[Wien] Non-zero :EFG at cubic point group!

[Stefaan Cottenier]

>   At a site with cubic  point group, electric field gradient is zero.
>       As I expected, I  have found nothing for the :EFG001 parameter  
>  using grep linux command, since  the point group of my first atom  
> in  the case.struct is m-3m.
>           However, I have  found for the above atom a non-zero EFG   
> after including spin-orbit coupling:
>       :EFG001:                         EFG         =      0.08163     
> *10**21  V / m**2

This is correct. Spin-orbit coupling breaks cubic symmetry, and small  
EFG's of that order of magnitude have been observed also experimentally.

>       I have then, as you can see below, observed that the  symmetry  
>  has not reduced yet:
>       :PCS001: PARTIAL CHARGES SPHERE =  1 S,P,D,F,      D-EG,D-T2G
>
>       :QTL001_up: 0.9684 2.6019 5.3193 0.7528 0.0000 0.0000 0.0000    
> 2.1118 3.2075 0.0000 0.0000 0.0000
>       :QTL001_dn: 0.9675 2.5964 5.3005 0.1819 0.0000 0.0000 0.0000    
> 2.1067 3.1938 0.0000 0.0000 0.0000

Also this observation is correct: when adding spin-orbit coupling, the  
ISPLIT parameters are not changed (I don't know whether this has a  
specific reason, I would expect it to be more consistent if they were  
changed -- can anybody comment?). If you want the splitting that is  
consistent with your point group, you have to change ISPLIT yourself  
(in simple cases this is most easily done by elongating your lattice  
slightly along the spin-orbit direction, init_lapw, and then restore  
the original dimensions of the lattice).

> NB: Here I would take the opportunity to make a little bit   
> modification on my last posted problem,   
> http://zeus.theochem.tuwien.ac.at/pipermail/wien/2006-August/007639.html,   
> where I wrote delta(d) formula at the top of e-mail. The sing of the  
>  last term, d_z2, has to be reversed. This cannot fix my last  
> problem  since all the data (unfortunately!) was computed using the  
> correct  formula. This means that I am still looking forward to  
> hearing from  you on that respect as well, since my last problem  
> still survives.

Still thinking...

Stefaan


Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm