[Wien] Unitary transformation for projected density of states

[Peter Blaha]

> I am trying to calculate DOS projected on a particular basis of a certain
> l-subshell (l=2,d) which is different from the basis used in the scf
> calculation.
> Namely I want to calculate the DOS in a local (atomic site) reference frame
> that is rotate rispect to one dictated by the point group symmetry.
> This is in principle possible using the density matrix spectral decomposition
> for the considered atom. This matrix  is to be transformed according to the
> required basis by a unitary transformation which is provided by the files
> $WIENROOT/SRC_templates/case.cf_*
> My question is:
> Is there anyone who can explain the format and the structure of these files?
> For example
> $WIENROOT/SRC_templates/case.cf_d_eg_t2g    ?

I'm by no means an expert and have never used it myself. Anyway:

This file represents (for l=2) a  10x10 matrix, which is blocked into
spin-up and down blocks.
For nonrelativistic cases this matrix is blockdiagonal with two identical
subblocks. (as for the eg-t2g example).

Each matrixelement is a complex number !, i.e. 2 numbers give real and imag.
part of it.

And of course they are ordered according to m=-2,-1,0,1,2

The * ends a certain representation, thus in the eg/t2g example
the first line says:   take the m=0 (d-z^2) orbital
    second line        take 1/sqrt2(Y22 + Y2-2)   (dx2-y2)
and this forms the eg orbital.



                                      P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-15671             FAX: +43-1-58801-15698
Email: blaha at theochem.tuwien.ac.at    WWW: http://info.tuwien.ac.at/theochem/
--------------------------------------------------------------------------