Modern Physics

Lecture Notes
Spring 2022

Rizwan Khalid
Department of Physics
Syed Babar Ali School of Sciences
and Engineering (SBASSE)
Lahore University of
Management Sciences
Office: 9-113A, Feynman Wing
E-mail: rizwan khalid@lums
Contents

1 The Structure of Spacetime 1

1.1 Equivalence of Inertial Reference Frames . . . . . . . . . . . . . . . . . 2
1.2 Michelson-Morley Experiment: . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.1 Michelson Interferometer . . . . . . . . . . . . . . . . . . . . . . 4
1.3 The Constancy of the Speed of Light . . . . . . . . . . . . . . . . . . . 6
1.4 Galilean Transformations . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.5 Lorentz Transformations . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.6 The Consequences of Lorentz Transformations . . . . . . . . . . . . . . 12
1.6.1 Non-relativistic limit . . . . . . . . . . . . . . . . . . . . . . . . 12
1.6.2 Loss of Simultaneity . . . . . . . . . . . . . . . . . . . . . . . . 13
1.6.3 Time Dilation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.6.4 Length Contraction . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.7 Invariant Interval . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.8 Addition of Velocities . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.9 Relativistic Momentum and Energy . . . . . . . . . . . . . . . . . . . . 22
1.9.1 3 Vectors and 4 Vectors . . . . . . . . . . . . . . . . . . . . . . 22
1.10 The Momentum 4 Vector . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.10.1 Proper Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.10.2 Conservation of Momentum . . . . . . . . . . . . . . . . . . . . 24
1.10.3 Time-like Component of Momentum p0 . . . . . . . . . . . . . . 25
1.10.4 The E-p Invariant . . . . . . . . . . . . . . . . . . . . . . . . . 26
1.10.5 Massless Particles . . . . . . . . . . . . . . . . . . . . . . . . . . 27
1.11 Compton Scattering . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

A Problems-Special Relativity 30
A.1 Problem 65 of Randy Harris . . . . . . . . . . . . . . . . . . . . . . . . 30
A.2 Problem 65 of Randy Harris . . . . . . . . . . . . . . . . . . . . . . . . 32

i
Chapter 1

The Structure of Spacetime

One of the most significant revolutions in all of science in the realization that we do
not inhabit a three dimensional world, but rather live in a four dimensional space-
time. Time is the fourth dimension, and must be treated as such. This idea was
developed by Einstein in the early twentieth century while trying to justify the null
result of the Michelson-Morley experiment. We will come to this experiment in just
a bit.
The fact that we live in a four dimensional space-time requires a radical revision
of some of our most cherished concepts from Newtonian times. The first of these
being the absoluteness of time, in other words the ability of all observers to agree on
a universal clock. With time being a dimension, this is no longer possible. We will
see that two physical events that occur at the same time according to one observer
may happen at different times according to other equally ‘good’ observers. This is
known as the relativity of simultaneity. The time-interval between physical events
is no longer the same and depends on the state of relative motion of the observers,
aka time dilation. Not only, volumes of objects are also no longer the same for all
observers. This is known as length contraction. It cannot be emphasized enough that
these are not optical illusions but rather actual measurable physical facts and are
tested on a daily basis for example in the muons that we observe on the surface of
the earth from cosmic rays interacting with the upper atmosphere or in the particle
accelerators that we use to inform us the structure of the Universe at high energy
scales.
One of the most celebrated consequences of Einstein’s theory of relativity is the
equivalence of energy and mass, i.e. E = mc2 . We will see that relativity also forces
us to modify or ‘correct’ our Newtonian formulas for energy and momentum.
The word ‘correct’ is very important. It must be realized that whereas Einstein’s
theory supersedes Newtonian mechanics, it is a correction to Newtonian mechanics.
For everyday velocities, the correction to Newtonian mechanics is rather insignificant.
We will see that corrections to newton’s formulas appear at the order of O(v 2 /c2 ),

1
where v is the speed of the object under consideration and c is the speed of light in
vacuum. The fastest fighter jet ever built had a speed of 4520mph which translates
to just a little over 2000m/s. Now this is a ‘tiny’ fraction of the enormous speed of
light in vacuum of 3 × 108 m/s. In fact, in this example, v 2 /c2 ∼ 10−10 or 1 part in
ten billion. Everyday speeds are far lower than this. It is because of this reason that
the ideas of Einstein’s theory of relativity seem to be a radical departure from our
intuition. Indeed it is always shocking to map out the consequences of these ideas.
For instance, if we did travel at speeds close to the speed of light, time would pass
very differently for us. In particular, assume that one twin goes on an adventurous
space-bound journey while the other stays at home. Then, if the space-bound twin
were traveling at speeds comparable to the speed of light, by the time she returns,
she can possibly find many centuries to have passed to only a few years that she
experienced.

1.1 Equivalence of Inertial Reference Frames

An inertial reference frame is one in which Newton’s P second law holds, i.e. a frame
in which the sum of all forces acting on an object ( F ) is the same as the time rate
of change of momentum
P of that object dp/dt. If the mass is constant, this reduces to
the simpler form F = ma.
For a frame to qualify as an inertial frame, it is necessary and sufficient that the
frame be non-accelerated. It is easy to see that in an accelerated frame, Newton’s
second law would need to be modified. For instance, imagine yourself holding He
filled balloon in a stationary car. The balloon would point straight up. Now as the
car begins its journey, the balloon would be seen to accelerate in the frame of reference
attached to the car. In particular, for a car that begins to move forward with respect
to ground observers, the balloon will be seen to accelerate towards the back of the car.
However, no force will be seen acting on the balloon in this direction, thus violating
Newton’s second law. Another way to violate Newton’s second law is via your own
experience sitting in a car. You do not accelerate with respect to the car when it
begins to move forward, but you do experience a force from the seats and seat belts
acting on you. In this example, there is a net force but no accompanying acceleration.
Let’s analyze the same balloon and sitting in a car example from the point of view
of an ‘inertial’ observer standing outside the car. From this perspective, the balloon
does not accelerate. It stays where it was before the car started to move forward. It
is the back of the car which accelerates towards the balloon and there is a force that
drags the car, the cause of which is ultimately the fuel burnt by the engine (it is a
good exercise to think about the cause-effect relationship in this example in detail,
but that you will have done in your mechanics course).
While it is important to understand what an inertial frame is, it is equally impor-

2
tant, if not more so, to understand and appreciate the fact that all inertial reference
frames are equivalent. One can be convinced of this physical fact by a few examples
from daily life. If we are sitting in a cruising aircraft, our experience with our sur-
roundings is very similar to our experience when we are on ground. For instance, we
can sit, stand, eat and drink just as if the aircraft were standing on the runway. In
fact, if one were to lower the blinds on the aircraft, it would be impossible to tell if
the aircraft were cruising or sitting on a runway. On the other hand, even a little
turbulence will announce its presence with sometimes comic results (spilling coffee,
etc). The same is the experience of train travelers. Take for instance the situation
of sitting in a stationary train and looking at an adjacent one. If the adjacent train
begins to move, it is common experience that travelers in the stationary train may
think their own train is moving. They can be jolted back to reality if they stare in the
opposite direction and see that their train is not moving with respect to the platform.
Motion is relative.
The crux of the equivalence of inertial frames lies in understanding that no physics
experiment, however sophisticated, can distinguish between two different inertial
frames. This also means that there is no preferred inertial frame of reference. It also
means that no Physics experiment can define an ‘absolute’ velocity for any frame.
Velocities of objects are always relative.

1.2 Michelson-Morley Experiment:

The Michelson-Morley experiment was designed to measure the speed of ether relative
to earth. Ether was postulated as being an substance that pervades all of space and
provides sort of an ideal reference frame. The idea of the ether traces back to Greek
mythology with everything on lowly earth made of the four elements of earth, fire,
water and air and the high heavens made of quintessence (the fifth element) or ether.
The standard English usage of ethereal has the same essence (pun intended) of refer-
ring to something as being other worldly. To understand why ether was postulated,
a little history of light is in order.
Ether was invented primarily due to propagation of light. Huygens had promul-
gated the idea of light being a wave. However, the standard view in those times was
the one Newton supported, that light was made up of tiny particles. In the absence of
any conclusive experimental confirmation, the scientific consensus was with Newton.
This was perhaps due to the fact that all mechanical waves were known to propagate
in some medium. Waves on a string could not be setup if there was no string, and
water waves needed water. In the absence of any medium, light must have been made
up of particles that require no medium to travel. Even though not known at that time,
this is consistent with the slowing down of light in a medium. After all, particles will
be dragged down in the presence of media.

3
 Young’s double slit experiment, however, confirmed Huygens’ point of view that
light was, after all a wave. Having solved this mystery, two questions came up. The
first was, how can light propagate from the sun to the earth in the absence of any
material medium. The solution was to postulate an ether. This interpretation was not
without problems as it was also known that light gets polarized, and only transverse
waves get polarized. How a malleable medium (ether) sustains a transverse wave was
an outstanding question. It becomes a whole lot harder to imagine ether when one
considered the wave equation for a material, turns out the wave speed is proportional
to the Bulk modulus of a medium. So ether must have been incredibly stiff......
Leaving aside these questions, the ether hypothesis was also somewhat comforting.
For instance, Maxwell’s equations described electromagnetic waves that were trans-
verse and traveled at the known pspeed of light. Maxwell’s equations described waves
that travel at a speed of c = 1/0 µ0 , where 0 is the electric permittivity of free
space and µ0 is its magnetic permeability. These constants were well known and so
the speed of Maxwell’s electromagnetic (EM) waves could be computed which turned
out to be the same as the known speed of light. Maxwell, therefore, correctly assumed
that light was an EM wave. Maxwell’s equations also tell us that the EM wave in free
space has to be transverse. This is consistent with the fact that light does exhibit
polarization.
But now we get to the question that is most relevant to our p discussion of the
special theory of relativity. We get a from Maxwell’s equations c = 1/0 µ0 which is
supposed to be the speed of light. The question is, in which frame does light travel
with this speed. We, of course, know that speeds and velocities are relative - they
depend on the choice of reference frame. A pen placed on a table in a train has a
velocity of exactly zero with respect to passenger seated in the train. The same pen
has a speed equal to the speed of the train with reference to someone sitting on a
bench observing the moving train. It will have a speed of the order of the orbital
speed of the earth with reference to the sun.
The answer to most physicists was obvious. Light is traveling in ether which
provides a perfect reference frame with respect to which light and perhaps everything
else is at rest. Michelson-Morley set out to measure the speed of earth in ether. For
the remainder of this section, we shall take v to be the speed of ether as observed
from the earth or ‘lab’ reference frame while c is assumed to be the speed of light in
the ether rest frame.

1.2.1 Michelson Interferometer

The Michelson interferometer was designed to precisely tell the difference in travel
times of light as it traveled along a path parallel to the ether and perpendicular to
it. A coherent source of light is split at point A in the figure below. One half of it
goes towards the point B and the other half towards point C. At B and C we have

4
mirrors. The reflected light passes the beam splitter at A and recombines to give an
interference pattern.

v in ether

If the travel time along the two paths is computed, one will see that it is different
assuming equal arm length interferometer and a non-zero speed of ether. In this
frame, ether is assumed to move to the right, along the AC arm of the interferometer.
Let t1 be the travel time to go from A to C and back. Now focus on the following
figure:

c-v v in ether

c+v
L

It is clear that,
L L 2Lc
t1 = + = 2
c−v c+v c −v 2
v2

2 2 −1 2L
= 2Lc(c − v ) = 1 + 2 + ···
c c
2
 
2L v
≈ 1+ 2
c c

In manipulating this formula, use has been made of the binomial series expansion:

(1 + x)n = 1 + nx + n(n − 1)x2 /2! + n(n − 1)(n − 2)x3 /3! + . . . ,

which converges for all n if |x| < 1. The great benefit of this series expansion is in
cases where x  1 and we can retain only the first few terms. In this case the first
two terms will suffice. We now focus on computing the travel time t2 from A to B
and back. For this case, ether and light are traveling at right angles to each other in
the ‘lab’ frame. Let v Lg be the velocity of light with respect to ground. Also let v Le

5
be the velocity of light in the ether frame and vectveg be the velocity of ether with
respect to the ground. We know from the velocity addition formula, that

v Lg + v ge = v Le .

Using the fact that the speed of ether relative to ground is v and that the speed of
light with respect to ether is c, one can construct a right angled triangle to determine
the speed of the light with respect to ground. This is achieved by referring to the
velocity addition formula above and recognizing that v Lg and vectveg = −vectveg are
at right angles to each other and so the head to tail rule puts the third vector v Le as
the hypotenuse.√The Pythagorean theorem now gives us the speed of light in the lab
frame as being c2 − v 2 . This would be the same for light going from A to B and for
the return journey. Therefore,

v2
 
2Lc 2L
t2 = 2 ≈ 1+ 2
c − v2 c c

Use is again made of the binomial series and only the first two terms are retained.
We can clearly see that t1 6= t2 and in fact ∆t ≈ Lv 2 /c3 . Since the speed of
light is roughly c, this time difference corresponds to an optical path difference of
Lv 2 /c2 . This path difference should lead to an interference pattern. As the apparatus
is rotated by 90o , the long arm of the interferometer becomes the short one and vice
versa. This results in a doubling of the optical path difference. Plugging in known
numbers, and taking into account experimental uncertainties, it was concluded that
the speed of earth in ether had to be much smaller than the known orbital speed of
earth.
It is good to keep in mind that experiments always produce bounds. The upper
bound on the ether speed was all that could have been hoped for in an experiment
like this. But if the shift in fringes had been significant, one would have obtained an
upper and lower bound on the ether speed.

1.3 The Constancy of the Speed of Light

Einstein reconciled the speed of light given by Maxwell’s equations and the null result
of the Michelson-Morley experiment by boldly postulating that light travels with the
same speed in all inertial reference frames. Now this is a radical departure from the
Newtonian idea of velocity addition. In essence, one is saying that close to the speed
of light, the velocity addition formula will need to radically change to account for
keeping the speed of light a constant with respect to all observers. There will be no
special formula for light, but a general one applicable to all objects. At small speeds,
this formula better agree with the well-known velocity addition formula that we are

6
familiar with. A better way think about this is to say that there is an upper limit
to speeds. There will be a maximum speed allowed. Granted that maximum is fairly
large (3 × 108 m/s). But it is an absolute maximum. Nothing can travel faster than
this speed. Light just happens to be something that travels at exactly this speed.
This constancy of the speed of light requires a radical revision of our concepts
of space and time and space and time taken together. We will now work out the
consequences of this postulate and see how we need to think about space-time.

1.4 Galilean Transformations

The Galilean transformations should be familiar from your elementary Mechanics
course. We relate the coordinates of two frames S and S 0 such that S 0 moves to the
right of S with speed v along the x axis. See Figure 1.4 for details.

y y
v P(x,y,z,t )
P(x,y,z,t)
S S
x x

The coordinates are related via,

x0 = x − vt
y0 =y
z0 =z
t0 =t (1.1)
The last of these equations is the mathematical statement of the fact that we can
define a universal clock good for all observers. By differentiating these equations with
respect to the relevant time coordinate we can find the relations for the transformation
of the velocities and accelerations.

dx0 dx0 dx
= = −v
dt0 dt dt
=⇒ vx0 = vx − v
d2 x0 d2 x
= 2
dt02 dt
or a0x = ax (1.2)

7
The last equation is a statement that Newton’s second law holds in both frames S
and S 0 .
Before moving forward, it is a good idea to begin drawing the trajectories of
particles in S and S 0 .

t=t
t S t S

x x
Moving
Particle Particle at brest in S
Particle at
rest in S moving with respect to S

Shown in green is a particle at rest in S. The time coordinate for this particle
changes but its position remains the same. The same particle in S 0 will have a “world-
line” as shown in the figure. The slope of this line would be;

dt0 1
0
=− , (1.3)
dx v
with the − sign signifying the fact that the particle moves to the left according to
frame S 0 .
One of the consequences of the Galilean transformations is the fact that lengths
remain the same as observed in S and S 0 .
To see this imagine you are finding the distance between two particles. If the two
particles are infinitesimally close to each other, the distance squared between them
would be;
ds2 = dx2 + dy 2 + dz 2
Notice that we must set dt = 0 when we are measuring distance between the
particles. This is because the distance between two objects makes sense if measured
at the same time. For instance, imagine Ali is in Islamabad one fine morning and
wishes to meet Bismah in Lahore. The distance between them makes sense only if
we observe it at a fixed time. In particular, it does not make sense to take the x-
coordinate of Ali in the morning and subtract it from the x-coordinate of Bismah in
the evening.

8
 Let’s formally show that the distance between points is independent of the coor-
dinate system used to measure the distance;

dx0 = dx − vdt
dy 0 = dy
dz 0 = dz
dt0 = dt (1.4)

Setting dt = dt0 = 0, one can see,

(ds)2 = (dx)2 + (dy)2 + (dz)2 (1.5)

= (dx0 )2 + (dy 0 )2 + (dz 0 )2 (1.6)

We can integrate the infinitesimal distances to show that lengths of objects in S

will be the same as their lengths in S 0 .

1.5 Lorentz Transformations

y=y ; z=z
t S t S
v

x x

It is clear that the Galilean transformations are not a description of reality. This
is because they would predict that light travels at different speeds in different frames
of reference. This is opposed to Einstein’s postulate about the constancy of the speed
of light. This postulate, in turn, is based on the null result of the Michelson-Morley
experiment, or at least at an attempt at explaining the said null result.
The way we proceed is to take this postulate of the constancy of the speed of light
at face value, work out the Lorentz transformations, and figure out the physical conse-
quences. We will then go to the lab to test the consequences for physical experiments,
and accept the postulate as a valid theory if it holds up to experimental scrutiny.
The mere reason that Einstein’s theory of relativity is being taught is testament
to the fact that it indeed does hold up to experimental scrutiny.
The transformation describing the world from the S to S 0 points of view has to be
linear. Otherwise space cannot be homogeneous. A priori, the homogeneity of space

9
sounds like a good idea. All we mean by homogeneity of space is that all points in
space are identical. A simple reason for homogeneity can be sought by considering
force free motion of a particle. In this case, both in S and S 0 the particles x and t or
x0 and t0 must be linearly related. This is possible only if the transformation from S
to S 0 is itself linear.
The most general linear transformation connecting S and S 0 is,

x0 = Ax + Bt + E
t0 = Cx + Dt + F, (1.7)

and we expect y 0 = y and z 0 = z.

We choose t0 = t = 0 to be the event at which x0 = x = 0. Then E = F = 0.
This is just a way for agreeing on what time to call zero according to all observers.
We will later see that the clock’s do not agree at other times. This means that they
must be ticking at different rates. This is just time dilation, something we will explore
in detail later. Now x0 = 0 corresponds to the origin of S 0 and has x-coordinate vt.
Therefore the first of Equations (1.7) implies,

0 = Avt + Bt
or B = −Av

So we can re-write this equation as;

x0 = A(x − vt) (1.8)

Similarly x = 0 is the origin of S which moves according to S 0 as x0 = −vt0 ; and

so the second of Equations (1.7) and Equation (1.8) give;

t0 = Dt
and − vt0 = −Avt

and so A = D. We now have the equations;

x0 = A(x − vt)
and t0 = Cx + At
 
0 C
or t =A t+ x
A
0
or t = A(t + Gx)

10
 It is now convenient to go to the standard form;
x0 = γ(x − vt)
t0 = γ(t − Gx) (1.9)
The inverse transformation corresponding to the first of Equations (1.9) must be
of the form;
x = γ(x0 + vt0 ) Role reversal between S and S 0
⇒ x = γ 2 (x − vt) + γvt0
x γ2
⇒ t0 = − (x − vt)
γv γv
 
0 1 1
or t = γt + γ 2 − x
γ v v
1 − γ2
 
1 1
⇒ G= − 1 =
v γ2 γ 2v

So finally x0 = (γ− vt)  
2 (1.10)
and t0 = γ t + 1−γ
γ2v
x 

Imagine that at t = t0 = 0 a light is emitted at x = x0 = 0; and x0 and x give the

coordinates of this light wavefront for t0 and t respectively. then, we already know
that;
)
x = ct
(1.11)
and x0 = ct0
Using Equations (1.11) in Equations (1.10) gives;

ct0 = γ(ct − vt) (1.12)

1 − γ2
 
0
t = γ t + 2 ct
γ v
1 − γ2 2
 
0
⇒ ct = γ ct + 2 c t (1.13)
γ v
Compare equation (1.12) and equation (1.13)
1 − γ2 2 1 − γ2 v2
c = −v ⇒ = −
γ 2v γ2 c2
1 v2 1 v2
or − 1 = − ⇒ = 1 −
γ2 c2 γ2 c2

11
 So our final result is;

1
γ=p (1.14)
1 − v 2 /c2

with the Lorentz transformations given by;

x0 = γ(x − vt)
 
0 v
and t = γ t − 2x (1.15)
c

and the inverse transformations being;

x = γ(x0 + vt0 )
 
0 v 0
t = γ t + 2x (1.16)
c

1.6 The Consequences of Lorentz Transformations

As alluded to earlier, now that we have set up the Lorentz transformations, we now
explore the various consequences of these transformation rules. In particular, we will
focus on the loss of simultaneity, time dilation and length contraction. For this entire
discussion, we will keep in mind the physical picture of one frame S 0 that moves to the
right of frame S with speed v. This motion is along the x axis of both frames which
agree to call t = t0 = 0 the event corresponding to the overlap of their spatial origins
x = x0 = 0. In this case, as was derived in the previous section, the Equations 1.15
give the Lorentz transformations connecting coordinates in the frame S 0 to the frame
S while Equations 1.16 give the inverse transformations. The gamma factor appearing
in these equations is given by Equation 1.14.
But before we discuss the relativistic effects of any kind, it is a good idea to
consider the non-relativistic limit of the Lorentz transformations.1

1.6.1 Non-relativistic limit

p case v  c, or the case of everyday
By the non-relativistic limit we mean the limiting
velocities. First let’s concentrate on γ = 1/ 1 − v 2 /c2 . It is obvious that larger
1
I would urge you to draw plenty of space-time plots while reading this section and make sure
you are able to draw the spacetime points that we talk about. This will lead to better understanding
of the material.

12
the v, the larger γ will be. A binomial expansion of γ is appropriate to look at the
non-relativistic limit:
−1/2
v2

γ = 1− 2
c
v2 3v 4
= 1 + 2 − 4 + ... (1.17)
2c 8c
In particular, it is interesting to see that for v  c, it will be sufficient to take
γ ' 1. It is now easy to see how the first of Equations 1.15 yields the Galilean result
for v  c. The second of Equations 1.15 also nicely reduces to the Galilean result
of t0 = t when v  c. The situation that we are familiar with due to our everyday
experience, i.e. the world of the non-relativistic particles, is nothing but a special case
of the far more general Special Theory of Relativity which applies to all objects that
may or may not be traveling at high speeds. In other words, Einstein’s theory would
lead to non-measurable corrections to Newtonian ideas whenever the speeds involved
are much smaller than the speed of light.

1.6.2 Loss of Simultaneity

We now move to the special relativistic effects; effects that are surprising to us pre-
cisely because they do not form part of our intuition that is informed through expe-
rience. As argued already, special relativistic effects will appear only when objects
travel at speeds comparable to the speed of light.
Let us begin investigating the relativity of simultaneity and ask the question;
“If two events, say A and B are simultaneous in some frame of reference, are they
also simultaneous in all frames of reference?”
Alternatively stated, the question is;
“Do all inertial observers agree on the termporal ordering of events?”
The simple answer to this question is “maybe”. In this section we will only look at
an example of two events that are simultaneous in one frame but happen at different
times according to different observers. This treatment will suffice to conclude that at
least in some cases, the special theory of relativity requires us to give up the sense of
a strict temporal ordering of events.
We will later see that in certain cases the temporal ordering of events is strictly
preserved by Lorentz transformations. But for now, as an example of where time
ordering is not preserved, take as example two events happening at the same time
in S. Call these events A and B. Let their coordinates2 in frame S be A(T, 0) and
2
Please note that for these lecture notes I will be writing the coordinates in the order (t,x,y,z).
More frequently, we will only talk to the (t, x) plane. This convention, while it differs from the
one used by Randy Harris, is preferred as it seamlessly takes one from considering the t, x plane to

13
B(T, L). More explicitly, tA = tB = T, xA = 0, xB = L, so that in frame S the two
events happen at the same time T but have a spatial separation of L.
Let’s use Equations 1.15 to find the time coordinates of A and B in frame S 0 .
Clearly;
vxA
t0A = γ(tA − ) = γT
c2
vxB vL
t0B = γ(tB − 2 ) = γT − 2 (1.18)
c c
Clearly, with v > 0, the event B happens before A. One physical example of this
could be given by considering A and B to correspond to light striking the left and
right ends of a room. Imagine that a light is placed at the position x = L/2. This
light is turned on at a time t < T . According to S, light must reach the ends of
the room at the same time which we call T . Let’s refer to the left end of the room
as the one with x = 0. Clearly, we have just identified the events A and B above
as light reaching the left and right ends of the room. The same events will not be
simultaneous in frame S 0 as our equations have shown. In fact, with S moving to the
left of S 0 , observers will certainly see light reaching the right end of the room before
the left as the right end of the room, according to observers in S 0 is moving towards
the light ray while the left end is moving away from the light ray. Observers in S 0
must agree that light is traveling at the same speed in both directions: this is what
special relativity is all about. With light traveling at the same speed, and it having
to cover less distance to reach the right end, it is no wonder that it indeed does reach
the left end later.
It should be noted that the mathematical (algebraic) demonstration of the loss of
simultaneity is true regardless of what events A and B physically are. They could,
for instance, correspond to a snail reaching the left and right ends of a room. The
loss of simultaneity is a fact of life: it is not specific to light. For light, making the
physical picture is easier as it is guaranteed to travel with the same speed according
to all observers.
Incidentally, the same calculation shows that if S 0 moves to the left of S, event
B will happen after A. Make sure you can visualize this via the light example given
above.

1.6.3 Time Dilation

Let’s look at a moving clock. From our setup, it is convenient to stare at the clock
placed at x0 = 0 Let events A and B correspond to this clock. The spacetime coordi-
nates according to the frame S 0 that we consider are, therefore, A(0, 0) and B(T, 0).
the four-dimensional spacetime. Furthermore, this is by-far the most prevalent notation in modern
usage.

14
In frame S we will get:

vx0A
tA = γ(t0A + ) = γT
c2
vx0
tB = γ(t0B − 2B ) = γT
c
0
=⇒ ∆t = ∆t/γ, (1.19)

or that time dilates. The moving clock runs slower. An interval of time ∆t0 in the
frame in which the clock is at rest dilates to a longer interval ∆t in the frame in which
the clock is moving.
Time dilation has been extensively tested in the laboratory. We know that short-
lived particles live much longer when they are moving at high speeds, their internal
clocks slow down when they are in motion. A typical example is that of muons, the
details of which you will work out in one of the Assignment Problems. One must
stress that the lifetime of short-lived particles is a characteristic of their nature. Its
just that their nature changes when they are traveling at high speeds. This increase
in their lifetime is entirely explained by the γ factor time dilation relation.
We have also made measurements of time dilation using very precise atomic clocks.
I invite you to read up on the Hafele-Keating experiments. A full calculation of those
also requires one to compute the gravitational time dilation. But it is a very satisfying
experiment that proves, if there was any doubt, that these relativistic effects will apply
equally to all physical systems.
I also invite you to think about the Twin paradox. One twin goes out in space
and travels around the cosmos. The stay at home twin may age decades to only a
few years, or even a few month aging of the twin that travels. Now why is this a
paradox? It is because according to the journeying twin, time dilation should apply
in the opposite direction. The resolution of this paradox is not so complicated. It
relies on us recognizing that it is, in fact, the journeying twin who has to accelerate,
change direction of motion, and only then can they head back ‘home’. Therefore, the
journeying twin’s calculations cannot be trusted as they are not made from an inertial
frame.

1.6.4 Length Contraction

We now discuss special relativistic length contraction, the fact that lengths along the
direction of motion will be contracted. Imagine a stick of length L0 placed in the S
frame. For definiteness, we will place this stick between xA = 0 and xB = L0 . We
must compute the x0 coordinates of the ends of this stick in frame S 0 in order to find
its length according to S 0 . However, this must be done at some time in S 0 which is the
same for events A and B. For definiteness and simplicity, let’s choose t0A = t0B = 0.

15
Then, from Equations 1.15, it is clear that;
 vxA 
t0A =γ tA − 2 = 0
c
=⇒ tA =0
and x0A =γ(xA − vtA ) = 0 (1.20)

Now inserting t0B = 0, xB = L0 in the Equations 1.15, we get

 vxB 
t0B = 0 =γ tB − 2
c
vL0
=⇒ tB = 2
c
0
and xB =γ (xB − vtB )
 
vL0
=γ L0 − v 2
c
=L0 /γ
=⇒ L =x0B − x0A = L0 /γ (1.21)

and since γ > 1, the length of an object L is smaller than its length at rest, L0 . Of
all the relativistic effects, this one is perhaps the most poorly understood. There is
nothing wrong in this calculation that we just did. But we must know that the visual
appearance of an object and its “true” geometric shape are two different things. This
difference is again due to the finite propagation speed of light. So, whereas, length
contracted images of extended objects are very common even in physics texts, one
must remember that the actual visual shape of an object will not be purely determined
via simple length contraction considerations. To take a firm example, whereas simple
length contraction would lead us to believe that a sphere would look like a rugby ball,
in fact, it has been demonstrated that an object with a spherical shape would always
look like a sphere. This phenomenon goes by the name of Terrell rotations and has
been widely discussed in the literature. I invite you to have a look at this if you are
interested3 .

3
See for instance, https://en.wikipedia.org/wiki/Terrell rotation. In particular the Mary L Boas
paper referenced in the Wikipedia article is very interesting.

16
1.7 Invariant Interval
In Lorentz Transformations, the Euclidean distance, ds2 = dx2 + dy 2 + dz 2 for dt = 0
is not invariant. This should be obvious as the transformation equations no longer
respect lengths. We have seen that lengths get modified by the Lorentz transforma-
tions.
One can ask, is there some “distance” between space-time points that does remain
invariant under Lorentz transformations? It turns out that there is;
ds2 ≡ dx2 + dy 2 + dz 2 − c2 dt2
can be shown easily to be a Lorentz invariant quantity. Since we have chosen the x
direction for the Lorentz transformation, it is clear that dy = dy 0 and dz = dz 0 .
Consider;
  2
0 2 2 0 2 2 2 vdx
(dx ) − c (dt ) = [γ(dx − vdt)] − c γ dt − 2
c
v 2 dx2 2vdxdt
 
2 2 2 2 2 2 2
= γ (dx + v dt − 2vdxdt) − c γ dt + −
c4 c2
v2 v2
   
= γ 2 1 − 2 dx2 − c2 γ 2 1 − 2 dt2
c c
But now realize that γ 2 = 1/(1 − v 2 /c2 ) so that the above simplifies to,
(dx0 )2 − c2 (dt0 )2 = dx2 − c2 dt2 ,
In other words, the invariant space-time interval is,
ds2 = dx2 + dy 2 + dz 2 − c2 dt2
It is also easy to show, using the Lorentz Transformations, that I ≡ (x1 − x2 )2 +
(y1 − y2 )2 + (z1 − z2 )2 − c2 (t1 − t2 )2 is the invariant spacetime interval between two
points with coordinates (t1 , x1 , y1 , z1 ) and (t2 , x2 , y2 , z2 ).
In the subsequent discussion we will suppress the y and z coordinates.
Let’s now consider a particle, moving with a constant speed u. Let’s say it begins
at point P1 (t1 , x1 ) and travels to P2 (t2 , x2 ). then
∆x x2 − x1
u= =
∆t t2 − t1
⇒ ∆x = u∆t
The spacetime interval between P1 and P2 is;
I(1, 2) = (∆x)2 − c2 (∆t)2
= (u2 − c2 )(∆t)2

17
 ct

o o
45 45 x

Figure 1.1: The light-cone. To visualize the cone, add another spatial axis to this
diagram.

Now for a physical particle u ≤ c with u = c for photons. Particles with mass
always have u < c. So for photons I = 0, and for massive particles I < 0, strictly.
Spacetime intervals for which I = 0 are called light-like intervals, while those for
which I < 0 are called time-like 4 . The lightcone is a convenient representation for
the path taken by light.5 Focus attention on Figure 1.1. x = 0 and t = 0 is just
some convenient choice of origin for our coordinate system. The red line represents
a photon that is traveling along the x-axis and passes through x = 0 at t = 0. This
photon travels from the “left” to the “right”. The blue line similarly represents a
photon that travels right to left. If we plot also the y-axis, we can see why this is
called the light-cone. The interpretation of the light-cone that if we find a particle at
x = 0, t = 0; its world-line in the past must have been entirely in the “past” light-cone
and all its possible future trajectories are in the future light-cone.
Let’s now consider two particles that appear out of nowhere at x = 0 and x = L
at t = 0 as shown in Figure 1.2. Shown in the figure are the future light-cones of these
particles. These future light-cones do not intersect before t = T = L/2c. Before this
much time has lapsed, the two particles cannot, even in principle, know of each other’s
existence. At t = 0, the two events shown in the Figure 1.2 are spacelike separated,
they have an invariant interval that is I = L2 > 0. As previously discussed, no
physical particle can have a worldline that connects these two events. This, of course,
is obvious for this figure, as the two events are simulutaneous and for a physical
particle to travel from 1 to 2 would require travel at an infinite speed.
This, of course is a dynamic situation. As the particles move, their possible futures
4
There is no agreement in the literature on what sign should be used to define the interval. Some
authors prefer to define ds2 = c2 dt2 − dx2 − dy 2 − dz 2 . This choice would mean that intervals for
which I > 0 would be called time-like. All that the special theory of relativity requires is for the
time coordinate to have the opposite sign as compared with the spatial coordinates.
5
Please keep in mind that this discussion assumes the particle begins at P1 and travels to P2 .

18
 ct

cT=L/2

1 2 x

Figure 1.2: Two particles pop into existence simultaneously at t = 0 and have a
spatial separation of L. They will not know, in principle, of each other’s existence for
at least t = L/2c.

will modify. For instance, let’s assume that the two particles are moving away from
each other at t = 0. Then their future light-cones will intersect at a time t > T as
shown in Figure 1.3.
If I(1, 2) > 0, events 1 and 2 are space-like separated as previously discussed.
Nothing happening at 1 can have a causal impact on 2, and vice versa. These are
the types of events that do not have a well defined time ordering. We saw earlier
examples of loss of simultaneity. Two events that were simultaneous in one frame,
were not simultaneous in another. Not only this, some observers reported 1 to happen
before 2, while others reported 2 to happen before 1. If you go back to that example,
you will recognize the two events were space-like separated.
On the other hand, if I(1, 2) ≤ 0, then event 2 can causally depend on 1 if t1 < t2 .

Figure 1.3: The possible futures of particles modify as they each move along their
world-line. This will impact “when” they can “see” each other.

19
This should be obvious from the fact that now a physical particle can travel from 1 to
2, thereby something happening at 1 can extend its influence to 2. Therefore, it must
be true that the time-ordering of events 1 and 2 is Lorentz invariant if I(1, 2) ≤ 0.
In other words, if I(1, 2) ≤ 0 and t1 < t2 for some inertial observer, then all inertial
observers must agree that t1 < t2 according to them. This statement is important
enough to warrant a mathematical proof.
Let A(t1 , x1 ) and B(t2 , x2 ) be events such that I(A, B) ≤ 0, with t2 > t1 . We need
to confirm that t02 > t01 where the primed coordinates refer to some Lorentz “boosted”6
frame. Then,
 
0 vx1
t1 = γ t1 − 2
c
 
0 vx2
t2 = γ t2 − 2
c
v
⇒ t02 − t01 = γ(t2 − t1 ) − 2 (x2 − x1 ) (1.22)
c
If v > 0 and x2 ≤ x1 , it is clear that t02 > t01 if t2 > t1 . If I(A, B) = I(1, 2) ≤ 0, we
have;
(x2 − x1 )2 − c2 (t2 − t1 )2 ≤ 0
or (x2 − x1 )2 ≤ c2 (t2 − t1 )2
(x2 − x1 )2
=⇒ ≤ c2
(t2 − t1 )2
Since, t2 > t1 , we get from x2 > x1 , the result that:
(x2 − x1 ) ≤ c(t2 − t1 )
In other words, the maximum value that (x2 − x1 ) can have is just shy of c(t2 − t1 ).
Plugging this in equation (1.22) gives,
v
t02 − t01 ≥ γ(t2 − t1 ) − 2 c(t2 − t1 )
  c
v
t02 − t01 ≥ γ − (t2 − t1 )
c
Now v/c ≤ 1 and γ ≥ 1. We can, therefore, claim that γ ≥ v/c. Moreover,
when v → c, γ grows without bound and we can legitimately claim that γ is strictly
larger than v/c or that γ − v/c > 0. This leads us to conclusion that in all cases for
I(1, 2) ≤ 0, if t2 > t1 then t02 > t01 . Or for events that can causally affect each other,
the temporal ordering is something that all observers agree on.
6
A Lorentz boost is what we have been calling a Lorentz transformation. Technically, a rotation
of corrdinates, for example in the xy-plane is also a Lorentz transformation. But here we specifically
talk about a “boost”, a Lorentz transformation that mixes the time coordinate with one (or more)
of the spatial ones.

20
1.8 Addition of Velocities
We can now move on the relativistically correct formula for velocity addition, the
starting point for our discussion of the special theory of relativity.
Let an object have a velocity u = (ux î + uy ĵ + zz k̂) in a frame S. We would like
to know u0 , the velocity of the same object in frame S 0 , such that S 0 moves to the
right of S with speed v.
We already know that,
dx0 = γ(dx − vdt)
 
0 vdx
dt = γ dt − 2
c
0
dy = dy
dz 0 = dz
So;
dy 0 dy
u0y = 0
=  
dt vdx
γ dt − 2
c
dy
u
or u0y =  dt =  y 
v dx vux
γ 1− 2 γ 1− 2
c dt c
Likewise,

uz
u0z =  
vux
γ 1− 2
c
 
dx
−v
0 dx0 γ(dx − vdt) dt
and ux = 0 =   =  
dt v v dx
γ dt − c2 dx 1− 2
c dt

So that we can finally conclude;

ux − v
u0x =  
vux
1− 2
c
As always, for the non-relativistic limit, these formulas reduce to the Galilean
ones. The formula for u0z will resemble that of u0y .

21
1.9 Relativistic Momentum and Energy
Just like space and time, our concepts of momentum and energy need to be modified.
But before we do this, let’s revisit the idea of vectors.

1.9.1 3 Vectors and 4 Vectors

Displacement, velocity and momentum etc are all vectors that we are familiar with.
The quintessential property of a vector is invariance of an appropriately defined length
when we do a coordinate transformation. For the case of the Galilean transformations,
this length is the “normal” length of the arrow we use to represent the vector.
In the case of the Lorentz transformations we already know that the usual Eu-
clidean lengths are nor invariant. Instead, the 4-distance that we called the invariant
interval is the appropriate measure to use. For the space-time vector, this invariant
is;

I = (∆~x)2 − c2 (∆t)2

This is the direct analogue of the length of displacement 3 vector. For an infinit-
simal 4-displacement, we can define the invariant measure to be;

ds2 = dx2 + dy 2 + dz 2 − c2 dt2

The first thing this view forces us to accept is the reality of living in a four dimen-
sional space-time continuum. So space and time together define the 4 dimensional
space that we inhabit.
On a related note, all vectors must transform similarly. In particular, for a Lorentz
transformation along the x-axis;
 
0 vct
x = γv x −
c
 
0 vx
ct = γv ct −
c

v
Let’s define βv = , so that
c

x0 = γv (x − βct)
ct0 = γv (ct − βx)

Clearly, defining ct as the time component of the space-time 4 vector has its
advantages. To summarize in 4D space-time it will be meaningful to talk of 4 vectors.

22
1.10 The Momentum 4 Vector
Following on from the previous discussion, define
dxµ ≡ (cdt, d~x) = (cdt, dx, dy, dz)
as the infinitsimal displacement in space-time. The index µ takes on the values
0, 1, 2, 3, with µ = 0 corresponding to the time-like direction. In this case dx0 = cdt
The usual way to define momentum is to take the displacement, divide it by dt to
get the velocity, and multiply it with the mass m. However one can show that,
 
d~r dx dy dz
m = m ,m ,m
dt dt dt dt
is a bad definition for momentum as this quantity is not conserved. In particular, it
is possible to show that d~r/dt will not be conserved in a Lorentz boosted frame even
if it is conserved in some given frame.
Moreover, the Lorentz transformation properties of d~r/dt are complicated. If
momentum is to be a genuine vector in the special theory of relativity, it must have
the same Lorentz transformation properties as the infinitesimal displacement dxµ .

1.10.1 Proper Time

Every object has associated with it a special clock. This is a clock that is at rest
with respect to that object. In other words every object carries with it it’s own clock.
The time measured by this is called the proper time. For an infinitesimal interval of
proper time we use dτ . Clearly,
dt = γu dτ,
p
2 2
where γu = 1/ 1 − u /c and u is the speed of the particle in frame S. dτ is special
as it is a kind of Lorentz invariant by definition. Also, define m to be the rest mass
of the particle. We will take the modern viewpoint in this course that the mass is
always the rest mass of a particle. So this again, like the proper time, is a Lorentz
invariant by definition. Define;
dxµ
pµ = m
 dτ 
dt dx dy dz
= mc , m , m , m
dτ dτ dτ dτ
Using dt = γu dτ , we can rewrite this as;
dxµ
pµ = mγu
 dτ 
dx dy dz
= mcγu , mγu , mγu , mγu
dt dt dt

23
 Since m and dτ do not change under Lorentz transformations, we are guaranteed
that;
0
p1 = γv (p1 − βp0 )
0
p0 = γv (p0 − βp1 )

Keep in mind that in this notation p0 = mcdt/dτ = mcγu and p1 = mdx/dτ =

mγu dx/dt = mγu ux .

1.10.2 Conservation of Momentum

Let’s assume that we consider a scattering process involving the particles A and B in
the initial state and C and D in the final state. This notation makes explicit the fact
that during collisions, particles can, in principle, change their identities. You will be
familiar, for example, with the elementary annihilation process e+ e− → γγ, where the
e− is an electron, e+ is a positron and the γ is a photon. The reverse of this process
is also allowed.
Such scattering processes do not involve any “external” influence, and so, from
our previous experience, energy and momentum must be conserved. Therefore, let’s
assume that in some frame S, we observe the following process:

A + B → C + D,

such that,
pµA + pµB = pµC + pµD .
In other words, we insist that in frame S the momentum is the same before and
after the collision. In writing this statement, we are actually insisting that the time-
like component of the momentum 4-vector is also conserved. So far, we have not
discussed the physical significance of this time-like component.
We would like to ask the question if this conservation of momentum carries over
to some other Lorentz frame or not. In some frame S 0 moving to the right of S with
speed v, we hope to get,
p0µ 0µ 0µ
A + pB = pC + pD
0µ

Since we consider Lorentz transformations in the x-direction, we need worry only

about p1 and p0 .
0
p1 = γv (p1 − βp0 )
0
p0 = γv (p0 − βp1 )

We begin by assuming momentum conservation in the S frame;

p1A + p1B = p1C + p1D

24
 Then, from the Lorentz transformation equations,
0 0
p1A + p1B = γv (p1A − βp0A ) + γv (p1B − βp0B )
0 0
=⇒ p1A + p1B = γv (p1C + p1D ) − β(p0A + p0B )

We have used in the second of these Equations the condition that momentum is
0
conserved in the frame S. In order to get conservation of p1 , or the x0 component of
momentum in the frame S 0 , it seems quite obvious looking at this second Equation,
that we must have conservation of the time-like component of momentum in the frame
S. For when we assume that, we get:
0 0
p1A + p1B = γv (p1C + p1D ) − β(p0C + p0D )
0 0 0 0
=⇒ p1A + p1B = p1C + p1D

So, we have some useful information; summarily by defining momentum as; pµ ≡

dxµ /dτ , we get a quantity that
1. Transforms like a 4 vector.

2. Is conserved in all frames if it is conserved in some given frame.

1.10.3 Time-like Component of Momentum p0

Let’s now investigate the time-like component of the momentum 4-vector.
dt
p0 = mc = mcγu
dτ
mc
or, p0 = p
1 − u2 /c2

Let’s look at a closely related quantity, cp0 ;

mc2
cp0 = p
1 − u2 /c2
−1/2
u2

2
= mc 1 − 2
c
1 u2
  4 
2 u
= mc 1 + 2
+O 4
2c c
In the non-relativistic limit;
1
cp0 = mc2 + mu2 .
2

25
We can see that we have retrieved the familiar formula for the kinetic energy of a
particle, 12 mu2 . Added to this is a constant mc2 , which we define as the rest energy.
In other words, it seems prudent to identify Eu = cp0 = γu mc2 as the relativistic
energy of a particle of mass m traveling at a speed u in some frame. So one notion
that we must add to our repository of new ideas is this somewhat mysterious rest
energy of a particle, i.e. mc2 . This is rest energy is not important in the non-
relativistic scenario as we get an almost exact conservation of mass in that limit. In
the relativistic case, on the other hand, we can have interactions where the mass is
not conserved at all. Consider, for instance, the annihilation of an electron-positron
pair. In the initial state we have particles that have masses but in the final state we
have particles that don’t have a mass. The relativistic energy, on the other hand, as
we have already seen is conserved. Einstein’s theory talks of an equivalence of energy
and mass, there is no difference between energy and mass, except the obvious one of
using different units to measure the two quantities7 .
It is, therefore, only natural to talk about the energy momentum 4 vector, which
is conserved whenever there is no external force acting on the system.

1.10.4 The E-p Invariant

It should be obvious by now that; (p1 )2 + (p2 )2 + (p3 )2 − (p0 )2 is an invariant quantity.
This can be seen directly by applying the Lorentz transformations. It is, in fact, true
that for any 4-vector, the square of the time-like component subtracted from the sums
of squares of the spatial components is always a Lorentz invariant quantity.
Let’s evaluate this invariant in the particle’s rest frame; for which u = 0, γu = 1.
Then pµ = (mc, 0, 0, 0), and the invariant is;

(0)2 + (0)2 + (0)2 − (mc)2 = −m2 c2

In any other frame, therefore,

(p1 )2 + (p2 )2 + (p3 )2 − (p0 )2 = −m2 c2

This can be conveniently written using the 3 vector notation;

E2
|p|2 − = −m2 c2
c2
or E 2 = m2 c4 + |p|2 c2
7
This is precisely why the ‘natural’ system of units sets velocities as being dimensionless quantities
(c = 1). Now energy and mass can be measured in the same units. Have a look at the structure of
all the equations we have written by setting c = 1. Instead of v, use β = v/c.

26
1.10.5 Massless Particles
We have seen that the energy momentum relation for a particle with mass m is;

E 2 = m2 c4 + |p|2 c2

We now pause and ask if it is possible to accommodate massless particles into this
framework. Let’s stare at the formula for the energy of a particle;

2 mc2
E = γu mc = p
1 − u2 /c2

First note that as m → 0, E → 0. Also, as u → c, E → ∞. The first conclusion from

this observation is that a massive particle cannot travel at the speed of light. This is
because the energy for such a particle would become infinite, and so to accelerate it
to the velocity of light would require infinite energy.
An associated conclusion of this fact is that photons must be massless as they are
particles of light, and so, by definition travel at the velocity of light.
The energy momentum relation for a massive particle, E 2 = m2 c4 + |p|2 c2 , can be
seamlessly generalized to the case for a massless particle. We can conclude that for
photons (and all other massless particles);

E
E = |p|c or = |p|
c
So the momentum 4 vector for a photon traveling along the z-axis would be given by,

pµ = (|p|, 0, 0, |p|).

As another example, assume that a photon is traveling in the xy plane making

an angle θ with respect to the x axis. Then, the momentum 4-vector would be,
|p|, |p| cos θ, |p| sin θ, 0.

1.11 Compton Scattering

Assume that the energy of a photon is given by,
hc
E = hf =
λ
We will later justify this from the photoelectric effect. Using E = |p|c ⇒ |p| = h/λ.
Compton scattering is the observation that x-ray photons, collide with free elec-
trons just as marbles collide with each other.

27
 y
After Collision
Before Collision g
g e q x
f
m
e

hc
Ei = + mc2
λ
h
p~i = î
λ
hc p
Ef = 0 + p2e c2 + m2 c4
λ   
h h
pf = 0 cos θ + pe cos φ î + 0 sin θ − pe sin φ ĵ
λ λ

Conservation of momentum suggests;

p i = pf
h
sin θ = pe sin φ (1.23)
λ0
h h
0
cos θ + pe cos φ = (1.24)
λ λ

Conservation of energy requires that,

Ei = Ef
hc hc p
+ mc2 = 0 + p2e c2 + m2 c4 (1.25)
λ λ
Let’s pause to think how we would do a controlled experiment. We can be quite sure
of the wavelengths of the initial and final photons as we can measure those accurately.
We can also measure the angle θ at which the photons are going out. What we have
no experimental control over is the electron as the original experiment used carbon
as a source of electrons. So, we need to use our equations to eliminate the electron’s
momentum pe and scattering angle φ.

28
 Squaring equation (1.24) gives,

h2 2 h2 h2
cos θ + − 2 cos θ = p2e cos2 φ = p2e (1 − sin2 φ)
λ02 λ2 λλ0
Using equation (1.23) in the above gives;

h2 h2 h2 h2
 
2 2 2
cos θ + 2 − 2 0 cos θ = pe 1 − 02 2 sin θ (1.26)
λ02 λ λλ λ pe

We can now use equation (1.25) to eliminate p2e from equation (1.26). From equation
(1.25);

hc hc p
+ mc2 − 0 = p2e c2 + m2 c4
λ λ
2 2 2 2 3
hc 2 4 hc hmc hmc3 h2 c2
⇒ 2
+ m c + 02 + 2 − 2 0 − 2 0 = p2e c2 + m2 c4
λ λ λ λ λλ
h2 h2 hmc hmc h2
⇒ + +2 − 2 0 − 2 0 = p2e
λ2 λ02 λ λ λλ
Plugging this in equation (1.26) gives;

h2 2 h2 h2 h2 h2 h2 hmc hmc h2
cos θ + − 2 cos θ + sin2 θ = + + 2 − 2 − 2
λ02 λ2 λλ0  λ 02
 λ2 λ02 λ λ0 λλ0
1 1 2h
⇒ 2mc − = (1 − cos θ)
λ λ0 λλ0
 
0 1 1 h
λλ − = (1 − cos θ)
λ λ0 mc
h
λ0 − λ = (1 − cos θ)
mc
This is a wonderful result. It tells us how much would the wavelength of the photon
change as a function of the scattering angle θ. In other words, this formula predicts
the wavelength of the scattered photon if it is found at the angle θ. This is a purely
kinematic result. It does’nt predict, for instance, the number of photons that are
expected to be scattered in a given direction. One requires a quantum field theory
treatment to compute such probabilities. Nevertheless, this is a most satisfying result
in that it confirms a particle picture for light.

29
Appendix A

Problems-Special Relativity

A.1 Problem 65 of Randy Harris

Consider the collision of two identical particles, each of mass m0 . In experiment A, a
particle moving at 0.9c strikes a stationary particle.
(a) What is the total kinetic energy before collision?
(b) In experiment B, both particles are moving at a speed u (relative to the
lab), directly towards one another. If the total kinetic energy before the collision in
experiment B is the same as that in experiment A, what is u?
(c) In both experiments, the particles stick together. Find the mass of the resulting
single particle in each experiment. In which more of the initial kinetic energy converted
to mass?

Experiment A
0.9c
m0 m0

Experiment B
u u
m0 m0

Solution:

(a) Only one particle has kinetic energy in experiment A and travels with speed

30
0.9c.
E = γ0.9c m0 c2
K.E = γ0.9c m0 c2 − m0 c2
= (γ0.9c − 1)m0 c2
 
1
= √ − 1 m0 c2 ≈ 1.29 m0 c2
1 − 0.81
(b) In experiment B , both particles have the same speed u and the same total
kinetic energy as experiment A.
K.E = 1.29 m0 c2
= 2[γ − 1]m0 c2
1.29
γ−1=
2
and γ ≈ 1.65
1
r = 1.65
u2
1− 2
c
u2
and 1 − 2 ≈ 0.367
c
u2
and 1 − 0.367 = 2
c
or u ≈ 0.8c
(c) Experiment A:

Ei = 2m0 c2 + 1.29m0 c2 = 3.29m0 c2

2
Ei1 = m20 c4 + p2i c2
So;
(m0 c2 + 1.29m0 c2 )2 = m20 c4 + p2i c2
(2.29)2 m20 c4 = p2i
4.24m0 c2 = p2i
pi = 2.06m0 c
Conservation of energy gives;
Ei = Ef
Ef = 3.29m0 c2

31
Conservation of momentum gives;

pi = pf = 2.06m0 c2

We already know,

Ef2 = m2f c2 + p2f c2

so; (3.29)2 m20 c4 = m2f c4 + (2.06)2 m20 c4
or m2f = 6.58 m20
or mf ≈ 2.57 m0

Experiment B:

Ei = 2m0 c2 + K.Ei
= 3.29 m0 c2
pi = 0
pf = 0
and Ef = Ei = 3.29 m0 c2
so; Ef2 = m2f c4 + p2f c2
Ef2 = m2f c4
and (3.29)2 m20 c4 = m2f c4
or mf = 3.29 m0

So there is a clear advantage in setting up experiments that have two colliding beams
oppositely directed. This is the configuration that we have been using for the post
few decades.

A.2 Problem 65 of Randy Harris

Both classically and relativistically, the force on an object is what causes a time rate
of change of its momentum: F = dp/dt.
(a) Using the relativistically correct expression for momentum, show that,

du
F = γu3 m
dt

(b) Under what condition does the classical equation F = ma hold?

32
 (c) Assume a constant force and that the speed is zero at t = 0, separate t and u,
then integrate to show that

1 F
u= p t
(1 + F t/mc) m
2

(d) Plot u versus t. What happens to the velocity of an object when a constant force
is applied for an indefinite length of time?

Solution: (a)

d
F = (γu mu)
dt 

dγu du
= mu + γu m
dt dt
−1/2
u2
 
d du
= 1− 2 mu + γu m
dt c dt
−3/2
u2
   
1 −2u du du
=− 1− 2 2
mu + γu m
2 c c dt dt
2
u du du
= m 2 γu3 + γu m
c dt 2 dt
du u 2
= γu m γ +1
dt c2 u
du u2
   
1
= γu m +1
dt c2 1 − uc22
u2 /c2 1 − u2 /c2 1
2 2
+ 2 2
= 2 2
= γu2
1 − u /c 1 − u /c 1 − u /c
du
so; F = mγu3
dt

(b) For u  c; clearly γu3 = 1 and the classical equation is good.

33
(c) Assume F is constant and u = 0 at t = 0. Then

du
F = mγu3
Z t Z u dt
F 1
dt = −3/2 du
m

0 0 u2
1 − c2
Z u
F 1
t= −3/2 du
m

0 2
1 − uc2

Set u = c sin θ
du = c cos θdθ
u=0 ⇒ θ=0
 
−1u
and u=u ⇒ θ = sin
c
Z sin−1 ( uc )
F c cos θ
So t= dθ
m 0 cos3 θ
Z sin−1 ( uc )
F 1
t=c dθ
m 0 cos2 θ
sin−1 ( u )
Ft  c
= tan θ
mc
0 
−1 u
= tan sin
c
u
=√
c2 − u 2
F 2 t2 2
⇒ 2 2
(c − u2 ) = u2
mc
F 2 t2 F 2 t2
 
2
⇒ =u 1+ 2 2
m2 mc
1 Ft
⇒ u= s  2 m
Ft
1+
mc

34
u
t u c

t
Effectively Linear

35