Re: today is my day for polarization questions...

[Alex Ankudinov <alex@leonardo.phys.washington.edu>]

Hello FEFF users,

Our code allows one to include x-ray polarization of incoming
beam using POLARIZATION, ELLIPTICITY and XMCD cards
(see detailes in the document).
ELLIPTICITY specifies the direction of incoming x-ray
and the ratio of intensities along two perpendicular
directions, one of which is specified by POLARIZATION card
and the other is found automatically.

ELLIPTICITY is not intended to be used for circular polarized light.
One has to use XMCD card, which assumes 100%  circular polarized light.

For general description of the polarizarion of monochromatic x-rays
 Stokes-Poincare components P1,P2, P3 can be used. Probably, at some
stage FEFF should be modified to use P1,P2,P3 to specify polarization
of incoming x-rays.

Answers to the Bruce's questions are below:

>     I am looking at some grazing incidence data from metals adsorbed
>     to a Langmuir monolayer, and I am trying to model it with FEFF
>     8.2. Since the monolayer is a 2D powder in the xy-plane, I am
>     modeling it with circular polarization in the xy-plane, like
> 
>     ELLIPTICITY 1.0 0.0 0.0 1.0
>     POLARIZATION 1.0 0.0 0.0

      This is correct use the FEFF card to average over polarizations
     in xy-plane.

> 
>     I was wondering, for an _L3_ edge, does FEFF use the complete
>     polarization dependence with the cross-terms between s and d
>     states (e.g. as shown in Stohr and Jaeger, Phys Rev B, vol. 27,
>     No.8 pp.5146), or does it use some approximation (neglecting some
>     terms)?  I looked at the papers that are quoted as references in
>     the FEFF manual, and the manual itself, but it does not have
>     anything on polarization. I figured you might know something about
>     it.

     By default, FEFF includes both p-->s and p-->d transition and
the cross term. The cross term contribution from FEFF7 and experiment
are compared on Fig.2 in Ankudinov,Rehr, Phys. Rev. B 56, p. R1712 (1997).
In the latest feff8.2 code (soon to be released for all users) one
can use MULTIPOLE card to add quadrupolar or magnetic dipolar
contributions, and also to turn off L-->L-1 contribution (e.g. p-->s).
Thus MULTIPOLE card allows one to turn on/off various channels of
x-ray absorption.

I'll be glad to answer further questions regarding use/implementation
of x-ray polarization in feff code.

Alex Ankudinov