Re: FEFF questions - OVERLAP and SS cards

["John J. Rehr" <jjr@leonardo.phys.washington.edu>]

Dear Yuri,

 Thanks for your posting of questions about the use of the
OVERLAP and SS cards in FEFF8.

>We have some troubles with using FEFF8. They are concerned with following
>cards:
>OVERLAP and SS.

  We're sorry the OVERLAP card is not easy to understand. No one, including
us likes it or uses it much, and we DO NOT recommend it, except possibly when
one knows absolutely nothing about the structure. We've kept it in the code
for that purpose only. In fact, even in the case when nothing is known,
it's probably better to start with a reasonable guess for a structure
that has the desired coordination chemistry by explicitly listing the
cartesian coordinates and atomic numbers using the ATOMS card in feff.inp
than to use OVERLAP and SS.

>As for OVERLAP card - the  examples from description for version 8.10
>(paragraphs 2.2)  cannot describe the full logic of using this card. The
>main difficult in understanding are:
>1) OVERLAP n means the overlapping of n-th coordination sphere, isn't it?
>2) n belongs to the interval (0..7), what does it mean?
>3) How can we calculate, for example EXAFS Cu, on 10 spheres?

1) Not precisely. OVERLAP n aims to determine which atoms are overlapped
with those in the nth coordination sphere. For example OVERLAP 0 tells
the code which atoms are overlapped with the atom at the origin, ipot=0.

2) n labels a potential index ipot which ranges from 0 to 7, 0 is
the central atom, 1 the atoms in the 1st shell of neighbors etc.,
as described in more detail in the FEFF8 doc.

3) Sorry, I'm not sure I understand your question as stated.  Do you
mean you want the EXAFS for Cu of up to 10 coordination spheres??
If that's the case using single-scattering calculations would be a
bad approximation. If you mean you want single scattering EXAFS from
10 Cu atoms in the 1st coordination sphere at 2.54 Ang using the SS
card, then you would need SS 29 1 10 2.54. If you just want the EXAFS
for Cu metal, forget the SS and OVERLAP cards and use the example
for Cu in Sec. 4.2.2.

  The example given in the document for the OVERLAP card (an advanced card) was
written to simulate what the FEFF6 and FEFF7 code does to get an 
overlapped atom potential for Cu metal in the case one knows absolutely nothing
about the structure. Thus if you really want to use the OVERLAP card
I suggest that you follow the example exactly as described the document.
But please let us know if it doesn't work.

>As for SS card (paragraphs 2.6) -  we quite understand the meaning of its
>parameters, but is it necessary to define this card each time before OVERLAP
>card description?

The SS card has nothing directly to do with the OVERLAP card per se.
The OVERLAP card is used to get the scattering potential and phase shifts
and the SS card to get single-scattering EXAFS from a given coordination
sphere.

>What additional cards do we need to use with OVERLAP and SS cards?
 TITLE and POTENTIALS
  
  Our recommendation: DO not use OVERLAP and SS cards. Find a trial
structure that matches the system you want to investigate and start
with that. 

  Sincerely,
  J. Rehr