[Pw_forum] smearing methods and relaxation

Nicola Marzari marzari at MIT.EDU
Thu Nov 30 16:32:58 CET 2006



Hi Jess, you should use a smearing instead (usually methfessel-paxton or
marzari-vanderbilt). You need to use a small value - try 0.4 eV
(remember, pwscf uses Ry, so divide by 13.6), and make sure
that whatever you want to calculate is converged with respect
to k-points.

0.4eV might be too large for a lot of critical properties - so decrease
it (and increase, by a lot, the kpoints) and see if things change.

				nicola


Jess Kondor wrote:
> 
>  Hello, everyone,
> 
>  I have a question about smearing methods and relaxation, in 
> particularly  tetrahedron method and structural optimization  in case of 
> metals.  VASP manual recommends do not use tetrahedron method in 
> relaxation of metals because this method is not variational with respect 
> to the partial occupancies, which means, that the calculated forces are 
> wrong by 5 to 10 % in metals.  Is it applicable in pwscf case? I am 
> studying magnetic system (which is metal without U correction, and 
> should be insulator with LSDA+U) and I am not sure it is good idea to 
> use 'tetrahedra' in this case. What do you suggest?
> 
>  Regards,
> 
>   JK


-- 
---------------------------------------------------------------------
Prof Nicola Marzari   Department of Materials Science and Engineering
13-5066   MIT   77 Massachusetts Avenue   Cambridge MA 02139-4307 USA
tel 617.4522758 fax 2586534 marzari at mit.edu http://quasiamore.mit.edu



More information about the Pw_forum mailing list