General Relativity

by

Kemal Akn
 General Relativity

Kemal Akn

Department of Physics, Istanbul Technical University

(Dated: May 4, 2017)

Abstract

These notes are prepared based on the course FIZ 494E - General Relativity taught by
Prof. Dr. Nee zdemir at Istanbul Technical University. This document is kind of a workspace
rather than lecture notes, there are lots of footnotes, boxes, visualized materials & derivations.

ITU - SPRING 2017

Suggested References
Direct links to electronic version of the following documents provided in the footnotes.
I Gravity: An introduction to Einsteins general relativity, J.B. Hartle, Addison Wesley 2003 I

I Introducting Einsteins Relativity, R. dInverno, Oxford: Clarendon, 1992II

I Spacetime and Geometry: Introduction to General Relativity, S.Carroll, Addison-Wesley, 2004 III

I A First Course in General Relativity, B.F. Schutz, Cambridge University Press, 2009 IV

I Einsteins General Theory of Relativity: With modern applications in cosmology, O. Grn, 2007V

I The Classical Theory of Fields, L.D. Landau, Lifschitz, 1973

I An Introduction to General Relativity, L.P. Hughston, K.P. Tod, Cambridge University, 1990

akinkem@itu.edu.tr
I
Gravity - J.B. Hartle
II
Introducting Einsteins Relativity - R. dInverno
III
Spacetime and Geometry - S. Carroll
IV
A First Course in General Relativity - B.F. Schutz
V
Einsteins General Theory of Relativity - O. Grn

1
Contents

1 Introduction 4
1.1 Historical Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 Special Relativity and Spacetime 7

2.1 Galilean Relativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.1.1 Translation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.1.2 Rotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.1.3 Boost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 Lorentz Transformations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

3 Derivation of Lorentz Transformation 13

3.1 Derivation From First Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.2 Landau - Lifshitz Derivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

4 Week IV & V 23
4.1 Velocity Transformations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

5 Tensor Calculus and Riemannian Geometry 38

5.1 Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
5.2 Tensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
5.3 Symmetric and Anti-symmetric Tensors . . . . . . . . . . . . . . . . . . . . . . . . . 45
5.4 Raising and Lowering Indices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.5 Appendix - Coordinate Transformation and Jacobian . . . . . . . . . . . . . . . . . 47

6 Parallel Displacement and Covariant Derivative 53

6.1 Covariant Derivative of Metric Tensor . . . . . . . . . . . . . . . . . . . . . . . . . . 55
6.2 Riemann Curvature Tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
6.2.1 Symmetry Properties of Riemann Tensor . . . . . . . . . . . . . . . . . . . . 56

7 Einstein Field Equations 57

7.1 Schwarzschild Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
7.2 Conformal Transformations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

8 Homeworks 58
8.1 Geometrized Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
8.1.1 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
8.2 Relativistic Velocity Addition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
8.2.1 Rapidity Parameter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
8.2.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
8.3 Acceleration in Special Relativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

2
 8.3.1 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
8.4 Transformation Law of Christoffel Symbols . . . . . . . . . . . . . . . . . . . . . . . 64
8.4.1 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
8.5 Static and Spherically Symmetric EFE . . . . . . . . . . . . . . . . . . . . . . . . . . 68
8.5.1 Christoffel Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
8.5.2 Riemann Tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
8.5.3 Ricci Tensor & Scalar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

List of Figures

1.1 Cube of Physics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2.1 Light Cone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2 The standard configuration in Special Relativity of two frames of reference, the
primed system in motion relative to the unprimed system only along the x-axis and
with speed v. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3 Two frames of reference S and S 0 moving relative to each other with a velocity v
have a flash of light originate at their respective origins at time t = 0. . . . . . . . . 12
3.1 The Lorentz factor plotted against velocity as a fraction of the speed of light. . . . . 17
3.2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
4.1 A curve parametrized with . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.2 A curve parametrized with s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.3 A unit hyperbola . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.4 Plot of tanh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.5 Two frames moving with a constant speed with respect to each other . . . . . . . . . 31
4.6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
8.1 Plot of tanh x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

List of Notes

o quizi ekle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
o dzgn szck bul . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

List of Boxes

3.1 Imaginary Time [1] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

5.2 A tensor is something that transforms like a tensor [8] . . . . . . . . . . . . . . . . . . . . 41
6.3 Historical Remark . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

3
 1 | Introduction

In Figure 1.1 a brief summary of all of physics is shown. Physics started with Newtonian mechanics
at one corner of the cube, and is now desperately trying to get to the opposite corner, where sits
the alleged Holy Grail1 . The three fundamental constants, c1 , h, and G, characterizing Einstein,
Planck or Heisenberg, and Newton, label the three axes. As each of these constants came into
physics, we took off from the home base of Newtonian mechanics.

Much of 20th century physics consisted of getting from one corner of the cube to another. Consider
the bottom face of the cube. When we turned on c1 we went from Newtonian mechanics to special
h, we went from Newtonian mechanics to quantum mechanics. When
relativity. When we turned on
we turned on both c1 and
h, we arrived at quantum field theory, one the greatest monument of
20th century physics. Newton himself had already moved up the vertical axis from Newtonian
mechanics to Newtonian gravity by turning on G. Turning on c1 , Einstein took us from that
corner to General Relativity, the main subject of this lecture. All the effort of the past few decades
is the attempt to cross from that corner to the Holy Grail of quantum gravity, when all three
fundamental constants are unified. [8]

Non-relativistic
Quantum Mechanics with Gravity Quantum
Gravity

Newtonian
Gravity Einstein
Gravity

Quantum
Quantum
Field Theory
G Mechanics


Newtonian
Mechanics c1 Einsteinian
Mechanics

Figure 1.1. The cube of physics.

Adapted from [8]

1
The Holy Grail is a vessel that serves as an important motif in Arthurian literature. Different traditions describe
it as a cup, dish or stone with miraculous powers that provide happiness, eternal youth or sustenance in infinite
abundance. Funny guy Zee!

4
1.1 Historical Introduction

4th Century (B.C) Aristotle

No action or movements takes place without a cause.
Massive objects move towards to the Earth due to their nature.
Massless objects fly up the sky.
There is no gravitation as an external force.
Gravity is intrinsic and moves the bodies to the center.
5th - 6th Century
The Indians know that Earth is sphere and attracts bodies.
9th Century
Arabs mention Earths attraction.
16th Century Copernic
Copernic stated heliocentric 2 model.
The planets move in circular trajectories, the Sun is at the center.
17th Century Kepler
Kepler investigated planetary motions.
1) Planets move in elliptical trajectories, the Sun being at one of the foci.
2) A line connecting the planet and the Sun sweeps equal areas in equal time intervals.
3) Square orbital period of a planet is proportional to the cube of the semi-major axis of its orbit.

Linear Semi-Latus Focal

Section Equation Eccentricity Eccentricity Rectum Parameter
e c l p

Circle x2 + y 2 = a2 0 0 a
s
x2 y 2 b2 b2 b2
Ellipse + 2 =1 1 a2 b2
a2 b a2 a2 a2 b2
Parabola y 2 = 4ax 1 - 2a 2a
s
x2 y 2 b2 b2 b2
Hyperbola 2 =1 1+ a2 + b2
a2 b a2 a2 a2 + b2

Table 1.1. Conic sections and their geometrical properties

17th Century Galileo

All objects accelerate the same under the gravity.
Late 17th Century Hook
The gravitational force is inversely proportional to the square of distance.
1686 Newton
2
helio- comes from the Greek root meaning the Sun

5
 M 1 M2
F (1.1)
r2

1797 Cavendish

GM1 M2
F = (1.2)
r2

where G is the gravitational constant3 , G = 6.67 1011 N m2 / kg

3
Also known as universal gravitational constant", or as Newtons constant"

6
 2 | Special Relativity and Spacetime

For the description of processes taking place in nature, one must have a system of reference. By
a system of reference we understand a system of coordinates serving to indicate the position of a
particle in space, as well as clocks fixed in this system serving to indicate the time. There exist
systems of reference in which a freely moving body, i.e. a moving body which is not acted upon by
external forces, proceeds with constant velocity. Such reference systems are said to be inertial.

If two reference systems move uniformly relative to each other, and if one of them is an inertial
system, then clearly the other is also inertial (in this system too every free motion will be linear
and uniform). In this way one can obtain arbitrarily many inertial systems of reference, moving
uniformly relative to one another. Experiment shows that the so-called principle of relativity is
valid. According to this principle all the laws of nature are identical in all inertial systems of
reference. In other words, the equations expressing the laws of nature are invariant with respect to
transformations of coordinates and time from one inertial system to another. This means that the
equation describing any law of nature, when written in terms of coordinates and time in different
inertial reference systems, has one and the same form. [6]

2.1 Galilean Relativity

The laws of physics are the same in all inertial frame of reference.
Consider Euclidean R3 space endoved with cartesian coordinates, let us look translation, rotation
and translation with constant speed(boost).

2.1.1 Translation

P (x, y, z) P 0 (x 0 , y 0 , z 0 )

x0 = x + d
y0 = y
z0 = z (2.1)

So that the position can be written as

r 0 = x 0 + y 0 + z 0 k

r 0 = (x + d) + y + z k

or, in short

r 0 = r + d = r + d (2.2)

7
Therefore, velocity is

dr 0 dx 0 dy 0 dz 0
v0 = = + + k
dt dt dt dt
dx dy dz
= + + k
dt dt dt
=v (2.3)

If the point P is static then v = 0 and since v = v 0 v 0 = 0

2.1.2 Rotation

: constant
x = x 0 cos y 0 sin
y = x 0 sin + y 0 cos
z = z0

x cos sin 0 x0

y = sin cos 0 y 0 (2.4)


z 0 0 1 z0
| {z }
O

Since the determinant of the rotation matrix in Eq. (2.4) is 1, O1 = OT . Let us verify this.

cos sin 0

OT = sin cos 0 (2.5)


0 0 1

Therefore,

cos sin 0 cos sin 0

O OT = sin cos 0 sin cos 0


0 0 1 0 0 1

cos2 + sin2 0 0

= 0 cos2 + sin2 0


0 0 1

=1 (2.6)

8

r = x + y + z k
= (x 0 cos y 0 sin ) + (x 0 sin + y 0 cos ) + z 0 k
(2.7)

We can use the same rotation matrix (2.4) for the unit vectors as well

cos sin 0 0

= sin cos 0 0 (2.8)



k 0 0 1 0
k

r 0 = x 0 (cos + sin ) + y 0 ( sin + cos ) + z 0 k

(2.9)

and the velocity is

v 0 = (vx cos + vy sin , vx sin + Vy cos , 0) (2.10)

and

dv 0
a0 = (2.11)
dt

2.1.3 Boost

x 0 = x v0 t
y0 = t
z0 = t
t0 = t (2.12)

dx dy dz
 
0
v = vo , ,
dt dt dt
= v v0

Galilean Group
1) Translation x 0 i = xi + d i where d i is constant and t = t + t0 . 4 parameters.
2) Rotation x 0 i = Ri j xj 3 parameters
3) Translation with constant velocity x 0 i = xi v i where v i is constant. 3 parameters
Therefore, there are 10 parameters that constitutes the Galilean Group.

9
2.2 Lorentz Transformations

Special relativity is a theory of the structure of spacetime, the background on which particles and
fields evolve. SR serves as a replacement for Newtonian mechanics, which also is a theory of the
structure of spacetime. In either case, we can distinguish between this basic structure and the
various dynamical laws governing specific systems: Newtonian gravity is an example of a dynami-
cal system set within the context of Newtonian mechanics, while Maxwells electromagnetism is a
dynamical system operating within the context of special relativity.

Spacetime is a four-dimensional set, with elements labeled by three dimensions of space and one of
time. An individual point in spacetime is called an event. The path of a particle is a curve trough
spacetime, a parametrized one-dimensional set of events, called the worldline. Such a description
applies equally to SR and Newtonian mechanics. I either case, it seems clear that time is treated
somewhat differently than space in particular, particles always travel forward in time, whereas
they are free to move back and forth in space.

In Newtonian mechanics, there is a basic division of spacetime into well-defined slices of all of
space at a fixed moment in time." The notion of simultaneity, when two events occur at the same
time, is unambiguously defined. Trajectories of particles will move ever forward in time, but are
otherwise unconstrained; in particular, there is no limit on the relative velocity of two such particles.

In SR the situation is dramatically altered: in particular, there is no well-defined notion of two

separated events occurring at the same time. That is not to say that spacetime is completely
structureless. Rather, at any event we can define a light cone, which is the locus of paths through
spacetime that could conceivably be taken by light rays passing through this event. The absolute
division, in Newtonian mechanics, of spacetime into unique slices of space parametrized by time, is
replaced by a rule that says that physical particles cannot travel faster than light, and consequently
move along paths that always remain inside these light cones.
ds2 = dx2 + dy 2 + dz 2

x x + dx
y y + dy
z z + dz

and in terms of these infinitesimally shifted quantities, line element can be written as

(s 0 )2 = (x + dx)2 + (y + dy)2 + (z + dz)2

(s 0 s)2 = ds2 = dx2 + dy 2 + dz 2 (2.13)

10
 Theory Position Velocity Time Acceleration

Newtonian Relative Relative Absolute Absolute

SR Relative Relative Relative Absolute
GR Relative Relative Relative Relative

Table 2.1. Comparison of Newtonian mechanics, SR, and GR

Figure 2.1. Light cone in 2D space plus a time dimension.

y y
v

O O  x x

z z
Figure 2.2. The standard configuration in Special Relativity of two frames of reference, the primed
system in motion relative to the unprimed system only along the x-axis and with speed v.

11
 y y
v

ct ct

O O x x

z z
Figure 2.3. Two frames of reference S and S 0 moving relative to each other with a velocity v have
a flash of light originate at their respective origins at time t = 0.

Since the speed of light is the same (c) in both systems, the wave front will satisfy both the following
equations

12
 3 | Derivation of Lorentz Transformation

3.1 Derivation From First Principles

The thought experiment we will use to derive the Lorentz transformations from first principles is
one of a flash of light originating at the origin of two frames of reference S and S 0 which are moving
relative to each other with a velocity v. We set up our experiment so that at time t=0 the origins
of the two frames of reference are in the same place.
Since the speed of light is the same (c) in both systems, the wave front will satisfy both the following
equations

x2 + y 2 + z 2 = (ct)2 (3.1)

x 02 + y 02 + z 02 = (ct 0 )2 (3.2)

Using Galilean transformations4 in Eq. (3.2) yields

(x vt)2 + y 2 + z 2 = (ct)2

x2 + y 2 + z 2 2xvt
| {z
+ v 2 t2} = (ct)2
not compatible
with Eq. (3.1)

The Galilean transformation does not solve our problem. Notice that the unwanted term includes
both space and time. If we stick with the reasonable assumption that y = y 0 ; z = z 0 5 , then the only
way we can avoid an unwanted term such as (2xvt + v 2 t2 ) above is to assume that t 0 is a function
of space (x) and time (t) not just the equation for x as is the case in the Galilean transformations.

So our main task is to determine how (x, t) and (x 0 , t 0 ) are related. If both coordinate systems are
inertial (that is, no relative acceleration), then a particle moving along a straight line in one system
must move along a straight line in the other. Otherwise we would introduce fictitious forces6 into
the system with a curved trajectory. Hence we require linear transformations.
The most general, linear transformation between (x, t) and (x 0 , t 0 ) can be written as

x 0 = a1 x + a2 t (3.3)

t 0 = b1 x + b2 t (3.4)

where a1 , a2 , b1 , b2 are constants that only depend on v, i.e the velocity between the coordinate
systems, and on c. Now, we will substitute these equations (Eqs. 3.3 and 3.4) in Eq. (3.2). Before
doing this, it is important to note that the origin of the primed frame (x 0 = 0) is a point that
moves with speed v as seen in the unprimed frame. Thus, its location in the unprimed frame (s) is

4
Galilean Transformations: x 0 = x vt , y 0 = y , z 0 = z , t 0 = t
5
This implies that there is no effect perpendicular to the relative motion
6
A fictitious force, also called a pseudo force or dAlembert force, is an apparent force that acts on all masses whose
motion is described using a non-inertial frame of reference, such as a rotating reference frame.

13
given by x = vt. Knowing that, we obtain

0
x = 0 = a1 x + a2 t
a2
x = t = vt
a1
a2
= v (3.5)
a1

Hence, Eq. (3.3) becomes

x 0 = a1 x + a2 t

= a1 x + (a1 v)t

x 0 = a1 (x vt) (3.6)

where we used the relation a2 = a1 v and eliminated a2 . Let us now plug this new reduced form
(Eq. 3.6) and Eq. (3.4) into Eq. (3.2) and equate distances in S and S 0 frame. Indeed, Eqs. (3.1)
and (3.2) defines the same distance.

[ a1 (x vt) ]2 + y 02 + z 02 [ c(b1 x + b2 t) ]2 = x2 + 
+y2 
+ z 2 c2 t2

(3.7)

| {z }
y 2
= +z 2


Expanding gives

a1 2 { x2 2 xvt + v 2 t2 } c2 { b1 2 x2 + 2 b1 x b2 t + b2 2 t2 } = x2 + c2 t2 (3.8)

By equating coefficients, we have

(a1 2 c2 b1 2 )x2 = 1x2 or simply a1 2 c2 b1 2 = 1 (3.9)

(a1 2 v 2 c2 b2 2 )t2 = c2 t2 or simply c2 b2 2 a1 2 v 2 = c2 (3.10)

(2a1 2 v + 2b1 b2 c2 )xt = 0 or simply b1 b2 c2 = a1 2 v (3.11)

Now we will manipulate these 3 equations to find relations between coefficients of linear transfor-
mation, i.e a1 , , a2 , b1 , b2 . Let us start with multiplying Eqs. (3.9) and (3.10)
n o n o
b1 2 c2 = a1 2 1 b2 2 c2 = c2 + a1 2 v 2

we get

b1 2 b2 2 c4 = a1 2 c2 + a1 4 v 2 c2 a1 2 v 2 (3.12)

14
and square of the Eq. (3.11) is

b1 2 b2 2 c4 = a1 4 v 2 (3.13)

Notice that LHS of the Eqs. (3.12) and (3.13) are the same, therefore

a1 2 c2 +  4 2
v c2 a1 2 v 2 =  4 2
 
a1 a1v

a1 2 (c2 v 2 ) = c2
c2 1
a1 2 = =
c2 v 2 v2
1
c2

One can obtain a1 as

1
a1 = s (3.14)
v2
1 2
c

Thus we can write a2 as

1
a2 = v p (3.15)
(1 v 2 /c2 )

Using the Eq. (3.9) that relates a1 and b1

1
b1 2 c2 = 1
(1 v 2 /c2 )
1 (1 v 2 /c2 )
b1 2 c2 =
(1 v 2 /c2 )
v2 1
b1 2 = (3.16)
c (1 v 2 /c2 )
4

Taking the negative square root we can write b1 as

v 1
b1 = 2
p (3.17)
c 1 v 2 /c2

and finally using the Eq. (3.11) that relates b1 b2 and a1

b1 b2 c2 = a1 2 v 2
b2 v 1 v2
= (3.18)
c2 1 v 2 /c2
p
1 v 2 /c2

15
b2 can be found as

1
b2 = p (3.19)
1 v 2 /c2

which is the same as a1 . By defining Lorentz factor as

1
p (3.20)
(1 v 2 /c2 )

and the coefficients of the linear transformation are

v
a1 = , a2 = v, b1 = and b2 = fl (3.21)
c2

Hence, we can write the so-called Lorentz transformations as

x 0 = (x vt)
y0 = y
(3.22)
z0 = z
vx
 
0
t = t
c2

and inverting primed and unprimed coordinates (v v) gives

x = (x 0 + vt 0 )
y = y0
z = z0 (3.23)
vx 0
 
t = t0 +
c2

Note that when v  c, i.e in the limit v/c 0, Eq. (3.23) reduces to

x 0 = x vt
y0 = y
(3.24)
z0 = z
0
t = t

So Galilean transformations are the limiting case of the Lorentz transformations when v/c 0.
The change of Lorentz factor with respect to velocity fraction is shown in Figure 3.1.

16
 10

1
=p
8 1 v 2 /c2

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
v/c

Figure 3.1. The Lorentz factor plotted against velocity as a fraction of the speed of light.

As Figure 3.1 shows, the Lorentz factor is essentially 1 (unity) until the velocity v becomes about
half of the speed of light, or about 1.5 108 m/s. Given that even our fastest space ships only
travel at a tiny fraction of the speed of light, it is not surprising that we have no direct experience
of the weird effects that a Lorentz factor deviating significantly from one produce. Of course we
see these effects in particle accelerators and cosmic ray showers, but human beings are a long way
from attaining speeds where the Lorentz factor will deviate from unity.

3.2 Landau - Lifshitz Derivation

See [6].

Three useful hyperbolic function formulae

cosh2 sinh2 = 1 (3.25)

tanh
sinh = p (3.26)
1 tanh2
1
cosh = p (3.27)
1 tanh2

The problem posed in standard configuration for a boost in the x-direction, where the primed
coordinates refer to the moving system is solved by finding a linear solution to the simpler problem

c2 t2 x2 = c2 t 02 x 02 (3.28)

17
The most general solution is, as can be verified by direct substitution using

x = x 0 cosh + ct 0 sinh , ct = x 0 sinh + ct 0 cosh (3.29)

o find the role of in the physical setting, record the origins of F progression, i.e. x 0 = 0, x = vt.
The equations become (using first x 0 = 0)

x = ct 0 sinh , ct = ct 0 cosh (3.30)

v
x v c 1
= tanh = sinh = q , cosh = q (3.31)
ct c 1 v2
1 v2
c2 c2

By introducing complex time component x4 = T = ict (see Box 3.1) we can define the transforma-
tion of time as a rotation as shown in Figure 3.2.

T=ict
T x

x= x cos T sin

Figure 3.2

Therefore, x and T can be written as

x = x 0 cos + T 0 sin (3.32)

T = x 0 sin + T 0 cos (3.33)

or, in matrix representation

x cos sin x0
= (3.34)
T sin cos T0

18
Since the determinant of the rotation matrix is 1, it is straightforward to show

x0 cos + sin x
= (3.35)
T 0 sin cos T

Therefore, one can express primed coordinates as

x 0 = x cos + T sin (3.36)

T 0 = x sin + T 0 cos (3.37)

Now, let the particle move with v such that x 0 = 0

0 = x cos + T sin (3.38)

x x v
tan = = =i (3.39)
T ict c

This expression may bring us a familiar expression with some manipulation

2
v

1 + tan2 = 1 + i (3.40)
c
v2
1 + tan2 = 1 (3.41)
c2

Here 1 + tan2 also equal to

sin2 1
1 + tan2 = 1 + = (3.42)
cos2 cos2

and therefore

1 v2
= 1
cos2 c2
1
cos = p (3.43)
1 v 2 /c2

which we have already defined as in Eq. (3.20) and knowing that

x 0 = x cos + T sin
sin
 
= cos x + T
cos
= (x + T tan )
iv
  
= x + ict
c

19
which gives us

x 0 = (x vt) (3.44)

What about T 0 ? Remember

T 0 = x sin + T cos

Again in cos brackets

sin
 
0
T = cos T x
cos
= cos (T x tan )
| {z }
iv
c
v
 
ict 0 = cos ict x i
c

and dividing both sides by ic and replacing cos yields

xv
 
0
t = t 2 (3.45)
c

However, the problem is that as v c ; cos = (1 v 2 /c2 )1 which can not be satisfied
since cos is in the interval [1, 1]. Thus, does not represent a real angle. If we want to express
as a real angle, hyperbolic functions should be used. In order to this let i. Therefore,

sin(i)
tan(i) = (3.46)
cos(i)

and using Eulers formula 7

ei(i) ei(i) e e
sin(i) = = = i sinh (3.47)
2i 2i
ei(i) + ei(i) e + e
cos(i) = = = cosh (3.48)
2 2

7
Eulers formula provides a powerful connection between analysis and trigonometry, and provides an interpretation
of the sine and cosine functions as weighted sums of the exponential function:

eix + eix
cos x = Re eix =


2
ix

eix eix
sin x = Im e =
2i
and also hyperbolic functions:
ex ex e2x 1 1 e2x
sinh x = = =
2 2ex 2ex
ex + ex e2x + 1 1 + e2x
cosh x = = =
2 2ex 2ex

20
and therefore

sin(i) i sinh
tan(i) = = = i tanh (3.49)
cos(i) cosh

cosh sinh 0 0

sinh cosh 0 0
= (3.50)



0 0 1 0

0 0 0 1

Box 3.1 | Imaginary Time [1]

One sometime participant in special relativity will have to be put to the sword: x4 = ict.
This imaginary coordinate was invented to make the geometry of spacetime look formally as
little different as possible from the geometry of Euclidean space; to make a Lorentz trans-
formation look on paper like a rotation; and to spare one the distinction that one otherwise
is forced to make between quantities with upper indices (such as the components p of the
energy-momentum vector) and quantities with lower indices (such as the components p of the
energy-momentum 1-form).

However, it is no kindness to be spared this latter distinction. Without it, one cannot know
whether a vector is meant or the very different geometric object that is a 1-form. Moreover,
there is a significant difference between an angle on which everything depends periodically
(a rotation) and a parameter the increase of which gives rise to ever-growing momentum
differences (the velocity parameter of a Lorentz transformation). If the imaginary time-
coordinate hides from view the character of the geometric object being dealt with and the
nature of the parameter in a transformation, it also does something even more serious: it hides
the completely different metric structure of (+ + +) geometry and ( + + +) geometry. In
Euclidean geometry, when the distance between two points is zero, the two points must be the
same point. In Lorentz-Minkowski geometry, when the interval between two events is zero, one
event may be on Earth and the other on a supernova in the galaxy M3l, but their separation
must be a null ray (piece of a light cone). The backward-pointing light cone at a given event
contains all the events by which that event can be influenced. The forward-pointing light cone
contains all events that it can influence. The multitude of double light cones taking off from
all the events of spacetime forms an interlocking causal structure. This structure makes the
machinery of the physical world function as it does (further comments on this structure in
Wheeler and Feynman 1945 and 1949 and in Zeeman 1964). If in a region where spacetime is

21
flat, one can hide this structure from view by writing

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

with x4 = ict, no one has discovered a way to make an imaginary coordinate work in the
general curved spacetime manifold. If x4 = ict cannot be used there, it will not be used
here. In this chapter and hereafter, as throughout the literature of general relativity, a real
time coordinate is used, x0 = t = ctconv (superscript 0 rather than 4 to avoid any possibility
of confusion with the imaginary time coordinate).

22
 4 | Week IV & V

Remember line element in 4 D spacetime is written as

Cartesian Coordinates: ds2 = (cdt)2 + dx2 + dy 2 + dz 2

Spherical Coordinates: ds2 = (cdt)2 + dr2 + r2 d2 + r2 sin2 d2 (4.1)
Cylindrical Coordinates: ds2 = (cdt)2 + dr2 + r2 d2 + dz 2

Note that expressing the metric in spherical coordinates does not mean that it describes a sphere.
In order to do so, replace r = a which is the radius of the sphere and deduce

ds2 = (cdt)2 + a2 (d2 + sin2 d2 )

Let us consider a parametrized curve with so that ():

Here vi is
() !
dxi
vi = ; i = 1, 2, 3 (4.2)
vi d

or explicitly,

dx y dy z dz
vx = ;v = ;v =
d d d
 
vi = v1, v2, v3

Figure 4.1. A curve parametrized with

Therefore, norm of v i can be written as

q
v v = |v|= (v 1 )2 + (v 2 )2 + (v 3 )2 (4.3)

The dual of v i is (v1 , v2 , v3 ) and given by vi = v j ji . For instance,

v1 = v j j1
(4.4)
= v 1 n11 + v 2 n21 + v 3 n31

and the norm is given by

v i v j ji = |v|2 (4.5)

where ij is symmetric metric and rank 2 tensor. We know that a line element in Euclidean space

23
is written as

ds2 = dx2 + dy 2 + dz 2

ds2 = ij dxi dxj (4.6)

It is trivial to deduce that ij = diag(1, 1, 1). However let us see this by writing Eq. 4.6 explicity

ds2 = 11 (dx1 )2 + 12 (dx1 dx2 ) + 13 (dx1 dx3 )

+ 21 (dx2 dx1 ) + 22 (dx2 )2 + 23 (dx2 dx3 )

+ 31 (dx3 dx1 ) + 32 (dx3 dx2 ) + 33 (dx3 )2 (4.7)

OR

ds2 = 11 (dx1 )2 + (12 + 21 ) (dx1 dx2 ) + (13 + 31 ) (dx1 dx3 )

+ 22 (dx2 )2 + (23 + 32 ) (dx2 + dx3 ) + 33 (dx3 )2 (4.8)

and using the fact that the metric should be symmetric8 , i.e ij = ji

0 0
ds2 = 11 (dx1 )2 + 2  * (dx1 dx2 ) + 2 
12
 * (dx2 dx3 ) + 22 (dx2 )2 + 33 (dx3 )2
23
 (4.9)

Note that 12 and 23 must be zero in order to obtain our pre-defined line element

ds2 = (dx1 )2 + (dx2 )2 + (dx3 )2 = dx2 + dy 2 + dz 2 (4.10)

1 0 0

ij =
0 1 0
(4.11)
0 0 1

and in spherical coordinates

1 0 0

gij = 2 (4.12)
0 r 0

0 0 r2 sin2

Similarly we can find the norm of a vector in Minkowski space as

ds2 = dx dx

8
Otherwise distance from A to B and B to A would be different.

24
where

1 0 0 0

0 1 0 0
= (4.13)


0 0 1 0

0 0 0 1

or, simply = diag(1, 1, 1, 1).

v = (v 0 , v 1 , v 2 , v 3 ) = (v t , v x , v y , v z ) (4.14)

remember line element in Minkowski space is given by

ds2 = (cdt)2 + (dx)2 + (dy)2 + (dz)2 (4.15)

dividing both sides of the equation by ds2 yields

2 2 2 2
dt dx dy dz
   
2
1 = c + + + (4.16)
ds ds ds ds

Now, let us define the proper time 9 such that

ds2 = (cd )2
(4.17)
dx 0 = dy 0 = dz 0 = 0

and parametrize the curve in spacetime with s itself as shown in Figure 4.2.

(s)

Figure 4.2. A curve parametrized with s

9
Proper time along a timelike world line is defined as the time as measured by a clock following that line. It is thus
independent of coordinates, and a Lorentz scalar. The proper time interval between two events on a world line is
the change in proper time. This interval is the quantity of interest, since proper time itself is fixed only up to an
arbitrary additive constant, namely the setting of the clock at some event along the world line. The proper time
between two events depends not only on the events but also the world line connecting them, and hence on the
motion of the clock between the events. It is expressed as an integral over the world line. An accelerated clock
will measure a smaller elapsed time between two events than that measured by a non-accelerated (inertial) clock
between the same two events. The twin paradox is an example of this effect. By contrast, coordinate time is the
time between two events as measured by an observer using that observers own method of assigning a time to an
event. In the special case of an inertial observer in special relativity, the time is measured using the observers
clock and the observers definition of simultaneity.

25
thus, 4-velocity can be written as

dt dx dy dz dx
        
v = c , , , = (4.18)
ds ds ds ds ds

and the norm of the 4-velocity is

" 2 2 2 2 #
dx dx dt dx dy dz
     

v v = = c tt + xx + yy + zz = 1 (4.19)
ds ds ds ds ds ds

We also know from Eq. (4.17) that ds2 = (cd )2 ,

2 2 2 2
dt dx dy dz
   
2
c + + + = c2 (4.20)
d d d d

Thus norm of 4-velocity (wrt to proper time ) is

v v = c2 (4.21)

where v = v . It is important to note that,

1. The minus() sign in RHS of the Eq. (4.21) comes from our metric convention (, +, +, +).

2. Norm of proper 4-velocity is always constant.

3. Norm of a vector in Euclidean space is always positive (+) or zero.

For the metric signature (, +, +, +), particles can be identified as

Spacelike : v v > 0

Timelike : v v < 0 (4.22)

Lightlike : v v = 0

Example. In Minkowski space, let us consider v = (1, 1) vector.

Norm of v is

v v = v v

= v 0 v 0 00 + v 1 v 1 11 (4.23)

=0

Therefore it is the 4-velocity of light-like particle!

Example. Consider the vector v = (1, 1) under following coordinate transformation in 1+1

26
spacetime.

u = ct + x

v = ct x

Let us start with finding x0 = ct and x1 = x in terms of the new set of coordinates which will
enable us to deduce the metric tensor.

u+v du + dv
ct = cdt =
2 2
(4.24)
uv du dv
x= dx =
2 2

Hence, line element can be written as

1 1
ds2 = (du + dv)2 + (du dv)2 = du dv (4.25)
4 4

while determining the metric we should be careful, metric tensor should be symmetric so that the
distance, for instance, from A to B equal to distance from B to A. Therefore,

0 1/2
g = (4.26)
1/2 0

and the norm of v = (1, 1) is

v v = g v v

= g00 v 0 v 0 + g12 v 1 v 2 + g21 v 2 v 1 + g11 v 1 v 1 (4.27)

= 1

Example. A particles trajectory in Minkowski space is given by the curve

= x = (x0 sinh , x0 cosh , 0, 0)

where x0 is constant and the parameter is in the interval of (, +).

a. Find the trajectory of this particle.

b. Find the 3 and 4 velocity of the particle, determine if the particles speed can be greater
than speed of light c or not. (Here c adopted as 1.)
c. Calculate proper time from = 0 to = 0 .
d. Determine whether the particle reach a phenomena observed at (t = 1, x = 0) or not.

Solution.

27
a. In this parametric form, it might be hard to see what is going to be its trajectory. Here
x0 = (c)t = x0 sinh , x1 = x0 cosh , and x2 = x3 = 0. In our case taking the square might be a
good move since we know the hyperbolic functions satisfy the relation

cosh2 x sinh2 x = 1 (4.28)

Thus, is the parametrization of

     
x2 t2 = x20 cosh2 x20 sinh2 = x20 cosh2 sinh2 = x20 (4.29)
| {z }
1

So the trajectory of the particle is

x2 t2 = x20 (4.30)

which is a hyperbola as shown in Figure 4.3.

x0 x0 x

Figure 4.3. A unit hyperbola

b. 3-velocity can be found using

dxi
vi = (4.31)
dt

In our case, we have only the x component, i.e

dx dx d
vx = = (4.32)
dt d dt

which are straight-forward to obtain as

x = x0 cosh dx = (x0 sinh ) d

(4.33)
y = x0 sinh dy = (x0 cosh ) d

28
Using these in Eq. (4.32)

(x0 sinh ) d d
vx = = tanh (4.34)
d (x0 cosh ) d

and

v = (tanh , 0, 0) (4.35)

Plot tanh is shown in Figure 8.1.

y
1
y = tanh

2 1 1 2

Figure 4.4. Plot of tanh

Notice that as tanh 1 and tanh 1. This can also be seen from

sinh e e
tanh = = (4.36)
cosh e + e

and in the limit of

 
e e2 1
lim = 1
e (e2 + 1)
  (4.37)
e 1 e2
lim = +1
+ e (1 + e2 )

4-velocity can be found using

dx
v = (4.38)
d

ds2 = (cdt)2 + dx2 (4.39)

c2 d 2 = c2 dt2 + dx2 (4.40)

29
but we choose c equal to be 1,

d 2 = dt2 + dx2 (4.41)

2 2
dt dx
 
1= (4.42)
d d

t = x0 sinh dt = (x0 cosh )d (4.43)

x = x0 cosh dx = (x0 sin )d (4.44)

(dt)2 + (dx)2 = (x20 cosh2 )d2 + (x20 sinh2 )d2 (4.45)

x20 (sinh2 cosh2 ) d2 (4.46)

| {z }
=1

= x20 d2 (4.47)

Therefore,

x20 d2 = d 2 (4.48)

|d = x0 d (4.49)

= x0 + x1 (4.50)

dx dx d
v = =
d d d
dt dx d
 
= , , 0, 0
d d d (4.51)
1
= (x0 cosh , x0 sinh , 0, 0)
x0
v = (cosh , sinh , 0, 0)

c. from = to = 0

30
 Z 0 Z 0
= d = x0 d = x0 0 (4.52)
=0 =0

4.1 Velocity Transformations

S (x, y, z) S

u (x , y  , z  ) v

u

O O

Figure 4.5. Two frames moving with a constant speed with respect to each other

= [x(t), y(t), z(t)]

(4.53)
= [x0 (t0 ), y 0 (t0 ), z 0 (t0 )]

Velocities in unprimed and primed frame are given by

dxi dx dy dz
 
u= = , , = (ux , uy , uz )
dt dt dt dt
(4.54)
dx 0i dx 0 dy 0 dz 0
 
0 0x 0y 0z

u = = , , = u ,u ,u
dt dt dt dt

From Lorentz transformations (Eq. 3.22) we know that

x 0 = (x vt) dx 0 = (dx v dt)

xv v dx
   
0 0
t = t 2 dt = dt 2
c c (4.55)
0 0
y = y dy = dy

z 0 = z dz 0 = dz

31
We can find velocities in the primed frame simply using dx 0 /dt 0 . Let us start with u 0x ,

ux
z }| {
dx 0 (dx v dt) dt( dx/dt v)
u 0x =   = (4.56)
v dx v
 
dt
dt 2 dt 1 2 dx/dt
c c

Thus,

ux v
u 0x = (4.57)
v ux
1 2
c

What about u 0y and u 0z ? Does the boost in x direction affect them? The answer is yes, because
this boost also effects the t 0 component which is same for all spatial coordinates.

dy 0 dy
u 0y = =
v dx
 
dt
dt 2
c
dy
=
v dx (4.58)
 
dt 1 2
c dt
uy
= 
v x

1 2u
c

Therefore,

uy
u 0y = (4.59)
ux v
 
1 2
c

and similarly,

uz
u 0z = (4.60)
ux v
 
1 2
c

Relation btw proper and coordinate time. Remember the line element in 4D spacetime is
given by

ds2 = (cdt)2 + dx2 + dy 2 + dz 2 = (c d )2 (4.61)

32
dividing both sides of the equation by (c d )2 yields
2 " 2 2 2 #
dt dx dy dz
  
+ + =1
d c d c d c d
" 2 2 2 #
dx dy dz
 
2
(dt) + + = (d )2
c c c

(4.62)




 2  2  2
1 dx dy dz
2 2

(d ) = (dt) 1 2 + +
c
dt dt dt



|
{z }

2

v
" #
v2
(d )2 = (dt)2 1 = (dt)2 2
c2

Hence, we obtained a very important relation

dt
d = (4.63)

OR

d = dt (4.64)

Example. Acceleration of a particle in its own rest frame is given by

 
a 0 = a 0t , a 0x = (0, g),

find the trajectory of the particle wrt the rest frame.

Solution. It is kind of a long winded problem, but our ultimate goal will be deducing the set
(x0 , x1 ) = (t, x). We are expecting to obtain a parametric equation that will allow us to determine
the trajectory of the particle. Important relations will be shown in boxed environments.

dx
a = a =
dt
(4.65)
0 0 dx 0
a =a =
dt

and a 0 is given as a 0 = (0, g). We can use the fact that norm of both a and a 0 must be equal,
since scalars are invariant under coordinate transformations. This provides us a constraint while
trying to determine the primed quantities in terms of unprimed quantities, or vice versa.

a a = a 0 a 0 (4.66)
| {z } | {z }
Scalar Scalar

33
Starting from the what is given,

2
a 0 a 0 = a 0t a 0 t + a 0x a 0 x = a 0x = g2 (4.67)

and since the norm g 2 is our invariant in this problem,

 2
a a = at at + at at = at + (ax )2 = g 2 (4.68)

hence,
q
ax = g 2 + (at )2 (4.69)

Another relation that might be useful for our current discussion is the fact that a u so that

a u = u a = 0 (4.70)

and it is well known that

u u = 1 (4.71)

as we have shown by Eq. (4.21) previously. Notice that here we adopted c = 1. Let us see this in
action,

ut ut + ux ux = 1 (4.72)

or,
 2
ut + (ux )2 = 1 (4.73)

Therefore,
q
ux = (ut )2 1 (4.74)

Now we can use Eqs. (4.70) and (4.74) to relate at and ut ,

q
t t
ua =u a = x x
(ut )2 1 ax (4.75)

by taking the square of both sides

 2  2   
2
t
u t
a = ut
1 (ax )2 (4.76)

but, we find ax in Eq. (4.69) as

q
a = x
g 2 + (at )2

34
So by simply substituting this in Eq. (4.76)
 2  2      2 
2
t t t
u a = u 1 g 2 + at
 2  2  2  2  2  2
ut at = ut g 2 + ut at g 2 at (4.77)
 2   
2
at = g2 ut 1

Thus,
q
a = g (ut )2 1
t
(4.78)

We have obtained the relation btw a and u of the same component, which are also related by
differentiation to each other

dut
q
at = = g (ut )2 1 (4.79)
d

and,

dut 1 dut
q =0 q = gd (4.80)
d (ut )2 1 (ut )2 1

in order to evaluate this, hyperbolic transformations might be useful,

ut = cosh q
(4.81)
dut = sinh q dq

which yields

sinh q dq sinh q dq
Z Z
q = = g d dq = g d (4.82)
cosh2 q 1 sinh q

Thus,

q = g + c0 (4.83)

where c0 is the integration constant. Now revert the transformation given by Eq. (4.81)
 
cosh1 ut = q = g + c0
(4.84)
ut = cosh (g + c0 )

35
We also know ux in terms of ut thanks to Eq. (4.74)
q
ux = (ut )2 1
q
ux = cosh2 (g + c0 ) 1 (4.85)

ux = sinh (g + c0 )

We can also find x0 = t by simply integrating Eq. (4.84)

dt
ut = = cosh (g + c0 )
d
Z Z
dt = cosh (g + c0 )d (4.86)
1
x0 = t = sinh (g + c0 )
g

and by similar integration of Eq. (4.85)

dx
ux = = sinh (g + c0 )
d
Z Z
dx = sinh (g + c0 )d (4.87)
1
x1 = x = cosh (g + c0 )
g

with these our equations we reached our goal at the beginning of the problem, obtaining a parametrized
curve that allow us to the determine the trajectory

1
x= cosh (g + c0 )
g
1
t = sinh (g + c0 )
g

Again using the relation cosh2 q sinh2 q = 1

1
x2 t2 = (4.88)
g2

which is a hyperbola as shown in Figure 4.6

36
 t

x
1 1

g g

Figure 4.6

37
 5 | Tensor Calculus and Riemannian Geometry

To work effectively in Newtonian theory, one really needs the language of vectors. This language,
first of all, is more succinct, since it summarizes a set of three equations in one. Moreover, the
formalism -of vectors helps to solve certain problems more readily, and, most important of all,
the language reveals structure and thereby offers insight. In exactly the same way, in relativity
theory, one needs the language of tensors. Again, the language helps to summarize sets of equations
succinctly and to solve problems more readily, and it reveals structure in the equations. [3]
A Riemannian space is endoved with non-singular10 (0,2) type metric tensor and line element is
given by

ds2 = g dx dx (5.1)

5.1 Vectors

If A ( = 0, 1, . . . , n 1) transforms as coordinate differential then A is called a contravariant

vector.

x 0
A = A
x 0
(5.2)
x 0
A 0 = A
x

If A ( = 0, 1, . . . , n 1) transforms as gradient of a coordinate then A is called a covariant

vector.

x
A 0 = A
x 0
(5.3)
x 0 0
A = A
x

Properties of vectors.

1. Sum or difference of two same kind vectors is also a vector of same kind.

A B = C (5.4)

In order to show this consider

x 0
A = A
x0
(5.5)
x 0
B = B
x0

10
det g = g 6= 0 and g = g

38
Therefore,

x 0 0 x 0
A B = (A B ) = C
x0 | {z } x 0 (5.6)
C 0

OR

x0 0
A = A
x
(5.7)
x0 0
B = B
x

Hence,
0
x0 0 0 x 0
A B = (A B ) = C
x | {z } x (5.8)
C 0

2. Product of two same kind vectors are 2nd rank same kind tensor.

A B = C
(5.9)
A B = C

x 0 x

0 x x 0 0 x x 0
(A ) (B ) = (A B ) = C (5.10)
x0 x0 x0 x0 | {z } x0 x0
C 0

3. Product of two different kind vectors are 2nd rank mixed tensor.
0
x 0 x0 0 x x 0 0 x x0 0
   
A B = A B = A B = C
x0 x x0 x x 0 x
x 0 x0 0 x x0 0 0 (5.11)
  

A B = A B = A B = A 0 B 0
x0 x x0 x
| {z }

Matrix A is symmetric if

AT = A
(5.12)
(Aij )T =Aji = Aji

Matrix A is anti-symmetric if

AT = A
(5.13)
(Aij )T = Aji

Any n n symmetric matrix has n(n + 1)/2 independent components.

39
Any n n anti-symmetric matrix has n(n 1)/2 independent components.

In n dimensional space,

x x 0
(5.14)
x 0 = f (x )

The Jacobian generalizes to any number of dimensions, so we get, reverting to our primed and
unprimed coordinates,
 0 
x 01 x 01 x 1

 
x1 x2 xn
 
 0 0 
x 02 x 2 x 2

 
x1 x2 xn
 
J (x) =  .. .. .. ..

 (5.15)

 . . . . 

x 0n x 0n x 0n

 
x1 x2 xn
 

or, in short
 
 x0 
J(x) =   (5.16)
 
 x 

and similarly for the primed coordinates,

 
x1 x1 x1

 
x 01 x 02 x 0n
 
 
x2 x2 x2
0


 x 01 x 02
x 0n


J x =  .. .. .. ..
 (5.17)
.


 . . . 

xn xn xn

 
x 01 x 02 x 0n
 

or, in short
 
0
 x 
J(x ) =  0  (5.18)
x

A worked example and further discussion is presented in the appendix of this section (see Section
5.5).

5.2 Tensors

To work effectively in Newtonian theory, one really needs the language of vectors. This language,
first of all, is more succinct, since it summarizes a set of three equations in one. Moreover, the
formalism -of vectors helps to solve certain problems more readily, and, most important of all, the
language reveals structure and thereby offers insight. In exactly the same way, in relativity theory,
one needs the language of tensors. Again, the language helps to summarize sets of equations suc-
cinctly and to solve problems more readily, and it reveals structure in the equations. [3]

40
A tensor is an object defined on a geometric entry called a (differential) manifold. In simple terms,
a manifold is something which locally looks like a bit of n-dimensional Euclidean space Rn .

Box 5.2 | A tensor is something that transforms like a tensor [8]

Long ago, an undergrad who later became a distinguished condensed matter physicist came to
me after a class on group theory and asked me, What exactly is a tensor? I told him that a
tensor is something that transforms like a tensor. When I ran into him many years later, he
regaled me with the following story. At his graduation, his father, perhaps still smarting from
the hefty sum he had paid to the prestigious private university his son attended, asked him
what was the most memorable piece of knowledge he acquired during his four years in college.
He replied, A tensor is something that transforms like a tensor.

But this should not perplex us. A duck is something that quacks like a duck. Mathematical
objects could also be defined by their behavior. We already saw in the preceding chapter that
a vector is defined by how it transforms: V 0i = Rij V j Consider a collection of mathematical
entities T ij with i, j = 1, 2, . . . , D in D-dimensional space. If they transform under rotations
according to

T ij T 0ij = Rik Rjl T kl (5.19)

then we say that T transforms like a tensor, and hence is a tensor. Indeed, we see that we are
just generalizing the transformation law of a vector.

41
Definition. A quantity S1 1 ... r transforms as
0 0 0
0 x 1 x 2 x r
S1 2 ... r = S 1 2 ... r . . . (5.20)
x1 x2 xr

under coordinate transformations, then S1 1 ... r is called rcovariant tensor.

Definition. A quantity S 1 1 ... r transforms as

x1 x2 xr
S 1 2 ... r = S 01 2 ... r . . . (5.21)
x 0 1 x 0 2 x 0 r

under coordinate transformations, then S1 1 ... r is called rcontravariant tensor.

...
Definition. A quantity S11 22 ... qp transforms as
0
01 2 ... p x1 x2 xp x01 x02 x p
S11 22 ...
... p
q = S 1 2 ... q . . . . . . (5.22)
x01 x02 x0q x1 x2 xq
...
under coordinate transformations, then S11 22 ... qp is called r rank tensor where r = p + q. This
tensor can also be represented as Type (p, q) tensor, where p represents contravariants and q rep-
resents covariants.

Definition. A quantity transforms as

= J p
(5.23)

under coordinate transformations, then has weight11 p.

Remember last weeks quiz:

quizi
Example. Kronecker delta is defined by ekle

1, if a = b xa

ab OR a b (5.24)
0,

if a 6= b xb

and similarly in the primed frame

x0a
0a b = (5.25)
x0b

11
In differential geometry, a tensor density or relative tensor is a generalization of the tensor concept. A tensor
density transforms as a tensor when passing from one coordinate system to another, except that it is additionally
multiplied or weighted by a power p of the Jacobian determinant of the coordinate transition function or its absolute
value.

42
Let us check whether a b is a tensor or not.

x x0c x0d
a
a
b = 0c

0d xb
x |x
{z }
0c d
xa x0d
= 0c d (5.26)
x0c x0c

Thus, a b is a Type (1, 1) OR rank 2 tensor.

Example. Show that F = A A is an anti-symmetric tensor.

By changing the indices, i.e ,

F = A A

= ( A A )

= F (5.27)

Since F = F , it is an anti-symmetric tensor.

Example. Show that g F = 0 where g is a symmetric tensor and F is an anti-symmetric

tensor.

By changing the indices, i.e ,

g F = g F = g (F ) = (g F ) (5.28)

Since there is a summation over indices results will be a scalar, and the only scalar that is equal to
its negative is 0.

Example. Find how an anti-symmetric F tensor changes under coordinate transformations.

A A
F =

x x

Similarly, in the primed frame

A 0 A 0
F 0 = 0 (5.29)
x x 0

43
We know that a covariant vector transforms like

x
A 0 = A
x 0

using this in Eq. (5.29) we obtain

x x
   
0
F = 0 A 0 A 0
x x x 0 x
x x x x
     
= A 0 A 0 (5.30)
x 0 x x x 0 x x

and by distributing derivatives, dzgn

szck
A x x 2 x A x x 2 x bul
F 0 = 0 0 + A 0 0 A
x x x x 0 x 0 x x x x 0 x 0
x x A A x x
= 0 0 Change: in the 2nd term
x x x x x 0 x 0
x x A A
 
=
x 0 x 0 x x
| {z }
F
x x
= F = F 0 (5.31)
x 0 x 0

Thus,

x x
F 0 = F (5.32)
x 0 x 0

transforms like Type (0, 2) covariant tensor.

Example. Show that F + F + F 0 where F = A A .

i iii i iii
z }| { z }| { z }| { z }| {
A A + A A + A A 0 (5.33)

| {z } | {z }
ii ii

where we used the fact that A = A .

Example. Show that Christoffel symbols are symmetric with respect to 2 lower indices, i.e
= .

Remember, Christoffel symbols are defined as

1
= g ( g + g g ) (5.34)
2

44
by making the change of lower indices

1
= g ( g + g g ) (5.35)
2

which is identically equal to Eq. (5.34).

5.3 Symmetric and Anti-symmetric Tensors

Let S be a set and A an abelian group12 . Given a map : S S A, is termed a symmetric

map if (s, t) = (t, s) for all s, t S.

The symmetrization of a map.

: S S A is the map (x, y) 7 (x, y) + (y, x)

The anti-symmetrization of a map.13

: S S A is the map (x, y) 7 (x, y) (y, x)

The sum of the symmetrization and the anti-symmetrization of a map is 2. Thus, away from
2, meaning if 2 is invertible, such as for the real numbers, one can divide by 2 and express every
function as a sum of a symmetric function and an anti-symmetric function. The symmetrization
of a symmetric map is its double, while the symmetrization of an alternating map is zero; sim-
ilarly, the anti-symmetrization of a symmetric map is zero, while the anti-symmetrization of an
anti-symmetric map is its double.

Let us assume that C is neither symmetric nor anti-symmetric tensor. We can obtain totally
symmetric / anti-symmetric tensor by using the symmetrization / anti-symmetrization.

Example.

0 1 0 2
C = & (C )T = C = (5.36)
2 3 1 3

Let us just sum them up and divide by two (factorial),

1 1 0 3
C() = (C + C ) = (5.37)
2! 2 3 3

12
An abelian group, also called a commutative group, is a group in which the result of applying the group operation
to two group elements does not depend on the order in which they are written. That is, these are the groups that
obey the axiom of commutativity. Abelian groups generalize the arithmetic of addition of integers, i.e closure,
associativity, identity element, inverse element, commutativity
13
Sometimes called skew-symmetrization

45
which is an symmetric matrix. Conversely, for the anti-symmetrization

1
C[] = (C C ) (5.38)
2!

For instance, C[] = 0. For rank 3,

1
C[] = (C C + C C + C C ) (5.39)
3!

is totally anti-symmetric.

1
C[] = (C C + C C + C C ) = 0 (5.40)
3!

5.4 Raising and Lowering Indices

Remember,

V =V & V = V dx (5.41)
x

Example.

V g = V

V g = V

V g = V (5.42)

and

V g = V

V g = V

V g = V (5.43)

Another specific examples are

V g = V = V (5.44)

and

g g = 1nn If there is no summation (5.45)

g g = = g11 g 11 + . . . gnn g nn = n (5.46)

46
5.5 Appendix - Coordinate Transformation and Jacobian

Adapted from Physics Pages - Coordinate Transformations

Reference: dInverno, Ray, Introducing Einsteins Relativity (1992), Oxford Uni Press. Section
5.4 and Problem 5.2.
Since one of the main aspects of the definition of a tensor is the way it transforms under a change
in coordinate systems, its important to consider how such coordinate changes work.
Well consider two coordinate systems, one denoted by unprimed symbols xi and the other by
primed symbols x 0i . In general, one system is a function of the other one, so we can write

x 0i = x 0i (x) (5.47)

where the index i runs over the n dimensions of the manifold (so we have a set of n equations),
and the symbol x without an index means the set of all components of x, so its equivalent to (but
shorter than) writing
 
x 0i = x 0i x1 , x2 , x3 , . . . , xn (5.48)

Now suppose we want to do an integral over a portion of the manifold that is bounded by some
subsurface in the manifold. As we know from elementary calculus, the differential volume (or
area) has a different form depending on which coordinate system were using. For example, in
3-d rectangular coordinates, the volume element is dxdydz, while in spherical coordinates it is
r2 sin drdd.
To see how this works we can start with one dimension. If we have an integral in rectangular
coordinates such as
Z x2
f (x)dx (5.49)
x1

dx
we can change coordinate systems if we define x = x(u). Then we have dx = du du. To transform
the limits of the integral, we need to invert the definition to get u = u (x). Then the integral
becomes
Z u(x2 )
dx
f (x (u) du (5.50)
u(x1 ) du

Essentially, this redefines the line element into the u coordinate system.
In two dimensions, wed start off with (well leave out the limits on the integrals since were really
interested only in the area element):
Z Z
f (x, y)dxdy (5.51)

47
Now if we want to switch to another coordinate system, we define

u = u (x, y) (5.52)

v = v (x, y) (5.53)

Consider now an elemental rectangle R in the xy plane. The rectangle has its lower left corner
at the point (x0 , y0 ) and has dimensions x and y, so that its area is xy. We want to see
how this rectangle transforms under the coordinate transformation above. The new elemental area
will not necessarily be a rectangle, but we can transform it point by point to get the new shape.
Starting with the lower left corner, this transforms to

(u0 , v0 ) = [u (x0 , y0 ) , v (x0 , y0 )] (5.54)

We can write the general transformation as a vector:

r (x, y) = u (x, y) i + v (x, y) j (5.55)

Here, r is the transformed location of the original point (x, y), written with respect to the rectan-
gular basis vectors.
The idea now is to consider what happens as x and y tend to zero. In this case, the transformed
version of R tends to a parallelogram whose sides are parallel to the transformation of the two sides
of R that touch at the point P0 = (x0 , y0 ) (the lower left corner of R we mentioned above). Consider
first the edge of R along the line y = y0 (the bottom of the rectangle). We can think of this edge
as a tangent to the rectangle at the point P0 . How does this tangent transform?
Well, the lower edge of R transforms as

r (x, y0 ) = u (x, y0 ) i + v (x, y0 ) j (5.56)

The tangent along this curve is then the derivative with respect to x, so we get

r (x, y0 ) = u (x, y0 ) i + v (x, y0 ) j (5.57)
x x x

Thus the tangent along the bottom edge of R at the transformed location of P0 is
  
 
i + v (x, y0 )

j
rx r (x, y0 ) = u (x, y0 ) (5.58)
x x=x0 x x=x0 x 
x=x0

By the same argument, the tangent at P0 along the left edge of R is found by setting x = x0 and
differentiating with respect to y, and we get
  
 
i + v (x0 , y)

j
ry r (x0 , y) = u (x0 , y) (5.59)
y y=y0 y y=y0 y 
y=y0

48
By the definition of a derivative, we can write these tangents in the form

r (x0 + x, y0 ) r (x0 , y0 )
rx = lim (5.60)
x0 x
r (x0 , y0 + y) r (x0 , y0 )
ry = lim (5.61)
y0 y

The vector r (x0 + x, y0 ) r (x0 , y0 ) connects the transformed lower left corner of R to the trans-
formed lower right corner. Similarly r (x0 , y0 + y) r (x0 , y0 ) connects the lower left corner to the
upper left corner. Thus these two vectors define the sides of a parallelogram that, for very small
x and y, is a good approximation to the transformed R. In this approximation, we can write

r (x0 + x, y0 ) r (x0 , y0 ) ' rx x (5.62)

r (x0 , y0 + y) r (x0 , y0 ) ' ry y (5.63)

The area of a parallelogram is A = s1 s2 sin , where s1 and s2 are two adjacent sides and is
the angle between them. If we have two vectors corresponding to the sides, the area is thus the
magnitude of the cross product of the vectors. So we get

A = |rx ry | xy (5.64)

Using the equations above, we can work out this cross product. Well use the notation uy u/y
to save space. We get
 

 i j 
k

 
rx ry =  ux vx 0  = (ux vy vx uy ) k (5.65)
 
 uy vy 0 

is itself a 2 2 determinant, and can be written as

The coefficient of k
 
 ux uy 
 
J(x, y)  (5.66)
 vx vy 


This is called the Jacobian of the transformation. The area element is thus

dA = J (x, y) dxdy (5.67)

Now this is all very well, but the differentials x and y are still in the original coordinate system.
How can we use this result to transform the integral that we began with? The trick is to assume
that the transformation is invertible, that is, that we can also write

x = x (u, v) (5.68)

y = y (u, v) (5.69)

49
We can run through the same argument again to get

dA = J (u, v) dudv (5.70)

with
 
 xu xv 
 
J (u, v) =  (5.71)
 yu yv 


That is:
Z Z Z Z
f (x, y)dxdy = f [x (u, v) , y (u, v)] |J (u, v)| dudv (5.72)

Note that weve taken the absolute value of J since were dealing with an area element, which must
be positive. It can also be shown that (the proof would make this post too long) the Jacobian
satisfies a very convenient property:

1
J (u, v) = (5.73)
J (x, y)

That is, the Jacobian of an inverse transformation is the reciprocal of the Jacobian of the original
transformation.
The Jacobian generalizes to any number of dimensions, so we get, reverting to our primed and
unprimed coordinates:
 
x1 x1 x1

 
x 01 x 02 x 0n
 
 
x2 x2 x2
0


 x 01 x 02
x 0n


J x =  .. .. .. ..
 (5.74)
.


 . . . 

xn xn xn

 
x 01 x 02 x 0n
 

For obvious reasons, this can be abbreviated to

 xa 
 
J =  0b 
 (5.75)
x

As a simple example, consider the transformation from rectangular to polar coordinates in 2-d.
From the above, the Jacobian we want is J (r, ) which requires expressing the old coordinates in
terms of the new ones. The transformation is

x = r cos (5.76)

y = r sin (5.77)

50
So we have
 
 cos r sin 
 
J (r, ) =  =r (5.78)
 sin r cos 

Thus the transformation of the area element is

dxdy rdrd (5.79)

For the inverse transformation, we have

q
r= x2 + y 2 (5.80)
y
= tan1 (5.81)
x

so
 
y
 2x 2
 
x +y x2 +y 2
 1 1
J (x, y) =  y/x = p = (5.82)
 
2 1/x

 1+(y/x)2
 2
x +y 2 r
1+(y/x)2 

Thus J (u, v) = 1/J (x, y) as required. In 3D,

x = r sin cos (5.83)

y = r sin sin (5.84)

z = r cos (5.85)

 
 sin cos r cos cos r sin sin 
 
 = r 2 sin
 
J (r, , ) =  sin sin r cos sin r sin cos  (5.86)
 
 cos r sin 0 

For the inverse:

q
r= x2 + y 2 + z 2 (5.87)
p
x2 + y 2
= tan1 (5.88)
z
y
= tan1 (5.89)
x

 
 x y z 
2 +y 2 +z 2 2 +y 2 +z 2 x2 +y 2 +z 2
 
 x x


    
 x/ z x2 +y2 y/ z x2 +y 2
J (x, y, z) =  x2 +y 2 /z 2 (5.90)


 1+(x2 +y2 )/z 2 1+(x2 +y 2 )/z 2 1+(x2 +y 2 )/z 2 
y/x2
 
1/x
0
 
1+(y/x)2 1+(y/x)2
 

51
Converting back to spherical coordinates proves a bit easier. Substituting the above transformation
equations, along with

r2 sin2 = x2 + y 2 (5.91)

helps to simplify things.

 
 x y z 
 r r r 
yz sin
J (x, y, z) =  r3 xz
 
sin r3 sin r

 (5.92)
 2 y2 x
 
r sin r2 sin2
0 

The determinant now comes out to

y sin y z yz x sin x z xz
   
J (x, y, z) = 2 2 2 2 (5.93)
r sin r r r r3 sin r sin r r r r3 sin
! !!!
1 2 z2 2 z2
= 4 2 y sin + 2 + x sin + 2 (5.94)
r sin r sin r sin
!
x2 + y 2 z2
= 4 2 sin + 2 (5.95)
r sin r sin
!
1 z2
= 2 sin + 2 (5.96)
r r sin
!
1 r2 sin2 + z 2
= 2 (5.97)
r r2 sin
!
1 x2 + y 2 + z 2
= 2 (5.98)
r r2 sin
1
= (5.99)
r2 sin

52
 6 | Parallel Displacement and Covariant Derivative

The covariant derivative is a way of specifying a derivative along tangent vectors of a manifold.
Alternatively, the covariant derivative is a way of introducing and working with a connection on
a manifold by means of a differential operator, to be contrasted with the approach given by a
principal connection on the frame bundle. In the special case of a manifold isometrically embedded
into a higher-dimensional Euclidean space, the covariant derivative can be viewed as the orthogonal
projection of the Euclidean derivative along a tangent vector onto the manifolds tangent space.
In this case the Euclidean derivative is broken into two parts, the extrinsic normal component and
the intrinsic covariant derivative component.

The covariant derivative is a generalization of the directional derivative from vector calculus. As
with the directional derivative, the covariant derivative is a rule, u v , which takes as its inputs:
(1) a vector, u, defined at a point P , and (2) a vector field, v, defined in a neighborhood of P . The
output is the vector u v(P ), also at the point P . The primary difference from the usual directional
derivative is that u v must, in a certain precise sense, be independent of the manner in which it
is expressed in a coordinate system.

A vector may be described as a list of numbers in terms of a basis, but as a geometrical object a
vector retains its own identity regardless of how one chooses to describe it in a basis. This persistence
of identity is reflected in the fact that when a vector is written in one basis, and then the basis is
changed, the components of the vector transform according to a change of basis formula. Such a
transformation law is known as a covariant transformation. The covariant derivative is required to
transform, under a change in coordinates, in the same way as a basis does: the covariant derivative
must change by a covariant transformation (hence the name).

Box 6.3 | Historical Remark

At the turn of the 20th century, the covariant derivative was introduced by G. Ricci and T.
Levi-Civita in the theory of Riemannian and pseudo-Riemannian geometry. Ricci and Levi-
Civita (following the ideas of E. B. Christoffel) observed that the Christoffel symbols used to
define the curvature could also provide a notion of differentiation which generalized the classi-
cal directional derivative of vector fields on a manifold. This new derivative the Levi-Civita
connection was covariant in the sense that it satisfied Riemanns requirement that objects in
geometry should be independent of their description in a particular coordinate system.

It was soon noted by other mathematicians, prominent among these being H. Weyl, J. A.
Schouten, and E. Cartan, that a covariant derivative could be defined abstractly without the
presence of a metric. The crucial feature was not a particular dependence on the metric, but
that the Christoffel symbols satisfied a certain precise second order transformation law. This
transformation law could serve as a starting point for defining the derivative in a covariant

53
 manner. Thus the theory of covariant differentiation forked off from the strictly Riemannian
context to include a wider range of possible geometries.

Example.
For a scalar field , covariant differentiation is simply partial differentiation

;a a (6.1)

For a contravariant vector field a ,

a ;b b a + a bc c (6.2)

For a covariant vector field a ,

a;c c a b ca b (6.3)

For a type (2,0) tensor field ab , we have:

ab ;c c ab + a cd db + b cd ad (6.4)

For a type (0,2) tensor field ab , we have:

ab;c c ab d ca db d cb ad (6.5)

For a type (1,1) tensor field a b , we have:

a b;c c a b + a cd d b d cb a d (6.6)

The notation above is meant in the sense

ab ;c (ec )ab (6.7)

One must always remember that covariant derivatives do not commute, i.e. a;bc 6= a;cb . It is
actually easy to show that:

a;bc a;cb = Rd abc d (6.8)

where Rd abc is the Riemann tensor. Similarly,

a ;bc a ;cb = Ra dbc d (6.9)

and

ab ;cd ab ;dc = Ra ecd eb Rb ecd ae (6.10)

54
Example. Show that
A A = A A

where A is a arbitrary vector field.

RHS can be written as

A A = A A ( A A )
(6.11)
= A A

Thus,

A A = A A (6.12)

Example. Show that

F + F + F = 0

where F = A A .

Performing covariant differentiation on F yields

F + F + F = F F F
| {z } | {z }

+ F F F (6.13)
| {z } | {z }

+ F F F
| {z } | {z }

where we used the fact that F = F and = . Therefore,

F + F + F = 0 (6.14)

and

F + F + F = 0 (6.15)

6.1 Covariant Derivative of Metric Tensor

Metric Compatibility. Given a metric tensor gab , a covariant derivative is said to be compatible
with the metric if the following condition is satisfied14

c gab = 0
14
Although other covariant derivatives may be supported within the metric, usually one only ever considers the
metric-compatible one. This is because given two covariant derivatives, and 0 , there exists a tensor for
transforming from one to the other
a xb = 0a xb Cab c xc
If the space is also torsion-free, then the tensor Cab c is symmetric in its first two indices.

55
Consider the following covariant differentiations,

c gab = c gab d ca gdb d cb gad = 0 (6.16)

a gbc = a gbc d ab gdc d ac gbd = 0 (6.17)

b gca = b gca d cb gda d ba gcd = 0 (6.18)

Let us try to recover the definition of Christoffel symbols from these differentiations. Invert the
sign of the Eq. (6.18), i.e
b gca + d cb gda + d ba gcd = 0

and sum them up

c gab d ca gdb d cb gad +a gbc d ab gdc

| {z } | {z } | {z }
(6.19)
d ac gbd b gca + d cb gda + d ba gcd = 0
| {z } | {z } | {z }

after some algebra, d ca can be obtained as

2d ca gdb = c gab + a gbc b gca

1
d ca = (gdb )1 (c gab + a gbc b gca ) (6.20)
2
1
d ca = g db (c gab + a gbc b gca )
2

6.2 Riemann Curvature Tensor

6.2.1 Symmetry Properties of Riemann Tensor

56
7 | Einstein Field Equations

7.1 Schwarzschild Solution

7.2 Conformal Tensors and Conformal Curvature Tensor

Conformal transformations are frequently used tools in order to study relations between various
theories of gravity and the Einstein relativity. In this paper we discuss the rules of these transfor-
mations for geometric quantities as well as for the matter energy-momentum tensor [9].

57
 8 | Homeworks

8.1 Geometrized Units

Question:
In the geometrized units c and G is taken to be equal to 1. Express,
h, and lp in geometrized units. Here lp is the Planck length:
1 year, 175 cm,
s
Gh
lp =
c3

Second in terms of meter:

c = 1 = 3 108 m/s

1 s = 3 108 m (8.1)

Kilogram in terms of meter:

G = 1 = 6.67 1011 m3 / kg s2

In order to eliminate the second and determine kg in terms of meter, we can use Eq. (8.1) directly.

1 = 6.67 1011 m3 / kg (3 108 m)2

= 0.74 1027 m kg1

Therefore,

1 kg = 0.74 1027 m (8.2)

All other conversion factors can be worked out by combining these two.

8.1.1 Solution

i.
1 year = 365 days = 365 24 hours = 3.154 107 seconds

Using the conversion as shown in Eq. (8.1),

1 year = 107 seconds 3 108 m

= 9.462 1015 m (8.3)

58
ii.

Since it is geometrized units

175 cm = 175 cm (8.4)

iii.

h = 1.05 1034 m2 kg s1

Again using our two fundamental conversion factor, i.e Eqs. (8.1) and (8.2), h can be written as

h = 1.05 1034 m2 (0.74 1027 m) (3 108 m)1

which is equal to

h = 2.59 1070 m2 (8.5)

Now, we are ready to express lp in geometrized units. Starting from

s
Gh
lp = (8.6)
c3

but c = G = 1, and using the Eq. (8.5)

s
1 2.59 1070 m2
lp =
13
lp = 1.61 1035 m (8.7)

59
8.2 Relativistic Velocity Addition

Question:
A cart rolls on a long table with velocity v. A smaller cart rolls on the first cart in the same
direction with velocity v relative to the first cart. A third cart rolls on the second cart in the same
direction with relative velocity v, and so on up to n carts. What is the velocity v of the n-th cart
in the frame of the table? What does vn tend to as n ?

In order to solve this problem, we should discuss the rapidity parameter.

8.2.1 Rapidity Parameter

A feature of the velocity addition formula is that if we combine two velocities less than the speed
of light, we always get a result that is still less than the speed of light. This means that no amount
of combining velocities can take you beyond light speed. Sometimes it is more convenient to talk
about in terms of the rapidity r, which is defined by the relation

v = c tanh (r/c) (8.8)

The hyperbolic tangent function tanh maps the real line from to + onto the interval (1, +1)
(See Figure 8.1). Thus, while velocity v can only vary between c and c, the rapidity r varies over
all real values. At small speeds rapidity and velocity are approximately equal. If s is also the
rapidity corresponding to velocity u, then the rapidity t of the combined velocities is given by the
simple addition

t=r+s (8.9)

This follows from the identity of hyperbolic tangents

tanh x + tanh y
tanh (x + y) = (8.10)
1 + tanh x tanh y

Rapidity is therefore useful when dealing with combined velocities in the same direction, and also
for solving problems with linear acceleration.
For instance, if we combine the speed v n times, the result is
h i
= c tanh n tanh1 (v/c) (8.11)

60
 y
1
y = tanh x

x
2 1 1 2

Figure 8.1. Plot of tanh x

8.2.2 Solution

As discussed above, the rapidity parameter adds linearly. Let us denote the rapidity of the addition
of velocities up to n-th cart with r. Therefore,

r = r1 + r2 + + rn (8.12)
| {z }
n

but since the relative velocities of each cart are equal, rapidities are also equal, i.e r1 = r2 = = rn .
Thus,

r = n r1 (8.13)

In order to find rapidity in terms of the velocities, let us go back to the definition:

vn = c tanh (r/c)

vn /c = tanh (r/c)

r = c tanh1 (vn /c) (8.14)

Now using the Eq. (8.13)

1

c tanh (vn /c) = n
c tanh1 (v/c) (8.15)

and by taking the tanh of both sides

vn h i
= tanh n tanh1 (v/c) (8.16)
c
n 1 + v/c
  
= tanh ln (8.17)
2 1 v/c
" n/2 #
1 + v/c

= tanh ln (8.18)
1 v/c

61
where we used the expansion

1 1+x
 
1
tanh x = ln (8.19)
2 1x

in Eq. (8.17). Another expansion that we shall use is

1 e2x
tanh x = . (8.20)
1 + e2x

Therefore, Eq. (8.18) becomes

" n/2 #
1 + v/c

1 exp 2 ln
vn 1 v/c
= " n/2 #
c 1 + v/c

1 + exp 2 ln
1 v/c
" n #
1 + v/c

1 exp ln
1 v/c
= " n #
1 + v/c

1 + exp ln
1 v/c

1 + v/c n 1 v/c n
   
1 1
1 v/c 1 + v/c
= n = (8.21)
1 v/c n
 
1 + v/c

1+ 1+
1 v/c 1 + v/c

and finally vn can be writtten as

1 [(1 v/c) / (1 + v/c)]n

vn = c (8.22)
1 + [(1 v/c) / (1 + v/c)]n

Notice that as n the terms in square brackets in Eq. (8.22) goes zero and the equation itself
goes to c 1 = c which is the speed of light. Thus,

n vn c (8.23)

8.3 Acceleration in Special Relativity

This homework is mostly based on Lecture Notes on SR by Stephen Siklos.

It is often said that SR can NOT deal with acceleration because it is only valid in inertial frames,
and therefore acceleration must be the interest of GR. We are only allowed to make transformations
between inertial frames and these frames should NOT accelerate, but of course, the observers in
the frame can move. SR can deal with anything kinematic but GR is required when gravitational
forces are present. However, since the amount of spacetime curvature is not particularly high on
Earth or its vicinity, SR remains valid for most practical purposes, such as experiments in particle

62
accelerators.

Accelerations in SR follow, as in Newtonian Mechanics, by differentiation of velocity with respect

to time. Because of the Lorentz transformation and time dilation, the concepts of time and distance
become more complex, which also leads to more complex definitions of acceleration.

8.3.1 Solution

We can write the acceleration in S 0 frame as

du 0 du 0 d
a0 = = 0 (8.24)
dt 0 d |{z}
|{z} dt
i ii

Let us look convenient forms for (i) and (ii).

Remember

uv
u0 = (8.25)
1 uv/c2

Differentiating wrt to yields

 
du uv
du 0 d 1 c2
(u v) (v/c2 ) du
d
= 2 (8.26)
d

uv
1 c2

In du/d paranthesis

du 0 (1 uv/c2 ) (uv/c2 + v 2 /c2 ) du

= 2 (8.27)
d d

1 uv c2

Therefore, (i) in Eq. (8.24) can be found as

v2
du 0 1
= c2  du (8.28)
d uv 2 d
1 2
c

and Eq. (8.24) becomes

1 v 2 /c2 du d
a0 = (8.29)
(1 uv/c2 )2 d dt 0

Now let us look for (ii). We know that

vx
t 0 = (t ) (8.30)
c2

63
differentiating wrt to t

dt 0 vu
 
= 1 2 (8.31)
dt c

OR

vu
 
0
dt = 1 2 dt (8.32)
c

Hence,

1 v 2 /c2 du d
a0 =



2 2 d
 (1 uv/c2 )2 dt
(1 uv/c ) 
1 v 2 /c2 1 du
=
2 2 2 2 dt
(1 uv/c ) (1 uv/c )

Rearranging yields

3/2
1 v 2 /c2
0
a = a (8.33)
(1 uv/c2 )3

OR

a
a0 = (8.34)
3 (1 uv/c2 )3

8.4 Transformation Law of Christoffel Symbols

Question:
Show that Christoffel symbols do Not transform tensorially.

1
= g ( g + g g )
2

and the transformation law is

0
" #
x x x 2 x
= 0 0 +
x x x x 0 x 0

8.4.1 Solution

In the primed frame using the definition given above, Christoffel symbols can be written as

1

= g 0 g 0 + g 0 g 0 (8.35)
2 | {z } | {z } | {z }
i ii iii

64
We know that the metric tensor transforms as

x x
g 0 x 0 =


g (8.36)
x 0 x 0

Therefore, the inverse of the metric transforms as

0 0
x x
g 0 x 0 =


g (8.37)
x x

which is the first term that we are going to use in Eq. (8.35). Now let us look for that how i, ii,
and iii transforms.

i.
!
0 x x
0 g = g
x 0 x 0 x 0
!
x x x x x
= 0 0 0 g + g 0 (8.38)
x x x x x x 0 x 0

where we used the transformation rule of the metric, i.e Eq. (8.36). Here, another important
x
thing is bringing an unprimed differentiation by the chain rule, i.e x
. It is kind of an educated
guess since we know that transformation of Christoffel symbols bring an another Christoffel plus
an inhomogeneous term (see Eqs. (8.41), (8.42), and (8.43)). Similarly,

ii.
!
0 x x
0 g = g
x 0 x 0 x 0
!
x x x x x
= 0 0 0 g + g 0 (8.39)
x x x x x x 0 x 0

iii.

!
x x
0 g 0 = g
x 0 x 0 x 0
!
x x x x x
= g + g 0 (8.40)
x 0 x 0 x 0 x x x 0 x 0

So we have found everything that we need, now let us plug these results into Eq. (8.35).

= 1
g 0
0 0 0
g + g g

2 | {z } | {z } | {z } | {z }
Eq. (8.37) Eq. (8.38) Eq. (8.39) Eq. (8.40)

65
Thus,

0 0
!
1 x x x x x x x
= g 0 0 0 g +g 0
2
x x x x x x


| {z } x x 0 x 0

!
x x x x x
+ g +g 0

x 0 x 0 x 0 |x{z } x x 0 x 0



!
x x x x x
0 0 0 g +g 0 (8.41)

x x x |x{z } x x 0 x 0

Note that *, **, and *** with 12 g in Eq. (8.41) constitutes a Christoffel symbol as follows

1
= g ( g + g g ) (8.42)
2

Another important thing is that these *, **, and *** terms all have the same Jacobian factor.
Knowing that Eq. (8.41) can be written as

0 0
z }| {
x x x x x 1
=

g [ g + g g ]
x x x 0 x 0 x 0 2
" ! !
1 x x x x
+ g g 0 + g 0
2 x x 0 x 0 x x 0 x 0
!# )
x x
g 0 (8.43)
x x 0 x 0

Or

0 0
( " ! !
x x x x x x x
= x x
0 0 0 +
1
g g 0

+ g 0
x x x x x 2 x x 0 x 0 x x 0 x 0
!# )
x x
g 0 (8.44)
x x 0 x 0

In Eq. (8.44) we have an object that transforms like a tensor and plus a inhomogeneous term.

66
Let us focus on this term, by distributing derivatives we get
( !
1 2 x x x 2 x
g g 0 + 0
2 x 0 x 0 x x x 0 x 0
!
2 x x x 2 x
+ g 0 + 0
x 0 x 0 x x x 0 x 0
!)
2 x x x 2 x
g + (8.45)
x 0 x 0 x 0 x 0 x 0 x 0

We can always change the dummy indices and relabel them anyway we like for convenience. Here
in the 2nd step , and in the 3rd step . Note that we relabeled the indices like
this, so that 2nd and 3rd terms become compatible with the 1st term. With these new indices we
have

1 2 x x x 2 x
g g 0 0 0 + 0 0 0
2 |x x {z x } |x x {z x }


2 x x x 2 x
+ g 0 0 0 + 0 0 0

|x x{z x } |x x
{z x }

2 x x x 2 x
g 0 0 0 + 0 0

0
(8.46)
|x x
{z x } |x x{z x }

after grouping the terms, we are left with

" # " #
1 x 2 x x 2 x
g g 2 0 = 0
2 | {z } x x 0 x 0 x x 0 x 0

x 2 x
= (8.47)
x 0 x 0 x 0

Notice that we have summed two terms labeled by green braces in Eq. (8.46) which have different
indices in the numerator. We are able to this since the metric tensor is symmetric by definition,
i.e g = g . Thus, here and are interchangeable. Let us go back to Eq. (8.44) for the final
result.

0 0
x x 0 0
x x x
 2 x

= x x x


 x 0 x 0 x 0 +
 x 0 x 0 x 0

x x x x
0 0
x x x x 2 x
= 0 0 + (8.48)
x x x x x 0 x 0

67
Therefore, by grouping the like terms, the final result can be obtained as

0
" #
x x x 2 x
= 0 0 + (8.49)
x x x x 0 x 0

We have shown that Christoffel symbols do Not transform tensorially!

8.5 Static and Spherically Symmetric EFE

Question:
Find the corresponding Christoffel symbols, Riemann tensor, Ricci tensor & scalar components
of the metric

ds2 = e2(t,r) dt2 + e2(t,r) dr2 + r2 d2 + r2 sin2 d2

Christoffel symbols are given by

1
= g ( g + g g ) (8.50)
2

Components of Riemann tensor are given by

R = + (8.51)

The contraction of Riemann tensor gives the Ricci tensor as

R = R (8.52)

and one more contraction yields Ricci scalar as

R = g R (8.53)

8.5.1 Christoffel Symbols

The non-zero Christoffel symbols are

t tt :

1
t tt = 0 00 = g 00 (t g00 + t g00 t g00 )
2

Therefore,

t tt = 0 00 = t (8.54)

68
t tr :

1
t tr = 0 01 = g 00 (t g01 + r g00 t g01 )
2

Thus,

t tr = 0 01 = r (8.55)

t rr :

1
t rr = 0 11 = g 00 (r g01 + r g10 t g11 )
2

Hence,

t rr = 0 11 = e2() t (8.56)

r tt :

1
r tt = 1 00 = g 11 (t g01 + t g10 r g00 )
2

Therefore,

r tt = 1 00 = e2() r (8.57)

r tt :

1
r tr = 1 01 = g 11 (t g11 + r g01 r g01 )
2

Thus,

r tr = 1 01 = t (8.58)

r rr :

1
r rr = 1 11 = g 11 (r g11 + r g11 r g11 )
2

Hence,

r rr = 1 11 = r (8.59)

r :

1
r = 2 12 = g 22 (r g22 + g12 g12 )
2

69
Therefore,

1
r = 2 12 = (8.60)
r

r :

1
r = 1 22 = g 11 ( g12 + g21 r g22 )
2

Therefore,

r = 1 22 = r e2 (8.61)

r :

1
r = 3 13 = g 33 (r g33 + g13 g13 )
2

Hence,

1
r = 3 13 = (8.62)
r

r :

1
r = 1 33 = g 11 ( g13 + g31 r g33 )
2

Hence,

r = 1 33 = r sin2 e2 (8.63)

1
= 2 33 = g 22 ( g23 + g32 g33 )
2

Therefore,

= 2 33 = sin cos (8.64)

1
= 3 23 = g 33 ( g33 + g22 g23 )
2

70
Therefore,

cos
= 3 23 = (8.65)
sin

8.5.2 Riemann Tensor

R = + (8.66)

where is the repeated index that will be summed over. Rt rtr :

Rt rtr = t t rr r t rt rt t r + rr t r

Therefore,
n o h i
Rt rtr = R0 101 = e2() t2 + (t )2 t t + r r r2 (r )2 (8.67)

Rt t :

Rt t = t t t t t t + t

Therefore,

Rt t = R0 202 = re2 r (8.68)

Rt t :

Rt t = t t t t t t + t

Therefore,

Rt t = R0 303 = re2 sin2 r (8.69)

Rt r :

Rt r = r t t r r t + t r

Hence,

Rt r = R0 212 = re2 t (8.70)

71
Rt r :

Rt r = r t t r r t + t r

Hence,

Rt r = R0 313 = re2 sin2 t (8.71)

Rr r :

Rr r = r r r r r r + r r

Thus,

Rr r = re2 r (8.72)

Rr r :

Rr r = r r r r r r + r r

Thus,

Rr r = re2 sin2 r (8.73)

R :

R = +

Thus,
 
R = 1 e2 sin2 (8.74)

8.5.3 Ricci Tensor & Scalar

Rtt :

Rtt = Rr trt + R tt + R tt

1
 n o (8.75)
2()
=e r2 2
+ (r ) r r + 2r + (t )2 + t2 t t
r

72
Rrr :

Rrr = Rt rtr + R rr + R rr
n o 
1
 (8.76)
= e2() t2 + (t )2 t t + r r r2 (r )2 + 2 r
r

Rtr :

Rtr = R tr + R tr
1 (8.77)
= 2 t
r

R :

R = Rt t + Rr r + R
(8.78)
= e2 {r(r r ) 1} + 1

R :

R = Rt t + Rr r + R

= e2 sin2 {r(r r ) 1} + sin2 (8.79)

= sin2 R

Therefore, Ricci scalar for the metric is

R = g R = g 00 R00 + g 11 R11 + g 22 R22 + g 33 R33

2 1 
 
2
R = 2e r2 2
+ (r ) r r + (r r ) + 2 1 e 2
(8.80)
r r

73
References

[1] C. W. Misner, K. S. Thorne, and J. A. Wheeler, Gravitation, W. H. Freeman, 1973.

[2] J. B. Hartle, Gravity: An Introduction to Einsteins General Relativity, Addison Wesley 2003

[3] R. dInverno, Introducting Einsteins Relativity, Oxford, 1992

[4] S.Carroll, Spacetime and Geometry: Introduction to General Relativity, Addison-Wesley, 2004

[5] B.F. Schutz, A First Course in General Relativity, Cambridge University Press, 2009

[6] L.D. Landau, E.M Lifshitz, The Classical Theory of Fields, 1973

[7] O. Grn, Einsteins General Theory of Relativity: With modern applications in cosmology,
2007

[8] A. Zee, Einstein Gravity in a Nutshell, Princeton University Press, 2013

[9] M. P. Dabrowski, J. Garecki, D. B. Blaschke, Conformal transformations and conformal in-

variance in gravitation, 2008, arXiv:gr-qc/0806.2683v3

74