Some applications require a complete extinction matrix K either for fixed
orientation or orientation-averaged. For fixed orientation K can be easily
obtained from the amplitude scattering matrix at forward direction -
Eqs.(2.140)-(2.146) in Mishchenko's book (2002). For complete orientation
averaging of mirror-symmetric particle K is trivially expressed through <Cext>
- Eq.(4.32).
The most problematic case is chiral particles and/or partial orientation
averaging. Then K has non-diagonal elements, e.g. K14 and K41. Computing
averaged K is, in principle, possible by doing averaging manually (with scripts
and multiple runs), but doing it in the framework of the existent orientation
averaging machinery in ADDA would be much more convenient.
Having the extinction matrix will also help to determine Cext for circular (or
arbitrary) incident polarization (issue 30) - Eq.2.159.
The first implementation of this feature should be based on Eqs.(2.140)-(2.146)
mentioned above, but they are valid only for the plane incident wave (like the
far-field expression for Cext). An interesting idea is to generalize the
definition of K to Gaussian beams. I am not sure if that any sense in the first
place, but that can only be made through some volume integral definition for K
(to be derived).
Original issue reported on code.google.com by
yurkinon 11 Mar 2014 at 2:06