ASTR 498N                 Lecture 9               Radiative transport

 

(online at www.astro.umd.edu/~drabin/)

 

                                                                      The clear light of truth.

 

 

We’ve now seen the basic equations of stellar evolution, including the equation expressing the radiative transport of luminosity:

 

                               

 

Today we’ll derive this equation and explore the opacity  that appears in it.

 

 

Microreview

 

Linear aborption coefficient                  

 

Cross section                                        

 

Mean free path                                      

                   

Mass absorption coefficient                  

 

 

Thermal conduction

 

Heat flows from hot to cold.  Experimentally, for a wide range of gases, liquids, and isotropic solids, the heat flux is found to be proportional to the temperature gradient:  , where K is the coefficient of thermal conductivity.  The microscopic basis for conduction is the random motion of particles.  We can estimate the form of K with a simple argument.

 

Let  be the mean energy of a particle at x.  At any moment, roughly 1/6 of the particles are headed in each of the 6 directions ±x, ±y, ±z.  Thus, per unit time and area, about  particles cross a plane of constant x from below, and a similar number from above.  The particles crossing from below experienced their last collision, on average, a distance l below the plane, where l is the mean free path, and therefore carry a mean energy      across the plane.  The net energy flux in the +x direction is

 

                               

 

We can write this as

 

                               

 

where the coefficient of thermal conductivity is

 

                               

 

and  is the specific heat per particle.  For a monatomic ideal gas,  and .  For a gas of neutral atoms, we can estimate the mean free path using  and  where a0  is the Bohr radius (O&C §9.2).  The collision cross section for an ionized gas is not so obvious.

 

What about a photon gas?  Rewrite the estimate above for conductive flux as

 

 

                     

 

where u is energy density and the wavelength averaged mass absorption coefficient  (the opacity) is effectively defined by its correspondence with the mean free path.

 

Finally, for blackbody radiation, u = aT4 where  is the radiation constant.  Using this in du/dT above gives

 

                               

 

In a spherical star, ,  and we obtain the equation for radiative energy transport:

 

                               

 

In view of our rough derivation of the coefficient of thermal conductivity, it is fortuitous that the equation comes out right, including the numerical factors.  However, the approximate derivation here brings out the underlying unity of heat conduction, whether by massive particles or photons.

 

 

Rosseland mean opacity

          (O&C §9.2)

 

 

As a homework problem (O&C Problem 9.16), you’ll derive from the radiative transfer equation the result

 

                               

 

For an isotropic radiation field, the radiation pressure is (Lecture 1) .  Further assume that the radiation is blackbody, Iλ = BλThen

 

                               

Integrating over all wavelengths,

 

                               

 

Now, if we define the Rosseland mean opacity by

 

 

 

we obtain

 

                               

 

which recovers our basic result while revealing the nature of the mean opacity.  Thinking in terms of transparency (1 /κλ) instead of absorption, the Rosseland mean transparency is the integral of the monochromatic transparency weighted by the temperature derivative of the Planck function.   Regions, if any, of particularly high transparency (low opacity tend to dominate the integral.  The weighting function de-emphasizes high and low frequencies.

 

Note that, when more than one source of opacity contributes significantly to the total opacity at some frequencies, the Rosseland mean of the total opacity is not in general equal to the sum of the individual Rosseland mean opacities. That is, if opacity sources κ1(ν) and κ2(ν) have Rosseland means  and , .

 

 

 

 

 

 

 

Sources of opacity

          (O&C §9.2)

 

Bound-bound transitions

          The absorption (or emission) of a photon when an electron makes an upward (or downward) transition between two bound energy levels is the familiar process that gives rise to discrete spectral lines.  Although bound-bound opacity is most important in cool stars and in the outermost layers of hotter stars, it makes a nontrivial contribution to the Rosseland mean over a fairly large region of the ρT plane.  In the Sun, the ratio of the total opacity with lines to that without can reach 23.  There is no convenient formula that encapsulates bound-bound opacity.

 

Bound-free transitions

          Bound-free absorption, also known as photoionization, occurs when an incident photon has enough energy to liberate an electron from a bound atomic state.  If we consider the ionization of an electron from a particular bound level, the cross section is zero until it spikes sharply (ionization edge) at = χ (χ is the ionization potential), then drops to higher energies as ν3.  Thus, the total opacity due to bound-bound transitions is a summation over sawtooth-like patterns followed by an integration over frequency.  Kramers showed in a semi-classical approximation that

 

                               

 

where the Gaunt factor gbf  is of order unity and the mysterious guillotine factor t is typically in the range [1,100].  The inverse process to photoionization is radiative recombination.

 

The negative hydrogen ion (H) is a dominant source of bound-free opacity in the outer layers of cooler stars.  There is a bound state for a second electron in the field of a proton, but the binding energy of the second electron is only 0.75 eV, compared to 13.6 eV for the first electron.  When hydrogen is predominantly neutral, metals (Z > 2), which often have first ionization potentials of only a few eV, provide most of the free electrons and therefore, through H, control the total opacity.

 

Free-free transitions

          A free electron in the vicinity of an ion can accelerate and in so doing absorb or emit a photon:  free-free absorption or free-free emission (also known as bremsstrahlung).  Kramers came to the rescue here too with the approximate forumula

 

                               

 

Note that  and  share the same functional dependence on ρ and T.

 

Electron scattering

          The cross section for electron scattering may be calculated classically by considering that an electromagnetic wave passing an electron has an electric field that causes the electron to oscillate, which then tends to re-emit a photon of the same frequency in another direction.  In terms of the classical electron radius ,

 

                               

 

Note that  does not depend on frequency, so .  The formula assumes a fully-ionized, pure hydrogen-helium mixture, which is generally a good approximation where electron scattering is important.  The cross section (and opacity) must be modified for relativistic effects at very high temperatures (  ) or if the electrons are significantly degenerate.

 

Electron conduction

          Because normally the mean free path of a photon is much longer than the mean free path of an electron or ion, heat conduction by particles can be ignored relative to radiative diffusion.  However, as we mentioned in our discussion of the Fermi-Dirac distribution, a highly degenerate electron gas becomes highly conductive because an electron on the surface of the Fermi sea cannot give up energy in a collision since all lower energy states are already occupied.  Electron conduction is important in the cores of evolved stars and in white dwarfs.

 


 

Opacity in the ρT plane

 


Rosseland mean opacity graphically