[Pw_forum] big atomic forces

[Xunlei Ding]

Dear Jess,
Is your system metals or /semiconductors and insulators ?/
Tetrahedron is good for scf calculation, but not for relaxation of 
metals, because of the forces from this method are not good for metals.

I found below on web. Hope it is useful.
For the calculation of the / total energy/ in bulk materials we 
recommend the tetrahedron method with Blöchl corrections. This method 
also gives a good account for the electronic density of states (DOS). 
The only drawback is that the methods is not variational with respect to 
the partial occupancies. Therefore the calculated forces are wrong by 5 
to 10 % in metals.

Best regards,
Ding

Jess Kondor wrote:

>
> Hello, everybody,
>
>  I have a question. I relaxed a surface using 8x8x1 k-point mesh and 
> 'm-v' smearing method with degauss=0.01 with criteria for forces 4.d-4 
> . Then I performed 'scf' calculation with increased number of k-points 
> (12x12x1) and turn on an option 'tprnfor = .true.' in order to 
> calculate forces. Forces are OK (all lower than 4.d-4). But when I 
> perform 'scf' with occupation method 'tetrahedra'  (other parameters 
> are the same), forces on atoms are much higher (some of them are 
> ~2.0d-3). Does it mean that geometry relaxation is not been completed?
>
> cheers,
>    Jess
>