(online at www.astro.umd.edu/~drabin/)
Radiative Transfer
(O&C §9.3)
Linear absorption
coefficient (opacity) kν
![]()
Note that [kν-1] = m and is a measure of the mean free path for photon absorption.
The opacity kν includes any process that removes energy from the
frequency-specific pencil of radiation described by Iν, whether by change in direction (scattering), change in frequency, or
conversion to another form of energy.
Optical depth τν
Mass absorption
coefficient κν
Cross section σν
(O&C §10.5)
Emission
coefficient
P Does
?
Equation of radiative transfer
or
where the source function Sν is defined as
![]()
The radiative transfer equation is fundamentally probabilistic.
Formal solution of the radiative transfer equation:
The solution is
formal because, in general, we don’t know Sνit is part of the solution. However, we can see why “source function” is
a sensible name; and, in some cases, we can make a reasonable assumption about
the form of Sν.
P O&C Problem 9.22
Black body radiation
(O&C §3.4)
Intensity (Planck function)
Wavelength of maximum Bλ Tλmax
= 2.898e-3 m K (Wien displacement
law)
Frequency of maximum Bν
Tc/νmax
= 3.670e-3 m K

Wien approximation ( )
Rayleigh-Jeans approximation ( )
(Note:
still Bν)
Flux density (Stefan-Boltzmann law and constant)
Stellar effective temperature Te
Source function for a gas in TE P Why?
P Why can a star never be strictly in thermodynamic equilibrium?
Maxwell speed distribution
(O&C §8.1)
The density function of particle speed in a nondegenerate, nonrelativistic gas in TE is
where gives the probability of finding a
particle of mass m in the velocity range v to v + dv
at temperature T. Since it’s a
probability,
.
P What is vmp, the most probable speed for a particle of mass m and temperature T? Write this also as a convenient numerical formula for electrons and for hydrogen atoms.
If we normalize all speeds to the most probable speed, ,
the density of x has a universal shape:

The Maxwell speed density is derived from the more fundamental 6-dimensional phase space density describing the equilibrium of an ideal gas,
(Maxwell-Boltzmann)
where .
does not depend on the spatial coordinates x
because there is no spatial (or time) dependence in a true equilibrium. Note that this density is a pure Gaussian.
P Derive the density function for the radial (line-of-sight)
velocity (which governs spectroscopic Doppler shifts) and for the kinetic
energy .
The Maxwell-Boltzmann density function is approximately applicable in a wide variety of circumstances. However, we will find that the underlying Maxwell-Boltzmann statistics are not adequate to describe the state of all stellar interiors. We will also need Fermi-Dirac statistics and Einstein-Bose statistics (the Planck radiation function, above, can be derived from Einstein-Bose statistics, which apply to a photon gas).
Boltzmann formula (excitation)
(O&C §8.1)
where u
(upper) and l (lower) label two energy levels of an atom or ion, n (m-3)
is the population of the level, g is the statistical weight of the level
(the number of states with different quantum numbers but the same energy), and is the
energy difference between the levels.
The Boltzmann formula holds for thermodynamic equilibrium at temperature T. Sometimes an excitation temperature Tex is defined through this relationship even when TE is not known to hold, but such a temperature applies only to the specific levels involved.
Saha equation
(O&C §8.1)
(Saha)
describing the relative number densities of two ionization states of the same ion, r-times ionized and (r +1)-times ionized, in terms of the ionization potential χr, the temperature T, and the partition functions ur +1 and ur, where
The partition function is a summation of the statistical weights of all the bound states of the ion, each weighted by a Boltzmann factor to account for the relative population the level relative to the ground state. Some authors (including O&C) use Z instead of u to denote a partition function.
The Saha equation is sometimes multiplied through by
kT and written in terms of Pe, the electron
pressure. This assumes the perfect gas
law, , and is not correct if, for example, the gas is
partially degenerate.
For numerical work, it’s convenient to have the Saha equation in numerical form, e.g.,
where n is in m-3 and χr is in eV. This is the origin of the odd factor 5040/T that crops up in astronomical papers. Notation: “log” means log10, “ln” means loge.
It’s a useful fact, illustrated for atomic nitrogen below (the label above the red points should be N II, not N), that one rarely needs to consider more than two stages of ionization at once in a stellar interior or atmosphere.

The Saha equation applies to thermodynamic equilibrium. As in the case of the Boltzmann formula, an ionization temperature Ti can be defined with reference to a specific ion and levels.
P Derive an expression for ,
the population of an individual level k of an r-times ionized
species relative to the total population of that ion.
P O&C Problems 8.10(a), 8.10(b), and 8.(12).