[Wien] Slab convergence

[Lukasz Plucinski]

  On 9/14/2010 4:00 PM, Laurence Marks wrote:
> It will probably have no effect. Do a grep -e :NEC to see if you are 
> losing density from high-energy core states and need to move them into 
> the valence regime. I would not increase RMT's, you could run into 
> other problems. Good luck.

Dear Laurence,

Do you mean that I should leave RMT at 2.5 ?

I checked on several examples and it seems that RMT is never set above 
2.5 by the automatic procedure... is there any reason for this ? In 
manual I see there are some rules concerning rkmax parameter, maybe this 
is the reason ?

So far I tested all my Fe/Au(001) calculations with higher RMTs, around 
2.72, so maybe this is the reason for convergence problems.

Non-relaxed Au20 slab of 60Angstroem converged with RMT 1.72 after 
approx. 50 iterations.

  I am going to take a more systematic approach, first optimize the 
lattice constant (volume) of bulk Au, then relax the small slab, e.g 
Fe1Au5.  Maybe this way things start to converge.

Regards,
Lukasz