Re: comparing results of LDOS and XANES calculations

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

Hi Faisal,

  Thanks for your comments.

> 1. I am attaching the sum of all the p-symmetry LDOS as compared to the mu 
> plot.  Sorry that the axes are not labeled.  The x-axis here is in electron 
> volts.  As you can see, there is significant disagreement between the pDOS 
> and the mu plots, especially in the region of ~ 10 eV where there is no sharp 
> p feature.  Where is this feature in the mu plot coming from, if it is not 
> present in the pDOS?

   Except for the big peak at -10 eV which I don't understand either, there 
does seem to be a reasonable correspondence between the peaks. So I don't 
know which feature you are discussing.  Perhaps it would help
if you were to scale the pDOS so the pDOS and XAS are comparable in
magnitude. Also you should plot the matrix elements, col 2 vs 5 in
XMU.  The XAS is related to the pDOS by an energy dependent factor,
in brackets below:

    mu(E) = mu_0(1 + chi(E)) = [mu_0(E)/rho_0(E)] rho (E) theta_Gamma(E-E_F)

> 2. In order to make a proper comparison  between the pDOS and the mu 
> functions, I suppose I need to add a step function to the pLDOS .  Is that 
> correct?  Would it instead be valid to amplify the amplitude of the pLDOS

  Yes - only the unoccupied DOS is related to the XAS. In the above
theta_Gamma(E-E_F) is a broadened step function.


   Could you please send me your feff.inp so I can understand better
what's going on e.g., at -10 eV?

   Thanks.
   J. Rehr