equations of Lorentz transformations
given in (1.4).
Home Work (2): Verify that (1.5) is invariant under (1.4). This
means that
0
ds 2 = ds2 (1.12)
Note that in special relativity ds is called the spacetime interval but
later when we study General Relativity we will call it the spacetime
metric or the line element.
The general form of Lorentz transformation can be written as
0
X
4
x = x (1.13)
=1
So that
0
X
4
dx = dx (1.14)
=1
Note that from now on we will label the coordinates with upper indices
and label coe¢ cients with upper, lower or mixed indices.
We will use the Einstein Summation Conversion by which repeated
indices are summed over without writing explicitly the summation sign,
so (1.13) becomes
x0 = x = 1x
1
+ 2x
2
+ 3x
3
+ 4x
4
(1.15)
where are constant coe¢ cients for a constant velocity belonging to
Lorentz transformations.
We note that
0
dx = dx
As mentioned above, the Lorentz transformations are required to pre-
serve the spacetime interval ds2 invariant, i.e.,
ds2 = (dx1 )2 + (dx2 )2 + (dx3 )2 + (dx4 )2

0 2 0 2 0 2 0 2 0
= dx1 + dx2 + dx3 + dx4 = ds 2 (1.16)
this leads to
0 1
0 0 i
B 0 1 0 0 C
=B
@
C (1.17)
0 0 1 0 A
i 0 0
So that we can write Eqn. (??) in matrix form as
0 01 1 0 10 1 1
x 0 0 i x
B x02 C B 0 1 C
0 0 C B x2B C
B 03 C = B C (1.18)
@ x A @ 0 0 1 0 A @ x3 A
x04 i 0 0 x4
Lorentz Transformations and Rotation

The Lorentz transformations can be expressed as a rotation with imag-
inary angle in 4d spacetime. If we set
= cosh ; = sinh (1.19)
then 0 1
cosh 0 0 sinh
B 0 1 0 0 C
=B
@
C
A (1.20)
0 0 1 0
sinh 0 0 cosh
this clearly correspond to rotation of the coordinates ict and x by angle
= i . For this reason we say that there is a homomorphic relation
ship between the Lorentz group and the rotation group. In fact rotation
in space forms a subgroup of the Lorentz group.
Figure 1-4 Length contraction
Consequences of Lorentz transformations

Length contraction
The spatial distance is not invariant under Lorentz transformation. It
can be easily proved that:
r
2
L = L0 1 :
c2
0 0
Assuming an observer in S measuring a distance along the x -axis
0 0 0
x = x 2 x1 :another observer in S will measure the same distance
as x = x2 x1 . from (1.18) we have
0 x
x =q (1.21)
2
1 c2
0
If L = x, and L0 = x therefore
r
2
L = L0 1 (1.22)
c2
0
Note that x is the distance measured at rest. This shows that the
spatial distance contract as the object moves faster. This is called
Lorentz-Fitzgerald contraction.
Time dilation
A gain the time interval in relativity theory is not invariant too. It
could be easily proved that
t0
t= q (1.23)
2
1 c2
where t is the time interval measured by the moving observer and

t0 is the time interval measured by the observer at rest. This can
be shown in standard text books (see for example Beiser, Concepts of
Modern Physics, 6th Edition, p. 6).
Home Work (3a): -measons are created at the elevation of 6000

m above sea level moving with a speed of 0:99c. The - meason are
known to live, on average, for only 2 s when at rest in the laboratory.
Despite this measons are found at sea level on Earth. Explain this.
Velocity addition
We can easily prove that the relativistic form of the velocity. This can
be obtained from the Lorentz transformations (1.4) where we have
dx dt
dx0 = q
2
1 c2
and
dx
dt c2
dt0 = q
2
1 c2
Consequently
0
0dx u
u = 0 = uv (1.24)
dt 1 c2
dx
where u = dt
is the velocity of the object with respect to rest frame and
0 0
dx
u = is the velocity of the object with respect to the frame moving
dt
0
with velocity v with respect to the rest frame. The minus sign should
be used if the velocities are parallel, and the plus sign is to be used if
the velocities are anti-parallel
Length contraction and time dilation were known before Einstein
but the law of velocity addition and the relativity of simultaneity be-
came known to Albert Einstein …rst.
Home Work (3b)
1. A rocket leaves a spaceship in the forward direction at a speed

of 0:5c as seen by its pilot. If the ship is cruising with a relative
speed of 0:4c with respect to an inertial observer what will be the
speed of the rocket with respect to the inertial observer.
2. Two beam riders are moving in opposite directions with the speed
of light with respect to an inertial observer. What is the relative
speed of one rider with respect to. the other?
Relativity of Simultaneity
In pre-relativity physics simultaneous events where always possible be-
cause time was an independent variable. But since space and time are
now interdependent, therefore simultaneous events are no longer pos-
sible for two reasons:
For observers who are at rest with respect to each other there is always
a …nite time interval between them that is caused by the …niteness of
the velocity of light. This is given by
L
t=
c
For observers in relative motion we have di¤erent coordinate-time mea-
sure depending on their relative velocities. Therefore the absolute si-
multaneity is lost according to special relativity. The situation can be
illustrated with the following example:
Consider the following experiment in the Einstein train. A light
source situated at the centre of the train A is switched on, and the
light beam travels forward and backward to reach the front and rear
ends of the train at the same time as viewed by an observer on the
train. Show that, as viewed from an observer on the road the arrival
of the light at these two points is not simultaneous, but reaches the
rear end of the train at a time vL=c2 before reaching the front end.
This means that a simultaneous event with respect to one observer
might not be simultaneous with respect to another observer. This is
an example of the relativity of simultaneity.
The Proper time

Because di¤erent observers measure di¤erent times, therefore, it is nec-
essary for the invariance of physical laws under Lorentz transformations
that a uni…ed time be found such a time should be di¤erent from the
coordinate time (dt) in that it has same value for all observers. This is
called the ‘Proper time’and is de…ned by:
2 ds2 dx2 + dy 2 + dz 2 c2 dt2

d = =
c2 c2
2 dx2 + dy 2 + dz 2
d = dt2
r c2
2
d = dt 1 (1.25)
c2
All inertial frames will measure same value of d . Clearly the proper-
time is what every observer will measure locally. That is to say it is
the time as measure by the observer at rest, this is why John Wheeler
called it the wrist-watch time.
In similar fashion the proper length (distance) can
Mass and Energy

The law of conservation of linear momentum requires that the mass of
a body should be velocity dependent, then the "relativistic mass" will
be given by
m0
m= q : (1.26)
2
1 c2
(for the derivation of this formula see A. Beiser, Concepts of Modern

Physics, 6th edition p. 22-24). So the mass of a moving body is larger
than the mass of the same body at rest. The additional mass is coming
from the kinetic energy of the moving particle. This can be seen from:
2
1
m = m0 (1 ) 2
c2
1 2
m = m0 (1 + + :::)
2 c2
1 m0 2
m = m0 + + ::: (1.27)
2 c2
K:E
m = m0 + 2 + :::
c
Note also that the kinetic energy in special relativity is given by
K:E = mc2 m0 c2 (1.28)

This inspired Albert Einstein to suggest the mass-energy equivalence
E = mc2
where m is the relativistic mass of the particle . Note that total con-
version of mass into energy can take place in events where particles and
anti-particles (like electrons and positrons) annihilate each other.
Energy and Momentum

The total energy of a particle of rest mass m0 and momentum p is given
by
E 2 = p2 c2 + m20 c4 (1.29)
This can be easily derived as follows. We have
m0 c2
E=q
2
1 c2
so that
m20 c4
E2 = 2
1 c2
and we have
m20 v 2 c2
p2 c2 = 2
1 c2
so that
E2 p2 c2 = m20 c4 :
Massless Particles
Massless particles are those for which the rest mass is zero like
photons. For massless particles like photons E = pc exactly, however,
we note that even for massive particles if the velocity is high (we call
them ultra-relativistic particles) then pc >> m0 c2 which means that
E ' pc: Most elementary particles acquire energies that allow them to
have high velocities and behave like ultra=relativistic particles.
Home Work (3d): Calculate the speed of a body that have total
energy equal twice its rest mass energy
Home Work (3e): A body at rest explode into two fragments
each of mass 1:0 kg that moves apart at speeds of 0:6c relative to the
original body. Find the mass of the original body.
Minkowski Spacetime
In 1907 Herman Minkowski suggested that a better realization of space-
time continuum could be achieved if we de…ne
x0 = ct; x4 = ict
x1 = x; x2 = y; x3 = z (1.30)
so that the spacetime interval of the special relativity becomes
ds2 = dx2 + dy 2 + dz 2 c2 dt2 (1.31)

The above metric (line element) describes a 4-dimensional ‡at
spacetime and is called Minkowski space.
Conventionally one can use also the line element multiplied by a
minus sign, i.e.,
ds2 = c2 dt2 dx2 dy 2 dz 2 (1.32)

Throughout this course we will adopt the quadratic form given in
(1.32) for the ‡at spacetime of Spacial relativity. This can be written
in more compact form
ds2 = dx dx (1.33)
where
Figure 1-5 Minkowiski Diagrams
0 1 0 1
00 01 02 03 1 0 0 0
B C B 0 1 0 0 C
=B
@
10 11 12 13 C=B
A @ 0
C (1.34)
20 21 22 23 0 1 0 A
30 31 32 33 0 0 0 1
The Minkowski space is a ‡at four-dimensional Euclidean space-

time. According to this realization spacetime diagrams can be drawn,
with reduced special dimensions, as such. These diagrams are called
spacetime diagrams or Minkowski diagrams. The spacetime of special
relativity is called Minkowski spacetime.
Spacetime diagrams
A spacetime diagram in 2d is formed of two axes one for time which may
be taken to be ct and the other for space which is taken by convention
as x. An event is represented in this diagram as a point. Accordingly
a line (collection of points) is a history. This history is called a world
line.
The line x = ct is a border line between the physical world (the
timelike world) in which events are connected with signals propagating
with a velocity less than or equal that of light, and the non-physical
world (the spacelike world) in which events are connected by signals
moving with velocities greater than that of light in free space. The
upper part of the diagram con…ne all the future events while the lower
part contains all the past events. The origin is a singular point that
stands for the present.
If we draw inclined lines in Minkowski space then we are presenting
another inertial frame of reference moving with velocity v with respect
to the original one. In Fig. (1-5) we have such a frame of reference
moving with velocity = tan 1 :Minkowski diagrams are very useful

in spacial relativistic calculations of positions and times.
The Light-Cones
From Minkowski diagrams we can form the so-called light-cone shown
in the Fig. (1-6) below;
Figure 1-6 3-d light-cone
The spacetime interval ds describing the geometry of the spacetime

will de…ne the causal status of events. In other word the causal status
will depend on the proper time d as de…ned above in (1.25).
If ds2 > 0 then this means that we have the events being connected
with a signal that moves with a speed less than that of light. Such a
spacetime is called time-like. These are causal spaces (e¤ects follows
the cause)
If ds2 < 0 then this means that we have the events being connected
with a signal that moves with a speed greater than that of light. Such
a spacetime is called space-like. These are non-causal spaces (cause
follows the e¤ect).
If ds2 = 0 then this means that we have the events being connected
with a signal that moves just with the speed of light. Such a spacetime
is called null space. They represent exactly the light paths.
4-Vectors
Since time has become an extra dimension Einstein realized that the
structure of the physical world is based in 4 dimensional space-time.
Accordingly the physical vector quantities should be regarded as having
Figure 1-7 Two di¤erent light-cones one representing a point at rest and
the other is a point moving with constant velocity
4 components and it is the norm of this 4-Vector which is invariant

with respect to all inertial frames.
A 4-vector is just the same as 3-vector but now have 4 components;
e.g.,
A = (A1 ; A2 ; A3 ; A4 ) = (A; A4 ) (1.35)

A: Stands for the spatial components. A4 : Stand for the time compo-
nent, also
X
4
A = en An = e1 A1 + e2 A2 + e3 A3 + e4 A4 (1.36)
n=1
The scalar components An can be obtained by scalar products:
An = en A (1.37)
where en are basis vectors that satisfy the orthonormality relations:
en em = nm : (1.38)
0 0
In a coordinate system S a vector A is de…ned as
0 0 0
A = A (x );
whereas a vector in S is de…ned as:
A = A(x );
and since
0
x = f (x ):
Therefore we can identify two types of vectors: covariant vectors and
contravariant vectors. These two types of vectors will be discussed later
in chapter three.
The scalar product of a 4-vector with itself is an invariant quantity

and is de…ned by
0 0
A A =A A ; (1.39)
and therefore the norm of a 4-vector is de…ned by
p
A= A A : (1.40)
Velocity 4-Vector
The velocity 4-vector components are de…ned by
dx
U =
d
dx dx x
U1 = Ux = = q =q
d dt 1
2
1
2
c2 c2
y
U2 = q
2
1 c2
z
U3 = q (1.41)
2
1 c2
dx4 icdt ic
U4 = = q =q
d dt 1
2
1
2
c2 c2
The 4-momentum
The Momentum 4-Vector is de…ned by
p = m0 u
where
m0 x
p1 = px = q
2
1 c2
m0 y
p2 = py = q (1.42)
2
1 c2
m0 z
p3 = pz = q
2
1 c2
and
m0 (ic)
p4 = pt = q ; (1.43)
2
1 c2
So we can write the 4-momentum as
iE
p = (p; ) (1.44)
c
where p is the 3-momentum.Maxwell Equations in 4-Vector form
We mentioned earlier that the Maxwell equations of the Electro-
dynamics satisfy Lorentz transformations; therefore it would not be
in con‡ict with special relativity. It will su¢ ce to accommodate the
Maxwell equations into the framework of special relativity to reformu-
late them in terms of 4-vectors. For this purpose we remind the reader
that the four Maxwell equations are
1 @E 4
r H= + J
c @t c
1 @H
r E=
c @t
r:E = 4 (1.45)
r:H = 0
The …rst two equations can be written in 4-vector form as follows
@ 4
F = J (1.46)
@x c
while the other two equations can be written as follows
@ @ @
F + F + F =0 (1.47)
@x @x @x
where F is the Electromagnetic Field Tensor given by
0 1
0 Hz Hy iEx
B Hz 0 Hx iEy C
F =B @ Hy
C (1.48)
Hx 0 iEz A
iEx iEy iEz 0
and J is the 4-current given by
J = (Jx ; Jy ; Jz ; ic ) (1.49)
This current satis…es the equation of continuity
@J
=0 (1.50)
@x
In order to verify that (1.46) and (1.47) are correct we make the
direct substitution
@F 1 @F 2 @F 3 @F 4 4
1
+ 2
+ 3
+ 4
= J
@x @x @x @x c
Substituting for = 1 we …nd
@F 11 @F 12 @F 13 @F 14 4 1
1
+ 2
+ 3
+ 4
= J (1.51)
@x @x @x @x c
and from (1.48) we …nd that
F 11 = 0; F 12 = Hz ; F 13 = Hy ; F 14 = iEx
and from (1.49) we …nd that
J 1 = Jx
Substituting in (1.51) we …nd that
@Hz @Hy @Ex 4
= + Jx
@y @z c @t c
Similarly, substituting for = 2; we …nd that
@Hx @Hz @Ey 4
= + Jy
@z @x c @t c
and for = 3 gives
@Hx @Hy @Ez 4
= + Jz
@y @x c@t c
So that if we multiply each of the above three equations by the respec-
tive unit vector and add them we …nd
1 @E 4
r H= + J
c @t c
which is the …rst of Maxwell’s equations written in (1.45).
Other values of will yield the other two equations and proper substi-
tutions for the values of ; and will yield the other two equations,
this will be left as homework for the students.
Problems
1. An observer on a spacecraft moving at 0:7c relative to the earth
…nds that a car takes 40:0 min to make a trip. How long does
the trip take to the driver of the car?
2. How fast must a spacecraft travel relative to the earth for each
day on the spacecraft to correspond to 2 d on the earth?
3. Show that the relativistic form of Newton’s second law, when F

is parallel to v, is
3=2
d 2
F = m0 1
dt c
4. The mass of a particle is triple its rest mass. What is its speed?
5. Dynamite liberates about 5:4 106 J/kg when it explodes. What

fraction of its total energy content is this?
6. How much work (in MeV) must be done to increase the speed of
an electron from 1:2 108 m/s to 2:4 108 m/s?
p
7. Show that 12 m 2 ; where m = m0 = 1 2 =c2 ; does not equal the
kinetic energy of a particle moving at relativistic speeds.
8. A moving electron collides with a stationary electron and an

electron-positron pair comes into being as a result (a positron
is a positively charged electron). When all four particles have
the same velocity after the collision, the kinetic energy required
for this process is a minimum. Use a relativistic calculation to
show that KEmin = 6m0 c2 ; where m0 is the rest mass of the
electron.
9. Verify that s
1 p2
p = 1+
1 2 =c2 m20 c2
10. Show that the relative velocity of two observers each moving with
velocity c with respect to the ground in the same direction is c.
11. Verify the second equation of Maxwell.
12. Show that the Lorentz transformation in 2d are equivalent to the

rotation of the x t plane by an imaginary angle i :
Chapter 2
The Principles of General Rel-

ativity
Why relativistic Gravity

Newtonian concept of gravity was based on the classical concept of
the independence of space and time and their absoluteness. It was
also based on the invariance provided by Galilean transformation. As
we know Galilean transformation preserves the invariance of the line
element
dr2 = dx2 + dy 2 + dz 2 (2.1)

and also preserves the invariance of Poisson equation
@2 @2 @2
+ + r2 = 4 G (2.2)
@x2 @y 2 @z 2
where is the gravitational potential of a distribution of matter of
density .
In special relativity we have seen that Lorentz transformations is
required to preserve the invariance of physical quantities, and therefore,
to preserve the invariance of the new line element
2
d = c2 dt2 (dx2 + dy 2 + dz 2 ) (2.3)
The Galilean transformations cannot guarantee the constancy of the
velocity of light and furthermore the Newtonian description of gravity
implies the Principle of “Action at a distance” which means that the
gravitational interactions are taking place at in…nite speed, something
31
32 The Principles of General Relativity
that special relativity cannot accept.

We face two alternatives:
1. Try to modify or replace Maxwell’s electrodynamics to …t Galileo

transformation, an approach that proved to be unsuccessful. Or
2. Modify Newtonian gravity.
So, in order to resolve the con‡ict between Newtonian gravity and

special relativity, a new concept of gravity has to be developed that
should be compatible with special relativity. Therefore the best ap-
proach is to modify Newtonian gravity to …t the requirements of special
relativity. A Direct answer is to extend Eqn.(2.2) to read
1 @2
r2 = 4 G (2.4)
c2 @t2
Di¢ culties
If is a scalar …eld then must also be a scalar …eld. On the other
hand the equivalence of matter and energy requires us to recognize any
energy density as a source of gravitational potential. So that is the
00-component of second rank tensor and Eqn. (2.4) should read
1 @ 2 ik
2 2
r2 ik = 4 GTik (2.5)
c @t
But in this case what is the meaning of ik ? The above equation is
a second order equation of motion describing the propagation of a …eld
with the velocity of light in spacetime, so such a …eld can be the EM
…eld itself. therefore one question we come across will be how would the
light be a¤ected by gravitation? The other question that appears here
is: special relativity treats inertial frames that are free. Gravitational
…eld exerts acceleration, i.e. force. How can we bring both together??
Inertial mass and Gravitational mass

After the discovery of the laws of motion by Newton it was noticed
that the mass of a particle has two independent meanings:
The inertial mass, i.e. the resistance of a particle to change its state of
motion obtained from Newton’s second law
Fa
ma =
a
And the gravitational mass obtained from the law of gravitation
Fg
mg =
g
The question is Does ma = mg ?
The Principles of General Relativity 33
Figure 2-1 Di¤rent inertial frames
If both masses are exactly equal then gravitational acceleration can

be directly related to gravitational force.
Einstein believed that the answer is YES. Experimental evidence is
con…rming this equivalence with high degree of accuracy but there is
no theoretical reason why these two masses should be equal.
The Principle of Equivalence

Gravity and Acceleration
1. An observer in a frame at rest in Gravitational …eld experience
acceleration. This he can check by releasing a particle to fall
freely. He will …nd the particle falling toward the center of the
gravity with acceleration.
2. An observer in a freely falling frame feels no gravitational force.

This he can check by releasing a particle to fall freely. He will
…nd the particle at rest with him ‡oating in space. The observer
will not feel the pull of the gravitational …eld.
3. An observer in an accelerated frame feels that he is pulled to the

direction opposite to the direction of acceleration. This resembles
a gravitational force.
4. An inertial observer in a gravity free frame feels no gravitational

force of course.
If we are in gravity free space in an accelerated capsule with accel-

eration g then we will feel that we have the same weight as on earth.
From 1-4 above we deduce that

Figure 2-2 Equivalence of gravity and acceleration
1. Locally Gravity is equivalent to Acceleration.

2. Gravity is related to spacetime properties.
This implies that gravity is a Geometrical property.

Note that we say “locally” because acceleration can be uniform while
gravity being a central …eld of force is non-uniform.
Einstein recognized that Gravity should be equivalent to Acceleration
and, therefore he postulated that
“Gravity is locally equivalent to acceleration”. This was called the
“Principle of Equivalence”.
General Covariance
As we have seen in chapter 1 the invariance of physical quantities in
special relativity under coordinate transformations of inertial frames
is necessary to maintain the universality of physics. Therefore, if we
will generalize the universality of physics to cover non-inertial frames
therefore we should look for physical quantities that preserve itself un-
der some general coordinate transformations. Also it is very important
that the form of the physical laws be preserved under such a general
coordinate transformations.
To guarantee these requirements Einstein postulated that
“The laws of physics should be invariant under general coordinate trans-
formations”. This was called The “Principle of General Covariance”.
Physical implications
The requirement that the laws of physics be the same in all inertial and
none-inertial frames, with the fact that spacetime is now to be realized
as a continuum, implies that the physical quantities get expressed in
terms of objects that allows for the integrity of the spacetime. From
our experience with special relativity we can infer that these objects
are to be described in 4 dimensional space no less. On the other hand
taking into consideration that now space and time might be curved
suggests that four vectors might not be su¢ cient to cover the whole
spacetime manifold. Four vectors can only be realized in a local coor-
dinate system.
The Line Element

Acceleration describes one non-linear relationship between space and
time. If gravity is to be taken as a property of the spacetime then
this means that it will be described in geometrical terms instead of the
physical terms. This of course will imply a deep relationship between
gravity and geometry.
However, non-linearity of the spacetime means that there will be cross-
terms in the line element and will require that physical quantities be
described in terms of tensors because vectors will not be su¢ cient for
its description.
Let us …rst try to generalize the Minkowski space into a more general
form that will be more suitable to accommodate the formulation of
gravitational physics. The line element can be written in the general
form
ds2 = g dx dx (2.6)
where we adopt the notations
x = (x0 ; x1 ; x2 ; x3 ) = (ct; x; y; z) g = g (x )
is the metric tensor.
Obviously this means that
ds2 = g00 dx0 dx0 + g01 dx0 dx1 + g02 dx0 dx2 + g14 dx0 dx3
+g10 dx1 dx0 + g11 dx1 dx1 + g12 dx1 dx2 + g13 dx1 dx3
This general form of the spacetime metric can describe a curved

spacetime. As we see these are 16 terms, but it may be reduced to 4
terms under certain conditions of spacetime symmetries. These sym-
metries imply, of course, some physical properties.
The disappearance of the cross terms will mean that the spacetime
is axi-symmetric. Then the metric will take the diagonal form
ds2 = g00 c2 dt2 + g11 dx2 + g22 dy 2 + g33 dz 2

In matrix form this can be written as
0 10 2 2 1
g00 0 0 0 c dt
B 0 g 0 0 C B dx2 C
ds2 = B@ 0
11 CB C (2.7)
0 g22 0 A @ dy 2 A
0 0 0 g33 dz 2
So that if gik = ik = diag( 1; 1; 1; +1) then the spacetime is
Minkowskian.
Note about conventions
It should be understood that the coordinates are normally written as
x1 = x; x2 = y; x3 = z; x4 = ict
But it is also possible to use
x0 = ct; x1 = x; x2 = y; x3 = z;
where ix0 = x4 . In every case we have 4 coordinates.
Examples of line elements

The line element in Minkowski spacetime is given in (1.33), this is the
‡at 4-d spacetime of special relativity.
The line element of a 2d ‡at space (plane) given in Cartezian coor-
denates (x; y) is
ds2 = dx2 + dy 2 (2.8)
Here g11 = g22 = 1:This can be easily convert it into polar coordinates
(r, ) as
ds2 = dr2 + r2 d 2 (2.9)
where now g11 = 1 and g22 = r2 :
The line element of a 2d-sphere of radius a (surface of a sphere in
2d) de…ned by and is given by
2
ds2 = a2 (d + sin2 d 2 ) (2.10)
we will use these line elements later.
Geometrical realization
A spacetime can be one of three types:
Flat: in which Euclidean geometry applies, e.g. the sum of the angels
of a triangle equals 180 degrees, the ratio of the circumference of a
circle to its radius is 2 , etc. Minkowski spacetime is an example of
‡at spacetime.
Figure 2-3 Di¤erent Geometries
Positively curved: in which the laws of non-Euclidean (Riemann)

geometry applies, the sum of the angles of a triangle is more than 180
degrees, but the ratio of the circumference of a circle to its diameter is
less than . An example of this is the surface of a sphere.
Negatively curved: in which the laws of non-Euclidean (Lobachevskian)
geometry applies, the sum of the angles of a triangle is less than 180
degrees, but the ratio of the circumference of a circle to its radius is
more than . An example of this is the surface of a saddle.
Note that some surfaces may appear to be spatially curved but in
reality it may be ‡at since the Euclidean laws apply to it. An example
for this is the surface of a cylinder.
Figure 2-4 Euclidean and Non-Euclidean Geometries
Also we should note that when we say curved spacetime we do not

necessarily mean that the pace only is curved but the time too. In fact
in a weak gravitational …eld the time is the more likely to be curved.
The curvature of time can be realized through the time interval being
longer than the corresponding interval in ‡at space, i.e., time running
slower.
Examples
Accelerated Observers
Consider the Minkowski spacetime with the standard line element
ds2 = c2 dt2 dx2 dy 2 dz 2 (2.11)

Let a particle be moving with acceleration g in this spacetime. Then
the equation of motion will be
0 1
d @ u A = g; dx
q u=
dt 1 u
2 dt
c2
If we make the coordinate transformations for a constant g
c2 gt0 gt0
x= cosh 1 + x0 cosh
g c c
0 0
c gt x gt0
y = y 0 ; z = z 0 ; t = sinh + sinh (2.12)
g c c c
we obtain
2
2 2 gx0
ds = c 1 + 2 dt02 dx02 dy 02 dz 02
c
How can we interpret this line element? Clearly this line-element
is describing an accelerated coordinate system in which time is bent
where
gx0 0
dt = 1 + 2 dt (2.13)
c
We notice that the origin of (x0 ; y 0 ; z 0 ) is given by
c2 gt0 c gt0
x= cosh 1 ; y = 0; z = 0; t = sinh
g c g c
This shows that the primed coordinate system is an accelerated sys-
tem with uniform acceleration g with respect to the un-primed system.
We will come back to this metric later in chapter 4.
Rotating Coordinates
Described in cylindrical coordinates, a ‡at Minkowski space time is
described by the line element
ds2 = c2 dt2 dr2 r2 d2 dz 2
0
A rotating coordinate system S is described by the coordinates
0 0 0 0
t = t; r = r ; = !t; z = z (2.14)
0
Accordingly the line element in S will be
0 0 0 0 0 0 0 0 02 0
ds 2 = (c2 ! 2 r 2 )dt 2 2!r 2 d dt dr 2 r 2d dz 2
Dropping the primes we then have for the invariant spacetime interval
in a rotating frame
2
ds2 = (c2 ! 2 r2 )dt2 2!r2 d dt dr2 r2 d dz 2
But since
2
ds2 = g dx dx = g00 dx0 + 2g0i dx0 dxi + gik dxi dxk
Therefore we conclude that in a rotating frame
0 2 2 !r2
1
(1 !c2r ) 0 c
0
B 0 1 0 0 C
g =B @ !r 2
2
C
A (2.15)
c
0 r 0
0 0 0 1
The fact that the frame is rotating is shown in the terms g02 and g20
and also in g00 . This expresses the fact that a rotating frame is not
Minkowskian. Let us investigate time intervals and distances in this
frame of reference. The interval is given by the di¤erence between the
coordinate time at two points on …xed r, ; and z. This is
ds2 = g00 c2 dt2 = (c2 ! 2 r2 )dt2 = c2 d 2
(2.16)
therefore the proper invariant time is given by
r
p !2 r2
d = g00 dt = 1 dt (2.17)
c2
!2 r2
Since c2
< 1 then d < dt:
Problems
1. The usual formula for the period T of a simple pendulum of length
l is s
l
T =2
g
where g is the acceleration due to gravity. Denoting the iner-
tial mass of the pendulum bob by mi and its gravitational mass
by mg , derive an alternative expression for T in terms of these
masses, the radius R of the Earth and its mass Mg .
2. By employing spherical polar coordinates show that the circum-
ference C of a circle of radius R inscribed on a sphere S 2 (as in
Fig. 2-4) obeys the inequality C < 2 R.
Chapter 3
Tensors and Tensor Calculus
Tensors are non-directional mathematical objects representing the dis-

tribution of certain …elds over a region in space. For this reason it
is sometimes called tensor densities. The consideration of tensors in
general relativity is necessary for two reasons:
1. Gravitational …eld is best described as a curved (non-Euclidean)

spacetime.
2. The formulations of physical laws should be independent of the
frame of reference. We will …rst consider vectors and then tensors.
Coordinate transformations
Consider two coordinate systems xi and x0i . We will assume that in
the overlapping region where the two coordinates overlap the functions
0
of xi and x i are twice di¤erentiable.
0 0
x i = x i (xk )
We can therefore write the following relations

0
0i@x i k
dx = dx (3.1)
@xk
0i
Clearly @x
@xk
plays the role of coupling coe¢ cients between the two co-
ordinate systems.
@xk 0 i
dxk = dx (3.2)
@x0 i
41
42 Tensors and Tensor Calculus
So that
0
@x i @xk i
= l (3.3)
@xk @x0 l
These transformations are valid if
0
@x i
jj k jj 6= 0 (3.4)
@x
This means that the transformation matrix should be non-singular.
Scalars
Quantities that do not change their magnitude under coordinate trans-
formations are called Scalars.
A scalar …eld is a function of spacetime coordinates that does not
change under any transformation of the type discussed above. Thus
0 0 0
[xi ] = [xi (x k )] = (x k ) (3.5)
The quantity ds2 , which is the square of the distance between two
neighboring points, is unchanged under coordinate transformations1 .
Hence
0 0 0
gik dxi dxk = gik dx i dx k (3.6)
Using (3.1)-(3.3) we …nd that
0 @xl @xm
gik = glm (3.7)
@x0 i @x0 k
This equation relates the metric tensor of two spacetimes. In a ‡at
spacetime det jgj = 1.
Home Work (4): Using (3.7) consider the transformation of the

system xi = (x; y) described in the Cartesian coordinates, the line ele-
0
ment in (2.8) into the system x i = (r; ) described by polar coordinates
0
as in (2.9). Find the components of gik .
Vectors
These are quantities that are completely speci…ed by their magnitude
and direction. Having the direction in their de…nition does not allow
preserving them under coordinate transformation. In fact there are two
types of vectors under coordinate transformation
1
In some books this requirement is taken as if it is an axiom.
Tensors and Tensor Calculus 43
Contravariant vector in curved space:

Contravariant vectors are actually tangent vectors to curves in space-
time. Let us suppose that the parametric representation of a curve
in a spacetime is given by functions xi ( ) where is a monotonic
parameter along . Then the four quantities
dxi
Ai (P ) jP
d
de…nes the components of the tangent vector to at the point P .
Conversely, given a set of four quantities Ai at P we can construct a
curve P by choosing = 0 at P and de…ning
xi ( ) = xiP + Ai (3.8)
in a small neighborhood of P . Then Ai will be the components of a
tangent vector to & P at P .
In another coordinate system x0i we can de…ne A0i (P )in the same way
as before. The relation between Ai and A0i is given by
0 0 0
0i@x i k @x i 1 @x i 2
A = A = A + A + ::::::: (3.9)
@xk @x1 @x2
Any set of components that transforms like (3.9) constitutes a Con-
travariant vector.
Covariant vectors
Let '(xi )be a scalar function of xi . The equations '(xi )= C, for various
values of C, describes a set of 3-dimensional surfaces in the spacetime.
d'
The normal to the surface at a point P is dx i so we de…ne:
@'(xk )
Bi = (3.10)
dxi
In the primed coordinate system we have
0 @'
Bk = : (3.11)
@x0 k
The relationships between the components of this normal vector in
coordinate systems xi and x0i are given by
0 @' @' @xi @xi

Bk = = = Bi (3.12)
@x0 k @xi @x0 k @x0 k
Any vector with components transforming in this way is called co-
variant vector. Clearly
0
@x k 0
Bi = B (3.13)
@xi k
The velocity of a particle rotating in a circle has two components

tangential and radial. The tangential velocity can be expressed as
a contravariant vector whereas the radial component is a covariant
vector.
The product Ai B i is a scalar. The proof is
0
0i @x i k @xl l k
A Bi0 = A Bl = k A Bl = Al Bl (3.14)
@xk @x0 i
Example (1)
Given the vector V being described in two dimensional Cartesian
coordinate with the components (V x ; V y ); express V in polar coordi-
nates with the components (V r ; V ):
0
Solution: We have two coordinate systems x = (x; y) and x =
(r; ) with r = (x2 + y 2 )1=2 and = tan 1 ( xy );
@r x @r y
Vr = V + V
@x @y
@ x @ y
V = V + V
@x @y
@r x @r y @ 1 @ 1
Now @x = r
= cos ; @y
= r
= sin and @x
= r sin
; @y
= r cos
:
Therefore
V r = cos V x + sin V y
1 1
V = Vx+ Vy
r sin r cos
and the covariant components will be
@x @y
Vr = Vx + Vy
@r @r
= cos Vx + sin Vy
and
@x @y
V = Vx + Vy
@ @
= r sin Vx + r cos Vy
Now it is clear that
V V = V r Vr + V V = V x Vx + V y Vy
Tensors
We learned that a vector in any n-dimensional space is an object that
represent the resultant of its components. A vector V has a magnitude
and de…nite direction. Vectors are used to describe linear quantities
which have their e¤ect is a well de…ned direction, like the electric …eld
for example. However, Tensors are objects with di¤erent character,
they do not have de…nite directions, instead a tensor like Tij is an
object de…ned by its sole components. Tensors describe quantities with
character of a distribution over a surface or a volume of any dimension.
An example of a tensor is the torsion tensor.
Contravariant tensors
A quantity T i1 i2 i3 :::in transforming like:
0 0 0 0
@x i1 @x i2 @x i3 @x in j1 j2 j3 ::::jn
0 i i i :::i
T =
1 2 3 n
::: T : (3.15)
@xj1 @xj2 @xj3 @xjn
is called a Contravariant tensor of rank n.
Covariant tensor
A quantity transforming like
0 @xj1 @xj2 @xj3 @xjn
Ti1 i2 i3 :::in = ::: Tj j j ::::j : (3.16)
@x0 i1 @x0 i2 @x0 i3 @x0 in 1 2 3 n
is called a covariant tensor of rank n.
Mixed tensors
A quantity transforming like
0 0
0 i i i :::i
n
@x i1 @x i2 @xl1 @xl2
T 1 2 3
=
j1 j2 j3 :::jm ::::
@xk1 @xk2 @x0 j1 @x0 j2
0i
@x n @xlm
::::::::::::: kn 0 jm Tlk11l2k:::l
2 :::kn
(3.17)
@x @x m
is called mixed tensor of rank(n + m).

Scalars are tensors of rank 0 and vectors are tensors of rank 1.
Note that:
1. If we multiply a scalar with a tensor then there will be no change

in rank.
2. A scalar is zero rank tensor, while a vector is a …rst rank tensor.
3. Conventionally a tensor is second rank or higher.

4. A scalar is usually a complex number [c-number].
5. A vector is a column or row matrix.
6. A tensor of rank m has mn components in n-dimensional space.
0 1
T 00 T 01 T 02 T 03
B T 10 T 11 T 12 T 13 C
T ij =B
@ T 20
C (3.18)
T 21 T 22 T 23 A
T 30 T 31 T 32 T 33
(T 00 is the time-time component, T 11 , T 22 , T 33 are spatial compo-
nents, others are mixed components)
The Metric Tensor

The metric tensor is a second rank tensor which describes the distrib-
ution of the geometrical structures in the space. This tensor appears
when we express the line element between two points in space separated
by an in…nitesimal distance dx
ds2 = gik dxi dxk (3.19)
Note that here we are using the Einstein summation convention. The
metric tensor gik transform like a 2nd -rank tensor.
@xk @xl
gij0 = gkl (3.20)
@x0i @x0j
We now introduce the inverse of the matrix gik : Assuming that
g = det jjgik jj 6= 0
Because of the quadratic form (3.19) has a signature of 2 therefore
g < 0. Writing the inverse matrix as g jk , we get
k
gij g jk = i: (3.21)
The determinant g does not transform like a scalar. We have
0 @xk @xl
g 0 = det jjgij jj = det jj gkl jj
@x0i @x0j
@xk @xl
= det jj 0 i jj: det jjgkl jj: det jj 0 j jj
@x @x
0 2
= [Jfx ! x g] :g;
where J is the Jacobian of the transformation
x ! x0
Since
dx1 dx2 dx3 dx4

transform as
0 0 0 0 0
dx1 dx2 dx3 dx4 = Jfx ! x g:dx 1 dx 2 dx 3 dx 4
It follows that
p
gdx1 dx2 dx3 dx4
is invariant.
Example on trasforming use of coordinate systems
We are given the line element in (2.8) describing a 2d space in
Cartesian coordinates. We here show how the componets of the metric
tensor gik will change if we change the coordinates into a polar system.
0
We have x = (x; y) and x = (r; ): The relations between the
coordinates are given by
x = r cos ; and y = r sin (3.22)
We also have the transformation relation

@xk @xl
gij0 = gkl
@x0i @x0j
Accordingly we have to calculate g11 ; g12 ; g21; and g22 : We have
0 @x1 @x1 @x2 @x1 @x1 @x2 @x2 @x2

g11 = g11 + 22g21 + g12 + g22 (3.23)
@x0 1 @x01 @x0 1 @x01 @x0 1 @x01 @x0 1 @x01
since g11 = g22 = 1 and g12 = g21 = 0; then
0
g11 = cos2 + sin2 =1
Similarly we …nd that

0
g22 = r2 cos2 + r2 sin2 = r2
So that
0 2
ds 2 = dr2 + r2 d (3.24)
Home Work (5):
k k
1. Given gij g ik = i, show that i transforms like a 2nd rank mixed
tensor.
2. Then if gij is a 2nd rank covariant tensor, show that g jk transform

like a second rank Contravariant tensor.
Raising and lowering indices

The metric tensor can be used for rising and lowering indices and there-
fore converting Contravariant index into a covariant one and vice versa.
For example
gik Ak = Ai (3.25)
Also
g ik Ai = Ak (3.26)
Magnitude of a Vector
We de…ne the magnitude of a vector as
A2 = gik Ai Ak (3.27)
= g00 A0 A0 + g01 A0 A1 + ::::::::: (3.28)
The angle between two vectors is given by
AB cos = Ai B i = scalar = A0 B 0 + A1 B 1 + A2 B 2 + A3 B 3
If
0 1
>0 then Ai is called time-like
gik Ai Ak = @ < 0 then Ai is called space-like A (3.29)
=0 then Ai is called null vector
Note that
1. The magnitude is a scalar and therefore is independent of the

coordinate system. If Ai Ai = 0 then the components Ai are
belonging to a null-vector. Therefore if the sum of the squares of
all the components is zero the vector is called null. This means
that for a null vector we have
(A0 )2 = (A1 )2 + (A2 )2 + (A3 )2
2. Because the signature of the line element is 2, a null vector does

not necessarily imply a zero vector.
Home Work (6) Show that the product A A is invariant under

coordinate transformations.
Symmetric and antisymmetric tensors

If Sik = Ski , then Sik is a symmetric tensor.
If Aik = Aki then Aik is antisymmetric tensor.
Any second rank tensor Tik can be divided into symmetric and
antisymmetric parts as follows
1 1
Tik = (Tik + Tki ) + (Tik Tki ) = T(ik) + T[ik] (3.30)
2 2
where
1
T[ik] = (Tik Tki )
2
is the antisymmetric part of the tensor. A totally antisymmetric tensor
can be constructed accordingly.
Generally for any order
1 X
T(i1 ::::in ) = Tip1::::::: ipn (3.31)
n! p
and
1 X
T[i1 ::::in ] = ( 1)p Tip1::::::: ipn (3.32)
n! p
For example the electromagnetic tensor is an antisymmetric tensor of

second rank.
@Ak @Ai
Fik = ;
@xi @xk
where Ak are the components of the electromagnetic vector potential.
The EM …eld tensor is given by
0 1
0 Hz Hy iEx
B Hz 0 Hx iEy C
F ik =B
@ Hy
C (3.33)
Hx 0 iEz A
iEx iEy iEz 0
Home Work (7): Verify that the four Maxwell equations can be
written in the forms
1
F[ij;k] = 0 and that F;kik = j i
c
Levi-Civita tensor:
Consider the antisymmetric symbol
0 1
+1; for permutation of (ijkl)
even
[ijkl] = @ 1; for permutation of (ijkl) A
odd
0; for repetition of indices
and de…ne the antisymmetric tensor
ijkl = ( g)1=2 [ijkl]

We show that this object is a tensor as follows
@xi @xj @xk @xl

ijkl
@x0m @x0n @x0p @x0q
@xi @xj @xk @xl
= ( g)1=2 [ijkl]: 0m 0n 0p 0q
@x @x @x @x
= ( g) [ijkl]Jfx ! x0 g
1=2
= ( g 0 )1=2 [ijkl]
= 0ijkl (3.34)
The Volume and surface elements

The 4-volume element is given by
1 ijkl
d = dxi dxj dxk dxl
4!
The 3-volume (hypersurface) element is given by
l 1 ijkl
d = dxi dxj dxk
3!
The 2-surface element is given by
ij 1 ijkl
d = dxk dxl
2!
Parallel Transport
When we move a vector normal to a surface parallel to itself in Euclid-
ean space then we can be sure that the vector will be kept normal at
all times with no change in its direction or magnitude. But when we
move a vector A normal to a surface parallel to itself (that is keep-

ing it normal to the surface at all points) from point P (xk ) to a point
Q(xk +dxk ) in non-Euclidean (curved) space, we can never be sure that
that the magnitude or the direction of the vector will not change.
Figure 3-1 Moving a tangent vector on the surface of a sphere
In fact the change in the respective components of the vector will

be proportional to the original component and to the displacement in
position from p to q (see the …gure).We may express this change in
general as
l
Bi = ik Bl xk
i
where jk are functions of spacetime and are called Christofell Symbols.
Figure 3-2 Parallel Transport of a Vector. In this …gure we have used

the symbol v instead of B.
Covariant di¤erentiation
@Bi
How would the quantity @xk
transform?
0
@Bk @ @xi
= Bi
@x0 m @x0 m @x0k
0
@Bk @xi @Bi @ 2 xi
0m = 0k 0m + Bi
@x @x @x @x0 m @x0 k
0
@Bk @xi @Bi @xn @ 2 xi
= + Bi
@x0 m @x0 k @xn @x0 m @x0 m @x0 k
0
@Bk @xi @xn @Bi @ 2 xi
0m = 0k 0m + Bi (3.35)
@x @x @x @xn @x0 m @x0 k
Because of the second term this quantity does not transform like a
tensor if the coordinates are second di¤erentiable. Such a case happens
when we deal with accelerated systems; otherwise if the coordinate
system is inertial (v =constant), then the coordinates will not be second
di¤erentiable and the derivative of a vector will transform like a tensor.
As we have seen above a translation of a vector a long a curve will not
guarantee its invariance unless we require parallel transport.
The di¤erence between the continuous and dotted vector in the …gure
above is given by
Bi (xk + xk ) Bi (xk ) k Bi k
Bi (xk + xk ) Bi (xk ) + Bi = x x
xk xk
@Bi l
= ik Bl xk
@xk
Therefore we may rede…ne the derivative of a vector by
@Bi l l
Bi;k = ik Bl = Bi;k ik Bl (3.36)
@xk
This is called the covariant derivative of a vector.
If Bi;k would transform like a tensor (of 2nd rank) then l
ik transform
like
0i @x0i @xn @xp m @ 2 xp @x0i

lk = np + (3.37)
@xm @x0k @x0l @x0k @x0i @xp
Home Work (8)(a) Prove the above expression. (b) Using (3.13)
Show that Bi;k transforms like a 2nd rank tensor.
The second term on right hand side shows that does not transform
like a tensor.
A scalar like Ai Ai is invariant under coordinate transformation, so
0 0
Ai A i ;k
= Ai A i (3.38)
;k
From the de…nition of the covariant derivative above and this relation-
ship we can prove that
@Ai
Ai;k = + i
lk A
l
(3.39)
@xk
The rule for a covariant derivative of a tensor of any rank can be

easily obtained; we introduce a (+ ) term for each Contravariant index
and a ( ) term for each covariant index. Examples:
@Tji
i
Tj;k = + i
pk Tjp p i
jk Tp (3.40)
@xk
and
@gik p p
gik;l = ( il ) gpk ( kl ) gip (3.41)
@xl
We will see later that corresponds to the gravitational accelera-

tion. Geometrically it is a measure of the curvature of the spacetime.
Home Work (9): (a) Prove (3.37), (b) Prove (3.38)
Problems
1. Which of the following expressions are invalid with respect to
the summation convention? Simplify those expressions that are
valid. (a) Aij B jk Ajl , (b) gik g ik , (c) Rik gik , (d) eiklm eiklm , T ik
glk (e).
2. Show that the product of two tensors is a tensor.
3. Aik is a tensor such that the matrix (Aik ) is non-singular. Show
that the components of the inverse matrix transform as a tensor.
4. Use two dimensional polar coordinates (r, ) on Euclidean plane.
Let Ar and A be the radial and transverse components of a vector
A at a typical point chosen with respect to locally Cartesian axes,
with direction coinciding with = constant and r =constant, re-
spectively. Show that a parallel transport of the vector at a point
P to a neighboring point Q(r + r; + ) gives the component
of the vector Q as Ar + A ; A Ar .
5. Show by direct enumeration that the number of algebraically in-
dependent components of Riklm is 20.
6. Prove from …rst principles that Bi; nk Bi; kn = Ri mkl Bm .
7. Calculate tangent vectors at typical points on curves in space-

time given by the following relations (x1 = x; x2 = y; x3 = z
and x0 = ct) and specify the type. (a) x = ct0 cos (t=t0 ) ; y =
ct0 sin (t=t0 ) ; z = ct; t0 =constant. (b) x = 0; y = 0; z 2
c2 t2 = 0. (c) x = ct cos (t=t0 ) ; y = ct sin (t=t0 ) ; z =constant,
t0 =constant.
Chapter 4
Spacetime Curvature
Spacetime Geometry
In chapter 2 we have shown that gravity is a geometrical property of

the spacetime rather than being an intrinsic property of mass. The
‡at, Euclidean, spacetime is geometry describing a gravity-free state.
A non-Euclidean (curved) spacetime describes gravity or gravitational
…eld. It is inaccurate to talk a bout spacetime wrapped with gravity
because gravity itself is the spacetime. This is the reason why gravity
cannot be eliminated, unlike the case of electric and magnetic …elds,
which can be eliminated by shielding.
In ‡at spacetime all of the Euclid theorems on geometry are valid,
namely the sum of the three angles of a triangle is 180 degrees, the
ratio of the circumference of a circle to its diameter is and the sum of
the four angles of a square is 360 degrees. . . etc.
In curved geometry the Euclidian theorems do not apply. For example
the sum of the angles of a triangles can be more than 180 degrees on
a positively surface like that of a sphere. Also on a curved surface of a
sphere the ratio of the circumference of a circle to its diameter is less
than . However not all curved-looking surfaces are necessarily curved,
for example the surface of a cylinder is ‡at and all Euclid’s theorems
apply perfectly well on it.
55
56 Spacetime Curvature
The Gaussian curvature

This is a quantity which is de…ned for 2-dimensional surface in 3-
dimentional space, in the following way. Consider a two dimensional
surface and visualize the intersection of a plane with this surface in
some region. At a point P on the line of intersection a tangent vector
of v1 may be drawn, and in the direction of v1 the intersection of the
surface and the plane de…nes, to the lowest order, a circle, with radius
say r1 . We then say that the curvature of the surface at P in the direc-
tion of the tangent vector v1 is k1 = r11 : In the direction perpendicular
to v1 the same is true; in the direction of a tangent vector v2 the inter-
section de…nes a circle with radius r2 ; say, and the curvature k2 = r12 .
The Gaussian curvature K at P is then
1
K = k1 k2 = (4.1)
r 1 r2
Gauss showed that K does not depend on the choice of the vectors
1, 2, or equivalently on the planes of intersection. In this sense K is
an invariant of the surface at the point considered. It may, however,
have a positive or negative sign. Consider, for example, a sphere S 2 of
radius a. Because S 2 is a homogeneous space (like S 3 is) the curvature
is the same everywhere, so that at any point on the sphere both k1 and
k2 are equal to 1=a and
1
S2 : K = >0
a2
the Gaussian curvature is positive. This is because at any point on

the sphere the centres of the two circles (de…ning the curvatures in two
perpendicular directions) are on the same side of the surface (‘inside’
it). In contrast, consider a point on the surface of a hyperboloid. It
is clear that if the plane of intersection with this surface is (zy) the
centre of the circle of intersection is outside the surface, but the circle
resulting from the intersection with the (xz) plane is inside it. Since
the centres are on opposite sides of the surface the Gaussian curvature
is negative,
hyperploid: K < 0
A ‡at surface (plane), of course, has K = 0. In these examples we

have considered surfaces (curved or ‡at) embedded in a ‡at Euclidean
space E3. In the context of cosmology we are considering the Gaussian
curvature K of a 3-dimensional space, not a 2-dimensional one, but
actually K, or a generalization of it, may be de…ned for a space of any
dimension.
Spacetime Curvature 57
Riemannian geometry
This is a non-Euclidean geometry describing spaces with positive cur-
vature, e.g. the surface of a sphere. The geometrical assumptions are
i i
(a) jk = kj
(b)gij;k = 0 (4.2)
In 4d space, ikl have 4 4 4 = 64 components. But under
condition (a) the number of independent components of ikj will be
reduced under to [(4 4) 6] 4 = 40.
The second condition implies
p p
gik;l = gik;l ( il ) gpk ( kl ) gip =0
gik;l = kjil + ijkl ; (4.3)
where
p
kjil =( il ) gpk
By cyclic interchange of the indices we get two more equations:
ljki + kjli = kl;i and ijlk + ljik = li;k

i
Using the symmetry condition on kj we get
2 ijkl = gik;l + gli;k gkl;i

Rising the index i we get
i 1 @gml @gmk @gkl

kl = g im k
+ l
(4.4)
2 @x @x @xm
Thus the Riemannian a¢ ne connections are completely speci…ed by
the metric of the spacetime and its …rst derivative.
i
Home Work (11): Calculate all the elements jk for the line
element
2
ds2 = dr2 + r2 d
The locally Inertial Coordinate sys-

tem
The symmetry condition in (4.2) enables us to choose special coordinate
in which the Christofell symbols are all zero at a given point. Suppose
we start with
m
np 6= 0 (4.5)
in the coordinate system ( xi ) at a point P . Let the coordinates at P
be given by xiP , now de…ne a new coordinate in the neighborhood of P
by
1
x0i = 0i
kl P (xk xkP )(xl xlP )
2
Then at P we have
@x0i i @ 2 x0i
x0P i = 0; jP = k; jP + 0i
kl P =0
@xk @xk @xl
Then from
0i @x0i @xn @xp m @ 2 xp @x0i

lk = np +
@xm @x0k @x0l @x0k @x0i @xp
we obtain
0i
lk =0 (4.6)
This shows that the primed system we have chosen is locally ‡at. Notice
that this means that we have absorbed the variations coming from the
Christofell symbols into the derivatives.
Further by linear transformations (e.g., Vierbein transformations)
we can arrange to have a coordinate system with
i
gik = ik = diag(1; 1; 1; 1); kl =0 (4.7)
at any chosen point P . Such a coordinate system is called Locally
Inertial Coordinate system.
Geodesics
What do we mean by a straight line in Euclidean space? We know that
it is a line with unchanging direction and shortest distance.
Unchanging Direction
Let be a curve given by real parameters so that a typical point on
has coordinates xi ( ). The tangent vector to the curve at P has the
components
dxi
Ui = (4.8)
d
Unchanging direction means that as we move along the curve the
rate of change of U is zero. i.e.
dU i
Ui = + i
kl U
k
xl = 0
d
in terms of the components of U this becomes
d2 xi i dx dxl
k
+ kl =0 (4.9)
d 2 d d
A line de…ned by this equation in a Riemannian space is called Geo-
desic.
Equation of geodesics
We would like now to …nd the trajectories of particles in curved space-
time. This means that we need to …nd the equation of the geodesics.
Now multiply (4.9) by
dxm
2gim
d
and use (4.4)
d2 xl dxm dxk dxl dxm

2gim + fgmk;l + gml;k glk;m g =0
d 2 d d d d
Using the symmetry and antisymmetry properties of k and m we
get
d2 xl dxm dxk dxl dxm

2gim + gmk;l =0
d 2 d d d d
Using gik = gki we get the …rst integral as
dxi dxm
gim = constant (4.10)
d d
The parameter can be replace with the spacetime interval s and
the constant can be normalized to be 1 (time-like geodesic) or 1
(space-like geodesic) or is zero then it is called null geodesic. The
null geodesic is the path of light. So that in the physical Riemannian
spacetime the equation of motion is given by
d2 xi k
i dx dx
l
+ kl =0
ds2 ds ds
dxi dxm
gim =1 (4.11)
d d
Shortest Distance
It is not di¢ cult to prove that (4.9) describes the shortest distance in a
Riemannian space. The distance between two points P1 and P2 along
is given by
Z1 1=2
dxi dxk
s(P1 ; P2 j ) = gik d (4.12)
d d
0
If we have to maintain s(P 1; P 2j ) at minimum then by the varia-

tional principle we have
s(P1 ; P2 j ) = 0 (4.13)
Using the procedure of the Euler-Lagrange equation we obtain an

equation of motion identical with the …rst equation in (4.11). Details
are shown in the box below.
Z1 1=2
dxi dxk
S(P1 ; P2 ) = gik d if S(P1 ; P2 ) = 0; then S is minimum
d d
2
We have to solve
2 3
Z1 1=2 Z
4 dxi dxk 5 = 0 or
gik Ld = 0: (4.14)
d d
2
This is a standard variational calculus exercise, the result is

1=2
d @L @L dxi dxk
( ) = 0 where L = gik : (4.15)
d @ x: i @xi d d
:i
where x = @xi =@ : It is easy to see that
d 1 dxk 1 @gmn 1 @xm @xn

gik : : =0
d L d 2 @xi L @ @
but ds = Ld , so
d2 xi i dx
k
dxl
+ kl =0 (4.16)
ds2 ds ds
This equation de…nes the path of particles in curved spacetime. It

is sometimes called the equation of motion in curved spacetime.
Examples
Example (1): Two dimensional Euclidean plane in Cartesian coordi-
nates. Here we have
ds2 = dx2 + dy 2
i
g11 = g22 = 1; g12 = g21 = 0; so all km = 0 and the geodesic
equations become
d2 x d2 y
= =0
ds2 ds2
solutions of this is x = as + c1 and y = bs + c2 , where a; b; c1 and
c2 are constants. Therefore, the geodesic is
y = Cx + D
which is a straight line.

Example (2): Consider the same plane but in polar coordinates:
we have
2
ds2 = dr2 + r2 d (4.17)
then Christofell symbols are
1 1 2 2 2 2 1 1
11 = 12 = 22 = 11 = 0; 21 = 12 = ; 22 = r
r
so the geodesic equations are
2
d2 r d
r =0 (4.18)
ds2 ds
and
d2 2 dr d
2
+ =0 (4.19)
ds r ds ds
We must verify that these equations describe straight lines. If dds = 0;
then =constant, and from (4.18) r = as + b which is a straight line.
0 0
Otherwise, if dds = 6= 0 then we divide (4.19) by to get
0
1d 2 dr
0 + =0
ds r ds
Integrating we obtain
0
ln + ln r2 = const.
therefore
0 d
r2 = r2 = h = const. (4.20)
ds
Dividing the equation for the metric (4.17) by ds2 and using (4.20)
gives
2 2
dr d
1 = + r2
ds ds
2
dr h2
= +
ds r2
hence
dr 1 2 1l2
= r h2 (4.21)
ds r
accordingly (4.17) gives
d h
= 2 (4.22)
ds r
dividing (4.21) and (4.22) we obtain
d h d 1 h
= = cos
dr r (r2 h2 )1=2 dr r
which has the solution
1 h
0 = cos
r
therefore
h
= cos( 0) (4.23)
r
which is the equation of a straight line in polar coordinates (see Fig.
4-1.)
Figure 4-1 Geodesic in a plane in polar coordinates
Example (3): A geodesic on the surface of a sphere S 2 : The metric

is given by
ds2 = a2 d 2 + a2 sin2 d 2 (4.24)
the non-zero connections are
2 2 1
21 = 12 = cot ; 22 = sin cos (4.25)
the geodesic equation for the coordinate is

d2 d d
+ 2 cot =0
ds2 ds ds
d 0
de…ning ds
= and integrating we obtain
0
sin2 = const. = h (4.26)
From (4.24) we have
02 02
1 = a2 + a2 sin2
substituting (4.26) we get
1=2
0 sin2 a2 h2
= (4.27)
a sin
Dividing (4.26) by (4.27) we obtain
0
d h csc2
0 = = 1=2
;
d 1
h2 csc2
a2
on integration this becomes

0 1
h cot
0 = cos 1 @ q A
1
a2
h2
therefore
h cot
cos ( 0) +q =0
1
a2
h2
this means that
A cos + B sin + C cot = 0
If we multiply by r sin then we obtain the form
x+ y+ z =0
with
x = A sin cos ; y = B sin sin ; z = C cos
It thus represents a plane section of the sphere through the origin, i.e.
a great circle. This is what we set out to prove.
Home Work (12): Calculate the radial null geodesics from r = 0,

t = 0 to the point t = T , r = R considering ( and constant) in the
spacetime with the line elements
2
ds2 = dt2 e2Ht [dr2 + r2 (d + sin2 d 2 )]
where H =constant:
Riemann Curvature Tensor

Lemma: if Ai is a contravariant vector, then
Ai;kl Ai;lk = Rmlk

i
Am ; (4.28)
where
i @ imk @ iml i n i n
Rmlk = + nl mk nk ml (4.29)
@xl @xk
Proof:
@Ai;k
Ai;kl = l
+ inl An;k n i
kl A;n
@x
= Ai;kl + imk;l Am + i m
mk A;l
i n i n m
+ nl A;k + nl mk A
n i n i m
lk A;n lk mn A
Using the equality of Ai;lk and Ai;kl we …nd
Ai;kl Ai;lk = ( i
mk;l
i
ml;k + i n
nl mk
i
nk
n
ml )A
m
This proves (4.28).

i
The tensor Rmlk plays an important part in the geometrical proper-
ties of the spacetime. The tensor is called Riemann tensor or curvature
i
tensor. The vanishing of all components of Rmlk implies a ‡at space-
time with considerable simpli…cations of the global properties.
In a locally inertial coordinate (e.g. free falling frame)
1
Riklm = [gim;kl + gkl;im gmk;il gil;mk ] (4.30)
2
and
1
Riklm;p = [gim;k;lp + gkl;imp gmk;ilp gil;mkp ] (4.31)
2
i
Symmetry Properties of Rmlk
Since the symmetry properties of a tensor do dot depend on the coor-
dinate system we can use (4.30) and …nd that
Riklm = Rkilm = Rikml = Rlmik

Riklm + Rimkl + Rilmk = 0 (4.32)
Number of components
A 4th rank tensor generally has 44 = 256 components. However, be-
cause of the symmetry properties (4.32), the number of independent
i
components of Rmlk is only 20.
Ricci tensor
The Ricci tensor is de…ned by
g im Riklm = Rkl
p p
@ 2 ln g @ ikl m@ ln g
= k l l
+ mkn
n
lm kl m
(4.33)
@x @x @x @x
It is clear that
Rkl = Rlk (4.34)
The Ricci scalar is de…ned as
R = g kl Rlk (4.35)
Constant Curvature
In a spacetime of constant curvature we have
R
Rpm = gpm (4.36)
n
where n = gii .
A spacetime satisfying (4.36) is called Einstein space. For such a
spacetime the Riemann tensor can be written as
R
Rptmn = (gnt gpm gmt gpn ) (4.37)
n(n 1)
by taking the divergence of (4.37) and using (4.41) we …nd for n 3
that
R = n(n 1)K = const: (4.38)
So that the Riemann tensor can be written as
Rptmn = K (gnt gpm gmt gpn ) (4.39)
In di¤erential geometry this is known as the curvature tensor for a

space with constant curvature.
The Einstein Tensor

This is a second rank tensor de…ned by
1
Gkl = Rkl gkl R (4.40)
2
The Einstein tensor is divergenless
1 ik
Gik
;k = (R
ik
g R);k = 0 (4.41)
2
To prove this use the Bianchi identities written in the form
mi mi mi
Rjk;l + Rlj;k + Rkl;j =0 (4.42)
and put m = k, i = j to get
m i
R;l Rl;m Rl;i =0 (4.43)
which is equivalent to (4.41).
Home Work (13): Consider the surface of 2-sphere S 2 :
2
ds2 = a2 d + sin2 d'2 :
2
Find the components of Rij then show that R = g ij Rij = a2
Problems
i i
1. Show that if kl 6= lk then the condition gik;l = 0 implies that
i 1
(kl) = g im (gmk;l + gml;k gkl;m ) + g im gkn n
[lm] + g im gl n
n [km]
2
k
2. Show that if [kl] 6= 0 it will transform as a tensor.
1 1 1
3. Calculate 00 ; 11 ; 22 for the line element
1
2 2 2GM 2 2GM 2
ds = c 1 dt 1 dr2 r2 d + sin2 d'2
c2 r c2 r
1
where M; G; c are constants. Compare the result for 00 for
GM
c2 r
<< 1 with the Newtonian gravitational acceleration.
4. Calculate Rki and R for the metric
(ds)2 = c2 (dt)2 e2H t (dr)2 + r2 (d )2 + sin2 (d )2
and show that this metric describes an empty spacetime if it is to
satisfy the modi…ed Einstein …eld equations with the cosmological
constant = 3H 2 =c2 .
Chapter 5
The Field Equations
Heuristic derivation of Einstein Equa-

tion
As we mentioned in chapter 2, the theory of special relativity requires
that the Poisson equation in 3d Euclidean space, i.e.
r2 ' = 4 G (5.1)
should take into account (a) the equivalence of gravitational and inertial
masses, and (b) that this equation should be invariant under Lorentz
transformation. This means that (5.1) should take the general form
g lm rl rm ik = 4 GTik (5.2)
where rl is a covariant derivative, Tik is the energy-momentum tensor
and ik is a second rank tensor that is related to the gravitational po-
tential. The hint that g44 being related to the Newtonian gravitational
potential may suggest that gik or a suitable multiple of it may replace
ik , but since the covariant derivative of gik vanishes rules out such a
suggestion.
There is another tensor that contains the …rst and second derivatives
of gik and plays an important part in the geometry of spacetime; this is
the Riemann curvature tensor Rijkl . However this is a 4th rank tensor,
but a better substitute would be the combination
aRik + bgik R (5.3)

with a and b are constants. Einstein at …rst choose b = 0 so his
equations read,
67
68 The Field Equations
Rik = const.Tik (5.4)

However, he later realized that an additional condition should be taken
into consideration that is
T;kik = 0 (5.5)
This is necessary to maintain that the total energy and momentum is
conserved. Using the symmetry of the Ricci tensor we see that (5.4)
follows as an identity if we choose in (5.3) b = 21 a. Accordingly we
get the Einstein equations to read
1
Rik gik R = Tik (5.6)
2
This is the Einstein Field equation for non-empty spacetime, where
8 G
= (5.7)
c4
is a constant which was calculated later by Einstein assuming that his
equation should reduce to Newton’s equation in the weak …eld limit
as we will see below. It is clear from (5.6) that the Einstein …eld
equation is equating geometry to the energy-momentum content of the
spacetime. This implies that the geometry of the spacetime is produced
according to the energy-momentum content of the spacetime; the LHS
of (5.6) is pure geometry whereas the RHS is pure matter/energy. This
was coined in a beautiful sentence by John Wheeler saying that "Matter
tells spacetime how to curve and spacetime tells matter how to move"
This, in some sense, indicates the equivalence of matter/energy and
geometry.
In 4d-spacetime, the set of equations (5.6) represent 16 di¤erential
equations that are reduced to 10 independent equations by the sym-
metry properties of Rik .
In fact the Einstein equations can be derived rigorously from the
gravitational action
Z
p
I= R gd4 x (5.8)
through the traditional methods of variations, but we will not present

such derivation here.
The Newtonian Approximation

It would be logical to assume that in the limit of slowly changing weak
gravitational …eld the Einstein …eld equations should reduce to the
The Field Equations 69
same form as that of the Newton’s equation of motion and his gravi-
tational law. This was the argument used in order to determine as
given in (5.6). Taking c = 1, we can assume that the metric tensor for
a weak gravitational …eld can be written as
gik = ik + hik (5.9)

where ik is the Minkowski spacetime metric tensor and hik is a small
departure from ik . Thus we will assume that
j hik j<< 1
This means that terms second order and higher in h will be ignored.
Also indices of h will be raised and lowered by ik rather than gik .
We de…ne
h hii = hik ik (5.10)

For Tik we will use the dust approximation and will take
T44 = (5.11)
All other components of Tik are zero.
Now let us contract (5.6) multiplying by g ik and sum over i and k
(with g ik gik = 4) we get
R 2R = T
so that
R= T (5.12)
therefore we can write
1
Rik = Tik gik T (5.13)
2
Using (5.9) we get
1 lm 1
Rik = hik;lm + h;ik hli;lk hlk;il (5.14)
2 2i
Ignoring the time derivatives we get
1 2
R44 = r h44 (5.15)
2
So that from (5.13) and (5.15) we get
1
T44 g44 T = =2 (5.16)
2
so that
r2 h44 = (5.17)
Assuming that h44 = 2 with being the Newtonian gravitational

potential, we have
2r2 = (5.18)
Comparing with the Poisson equation for the gravitational …eld
r2 = 4 G
We …nd that
=8 G (5.19)
If c were introduced we would get
8
= G
c4
So that the full Einstein …eld equations read
1 8 G
Gik Rik gik R = Tik (5.20)
2 c4
These are the Einstein …eld equations for a spacetime with a source
inside it. It is important to notice that these equations unify Mat-
ter and Geometry. The right hand side contains purely geometrical
terms, whereas the left hand side contains the energy-momentum ten-
sor, which represents matter. It was the Einstein dream to convert the
left hand side into a geometrical expression, but was never fruitfully
realized.
Energy-Momentum Tensor
The energy momentum tensor plays a central role in general relativity;
it is the source for the Einstein …eld equations. Consequently, the
geometrical properties of the spacetime will be determined according
to the form of Tik .
The form of Tik is di¤erent for the di¤erent forms of matter and energy.
It is normally derived from the action principle.
Basically Tik has 4 4 = 16 components. But practically all except
the diagonal components are zero, such that Tik becomes
0 1
T00 0 0 0
B 0 T11 0 0 C
Tik = B
@ 0
C (5.21)
0 T22 0 A
0 0 0 T33
T00 = energy density,T11 ; T22 ; T33 are pressures:Others are shears

Here we are not going to consider the derivation of the energy mo-
mentum tensor but will give the forms that are most common after we
outline the derivation method.
The general formula for Tik is
" #
@L 1 @L mn @L 1 ik
T ik = 2 + g gmn;l Lg (5.22)
@gik;l ;l 2 @gik;l @gik 2
where L is the Lagrangian density.

If L does not depend on gik;l explicitly we get
@L
T ik = 2 Lg ik (5.23)
@gik
and
@L
Tik = 2 Lgik (5.24)
@g ik
Special Cases
(A) EM …eld
The action of the EM …eld is
XZ 1
Z
p X Z
lm 4
A= ma da Flm F gd x ea Ai dxi (5.25)
a
16 a
In this case we have
1
L= Fik F ik
16
@L 1 1 m
= F;l Fkm g lm = F Fkm
@g ik 8 8 i
1 1
Tik = Fim Fkm Flm F lm gik (5.26)
(em) 4 4
This tensor has zero trace.
Home Work: Show that Tik has zero trace.

(em)
(B) Radiation Field

For a blackbody radiation –which is isotropic- we have
u u u
Tik = (T11 ; T22 ; T33 ; T44 ) = diag ; ; ;u (5.27)
(em) 3 3 3
where
u = energy density of radiation = E 2 + H 2 =8

(C) Dust
Consider a stream of particles with no pressure, i.e., with relative
random motion. Such a system is called dust.
In a general frame in which the dust particle has velocity v i , we have
T ik = 0v
i k
v ; (5.28)
(m)
with 0 = T is a scalar, and is the rest mass density of the dust. v i

are the components of the velocity.
(D) Fluid
For the individual components of T ik we have
1
T 11 = T 22 = T 33 = P = 0v
2
(5.29)
3
where v is the average velocity, and
T 44 = 0 + 3P = (5.30)
In a general (non-rest, non-locally inertial) frame of reference we
may write
T ik = (p + )v i v k pg ik ; (5.31)
(matter)
and
T 44 = 0 (5.32)
(radiation)
The Einstein Equations for Empty

Spacetime
If the spacetime is completely empty then
Tik = 0 (5.33)
Accordingly (5.20) becomes
1
Gik Rik gik R = 0 (5.34)
2
These equations represent the gravitational …eld outside the gravitating
source (i.e. outside the massive distribution). Note that if the space-
time is empty then this does not imply that it should be necessarily
‡at, so that we can have a curved spacetime but without energy matter
content and the Einstein …eld equations still will have a non-trivial so-
lution. Also we note that, in the weak …eld solution, the gravitational
potential is produced as a result of the time curvature rather than the

space curvature.
One conceptual note is due here in connection with the meaning
of empty spacetime. Usually it was always part of the mainstream
thought of the Greeks and modern Europeans that space is allocated
once objects material exist. Empty space has no meaning on its own.
But in general relativity we see that space time has its own identity
and for this reason empty spacetime has its own character independent
of the existence of material object. Physically empty spacetime that
does not contain any matter or energy can be imagined to be the region
outside the energy-momentum source, for example, the region outside
a mass distribution. This will be discussed in the next section.
Chapter 6
The Schwarzschild Solution
The solutions of Einstein …eld equations depend on two factors:
1. De…ning Tik for the system.
2. The boundary condition imposed.
There are two types of solutions:
1. Matter …elds (localized source) solutions.
2. Cosmological solutions: concerns global distribution of matter

and energy.
Karl Schwarzschild obtained what was called later the Schwarz-

schild solution in 1916, soon after Einstein published his work on gen-
eral relativity. It describes the geometry of the spacetime in empty
space outside a spherically symmetric distribution of mass.
In Newtonian gravitation, the force of gravity is thought to emanate
from the massive body of mass M radially causing a gravitational po-
tential at a distance r from the center of the body given by
GM
= (6.1)
r
This potential is centrally symmetric. According to Newton the

gravitational interaction is transmitted with in…nite speed .
75
76 The Schwarzschild Solution
Spherically Symmetric Spacetime

The most general form of the spherically symmetric spacetime is given
by
2 2
ds2 = Ac2 dt2 Bdr2 Cr2 d + sin2 d + Ddtdr (6.2)
where A; B; C; and D are general functions of r and t. The sub-

spaces r =constant, t =constant are spheres on which a typical point
is labeled by the coordinates ; . The center of symmetry is the point
O given by r = 0.
It is always possible to choose a coordinate transformation that will
enable us make a suitable choice of the coe¢ cients in these coordinates.
Consider now a coordinate transformation that makes
C = 1 and D = 0 (6.3)
Accordingly the metric (6.2) becomes

2 2
ds2 = Ac2 dt2 Bdr2 r2 d + sin2 d (6.4)
This metric represent an isotropic and homogeneous spacetime. We

note that a purely radial element in this metric (6.4) measures A1l2
times the corresponding measure in Euclidean space, while an angu-
lar element (transverse) will measure the same as the corresponding
Euclidean measure.
We can rewrite the line-element (6.4) in the form
2
ds2 = e c2 dt2 e dr2 r2 d + sin2 d'2 (6.5)
where ;and ;are functions of t and r.
The Schwarzschild Solution

The Schwarzschild solution is the solution of the Einstein …eld equation
in empty space outside a spherical mass M . In tis case Tik = 0:This
means that we have to solve the equations
1
Rik gik R = 0 (6.6)
2
But if Tik = 0 then it can be easily shown that R = 0 too. This goes
as follows:
Consider
1 8 G
Rik gik R = Tik (6.7)
2 c4
The Schwarzschild Solution 77
Multiply both sides by g ik and sum over i and k we get
R 2R = T
where T = Tii is the trace of the energy-momentum tensor. However,

if Tik = 0 the T = 0 too, so that R = 0.
Therefore, we have to solve the Einstein equation
Rik = 0 (6.8)
In order to do this we have to calculate …rst the Chistofell symbols for

the metric in (6.5). We have
g00 = e ; g11 = e ; g22 = r2 and g33 = r2 sin2 (6.9)

and
2
g 00 = e ; g 11 = e ; g 22 = r 2
and g 33 = (r sin ) (6.10)
Using
s 1 @gmi @gmk @gik
ik = g ms + (6.11)
2 @xj @xi @xm
and the explicit forms (6.9) and (6.10) we …nd that
1 1@ 0 1@ 1 @
10 = ; 00 = ; 100 = e ;
2 @t 2 @t 2 @r
1 1
22 = re ; 33 = r sin2 e ;
3 3 2 1 1 1@
23 = 32 = cot ; 33 = sin cos ; 10 = 01 = :(6.12)
2 @t
where the prime denotes di¤erentiation with respect to R and dot is
di¤erentiation with respect to T
The (00) and (11) components of gives
0
1 1
R00 = e + =0 (6.13)
r r2 r2
and
0
1 1
R11 = e + + =0 (6.14)
r r2 r2
From these we get
0 0 @( + )
+ =0) =0 (6.15)
@r
This means,
+ = f (t) (6.16)
The arbitrary function f (T )can, however, be set to equal zero since
we still have an arbitrary time transformation. Therefore we can take
without loss of generality
+ =0 (6.17)
From R01 = 0 we have
=0 (6.18)
Thus both and (= )are functions of R only. The equations (6.13)

and (6.14) then yield the solution
A
e =e =1 ; (A = const:) (6.19)
r
This means that the line-element (6.5) becomes
1
A A 2
ds =2
1 2
c dt 2
1 dr2 r2 (d + sin2 d 2 ) (6.20)
r r
If the above metric is to yield the Newtonian potential form, then

A = 2GM=c2 . So that
1
2GM 2GM 2
ds2 = 1 c2 dt2 1 dr2 r2 (d + sin2 d 2 )
c2 r c2 r
(6.21)
This is known as the Schwarzschild line element. This solution is man-
ifestly static.
Figure 6-1 spacetime dip

Schwarzschild Black Hole

The line element in (6.21) describes the gravitational …eld outside a
mass M . This line element is characterized by the following features:
1. There exist a singularity at R = 2GM=c2 . This singularity

describes a spherical event horizon that contains the mass M
within, and at which the temporal component of the metric ten-
sor vanishes, that is to say that the temporal separation between
any two events taking place on this surface is in…nitely dilated.
In fact we can say that physical events at this surface are frozen.
The radius of the event horizon is called Schwarzschild radius
2GM
r = Rs = (6.22)
c2
This is not a real singularity since it can be removed by certain
change of the coordinates.
2. At r = Rs the escape velocity is c, so no particle or EM signal
can escape the even horizon. Any light signal that falls on the
event horizon will be captured and will not be re‡ected away.
Therefore, such a system will be a "Black Hole".
3. Outside the event horizon, the temporal separation between two
point is dilated according to the relationship
r
2GM 0
dt = 1 dt (6.23)
c2 r
where dT is the coordinate time at in…nity. i.e., the time interval
near a body of mass M and at a distance r from its center is
given by
0 t
t =q (6.24)
2GM
1 c2 r
This leads to the phenomena of gravitational redshift which we

will discuss in the next chapter.
4. There is a real singularity at r = 0 at which the density becomes
in…nite.
Home Work (14): (a) Theoretically the density of a black hole

is in…nity since it is assumed that the mass will continue to collapse
into a point. But assuming that the black hole mass is homogeneously
distributed within the volume enclosed by the event horizon, calculate
the density of such a black hole as a function of its mass. (b) On the
same verify the Laplace claim that a star of the same density as the
Earth (5:4 gm/cm3 ) with diameter 250 times that of the Sun would
not be able to shine.
Orbits in Schwarzschild Spacetime

We will consider the motion of planets and light rays in the gravita-
tional …eld of a mass M described by the Schwarzschild spacetime.
The geodesic equation is
d2 xi i dx dxk j
+ jk =0 (6.25)
ds2 ds ds
For the Schwarzschild spacetime we will use the following Christofell
symbols
0 0 m 1 m (r 2m) 1 m
10 = 01 = ; 00 = ; 11 =
r(r 2m) r3 r(r 2m)
1
22 = (r 2m); 133 = (r 2m) sin
2 2 1
12 = 21 = 313 = 331 = ; 233 = sin cos
r
3 3
23 = 32 = cot ; all others = 0: (6.26)
where m = GMc2
.
For i = 0 (using our usual convention of coordinates) this give us
:: 2m ::
t+ tr = 0 (6.27)
r(r 2m)
or
d 2m :
1 t =0
ds r
and the solution is
2m :
1 t = b = constant (6.28)
r
For i = 2 we get
:: 2:: :
+ r sin cos =0 (6.29)
r
For i = 3 we get
2: : :: : :
+ r cot =0 (6.30)
r
For i = 1 we take the expression
1
2m 2m
ds2 = 1 c2 dt2 1 dr2 r2 (d 2 +sin2 d 2 ) (6.31)
r r
and divide both sides by ds2 = c2 d 2
to get
1 :2
2m :2 2m r r2 : 2 :2
1= 1 t 1 ( + sin2 ) (6.32)
r r c2 c2
The use of this form is useful in case we want to consider the trajectories
of light rays in the background of Schwarzschild spacetime in that case
we only need to set the LHS to 0 instead of 1.
Now consider the geodesic passing through

:
a point P on the equator
= 2 tangent to the equatorial plane = 0: Then Eqn. (6.29) gives
::
= 0 which imply that planar motion is guaranteed for the relativistic
orbits:In this case Eqn. (6.30) will give us
:: 2: :
+ r =0 (6.33)
r
which means that
d :
r2 =0 (6.34)
ds
Therefore :
r2 = a = constant (6.35)
this insures that the angular momentum is conserved.
Now sub (6.28) and (6.35) into (6.32) using
: dr d dr a
r= =
d dt d r2
we get
1 1 2
2m 2 2m a2 dr a2
1= 1 b 1 (6.36)
r r c2 r 4 d c2 r 2
Now since
2 2
1 dr d 1
=
r4 d d r
then if we multiply by (1 2m=r) =a2 we get
2
d 1 1 c2 (b2 1) 2mc2 2m
+ = + + 3
d r r2 a2 ra2 r
Di¤erentiating with respect to gives
d 1 d2 1 2 d 1 2mc2 d 1 6m d 1
2 + = +
d r d 2 r rd r a2 d r r2 d r
Assuming that no circular orbit is allowed means that we can cancel
d 1
d r
on both sides. Consequently we obtain
d2 1 1 mc2 3m
+ = 2 + 2 (6.37)
d 2 r r a r
This is the di¤erential equation for the orbits in Schwarzschild space-
time.
If we have light rays propagating in the Schwarzschild spacetime
then the equation for the orbit will be
d2 1 1 mc2 3m
+ = 2 + 2 (6.38)
d 2 r r a r
Figure 6-2 Orbits in Scwarzschild spacetime
Precession of Planetary orbits

In Newtonian theory of gravity the di¤erential equation for a planet
moving in non-circular orbit is given by
d2 u 1
2 +u = (6.39)
d p
where u = 1=r and p = a2 =GM = a2 =mc2 : p is the semi-latus rectum

of the ellipse and is given by
p = a0 1 e2 (6.40)
where a0 is the semi-major axis and e is the eccentricity.

The general solution of Eqn. (6.39) is
1 1
u= = [1 + e cos ( )] (6.41)
r p
where is a constant of integration. A suitable choice of coordinates
allows us to set = 0: So that
1 1
u= = [1 + e cos ] (6.42)
r p
This is the equation of an ellipse that has eccentricity e. If e = 0 then
the orbit is perfect circle.
The perihelion of the planets where known to precess as shown in
Fig (6-3). However, not all the observed amount of precession could
be explained by the perturbational e¤ects of the other planets on the
speci…c one under consideration. It remained for general relativity to
Figure 6-3 Precession of the perihelion of the Planets
explain the di¤erence between theory and observations in this respect.

In what follows we will see how this can be done.
If we substitute (8.45) for the last term of (8.40) and neglect the
term in e2 we obtain
d2 1 1 mc2 3m3 c4
+ = 2 + (1 + 2e cos )
d 2 r r a a4
The solution is
1 mc2 3em2 c2
= 2 1 + e cos + sin
r a a2
for small we can expand in to get
2
1 mc 3m2 c2 m 4
= 2 1 + e cos 1 +O
r a a2 r
we now de…ne
3m2 c2
1 = (6.43)
a2
So that
1 mc2
= 2 [1 + e cos ( )]
r a
The period now is not 2 but
2 6 m 2 c2
= 2 +
a2
= 2 + (6.44)
where is the amount of the precision.
It was known that Mercury’s orbit is precessing at a rate of about
5600:73 0:41"per century and with most accurate calculations as-
tronomers could explain 5557:62 0:4" as being due to the e¤ect of
the other planets, mainly Jupiter. However it remains to explain the
remaining 43:11 0:4" per century and clearly the above calculation
shows that general relativity can explain this di¤erence.
Chapter 7
Experimental Test of General

Relativity
The real …eld for testing general relativity is the whole universe because
the gravitational …eld, unlike the electromagnetic …eld, can only be
obtained naturally and cannot be generated in the laboratories. The
Earth’s gravitational …eld is too weak in comparison with the strength
required by general relativity to manifest its predictions. In order to
get sense of the time dilation caused by gravity for example we write
Eq. (6.24) as
t1
t= q (7.1)
2GM
1 c2 r
where t1 refers to the time at a point very far from the mass M. So
1l2
t t1 2GM
= 1 (7.2)
t1 c2 r
for small value for M=r we can approximate this to
t t1 GM
' (7.3)
t1 c2 r
For the Earth we have GM E

c2 RE
' 6 10 10 and for the Sun we have
GMSun
c2 RSun
' 2 10 6 : The smallness of these numbers conveys some idea of
the di¢ culty involved in testing general relativity. However, there are
several predictions of general relativity, which can be tested in the vast
arena of our universe including those massive stars, which are found
near the center of the Galaxy.
85
86 Experimental Test of General Relativity
Figure 7-1 Gravitational redshift
The Gravitational redshift

Consider two observers A and B at rest in a static spacetime. Let
the two observers have coordinates ai and bi where a and b ( =
1; 2; 3) are constant along their respective world lines. Suppose the
null geodesics leave A at ai , and is received by B at bi (see …gure
7-1).
Let the parametric representation of be xi ( ) where is an a¢ ne
parameter so chosen that
xi (0) = ai ; xi (1) = bi ; x0 (0) = ctA ; x0 (1) = ctB (7.4)

What does our geodesic correspond to in physical terms?
It describes a light ray leaving observer A at time tA and reaching
observer B at time tB . Because of the static nature of the line element
we also have another null geodesic solution given by
x = x ( ); = 1; 2; 3;
x = x0 ( ) + ;
0
= const: (7.5)
This describes a light ray leaving A at tA + =c and reaching B at
tB + =c.
Now in the rest frame of A the time interval =c corresponds to our
proper time interval (measured by A) of
[g00 (a )]1=2 (7.6)

c
If n light waves have left A in this time interval, then the frequency of
these waves as measured by A is
Experimental Test of General Relativity 87
cn 1=2
A = [g00 (a )]
Since the same number of waves is received by B in the corresponding

time interval we get the ratio of frequencies measured by B and A as
1=2
B g00 (a )
= (7.7)
A g00 (b )
This also the ratio of the wavelengths
A : B
measured by A and B respectively.

If in the Schwarzschild solution A is an observer located on the surface
of a star at R = RE say, and B is the distant observer with R >>
2GM=c2 we get
1=2
B 2GM
1 (7.8)
A c 2 RE
Thus the spectral lines from a massive compact star should be red-
shifted. For 2GM=c2 RE small compared to unity, the redshift is
B A GM
z= (7.9)
A c 2 RE
White dwarf stars like Sirius B and 40 Eridani B does show redshifts
in the range of 10 4 to 10 5 , which is of the right order of magnitude.
More reliable and quantitatively accurate measurement, however are
possible only in a terrestrial experiments. For example in 1960 Pound
and Rebka measured the change in the frequency of a gamma-ray pho-
ton emitted by an excited iron nucleus as it fell from a height of 60 70
feet. As such a photon falls though a height h, the Newtonian poten-
tial increases by gh where g is the acceleration due to gravity at the
Earth’s surface. From (7.9) we see that the photon should undergo a
blueshift, that is its frequency increases by a fraction gh=c2 . Although
this fraction is as small as 10 15 , it can be measured by modern lab-
oratory techniques (Mossbaeur spectrometry). The Pound and Rebka
experiment and other’s later works have con…rmed the gravitational
redshift e¤ect.
The Bending of Light

According to the relation of mass energy E = mc2 one can take the
photon of frequency to be equivalent to a particle of mass h =c2
accordingly it is expected that such a particle will be a¤ected by the
Figure 7-2 Light Bending
gravitational …eld of massive bodies and will follow a trajectory in

such …elds. In 1911 Einstein published a paper with such a content
and concluded that the sun will bend the trajectories of light passing
by its rim and calculated this bending to be about 0:85". Few years
after devising general relativity Einstein revised his calculation and
found that the bending is double this value. Here we will perform the
calculation of the angle of bending of the light ray for any massive body
and will …nally calculate the e¤ect for the Sun.
The point Q is a star, which is being observed by an observer on
the Earth E when the light ray from the star is likely to graze the
surface of the Sun S (or any other massive body). If a light ray from Q
to E is bent by the Sun’s gravitational attraction, the star may appear
0 0
to be in another direction Q . What is the angle QEQ ??
We start from the equation of null geodesics obtained in the previous
chapter. We have
d2 1 1 3m
2 + = 2 (7.10)
d r r r
If no massive object exist in the pathway of the light ray then we have
a straight path as in …g. (7-3)
d2 1 1
+ =0
d 2 r r
which has the solution
1 1
= cos (7.11)
r r0
r 0
Figure 7-3 No gravitational mass
In presence of the massive object the situation will be as shown in

Fig. (7-1). If we substitute (7.11) on the RHS of (7.10) we get
d2 1 1 3m
2 + = 2 cos2
d r r r0
which has the general solution
1 1 m
= cos + 2 (1 + sin2 )
r r0 r0
For the straight path if r ! 1 then ! 2 : But for the curved path
if r ! 1 then ! 2 + , where is the angle of bending. Therefore
we can write
1 m
0 = cos + 2 (1 + sin2 )j = =2+
r0 r0
This gives
1 m
sin + 2 (1 + cos2 ) = 0
r0 r0
We expect that will be very small angle for moderately massive
objects like the Sun so that we can use the approximation
1 2m
+ 2 =0
r0 r0
which means that
2m
= (7.12)
r0
since the de‡ection of the light ray is occurring twice once before en-
countering the closest distance with the Sun and once after that then
we should take the total de‡ection to be 2 : This means that the total
de‡ection written in terms of the mass of the de‡ecting object is
4m 4GM
=2 = = 2 (7.13)
r0 c r0
For the case of the Sun r = 7 105 km, Msun = 2 1030 Kg, so that
4GM
= = 1: 75" (7.14)
c2 r0
In principle this can be measured by looking at a star when it is

occulted by the Sun at solar eclipse, and at other times of the year
when the star is not in the same part of the sky as the Sun. Since
the change in the angle is very small very careful observations are in-
volved. The present observations are consistent with the prediction in
Eqn. (7.13) within the experimental limits. These limits appear to be
narrow enough to rule out the Newtonian prediction, which is exactly
half of that in Eqn. (7.13). In strict Newtonian theory no interaction
of light with gravitation is postulated. But we may obtain this result
by assuming that a photon or a pulse of light of frequency has energy
h and gravitational mass h =c2 .
There are, however several di¢ culties and uncertainties in this obser-
vations. Perfect solar eclipses are rare and far between to be able to
repeat this experiment often enough. It is usually not possible to go
closer to the Sun’s disk than twice its radius. There is a refraction
of rays by the solar corona complicating the measurements. For these
reasons measurements during solar eclipse are not very reliable and a
better Radar Echo Delay techniques has been developed.
Gravitational Lensing
The power of large masses to bend light paths invites comparison with
optical lenses, hence the topic of ‘gravitational lensing’. The analogy,
however, is not complete, as we shall see.
First, consider light coming from in…nity and bent by a ‘lens’L to
reach an observer O. Call the distance of closest approach between
the light path and the lens b –an ‘impact parameter’. The bending of
the light path may be simply represented as taking place at one point,
discontinuously, as shown in Fig. (7-4). (Note that the angles involved
here are extremely small; of the order of a few seconds of arc at most,
as we saw above. The diagram is therefore greatly exaggerated.) Light
is detected by the observer at a distance D from L:
b
D= (7.15)
4m
However, from (7.13) above = = b
so
b2
D= (7.16)
4m
Figure 7-4 Light getting bent
where, as always, m = M G=c2 , M being the mass of the lens. D

therefore increases with b: rays ‘further out’from L reach the axis at
a di¤erent point. The lens does not focus the rays to one point.
Consider the symmetric situation where the source S, lens L and
observer Oare in a straight line, as in Fig. (7-5). In a planar cross
section the source S appears to the observer O as two images. There
is cylindrical symmetry, however, so the image is actually a ring –the
‘Einstein ring’.
Figure 7-5 Formation of Einstein Ring
What is its (angular) radius? The general situation is sketched in

Fig. (7-5): here the source S, lens L and observer O are not in a
straight line. The distances between the planes are as marked: DLS is
the distance from the lens to the source, etc. These distances depend
on the angles, also as marked. Then
1 Ds = s Ds + aDLS (7.17)
but from (7.13) we have
4m 4m
= =
b 1 DL
so
4 mDLS
1 = s +
1 DL Ds
which on relabelling 1 ! becomes

2
E
= s+ (7.18)
where
4mDLS
E = (7.19)
DL Ds
and E is called the ‘Einstein angle’.
Figure 7-6 Gravitational Lensing
The solutions to equation (7.18) give the angular positions of images

in the sky – there are generally two solutions. In the situation where
DLS; DS >> DL (the source is at an ‘in…nite’distance) we have
2 4m
E (7.20)
DL
The signi…cance of the Einstein angle is that in the symmetric case,

when the source, lens and observer are collinear, S = 0 so = E :
the image is a ring with this angular radius. In the case of lensing in
the galaxy (of one star by another, not the Sun), E 10 300 , an
angle which is too small to be resolvable by current telescopes. This is
an example of microlensing. On the cosmological scale, however, (the
lensing of one galaxy by another) E 100, which is resolvable, and
indeed there is observational evidence for lensing on this scale, known
as macrolensing. (For the …gures quoted above, see HW 15.)
Figure 7-7 Light ray path
In a cosmological context lensing may be used –at least in principle,

and practice seems now not far behind – to …nd the masses and dis-
tances away of galaxies acting as lenses. This is important information,
for example in the study of dark matter; the determination of masses
by this method is non-dynamical, in contrast with the usual method
of detecting dark matter. It will be shown here, in a very simple (and
somewhat unrealistic) model how this works.
In the approximation DLS ~DS >> DL = D Equation (7.18) holds
with
4m
E = (7.21)
D
where m = M G=c2 , M being the mass of the lens, and D being its
distance away. Hence
2 2
s E = 0 (7.22)
This equation has two solutions, 1 and 2 (which are of di¤erent signs).
Their product is
4m
j 1 2j = E = (7.23)
D
We now …nd one more relation between 1 and 2 . This depends on
the di¤ering path lengths of the two rays. Consider a point P in the
observer plane on the line SL extrapolated, as in Fig. (7-7). From P
an Einstein ring is seen: so in the planar diagram light fronts arriving
along the two paths do so at the same time. In addition, to a good
approximation, , the angle between these two rays, is equal to 1 +
2 , the corresponding angle at O. Let OP = d, CP = l1 , DP = l2 .
Also let CL = LD = h (and LO = D, above). Then by Pythagorean
q 1l2
2 (h + d)2 (h + d)2
l1 = D2 + (h + d) = D 1 + D+ (7.24)
D2 2D
similarly
(h d)2
l2 D+ (7.25)
2D
so
2hd
l1 = 2h
l2 = (7.26)
D
But 2h = D( 1 + 2 ), so l1 l2 = D( 1 + 2) and the time di¤erence
between the two signals arriving at P is
l1 l2 D
t= = ( 1 + 2) (7.27)
c c
On the other hand S (compare Figs. 7-6 and 7-7) and from the
theory of quadratic equations we have from (7.22)
2 2
j 1 + 2j = s (7.28)
(recall once more that 1 and 2 are of di¤erent signs), hence =

j 1 2 j and (7.27) gives
D 2 2
t= j 1 2j (7.29)
c
hence
2c t 2
j 1 2j = (7.30)
D
This is a second relation between 1 and 2 ; so by measuring 1 ; 2 and
t, the Equations (7.23) and (7.30) allow us in principle to determine D
and m, the distance away and mass of the lens. It was mentioned above
that this model is too simplistic; two e¤ects which have been ignored are
the echo delay e¤ect (the subject of the next section) and the expansion
of the Universe. Nevertheless we have shown how gravitational lensing
is a useful tool for cosmology.
Home Work (15): Calculate the Einstein angle for (a) Lensing
within the galaxy of a star by an object of one solar mass between us
and the star, (b) lensing of a source at a cosmological distance by our
Galaxy (M 109 M ).
Problems
1. Find the radius at which light would travel in a circular path
round a body of mass M , (i) according the General Relativity,
(ii) according to Newtonian theory.
2. Show that, according to Newtonian theory, for a planet of mass

in a non-circular orbit around a Sun of mass M , the laws of
conservation of energy (kinetic plus potential) and angular mo-

mentum L = a require that the equation for the orbit, r( ), be
(with u = 1=r)
d2 u GM
2 +u =
d a2
Chapter 8
Cosmology
Until the beginning of the twentieth century Cosmology was not an

exact science; most of our information about the universe on the large
scale were part of the religious teachings. By the …rst two decades of
the twentieth century astronomers were able to explore the large scale
universe due to progress made in telescope technology. Edwin Hub-
ble was able to discover that the universe is not just one island of the
galaxy called Milky Way, but is actually constituted of a large num-
ber of galaxies some similar to our own Milky Way galaxy. However,
until the end of the second decade of the twentieth century the uni-
verse was thought to be static, but after twenty years of observations
Hubble discovered that the line spectra of far galaxies shows a shift
mostly toward the red end of the spectrum. If the Doppler e¤ect will
be adopted to explain this redshift then we should say that galaxies
are receding from us. Moreover it was observed that the shift is pro-
portional to the distance of the galaxy; the more far the galaxy is the
more redshift it shows. The led to the conclusion that the universe is
expanding according to the law
v = Hr (8.1)
where H is called the Hubble Constant.

But does this means that we are at the center of the universe? Not
necessary, since if we assume that the universe looks the same from any
point within then we should expect that wherever we go we will see the
same thing, which implies that the universe is spatially homogenous.
This is called the Cosmological Principle.
The theory of general relativity can realize a model for the whole
universe, by which it was found that the universe can be …nite and
unbound. Generally cosmological models comes as a solution to the
97
98 Cosmology
Einstein …eld equation after specifying the energy-momentum tensor

being de…ned. There are three classical solutions, these are:
1. The Einstein static universe.
2. The de Sitter vacuum solution.
3. The Friedman-Roberson-Walker models.
The Einstein Static Universe

This is a model that was suggested by Einstein himself soon after he
published general relativity. At that time it was thought that the uni-
verse is static and never changing. In addition to this Einstein thought
that the universe must be …nite but unbound. Accordingly, the uni-
verse will collapse under its own attractive gravitational forces between
its parts. Einstein further assumed that the universe is spherically sym-
metric and that matter is homogeneously and isotropically distributed
throughout the universe.
It is not di¢ cult to guess the form of the line element for such a sys-
tem. The spherical symmetry suggests that we should adopt spherical
polar coordinates and the homogeneity means that g44 is constant. On
the other hand isotropy of the material distribution suggests that no
term of the form dtdx should appear in the line element. This enabled
Einstein choose the form of the line element as
ds2 = c2 dt2 dx dx (8.2)

where are functions of space coordinates only. The spherical sym-
metry of the spatial part can be realized by having a 3d spherical
surface S 3 embedded in a 4-d space. In this case we will need to have
4 spatial dimensions in order to de…ne the 3d surface. We have the
coordinates being de…ned by
x1 = a sin cos ; x2 = sin sin cos ; x3 = a sin sin sin ; x4 = a sin

(8.3)
The line element of the Einstein universe can be written as
ds2 = c2 dt2 d h2 i
2 2 2 2
= c dt a2 d 2 2
+ sin d + sin 2
d' (8.4)
where 0 , 0 ; 0 ' 2 and a is the radius of

the3 sphere Using the elliptical coordinate
sin = r ) cos d = dr (8.5)

Cosmology 99
the form of the line element in (8.3) can be written as
dr2 2 2
ds2 = c2 dt2 a2 + r2 d + sin2 d' (8.6)
1 r2
Dust Universe
Let us now apply the Einstein …eld equation to the line element in (8.4)
assuming that the spatial part of the universe is …lled with pressureless
dust. First we have
a2
g00 = 1; g11 = 2
; g22 = a2 r2 ; g33 = a2 r2 sin2
1 r
1 r2 1 1
g 00 = 1; g 11 = , g 22
= , g 33
=
a2 a2 r 2 a2 r2 sin2
(8.7)
From which we can calculate the Christofell symbols
1 r 2 3 1 1
11 = ; 12 = 13 = ; 22 = r(1 r2 )
1 r2 r
1
33 = r(1 r2 ) sin2 ; 2
33 = sin cos ; 3
32 = cot
(8.8)
Accordingly the Ricci scalar and the components of Ricci tensor
can be calculated as follows
6
R = g ik Rik =
a2
2
R00 = 0; R11 = 2 (8.9)
a
So that
1 3
R= 2 R00
2 a
1 1 2 1 1 1
R1 R = R2 R = R33 R= 2 (8.10)
2 2 2 a
From the above results we can calculate the RHS of the
Einstein equations
1 jk 8 G j
Rij g gik R = Ti
2 c4
1 j 8 G j
Rij iR = Ti
2 c4
1 8 G 0
R00 R= T0
2 c4
1 8 G
R R= T ( = 1; 2; 3) (8.11)
2 c4
100 Cosmology
So that
1 8 G 0
R00 R= T0 (8.12)
2 c4
and
1 8 G
R R= T ( = 1; 2; 3) (8.13)
2 c4
Now, if we assume that the universe is …lled with pressureless dust then
will have
T00 = energy density = 0 c2 ; and

T11 = T22 = T33 = 0 (8.14)
This means that the …eld equation will yield
3 8 G 1
= 0; =0 (8.15)
a2 c 2 a 2
Clearly no sensible solution can be drawn from this result. This means
that the original Einstein …eld equations do not lend itself to a static
spacetime solution.
The Cosmological Constant

To remedy the dead lock posed by the results (8.15) Einstein modi…ed
his original equations by adding a “Cosmological Constant” term on
the RHS so that his new set of equations would read
The solution was obtained after modifying the original Einstein
equations by adding a cosmological constant term gik to the R.H.S of
the …eld equations so that it would read
1 8 G
Rik gik R + gik = Tik (8.16)
2 c4
The addition of a positive cosmological constant term to the RHS of

the …eld equation is equivalent to add repulsive force acting against the
gravitational self-attraction of the parts of the universe. Without such
a repulsive force Einstein thought that the universe would collapse.
Using the modi…ed Einstein equations (8.16)
3 8 G
+ = 0 (8.17)
a2 c2
and
1
+ =0 (8.18)
a2
From which we deduce that
Cosmology 101
1=2 1=2
1 c 1
a= = (8.19)
2 G 0
Clearly this relationship means that a highly dense universe will be

closed, and an in…nite density universe will have zero size.
If we take the present matter density 0 10 31 gcm 3 ; then from
(8.19) we get
a 1029 cm and 10 58
cm 2
(8.20)
The smallness of the value of the cosmological constant at present is
a striking feature of the universe. This is the subject of many recent
studies and investigations1 .
The Einstein universe went out of interest soon after Edwin Hubble
discovered that the universe is expanding. However recently the inter-
est in the Einstein universe revived in the context of studying quantum
…led theoretic e¤ects and thermodynamical e¤ects using this model.
Home Work (15): Show that the cosmological constant stands

for a long-range repulsive force between galaxies.
Hint: See Sahni and Starobinsky cited below.
The de Sitter solution

In 1917 de Sitter published another solution to the Einstein …eld equa-
tion with the line element given by
1
H 2 r2 H 2 r2 2
ds2 = c2 1 dt2 1 dr2 r2 d + sin2 d'2
c2 c2
(8.21)
where H is a constant related to by
3H 2
= (8.22)
c2
It is very easy to show that the de Sitter universe is empty!!
Home Work (16): Show that the de Sitter line element is a solu-
tion of the modi…ed Einstein …eld equations with for empty spacetime
and with a cosmological constant.
The line element in (8.21) can also be written in the form

1
V. Sahni and Starobnisky "On the cosmological constant” arXive: astr-ph
9904398.
102 Cosmology
2
ds2 = c2 dt2 e2Ht dr2 + r2 d + sin2 d'2 (8.23)
From this form it is clear that the de sitter universe is an expand-

ing spacetime. If we compare with the form in (8.4) for the Einstein
universe, we …nd that it is possible to write
a(t) = eHt (8.24)
This shows that the de Sitter universe is an in‡ating universe; this

property makes such a model a good candidate to describe the very
early stages of the universe dominated by quantum vacuum.
One may compare the form of the de Sitter universe given in (8.21)
with the line element of Schwarzschild given in (6.21) we …nd that
2GM
= H 2 r2 (8.25)
r
According to the Newtonian gravitation, the escape velocity from

the surface of a massive object of mass M is
2 2GM
Vesc = (8.26)
r
accordingly we get
V = Hr (8.27)
This is exactly the Hubble law for the velocity of galaxies.
The Robertson-Walker metric

The most general and well-known time-dependent metric is known as
the Robertson-Walker metric, which was suggested in 1935 to describe
a time-dependent homogenous and isotropic spacetime. This comes up
as a generalization to the Einstein static metric. The metric is given
by
dr2 2
ds2 = c2 dt2 S 2 (t) + r2 d + sin2 d'2 : (8.28)
1 kr2
The function S(t) is the scale factor and yet to be determined. Also
k is known as the curvature constant in this solution but it has to be
determined too. This can be normalized to take the values 1; 0 and 1.
We will take this occasion to talk about the scale factor S(t). On
a surface of, say a sphere, we can de…ne any point to be an origin.
Cosmology 103
The distance between two point S(t) can be a measure of the size of
the spherical surface and accordingly can be taken to stand for the
“radius”of the surface.
Also, since the distance in curved spacetime is coordinate-dependent,
therefore, some kind of a universal measure should be de…ned so as to
get a proper meaning for the distance; this is possible if we de…ne a
“proper distance”given by
D = S(t)r (8.29)
where S(t) is the scale factor and r is the radial position coordinate.
Friedmann Models
Earlier to the time when Robertson and Walker suggested their solu-
tions, Alexander Friedmann suggested a time-dependent solution for
the Einstein …eld equations in 1924. This model pre-assumes that the
universe is homogeneous and isotropic with the galaxies behaving like
particles of pressureless dust. The line element reads
dr2 2
ds2 = c2 dt2 S 2 (t) + r2 d + sin2 d'2 : (8.30)
1 kr2
where again here we have k = 1; 0; 1.

Accordingly here we have three models for the universe:
1. Closed model with positive curvature (k = 1): where the universe

starts from a singularity S(t = 0) = 0 and expands until reaching
a maximum radius after which the universe would collapse to the
state similar to the one started with.
2. Open ‡at model (k = 0): where the universe starts from a sin-
gularity S(t = 0) = 0 and expands in an decelerating mode till
reaching a coasting speed. This is also called the Einstein –de
Sitter universe.
3. Open with negative curvature (k = 1): where the universe

starts from a singularity S(t) = 0 and expands in an accelerating
mode inde…nitely.
Cosmological Red shift

The equation for null geodesic (path of light rays) can be described by
ds = 0. For a constant and this means:
104 Cosmology
Figure 8-1 The Friedmann Models
dr2
c2 dt2 = S 2 (t) (8.31)
1 kr2
Integrating we have
t0Z
+ t0 Zr1
cdt dr
= p
S(t) 1 kr2
t1 + t1 0
Assuming that S(t) is a slowly varying function of time t, we can write
c t0 c t1 t0 S(t0 )
= ! = ;
S(t0 ) S(t1 ) t1 S(t1 )
or
S(t0 )
1+z = (8.32)
S(t1 )
where
t0
z= 1 (8.33)
t1
is called the cosmological redshift parameter. Note that this should not
be confused with the gravitational red shift, which is asymmetric.
The Einstein …eld equations

In order to obtain the Einstein …eld equations we should …rst calculate
the following objects
S 2 (t)
g00 = 1; g11 = ; g22 = r2 S 2 (t); g33 = r2 S 2 (t) sin2
1 kr2
(8.34)
and
Cosmology 105
1 S_ 1 @S dS
1
01 = 2
02 = 3
03 = 2
= ; S_ =
c S S @x0 dt
1 kr 2 3 1
11 = 2
; 12 = 13 = :
1 kr r
1 2 1
22 = r(1 kr ); 33 = r 1 kr2 sin2
2
33 = sin cos
3
23 = cot :
_ 2
S Sr S S_ _ 2 sin2
S Sr
0 0 0
22 = ; 11 = ; 33 = (8.35)
c c(1 kr2 ) c
Then
3 S•
R00 =
c2 S !
1 2 3 1 S• 2S_ 2 + 2kc2
R1 = R2 = R3 = 2 +
c S S2
!
6 S• S_ 2 + kc2
R= 2 + (8.36)
c S S2
From which we calculate the components of the Einstein tensor
!
1 3 S_ 2 + kc2
G00 = R00 R=
2 c2 S2
!
1 1 2S• S_ 2 + kc2
G11 = G22 = G33 = R11 R= +
2 c2 S S2
Then from the Einstein equations
8 G j
Gji = T (8.37)
c4 i
we obtain
S_ 2 + kc2 8 G 0
= T (8.38)
S 2 3c2 0
and
2S• S_ + kc2 8 G 8 G 8 G
+ 2
= 2 T11 = 2 T22 = 2 T33 (8.39)
S S c c c
106 Cosmology
The Equation of Fluid

The Friedmann model assumes that the material content of the universe
is a ‡uid governed by:
dE + P dV = T dS = dQ (8.40)
where dE is the internal energy, pdV is the work done and dQ is the
quantity of heat involved. dQ is considered zero because there is no
transfer of heat outside the universe. Let be the mass density, then
4 3 2 4 3
E= S c; M= S
3 3
dE dS 4 d
= 4 S 3 c2 + S 3 c2
dt dt 3 dt
4 3 dV dS
V = S ; =4 S2 (8.41)
3 dt dt
Assuming that T ds = 0, we get
!
S_ P
_ +3 + = 0; (8.42)
S c2
This is called the Fluid Equation.
The Friedmann Equations

The Friedmann Equations are the followings
S_ 2 + kc2 8 G 0
= T (8.43)
S 2 3c2 0
and
2S• S_ 2 + kc2 8 G 8 G 8 G
+ 2
= 2 T11 = 2 T22 = 2 T33 (8.44)
S S c c c
Note that the above two equations plus the ‡uid equation are not
su¢ cient to determine the state of the universe because we need an
equation of state to de…ne the relation between and P . This equation
of state depends on the type of matter …lling the universe. We will give
such equation when we come to solve the Friedmann equations.
First let us identify that
T00 = c2 (8.45)
and
T11 = T22 = T33 = P (8.46)
so that (8.43) becomes
!2
S_ 8 G kc2
= (8.47)
S 3 S2
Cosmology 107
This is called the …rst Friedmann equation.
Now according to (8.46) Eqn. (8.44) becomes
2S• S_ 2 + kc2 8 G
+ = P (8.48)
S S2 c2
Using (8.47) we get
S• 4 G 3P
= + (8.49)
S 3 c2
This is called the second Friedmann equation or the Acceleration Equa-
tion.
Solving the Friedmann Equations

As mentioned earlier in order to solve the Friedmann equations we need
to de…ne the state of matter and energy of the system. This is called
the equation of state. For this purpose we have two cases:
The pressureless dust model

In this case we have P = 0, then the equation of ‡uid becomes
S_
_ +3 =0 (8.50)
S
This can be written in the form
1 d
S3 = 0
S 3 dt
which means that
S 3 = constant
so that
1
S3
which means that
3
0 S0
= (8.51)
S3
where 0 is the present energy density, and S(t0 ) = 1:
How does S change with time? This needs us to go back to Fried-
mann equations. Taking k = 0 (‡at universe) in (8.47) we have
!2
S_ 8 G 0
= : (8.52)
S 3 S3
108 Cosmology
so that
8 G
S_ 2 = 0
(8.53)
3 S
q
To solve this take S(t) t , then
L.H.S =) t2q 2 and R.H.S =) t q
2
2q 2 = q ) q = (8.54)
3
so that
S(t) / t2=3 : (8.55)

This means that setting S(t0 ) = 1, and According to (8.51) then
0 t20
(t) = = 0 2 (8.56)
S 3 (t) t
The radiation model
In this case we assume that the universe is …lled with massless radiation
only. The equation of state is given by
c2
P = (8.57)
3
and
c2
T00 = c2 , T11 = T22 = T33 = (8.58)
3
The ‡uid equation gives us
S_
_ +4 =0 (8.59)
S
which means that
d
( S 4) = 0
dt
which means that
1
(8.60)
S4
Consequently
1=2
t
S(t) = S(t0 ) (8.61)
t0
and
Cosmology 109
2
0 t0
(t) = = 0 (8.62)
S 4 (t) t
Note that: (t) is the same in both cases (dust and radiation).
But a universe dominated by radiation expands more slowly than the
one dominated by dust. The reason is that the pressure causes more
deceleration.
Mixture of dust and radiation

The real universe normally contains both matter and radiation. For
dust we have
1
dust (8.63)
S3
and for radiation we have
1
rad (8.64)
S4
The total density
= dust + rad
This means that the scale factor will have a more complicated be-
havior and so to convert (S) into (t) is much harder. It is possible to
obtain exact solutions for this situation but they are very messy. We
will consider the simpler situation where one or other of the densities is
by far the larger. In that case we can say that the Friedmann equation
is accurately solved by just including the dominant component. That
is we can use the expansion rates we have already found. For example
suppose that radiation is much more important then one would have
S(t) / t1=2 ; rad / 1=t2 ; dust / 1=S 3 / 1=t3=2

Notice that the density in dust falls o¤ more slowly than that in ra-
diation so the situation of radiation dominating cannot last forever,
however small the dust component might be originally it will eventu-
ally come to dominate. We can say that domination of the universe by
radiation is an unstable situation.
In the opposite situation, where it is the dust which is dominant we
obtain the solution
S(t) / t2=3 ; rad / 1=t2 ; dust / 1=S 4 / 1=t8=3

Dust domination is stable situation; the dust becomes increasingly
dominant over the radiation as time goes by.
Figure (8-2) shows that the evolution of a universe containing dust and
radiation, with the radiation initially dominating. Eventually the dust
comes to dominate, and as it does so the expansion rate speeds up from
110 Cosmology
Figure 8-2 Radiation and Matter
S(t) / t1=2 to the S(t) / t2=3 law. It is very possible that this is the
solution, which applies in our present universe.
So if S(t) is small (the early state of the universe)
dust < rad
But for
dust > rad
This means that if we start with a universe containing mixture of mat-

ter (dust) and radiation then for small values of S(t) radiation density
will dominate [radiation dominated universe], but as S(t) becomes large
matter dominates [matter dominated universe].
Home Work(8.2): Starting from the basic Friedmann equations cal-
culate S(t), (t) for dust and radiation for the cases k = 1.
The density of the universe

In analyzing the possible futures of the universe we will assume that the
universe is dominated by dust only. In practice this is not a restrictive
assumption and the results actually apply to real universe.
We have already assumed that if we assume that the constant k
in the Friedmann equations is zero, then the universe expand for ever,
S(t) / t2=3 , but slows down arbitrarily at late times. So we know the
fate of the universe in that case.
But what happen if k 6= 0? The principal question to ask is whether it
is: possible for the expansion of the universe to stop which means that
S = 0, i.e., H = 0. Looking at the Friedmann equation
8 G k
H2 =
3 S2
Cosmology 111
It is immediately apparent that this is not possible if k is negative,

for then both terms on the RHS of Friedmann equation are positive.
Consequently such a universe must expand forever. That enables us to
study the late-time behavior, because we can see that the term k=S 2
falls o¤ more slowly with the expansion than does
dust / 1=S 3
Since S becomes large for dust-dominated solution for negligible k,

the k=S 2 term must eventually come to dominate. When it does, the
Friedmann equation becomes
: !2
S k
=
S S2
Cancel o¤ the S 2 on both sides and you will …nd the solution is S / t.
So when the last term comes to dominate, the expansion of the universe
becomes yet faster. In this case, the velocity does not tend to zero at
late times, but instead becomes constant. This is sometimes known as
free expansion.
Things are very di¤erent if k is positive. It then becomes possible
for H to be zero, by the two terms on the RHS of Friedmann equation
canceling each other out. Indeed this is inevitable, because the nega-
tive in‡uence of the k=S 2 term will become more and more important
relative to the density term as time goes by. In such a universe, there-
fore, the expansion must come to an end after a …nite amount of time.
As gravitational attraction persists, the re-collapse of the universe be-
comes inevitable.
In fact the collapse of the universe is fairly easy to describe, because
the equations governing the evolution are time reversible. That is if
one substitutes t by t, they remain the same. The collapse phase is
therefore just like the expansion in reverse, and so after a …nite time the
universe will come to an end in a Big Crunch. These three behaviors
are illustrated in Figure (8-1).
Home Work (17) Discuss the fate of the universe assuming that
the density term is proportional to 1=S 2 .
Problems
1. Consider the surface of 2d sphere of radius R shown in the …gure
below. Write the general for the circumference of a circle as
a function of r, remembering that r = R. Demonstrate that
this gives the normal ‡at relationship for small r. Evaluate the
relation for the case when the circle is at the equator.
112 Cosmology
2. Suppose that galaxies are evenly distributed points on the surface

of the sphere, with a number density n per unit area. Show that
the total number N of galaxies inside the radius r is given by
N = 2 nR2 1 cos Rr
3. Expand this for r << R to show that the ‡at space result that the
number is 2 r2 n is recovered. Do you see more or fewer galaxies
out to the same radius if the universe is spherical rather than
‡at?
Figure 8-3 Home Work
1. Suppose we have a more general equation of state p=( 1) c2 ;where

is a constant in the range 0 < < 2. Find the solution for
(a); a(t) and hence (t) for universes containing such matter.
Assume k = 0 in the Friedmann equations.
2. Using your answer to problem (2), what value of would be

needed so that has the same time dependence as the curvature
term ak2 ? Find the solution a(t) to the full Friedmann equation
for a ‡uid with this , assuming negative k.
3. Show that the equation for the conservation of energy leads to

the equation of ‡uid.
4. Assuming a typical Galaxy weighs 1011 Msun , and using the criti-
cal density given in the equation , use your answer to the previous
problem to estimate the number of galaxies in the observable uni-
verse. How many protons are there in the observable universe?
5. An ideal gas has pressure p = 13 nhv P i; the average being over the
direction of particle motion. Here n is the number density. Using
the equation for the total relativistic energy of a massless particle,
show that this gives p = 31 nhEphoton i . Hence demonstrate the
equation for radiation.
Cosmology 113
6. Con…rm the result, quoted in the main text that pressureless

universe with cosmological constant has a deceleration parameter
given by q0 = 20 (t0 ).
7. (a) The energy density in radiation is related to its temperature

by
2
rad = rad c
Compute the temperature when the universe was one second old, using
the Friedmann equation and the radiation-dominated solution a(t) _
t1=2 .
(b) What was the corresponding mass density at that time. Com-
pare it with that of water. How old was the universe when its density
matches that of water.
1. In the dust dominated open universe the present age of the uni-
verse is given by the equation
1 0 1 2 0
H0 t0 = 3=2
cosh
1 0 2 (1 0) 0
Demonstrate that in the limiting case of an empty universe 0 ! 0;we

get H0 t0 = 1, and in the limiting case of a ‡at universe 0 ! 1 we
recover the result H0 t0 = 32 : [Useful formula:
p cosh 1 (x) t ln(2x) for
large x, and cosh 1 [(1 + x)=(1 x)] t 2 x + 2x3=2 =3 for small x].
114 Cosmology
Chapter 9
The Evolution of the Universe
It is known that the universe contains 92 chemical elements that seem

to have been formed naturally. The lightest of these natural elements
is the Hydrogen with its atom being composed of one proton and one
electron. The next heavier element is Helium with its atom being
composed of two protons and two neutron forming the nucleus and two
electrons rotating aroun it. The heviest element is Uranium with its
atomic nulcleus being composed of 92 protons and 146 neutrons. These
elements are known to have isotopes which have the same number of
protons and di¤er only in the number of neutrons. These elements
and their isotopes have di¤rent average abundance in the composition
of the universe. Hydrogen being the most abunndant with a value of
about 76% next is the Heium with nearly 23% and the rest of the elents
only have about 1% abundance. The Question that jumps to the mind
why is this?
In 1948 George Gamow thought of a senario for the development
of the universe by which the abundance of the natural element might
be justi…ed. He started with a tiny space full of a soup of all known
elementary particles being at thermal equilibrium at very high temper-
ature. Gamow employed all relevant laws of physics known at his time
including general relativity, particle physics, thermodynamics and sta-
tistical mechanics. Particles were continuously created ou of the avail-
able radiation and soon get annihilated to become radiation again. As
the universe expanded hevy particles went out of the equilibrium so
that they went into the process of annihiliation without geting created
again. The latest particles which left the thermal equilibrium were the
lightest ones. With such such a scenario Gamow were able to explain
the abundance of the light elements H, He and Li and their isotopes.
However he found that at certain stage of the development of the uni-
115
116 The Evolution of the Universe
Figure 9-1 The emission of the CMBR photons
verse when the average temperature was about 5000 K the electrons
got combined to the neuclei to form atoms of these elements. At this
recombination event the newly formed atoms emited photons with the
equivalent binding energy of those electrons and the universe suddenly
became transparent. These photons had the chance then to travel
very long distances without being interrupted by any objects. Con-
sequently such photons formed a smooth homogenous and isotropic
themal background for the universe. As the universe expanded since
then the wavelength of those photons get redshifted (became longer)
and they became a thermal bath of microwaves forming what became
known as the Cosmic Microwave Backround Radiation (CMBR). The
equivalent antenna temperature of these radiation is now about 2:73
K.
Indeed this kind of radiation was a predition for the Gamow theory
and during the sixties of last century prominent physicists like and Jim
Peeble worked hard on searching for these radiation. They invested
substantial amount of money and large e¤orts to detect thos photons
predicted by the Gamow scenario.
Bibliography
[1] J.V. Narlikar, Lectures on General Relativity, Wiley 1979.
[2] J.V. Narlikar, Introduction to General Relativity, Cambridge Uni-

versity Press 1999.
[3] A. Liddle, Introduction to Modern Cosmology, Wiley 2003.
[4] L. Ryder, Introduction to General Relativity, Cambridge University

Press, 2009.
[5] S. Weinberg, Gravitation and Cosmology, Wiley 1972.
[6] S. Carroll, Spacetime and Geometry, Addison Wesley, 2004.
[7] Barbara Ryden, Introduction to Cosmology, Addison Wesley, 2003.
117
View publication stats

Lecture Notes On General Relativity and Cosmology Mhadrat Fy Nzryt Alnsbyt Alamt

Uploaded by

Copyright:

Available Formats

Lecture Notes On General Relativity and Cosmology Mhadrat Fy Nzryt Alnsbyt Alamt

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lecture Notes On General Relativity and Cosmology Mhadrat Fy Nzryt Alnsbyt Alamt

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Lecture Notes on General Relativity and Cosmology ‫ﻣﺤﺎﺿﺮات ﻓﻲ ﻧﻈﺮﻳﺔ اﻟﻨﺴﺒﻴﺔ‬

Book · February 2019

Mohammed Basil Altaie

Dark Matter is it Necessary? View project

Back-reaction of Quantum Fields in curved spacetimes View project

The user has requested enhancement of the downloaded file.

Prof. M.B. Altaie

1 Review of Special Relativity 9

2 The Principles of General Relativity 31

Examples of line elements . . . . . . . . . . . . . . . . 36

3 Tensors and Tensor Calculus 41

5 The Field Equations 67

6 The Schwarzschild Solution 75

7 Experimental Test of General Relativity 85

9 The Evolution of the Universe 115

This is a brief introduction to the theory of general relativity and mod-

Acknowledgement 1 I would like to thank my ex-postgraduate stu-

Review of Special Relativity

1. Newton’s mechanics and his theory of Gravity

2. Maxwell’s electromagnetic theory and classical optics.

Figure 1-1 Refrence Frames

3. The kinetic theory of gases, thermodynamics and the Boltzmann

The basic assumption in Newtonian mechanics is that space and

These relations were called Galileo Transformations. It expresses the

Figure 1-3 Di¤erent trajectories from di¤erent frames of reference

Acceleration and Newtonian Mechanics

be written in several forms di¤erent units. In vacuum these equations

Implications of Lorentz Transformations

The Postulates of Special Relativity

Postulate (1): The Principle of Special Relativity

Postulate (2): The velocity of light

The spacetime interval

ds2 = dx21 + dx22 + dx23 + dx24 : (1.5)

Derivation of the Lorentz Transfor-

this correspond to a rotation in the x1 x4 -plane which has the solutions

In matrix form this can be written as

where is the angle of rotation in the plane which would be a function

it then follows that

Substituting in (1.7) we get

The general form of Lorentz transformation can be written as

ds2 = (dx1 )2 + (dx2 )2 + (dx3 )2 + (dx4 )2

Lorentz Transformations and Rotation

= cosh ; = sinh (1.19)

Figure 1-4 Length contraction

Consequences of Lorentz transformations

where t is the time interval measured by the moving observer and

Home Work (3a): -measons are created at the elevation of 6000

Home Work (3b)

1. A rocket leaves a spaceship in the forward direction at a speed

The Proper time

2 ds2 dx2 + dy 2 + dz 2 c2 dt2

Mass and Energy

(for the derivation of this formula see A. Beiser, Concepts of Modern

K:E = mc2 m0 c2 (1.28)

Energy and Momentum

so that the spacetime interval of the special relativity becomes

ds2 = dx2 + dy 2 + dz 2 c2 dt2 (1.31)

ds2 = c2 dt2 dx2 dy 2 dz 2 (1.32)