[Pw_forum] SiO2 xanes calculation

[maxradin at umich.edu]

>  >Hi,
>  >
>  >I'm trying to repeat the ultrasoft core-hole calculation of the Si
>  >>K-edge
>  >in Matteo's paper:
>  >
>  >"First-principles calculations of x-ray absorption in a scheme
>  >based on ultrasoft pseudopotentials: From alpha-quartz to
>  >high-Tc  compounds"
> 
>  >I haved used the same parameters given in that article (cell size,
>  >>cutoff
>  >energies, kpoints) but the spectrum that I get does not at all look 
>  >right.
>  >Is it possible that this discrepancy is due to the choice of
>  >pseudopotentials? For the absorbing atom I am using
> 
> Dear Max,
> 
>    the pseudo you use are correct, however this is not the case for
> the k-points in your Xanes calculation. In the paper it is stated
> that you need "the Lanczos method was performed
> using a centered 3x3x3 k-points grid of the 72
> atoms supercell." (see just below Eq. 27).
> 
> The second thing is that in an insulator, there is no Fermi level.
> Are you sure you're cutting eliminated the occupied states at the right 
> energy position ?
> 
> Can you try yo see if you reproduce the results of the QE XSpectra 
> example directory ?
> 
> all the best,
> 
> Matteo

After closer inspection, the major problem was due to unit conversion
(lattice constants were in Angstroms instead of Bohr). But regarding the
occupied states: as I understand it, for insulators, one should do an scf
calculation (without smearing) followed by an xspectra fermi level
calculation to find the energy to cut the occupied states. In the case of a
gamma only
kpoint grid, could one simply use the energy of the highest occupied band 
instead of doing the fermi level calculation? 

Thanks,

Max