Relativistic Effects

Relativistic Effects
1 Introduction
The radio-emitting plasma in AGN contains electrons with

relativistic energies. The Lorentz factors of the emitting elec-
trons are of order 10 2 – 6 . We now know that the bulk motion
of the plasma is also moving relativistically – at least in some
regions although probably “only” with Lorentz factors about
10 or so. However, this has an important effect on the prop-
erties of the emitted radiation – principally through the ef-
fects of relativistic beaming and doppler shifts in frequency.
This in turn affects the inferred parameters of the plasma.
Relativistic Effects ©Geoffrey V. Bicknell

2 Summary of special relativity
For a more complete summary of 4-vectors and Special Rel-
ativity, see Rybicki and Lightman, Radiative Processes in As-
trophysics, or Rindler, Special Relativity
High Energy Astrophysics: Relativistic Effects 2/92

2.1 The Lorentz transformation
y S y′ S′
V
x x′
The primed frame is moving wrt to the unprimed frame with

a velocity v in the x –direction. The coordinates in the primed
frame are related to those in the unprimed frame by:

 Vx
x′ = Γ ( x – vt ) t′ = Γ t – ------
 c2 
y′ = y z′ = z (1)
v 1
β = -- Γ = -------------------
c
1 – β2
We put the space-time coordinates on an equal footing by
putting x 0 = ct . The the x – t part of the Lorentz transforma-
tion can be written:
x′ = Γ ( x – βx 0 ) x 0 ′ = Γ ( x 0 – βx ) (2)
The reverse transformation is:

x = Γ ( x′ + Vt′ ) y = y′ z = z′
 Vx′ (3)
t = Γ t′ + --------
 c2 
i.e.,
x = Γ ( x′ + βx 0 ′ ) x 0 = Γ ( x 0 ′ + βx′ ) (4)

2.2 Lorentz–Fitzgerald contraction
y y′
S S′
L0
x x′
x2 ′ – x1 ′ = L0
x2 ′ – x1 ′ = Γ [ ( x2 – x1 ) – V ( t 2 – t 1 ) ]
L 0 = ΓL ⇒ L = Γ – 1 L 0

2.3 Time dilation
y y′
V
S S′
x x′

Consider a clock at a stationary position in the moving frame
which registers a time interval T 0 . The corresponding time
interval in the “lab” frame is given by:
T = t 2 – t 1 = Γ [ ( t 2 ′ – t 1 ′ ) – V ( x2 ′ – x1 ′ ) ⁄ c2 ]
(5)
= Γ ( t 2 ′ – t 1 ′ ) = ΓT 0
i.e. the clock appears to have slowed don by a factor of Γ

2.4 Doppler effect
P1 V l = V∆t P2
θ
To
ob
d = V∆t cos θ ser
D ve
r
The Doppler effect is very important when describing the ef-

fects of relativistic motion in astrophysics. The effect is the

combination of both relativistic time dilation and time retar-
dation. Consider a source of radiation which emits one period
of radiation over the time ∆t it takes to move from P 1 to P 2 .
If ω em is the emitted circular frequency of the radiation in the
rest frame, then
2π
∆t′ = ---------- (6)
ω em
and the time between the two events in the observer’s frame
is:

2π
∆t = Γ∆t′ = Γ ---------- (7)
ω em
However, this is not the observed time between the events be-
cause there is a time difference involved in radiation emitted
from P 1 and P 2 . Let
D = distance to observer from P 2 (8)
and
t 1 = time of emission of radiation from P 1
(9)
t 2 = time of emission of radiation from P 2

Then, the times of reception, t 1rec and t 2rec are:
rec D + V∆t cos θ

t 1 = t 1 + ---------------------------------
c
(10)
rec D
t 2 = t 2 + ----
c
Hence the period of the pulse received in the observer’s frame
is

rec rec  D  D + V∆t cos θ
t 2 – t 1 = t 2 + ---- – t 1 + ---------------------------------
 c  c 
V
= ( t 2 – t 1 ) – ---- ∆t cos θ (11)
c
V
= ∆t 1 – ---- cos θ

 c 
Therefore,

2π 2π  V
----------- = Γ ---------- 1 – ---- cos θ
ω obs ω em  c 
ω em ω em (12)
⇒ ω obs = ----------------------------------- = ---------------------------------
 V  Γ ( 1 – β cos θ )
Γ 1 – ---- cos θ
 c 
The factor Γ is a pure relativistic effect, the factor

( 1 – β cos θ ) is the result of time retardation. In terms of lin-
ear frequency:
ν em
ν obs = --------------------------------- (13)
Γ ( 1 – β cos θ )

The factor
1
δ = --------------------------------- (14)
Γ ( 1 – β cos θ )
is known as the Doppler factor and figures prominently in the
theory of relativistically beamed emission.
2.5 Apparent transverse velocity

Derivation
A relativistic effect which is extremely important in high en-
ergy astrophysics and which is analysed in a very similar way
to the Doppler effect, relates to the apparent transverse veloc-
ity of a relativistically moving object.

P1 V l = V∆t P2
θ
To
ob
d = V∆t cos θ ser
D ve
r
l ⊥ = V∆t sin θ

Consider an object which moves fromP 1 toP 2 in a time ∆t in
the observer’s frame. In this case, ∆t need not be the time be-
tween the beginning and end of a periodic. Indeed, in prac-
tice, ∆t is usually of order a year. As before, the time
difference between the time of receptions of photons emitted
at P 1 and P 2 are given by:
V
∆t rec = ∆t 1 – ---- cos θ
 (15)
 c 
The apparent distance moved by the object is
l ⊥ = V∆t sin θ (16)

Hence, the apparent velocity of the object is:
V∆t sin θ V sin θ
V app = ------------------------------------- = -------------------------------
 V   V
∆t 1 – ---- cos θ 1 – ---- cos θ
 c   c 
V (17)
V app ---- sin θ
c
------------ = -------------------------------
c 1 – V ---- cos θ
 c 
In terms of

V app V
β app = ------------ β = ----
c c
(18)
β sin θ
β app = ------------------------
1 – β cos θ
The non-relativistic limit is just V app = V sin θ, as we would
expect. However, note that this result is not a consequence of
the Lorentz transformation, but a consequence of light travel
time effects as a result of the finite speed of light.

Consequences
For angles close to the line of sight, the effect of this equation
can be dramatic. First, determine the angle for which the ap-
parent velocity is a maximum:
dβ app ( 1 – β cos θ )β cos θ – β sin θβ sin θ
-------------- = ---------------------------------------------------------------------------------
dθ ( 1 – β cos θ ) 2
(19)
β cos θ – β 2
= --------------------------------
( 1 – β cos θ ) 2
This derivative is zero when
cos θ = β (20)

At the maximum:
β sin θ β 1 – β2 β
β app = ------------------------ = ----------------------- = ------------------- = Γβ (21)
1 – β cos θ 1 – β2 1 – β2
If Γ » 1 then β ≈ 1 and the apparent velocity of an object can
be larger than the speed of light. We actually see such effects
in AGN. Features in jets apparently move at faster than light
speed (after conversion of the angular motion to a linear
speed using the redshift of the source.) This was originally
used to argue against the cosmological interpretation of qua-

sar redshifts. However, as you can see such large apparent ve-
locities are an easily derived feature of large apparent
velocities.

0.98
0.95
β app
0.9
0.8
Plots of β app for various indicated values

of β as a function of θ .
The following images are from observations of 3C 273 over
a period of 5 years from 1977 to 1982. They show proper mo-
tions in the knots C 3 and C 4 of 0.79 ± 0.03 mas/yr and
0.99 ± 0.24 mas/yr respectively. These translate to proper mo-
tions of 5.5 ± 0.2h – 1 c and 6.9 ± 1.7h – 1 c respectively.

From Unwin
et al., ApJ,
289, 109

2.6 Apparent length of a moving rod
The Lorentz-Fitzgerald contraction gives us the relationship
between the proper lengths of moving rods. An additional
factor enters when we take into account time retardation.
P1 V ∆x P2
θ
L To
ob
d = ∆x cos θ ser
D ve
r
l ⊥ = ∆x sin θ

Consider a rod of length
L = Γ –1 L0 (22)
in the observer’s frame. Now the apparent length of the rod is

affected by the fact that photons which arrive at the observer
at the same time are emitted at different times.P 1corresponds
to when the trailing end of the rod passes at time t 1 and P 2
corresponds to when the leading end of the rod passes at time
t 2 . Equating the arrival times for photons emitted from P 1
and P 2 at times t 1 and t 2 respectively,

D + ∆x cos θ D
t 1 + ----------------------------- = t 2 + ----
c c
(23)
∆x cos θ
⇒ t 2 – t 1 = -------------------
c
When the trailing end of the rod reaches P 2 the leading end
has to go a further distance ∆x – L which it does in t 2 – t 1
secs. Hence,

V∆x cos θ
∆x – L = -----------------------
c
L (24)
⇒ ∆x = -------------------------
V
1 – ---- cos θ
c
and the apparent projected length is
L sin θ L0
L app = ∆x sin θ = ------------------------ = --------------------------------- = δL 0(25)
1 – β cos θ Γ ( 1 – β cos θ )
This is another example of the appearance of the ubiquitous
Doppler factor.

2.7 Transformation of velocities
The Lorentz transformation
x = Γ ( x′ + Vt′ ) y = y′ z = z′
 Vx′ (26)
t = Γ t′ + --------
 c 2 
can be expressed in differential form:

dx = Γ ( dx′ + Vdt′ ) dy = dy′ dz = dz′
 Vdx′ (27)
dt = Γ dt′ + -----------
 c2 

so that if a particle moves dx′ in time dt′ in the frame S′ then
the corresponding quantities in the frame S are related by the
above differentials. This can be used to relate velocities in the
2 frames via
dx′
------- + V
dx Γ ( dx′ + Vdt′ ) dt
------ = ---------------------------------- = -----------------------
dt  Vdx′ V dx′
Γ dt′ + ----------- 1 + ----- -------
 c2  c 2 dt
(28)
vx′ + V
⇒ v x = ---------------------
V vx′
1 + -----------
c2

For the components of velocity transverse to the motion of S′ ,
dy dy′ vy′
------ = v y = ---------------------------------- = ------------------------------
dt  Vdx′  V v x ′
Γ dt′ + -----------
 c2  Γ  1 + -----------
 c2 
(29)
dz dz′ vz ′
----- = v z = ---------------------------------- = ------------------------------
dt  Vdx′  V v x ′
Γ dt′ + -----------
 c2  Γ  1 + -----------
 c2 
In invariant terms (i.e. independent of the coordinate system),
take

v || = Component of velocity parallel to V
(30)
v ⊥ = Component of velocity perpendiciular to V
then
v || ′ + V v⊥ ′
v || = ---------------------------- v ⊥ = ------------------------------------- (31)
1 + V v || ⁄ c 2 Γ ( 1 + V v || ⁄ c 2 )
The reverse transformations are obtained by simply replacing

V by – V so that:
v || – V v⊥
v || ′ = ---------------------------- v ⊥ ′ = ------------------------------------- (32)
1 – V v || ⁄ c 2 Γ ( 1 – V v || ⁄ c 2 )
and these can also be recovered by considering the differen-
tial form of the reverse Lorentz transformations.
2.8 Aberration
y y′
S v S′ v′
v⊥ v⊥ ′
θ θ′
v || v || ′
x x′

Because of the law of transformation of velocities, a velocity
vector make different angles with the direction of motion.
From the above laws for transformation of velocities,
v⊥ v⊥ ′ v′ sin θ
tan θ = ------ = -------------------------- = ------------------------------------ (33)
v || Γ ( v || ′ + V ) Γ ( v′ cos θ + V )
(The difference from the non-relativistic case is the factor of

Γ .)
The most important case of this is when v = v′ = c . We put
v || = c cos θ v ⊥ = c sin θ
(34)
v || ′ = c cos θ′ v ⊥ ′ = c sin θ′

and
V
β = ---- (35)
c
and the angles made by the light rays in the two frames satis-
fy:
c cos θ′ + V cos θ′ + β
c cos θ = ---------------------------- ⇒ cos θ = --------------------------
V 1 + β cos θ′
1 + ---- cos θ′
c
(36)
c sin θ′ sin θ′
c sin θ = ------------------------------------- ⇒ sin θ = -----------------------------------
 V  Γ ( 1 + β cos θ′ )
Γ 1 + ---- cos θ′
 c 

Half-angle formula
There is a useful expression for aberration involving half-an-
gles. Using the identity,
θ sin θ
tan --- = --------------------- (37)
2 1 + cos θ
the aberration formulae can be written as:
 θ  1 – β 1 / 2  θ′
tan --- = ------------ tan ---- (38)
 2  1 + β  2

Isotropic radiation source
Consider a source of radiation which emits isotropically in its
rest frame and which is moving with velocity V with respect
to an observer (in frame S ). The source is at rest in S′ which
is moving with velocity V with respect to S .
y 1 y′
– 1
sin ---
S Γ S′
x x′

Consider a rays emitted at right angles to the direction of mo-
π
tion. This has θ = ± --- . The angle of these rays in S are given
2
by
1
sin θ = ± --- (39)
Γ
(40)
These rays enclose half the light emitted by the source, so that
in the reference frame of the observer, half of the light is emit-
ted in a forward cone of half-angle 1 ⁄ Γ . This is relativistic
1
beaming in another form. When Γ is large: θ ≈ --- .
Γ
3 Four vectors
3.1 Four dimensional space-time

Special relativity defines a four dimensional space-time con-
tinuum with coordinates
x 0 = ct x1 = x x2 = y x3 = z (41)
An event is the point in space-time with coordinates x µ

where µ = 0, 1, 2, 3 .
The summation convention
Wherever there are repeated upper and lower indices, sum-
mation is implied, e.g.

3
Aµ B µ = A
∑ µ B µ (42)
µ=0
The metric of space-time is given by
ds 2 = η µν d x µ d x ν (43)
where
–1 0 0 0 –1 0 0 0
η µν = 0 1 0 0 Inverse = η µν = 0 1 0 0 (44)
0 0 1 0 0 0 1 0
0 0 0 1 0 0 0 1

(45)
Hence the metric
ds 2 = – ( d x 0 ) 2 + ( d x 1 ) 2 + ( d x 2 ) 2 + ( d x 3 ) 2 (46)
This metric is unusual for a geometry in that it is not positive
definite. For spacelike displacements it is positive and for
timelike displacements it is negative.
This metric is related to the proper time τ by
ds 2 = – c 2 dτ 2 (47)
Indices are raised and lowered withη µν, e.g. if A µ is a vector,
then

A µ = η µν A ν (48)
This extends to tensors in space-time etc. Upper indices are

referred to as covariant; lower indices as contravariant.
3.2 Representation of a Lorentz transformation

A Lorentz transformation is a transformation which preserves
ds 2. We represent a Lorentz transformation by:
xµ′ = Λµν x ν
(49)
η µν = Λ σ µ Λ τ ν η στ

That is, a Lorentz transformation is the equivalent of an or-
thogonal matrix in the 4-dimensional space time with indefi-
nite metric.
Conditions:
• detΛ = 1 – rules out reflections ( x → – x )
• Λ 0 0 > 0 – isochronous
For the special case of a Lorentz transformation involving a
boost along the x –axis

γ – βΓ 0 0
µ
Λ ν = – βΓ Γ 0 0 (50)
0 0 1 0
0 0 0 1
3.3 Some important 4-vectors

The 4-velocity
This is defined by
d x µ d x 0 d xi
u µ = --------- = ---------, -------- (51)
dτ dτ dτ

The zeroth component
d x0 dt dt
--------- = c ----- = c ----------------------------------------------------------------------------------------------------
dτ dτ
dt 2 + – c – 2 [ ( d x 1 ) 2 + ( d x 2 ) 2 + ( d x 3 ) 2 ]
c (52)
= ------------------- = cγ
v2
1 – -----
c2
Note that we use Γ for the Lorentz factor of the transforma-
tion and γ for particles. This will later translate into Γ for bulk
motion and γ for the Lorentz factors of particles in the rest-
frame of the plasma.

The spatial components:
d xi d x i dt
-------- = -------- ----- = γ v i (53)
dτ dt dτ
so that
u µ = [ γ c, γ v i ] (54)
The 4-momentum
The 4-momentum is defined by
E i
p µ = m 0 u µ = [ γ mc, γ mv i ] = ---, p (55)
c
where
E = c 2 p 2 + m 2 c 4 = γmc 2 (56)
is the energy, and
p i = γ m0 v i (57)
is the 3-momentum.
Note the magnitude of the 4-momentum
η µν p µ p ν = – ( p 0 ) 2 + ( p 1 ) 2 + ( p 2 ) 2 + ( p 3 ) 2
2 (58)
E
= –  --- + p 2 = – m 2 c 2
 c

3.4 Transformation of 4-vectors
Knowing that a quantity is a 4-vector means that we can eas-
ily determine its behaviour under the effect of a Lorentz
transformation. The zero component behaves like x 0 and the
x component behaves like x . Recall that:
x = Γ ( x′ + βx 0 ′ ) x 0 = Γ ( x 0 ′ + βx′ ) (59)
Therefore, the components of the 4-velocity transform like
U 0 ′ = Γ ( U 0 – βU 1 )
(60)
U 1 ′ = Γ ( – βU 0 + U 1 )
Hence,

v 1
cγ′ = Γ ( cγ – βγ v 1 ) ⇒ γ′ = Γγ  1 – β -----
 c
γ′v 1 ′ = Γ ( – βcγ + γ v 1 ) ⇒ γ′v 1 ′ = Γγ ( v 1 – cβ ) (61)
γ′v 2 ′ = γ v 2
γ′v 3 ′ = γ v 3
Transformation of Lorentz factors
Putting v 1 = v cos θ gives
v
γ′ = Γγ 1 – β -- cos θ
 (62)
 c 

This is a useful relationship that can be derived from the pre-
vious transformations for the 3-velocity. However, one of the
useful features of 4-vectors is that this transformation of the
Lorentz factor is easily derived with little algebra.
Transformation of 3-velocities
Dividing the second of the above transformations by the first:
Γγ ( v 1 – cβ ) ( v 1 – cβ ) ( v 1 – V)
v 1 ′ = ------------------------------- = ------------------------ = ------------------------ (63)
v 1 v 1 V v x

Γγ 1 – β -----   1 – β -----  
 c  c  1 – ---------
 c2 
Dividing the third equation by the first:

γ v 2 v 2
v 2 ′ = ------------------------------- = ---------------------------- (64)
v 1 V v x

Γγ 1 – β -----  
 c Γ  1 – ---------
 c2 
and similarly for v 3 . These are the equations for the transfor-
mation of velocity components derived earlier.

4 Distribution functions in special relativity
In order to properly describe distributions of particles in a rel-

pz
Distribution of momenta in
momentum space.
p
py
px

ativistic context and in order to understand the transforma-
tions of quantities such as specific intensity, etc. we need to
have relativistically covariant descriptions of statistical dis-
tributions of particles.
Recall the standard definition of the phase space distribution
function:
3 3 No of particles within an elementrary volume

f d xd p = (65)
of phase space

4.1 Momentum space and invariant 3-volume
The above definition of f ( x, t, p ) is somewhat unsatisfacto-
ry from a relativistic point of view since it focuses on three
dimensions rather than four.
Covariant analogue of d 3 p
The aim of the following is to replace d 3 p by something that
makes sense relativistically.
Consider the space of 4-dimensional momenta. We express
the components of the momentum in terms of a hyperspheri-
cal angle χ and polar angles θ and φ .

p 0 = mc cosh χ
p 1 = mc sinh χ sin θ cos φ
(66)
p 2 = mc sinh χ sin θ sin φ
p 3 = mc sinh χ cos θ
The Minkowksi metric is also the metric of momentum space
and we express the interval between neighbouring momenta
as η µν d p µ d p ν . In terms of hyperspherical angles:
η µν d p µ d p ν = – ( d ( mc ) ) 2 +
(67)
+ m 2 c 2 [ ( dχ ) 2 + sinh2 χ ( dθ 2 + sin2 θdφ 2 ) ]
This is proved in Appendix A.
The magnitude of the 3-dimensional momentum is
p = mc sinh χ (68)
A particle of mass m is restricted to the mass shell
m = constant . This is a 3-dimensional hypersurface in mo-
mentum space. From the above expression for the metric, it is
easy to read off the element of volume on the mass shell:
dω = ( mc ) 3 sinh2 χ sin θdθdφ (69)

This volume is an invariant since it corresponds to the invar-
iantly defined subspace of the momentum space,
m = constant .
On the other hand, the volume element d 3 p refers to a sub-
space which is not invariant. The quantity
d 3 p = d p1d p2d p3 (70)

depends upon the particular Lorentz frame. It is in fact the
projection of the mass shell onto p 0 = constant . However, it
is useful to know how the expression for d 3 p is expressed in
terms of hyperspherical coordinates.
In the normal polar coordinates:
d 3 p = p 2 sin θdpdθdφ (71)

Putting p = mc sinh χ in this expression,
d 3 p = ( mc ) 3 sinh2 χ cosh χ sin θd χdθdφ = cosh χdω (72)
That is, the normal momentum space 3-volume and the invar-
iant volume dω differ by a factor of cosh χ .
4.2 Invariant definition of the distribution function

The following invariant expression of the distribution func-
tion was first introduced by J.L. Synge who was one of the in-
fluential pioneers in the theory of relativity who introduced
geometrical and invariant techniques to the field.
We begin by defining a world tube of particles with momen-
tum p µ (4-velocity u µ ).

The distribution function f ( x µ, p µ ) is defined by:
t uµ
y
World tube of particles with 4-
nµ
velocity u µ . The cross-sectional
3-area of the tube is dΣ 0 .
dΣ
dΣ 0 x

The three-area of the world tube, dΣ 0 is the particular 3-area
that is normal to the world lines in the tube. Using dΣ 0 , we
define the distribution function, f ( x µ, p µ ) , by the following
definition:
Number of world lines within
the world tube with momenta = f ( x µ, p µ )dΣ 0 dω (73)
within dω
This is expressed in terms of a particular 3-area, dΣ 0 .

Now consider the world lines intersecting an arbitrary 3-area
(or 3-volume) dΣ that has a unit normal n µ . The projection
relation between dΣ and dΣ 0 is
dΣ 0 = dΣ × ( – c – 1 u µ n µ ) (74)
Proof of last statement

First, let us define what is meant by a spacelike hypersurface.
In such a hypersurface every displacement, d x µ , is spacelike.
That is, η µν d x µ d x ν > 0 . The normal to a spacelike hyper-
surface is timelike. The square of the magnitude of a unit nor-
mal is -1:
η µν n µ n ν = – 1 (75)
Example:
The surface
t = constant (76)
is spacelike. Its unit normal is:
n µ = ( 1, 0, 0, 0 ) (77)
We can also contemplate a family of spacelike hypersurfaces
in which, for example
t = variable (78)

This corresponds to a set of 3-volumes in which x 1, x 2 and
x 3 vary, that are swept along in the direction of the time-axis.
As before the unit normal to this family of hypersurfaces is
n µ = ( 1, 0, 0, 0 ) and the corresponding 4-velocity is
dt
u = c -----, 0, 0, 0 = ( c, 0, 0, 0 ) = cn µ
µ  (79)
 dτ 
In the present context, we can consider the set of 3-spaces
dΣ 0 corresponding to each cross-section of a world tube as a
family of such spacelike hypersurfaces. Each hypersurface is
defined as being perpendicular to the 4-velocity so that, in

general, u 0µ ≠ ( c, 0, 0, 0 ) . However, it is possible to make a
Lorentz transformation so that in a new system of coordinates
n 0µ = ( 1, 0, 0, 0 ) and u 0µ = ( c, 0, 0, 0 ) .
The significance of dΣ
What is the significance of a surface dΣ as indicated in the
figure? This is an arbitrary surface tilted with respect to the
original cross-sectional surface dΣ 0 .This surface has its own
unit normal n Σµ and 4-velocity, u Σµ = cn Σµ .

In the coordinate system in which the normal to dΣ 0 has
components n 0µ = ( 1, 0, 0, 0 ) , let us assume that the 4-ve-
locity of dΣ is u Σµ = ( γ c, γv ) . That is, dΣ represents a sur-
face that is moving with respect to dΣ 0 at the velocity v with
Lorentz factor, γ . The unit normal to dΣ is
n Σµ = ( γ , γβ
β) (80)
Relation between volumes in the two frames

Let dΣ be the primed (moving) frame. The element of vol-
ume of dΣ is

dΣ = d x 1′ d x 2′ d x 3′ (81)
At an instant of time in the primed frame denoted by dt′ = 0
d x 1 = γ ( d x 1′ + vdt′ ) = γd x 1′
(82)
d x 2 = d x 2′ d x 3 = d x 3′
Hence,
dΣ 0 = d x 1 d x 2 d x 3 = γd x 1′ d x 2′ d x 3′ = γdΣ (83)
Expression of the Lorentz factor in invariant form

Consider the invariant scalar product

n Σµ n 0, µ = ( γ , γβ ) ⋅ ( – 1, 0, 0, 0 ) = – γ (84)
Dropping the Σ subscript on n Σµ and using n 0, µ = c – 1 u µ we

have
γ = –c –1 u µ nµ (85)
Hence,
dΣ 0 = ( – c – 1 u µ n µ )dΣ (86)
Note that the “projection factor” is greater than unity, perhaps

counter to intuition.

Number of world lines in terms of dΣ
We defined the distribution function by:
the world tube with momenta = f ( x µ, p µ )dΣ 0 dω (87)
within dω
Hence, our new definition for an arbitrary dΣ :
the world tube crossing dΣ = f ( x µ, p µ ) ( – c – 1 u µ n µ )dΣdω
with momenta within dω

Counting is an invariant operation and all of the quantities ap-
pearing in the definition of f are invariants, therefore
f ( x µ, p µ ) = Invariant (88)
4.3 An important special case

Take the normal to Σ to be parallel to the time direction in an
arbitrary Lorentz frame. Then
n 0 = ( 1, 0, 0, 0 ) (89)
and
–uµ n µ = –( –u 0 n 0 ) = u 0 (90)

Also
dΣ = d 3 x (91)
Now
p 0 = mc cosh χ ⇒ u 0 = cosh χ (92)

Therefore,
f ( x µ, p µ ) ( – u µ n µ )dΣdω = f ( x µ, p µ ) cosh χd 3 xdω

(93)
= f ( x µ, p µ )d 3 xd 3 p

Our invariant expression reduces to the noninvariant expres-
sion when we select a special 3-volume in spacetime. Thus
the usual definition of the distribution is Lorentz-invariant
even though it does not appear to be.
5 Distribution of photons
5.1 Definition of distribution function

We can treat massless particles separately or as a special case
of the above, where we let m → 0 and cosh χ → ∞ in such a
hν
way that mc cosh χ → ------ . In either case, we have for pho-
c
tons,

No of photons within d 3 x
f ( x µ, p µ )d 3 xd 3 p = (94)
and momenta within d 3 p
and the distribution function is still an invariant.
5.2 Relation to specific intensity

From the definition of the distribution function, we have
Energy density of photons
= hνf d 3 p = hνf p 2 dpdΩ (95)
within d 3 p

The alternative expression for this involves the energy densi-
ty per unit frequency per unit solid angle, u ν ( Ω ) . We know
that
Iν
u ν ( Ω ) = ----- (96)
c
Hence, the energy density within ν and within solid angle dΩ
is
u ν dνdΩ = c – 1 I ν dνdΩ (97)
Therefore,

 hν 2  hν
hνf p dpdΩ = hνf ------ d ------ dΩ = c – 1 I ν dνdΩ
2
 c  c
(98)
c2 I ν
⇒ f = ------ ------
h4 ν3
This gives the very important result that, since f is a Lorentz
invariant, then
Iν
------ = Lorentz invariant (99)
ν3
Thus, if we have 2 relatively moving frames, then

I ν′ Iν ν 3
------------ = ------ ⇒ I ν =  ----- I ν′ (100)
( ν′ ) 3 ν 3  ν′ 
Take the primed frame to be the rest frame, then
I ν = δ 3 I ν′ (101)
where δ is the Doppler factor.
5.3 Transformation of emission and absorption

coefficients
Consider the radiative transfer equation:

dI ν jν
--------- = S ν – I ν S ν = ------ (102)
dτ ν αν
Obviously, the source function must have the same transfor-

mation properties as I ν . Hence
Sν
------ = Lorentz invariant (103)
ν3
Emission coefficient
The optical depth along a ray passing through a medium with
absorption coefficient α ν is, in the primed frame

In S′ In S
l′ l
θ′ θ
l′α ν′
τ = ------------ (104)
sin θ′
The optical depth in the unprimed frame is

lα ν
τ = ----------- (105)
sin θ
and is identical. The factor e – τ counts the number of photons
absorbed so that τ is a Lorentz invariant. Hence
lα ν sin θ′
------------ ------------ = 1 (106)
l′α ν′ sin θ
The aberration formula gives

sin θ
sin θ′ = --------------------------------- = δ sin θ (107)
Γ ( 1 – β cos θ )

and the lengths l and l′ are perpendicular to the motion, so
that l = l′ . Hence,
αν να ν
-------- = δ – 1 → ------------- = 1 (108)
α ν′ ν′α ν′
i.e. να ν is a Lorentz invariant.

The emission coefficient
Sν jν  jν
------ = ------------- =  ------ ( να ν ) – 1 = Lorentz invariant (109)
ν3 ν3αν  ν 2

Hence, j ν ⁄ ν 2 is a Lorentz invariant.
5.4 Flux density from a moving source

The flux from an arbitrary source is given by
V D
dΩ
Observer
1
F ν = ∫ I ν cos θ dΩ ≈ ∫ I ν dΩ = ------- ∫ j ν dV (110)
D 2
Ω Ω V
Now relate this to the emissivity in the rest frame. Since
jν
------ = Lorentz invariant
ν2
(111)
ν 2
j ν =  ----- j′ ν′ = δ 2 j′ ν′
 ν′
Therefore,
1 δ 2
F ν = ------- ∫ δ 2 j′ ν′ dV = ------- ∫ j′ ν′ dV (112)
D2 V D2 V
The apparent volume of the source is related to the volume in
the rest frame, by
dV = δdV′ (113)
This is the result of a factor of δ expansion in the direction of
motion and no expansion in the directions perpendicular to
the motion. Hence the flux density is given in terms of the rest
frame parameters by:
1 δ 3
F ν = ------- ∫ δ 2 j′ ν′ dV = ------- ∫ j′ ν′ dV′ (114)
D2 V D2 V
Effect of spectral index
For a power-law emissivity (e.g. synchrotron radiation),

 ν′ – α
j ν′ = j′ ν ----- = δ α j′ ν (115)
 ν
Therefore,
δ3 + α
F ν = --------------- ∫ j′ ν dV ′ (116)
D2 V
This gives a factor of δ 3 + α increase for a blue-shifted source

of radiation, over and above what would be measured in the
rest frame at the same frequency.

Example:
Consider the beaming factor in a Γ = 5 jet viewed at the an-
gle which maximises the apparent proper motion.
The maximum β app occurs when cos θ = β .
1
Γ = 5 ⇒ β = 1 – ------ = 0.9798
52
(117)
1 1
δ = --------------------------------- = ------------------------ = Γ
Γ ( 1 – β cos θ ) Γ( 1 – β2 )
Hence,

δ 3 + α = 5 3.6 = 330 (118)
for a spectral index of 0.6

5.5 Plot of δ
Plot of the
Doppler fac-
tor as a func-
tion of view-
ing angle.
δ

Appendix A
Line element in momentum space in terms of hyper-

spherical angles
We have the hyperspherical angle representation of a point in
momentum space:
p 0 = mc cosh χ
p 1 = mc sinh χ sin θ cos φ
(119)
p 2 = mc sinh χ sin θ sin φ
p 3 = mc sinh χ cos θ
We can write:

d p0 d ( mc )
d p 1 = A dχ (120)
d p2 dθ
d p3 dφ
where
cosh χ mc sinh χ 0 0
sinh χ mc cosh χ mc sinh χ – mc sinh χ
A = × sin θ cos φ × sin θ cos φ × cos θ cos φ × sin θ cos φ (121)
sinh χ mc cosh χ mc sinh χ mc sinh χ
× sin θ sin φ × cos θ sin φ × cos θ sin φ × sin θ cos φ
sinh χ cos θ mc cosh χ cos θ – mc sinh χ sin θ 0

Hence
d ( mc )
η µν d p µ d p ν = d ( mc ) dχ dθ dφ A† A dχ (122)
dθ
dφ
where

sinh χ sinh χ
– cosh χ sinh χ cos θ
× sin θ cos φ × sin θ sin φ
mc cosh χ mc cosh χ
– m c sinh χ mc cosh χ cos θ
A† = × sin θ cos φ × cos θ sin φ (123)
mc sinh χ mc sinh χ
0 – mc sinh χ sin θ
× cos θ cos φ × cos θ sin φ
– mc sinh χ mc sinh χ
0 0
× sin θ cos φ × sin θ cos φ
On matrix multiplication we obtain:
η µν d p µ d p ν = – [ d ( mc ) ] 2 + ( dχ ) 2 + sinh2 χ [ ( dθ ) 2 + sin2 θ ( dφ ) 2 ](124)

–1 0 0 0
0 1 0 0
A† A =
0 0 sinh2 χ 0
0 0 0 sinh2 χ sin2 θ

Relativistic Effects

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Relativistic Effects

Uploaded by

Copyright:

Available Formats

Relativistic Effects

The radio-emitting plasma in AGN contains electrons with

Relativistic Effects ©Geoffrey V. Bicknell

High Energy Astrophysics: Relativistic Effects 2/92

The primed frame is moving wrt to the unprimed frame with

High Energy Astrophysics: Relativistic Effects 3/92

High Energy Astrophysics: Relativistic Effects 4/92

High Energy Astrophysics: Relativistic Effects 5/92

High Energy Astrophysics: Relativistic Effects 6/92

High Energy Astrophysics: Relativistic Effects 7/92

i.e. the clock appears to have slowed don by a factor of Γ

High Energy Astrophysics: Relativistic Effects 8/92

The Doppler effect is very important when describing the ef-

High Energy Astrophysics: Relativistic Effects 9/92

High Energy Astrophysics: Relativistic Effects 10/92

D = distance to observer from P 2 (8)

High Energy Astrophysics: Relativistic Effects 11/92

rec D + V∆t cos θ

High Energy Astrophysics: Relativistic Effects 12/92

High Energy Astrophysics: Relativistic Effects 13/92

The factor Γ is a pure relativistic effect, the factor

High Energy Astrophysics: Relativistic Effects 14/92

2.5 Apparent transverse velocity

High Energy Astrophysics: Relativistic Effects 15/92

High Energy Astrophysics: Relativistic Effects 16/92

High Energy Astrophysics: Relativistic Effects 17/92

High Energy Astrophysics: Relativistic Effects 18/92

High Energy Astrophysics: Relativistic Effects 19/92

High Energy Astrophysics: Relativistic Effects 20/92

High Energy Astrophysics: Relativistic Effects 21/92

High Energy Astrophysics: Relativistic Effects 22/92

Plots of β app for various indicated values

High Energy Astrophysics: Relativistic Effects 24/92

High Energy Astrophysics: Relativistic Effects 25/92

High Energy Astrophysics: Relativistic Effects 26/92

in the observer’s frame. Now the apparent length of the rod is

High Energy Astrophysics: Relativistic Effects 27/92

High Energy Astrophysics: Relativistic Effects 28/92

High Energy Astrophysics: Relativistic Effects 29/92

can be expressed in differential form:

High Energy Astrophysics: Relativistic Effects 30/92

High Energy Astrophysics: Relativistic Effects 31/92

High Energy Astrophysics: Relativistic Effects 32/92

The reverse transformations are obtained by simply replacing

High Energy Astrophysics: Relativistic Effects 34/92

(The difference from the non-relativistic case is the factor of

High Energy Astrophysics: Relativistic Effects 35/92

High Energy Astrophysics: Relativistic Effects 36/92

High Energy Astrophysics: Relativistic Effects 37/92

High Energy Astrophysics: Relativistic Effects 38/92

3.1 Four dimensional space-time

An event is the point in space-time with coordinates x µ

High Energy Astrophysics: Relativistic Effects 40/92

High Energy Astrophysics: Relativistic Effects 41/92

High Energy Astrophysics: Relativistic Effects 42/92

This extends to tensors in space-time etc. Upper indices are

3.2 Representation of a Lorentz transformation

High Energy Astrophysics: Relativistic Effects 43/92

High Energy Astrophysics: Relativistic Effects 44/92

3.3 Some important 4-vectors

High Energy Astrophysics: Relativistic Effects 45/92

High Energy Astrophysics: Relativistic Effects 46/92

High Energy Astrophysics: Relativistic Effects 48/92

High Energy Astrophysics: Relativistic Effects 49/92