3 Theory of Relativity

In this section we will discuss how to treat gravity. But first, lets discuss the Special Theory of Relativity

3.1 Special Theory of Relativity (STR)

The Galilean Relativity states the laws of motion are same in all inertial frames. Maxwell theory of light, in
which the speed of light is constant in vacuum, do not refer to any particular frame of reference. Therefore,
the Maxwell’s theory is in direct conflict with the Galilean relativity. For example, the speed of light is
constant irrespective of the relative velocity between the observer and the source. So either the Maxwell’s
theory of light is wrong or the Galilean transformations are not the correct transformations. In 1902-1905
Michaelson and Morley through an experiment found that the speed of light in inertial frames is indeed
independent of the relative velocity between source and observer. Taking the result of the Michaelson and
Morley experiment at face value, Einstein proposed a new theory, the Special Theory of Relativity, to make
sense of all these. The STR is based on the following two postulates by Einstein

• laws of physics have the same form (invariant) in all inertial frames
• for all inertial observers, the speed of light in vacuum is a constant c and is independent of the relative
velocity with the source of the light
Einstein’s task was to find a set of transformation rules between inertial frames that satisfy the above two
statements. Such transformations were already worked out by Hendrik Lorentz.

3.1.1 Lorentz Transformations

With reference to the figure 8, consider two frames S ⌘ (x, y, z) and S 0 ⌘ (x0 , y 0 , z 0 ), such that the axes of
the two frames are parallel. Assume that at t = 0 the origins of the two frames coincide. At some other
time the relative velocity between the frames is v along the x-axis. In the S frame, an event is observed
to takes place at time t and location (x, y, z). In the S 0 frame the same event is observed to take place at
time t0 at the location (x0 , y 0 , z 0 ) . The Lorentz Transformation (LT) relations between the space and time
coordinates are
vx x
t0 = t = t ,
cc c
x0 = x vt = x ct ,
0
y = y, z0 = z

where
1
= v/c , =p .
1 2

Throughout the course we will work in the natural unit ~ = 1, c = 1. With this choice, the LT reads

t0 = t x ,
0
x = x t , (16)
0 0
y = y, z =z

and
1
= v, =p . (17)
1 2

A set of reverse transformation (from S 0 ! S ) also exists – they can be obtained from the above set
by interchanging the primed and unprimed coordinates, and changing the direction of the velocity. For
the above set of relations, we have assumed the velocity along the x direction. For velocity in arbitrary
directions the relations are a little more complicated, but we do not need them for our discussions.

13
 Figure 8: Two inertial frames with relative speed v.

For convenience, we introduce the following notations:

x0 ⌘ t, x1 ⌘ x, x2 ⌘ y, x3 ⌘ z
and, similarly for x01 , x02 , x03 , x04 . So the LT relations can be written as
0 00 1 0 1 0 01
x 0 0 x
Bx01 C B 0 0 C Bx1 C
B 02 C = B CB C , (18)
@x A @ 0 0 1 0A @x2 A
x03 0 0 0 1 x3

Defining the two column vectors by x0µ ⌘ (x00 , x01 , x02 , x03 )T x⌫ ⌘ (x0 , x1 , x2 , x3 )T , the matrix by ⇤µ⌫ ,
and using the Einstein summation convention (repeated indices are summed over) we can write the above
relation as
X3
x0µ = ⇤µ⌫ x⌫ = ⇤µ⌫ x⌫ , µ, ⌫ = 0, 1, 2, 3 (19)
⌫=0

Here µ, ⌫ are called the Lorentz indices. Comparing the preceding equations we find ⇤00 = , ⇤01 = ,
etc.
One crucial observation is to be made here – the space and the time coordinates are no longer separate
entities – they transform together as single entity. It is called the spacetime. In analogy with the location of
a point in three dimensional space, lets introduce the concept of an event, defined as a set of four coordinates
xµ ⌘ (t, x, y, z) in the 4-dimensional spactime. The xµ is a vector in 4D spacetime that transforms like
(19) under a LT. It is noting but a location of an event in 4D spacetime. And it is analogous to specifying
the location of a point in a 3D space. Therefore, xµ is called a 4-vector
xµ = (x0 , x1 , x2 , x3 ) = (x0 , ~x) ,
where ~x is the usual three vector. Conventionally, the x0 is called the time component and x1 , x2 , x3 are
the space components.

3.1.2 4-vectors
The concept of 4vector extends beyond the location of an event in spacetime. Any four component object
~ that transforms like (19)
Aµ ⌘ (A0 , A1 , A2 , A3 ) = (A0 , A)

A0µ = ⇤µ⌫ A⌫ , (20)

is called a four-vector. Remember that any set of four objects do not constitute a 4vector unless they
transform like the above.
In analogy with xµ , the 0th component A0 is called the time component and the other three components
are called the space components. In addition to xµ another 4vector that you should remember by heart is
the four momentum of a particle denoted as pµ . The time component of pµ is its energy E, and the space
components are just the components of its ordinary three-momentum p~. So
pµ = (E, p~) . (21)

14
 p
If m is the mass of the particle then E = p~2 + m2 . Hence the squared of the 4momentum is

p µ pµ = E 2 p~2 = m2 .

For massless particle, like the photon,

pµ = (~!, ~~k) or = (!, ~k) in natural unit (22)

where ! is the angular frequency, and k = 2⇡/ is the wave vector. In this case

pµ pµ = 0

3.1.3 Minkowski metric

The distance between twop nearby points, (x, y, z) and (x + dx, y + dy, z + dz), in a three-dimensional
Cartesian coordinate is dx2 + dy 2 + dz 2 . The distance remains same (or invariant) when measured from
a di↵erent coordinate system that is translated, or rotated about any axis. This motivates us to look for
an expression of distance between two events, (x0 , x1 , x2 , x3 ) and (x0 + dx0 , x1 + dx2 , x2 + dx2 , x3 + dx3 ),
in the spacetime that remains invariant under the LT. Consider the quantity

ds2 = (dx0 )2 + (dx1 )2 + (dx2 )2 + (dx3 )2 .

measured from a coordinate system S (you can refer to the previous figure.) Suppose, we perform a LT of
the coordinate system to S 0 and the events are x0µ and x0µ + dx0µ . Let say the quantity ds2 is transformed
to (ds0 )2
ds02 = (dx00 )2 + (dx01 )2 + (dx02 )2 + (dx03 )2 .

Using LT relation dx0µ = ⇤µ⌫ dx⌫ we can show (try to do it yourself – will be roughly done in the class)

ds02 = ds2 ,

So ds2 = (dx0 )2 + (dx1 )2 + (dx2 )2 + (dx3 )2 is an invariant quantity. The ds is called spacetime interval,
or line element, or metric. This is analogous to the concept of length in a 3D space.
With the aid of the summation convention, the spacetime interval can be written in a compact form as

ds2 = (dx0 )2 + (dx1 )2 + (dx2 )2 + (dx3 )2 = ⌘µ⌫ dxµ dx⌫ , (23)

where in the last line we have introduced the Minkowski metric or simply the metric of spacetime
0 1
1 0 0 0
B0 1 0 0C
⌘µ⌫ =B
@0
C . (24)
0 1 0A
0 0 0 1

3.1.4 Definition of LT
For any 4-vector Aµ we can define ⌘↵⌫ A⌫ A↵ as the equivalent of length2 (see (23)). Since this is a scalar
it is invariant under the LT, i.e., A⌫ ⌘↵⌫ A↵ = A0µ ⌘ µ A0 . Equating and simplifying

A⌫ ⌘↵⌫ A↵ = A0µ ⌘ µA
0

= ⇤µ⌫ A⌫ ⌘ µ ⇤ ↵ A↵ ,

hence
⇤µ⌫ ⌘ µ⇤ ↵ = ⌘↵⌫ . (25)

15
Any matrix ⇤ that satisfies this condition is LT matrix. In fact, it defines Lorentz transformation.
The rotation of a 3-dimension coordinate system, for example the rotation of Cartesian coordinate in
the x y plane constitutes a LT (easy to prove, try!)
0 1
1 0 0 0
B0 cos ✓ sin ✓ 0C
⇤µ⌫ = B@0
C, (26)
sin ✓ cos ✓ 0A
0 0 0 1
where ✓ is the angle of rotation. Another example of LT is boost, which is thought rotations between space
and time directions” 0 1
cosh sinh 0 0
B sinh cosh 0 0C
⇤µ⌫ = B
@ 0
C. (27)
0 1 0A
0 0 0 1

3.1.5 Contravariant and covariant vectors

~ is A.
The dot product or scalar product of a Cartesian 3-vector A ~ A.
~ The analog of a scalar product for a
4-vector Aµ is ⌘µ⌫ Aµ A⌫ . By defining
Aµ ⌘ ⌘µ⌫ A⌫ (28)
we can express the scalar ⌘µ⌫ Aµ A⌫ as

⌘µ⌫ Aµ A⌫ = (A0 )2 + (A1 )2 + (A2 )2 + (A3 )2 = Aµ Aµ . (29)

The Aµ Aµ form is analogous to A.~ A.

~
The vector with the upper indices is called contravariant 4-vector: Aµ = (A0 , A1 , A2 , A3 ), and the one
with the lower indices Aµ = (A0 , A1 , A2 , A3 ) is called the covarint 4-vector. Note that we can write the
components of a covariant vector in terms of the components of contravarint 4-vector as

Aµ ⌘ ⌘µ⌫ A⌫ = ( A0 , A1 , A2 , A3 ) .

The time component of contravariant vector is negative of the time component of its covariant form. For
two di↵erent 4-vectors Aµ and B µ we can write scalar product as

⌘µ⌫ Aµ B ⌫ = A0 B 0 + A1 B 1 + A2 B 2 + A3 B 3 = ~B
A0 B 0 + A. ~ = Aµ B µ . (30)

Lets define the LT of covariant vector as

A0µ = ⇤µ⌫ A⌫ . (31)

It then follows from the invariance of the scalar Aµ Aµ

A0µ A0µ = ⇤µ⌫ A⌫ ⇤µ A = A⌫ A⌫

which implies that

⇤µ⌫ ⇤µ = ⌫ (32)

Corresponding to each covariant vector there exist a contravariant vector and vice-versa. But there are
some vectors that appear more often in the contravariant form and others in covariant. For example, dxµ
is more often contravariant, and the gradiant operator
@
@µ ⌘ (33)
@xµ
is more often covariant. The transformation rule of the gradiant operator is
@ @x⌫ @
= ,
@x0µ @x0µ @x⌫

16
As x0⌫ = ⇤⌫µ xµ , it follows that
@x⌫
= ⇤µ⌫ , (34)
@x0µ
hence,
@ @
@µ0 = 0µ
= ⇤µ⌫ ⌫ = ⇤µ⌫ @µ . (35)
@x @x
Equivalently,
@ 0µ = ⇤µ⌫ @ ⌫ . (36)
It is easy to see that the divergences @µ Aµ and @ µ Aµ are invariant. Another useful invariant operator is
the d’Alembertian, the product of the gradiant with itself

@2
⇤2 = ⌘ ↵ @↵ @ = + r2 . (37)
@t2

3.1.6 Raising and lowering indices

The metric ⌘µ⌫ lowers the index of a covarint vector and makes it contravarint. The inverse of ⌘µ⌫ is ⌘ µ⌫
defined as
⌘ µ⌫ ⌘⌫⇢ = ⇢µ , µ
⇢ = 1 for µ = ⇢, or 0 otherwise (38)
⌘ µ⌫ can be used to raise the index of a contravarint vector

⌘ µ⌫ A⌫ = Aµ . (39)

From the diagonal expression of ⌘µ⌫ it is obvious that

⌘µ⌫ = ⌘ µ⌫ . (40)

3.1.7 Light-cone
The trajectories of particles in 4D spacetime are called world lines. The world line of light is special.
Consider the line element

ds2 = (dx0 )2 + (dx1 )2 + (dx2 )2 + (dx3 )2 ⌘ (dx0 )2 + d~x2 .

The wavefront of light has velocity c = |d~x/dx0 | = 1 (since we have c = 1) which means the worldline of
light corresponds to
ds2 = 0 . (41)
One can draw the world line of a light ray in a space vs time plot. Since it is impossible to plot in four
dimension, we will resort to a two dimensional plot where the horizontal axis corresponds to space and the
vertical axis is time. The world line of a light ray makes 45o angle with respect to the space axis as shown
in 9. This is also called the lightlike world line. In 4-dimension this would actually be a cone, hence 9 is
called the light cone. For ordinary particles |dx/dt| < 1, therefore, the world lines for any massive particles
makes larger angle with the space axis.
Due to the relative negative sign between the time and space components, ds2 in eq.(23) can be of three
types

• timlike, ds2 < 0: Two events separated by ds2 < 0 are causally connected, i.e., information can travel
from one event to the other. An example is two events taking place at the same spatial position but
at di↵erent times.
• spacelike, ds2 > 0: Spacelike event are causally disconnected. Example would be two events are
simultaneous in time but separated in space.

• null or lightlike, ds2 = 0: Causally connected and lie in the world line of light.

17
 Figure 9: Light cone and your world line.

3.2 Proper time

It is annoying that the timelike worldline of an ordinary particle is negative. It is customary to define a
quantity
(d⌧ )2 = (ds)2 = ⌘µ⌫ dxµ dx⌫ . (42)

Suppose there is a clock attached to the rest frame of an observer. The observer observes two consecutive
ticks in the clock, t1 = 0 and t2 = 0 + dt. Since the clock is at rest dx1 = 0, dx2 = 0, dx3 = 0, and
dx0 = t2 t1 = dt. So, the interval between the two ticks is

(d⌧ )2 = (ds)2 = (dx0 )2 = dt2 , d⌧ = dt , (43)

i.e., d⌧ is the measure of the ellapsed time dt in the clock of the observer. For this reason, d⌧ is known as
the proper time of an observer. Proper time is invariant, i.e., d⌧ 0 = d⌧ .
Exercise: There is a clock in the rest frame of an observer O. The time interval between two consicutive
ticks on the clock measured by the observer is dt. Another observer O0 is moving with a velocity v with
respect to the O’s frame. Show using the invariance of proper time, that the observer O0 would notice that
the clock is moving slower, as if the time is dialated.
Exercise: The length contraction exercise.

3.2.1 Realativistic Kinematics

In three-dimension, the position of a particle x is given as a function of time t. In four-dimension this is no
longer possible since the time itself is a coordinate. One therefore describes the motion with respect to a
parameter which varies along the world line of the particle. The natural choice is the proper time ⌧ . Thus
the world line is described by xµ = xµ (⌧ ).
We now define four velocity U µ by
dxµ
Uµ ⌘ . (44)
d⌧
From the definition of proper time (42) (simple, work it out)
p
d⌧ = dt (1 2) . (45)

18
Then we can write the components as
dt 1
U0 = p =p =
1 2 dt 1 2
(46)
dxi 1 dxi
Ui = p =p = .
1 2 dt 1 2 dt

The norm of the 4velocity can be easily shown to be U µ Uµ = 1 (this would have been equal to c2 had
we not put c = 1). The negative norm of the four velocity reflects the fact that it is defined for timelike
world lines.
Exercise: Work out the relativistic addition of velocities.
The four acceleration can be defined as
dU µ
aµ ⌘ , (47)
d⌧
and the relativistic four-force is
dU µ
f µ = maµ = m . (48)
d⌧
In an inertial frame, for speed of a particle that is much smaller than 1 (in other words, speed of light),
this boils down to Newton’s laws. The constant m defined here is called the rest mass of the particle. The
four acceleration is perpendicular to the four velocity. It can be easily shown that

aµ U µ = 0 (49)

The momentum of a massive particle can be defined as

dxµ
pµ ⌘ m = mU µ . (50)
d⌧
Using this definition we can write

pµ p µ = m2 U µ U µ = m2
) E 2 = (p0 )2 = |~
p|2 + m2 (51)

This is the famous relativistic energy-momentum relation. Here we have the energy E = p0 . It is easy to
see that in the limit of small velocities, beta ⌧ 1, and ' 1 + 2 /2, p0 . and pi are
1
p0 ' m + mv 2 , pi = mv i .
2
The second term of p0 is Newtonian kinetic energy and pi are the ordinary momentum. This is why pµ
is called the energy-momentum vector. Note that the energy has two components m called the rest mass
energy and the rest is the kinetic energy. For particles at rest the only energy is the rest mass energy

E = m, (52)

which is the most famous equation in science (we have taken c = 1, otherwise E = mc2 ). Expression of
only the relativistic kinetic energy of a particle is obtained by subtracting the rest mass energy as
p
T = p2 + m m . (53)

For a massless particle like the photon, the four-momentum vector is given by

pµ = (~!, ~~k) , (54)

where ! is the frequency, and ~k is the wave vector, and we have purposefully kept the Planck constant ~.
For photon, p2 = pµ pµ = 0.

19