Prof Wladimir Lyra ———– SELECTED TOPICS IN ASTROPHYSICS

Class notes - 2 (Aug 26th, 2015)

1 Principles of Radiative Transfer

The theory of radiative transfer is a macroscopic description of radiation fields. One of the most
primitive concepts is that of the energy flux, an intuitive measurement of energy flow. It is defined
as the energy passing through an area over time

dE ∝ dAdt (1)
where the flux is the proportionality factor, so that

dE = FdAdt (2)
We can also define a frequency range, so that the flux is the net energy crossing a unit area per
unit time in a frequency range.

dEν = Fν dAdtdν (3)

Considering an isolated star, if we put spherical surfaces s and S of radius r and R around it, we
know by conservation of energy that the total energy passing through s and S must be the same.
Thus,

F (r )4πr2 = F ( R)4πR2 , (4)

or
 2
R
F (r ) = F ( R ) . (5)
r
If we consider R as the radius of the star and r an arbitrary location away from the stellar surface,
Eq. (??) says that the flux falls with the square of the distance. This is merely a statement of
conservation of energy.

1.1 Intensity
The flux is a measurement of all rays that pass through a given area. A more fundamental descrip-
tion of the radiation field should consider the energy coming from individual rays. If we were
dealing with quantum mechanics, that would be a single photon. However, as said, the radiative
transfer theory is a classical (or semi-classical) theory, useful for dealing with radiation fields in a
macroscopic way. We wish then to consider the energy coming from individual rays. However, a
single ray carries essentially no energy, so we need to consider the energy carried by a set of rays,
that differ infinitesimally from the given ray. For that, construct an area dA normal to the direction
of the ray, and consider all rays that pass though dA whose direction is within a solid angle dω of
the given ray.

dE ∝ dA · dωdt (6)

1
and the proportionality is called specific intensity, or simply intensity. As with flux, we can also
define it in a frequency range

dEν = Iv dA · dωdtdν (7)

or

dEν
Iv = lim (8)
∆→0 dA · dωdtdν

Infinitesimally, the set of rays approach, in the limit, the energy of a single ray. The intensity has
dimension of

[ Iν ] = energy (time)−1 (area)−1 (solidangle)−1 (frequency)−1 (9)

= ergs s−1 cm−2 ster−1 Hz−1 (10)

Notice that both the area and the solid angle infinitesimals are vectors, pointing to their respective
normals, n̂ and n̂0 . Defining

dA = n̂dA (11)
0
dω = n̂ dω (12)
(13)

we have

dEν = Iv n̂ · n̂0 dAdωdtdν = Iv cos θdAdωdtdν (14)

where theta is the plane angle between n̂ and n̂0 .

1.1.1 Constancy of intensity along a ray

When we take the limit ∆ω → 0, the beam stops spreading out, approaching a single ray. Due to
energy conservation, the energy is constant along a ray, and thus the intensity does not depend on
distance. To prove this, consider a ray L, and choose two points along it, P1 and P2, separated by a
distance r. Then construct area elements around P1 and P2, call them dA1 and dA2 , and consider
the solid angle dω1 that goes from dA1 to dA2 . The energy flowing through this area into this
solid angle is

dE1 = I1 dA1 dω1 dt (15)

where dω1 is the angular size of dA2 as seen from dA1 , i.e., dω1 = dA2 /r2 .
The energy that crosses dA2 coming from dA1 is conversely

dE2 = I2 dA2 dω2 dt (16)

where dω2 = dA1 /r2 is the angular size of dA1 as seen from dA2 . Because energy is conserved,
dE1 = dE2 , and we have

I1 dA1 dω1 = I2 dA2 dω2 (17)

2
Substituting the definitions of the solid angles, we conclude that I1 = I2 , and the intensities at P1
and P2 are equal.
This means that the surface of the Sun is as bright seen from Earth as it is seen from Neptune.
This may seem counterintuitive at first, until we realize that a wall does not get brighter as we
approach it. Intensity is an intrinsic property of resolved sources of radiation.

1.2 Flux
Flux is the net energy in a frequency interval that passes per unit time through a unit area in a
frequency range

dEν
Fν = . (18)
dAdνdt
Comparing that with the definition of intensity, we have that
Z Z
0
Fν = Iv n̂ · n̂ dω = Iv cosθdω (19)

i.e., the flux is the directional integral of the intensity. For an isotropic field, where Iν does not
depend on direction
I Z 2π Z π/2
Fν = Iν cos θdω = Iv dφ cos θ sin θdθ = πIν (20)
0 0
The flux is integrated in solid angle, so the directional information is eliminated. The inten-
sity can only be defined for extended objects, angularly resolved, because we need to specify the
interval of solid angles. Flux instead can be defined for any source, including point sources. There-
fore, there is a transition between measuring intensity and flux as we distance an observer from a
source.

1.2.1 Bolometric Flux, Stefan-Boltzmann law, and stellar luminosity

The bolometric flux is the flux integrated in all wavelengths.
Z ∞
Fbol = F = Fν dν (21)
0

dE
F= (22)
dAdt
The quantity dE/dt in physics is called power, yet in connection to stars we prefer to call it lumi-
nosity. So, Flux × Area = Luminosity.
For a source of constant luminosity, as most sources in astrophysics are (in the timescales we
measure them), the product Flux × Area is constant. Therefore, the flux falls with area following
an inverse square law, recovering Eq. (??).
Notice that the luminosity can be determined knowing the radius of the star and the flux in its
surface.

L? = Area × Flux = 4πR2 × F? (23)

where we have to determine the flux at the surface of the star. Approximating a star as a blackbody,
we can write

3
 Iν = Bν (24)
where Bν is Planck’s law

2ν2 hν
Bν = (25)
c2 exp (hν/KT ) − 1
Because we only see one hemisphere of the star, we need to integrate through the half-sphere.
Z 2π Z π/2
Fν = Bν cos θ sin θdθdφ = πBν (26)
0 0
For the bolometric flux we integrate in frequency
Z ∞ Z ∞
2πh ν3 dν
F= Bν dν = 2 (27)
0 c 0 exp (hν/KT ) − 1
Substituting x = hν/KT, we have
4 Z ∞ x3 dx

2πh KT
F= 2 (28)
c h 0 ex − 1
The integral can be evaluated to π 4 /15. The result is thus
 4 2 
K π
F= T 4 = σT 4 (29)
60c2 h̄3
This result is called Stefan-Boltzmann law. The constant that appears
 4 2 
K π
σ= (30)
60c2 h̄3
is the Stefan-Boltzmann constant. Its numerical value is 5.6704 × 10−5 erg cm−2 s−1 K−4 .
Based on this, we can write the stellar luminosity

L? = 4πR2? σT?4 . (31)

1.3 Wien’s displacement law

Wien’s displacement law states that the peak wavelength of a blackbody radiation, multiplied by
its temperature, is constant.

λmax T ≡ const = 2.898 × 10−3 mK (32)

This law follows immediately from Planck’s curve, by taking the derivative wrt frequency (or
wavelength)

c2 dIν 3ν2 ν3 h
= − 2
ehν/KT (33)
2h dν exp (hν/KT ) − 1 (exp (hν/KT ) − 1) KT
and equating it to zero, we find the frequency νM where the intensity peaks.
2  
νM νM exp (hνM /KT ) h
3− =0 (34)
exp (hνM /KT ) − 1 (exp (hνM /KT ) − 1) KT

4
Substituting x = hνM /KT, it reduces to

xe x − 3e x + 3 = 0 (35)
which we can solve numerically to find x ≡ const, i.e, νM /T ≡ const and thus λ M T ≡ const as in
Wien’s law.
Eq. (??) and Eq. (??) already explain a lot of features of the HR diagram. The luminosity is a
function of the radius (more emitting surface), so dwarfs are less luminous than giants. Betelgeuse
is about 1000 times bigger than the Sun, so the area by itself would account for a million-fold
factor in luminosity. The luminosity is also a very strong function of the temperature. A change
in temperature by a factor 10, as in the main sequence from M to O stars, implies a 104 change in
luminosity.

1.4 Magnitudes
The flux depends on distance, so the magnitudes we measure are apparent. For absolute magni-
tude, we need a standard distance, that we define as D=10 pc.

m = −2.5 log F + C (36)

with

L?
F= (37)
4πd2
The absolute magnitude is
 
L?
M = −2.5 log +C (38)
4πd2
We can define also the distance modulus, which is the difference between apparent and absolute
magnitude

L? 4πD2
   
d
m − M = −2.5 log = 5 log (39)
4πd2 L? 10pc
The distance is thus

d(pc) = 100.2(m− M)+1 (40)

It is also useful to express this as
00
M = m + 5 log π + 5 (41)
00
where π =1/d(pc) is the parallax angle (do not confuse it with the usual circle-related number!).
The equation above depends only on measurable quantities and is useful to have in handy when
observing.

5
1.5 Mean intensity
We can define also the mean intensity Jν , which is the directional average of the intensity

H
Iν dω
Jν = H (42)
dω
1
I
= Iν dω (43)
4π
This quantity is related to the energy density of the radiation field. Consider it the energy per unit
volume in a frequency range. It is useful as a step to define it also per unit solid angle

dEν = uν,ω dVdνdω (44)

The volume defined by a light beam is dA c dt, where c dt is the distance light travels in a unit time
dt. So,

dEν = uν,ω c dA dt dν dω = Iν dA dt dν dω (45)

Comparing both identities, we find uν,ω = Iν /c, and integrating it in solid angle

1
I
uν = Iν dω (46)
c
4π
uν = Jν (47)
c
For a black body,

4π 4σ 4
Z
u= Bν dν = T = aT 4 (48)
c c
The constant

4σ
a= = 7.5667 × 10−15 ergs cm−3 K−4 (49)
c
is called the radiation constant.

1.6 Solved problem

A spherical star of radius R emits radiation of intensity Iν , in all directions. For a distance r,
describe the radiation field, obtain the mean intensity, and the flux.
The observer sees an anisotropic radiation field, with the star subtending a finite angular size, and
elsewhere having zero intensity. The field is thus

Iν for 0 ≤ θ ≤ θ M , 0 ≤ φ ≤ 2π
Iν (θ, φ) = (50)
0 for θ M < θ ≤ π/2, 0 ≤ φ ≤ 2π
where θ M = asin( R/r ).
The mean intensity is

6
 1 Iν 2π θM
I Z Z
Jν = Iν dω = dφ sin θdθ (51)
4π 4π 0 0
1 1
= Iν (1 − cos θ M ) = Iν (1 − cos (asin( R/r ))) (52)
2 2
q
using (1 − cos(asin( R/r ))) = 1 − sin2 (asin( R/r )) = 1 − ( R/r )2 = 1/r (r2 − R2 )−1/2
p

Iν
Jν = (r − (r2 − R2 )−1/2 ). (53)
2r
The flux is

Z Z 2π Z θM
Fν = Iv cosθdω = Iν dφ cos θ sin θdθ (54)
0 0
= πIν sin2 θ M = πIν sin2 (asin( R/r )) (55)
 2
R
= πIν (56)
r