[Wien] NON-ZERO forces for zinc-blend GaAs

[Ulrich Keßler]

Oleg Rubel wrote:

>Dear Wien2k Users,
>
>I met a problem with NON-ZERO forces while performing SCF calculation for
>a zinc-blend GaAs supercell.
>
>
>Structure: My supercell belongs to a rhombohedral crystal system with 4
>non-equivalent atoms and the edges along [-1 1 0], [101], and [111]
>directions (please see the structure file below). (I need this structure
>for further calculation of {111} antiphase boundaries.) This structure
>cannot be reduced down to a primitive 2-atom basis due to explicit
>specifying of Ga and As atoms as non-equivalent (As1, As2, Ga1, Ga2).
>
>
>Problem: When I perform 'run_lapw -ec 0.00005', I do see non-zero forces,
>though I expect the opposite since the structure is perfect (I verified
>that by measuring of interatomic distances and angles in XCRYSDEN):
>
>  
>
Dear Oleg,

are you sure, your forces really should be zero?

I didn't look at your cell carefully, but I think in your case symmetry 
does not force all bonds being of equal length.

For example, in wurtzite the Zn-S-bonds are not exactly of the same length.
Or the metals, that crystallize in hcp: most of them exhibit different 
M-M-distances.

Yours,
Uli