2.

Special Relativity 11

determination of the constant k appearing in it would find its values to be

c, within experimental error.

2.1 Lorentz Invariance

The Special Theory of Relativity, which holds in the absence of gravity,
plays a central role in physics. Even in the strongest gravitational fields
the laws of physics must conform to it in a sufficiently small locality of any
spacetime event. That was a fundamental insight of Einstein. Consequently,
the Special Theory plays a central role in the development of the General
Theory of Relativity and its applications.

2.1.1 LORENTZ TRANSFORMATIONS

The Lorentz transformation leaves invariant the proper time or differential
interval in Minkowski spacetime

(2.4)

as measured by observers in frames moving with constant relative veloc-

ity (called inertial frames because they move freely under the action of
no forces). The Minkowski manifold also implies an absolute spacetime in
which spacetime events that can be connected by a Lorentz transformation
lie within the cone defined by dr == o. "Absolute" means unaffected by any
physical conditions. This was the same criticism that Einstein made of New-
ton's space and time, and the one that powered his search for a new theory
in which the expression of physical laws does not depend on the frame of ref-
erence, but, nevertheless, in which Lorentz invariance would remain a local
property of spacetime. ( Local in this context, means "in a sufficiently small
region of spacetime".) We will develop General Relativity, which extends
the relativity principle to arbitrary frames in a gravity-endowed universe,
not just un-accelerated frames in relative uniform motion. Here we review
briefly the Special Theory.
A pure Lorentz transformation is one without spatial rotation, while a
general Lorentz transformation is the product of a rotation in space and a
pure Lorentz transformation. The pure Lorentz transformation is sometimes
also referred to as a boost. For convenience, define

(2.5)

(In spacetime a point, such as that above, is sometimes referred to as an

event.) The linear homogeneous transformation connecting two reference
frames can be written

(2.6)
12 Special and General Relativity

(We shall use the convenient notation introduced by Einstein whereby re-
peated indices are summed-Greek over time and space, Roman over space
alone.)
Any set of four quantities AfL (J-t == 0,1,2,3) that transforms under a
change of reference frame in the same way as the coordinates is called a
contravariant Lorentz four-vector,

(2.7)
The invariant interval (also variously called the proper time, the line
element, or the separation formula) can be written

(2.8)
where TJfLv is the Minkowski metric which, in rectilinear coordinates, is

1 0 o 0
o -1 o 0
(2.9)
TJfLV == o 0 -1 0
o 0 o -1

The condition of the invariance of dT2 is

(2.10)

Since this holds for any dx", dx f3 we conclude that the AfLv must satisfy the
fundamental relationship assuring invariance of the proper time:

(2.11)

Transformations that leave dT2 invariant also leave the speed of light the
same in all inertial systems; this is so because if dr == 0 in one system, it is
true in all, and the content of dr == 0 is that dx]dt == 1 (with units c == 1).
Let us find the transformation matrix AfL for the special case of a boost
Q

along the x-axis. In this case it is clear that

(2.12)

and, moreover, that x/o and x/ 1 cannot involve x 2 and x 3 . So,

x/o == AOoxo + A01 X 1

x/ 1 == A 1 O+ A 1 1 ,
0x 1x (2.13)

with the remaining A elements zero. So the above quadratic form in A yields
the three equations,

1 (A0)2 - (A10)2
-1 (A01 ) 2 - (A11 ) 2 (2.14)
o AOoAo1 - A10A 11
 2. Special Relativity 13

To get a fourth equation, suppose that the origins of the two frames in
uniform motion coincide at t == 0 and the primed x-axis, x,l, is moving
along xl with velocity v. That is, xl == vt is the equation of the primed
origin as it moves along the unprimed x-axis. The equation for the primed
coordinate is
(2.15)

or
AIO== -AIIV. (2.16)

The four equations can now be solved with the result,

Ao0-- Al1 -- 'VI
AIO== AOl == -V1 (2.17)
A22 ==A33 == 1,
where
1 == (1 - V
2)-1/2
== coshB, V1 == sinhB, v == tanhB. (2.18)

So
X'D == xO cosh B- xl sinh B
X,l == -x o sinh B + xl cosh B (2.19)
x,2 == x 2,x,3 == x 3 .

The combination of two boosts in the same direction, say VI and V2, corre-
sponds to B == Bl + B2 A boost in an arbitrary direction with the primed
axis having velocity v == (vI, v 2 , v 3 ) relative to the unprimed is
AOo == "Y
AO..J == ; == - .v j 1 (2.20)
AJ k == Akj == 8~ + (1- l)v jv k /y2 .
For a spatial rotation, say in the x-y plane, the transformation for a
positive rotation about the common z-axis is
x,l == Xl cosw + x 2 sinw
x,2 == -Xl sinw + x 2 cosw (2.21)
x'D == z", x,3 == x 3 .

Transformation of vectors according to either of the above, or a product

of them, preserves the invariance of the interval dT2 For convenience they
can be written in matrix form as
cosh B - sinhB 0 0
- sinh B coshB 0 0
A== o 010
(2.22)
o 001
14 Special and General Relativity

1 0 0 0
0 cosw sm ce 0
R== (2.23)
0 -sinw cosw 0
0 0 0 1

2.1.2 TIME DILATION

Let there be two identical clocks at rest at x == O. The clocks tick at equal
intervals dt. The propertime in this frame is

dr == J dt 2 - dx 2 == dt . (2.24)

An observer takes one of the clocks and moves away along the x-axis with
velocity v. He sees the first clock at dx' == -v dt . The propertime expressed
in his frame is

dr' == J dt,2 - dx,2 == dt' ~ (2.25)

But the two events are connected by a Lorentz transformation and so their
propertimes are the same. Consequently

dt' ==dt/~. (2.26)

This expresses the dilation of time as measured by observers in uniform

relative motion. The converse holds for length measurements.

2.1.3 COVARIANT VECTORS

Two contravariant Lorentz vectors such as

(2.27)

and BfL may be used to create a scalar product (Lorentz scalar)

A' . B' == TJfLvA'fL B'v == TJfLV AfLaAV/3AaB/3 == TJa/3A aB/3 == A . B . (2.28)

Because of the minus signs in the Minkowski metric we have

A . B == A 0 B O - A . B , (2.29)

and the covariant Lorentz vector is defined by

(2.30)

A covariant Lorentz vector is obtained from its contravariant dual by the

process of lowering indices with the metric tensor,

(2.31)
 2. Special Relativity 15

Conversely, raising of indices is achieved by

(2.32)

It is straightforward to show that

(2.33)

where

6/1 ==
v
{I if J-L == v
0 otherwise
(2.34)

is the Kroneker delta. It follows that

(2.35)

The Lorentz transformation for a covariant vector is written in analogy

with that of a contravariant vector:

(2.36)

To obtain the elements of A; we write the above in two different ways,

(2.37)

This holds for arbitrary A/1 so

A/1v -- TJ/1Q AQf3TJ f3v . (2.38)

Using (2.33) in the above we get the inverse relationship

(2.39)
Multiplying (2.38) by A/1a , summing on J-L, and employing the fundamen-
tal condition of invariance of the proper time (2.11) we find

(2.40)

We can now invert (2.6) and find that A; is the inverse Lorentz transfor-
mation,

(2.41)

The elements of the inverse transformation are given in terms of (2.17)

or (2.20) by (2.38). We have

Aaa == All == 1 ,
Ala == Aa1 == v1 , (2.42)
A 22 = A 33 = 1.
16 Special and General Relativity

A boost in an arbitrary direction with the primed axis having velocity

v == (VI, V 2, v 3 ) relative to the unprimed is

Aoo == 1,
Aoj == A.o == vj 1, (2.43)
J . .
Ajk == AkJ == 6~ + (1 - l)v jv k /v2 .

The velocity is a vector of particular interest and defined as

dxJ-t
uJ-t == dT . (2.44)

Because dr is an invariant scalar and dx!' is a vector, uJ-t is obviously a

contravariant vector. From the expression for the invariant interval we have

dr
dT==~dt, v== - (2.45)
dt
with r == (Xl, x 2 , x 3 ) ; it therefore follows that
dt
u O == - == 1, (2.46)
dT
or

uJ-t 1(1, vI, v 2 , v 3 ) , UJ-t == 1(1, -VI, -v 2 , -v 3 ) ,

uJ-tuJ-t 1. (2.47)
The transformation of a tensor under a Lorentz transformation follows
from (2.7) and (2.36) according to the position of the indices; for example,
(2.48)
We note that according to (2.11), the Minkowski metric TjJ-tv is a tensor;
moreover, it has the same constant form in every Lorentz frame.

2.1.4 ENERGY AND MOMENTUM

The relativistic analogue of Newton's law F == ma is

d 2xJ-t
FJ-t==m-- (2.49)
dT
and the four-momentum is
dxJ-t
pJ-t == m dT . (2.50)

Hence, from (2.44) and (2.45)

pO == E == m1
p == mrtv . (2.51)
 2. Special Relativity 17

2.1.5 ENERGY-MOMENTUM TENSOR OF A PERFECT FLUID

A perfect fluid is a medium in which the pressure is isotropic in the rest

frame of each fluid element, and shear stresses and heat transport are ab-
sent. If at a certain point the velocity of the fluid is v, an observer with
this velocity will observe the fluid in the neighborhood as isotropic with an
energy density E and pressure p. In this local frame the energy-momentum
tensor is
E 000
o p 0 0
(2.52)
o 0 p 0
o 0 0 p

As viewed from an arbitrary frame, say the laboratory system, let this fluid
element be observed to have velocity v. According to (2.41) we obtain the
transformation

(2.53)

The elements of the transformation are given by (2.42) in the case that
the fluid element is moving with velocity v along the laboratory x-axis, or
by (2.43) if it has the general velocity v. It is easy to check that in the
arbitrary frame

(2.54)

0--+----"'------'9---------+-----------,
o 2 3

FIGURE 2.1. Future light cones at radial distances both inside and outside a
black hole Schwarzschild radius r s == 2G M / c2 . The possible futures of any event
at the vertex of each cone, lies within the cone. Light propagates along the cone
itself. On the scale of distance relative to the Schwarzschild radius of the black
hole, the cones narrow and are tipped toward the black hole. At the critical radius,
the outer edge of the cone is vertical; not even light can escape. Within the black
hole, light can propagate only inward, as with anything else. Future points in
spacetime of a material particle at the vertex of a cone can lie only within the
cone.
18 Special and General Relativity

and reduces to the diagonal form above when v ~ O. We have used the
four-velocity u/-L defined above by (2.46). Relative to the laboratory frame
it is the four-velocity of the fluid element.

2.1.6 LIGHT CONE

The vanishing of the proper time interval, dr ~ 0, given by (2.4) defines a
cone in the four-dimensional space x!' with the time axis as the axis of the
cone. Events separated from the vertex event for which the proper time,
(or invariant interval) vanishes idr ~ 0), are said to have null separation.
They can be connected to the event at the vertex by a light signal. Events
separated from the vertex by a real interval dT2 > 0 can be connected by
a subluminal signal-a material particle can travel from one event to the
other. An event for which dT2 < 0 refers to an event outside the two cones; a
light signal cannot join the vertex event to such an event. Therefore, events
in the cone with t greater than that of the vertex of the cone lie in the
future of the event at the vertex, while events in the other cone lie in its
past. Events lying outside the cone are not causally connected to the vertex
event.
3

General Relativity

"Scarcely anyone who fully comprehends this theory can es-

cape its magic."
A. Einstein
"Beauty is truth, truth beauty-that is all
Ye know on earth, and all ye need to know."
J. Keats Ode on a Grecian Urn

General Relativity-Einstein's theory of gravity-is the most beautiful

and elegant of physical theories. It is the foundation of cosmology-the
subject that traces the evolution of the universe from its first intensely hot
and dense beginning to its possible futures. 1 General Relativity is also the
foundation for our understanding of compact stars. Neutron stars and black
holes Can be understood correctly only in General Relativity as formulated
by Einstein [15,16]. Dense objects like neutron stars could also exist in New-
ton's theory, but they would be very different objects. Chandrasekhar found
(in connection with white dwarfs) that all degenerate stars have a maximum
possible mass [17,18]. In Newton's theory such a maximum mass is attained
only asymptotically when all Fermions, whose pressure supports the star,
are ultra relativistic. Under such conditions, stars populated by the three
heavy quarks-known as charm, truth, and beauty-would exist. However,
such stars do not occur in Einstein's theory because the maximum-possible-
mass star is not sufficiently dense, even at its center; therefore they cannot
exist in nature.
Einstein sought the answer to a question that must have seemed to his
contemporaries, and all who preceded him, as of no consequence: What
meaning is attached to the absolute equality of inertial and gravitational
masses? If all bodies move in gravitational fields in precisely the same way,
no matter what their constitution or binding forces, then this means that
their motion has nothing to do with their nature, but rather with the na-

1 Einstein, himself, never did apply his theory to the evolution of the cosmos.
Indeed, when he discovered the theory (1915), it was a canon of western thought
that the world lasted from "everlasting to everlasting." Edwin Hubble's 1927
discovery of universal expansion shook that belief.