UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof.

Steven Errede

LECTURE NOTES 17
Proper Time and Proper Velocity

As you progress along your world line {moving with ordinary velocity u in lab frame
IRF(S)} on the ct vs. x Minkowski/space-time diagram, your watch runs slow {in your rest
frame IRF(S')} in comparison to clocks on the wall in the lab frame IRF(S).
The clocks on the wall in the lab frame IRF(S) tick off a time interval dt, whereas in your
rest frame IRF( S ) the time interval is: dt dt u 1 u2 dt

n.b. this is the exact same time dilation formula that we obtained earlier, with:

u 1 1 u c 1 1 u2 and: u u c
2

We use u u relative speed of an object as observed in an inertial reference frame
{here, u = speed of you, as observed in the lab IRF(S)}.

We will henceforth use v v relative speed between
two inertial systems e.g. IRF( S ) relative to IRF(S):

Because the time interval dt occurs in your rest frame

IRF( S ), we give it a special name: d dt = proper
time interval (in your rest frame), and: t = proper time (in your rest frame).

The name proper is due to a mis-translation of the French word propre, meaning own.
Proper time is different than ordinary time, t.
Proper time is a Lorentz-invariant quantity, whereas ordinary time t depends on the
choice of IRF - i.e. ordinary time is not a Lorentz-invariant quantity.

The Lorentz-invariant interval: dI dx dx dx dx ds2 c 2 dt 2 dx2 dy2 dz2

Proper time interval: d dI c ds c dt dx dy dz

2 2 2 2 2 2 2
c 2 dt 2 dt
= 0 in rest frame IRF(S)
2 t2
Proper time: 2 1 d dt t2 t1 t
1 t1

Because d and are Lorentz-invariant quantities: d d and: {i.e. drop primes}.

In terms of 4-D space-time, proper time is analogous to arc length S in 3-D Euclidean space.
Special designation is given to being in the rest frame of an object.
The rest frame of an object = the proper frame.

Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 1

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

Consider a situation where you are on an airplane flight from NYC to LA. The pilot comes on
the loudspeaker and announces in mid-flight that the jet stream is flowing backwards today, and
that the planes present velocity is u 0.8c u 0.8!! , due west.

What the pilot means by velocity is the spatial displacement d per unit time interval dt .
The pilot is referring to the planes velocity relative to the ground (n.b. here, we make the
simplifying assumption that the earth is non-rotating/non-moving, so that we can use IRFs)

Thus, d and dt are quantities as measured by an observer on the ground (e.g. an airplane
flight controller, using RADAR) in the ground-based (lab) IRF(S).

d d and dt are measured in the
Thus: u = ordinary velocity in the lab IRF(S)
dt ground-based (lab) IRF(S)

You, on the other hand are in your own rest frame IRF(S') in the airplane, sitting in your seat.
You know that the distance from NYC to LA is L 2763 miles (as measured on the ground,
referring to your trusty Rand-McNally Road Atlas {back pages} that you brought along with you).
So you, from your perspective, might be more interested in the quantity known as your

proper velocity , defined as:

d Spatial displacement, as measured on the ground
Proper 3-Velocity: = hybrid measurement = (in lab IRF(S)) per unit time interval, as measured
d in your (or an objects) rest frame (in IRF(S')).

1 1
dt 1 u2 dt 1 u c dt and: u , u u c
2
Since: d dt
u 1 u2


d d d d 1 1 n.b.
Then: u , but: u uu u u
d 1
dt dt dt 1 u
2
1 u c
2
0 u
u

If u 0.8c u 0.8 , then: u 1 1 u2 1 1 0.82 53 , hence: u u 53 0.8c 5 4
3 5 c 43 c !!!

Of course, for non-relativistic speeds u c , then: u to a high degree.

From a theoretical perspective, an appealing aspect of proper 3-velocity is that it Lorentz-
transforms simply from one IRF to another IRF.

= 3-D spatial component(s) of a relativistic 4-vector,

dx
The {contravariant} proper 4-velocity is:

whose zeroth/temporal/scalar component is:
d
1
dx 0
cdt dt dt c c u
0 c uc uc with: 1 u2
d d 1
dt dt 1 u2 1 u c
2

u u c u

2 Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

The proper 4-velocity vector is:

dx 0

d
u c
0
c dx1
1
dx u u x ux d
0 , u c,
or: 2 u 2
d u u y dx
u y
3 u u z u z d

3
dx

d

The numerator of the proper 4-velocity dx is the displacement 4-vector (as measured in the
ground-based (lab) IRF(S). The denominator of the proper 4-velocity d = proper time interval
(as measured in your (or an objects) rest frame IRF(S').

The Lorentz Transformation of a Proper 4-Velocity :

Suppose we want to Lorentz transform your proper 4-velocity from the lab IRF(S) to another
(different) IRF(S") along a common x -axis, in which IRF(S") is moving with relative velocity

v vx with respect to lab IRF(S):

Most generally, in tensor notation: v with v = Lorentz boost tensor. Thus:

0
0 1
0 0 0 0 1
1 1
0 0 1 with: 1 2
1 0
v 2
0 0 1 0 2 2 v
3 3 2
0 0 0 1 3

3 c

dx dx
Where: and:
d d
Compare this result to the same Lorentz transformation of ordinary 3-velocities, along a
common x -axis. We use the Einstein velocity addition rule:

dx ux v
u u x x u y y u z z ux
dt 1 u x v c 2
dy uy 1 v
u ux x uy y uz z u y with: and:

dt 1 u x v c 2 1 2 c
dz uz
uz

dt 1 u x v c 2
{See Griffiths Example 12.6 (p. 497-98) and Griffiths Problem 12.14 (p. 498)}

Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 3

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

Now we can see why Lorentz transformation of ordinary velocities is more cumbersome
than Lorentz transformation of proper 4-velocities:

d numerator,d d
For ordinary 3-velocities u , we must Lorentz transform both
dt denominator,dt dt
dx
For proper 4-velocities we only need to transform the numerator, d d .
d

Relativistic Energy and Momentum - Relativistic 4-Momentum:

In classical mechanics, the 3-D vector linear momentum p= mass velocity v , i.e. p mv .
How do we extend this to relativistic mechanics?

d
Should we use the ordinary velocity u for v ,
dt

d
or should we use the proper velocity for v ??
d

In classical mechanics, and u are identical.

In relativistic mechanics, and u are not identical.

We must use the proper velocity in relativistic mechanics, because otherwise, the law of
conservation of momentum would be inconsistent with the principle of relativity {the laws of

physics are the same in all IRFs} if we were to define relativistic 3-momentum as: p mu . No!!

Thus, we define the relativistically-correct 3-momentum as:

mu mu 1 u
p m u mu with: and: u
1 2
u 1 u c
2
1 2
u
c

Relativistic 3-momentum: p m u mu is the spatial part of a
0
relativistic 4-momentum vector: p m , i.e. p p , p .

The temporal/zeroth/scalar component of the relativistic 4-momentum vector is: p E c

0

mc mc 1
But: p 0 m 0 u mc with: and: u u c
1 u2 1 u c 1 u2
2

c c
Thus: p E c m u mc where: u c
0 0 0

1 u2 1 u c
2

Since: p m u mu , then: p p u m u u mu u u mc u u mc u E c .

4 Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

mc 2 mc 2 1
Relativistic Energy: E u mc with: and: u u c
2

1 u2 1 u c 1 u2
2

p0 E c
1
p px
Therefore, the components of the relativistic 4-momentum are: p 2
p
p y
p 3 pz

The 4-vector dot/scalar product p p is a Lorentz-invariant quantity (same in all IRFs):

p p E c px2 p y2 pz2 E c p 2 mc
2 2 2

This can be rewritten in the more familiar form as: E pc mc or: E pc mc 2 .

2 2 2 2 2 2

1
Since: E u mc then: u2 mc 2 pc mc 2 or: pc u2 1 mc 2 . But:
2 2 2 2 2 2

1 u2
1 1 1 u2 u2
hence: pc u2 1 mc 2 2
mc 2
2 2 2 2 2 2 2
1 mc mc
1 u 1 u 1 u
2 2


or: pc u u mc 2 . However: E u mc 2 Thus, we {again} also see that: p p u E c .

Note that the relativistic energy E of a massive object is non-zero even when that object is
stationary - i.e. in its own rest frame when: p = 0, u 0 and: u 1 1 u2 1 .
Then: Erest mc = rest energy = rest mass * c2. Einsteins famous formula!
2

If u 0 , then the remainder of the relativistic energy E is attributable to the motion of the particle
i.e. it is relativistic kinetic energy, Ekin .

Total Relativistic Energy: E Etot Ekin Erest u mc 2 but: Erest mc 2

Ekin Etot Erest u mc mc u 1 mc

2 2 2

1
1
Relativistic Kinetic Energy: Ekin u 1 mc 2

1 mc
2
1 mc 2

1 u 1 u c
2 2

Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 5

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

1 3 mu 4 1
In the non-relativistic regime u c , then: Ekin mu 2
2
mu 2 (classical formula).
2 8 c 2
2
p
However, for u c then: p mu and thus: Ekin (classical formula).
2m

Note that total relativistic energy, Etot and total relativistic 3-momentum, ptot ptot
are separately conserved in a closed system.

If the system is not closed, (e.g. external forces are present) then Etot and ptot will not {necessarily}
be conserved. Simply expand/enlarge the definition of the system until it is closed {e.g.

include whats producing the external forces}, then the (new) Etot and ptot will be conserved.

Note the distinction between a Lorentz-invariant quantity and a conserved quantity.

Same in all inertial reference frames Same before vs. after

a process/an event

Rest mass m is a Lorentz-invariant quantity, but it is not {necessarily} a conserved quantity.

Example: The {unstable} charged pi-meson decays (via weak charged-current interaction, with
mean/proper lifetime 26.0 ns ) to a muon and muon neutrino: v . The charged
pion mass m is not conserved in the decay { m (m mv ) }, however the relativistic

energy of the charged pion E p2 c 2 m2 c 4 is a conserved quantity: E E Ev ,

but E is not a Lorentz-invariant quantity.

Since the scalar product of any relativistic 4-vector a with itself is a Lorentz-invariant quantity
(i.e. = same numerical value in any IRF): then here, for v decay:

p p p p p 0 p p E c p2 E
2 2 2 2
But: c p2 m c

2 2
Thus: p p p2 m c p2 m c

6 Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

Griffiths Example 12.7: Relativistic Kinematics

Two relativistic lumps of clay {each of rest mass m} collide head-on with each other.
Each lump of clay is traveling at relativistic speed u 53 c as shown in the figure below:

u1 53 cx u2 53 cx
x
m m
The two relativistic lumps of clay stick together (i.e. this is an inelastic collision).
What is the total mass M of the composite lump of clay after the collision?
Conservation of momentum - before vs. after:
Since the two lumps of clay have identical rest masses and equal, but opposite velocities:
before 1 before
ptot p1 p2 but: p1 p2 u mu1 where: u ptot 0
1 2
u

Conservation of energy - before vs. after:

mc 2 mc 2
Before: Each lump of clay has total energy: E u mc 2 u mc 2
1 1 u c
2 2
u

mc 2
mc 2 mc 2 5 2
E mc
3
2
9 16 4
1 1
5 25 25

5 5
before
Thus: Etot Etot1 Etot2 2 u mc 2 2 mc 2 mc 2
4 2
5 2
However, Etot is {always} conserved in a closed system. Etot
after
Etot
before
mc
2
after before
And ptot is also {always} separately conserved in a closed system. ptot ptot 0

after 1 1
u after 0 since: ptot uafter Mu after 0 . n.b. uafter 1
1 1 uafter c
2 2
uafter

5 2 5
after
Then: Etot uafter Mc 2 Mc 2 mc EToT
before
M m 2m !!! Does this sound crazy??
2 2

Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 7

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

This is what happens in the everyday world of particle physics! Its perfectly OK !!!
e.g. The production of a neutral rho meson in electron-positron collisions: e e 0 .
The rest mass of the neutral rho meson is: M 0 770 MeV c 2 Electron rest mass: me 0.511 MeV c 2

p 0 0
p1 px p2 px
x
me me
M 0

Run the collision process backwards in time, e.g. the decay of a neutral rho meson: 0 e e

p 0 0
p1 p x p2 p x
x
me me
M 0

The production of a neutral rho meson e e 0 manifestly involves the EM interaction.

Similarly, the time-reversed situation: the decay of a neutral rho meson 0 e e manifestly
also involves the EM interaction.
The EM interaction is invariant under time-reversal, i.e. t t , thus {in the rest frame of the
neutral rho meson} the transition rate e e 0 (#/sec) vs. the decay rate 0 e e
(#/sec) are identical {for the same/identical electron / positron momenta in neutral rho meson
production vs. decay}. Experimentally: 0 e e 7.02KeV 1.70 1018 sec 1 .

For our above macroscopic inelastic collision problem, microscopically what would the new
matter of the macroscopic mass M be made up of, since M M 2m 52 m 2m 12 m ???

In a classical analysis of the inelastic collision of two relativistic macroscopic lumps of clay
{each of mass m} the composite / stuck-together single lump of clay of mass M 52 m 2m would
be very hot it would have a great deal of thermal energy in fact !!!

5 2
Mc 2 mc 2 mc 2 Ethermal Ethermal = 0.5mc2!!! E = mc2 = Einsteins energy-mass formula
2 classical mass
of composite
lump

8 Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

Conserved Quantities vs. Lorentz-Invariant Quantities in Collisions/Scattering Processes:

Before: pbefore Ebefore c , pbefore After: pafter

Eafter c , pafter . Neither is a Lorentz invariant

quantity. However, total relativistic energy E and total relativistic momentum p are separately

conserved quantities: Eafter Ebefore Mc 2 and: pafter pbefore 0 . The scalar 4-vector dot-
product is a Lorentz invariant quantity, which is also a conserved quantity i.e. its value is the
same before vs. after the collision/scattering process:

p p p p p 0 p p E c p 2 M 2 c 2 0 M 2 c 2
2 2

Griffiths Example 12.8: Relativistic Kinematics Associated with v Decay.

Pion rest mass: m 139.57 MeV c 2 Pion mean lifetime: 26.033 nsec 26.033 109 sec
Muon rest mass: m 105.66 MeV c 2 Muon neutrino rest mass: mvu 0 (assumed).
In the rest frame of the meson:

p 0
p px pv px
x
m mv 0
M
Energy Conservation: Momentum Conservation:
before
before
Before: Etot m c 2 ptot 0
after
After: after
Etot E Ev m c 2 ptot p pv 0 p pv px

But: Ev pv c pv c since: mv 0 . p p = pv pv

And: E2 p2 c 2 m2 c 4 or: p2 c 2 Eu2 m2 c 4 p c E2 m2 c 4

p p = pv pv = E2 m2 c 4 c

E2 m2 c 4
Then: E after
tot E Ev E pv c but: pv p
c
after
Etot E Ev E pv c E E2 m2 c 4 Etot
before
m c 2

E E2 m2 c 4 m c 2 Solve for E :

2
E2 m2 c 4 m c 2 E m2 c 4 2 m c 2 E E2 or: 2m c 2 E m2 c 4 m2 c 4

Thus: E
m2 c 4 m2 c 4

m
2
m2 c 2 and: pv p
E2 m2 c 4
with: p pv
2
2m c 2m c
as viewed from the rest frame of the meson.

Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 9

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

In classical collisions, total 3-momentum ptot and total mass, mtot are always conserved:
before after
ptot ptot , mtot
before
mtot
after
.

tot
In classical collisions, if total kinetic energy Ekin is not conserved inelastic collision.

An inelastic (i.e. a sticky) collision generates heat at the expense of kinetic energy.
An inelastic collision of an electron (e) with an atom {initially in its ground state} may
leave the atom in an excited state, or even ionized, kicking out a once-bound atomic electron!
Internal {quantum} degrees of freedom can be excited in inelastic e - atom collisions.

An explosive collision generates kinetic energy at the expense of chemical (i.e. EM)
energy, or nuclear (i.e. strong-force) energy, or weak-force energy. . . .
If kinetic energy is conserved (classically), elastic (i.e. billiard-ball) collision.

In relativistic collisions, total 3-momentum and total energy are always conserved
(in a closed system) but total mass and total kinetic energy are not in general conserved.
* Once again, in relativistic collisions, a process is called elastic if the total kinetic energy
is conserved total mass is also conserved in relativistic elastic collisions.
* A relativistic collision is called inelastic if the total kinetic energy is not conserved.
Total mass is not conserved in a relativistic inelastic collision.

Griffiths Example 12.9:

Compton Scattering = Relativistic Elastic Scattering of Photons with Electrons.

An incident photon of energy E0 p0 c elastically scatters (i.e. bounces off of/recoils) from
an electron, which is initially at rest in the lab frame. Determine the final energy E of the
outgoing scattered photon as a function of the scattering angle of the photon:

Consider conservation of relativistic momentum in the transverse ( ) (i.e. y -axis) direction:

pbefore
tot
0 pafter
tot

pbefore
tot
pbefore

pbefore

00 0
e

y direction
y direction

pafter
tot
pafter

pafter

0 p pafter
after

e e

10 Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

y direction
y direction

Since: p pafter
after
e

Or: pafter

pafter
e

Or: p sin pe sin

But: p E c
E
sin pe sin
c
E
Solve for sin : sin sin
p c
e

Conservation of relativistic momentum in the longitudinal (i.e. x ) direction gives:

E0
pbefore
tot (n.b. pebefore
0 , since e initially at rest, hence pbefore
e
0)
c

pafter
tot
pafter

pafter

p cos pe cos
e

Since: pbefore
TOT
pafter
TOT
then: E0 c p cos pe cos

2
E
sin cos thus: cos 1 sin 1
2
sin 2
c
But:
p c
e e

2
E0 E
p cos pe 1 sin 2
c p c
e

pe2 c 2 E0 E cos E2 sin 2 E0 2 E0 E cos E2

2 2
Or:

before
E
after
tot
E
tot

before
Conservation of Energy: Etot Etot
after
E me c E Ee E pe2 c 2 me2 c 4
0 2

2
E0 me c 2 E E0 2 E0 E cos E2 me2 c 4

1
Solve for E (after some algebra): E
1 cos me c 2 E0

E = energy of recoil photon in terms of initial photon energy E0 , scattering angle of photon
and rest energy/mass of electron, me c 2 .

Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 11

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

We can alternatively express this relation in terms of photon wavelengths:

Before: E0 hf0 hc 0 Useful constants:

After: E hf hc hc 1239.841eV -nm 1240eV -nm

hc
2
Get: 0 1 cos me c 2 0.511MeV 0.511 106 eV
me c
hc
Define the so-called Compton wavelength of the electron: e 2
2.426 1012 m
m c
e
Then: 0 e 1 cos

The Compton Differential Scattering Cross Section:

As we learned in P436 Lecture Notes 14.5 (p. 9-22) non-relativistic photon-free electron
scattering E0 me c 2 is adequately described by the classical EM physics-derived
{unpolarized} differential Thomson scattering cross section:

d Tunpol , 1 2 e2
re 1 cos 2 where: re
Classical
2.82 1015 m
e
electron
d 2 4 o me c 2
radius

However, when E0 me c 2 from the above discussion of the relativistic kinematics of photon-
free electron scattering, it is obvious that the classical theory is not valid in this regime. The
fully-relativistic quantum mechanical theory that of quantum electrodynamics (QED) is
required to get it right... Without going into the gory details, the results of the QED calculation
associated with the two Feynman graphs {the so-called s- and u-channel diagrams} shown on
p. 5 of P436 Lect. Notes 14.5 for the Compton differential scattering cross section known as
the Klein-Nishina formula for relativistic {unpolarized} photon-free electron scattering is:

d Cunpol , 1 1 x2 1 cos
2

re2 1 cos 2 1
e

d 1 x 1 cos 1 cos 1 x 1 cos

2 2
2

where: x E me c hf me c . In the non-relativistic limit x 0 , the relativistic Compton

0 2 0 2

scattering cross section agrees with the classical Thomson scattering cross section, as shown in the
figure below of the normalized differential scattering cross section r12 d Cunpol

d cos vs. .
e e

Note that as x the relativistic Compton differential scattering cross section becomes
increasingly sharply peaked in the forward direction, 0 .

12 Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

The Relativistic Doppler Shift for Photons/Light:

A rapidly moving atom isotropically emits monochromatic light (photons of frequency f ) in its
own rest frame IRF. What is the frequency f of the emitted photons as observed in the lab frame

IRF as a function of the lab angle between the atoms velocity v z and the direction of

observation { = photons momentum vector p } in the lab frame?

Rest frame of atom, IRF: Lab frame, IRF:

y p

v z

Without any loss of generality, we can choose the lab velocity v of the atom to be along the z

axis in the lab frame IRF {note that: z z v }.
hf
The energy of the photon in the atoms rest frame IRF is: E pc hf where: p p c
is the magnitude of the photons momentum in the atoms rest frame IRF. We can also assume

without loss of generality that the emitted photons momentum vector p lies in the y-z plane
of the atoms rest frame IRF. In the atoms rest frame IRF, the emitted photon makes an angle
with respect to the z axis.

Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 13

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

Hence: pz p cos E c cos and: py p sin E c sin and: px 0 .

The 4-momentum vector of the emitted photon in the atoms rest frame IRF is thus:

p E c , px , py , pz hf
c


, 0, hfc sin , hfc cos hf
c 1, 0,sin , cos
We then carry out a 1-D Lorentz transformation from the atoms rest frame IRF to the lab frame

IRF, boosted along the z z v axis (see e.g. Physics 436 Lect. Notes 16, p. 11), where:
v c and: 1 1 2 :

E c 0 0 E c 0 0 1

px 0 1 0 0 px hf 0 1 0 0 0

p

v p v

py 0 0 1 0 py c 0 0 1 0 sin

pz 0 0 pz 0 0 cos
cos 1 cos

hf 0 hf 0

c sin c sin

cos cos

Thus, in the lab IRF, the emitted photons 4-momentum vector is:
hf
p E c , p x , p y , p z 1 cos , 0,sin , cos
c
The emitted photons energy as observed in the lab IRF is: E hf hf 1 cos .
The frequency of the emitted photon observed in the lab IRF is: f f 1 cos .
Experimentally, the atoms rest frame photon emission angle is {often} not measureable;
the lab frame photon emission angle is what is measured experimentally. Hence, in order for this
formula to be useful, we must re-write this expression in terms of the lab frame photon emission
angle . The relationship between the atoms rest frame photon emission angle and the lab
frame photon emission angle can be obtained by analyzing the 3-momentum components of the

photon in the atoms rest frame p vs. the lab frame p , as shown in the figures below:

Rest frame of atom, IRF: Lab frame, IRF:

y p y p
p
p
py p sin p y p sin

z z v
pz p cos pz p cos

14 Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

In the lab frame IRF, the 4-vector momentum components of the emitted photon are:
E hf hf 1 cos E 1 cos
px 0 px
p y p sin p sin py
pz p cos p cos

We see that: pz p cos p cos . But for photons: p E c and: p E c

Thus: pz E c cos E c cos . But: E E 1 cos

Hence: pz E 1 cos c cos E c cos
Or: E 1 cos cos E cos 1 cos cos cos .

cos cos
Thus: cos .
1 cos 1 cos

However, we need an expression for cos in terms of cos . Solve for cos :
1 cos cos cos cos cos cos cos
cos cos cos cos cos 1 cos cos
cos cos
cos
cos 1 1 cos

cos 1 v
Thus: f f 1 cos f 1 where: and:
1 cos 1 2 c
1 n.b. RHS expressed
1 cos 1 1 cos entirely in lab frame
Or: f 1 f f 1 cos f
1 2
IRF variables i.e.
1 cos experimentally
measured quantities
Similarly, we can also obtain a relation for sin using:

p y p sin p sin py . But for photons: p E c and: p E c

Thus: p y E c sin E c sin But: E E 1 cos
Hence: p y E c 1 cos sin E c sin 1 cos sin sin

sin
Thus: sin
1 cos
cos
And: cos
1 cos
sin sin
Hence: tan
cos cos

Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 15

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

cos
Since: cos
1 cos

Then:
sin sin sin
sin
1 cos cos 1 cos cos 2
1
1 cos 1 cos
sin 2 1 cos sin
1 cos sin
1 2

1 cos

sin
Thus: sin
1 cos
cos
And: cos
1 cos
sin sin
Hence: tan
cos cos

Comments:
cos cos
From: cos and: f f 1 cos f 1
1 cos 1 cos

1.) The photon will be emitted in the forward direction in the lab frame IRF { 0 90 }
when the numerator: cos 0 , i.e. when: cos {n.b. cos 0 for 90 180 }.

2.) The photon will be emitted in the backward direction in the lab frame IRF { 90 180 }
when the numerator: cos 0 , i.e. when: cos .

3.) When 1 v c , all photons are emitted in the forward direction in the lab frame IRF.

4.) When: 0 , then: 0 , and: f f 1 .

n.b. so-called
5.) When: 90 , then: f f . When: 90 , then: f f 1 2 f . transverse
Doppler
shift(s)

6.) When: 180 , then: 180 , and: f f 1 .

16 Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

Special Relativity and Stellar Luminosity:

In its own rest frame IRF, a star isotropically radiates {thermal/black-body} photons. For
simplicitys sake {here}, we will assume that all radiated photons in the stars rest frame IRF
have the same energy E . The total rate of emission of such photons into 4 steradians in the
rest frame IRF of the star {assumed to be constant} is: R dN dt # sec . Note that in the
stars rest frame IRF, the temporal interval dt d the proper time interval. The total
luminosity of the star in its rest frame IRF is: L E R E dN dt Joules sec Watts .

The differential rate of emission of such photons into solid angle element d d cos d
in the stars rest frame IRF is:

dR d dN d dN dt
# sec sr
d d dt d

The differential luminosity of the star as measured in its rest frame IRF is:

dL d E R dR d dN d dN dt
E E E Joules sec sr Watts sr
d d d d dt d

Suppose that you are an astronomer on earth, observing this star through a telescope. Your
inertial reference frame is the lab frame IRF. {For simplicitys sake here, we neglect/ignore the

motion of the earth}. If the star is moving with velocity v z z , then what is the differential
luminosity dL d in the earths lab frame IRF i.e. how is dL d related to dL d ?

There are three inertial reference frame effects that must be taken into account here:

Time dilation: dt dt Rate of emission in rest frame IRF rate of emission in lab frame IRF
Angle transformation: d d
Lorentz transformation from rest frame IRF of star to lab frame IRF.
Doppler effect: E E
dt 1
The time dilation effect is: dt dt or: , where: 1 1 2 and: v c .
dt
The 1-D Lorentz transformation from the stars rest frame IRF to the lab frame IRF on earth, for

v z z is the same as that for the above relativistic Doppler shift example:

The 4-momentum vector associated with a photon emitted from the surface of the star in the rest
frame IRF of the star is: p E c , px , py , pz Ec , 0, Ec sin , Ec cos Ec 1, 0,sin , cos

The Lorentz transformation is:

E c 0 0 E c 1 cos
p
px 0 1 0 0 x E 0

p v p
v

py 0 0 1 0 py c sin

pz 0 0 pz cos

Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 17

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

Thus, the photons 4-momentum vector as seen by an astronomer in the lab frame IRF is:
E
p E c , p x , p y , p z 1 cos , 0,sin , cos
c
Here {again}, we need to express this result in terms of lab frame IRF measured variables {only}:
cos sin
cos , sin and: E 1 cos E
1 cos 1 cos

Now, the lab frame IRF vs. the stars rest frame IRF solid angle elements are, respectively:
d d cos d and: d d cos d .

The infinitesimal solid angle element d d cos d and the infinitesimal area element
da R 2 d Rd cos Rd {where the infinitesimal and arc lengths S Rd cos and
S Rd , respectively} associated with the lab frame IRF are shown in the figure below.

As per the above discussion of the relativistic Doppler shift, for v z z , without any loss of

generality we can choose the observation point P r Rr to lie in the y-z plane ( 90 ):

z da R 2 d Rd cos Rd
r cos z sin y
d d cos d

v R
cos y sin z

x
Star y
90

x

For the choice of observation point P r Rr lying in the y-z plane ( 90 ), note that x
is to r cos z sin y . Thus, since transverse components of a 4-vector are unaffected by a
Lorentz transformation e.g. from the rest frame IRF to the lab frame IRF, then , hence
and d d . Thus, in order to determine the relationship between solid angle element
d d cos d and d d cos d , we only need to determine how d cos is related to
d cos . From above, we already have the relation:
cos
cos
1 cos
18 Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois
2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

Thus, using the chain rule of differentiation:

d cos d cos 1 cos

d cos d cos 1 cos 1 cos 1 cos 2
1 cos cos 2 1 2 1 1

1 cos 1 cos 1 cos
2 2 2 2

Hence:
d d cos d d cos 1 1
2
d d cos d d cos 1 cos 2
1

We also already have the relationship between E and E from the Lorentz transformation result:
E 1
E 1 cos E or:
E 1 cos

The lab frame vs. rest frame differential luminosity of the star are related to each other by:

dL dR d dN dt d E d dN
E E E
d d d dt dt d E d dt
dt d E dR dt d E dL Watts
E
dt d E d dt d E d sr
Thus:
dL dt d E dL 1 1 1 1 dL Watts

2
d dt d E d 1 cos 1 cos d sr
2

dL 1 1 dL Watts n.b. peaks sharply in = 0 (forward)

Or: 4 direction, and 0 in = (backward)
d 1 cos d sr
3
direction as 1 ( ).

The differential luminosity of the star in its own rest frame IRF is thus:

dL 3 dL Watts n.b. RHS expressed entirely in lab

4 1 cos frame IRF variables i.e.
d d sr experimentally measured quantities

dL dL
The total luminosity of the star in its own rest frame IRF is: L d 4 Watts
d d
since the emission of photons in the stars own rest frame is isotropic.
Hence the total luminosity of the star in its own rest frame IRF is:

dL 3 dL n.b. RHS expressed entirely in lab

L 4 4 4 1 cos Watts frame IRF variables i.e.
d d experimentally measured quantities

Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 19

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

Relativistic Dynamics
Newtons 1st Law of Motion: An object at rest remains at rest, an object moving with speed v
continues to move at speed v, unless acted upon by a net/non-zero/unbalanced force
the Law of Inertia is built/incorporated into in the Principle of Relativity.
Newtons 2nd law of motion {classical mechanics} retains its validity in relativistic mechanics,
provided that relativistic momentum is used:

dp r , t
F r,t ma r , t
dt
Griffiths Example 12.10: 1-D Relativistic Motion Under a Constant Force.

A particle of (rest) mass m is subject to a constant force: F r , t F Fx constant vector .

If the particle starts from rest at the origin at time t = 0, find its position x t as a function of t.
dp t dp t
Since the relativistic motion is 1-D, then: F = constant, or: F = constant.
dt dt
p t Ft constant of integration . The particle starts from rest at t = 0. p t 0 0
constant of integration = 0. p t Ft {here}

mu t 1
Relativistically: p t u t mu t Ft where: u t
1 u t c 1 u t c
2 2

Solve for u t : m 2u 2 F 2t 2 1 u 2 c 2 F 2t 2 F 2t 2 c 2 u 2 m F t
2 2 2

c 2 u 2 F 2t 2

Ft m
2
F 2t 2 Ft m Relativistic particle
Or: u 2
2
u t = velocity for constant
m F t c 1 Ft mc
2
1 Ft m c
2 2 2 2
applied force F
n.b. when: Ft mc 1 i.e. Ft m c then: u t Ft m Classical dynamics answer.

Note also that as t : u t c !!! (Relativistic denominator ensures this!)

Ft m dx t t t t
Since: u t Then: x t 0 u t dt F m 0 dt
1 Ft m c dt 1 F mc t
2 2 2

2
F mc
t
mc 2
The motion is hyperbolic: x t 1 F mc t 2 1 F mc t 1
2 2 2

m F 0 F
n.b. Had we done this in classical dynamics, the result would have been parabolic motion:
F 2
x t t
2m

20 Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

Thus, in relativistic dynamics e.g. a charged particle placed in a uniform electric field E ,

the resulting motion under a constant force F qE is hyperbolic motion (not parabolic motion,
as in classical dynamics) see/compare two cases, as shown in figure below:

Relativistic Work:

Relativistic work is defined the same as classical work: W F d

The Work-Energy Theorem (the net work done on a particle = increase in particles kinetic energy)
also holds relativistically:

dp dp d dp d
W F d d dt u dt Ekin since: u
dt dt dt dt dt

dp d mu mu mu
But: u u since: p u mu
dt dt 1 u c 2 1 u2 1 u c
2

Thus:

2
mu 2 u m u u du
dp m du c du mu du c
u 2 u 3 u 3

dt 1 u c dt 1 u c 2 2 dt 1 u c dt 1 u c 2 2 dt
2

2

2

u c du 1 u c u c du
2
1 mu du
3
mu 3 mu 3

1 u c dt dt 2 dt
2
1 u c 2 2 1 u c

2
2
1 u c
2



mu du d mc 2
3

1 u c 2 2 dt dt 1 u c 2

Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 21

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

1 dp d dE
But: u u u mc tot
2

1 u c
2
dt dt dt

dp dp d dp
W F d d dt udt Ekin
Thus: dt dt dt dt
dE
tot dt Etotfinal Etot
initial
Etot
dt

But: Etot Ekin Erest Ekin mc 2 n.b. Etot u mc 2 u 1 mc 2 mc 2 , Ekin u 1 mc 2

Ekin

Etotfinal Etot

initial
Ekin
final

mc 2 Ekin
initial
mc 2 Ekin

final

Ekin
initial

(final-initial) difference in
total energy = (final-initial)
Etot Ekin
difference in kinetic energy
i.e. W Etot E final
tot Einitial
tot Ekin Efinal
kin E
initial
kin
= work done on particle.

As we have already encountered elsewhere in E&M, Newtons 3rd Law of Motion

(For every action (force) there is an equal and opposite reaction) does NOT (in general) extend
to the relativistic domain, because e.g. if two objects are separated in 3-D space, the 3rd Law is
incompatible with the relativity of simultaneity.
As
Suppose the 3-D force of A acting on B at some instant t is: F r ,t F r ,t
AB B B observed

r , t F r , t
e.g. in lab
and the 3-D force of B acting on A at the same instant t is: FBA A A IRF(S)

Then Newtons 3rd Law does apply in this reference frame.

However, a moving observer {moving relative to the above IRF(S)} will report that these
equal-but-opposite 3-D forces occurred at different times as seen from his/her IRF(S'), thus in

his/her IRF(S'), Newtons 3rd Law is violated (the two 3-D forces F rB , t and F
AB
rA , t at
BA
the same time t in IRF(S') are quite unlikely to be equal and opposite, e.g. if they are changing
in time in IRF(S)).
Only in the case of contact interactions (i.e. 2 point particles at same point in space-time =

(xA, tA)) where the two 3-D forces F r , t and F r , t are applied at the same point (xA) at
AB B BA A

the same time, and in the {trivial} case where forces are constant, does Newtons 3rd Law hold!

dp r , t
F r,t
dt

The observant student may have noticed that because F r , t is the derivative of the

(relativistic) momentum p r , t with respect to the ordinary (and not the proper) time t, it
suffers from the same ugly behavior that ordinary velocity does, in Lorentz-transforming

dp r , t
it from one IRF to another: both numerator and denominator of must be transformed.
dt

22 Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

Thus, if we carry out a Lorentz transformation from IRF(S) to IRF(S), along the x -axis

where v vx is velocity vector of IRF(S) as observed in IRF(S), and u is the velocity vector of
a particle of mass m as observed in IRF(S):

1 v
Then: where: with: v vx
1 2 c

The and factors are needed for the

Lorentz transformation of kinematic quantities
from IRF(S) IRF(S).

First, let us work out the y and z (i.e. the transverse) components of the 3-D force F r , t
as seen in IRF(S) {they are simpler / easier to obtain. . . }:
dp dp dx
Noting that: F , F and that: dt dt dx and: u x
dt dt c dt
dp y
dpy dp y dt Fy
In IRF(S): Fy
dt dt
1 u x c
dx
dx 1

c c dt
dpz
dp dpz dt Fz
Similarly: Fz z
dt dt dx dx 1 u x c
1
c c dt

Now calculate the x -component of the force F r , t in IRF(S):
dpx dp 0 dE
dp dp 0 Fx c tot
dp dt where: p 0 Etot
In IRF(S): Fx x
x
dt dt
dt 1
dx 1 ux c c
dt dx c dt
c
dp
dEtot dEtot dp
But we have calculated above / earlier: u F u u F since: F
dt dt dt dt

Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 23

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

Fx

Fx u F c Relativistic ordinary x, y, z force components
observed in IRF(S) acting on particle of mass m, for a
1 u c
x Lorentz transformation from lab IRF(S) to IRF(S).

Fy IRF(S) moving with velocity v vx relative to
Fy
1 u x c IRF(S) (as seen in IRF(S)), 1 1 2 , v c .
Fz
Fz Particle of mass m is moving with ordinary velocity
1 u x c u as seen in IRF(S).

We see that only when the particle of mass m is instantaneously at rest in lab IRF(S)

(i.e. u t 0 ) will we then have a simple Lorentz transformation of the ordinary force F F :

Fx Fx F F n.b. || force components are same/identical !!!

u 0: Fy Fy Where the subscripts ( ) refer to the parallel
(perpendicular) components of the force with respect to
Fz Fz F F the motion of IRF(S) relative to IRF(S), respectively.

Note that for u 0 , the component of F to the Lorentz boost direction is unchanged.

For u 0 , the component of F to the Lorentz boost direction is reduced by the factor 1 .

Proper Force The Minkowski Force:

In analogy to the definition of the proper time interval d and the proper velocity

d d versus the ordinary time interval dt and the ordinary velocity u d dt ,

we define a proper force K (also known as the Minkowski force), which is the derivative of the

relativistic momentum p with respect to proper time d :

dp dt dp dt dt 1 1
K but: u
d d dt d dt 1 u
2
1 u c
2

dp dt dp dp
K u F where: F
d d dt dt
1 1 1 1 u
Thus: K u F F F where: u and: u
1 u2 1 u c 1 u 1 u c c
2 2 2

24 Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

We can 4-vectorize the Minkowski Force, because its plainly / clearly a 4-vector:

dp 0 1 dEtot Etot Proper rate at which energy of particle

ct: K
0
since: p 0 K0 = increases (or decreases)
d c d c = (Proper power delivered to the particle)/c !
dp1 1 1
x: K
1
u F1 F1 F1
d 1 u
2
1 u c
2

dp 2 1 1 1 1
y: K2 uF2 F2 F2 with: u
d 1 u
2
1 u c
2
1 2
u 1 u c
2

dp 3 1 1
z: K3 uF3 F3 F3
d 1 u
2
1 u c
2

dp
Thus: K Minkowski 4-vector force = proper 4-vector force.
d
Relativistic dynamics can be formulated in terms of either ordinary quantities or proper
(particle rest frame) quantities. The latter is much neater / elegant, but it is (by its nature)
restricted to the particles rest frame IRF(S) {n.b. We can always Lorentz boost this proper
result to any other inertial reference frame. . . }
There is a very simple reason for this! Since we humans live in the lab frame IRF(S)
we want to know everything about particles trajectory, the forces acting on it, etc. in the lab
because this is the only IRF that we can (easily) carry out physical measurements in often, it is
not possible to make physical measurements e.g. in a particles rest frame / proper frame,
especially if the particles are in relativistic motion (e.g. at Fermilab/LHC/ hadron colliders).
In the long run, we will (usually) be interested in the particles trajectory as a function of
ordinary time, so in fact the ordinary 4-force F dp dt is often more useful, even if it is
more painful / cumbersome to calculate / compute
We want to obtain the relativistic generalization of the classical Lorentz force law

FC qE qu B { u = particles ordinary velocity in IRF(S)}. Does the classical formula FC

correspond to the ordinary relativistic force F , or to the proper / Minkowski force K ?

Thus, for the relativistic Lorentz force, should we write: F qE qu B q E u B ???

Or rather, should the relativistic Lorentz force relation be: K qE qu B q E u B ???
Since proper time and ordinary time are identical in classical physics / Euclidean /
Galilean 3-space, classical physics cant tell us the answer.

It turns out that the Lorentz force law is an ordinary relativistic force law: F q E u B
Well see why shortly Well also construct the proper / Minkowski EM force law, as well . . .

Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 25

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

But first, some examples:

Griffiths Example 12.11: Relativistic Charged Particle Moving in a Uniform Magnetic Field
Weve discussed this before, from a classical dynamics point of view:
The typical trajectory of a charged particle (charge Q, mass
m) moving in a uniform magnetic field is cyclotron motion.

If the velocity of particle ( u ) lies in the x-y plane and B Bo z ,

then F Qu B QuBo r QuBo r as shown on the right:

The magnetic force points radially inward it provides the

centripetal acceleration needed to sustain the circular motion.
However, in special relativity the centripetal force is not mu 2 R
dp d R d 1 Rd u
(as it is in classical mechanics). Rather, it is: F p p p p .
dt dt R dt R dt R
Top View: Vector Diagram:

u u2
F p r n.b. Classically: p mu thus, classically: F m r
R R

u u
Thus, relativistically: QuBo r p r or: QuBo p or: p QBo R
R R
The relativistic cyclotron formula is identical to the classical / non-relativistic formula!

However here, p is understood to be the relativistic 3-momentum: p m u mu .

26 Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

Griffiths Example 12.12: Hidden Momentum

Consider a magnetic dipole moment m modeled as a rectangular loop of wire (dimensions
w ) carrying a steady current I. Imagine the current as a uniform stream of non-interacting
positive charges flowing freely through the wire at constant speed u. (i.e. a fictitious kind of

superconductor.) A uniform electric field E is applied as shown in the figure below:

The application of the external uniform electric field E Eo y changes the physics the
electric charges are accelerated in the left segment of the loop and decelerated in the right
segment of the loop. [n.b. admittedly this is not a very realistic model, but other more realistic
models do lead to the same result see e.g. V. Hnizdo, Am. J. Phys. 65, 92 (1997)].
Find the total momentum of all of the charges in the loop.
The momenta associated with the electric charges in the left and right segments of the loop

cancel each other (i.e. p (in left segment) = p (in right segment), so we only need to consider the
momenta associated with the electric charges flowing in the top and bottom segments of the loop.
Suppose there are N charges flowing in the top segment of the loop, moving in x direction


with speed u u E 0 {because they underwent acceleration traveling on the LHS segment}
and N charges flowing in the bottom segment of the loop, moving in the x direction with


speed u u E 0 {because they underwent deceleration traveling on the RHS segment}.
Note that the current I u must be the same in all four segments of the loop, otherwise
charges would be piling up somewhere.

In particular: I I (top segment of loop) = I (bottom segment of loop), i.e. I I I .

Qtot NQ Q Q
Since: then: I u N u = I u N u

Q Q I
N u N u I N u N u
Q

Classically, the linear momentum of each electric charge is pclassical mQ u where mQ = mass
of the charged particle.

Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 27

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

The total classical linear momentum of the charged particles flowing to the right in the top
N
segment
segment of the loop is: ptopclassical mQ u N mQ u x
i 1

The total classical linear momentum of the charged particles flowing to the left in the bottom
N

segment of the loop is: pbottom
classical
segment
mQu N mQu x
i 1

The net (or total) classical linear momentum of the charged particles flowing in the loop is:
tot left segment top segment right segment bottom segment top segment bottom segment
pclassical pclassical pclassical pclassical pclassical pclassical pclassical
N mQ u x N mQ u x N u N u mQ x I Q I Q mQ x 0!!!
tot
Thus, pclassical 0 as we expected, since we know the loop is not moving.
However, now let us consider the relativistic momentum:
1 1
prel u mQ u (even if u u c ) where: u
1 u 1 u c
2 2

The total relativistic momentum of the charged particles flowing to the right in the top
1 1
segment of the loop is: ptoprel segment u N mQ u x where: u .
1 2 1 u c
2

The total relativistic momentum of the charged particles flowing to the left in the bottom
1 1
segment of the loop is: pbottom segment
u N mQ u x where: u .
1 2 1 u c
rel 2

The net / total relativistic momentum is:

tot top segment bottom segment
prel prel prel
u N mQ u u N mQ u x u N u u N u mQ x

tot I
But I I I gave us: N u N u
I

prel u u mQ x 0 because u u !!!
Q Q
Charged particles flowing in the top segment of the loop are moving faster than those flowing in
the bottom segment of the loop.

The gain in energy u mc 2 of the charged particles going up the left segment of the loop
= the work done on the charges by the electric force ( W QEo w ) (w = height of the rectangle).

Thus, for a charged particle going up the left segment of the loop, the energy gain is:
QEo w

E u mQ c 2 u mQ c 2 u u mQ c 2 W QEo w u
u
mQ c 2
Where Eo = the magnitude of the {uniform/constant} electric field.

28 Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

tot I Q Eo w I E I w
prel
u u mQ x
m Qc
2
m Q x o 2 x
Q
Q c

TOT Eo IA TOT mEo

But: w A = area of the loop. prel 2 x but: m m IA prel 2 x
c c

But: m m z (see picture above) and: E Eo y i.e. m {here}.
tot 1
x

Thus, vectorially we {actually} have: rel p
c2

m E
where: m

E mEo
z y

Thus a magnetic dipole moment m in the presence of an electric field E carries relativistic

linear momentum p , even though it is not moving !!!

n.b. it also (therefore) carries relativistic angular momentum Lrel r prel .
How big is this effect? Explicit numerical example - use everyday values:
Eo = 1000 V/m
I = 1 Amp
A = (10 cm)2 = 0.01 m2
m = IA = 0.01 A-m2

mEo 102 103

tot
prel 1016 kg -m /s Tiny !!! The 1/c2 factor kills this effect !!!
3 10
2 2
c 8

This so-called macroscopic hidden linear momentum is strictly relativistic, purely mechanical

But note that it precisely cancels the electromagnetic linear momentum stored in the E and B
fields!!! (Microscopically, the momentum imbalance arises from the imbalance of virtual photon
emission on top segment of the loop vs. the bottom segment of the loop.)
Likewise, the corresponding hidden angular momentum precisely cancels the

electromagnetic angular momentum stored in the E and B fields.
Now go back and take another look at Griffiths Example 8.3, pages 356-57. (The coax cable
carrying uniform charge / unit length and steady current I flowing down / back cable.)
Lets pursue this problem a little further
Suppose there is a change in the current, e.g. suppose the current drops / decreases to zero.
dI
For simplicitys sake, assume K (i.e. the current decreases linearly with time)
dt

Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 29

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

Classically: I t I t I t (as before)

Q Q Q
I t N t u N t u N t u

Q
I t N t u We assume that u, u+ and u are
unaffected by the change in the
current with time.

dI dI t dI t Q dN t Q dN t
Then: u u K
dt dt dt dt dt
dN t dN t K
u u = constant (no time dependence on RHS of equation)
dt dt Q

Then:

dpclassical t dN t K mQ K mQ
mQ u x x x
dt dt Q Q
Constant
dpclassical t dN t K mQ K mQ
mQ u x x x
dt dt Q Q

The net / total classical time-rate of change of linear momentum is:

tot
tot dpclassical t dpclassical t dpclassical t K mQ K mQ
Fclassical t x x 0
dt dt dt Q Q
tot
tot dpclassical t 0
Thus: Fclassical t as we expected, since the loop is not moving.
dt
Now, lets investigate this situation relativistically:
For individual charges
Since: prel u mQ u prel u mQ u and: prel u mQ u
with mass mQ

1 1
Then: ptoprel segment t u N t mQ u x where: u .
1 2 1 u c
2

1 1
And: pbottom segment
t u N t mQu x where: u .
1 2 1 u c
rel 2

dptoprel segment t dN t K mQ
And: u mQ u x constant u x
dt dt Q

dpbottom segment
t dN t K mQ
And: rel
u mQ u x constant u x
dt dt Q

30 Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

The net / total time rate of change of relativistic linear momentum is:
tot
dprel t dptoprel segment t dpbottom segment
t
u
K mQ K mQ
x u x
rel

dt dt dt Q Q
K mQ

u u
x 0 u u
Q

QEo w
From above (p. 20): u
u
mQ c 2
where: Eo = electric field amplitude

tot
dprel t Q Eo w K m Q E K w E KA A w = cross-
x o 2 x o 2 x
dt m Qc2 Q c c sectional area of the loop

dI dm dI
Now: K and: m IA A (Since A = constant).
dt dt dt

dm dI
KA A = time rate of change of the magnetic dipole moment of the loop.
dt dt
tot
dprel t 1 dm t E x

dt c 2 dt
o but: m m z E E y
o

tot

Frel t
dprel t 1 dm t
2

E 0 (assuming external E -field is constant in time)

dt c dt

Thus, a net hidden force acting on the magnetic dipole, when dI dt 0 .

One might think that this net hidden force would be exactly cancelled / compensated for by
a countering force due to the electromagnetic fields, as we saw in the static case ( dI dt 0 ),
with a steady current I. But it isnt!! Why??
As we saw for M(1) magnetic dipole radiation, a time-varying current in a loop produces EM
radiation. Essentially there is a radiation reaction / back-force that acts on the antenna a
radiation pressure much like the recoil / impulse from firing a bullet out of a gun the short
explosive pulse launches the bullet, but the gun is also kicked backwards, too.
The same thing happens here when dI dt 0 - the far zone EM radiation fields are produced
(i.e. real photons) while dI dt 0 and carry away linear momentum, and since dI dt 0 ,
a net force imbalance on the radiating object! (n.b. e.g. by linear momentum conservation, a
laser pen has a recoil force acting on it from emitting the laser radiation a radiation back reaction)

Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 31

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

Likewise, the net hidden time rate of change of relativistic angular momentum is:
tot
rel t
dLtot dprel t 1 dm t
r 2 Eo r x
dt dt c dt
Which will also not be exactly cancelled either, for the same reason the EM radiation field can
/ will carry away angular momentum
dLtot
In reality, in order to calculate rel
t , we need to go back
dt
and integrate infinitesimal contributions along the (short)
segments of upper and lower / top and bottom segments of the

loop because r p rp sin , between r and p .

dp dp
Same for r r sin .
dt dt

Will get result that has geometrical factor of order 1 .

Conclusions wont be changed by this, just actual #.

dL
As we know, the time rate of change of angular momentum: = torque.
dt
Thus, the time rate of change of the net / total hidden relativistic angular momentum
tot
dLrel t tot
= net hidden relativistic torque, rel t .
dt
tot
tot rel t
dLtot dprel t tot 1 dm t
Thus: rel t r r Frel t 2 r E 0
dt dt c dt

Which is not completely / exactly cancelled when dI t dt 0 !!!

Linear momentum, angular momentum, energy, etc. are all conserved for this whole system,
its just that the EM radiation emitted from the antenna is free-streaming, carrying away all these
quantities with it!
In the static situation I = constant, the hidden relativistic linear momentum and angular
momentum is exactly cancelled by the linear momentum and angular momentum (respectively)

carried by the (macroscopic) static electromagnetic fields E and B . Microscopically, the field
linear and angular momentum is carried by the static, virtual photons associated with the

macroscopic E and B fields, cancelling the (macroscopic) hidden linear and angular

relativistic momentum of the magnetic dipole in a uniform E -field.

In the non-static situation dI t dt 0 , virtual photons undergo space-time rotation,

becoming real photons, which carry away {real} linear and angular momentum. Hidden
relativistic linear and angular momentum is no longer exactly cancelled by the (now) real field
linear and angular momentum associated with the EM radiation fields. It is only partially
cancelled by remaining / extant virtual / near-zone / inductive zone EM fields.

32 Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

Griffiths Problem 12.36: Relativistic Ordinary Force

In classical mechanics Newtons 2nd Law is: F ma .

dp
The relativistic ordinary force relation; Frel cannot be so simply expressed.
dt

dp d d 1 1
Frel u mu mu where: u
dt dt dt 1 u c 2 1 u c
2


du 1 du
2u du
dt 1 c2 dt
Frel m u 3 where:
a = ordinary acceleration.
1 u c 2
dt
2
1 u c
2 2

m u u a m u u a
Frel a 2 a 2 Q.E.D.
1 u c
2

c 1 u c
2
1 u c
2
c u 2
Griffiths Problem 12.38: Proper Acceleration
We define the proper four-vector acceleration in the obvious way, as:

d d 2 x 0 dx

d

d 2
, where:

d
0 , u c, u u = proper four-velocity

a.) Find 0 and in terms of u and a ( = ordinary velocity, ordinary acceleration):

d 0 d 0 dt 1 d c since: d 1 dt dt 1
0

d dt d 1 u c dt 1 u c u d
u
1 u c
2 2 2

1 1
2 c 2 2 u a 1
c u a du

0
where: a

3
1 u c
2 c 1 u c 2 2 dt
1 u c
2 2

Similarly:

d d dt 1 d u since: u and: 1

d dt d 1 u c dt 1 u c
u u
1 u c
2 2 2

1 1
2 u a
1 a 2 c 2
u

3
1 u c 1 u c
2 2
1 u c
2 2

Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 33

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

1 u u a Frel
a 2 see Problem 12.36 above.
1 u c 2
c u 2 u m

b.) Express in terms of u and a :
2
u a
2

2
0 2 1

1
1
a 1 u c c 2 u u a
2

4 2 4
c 1 u c 2 1 u c
1 1 2
2

2 1 2
2
2 u a a 1 u c 1 u c u a 4 u 2 u a
2 2
2

1 u c 2 4 2
c c c

2 u a u u
2
2 2
1

2
a 1 u c
2
1 2 2


1 u c
2 4
c2 c c

2
1 u a
Or:

a 2
2
n.b. Lorentz-invariant quantity same in all IRFs.
1 u c 2 4
c u 2

c.) Show 0 .

Recall that the dot-product of any two relativistic four-vectors is a Lorentz-invariant quantity.

Thus, if we deliberately/consciously choose to evaluate 0 in the rest
2

frame of an object, where 0 , 0 , 1 and 0 c , then:

0
0 c 2 = constant.
2 2

d
Note that is also the dot-product of two relativistic four vectors { and }.
d

d d d
Note also that:
d


d
2
d

d d
But: c = constant (from above). Thus:
2

d
c 2 0
d
0 .

dp
d.) Write the Minkowski / proper force version of Newtons 2 law, K in terms of the nd
d
proper acceleration .

dp d d

K
d

d
m m

d
m

34 Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois

2005-2015. All Rights Reserved.
 UIUC Physics 436 EM Fields & Sources II Fall Semester, 2015 Lect. Notes 17 Prof. Steven Errede

e.) Evaluate the Lorentz-invariant 4-vector dot product K :

K m but: 0 from part c.) above.

K 0

Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois 35

2005-2015. All Rights Reserved.