Chapter 11: Special Theory of Relativity

(Ref.: Marion & Heald, Classical Electromagnetic Radiation, 3rd ed.,

Ch. 14)
Einsteins special theory of relativity is based on the following two
postulates:
1. Laws of physics are invariant in form in all Lorentz frames (In
relativity, we often call the inertial frame a Lorentz frame.)
2 The speed of light in vacuum has the same value c in all Lorentz
2.
frames, independent of the motion of the source.
The basics of the theory are covered in Appendix A on an
elementary level with an emphasis on the Lorentz transformation and
relativistic momentum/energy.
gy Here, we examine relativityy in the
four-dimensional space of x and t, which provides the framework for
us to examine the laws of mechanics and electromagnetism. The
contents
t t off the
th lecture
l t notes
t depart
d t considerably
id bl from
f Ch 11 off
Ch.11
Jackson. Instead, we follow Ch. 14 of Marion and Heald.
Section 1: Definitions and Operation Rules of Tensors of
Different Ranks in the 44-Dimensional
Dimensional Space
Consider two Lorentz frames, K and K . Frame K moves along
the common z -axis axis with constant speed v0 relative to frame K .
Assume that at t = t = 0, coordinate axes of frames K and K
overlap Postulate
overlap. Postulate 2 leads to the following Lorentz transformation
equations for space and time coordinates. [derived in Appendex A
and B, Eq. (A.15) or Eq. (B.3), where the relative motion is assumed
to be along the x-axis.]
x x
x = x ( x, y, z, t )
i (x, y, z, t)
y = y K K
z, z
z = 0 ( z v0t ) v0 (1)
v0 y y Frames K and K
t = (t z )
0 c2 coincide at t = t = 0.
v02 12
where 0 (1 2 ) is the Lorentz factor for the transformation.
c
 11.1 Definitions and Operation Rules of (continued)

A note about notation: In many books,

books the relative speed between
two frames is denoted by v and the particle velocity in a given frame
is denoted byy u. This eventuallyy leads to two definitions for the same
notation :
2 1 Lorentz factor for the transformation,
v
(1 2 ) 2
c Jackson
J k (11
(11.17)
17)
2 1 Lorentz factor of a particle in a given frame,
(1 u2 ) 2 .
c Jackson (11
(11.46)
46) and (11 .51)
51)
To avoid confusion with the notation (e.g. when we perform a
Lorentz transformation of the Lorentz factor of a particle),
particle) we will
denote the relative speed between two frames by v0 and the particle
velocityy byy v throughout
g this chapter,
p , and thus define
v02 12
0 (1 2 ) [Lorentz factor for the transformation]
c
2 1
v
(1 2 ) 2 [Lorentz factor of a particle in a given frame].
c
 11.1 Definitions and Operation Rules of (continued)

Define a position vector of a point in the 44-dimensional

dimensional space of x
and t as
x ( x, y, z , ict ) = (x, ict )
1 0 0 0
4-vector spatial vector
0 1 0 0
and
d a 4-D t i as a =
4 D matrix , 0 = v0 / c
0 0 0 i 0 0
= 1 4, row number 0 0 i

= 1 4, column number 0 0 0

then, the Lorentz transformation in (1) can be written

x 1 0 0 0 x
y 0 1 0 0 y 4
= or x = a x (2)
z 0 0 0 i 0 0 z =1
ict 0 0 i 0 0 0 ict

4
and the inverse Lorentz transformation is: x = a x . (3)
=1
 11.1 Definitions and Operation Rules of (continued)

The a matrix in (2) shows that the Lorentz transformation is

an orthogonal transformation because it satisfies
definition of orthogonal
a a = (4)
transformation*
x byy ((2)) x byy ((2)) byy ((4))
Thus, x2 = a x a x = a a x x = x2
,
x 2 + y 2 + z 2 c 2 t 2 = x 2 + y 2 + z 2 c 2 t 2 ,
which is a statement of postulate 2 [see Eqs. (B.1) and (B.2) in
Appendix B.]
*See H. Goldstein, "Classical Mechanics," 2nd edition, p.134.
Just as the 3-dimensional vectors (and tensors in general) are
defined by their transformation properties in the x-space, we may
d fi 4-vectors
define 4 t (and
( d 4-tensors
4t in
i general) l) by
b their
th i transformation
t f ti
properties in the (x, t ) space and find rules for their operation.
 11.1 Definitions and Operation Rules of (continued)

1 Any set of 4 quantities A ( = 1 4) or A = ( A1 , A2 , A3 , A4 ),

1. ) which
transform in the same way as x , i.e.
A = a A , (5)

is called a 4-vector (or 4-tensor of the first rank).
Th position
The t x [ ( x, y, z , ict
iti vector i t )] off a point
i t in
i the
th 4-D
4D
space is obviously a 4-vector. As another example, the momentum
vector of a particle in the 44-D
D space
space, defined as
p ( p x , p y , p z , iE
c ) = (p , iE ),
c
is a 4-vector because it transforms as [see Eq. (A.28), Appendix A.]
px 1 0 0 0 px
p p
y = 0 1 0 0 y

4
or p = a p
pz 0 0 0 i 0 0 p z =1
iE iE
c 0 0 i 0 0 0 c
 11.1 Definitions and Operation Rules of (continued)

2. If a quantity is unchanged under the Lorentz transformation,

it is called a Lorentz scalar (or 4-vector of the zeroth rank). The
Lorentz scalar is also called a Lorentz invariant.
The Lorentz scalar is in general a function of the components
off a 44-vector.
t ForF example, l we have
h j t shown
just h that
th t
2 = x2
x

Hence, x2 is a Lorentz scalar.

 11.1 Definitions and Operation Rules of (continued)

4 D operator, [ x ,
3. Define the 4-D , , ],
y z (ict
i )
as the counterpart
of the operator in the x-space. Then, the 4-gradient of a Lorentz
scalar,
l [
x , , , ],
y z (ict ) ] is
i a 4-v
4 ector.
t
x
Prooff : ( ) =
x =
x x

= a


= a
x

( ) ((6))

= a by (3) Transforms as a 4-vector.
A
4. The 4-divergence of a 4-vector, A x , is a Lorentz scalar.

Proof : A: 4-vector
4 t
= a A
x by (5) A : component of A

x A A A
A = Ax = x x = a a x = x = A (7)

= a by (3) = by (4)

A is unchanged under the Lorentz transformation

 11.1 Definitions and Operation Rules of (continued)

2 2 2 2
5 The 44-Laplacian
5. Laplacian operator,
operator 2
=
+ 2
+ 2 1
2 2, (8)
x 2 y z c t
is a Lorentz scalar operator, i.e. 2 = 2 [ : a Lorentz scalar].
Proof : We have shown in item 4 that the divergence of a 4-vector
is a Lorentz scalar, i.e. A = A. Let be a Lorentz scalar, then
A = and A = are both 4-vectors (see item 3). Hence,
A = A = 2 = 2 .
6. The dot product of two 4-vectors, A B A B , is a Lorentz scalar.

Proof :
A B = by (4)
A B = A B = a A a B = a a A B

= A B = A B (9)

A A = A A A2 = A2 a useful property of the
orthogonal transformation
 11.1 Definitions and Operation Rules of (continued)
Examplep : In frame K , a particle's
p position
p changes
g byy dx in a
time interval dt. Then, d x = ( dx, dy, dz, icdt ) is a 4-vector. Hence,
d x d x (= dx dx ) is a Lorentz invariant, i.e. in frame K , d x d x

(= dx dx ) is given by d x d x.

SSpeciall case : The
Th particle
i l is
i at rest in
i frame
f K (the
( h rest fframe off
the particle). Hence, dx = 0 and d x = (0, 0, 0, icd ), where we have
denoted the differential time in frame K by d instead of dt , because
frame K is a unique frame. d is called the proper time of the particle.
dx = dx dx c 2 d 2 = dx 2 + dy 2 + dz 2 c 2 dt 2
dx

2
d = dt 1 2 = dt
v
[[a Lorentz invariant]] ((10))
c
dy
where v = dx
dt
e x + dt
e y + dz e is the velocity of the
dt z
particle in frame K .
Discussion : (i) For the special case that K is the rest frame of the
particle, v is also the relative velocity of the 2 frames. Hence, = 0 .
 11.1 Definitions and Operation Rules of (continued)

(ii) The Lorentz transformation applies only to inertial frames. If the

particle has an acceleration, d (= dt
) in (10) is the differential time in
the instantaneous rest frame of the particle, in which the particle has
zero velocity but a finite acceleration. In general, the speed of the rest
frame (hence ) is a function of time and d = dt(t ) [Jackson, (11.26)].
(iii) Consider an example with constant particle velocity. The muon
has a lifetime of 2.2 sec in its rest frame between birth and decay. If
the
h lif
lifetime
i isi measuredd in
i a Lorentz frame
f i which
in hi h theh muon has
h a
constant , then by (10), the rest-frame lifetime ( d ) and the measured
lifetime td are related by
decay tdecay dt t
birth d = d = d .
tbirth
This expresses the phenomenon of time dilation; namely, when the
time interval of a clock's rest time (e.g. d above) is observed in a
t
mo ing frame,
moving b a factor of . The invariance
frame it is greater by in ariance of d (=
( d )
td
means that will have the same value in all Lorentz frames.
 11.1 Definitions and Operation Rules of (continued)

7 A 4-tensor of the second rank (T ) is a set of 16 quantities

7. quantities,
T ( , = 1 4), which transform according to
= a a T
T (11)
,
8. The dot pproduct of a 4-tensor of the second rank and a 4-vector,
(T A) T A , is a 4-vector.

a a T a A

Proof : (T A) = T
A = a a a T
A
, ,

= a T A = a (T A) (12)

Transform as a 4-vector.
 11.1 Definitions and Operation Rules of (continued)

T
9. The
h 4-divergence
di off a 4-tensor off the
h second
d rank,k ( T ) x
,

is a 4-vector.
Proof :

T x
( T ) = x = x a a T = x x a a T
,
a
T T
= a a a x = a x = a ( T ) (13)
, ,
Transform as a 4-vector.
 11.1 Definitions and Operation Rules of (continued)

10 A 44-tensor
10. t off the
th thi
thirdd rankk is
i a sett off 64 quantities,
titi
G ( , , = 1- 4), which transform according to
= ai a j a k Gijk
G (14)
ijk

P bl m 1: If F is a 4-tensor
Proble 4 tensor of the second rank,
rank show
sho that
F
x ( , , = 1- 4) is a 4-tensor of the third rank.
= a j a k F jk
Solution : F
jk
ai

F F jk xi F jk
x = a j a k x x = ai a j a k x (15)
jk i i ijk i

Transform as a 4-tensor
of the third rank
rank.
 11.1 Definitions and Operation Rules of (continued)

Problem 2 : Show that the set of equations,

F F F
x + x + x = 0 ( , , = 1- 4) (16)

is invariant in form under the Lorentz transformation.

F F jk
Solution : Rewrite ((15):
) x = ai a j a k xi
ijk
Change indecies in (15) as follows:

F Fij
{i k , k j, j i
, ,

x = ai a j a k x (17)

{k j, i k,
ijk k
, ,
Ch
Change iindecies
d i in i (17) as follows:
f ll
j i
F F
x
= a a a ki
i j k x j (18)
ijk
Combine (15), (17), and (18), =0 by (16)

F
x


F
+ x


F F jk
+ x = ai a j a k x
ijk i ( Fij
+ x
k
Fki
)
+ x = 0
j
 11.1 Definitions and Operation Rules of (continued)

11. If a physical law can be expressed as a relation between 4-tensors

4 tensors
of the same rank, then it's form is invariant in all Lorentz frames.
Example 1: If the physical law in frame K is of the form A = B,
then, A = a A = a B = B , i.e. A = B A = B. (19)

B
Example 2: If the physical law in frame K is of the form T = F,
= a a T = a a F = F
then, T , i.e.

F

T = F T = F . (20)
In the following section, we examine relativistic mechanics in
4-vector formalism. In Sec. 3, we will demonstrate that laws of
electromagnetism are in variant under the Lorentz transformation
by expressing them as relations between tensors of the same rank.
F
From the
h LLorentz transformation
f i off these
h tensors, we also
l obtain
b i
the transformation equations for various electromagnetic quantities.
 Section 2: Relativistic Mechanics
We beging with a note on the terms "conservation",, "invariance",,
and "covariance".
The conservation of a quantity means that it remains unchanged
in time in a given Lorentz frame. For example, the relativistic
momentum and energy of an isolated system of particles are both
conserved d after
ft a collision.
lli i This
Thi isi a fundamental
f d t l law
l tot be
b discussed
di d
in this Section.
The invariance of a quantity means that it is invariant in value
under a Lorentz transformation. Such a quantity is called a Lorentz
invariant or Lorentz scalar. For example, p , the dot product
p of two
4-vectors is a Lorentz invariant. However, it may or may not be a
conserved quantity. An example will be provided in this section.
The term covariance refers to physical laws. A physical law is
"covariant" if it is "invariant in form under the Lorentz transformation."
A will
As ill be
b shown,
h the
th new laws
l off relativistic
l ti i ti mechanics
h i andd existing
i ti
laws of electromagnetism are all covariant.
 11.2 Relativistic Mechanics (continued)

The 4 - Momentum (p) of a Single g Particle :

As shown in (A.28), if we define the momentum of a particle as
p mv and energy as E mc 2 (m is called the rest mass*), then the
4-momentum, p ( px , p y , pz , iE
c ), is a 4-vector, which transforms as
px = px i Px , Py , Pz , E (
(21.1)
)
p = p K z (21.2)
y y
v0
pz = 0 ( pz c2 E ) i Px , Py , Pz, E (21 3)
(21.3)
E = 0 ( E v0 pz ) K z
(21.4)
v0
* h
*Throughout
h this
hi chap
h ter, m andd M ddenote the
h rest mass.
Discussion: In Appendix A, we first define p = mv and E = mc 2 ,
then show that the law of conservation of momentum and energy is
covariant. Conversely, from the requirement of the covariance of this
law, we can deduce the definitions of p = mv and
conservation law
E = mc 2 (see Jackson Sec. 11.5).
 11.2 Relativistic Mechanics (continued)
The dot pproduct of two 4-vectors is a Lorentz scalar, hence
2 2
p p = p p p 2 E2 = p2 E2 (22)
c c
i e E 2 p 2c 2 is a Lorentz scalar (invariant).
i.e. (invariant)
If frame K is the rest frame of the particle (i.e. p = 0, E = mc 2 )
2 E2
th p = (0,
then (0 0,
0 0, i ) andd p p = p p gives
0 imc i p 2 = m 2c 2 , or
c
E 2 p 2c 2 = m2c 4 (23)
Since E 2 p 2c 2 is a Lorentz invariant, (23) shows that the rest
mass m is a Lorentz invariant. This has in fact been assumed in
Sec. 2 of Appendix A, where we derive the Lorentz transformation
equations for p ( = mv) and E ( = mc 2 ). (23) is a useful formula
f iit relates
for l the
h particle's
i l ' totall energy (E ) to its
i momentum (p ). )
(Momentum in particle physics is often expressed in unit of GeV/c.)
F a relativistic
For l ti i ti particle,
ti l we can still
till speakk off its
it (macroscopic)
( i )
kinetic energy K , defined as: K = E mc 2 = ( 1)mc 2 . (24)
 11.2 Relativistic Mechanics (continued)

The 4 - Momentum (P) of a System y of Particles

Consider a system of particles, each with the 4-momentum
p j = ( pxj , p yj , pzj , iiE j / c) = (p j , iE
i j / c), j = 1,, 2,, 3,
Since the Lorentz transformation is a linear transformation, the
sum of anyy number of 4-vectors also obeys
y the Lorentz tansformation.
Thus, P = p j is a 4-vector and its components transform as
j
pxj = pxj (25 1)
(25.1)
j j
pyj = p yj (25.2)
j j
p = ( p v0 E ) (25.3)
j zj 0
j
zj c 2
j
j
E = ( E v p ) (25.4)
(25 4)
j j 0
j
j 0
j
zj

and P P = ( p j ) ( p j ) = ( p j ) ( p j ) ( E j c) 2 ((26))
j j j j j
is a Lorentz invariant.
 11.2 Relativistic Mechanics (continued)

pxj = pxj
j j
pyj = p yj
j j
Rewrite (25): p = ( p v0 E )
zj 0 zj c 2 j
j j j
E = ( E v p )
j j 0
j
j 0
j
zj

We see from (25) that only when all the components of P (i.e. the
three components of total momentum plus the total energy) are each
conserved in frame K will all the components of P be conserved.
If one component of P is not conserved, a rotation of the spatial
coordinate system can make any component of P (momentum or
energy) unconserved in the new spatial coordi
coordinate
nate system.
system Thus
Thus, the
relativistic law of conservation must take the form as described below
in order for it to be a covariant law.
law
 11.2 Relativistic Mechanics (continued)

Law of Conservation of Momentum and Energy :

For reasons just discussed, in relativity, the conservation of
momentum and energy gy comes in one law rather than separate p laws
for the momentum and energy as in nonrelativistic mechanics. The
law states that, for an isolated system of particles,
P ( before collision ) = P ( after collision ) , (27)
which implies that p xjj , p yj , pzjj , and E j are each conserved,
j j j j
i.e.
p j ( before collision ) = p j ( after collision ) (28)
j j
E j ( before collision ) = E j ( after collision ) (29)
j j
Since the law in (27) is expressed as a 4-vector relation, it has
the same form in all Lorentz frames [see (19)]. Thus, in frame K ,
we have P ( before collision ) = P ( after collision ) .
 11.2 Relativistic Mechanics (continued)

If P is conserved,, the dot product

p P P must also be conserved. Thus,,
Ej 2 Ej 2
( p j ) ( p j ) ( c
) = ( p j ) ( p j ) ( c
) (30)
j j j j j j
before collision after collision
Discussion :
(i) P P for an isolated system is both a Lorentz invariant [see (26)]
and a conserved quantity [see (30)]. If the system is not isolated, it is
still a Lorentz invariant
invariant, but no longer a conserved quantity.antity
(ii) P ( before collision ) = P ( after collision ) in (27) is a fundamental
law (rather than a derived relation),
relation) in which the nonrelativistic law
of conservation of momentum has been extended to include the energy,
E = mc 2 . A veryy important
p aspect
p of this law is that it applies
pp to all
processes in an isolated system, such as elastic and inelastic collisions,
nuclear reactions, and particle decays. As a result, the total rest mass
off the
h system may not be b conserved, d as iis illustrated
ill d in
i the
h following
f ll i
two problems.
 11.2 Relativistic Mechanics (continued)

Problem 1: Two identical particles of rest mass m and equal and

opposite velocities v collide head-on inelastically to form a single
particle. Find the mass and velocity of the new particle.
Solution :
m, v v m, (before)
The total momentum before
M cm (after)
the collision is mv mv = 0.
So the collision occurs in the center-of-momentum (CM) frame, i.e. the
frame in which the sum of the momentum of all particles vanishes. We
denote the mass of the new particle as M cm to indicate that it is created
i the
in th CM frame.
f However,
H the
th CM frame
f is
i the
th only
l frame
f we willill use.
So m, , and v of the colliding particles are also CM frame quantities.
The new particle
particle'ss mass is designated M cm here for later convenience
when we compare it with the new particle's mass in problem 2.
Conservation of momentum The new particle is stationary. nary

Conservation of energy m + m = M cm M cm = 2 m
 11.2 Relativistic Mechanics (continued)

Discussion:
Di i I this
In thi problem,
bl we m, v v m, (before)
find Mcm= 2m > 2m, i.e. rest mass M cm (after)
has been created from the kinetic energy [( 1)mc2] of the colliding
[(-1)mc
particles. There is no need to know whats inside the new particle. We
only need to know its rest mass and hence the energy associated with
it. A hot object has a rest mass greater than when its cold. The
difference in rest mass due to an increase in temperature can in
principle be measured by its acceleration under a known force,
force and we
know that at least some of the added mass is in the form of thermal
energy. In many other cases, it
itss not possible to know what
whatss inside.
Nuclear fusion and fission reactions are examples of non-
conservation of rest mass. The total rest mass is reduced after the
reaction and the mass deficit appears as kinetic energies and radiation.
In fact, all reactions (chemical or nuclear) in which energy is absorbed
(e g photosynthesis) or released (e.g.
(e.g. (e g digestion of food) involve a
corresponding change of the reactants total rest mass.
 11.2 Relativistic Mechanics (continued)
Problem 2 : A p particle of rest mass m and velocityy v collides with
a stationary particle of the same rest mass and is absorbed by it. Find
the rest mass and velocity of the new particle.
Solution : The collision occurs in the stationary-target (ST) frame. So,
we denote the new particle mass by M st , velocity by Vst , and Lorentz
f t by
factor b st [ = (1 Vst2 / c 2 ) 1/ 2 ].
] (m, , v are also
l ST frame
f quantiti
tities.))
Conservation of momentum mv = st M st Vst (31)

Conservation
C ti off energy ( + 1)m = st M st (32)
(31)
((32))
Vst = +1
v m, v m ((before))
(32) M st = +1 m M st , st Vst (after)
st
2
2 ( +1) Vst2 2 2v 2
M st2 =m 2
= m ( + 1) 2
(1 2 ) 2
= m ( + 1) 2
[1 2 2]
st2 c c ( +1)
2 2 2 v2 2 2 2 2
= m ( + 2 +1 2 ) = m [ v
(1 2 ) + 2 + 1] = 2m 2 ( + 1)
c c
M st = 2( + 1)m
 11.2 Relativistic Mechanics (continued)

Discussion :
In problem 1 (CM frame), the new m, v v m, (before)
particle's mass is M cm = 2 m. (33) M cm (after)
In problem 2 (ST frame), the new m, v m (before)
particle's mass is M st = 2 (1 + )m. (34) M st , st Vst (after)
Note that is the Lorentz factor of the particle(s) before collision.
In particle physics experiments, M cm c 2 or M st c 2 is the energy
available for the creation of new particles (why not st M st c 2?).
The rest energy of the electron or positron is mc 2 = 0.511 MeV. If
2 TeV of energy is needed for particle creation (i.e. M cm c 2 = 2 TeV
or M st c 2 = 2 TeV), then the required of the colliding particle(s) is
by (33), M cm c 2 = 2 mc 2 = 2 TeV 1.957 106 [CM frame]
.
by (34), M st c = 2 (1 + )mc = 2 TeV 7.66 10 [ST frame]
2 2 12

The energy associated with is to be obtained in an accelerator.

 11.2 Relativistic Mechanics (continued)

Thus,
Th
kinetic energy needed in CM frame 2(1.957106 1) 7
gy needed in ST frame
kinetic energy
= 5 10
7 661012 1
7.66
This shows that far less kinetic energy is needed in the CM
frame than in the ST frame. In fact, all the kinetic energy
gy of the
two colliding particles [2 (1.957 106 1) 0.511 MeV = 2 TeV]
is put in use in the CM frame, while in the ST frame, 99.99995%
of the kinetic energy of the incident particle is wasted! This is why
the International Linear Collider (ILC) project plans to accelerate
both electrons and positirons to energies up to 1 TeV so that the
the collision occurs in the CM frame.
Question: Why use a long linear accelerator instead of a more
compact circular accelerator?
 Section 3: Covariance of Electrodynamics
In the special theory of relativity, Newton
Newton'ss law has been radically
modified. The electromagnetic laws do not need any modification
because they are already covariant. However, the covariance of these
laws (such as Maxwell equations) is not immediately clear from the
equations by which they are usually represented.
Our purpose ini this
hi section
i is i to prove that
h the h EM llaws are indeed
i d d
covariant by casting them into relations between 4-tensors of the same
rank [see (19) and (20)].
(20)] We will do this backward
backward. That is,
is we first
define 4-tensors in terms of known EM quantities and form equations
with 44-tensors
tensors of the same rank, then show that one or more existing
EM laws are implicit in each equation. This will prove that the laws
are convariant and the defined quantities are legitimate 4-tensors.
Furthermore, Lorentz transformations of these tensors will yield
the tranformation equations for various EM quantities.
Note: Jackson
k switches
i h to theh Gausian i unit
i system starting
i fro
f m
Ch. 11. From here on, we also switch to the Gaussian unit system.
 11.3 Covariance of Electrodynamics (continued)
1. Define a 4-current as
J ( J x , J y , J z , ic ) = (J , ic ) (35)
and use it to form a ((covariant)) 4-scalar relation
J = 0 (36)
Then, (36) gives the law of conservation of charge
(ic )

x
J x + y J y + z J z + (ict ) = 0 J + t = 0 (5.2)
Thus the definition of J in (35) as a 44-vector
Thus, vector leads to the
covariant representation [(36)] of the EM law in (5.2). This
in turn jjustifies the definition of J as a 4-vector. The Lorentz
transformation of J then gives
J x = J x i Jx, J y, Jz,
J y = J y K z

J z = 0 ( J z v0 ) (37)

i Jx, J y, Jz,
v0
= 0 ( c2 J z )
K z
v0
 11.3 Covariance of Electrodynamics (continued)

2. potential as A ( Ax , Ay , Az , i )
2 Define a 44-potential (38)
2 A = 4 J (39)
and write the covariant relations: c
A = 0 (40)
2 A 1 2 A = 4 J (6.15)
2
c t 2 c
(39) 2
2
c 2 t 2 = 4
1 (6.16)
(40) A + 1c t = 0 [Lorenz condition] (6.14)
This again shows the consistency of A being a 4-vector and
(6.14)-(6.16) being covariant laws. The Lorentz transformation
Ax = Ax i Ax , Ay , Az ,
Ay = Ay z
K
of A then gives v (41)
Az = 0 ( Az c0 )
i Ax , Ay , Az ,
= 0 ( v0 Az ) K z
c v0
 11.3 Covariance of Electrodynamics (continued)

Note: The source-free

source free wave equation can be directly put into the
2
covariant form: 2
2 2
1 =0 2
= 0. (42)
c t
3. Define a 4-wavenumber as
k (k x , k y , k z , ic ) = (k , i
c
) (43)
Then, k x = k x k x t = k x t
Invariance of the phase
By the same argument, we find that k defined in (43) is a
legitimate 4-vector. Thus, its Lorentz transformation gives
k x = k x
k y = k y i kx , k y , kz ,
K z
v0 (44)
k z = 0 (k z c 2 ) i k x , k y , k z ,
= 0 ( v0 k z ) K z
relativistic Doppler shift v0
 11.3 Covariance of Electrodynamics (continued)

4. Define a field strength tensor of the second rank F [Marion, (14.62)]:

0 Bz By iE x

Bz 0 Bx iE y
F (45)
B Bx 0 iE z
y
iE x
iE y iE z 0

E = 4
Then, F = c J
4
1 E = 4 J
B c t c
In the covariant set of equations [see (16)]
F F F F 's are elements
x
+ x + x = 0 ( , , = 1- 4) ,

off F in
i (45).
(45)
set ( , , ) = (1, 2,3) B = 0
sett ( , , ) = (1,
(1 2,
2 4),
4) (1,3,
(1 3 4), (2 3 4) E + 1c Bt = 0.
4) andd (2,3, 0
 11.3 Covariance of Electrodynamics (continued)

F F F
Th covariant
The i t equations,
ti F = c J andd x
4 + x + x = 0,
0

give the set of Maxwell equations in free space. This shows that
Maxwell equatins are covariant as well as justifies the definition of
= a a F gives
F as a tensor of the second rank. Thus, F
,
the transformation equations for E and B (see Marion, Sec. 14.6.)
E = E i E , E , B , B
( )
v0

E = 0 E + c B
K
B = B (46)

i E , E , B , B

(
B = B v 0 E
0 c ) K
v0
In (46), v 0 is the velocity of frame K relative to frame K , and
" " and " " refer to the direction of v 0 .
See Appendix C for a summary of transformation equations.
 11.3 Covariance of Electrodynamics (continued)

5. equation* d P = e F P (d is a Lorentz scalar),

5 The covariant equation*, scalar)
d mc
gives (Marion, p.439)
d p = e(E + 1 v B) relativistic ti
l ti i ti equation (47)
dt c
of motion


mc 2 d = ev E This equation is
dt implicit in (47).
*In order for this equation to be covariant, the charge e must be
a Lorentz invariant. This has been experimentally established
(see Jackson, p.554).
6. In a similar manner, we can demonstrate the covariance of the
conservation laws for field/mechanical momentum and field/
mechanical
h i l energy, as given i by
b Jackson
J k (6 111) and (6.122):
(6.111) (6 122)
d (E + Efield ) = s n Sda (6.111)
dt mech See Marion,
d Sec. 14.12. (6.122)
dt ( p mech + p field ) =
s T n da

 11.3 Covariance of Electrodynamics (continued)

Consider the ggeneral form of the relativistic equation

q of motion in
(47), d p = F, where F is any force, such as the gravitational force.
dt
Special case 1: F v (one-dimensional
(one dimensional problem)
F= d ( mv )
dt
= mv ddt + m ddtv = m ddtv (2 v2
2 + 1) = 3m ddtv (48)
c

d = d (1 v2 )1/ 2 2 v2 v2 / c2
dt dt c2 2 +1 = 2 2 +1
2 3/ 2 2v dv c 1v / c
1 v
= 2 (1 2 ) ( 2) = 3 v2 dv v 2 / c 2 +1v 2 / c 2 1
c c dt c dt = = = 2
1v 2 / c 2 1v 2 / c 2
F = 3ma Constant
C t t force
f does
d nott cause constant
t t acceleration.
l ti
Special case 2: F v ( = const., as in uniform circular motion)
F = d p = d ( mv ) = m d v (49)
dt dt dt
Questions: (i) It is sometimes said that a particle has two masses,
3m and m. Why? (ii) The acceleration is not necessarily
parallel to the force. Give an example. (iii) Relate (48) to (A.23).
 11.3 Covariance of Electrodynamics (continued)

Problem 1 : A p police radar operates

p on a frequency
q y of . What is
the frequency received by the police after the signal is reflected from
a car moving at the velocity v 0? v0

Solution : We do it in 2 steps. , k k
police car
Step 1. In the police frame, the
radar sends a wave ( , k ) toward shown in police frame
the car, which is moving at velocity v 0 (direction shown in the figure).
T
Transformming
f i to t the
th car fframe bby (44)
(44), we obtainbt i
= 0 ( v0 k z ),
where
h k z is i the
th componentt off k alongl i k z = k cos (see
v 0 , i.e. ( figure.)
fi )
Thus,
k =/c 0 = (1 v 02 c 2 )1/ 2
v0 cos
= 0 ( k z v0 ) = 0 ( kv0 cos ) = 0 (1 c )
This is
Thi i the
th Dopper-shifted
D hift d frequency
f d t t d by
detected b the
th car. It is
i also
l
the frequency of the wave reflected by the car as seen in the car frame.
 11.3 Covariance of Electrodynamics (continued)

Step 2. In the car frame (see figure), the car sends the reflected
wave ( , k ) back to the car at the frequency
v0 cos ppolice
= 0 (1 c )
k
car

, k
In the car frame, the police is
v0
moving at velocity v 0 (direction
shown in car frame
shown in the fugure) relative to
the car. Tranformming g to the police
p frame byy (44),
( ), we obtain the
frequency observed by the police (Doppler shifted again)
v0 cos
= 0 ( k z v0 ) = 0 ( k v0 cos ) = 0 (1 c )
v0 cos 2 v0 cos
= 02 (1 c ) (1 2 c ) since v0 c.
If the radar frequency is f (= / 2 ) = 109 Hz and the car moves
away from the police ( = 0) at v0 = 150 km/hr, the police would
22vv
detect a frequency f (= / 2 ) shifted by f f c0 278 Hz.
 11.3 Covariance of Electrodynamics (continued)

Problem 2: An observer in the laboratoryy sees an infinite electron

beam of radius a and uniform charge density , moving axially at
velocity v0. What force does he see on an electron at a distance r ( a)
from the axis? Assume the electron moves axially at the velocity v0.
Solution: The problem can be readily solved in the lab frame. Here,
we will take a long route for an exercise on some of the transformation
equations just derived.
a , J , v (= v )
The current density J z in the lab frame is z z 0

J z = v0 . [ has a negative value.] K z

Byy ((37),
), we have,, in the beam frame a
J z = 0 ( J z v0 ) = 0, (= / 0 ), J z = vz = 0
v0 v02 K z
= 0 ( 2 J z ) = 0 (1 2 ) = . v0
c c 0
We see that the lab frame is greater than the beam frame by
the factor 0 . This is because every unit length of the beam in its rest
frame is contracted by this factor when viewed in the lab frame.
 11.3 Covariance of Electrodynamics (continued)

frame, J z = 0,
In the beam frame 0 = 0 ; hence,
hence there is only a radial
electric field. Gauss law, s E da = 4 v d 3 x, E , B (E = B = 0)

2 K
th gives
then i 2 r Er = 4 ( r ), f r a
) for
E (E = B = B = 0)
2 r
Er = 2 r = . [r = r , = 0 ] K
v0
0
We now transform E ( = Er er ) into lab-frame E and B by using
the reverse transformation equations in (46), in which we set v 0 = v0e z .
E = (E v0 B ) = E = 2 r e = 2 re
0 c 0 0
0
r r
v0 v0e z 2 r v0

B = 0 ( B
+ E ) = 0 ( ) e r = 2 re
c c 0 c
Thus, the force f on an electron (in the lab frame) is
f = e ( E + c v B ) = e 2 re r + c (v0e z ) ( c 2 re )
1 1 v0

v02 2 e r e e is positive
= 2 e r (1 2 )er = 2 er
positive. For an
c 0 electron beam, is negative.
 Appendix A: Relativity in College Physics
(Ref Halliday,
(Ref. Halliday Resnick,
Resnick and Walker,
Walker Fundamentals
Fundamentals of Physics
Physics))
Section 1: The Lorentz Transformation
The Galilean Transformation: Consider 2 inertial frames S and
S . Frame S moves along the common x -axis* with constant speed v0
relative to frame S . At t = 0,, the coordinates coincide and,, at time t ,
the position of point P is (x, y, z ) in S and (x, y, z ) in S . Then the
Galilean transformation gives
t is unchanged in
x = x v0t , y = y, z = z , t = t the transformation. (A.1)

* In the main text, the z -axis is the
y y
di i off relative
direction l i motion. i To
T
S x S
be consistent with the references x
cited in this appendix,
appendix here we
iP
assume that the relative motion
is along the x-axis. y y
Question: How do you determine x x
a reference frame is inertial? v0t v00

 11.A.1 The Lorentz Transformation (continued)

Postulates: The laws of classical mehanics do not vary

Einstein s Postulates:
Einstein's
in form under the Galilean transformation. For example, (A.1) shows
F = ma in frame S transforms to F = ma in frame S . However, when
the same transformation is applied to the wave equation in vacuum,
2 2

2 2 = 0, its form changes completely (see Jackson, p. 516.)
1
c t
So, when Einstein began his work on relativity, there were two
approaches to make ALL the laws of physics invariant in form in all
inertial frames: (1) Modify the theory of electromagnetism so that it is
invariant in form under the Galilean transformation; or (2) Modify the
Galilean transformation and the laws of mechanics so that the laws of
both mechanics and electromagnetism are invariant in form under the
pp
new transformation. Einstein took the second approach. His special
p
theory of relativity is based on 2 postulates:
1. Laws of physics are invariant in form in all inertial frames.
2. The
h speedd off light
li h in
i vacuum has
h theh same value
l c ini all
ll inertial
i i l
frames, independent of the motion of the source.
 11.A.1 The Lorentz Transformation (continued)
Event and Simultaneity: An event is something (such as the
emission of a light pulse by a source) which happens at position (x, y,
z) and time t. An event [described collectively as (x, y, z, t) in a given
f
frame] ] will
ill have
h diff
different
t coordinates
di t in i different
diff t frames.
f Th
The
frames mentioned here and later are all inertial frames.
The time of an event can be measured by methods we normally
think of. But, in relativity, time measurement often requires high
precision (which can at least be done in a thought experiment) and
we must bear in mind the frame in which the time is measured. The
simplest way to measure time is to read the clock at the position of
the event.
event If the clock is away from the event,
event the time of the event
is the time shown on the clock (at the instant the light signal of the
event reaches the clock) minus the time delay due to the travel of the
signal (at speed c) from the events position to the clocks position.
The position of the event and the measured time of the event all refer
to the frame in which the observer and the clock are both at rest (but
the source which generates the event is not necessarily at rest.)
 11.A.1 The Lorentz Transformation (continued)
Two events are simultaneous in a reference frame if they have the
same time coordinate in that frame, whether or not they have the same
spatial coordinates. Simultaneity can be experimentally tested as
follows. If two events are detected at the same instant by an observer
located midway, they are simultaneous in the observers frame.
Within a given
gi en frame,
frame the concept of space and time in the special
theory of relativity is not different from our usual concept of space
and time.
time However,
However radical differences arise when space and time
coordinates of an event measured in one frame are compared with
those measured in another frame. In makingg the comparison,
p , we find
that space and time are entangled with each other in relativity. For
example, two simultaneous events occurring at different positions in
frame S will no longer be simultaneous in frame S, and their time
difference in S depends upon their spatial separation in S. In relativity,
space and d time
ti coordinates
di t transform
t f according
di t the
to th Lorentz
L t
transformation, which are derived below from postulate 2.
 11.A.1 The Lorentz Transformation (continued)

Time Dilation: Consider a pulse of light emitted by a source on a

train (event 1). It travels vertically upward for a distance D, then is
reflected back by a mirror, and later detected at the source (event 2).
In the train frame (Fig. 1), the time interval between the 2 events is
t0 = 2cD . (A.2) mirror

In the lab frame (Fig. 2), the train, mirror,

and source are all moving at speed v0 , but the D Fig. 1
light still travels at speed c (by postulate 2).
2) Event 1 Event 2 (at the same
location as event 1)
So, the time interval of the 2 events is source
t = 2cL , (A.3)
mirror
where L = [( 12 v0 t )2 + D 2 ]1/ 2 (A.4)
Fig. 2
Eliminating D and L from L D L
(A.2)-(A.4), we obtain Event 1 Event 2
t = 0 t0 , (A.5)
(A 5) source source
2 1/ 2 v0 t
where 0 [1 v02 /c ] . (A.6)
 11.A.1 The Lorentz Transformation (continued)

Q
Question : Whyy is D the same in both frames?
Lengths perpendicular to the direction of motion are the same in
both frames, i.e. the y and z coordinates transform as:
y = y, z = z (A.7)
The proof of this is by contradiction.
S
Suppose that
h we have
h two identically
id i ll vx
manufactured pieces of pipe (see figure).
They cannot fit inside each other because vx
they have identical radius. Imagine that one
from H. C. Ohanian, "Physics"
pipe is at rest on the ground and the other
is at rest on the train. If the motion of the train relative to the ground
were to bring g about a transverse contraction of the train p pipe,
p then by y
symmetry, the motion of the ground pipe relative to the train would
have to bring about a contraction of the ground pipe. But these two
effects are contradictory, since in one case the train pipe would fit
inside the ground pipe, and in the other case it would fit outside.
 11.A.1 The Lorentz Transformation (continued)
Going back to (A.5): t = 0t0. In this equation, t0 is the time
interval of 2 events measured in a special frame in which the 2 events
occur at the same position. It is called the proper time. Viewed in any
other
h frame,
f these
h 2 events willill occur at different
diff positions
ii and,
d by
b
(A.5), their time interval (t) will be greater than the proper time by a
factor of 0. This is known as the effect of time dilation.
dilation
The muon has an average lifetime of 2.2 sec (between birth and
decay) in its rest frame.
frame In a 1977 experiment at CERN, CERN muons were
accelerated to a speed of 0.9994c, corresponding to a 0 value of 28.87.
Within experimental
p error, the measured average g lifetime of these
muons was indeed 28.87x2.2 = 63.5 sec. In another experiment, two
synchronized clocks with near perfect precision showed slightly
different readings after one had been flown around the world. The
difference was again in agreement with (A.5).
Ti dilation
Time dil ti runs counter t tot our intuition,
i t iti b
because it is
i rooted
t d in
i a
postulate which also runs counter to our intuition.
 11.A.1 The Lorentz Transformation (continued)

The Twin Paradox: Supposepp someone travels on a spaceship

p p with
a Lorentz factor of 0 = 20 (in the earth frame) and his twin brother
stays on earth. Then, by time dilation, every day measured by the
traveling
li twin i in
i the
h spaceship
hi frame
f ( hi is
(this i his
hi proper time)
i ) will
ill be
b
20 days when measured by the earth twin in the earth frame. So the
earth twin ages faster and his traveling brother will be 19 years
younger when he returns to earth after an 1-year journey (neglect the
spaceships
p p acceleration/deceleration pperiods).) The pparadox is: if the
traveling twin measures the age of his earth twin, will he conclude that
he himself ages faster by the same argument of time dilation?
Th
There i no paradox
is d at all.ll Only
O l theh earthh twins
i measurement is i
correct because he is always in an inertial frame. The traveling twin
will have to be accelerated and decelerated in the spaceship.
spaceship During
these periods, he cannot use the special theory of relativity (Einsteins
2 ppostulates refer to inertial frames.)) In fact, he will confirm the
measurement of his earth twin if he uses Einsteins general theory of
relativity, which deals with accelerating reference frames.
 11.A.1 The Lorentz Transformation (continued)
Length Contraction: Assume that planet neptune is stationary in
the earth frame and at a distance L0 from earth (Fig. 1). A spaceship is
traveling at speed v0 to neptune. The duration of the trip, measured on
earth,
th is
i t = L0 / v0 . (A 8)
(A.8)
In the spaceship's frame
(Fig 2)
(Fig. 2), both earth and neptune Fig.
g 1
v0
move at speed v0 . The duration earth spaceship neptune

p, t0 , is the interval
of the trip, v0 v0 Fig. 2
between the departure of the from Giancoli, "Physics for Scientists and Engineers"

earth and the arrival of the neptune. This is the "proper time" of the
spaceship
hi because
b both
b h events occur at the h same position. ii Thus,
h
by (A.5) t0 = t / 0 , (A.9)
t0 can be used to calculate the earth-naptune
earth naptune distance as viewed
on the spaceship L = v0 t0 . (A.10)
Eliminating t and t0 from (A.8)-(A.10),
(A.8) (A.10), we obtain
L
L = 0 (A.11)
0
 11.A.1 The Lorentz Transformation (continued)
), L = L0/0, L0 is the length
In ((A.11), g of an object
j ((or,, in the above
example, the earth-neptune distance) measured in the rest frame of the
object (i.e. the frame in which the object is at rest). Length measured
i this
in hi special
i l frame
f i called
is ll d the
h proper length.
l h Viewed
Vi d in
i any other
h
frame, the object will be moving and, by (A.11), its length will be less
than the proper length by a factor of 0. This is known as the effect of
length contraction. Note that the contraction effect applies only to
lengths
g along g the direction of motion.
Length contraction is a direct consequence of time dilation [see
(A.9)]. It is therefore not surprising that time dilation can be inferred
from length contraction. If, for example, the spaceship has a 0 value
of 2. The earth-neptune distance, as measured in the spaceship, would
be half of that measured on earth. earth But the speed of earth/neptune
relative to the spaceship is still v0. So, to the spaceship, the journeys
duration is only half of that measured on earth. Hence, one minute
elapsed in the spaceship will be 2 minutes elapsed on earth.
 11.A.1 The Lorentz Transformation (continued)
The Lorentz Transformation: Assume frames S and S coincide at
t = 0 and S moves along the common x -axis with speed v0 relative to S
(see figure). A point P has coordinates (x, y, z , t ) in S and (x, y, z , t )
i S . The
in l th x, when
Th length h measuredd in i S , isi x (length
(l th contraction).
t ti ) So, S
0
x = v0t + x or x = ( x v0t ) . (A.12)
0
By symmetry or by similar argument, x = ( x + v0t ) (A.13)
) [usingg 02 1 = 02 v02 c 2 ],
Eliminatingg x from ((A.12)) and (A.13)
(
v
t = (t 02 x) (A.14)
c
(A.7),
(A 7) (A.12),
(A 12) and (A.14) (A 14) give y y
the Lorentz transformation: S x S
x = 0 ( x v0t ) x
y = y iP

z = z y y (A.15)
v0 x
t = 0 (t x) x
c 2 v0t vv00
 11.A.1 The Lorentz Transformation (continued)
Transformation of Coordinate Difference between 2 Events :
Since the Lorentz transformation is linear, the coordinate differences
between 2 events:
i S : x = x2 x1 , y = y2 y1 , z = z2 z1 , t = t2 t1 (A.16)
in (A 16)
in S : x = x2 x1 , y = y2 y1 , z = z2 z1 , t = t2 t1 (A.17)
transform
f in
i the
h same manner. Thus,
Th
x = 0 ( x v0 t )
y = y

z = z (A.18)
v0
t = 0 ( t 2
x)
c
i Event 1 x = x2 x1 i Event 1 x = x2 x1
(x1 , y1 , z1 , t1 ) (x1 , y1 , z1 , t1 )
y = y2 y1 y = y2 y1
i Event 2 i Event 2
z = z2 z1 (x2 , y2 , z2 , t2 ) z = z2 z1
(x2 , y2 , z2 , t2 )
t = t2 t1 t = t2 t1
x x
S S v0
 11.A.1 The Lorentz Transformation (continued)

Discussion on simultaneity: Consider the transformation equation

for the time interval between two events
v
t = 0 (t 20 x) [[from (A.18)]
( )] A
c c
B
c source
It indicates that 2 simultaneous S x
events in frame S (t = 0) which v0 c c v0
A B
occur at different positions (x 0) source
S x
will not be simultaneous in frame S v0
(t 0). This can be explained on the basis of postulate 2 through the
followingg example.
p
In frame S , a pulse of light emitted midway between points A and
B ((see figure)
g ) will reach A and B at the same time, i.e. the two events
(arrivals of the signals at A and B ) are simultaneous in frame S . In
frame S , the signal still travels at speed c in both directions, but B is
moving toward the light and A away from it. So, the signal will reach
B first and the two events are no longer simultaneous.
 11.A.1 The Lorentz Transformation (continued)

The examplep discussed above

2x 2x
can be examined quantitatively as
c c
follows. A B
source
Assume that,
h iin frame
f S , the
h S x
two events are spatially separated
2x0 2x0
by a distance x. Observed in S ,
v0 c c v0
the distance is shorter by a factor A source B
of 0 due to length
g contraction,, i.e. S x
v0
x = x .
0
Th in
Thus, i frame
f S , the
th signals
i l reachh A and d B with
ith a time
ti
difference of
x x
2 2
t = t B t A = c+v0 cv0 = x v0 = 0v0 x.
0 0 0 c 2 v02 c2
This is precisely the prediction of (A.18),
v v
t = 0 (t 02 x) = 020 x. [t = 0 in frame S ]
c c
 11.A.1 The Lorentz Transformation (continued)

Problem 1: In frame S , events A and B occur at different p positions,,

and event B occurs after event A. Is it posible for event B to precede
event A in another frame S moving at speed v0 relative to frame S?
If so, does this mean that an effect can precede its cause?
Solution: In frame S , let the 2 events have a spatial interval
x = xB x A andd ti
time interval
i t l t = t B t A . Then
Th the
th time
ti interval
i t l
v
in frame S , t = t B t A , is given in (A.18): t = 0 (t 20 x).
2
c
We see that if t < v0 x / c , then t < 0, which means that the
order of independent events in frame S may be reversed in frame S .
S
Suppose , however,
h th
thatt the
th events
t are connected,
t d i.e.
i eventt B is
i
caused by event A. This would require a body, or a signal, to travel
v0 x
f
from A tto B. Rewrite
R it (A.18)
(A 18) as t = 0 t (1 2 t ).
) Since
Si th fastest
the f t t
c
speed for a signal to travel from A to B is x = c, we must have v > c
t 0
in order for t < 0. This is not possible [see (A.6)] and thus the order
of connected events (cause and effect) cannot be reversed.
 11.A.1 The Lorentz Transformation (continued)

Problem 2: Show that the effects of time dilation and length

contraction are implicit in the Lorentz transformation.
Solution : The time interval between 2 events transform as
v0 v0
t = 0 (t 2 x) or t = 0 (t + 2 x ), where the x-direction
c c
y If t is the pproper
is arbitrary. p time in S , then the 2 events occur at
the same position (x = 0). So we use the latter equation and obtain
t = 0 t (time dilation).
The difference in the x coordinates of the 2 events transform as
x = 0 ( x v0 t ) or x = 0 (x + v0 t ). Again, the question is
which
hi h equationi to use. If x is i the
h "proper
" h" iin S , then
llength" h the
h
two end points are at rest and their coordinates do not have to be
measured simultaneously (i.e. (i e we do not know t .)) But since the
rod is moving in S , its end points must be measured simultaneously
in S (t = 0).) So we use the former equation
q and obtain
x = x (length contraction).
0
 11.A.1 The Lorentz Transformation (continued)
Transformation of Velocity : The velocity of a particle is given by
v = lim xt (in frame S ); v = lim xt (in frame S ). (A.19)
t 0 t 0
v v
x = vx t i i
Let t 0, S x S x
( x v t ) v v v0
vx = x = 0 0
t (t 0 x) 1 v0 v
v = x 0
x = 0 (x v0 t ) 0 2 2 x
y = y c c
y y vy
z = z vy = t = v0 = v (A.20)
v0 0 (t 2 x) 0 (1 2 vx ) 0

t = ( t x ) c c
0
c2 v = z = z =
vz
z t 0 (t v0 x) 0 (1 v0 vx )
c2 c2
Problem: A spaceship moves away from the earth at speed v0 .
A pulse of light is emitted from the earth in the direction toward the
sapceship. What is the speed of light measured on the spaceship?
v v c v
Solution: vx = c vx = xv0 0 = v00 = c
1 2 vx 1 c
c
 11.A.1 The Lorentz Transformation (continued)

Transformation of Acceleration : For simplicity

simplicity, we first
consider the transformation of acceleration in the direction of
dvx
relative motion (i.e.
(i e dtd
)
vx v0

vx = v0 [ from (A.20)] i vx ,
dvx
1 2 vx dt
c
( v v )( v0 ) dv S x
dvx x 0 2 x dvx

dvx = v0 c = 2 v0 2 dvx
1 2 vx v
(1 2 vx )
0 2 0 (1 2 vx )
i vx ,
c c c dt
v0 S x
t = 0 (t 2
x) [from (A.15)] v0
c
v0
dt = 0 (dt dx) dx = vx dt
c2
dvx 1 dvx 1 dvx
Hence, = v v = 3 v0 3 (A.21)
dt 0 ( dt 2 dx ) 0 (1 2 vx )
0 2 0 2 0 (1 2 vx ) dt
c c c
Question: In what frame does the acceleration have the largest
value? (see next problem.)
 11.A.1 The Lorentz Transformation (continued)

By the same method, we may obtain the transformation equations

for acceleration in arbitrary directions (see Jackson Problem 11.5).
a
a = 1 v
3a i
03 1 v02v S
c
a (A.22)
a = 1 a v 0 ( a v ) v
3 c 2 i
02 1 v02v S
c
v0
where " " and " " refer to the direction of v 0 .
 11.A.1 The Lorentz Transformation (continued)
Problem : A rocket is launched from the earth into outer space.
It moves on a straight line with a constant acceleration (a) with
respect to its rest frame (Why is a specified in the rocket's frame?)
Calculate the time required for the rocket to accelerate from zero
speed to the final speed v f , according to earth and rocket clocks.
Solution : Let S be the earth frame, S be the rocket rest frame,
and the one-dimensional motion be along the x-axis. The inverse
transformation of (A.21) gives (omitting subscript "x"):
i
a = dv = 3 1
dv . v = v , dv = a
dt 0 (1+v0v / c2 )3 dt 0 dt 03
Lorentz transformations
transformations apply only to inertial S x
i
frames. So, S is the instantaneous rest frame of v = 0, dv = a
dt

the rocket, but S does not accelerate with the S x
rocket. In S , we have v = 0 and dv / dt = a. v0
This gives a (acceleration in S ) = a / 03 . Since S is the rest frame of
the rocket, 0 for the transformation equals of the rocket in S . Thus,
a = a / 03 = a / 3 , where = (1 v 2 / c 2 )1/ 2 .
 11.A.1 The Lorentz Transformation (continued)

From the expression

p of the acceleration in the earth frame (S )),
a = a / 3 = a(1 v 2 / c 2 )3/ 2 , (A.23)
we may evaluate the total acceleration time as measured on earth
T vf 1 vf 3 vf dv vf
T = 0 dt = 0 a dv = 0 a dv = a 1
2 2 3/ 2 = 2 2 1/ 2 .
0 (1v / c ) a(1v f / c )
The rocket
Th k t does
d nott remain
i in
i a single
i l inertial
i ti l frame.
f So,
S tot find
fi d the
th
total acceleration time as measured on the rocket, we must still work in
the earth frame by using the relation dt = dt/ .
T =
T
=
T1 v v
= f 1a dv = 1 f dv = c In(1+v f / c ).
0 dt 0 dt 0 a 0 1v2 / c2 2a 1v f / c
We find that, in the limit v f / c 0, both
T and T approach the expected value of
v f / a. However,
H T / T increases
i rapidly
idl as f
v f / c 1 due to the effect of time dilation
2
1
2 2 T /T
(
(see figure).
fi ) In
I the
th figure,
fi f = (1 v f / c )
is the time dilation factor at the final speed v f . vf /c
 Section 2: Relativistic Momentum and Energy
(Ref.: H. C. Ohanian, Physics,
Physics, 2nd ed., pp.1013-1014.)
pp.1013 1014.)
The law of conservation of momentum states that, for an isolated
system
y of particles,
p mi v i (before
( collision)) = mi ui (after
( collision).
)
Under the Galilean transformation, the statement is true in all
(inertial) frames. However, under the Lorentz transformation, mi vi ,
though conserved in one frame, will in general not be conserved in
another frame. Thus, postulate 1 is violated if we continue to define
the moment
momentum m as mv. The theory
theor of relativity
relati it takes a major step by b
redefining the momentum and energy as
p = mv Note: (1 v 2 / c 2 )1/ 2 is the Lorentz factor (A 24)
(A.24)
of a particle. It is to be distinguished from the
E = mc 2 (A.25)
Lorentz factor 0 for the transformation.
For simplicity, we will consider only one-dimentional motion
along the x axis. The momentum and energy of a particle are then
mvx mc 2
px = 2 2
and E =
1vx c 1vx2 c2
 11.A.2 Relativistic Momentum and Energy (continued)
vx v0
(A 20) the velocity in frame S is vx =
From (A.20), 2 . Hence
Hence, the
1vxv0 c
momentum of the particle is (assuming m has the same value in S )
mvx m(vx v0 )
px = = 2
1
1vx2 c2 1vxv0 c 1(1 c 2 )[(vx v0 ) (1vxv0 c 2 )]2
m(vx v0 )
= i vx , px , E
2
(1vxv0 c 2 )2 ( vx v0 ) c2
S x
Since (1 vx v0 c 2 ) 2 ( vx v0 )2 c2
2 2 i vx , px , E
= (1 v0 c )(1 vx2 2
c ), px becomes x
S
px = 1 mvx

v0 m v0
1v02 c2 1vx2 c 2 1v02 c 2 1vx2 c 2
v0
= 0 ( px 2 E) (A.26)
c
Similarly we derive the Lorentz transformation equation for energy:
Similarly,
E = 0 ( E v0 px ) (A.27)
 11.A.2 Relativistic Momentum and Energy (continued)
By the same method, we can extend the motion to 3 dimensions and
derive the Lorentz transformation equations for p and E. The result is
px = 0 ( px v0 E ) i p, E
2
c
py = p y S x
(A.28)
pz = pz i p, E
E = ( E v p ) S x
0 0 x
v0
i p, E
S

i p , E
S
v0

28) shows that p and E in S is a linear combination of p and E

(A.28)
(A
in S , with constant coefficients (i.e. the coefficients are independent of
p and E of the pparticle). ) The same equations
q will therefore hold true for
the total momentum and energy ( p j , E j ) of a system of particles,
pjx = 0 ( p jx v0 E j )
c2
pjy = p jy
(A.29)
pjz = p jz
E = ( E v p )
j 0 j 0 jx
 11.A.2 Relativistic Momentum and Energy (continued)

pjx = 0 ( p jx v0 E j )
c2
pjy = p jy
Rewrite (A.29) .
pjz = p jz
E = ( E v p )
j 0 j 0 jx
F
Form this
hi set off equations,
i we see thath if (and
( d only l if) the
h totall
momentum ( p j ) and total energy ( E j ) of a system of particles
are both conserved in S , the total momentum and total energy will
be both conserved in S .
definitions p = mv and E = mc 2 , will
This shows that the new definitions,
preserve the conservation law under the Lorentz transformation.
However, the conservation law must now be extended to include
both the momentum and energy. The "relativistic law of conservation
of momentum and energy", the "relativistic equation of motion", and
th i consequences have
their h been
b discussed
di d in
i Sec.
S 2 off th the main
i text.
t t
 11.A.2 Relativistic Momentum and Energy (continued)

Problem: A pparticle of rest mass m moves on the x-axis is attracted

to the origin by a force m 2 x ( = const ). It performs oscillations of
amplitude a. Express the relativistic oscillation period as a definite
integral, and obtain the 2 leading terms of this integral for small a.
Solution: The period is given by
a
= 4 0 dx
d ,
v
where the velocity v can be calculated from the energy equation
mc 2 (1 v 2 / c 2 ) 1/ 2 + 12 m 2 x 2 = mc 2 + 12 m 2 a 2
p
In this manner the expression for the p
period becomes
a 1+ 2 ( a 2 x 2 ) / 2c 2
= 4 0 dx 1/ 2
2 2 1/ 2 2 ( a 2 x2 )
( a x ) 1+
4 c2
We expand the integrand in powers of 2 (a 2 x 2 )/c 2 and get
2a 2
2
= 1+ 3 +
16c 2
 Appendix B: A Formal Derivation of the
Lorentz Transformation
In Appendix A, we first derive the effects of time dilation and
g contraction from postulate
length p 2 and, from these effects, derive
the Lorentz transformation. Here, we present a more formal (while
physically less transparent) approach, whereby the Lorentz trans-
f
formationi is
i derived
d i d directly
di l fromf postulate
l 2.
2 The
Th following
f ll i
paragraphs are taken from Alonso & Finn "Physics," p.92.
Referring to the figure to the right,
right
suppose that at t = 0 a flash of light is i
emitted at the common pposition of the
two observers. After a time t , observer
O will note that the light has reached
point
i A andd will i r = ct , where
ill write h c v0
is the speed of light. Since x 2 + y 2 + z 2
= r 2 , we may also write
x 2 + y 2 + z 2 = c 2t 2 (B.1)
 11.B A Formal Derivation (continued)

y observer O, whose pposition

Similarly,
is no longer coincident with that of O, will i
note that the light arrives at the same point
A in ti t , but
i a time b t also
l with ith velocity
l it c.
Therefore he writes r = ct , or v0
x 2 + y 2 + z 2 = c 2 t 2 (B 2)
(B.2)
Our next task is to obtain a transformation relating (B.1) and (B.2).
The symmetry of the problem suggests that y = y and z = z. Also since
OO = v0t for observer O, it must be that x = v0t for x = 0 (point O).
gg
This suggests makingg x = k ( x v0t ), where k is a constant to be deter-
mined. Since t is different, we may also assume that t = a (t bx ), where
a and b are constants to be determined (for the Galilean transformation,
k = a = 1 and b = 0). Making all these substitutions in (B.2), we have
k 2 ( x 2 2v0 xt + v02t 2 ) + y 2 + z 2 = c 2 a 2 (t 2 2bxt + b 2 x 2 ) or
(k 2 b 2 a 2c 2 ) x 2 2(k 2v0 ba 2c 2 ) xt + y 2 + z 2 = (a 2 k 2v02 / c 2 )c 2t 2 .
 11.B A Formal Derivation (continued)

This result must be identical to (B.1).

(B 1) Therefore
k 2 b2 a 2c 2 = 0
k 2v0 ba 2c 2 = 0
a 2 k 2v02 / c 2 = 1
Solving this set of equations,
equations for k , a, and b, we have
2
k =a= 1
2 2
and b = v0 / c
1v0 / c
Inserting these values of k , a, and b in x = k ( x v0t ) and
t = (a bx), we obtain the Lorentz transformation
x = xv0t
1v02 / c 2
y = y
z = z (B.3)
2
t = t v0 x / c
1v02 / c 2
 Appendix C: Summary of Lorentz
Transformation Equations
v02 12
For all equations, 0 (1 2 ) . By symmetry, equations for
c
the inverse transformation differ only by the sign of v0 (or v 0 ).
x = x x x ( x, y, z, t )
y = y i ( x, y, z, t)
K K
1. z = ( z v t ) z, z
0 0
v
v0
t = 0 (t 02 z ) y y Frames K and K
c coincide at t = t = 0.
vx = vx
0 (1 v02 vz ) v
c i
v = v y K
2. y (1 v0 v )

0
c2 z
v v v
v = z 0 i
z v K
1 02 vz
c v0
 11.C Summary of Lorentz Transformation Equations

a = 1 a
3a
03 1 v02v
v
i
c K
3.
a = 1 a v 0 ( a v ) a
3 c2

0 1 2
2 v 0 v v
c i
K
where " " and " " refer to the direction of v 0 . v0
Special case: one dimensional motion i vz , az
az = 1 az K z
3 v
0 (1 2 vz )
0 3
c
i vz , az
px = px i Px , Py , Pz , E K z
p = p v0
y y K z
4.
p z = 0 ( pz v02 E )
c i Px , Py , Pz , E
E = 0 ( E v0 p z ) K z
v0
 11.C Summary of Lorentz Transformation Equations

i Jx, J y, Jz,
J x = J x z
J = J K
y y
5. i J x , J y , J z ,
J = 0 ( J z v0 )
z K z
= 0 ( v02 J z ) v0
c

Ax = Ax
A = A i Ax , Ay , Az ,
y y
K z
6. v
Az = 0 ( Az c0 )

= ( v0 A ) i Ax , Ay , Az ,
0 c z
K z
v0
 11.C Summary of Lorentz Transformation Equations

k x = k x i kx , k y , kz ,
k = k K z
y y
7. v0
k z = 0 (k z c 2 ) i k x , k y , k z ,
= ( v k ) K z
0 0 z v0

E = E
i E , E , B , B
( )
v0
E = 0 E + c B K
8.
B =B i E , E , B , B

0 (
B = B v 0 E
c )
K
v0
where " " and " " refer to the direction of v 0 .