The ABINIT Users Mailing List ( CLOSED )

[verstraete@pcpm.ucl.ac.be]

Hello,

Firstly, your tsmear is quite high. It might not influence the energy too 
much, but unless you're using cold smearing it probably deforms the forces.

To get a good dos, 2 possibilities:
1) use a lot of kpoints. For a metallic system at least a 10x10x10 grid. 
For prtdos 1 use a lower tsmear or the whole DOS will be smeared out of 
visibility. I think this is your main problem. If you want a clean DOS, do 
a first calculation with whatever smearing you want, then redo a single 
SCF step with occopt 7 and a small tsmear (~0.001 to 0.01) to insure the 
full detail of the DOS can be seen.

2) use the tetrahedron method prtdos 2 (much better and uses fewer 
kpoints). However, I recommend a non-shifted and even numbered grid of 
kpoints; this insures you'll hit on the special points in most systems.
Take care: if your system doesn't have inversion symmetry, there is still 
a bug in 4.1.2 (use prtdos 1). The fix has been found by Michel Cote, but 
isn't implemented yet.

Matthieu

On Thu, 12 Jun 2003, Dominik Legut wrote:

> Dear Abinit users,
> 
> I have a general question concerning DOS.
> 
> I have converged the total energy in my calculation  to the precision
> of 3E-06 Ha per atom with respect to the number of k-points (40), cutoff
> (30 Ha) and tsmear parameter (0.05).
> 
> If I would like to get precise DOS, how much do I need to change
> the parameters mentioned above ?
> 
> I plotted the DOS (with parameters indicated above) and compared it to the
> FLAPW
> calculation; I found considerable discrepancies between those two.
> 
> I also increased the number of KP to be the same in both calculations
> (ABINIT and FLAPW cases) and compared both DOSs again, but they do not
> change too much (almost not at all) - there are still large differences
> between the FLAPW and ABINIT results.
> 
> Can anyone comment on this?
> Thank you in advance.
> 
> Dominik Legut.
> 
> 
> ---------------------------------------------------------------------
>  Dominik Legut                               tel.+420/5/32290461
>  Institute of Physics of Materials           fax.+420/5/41218657
>  Academy of Sciences of the Czech Republic   email: legut@ipm.cz
>  Zizkova 22, Brno
>  CZ-616 62
> ---------------------------------------------------------------------
> 
> 
> 

-- 
===================================================================
Matthieu Verstraete               mailto:verstraete@pcpm.ucl.ac.be 
PCPM, Boltzmann, pl. Croix du Sud, 1            tel: 010/ 47 86 81
B-1348 Louvain-la-Neuve  Belgium                fax: 010/ 47 34 52