GGA and VOSKOWN

Message

danielk · #1 Post by **danielk** » Fri Dec 15, 2006 11:21 pm

In the manual, VOSKOWN = 1 is recommended for PW91. Is it also recommended for PBE? I found a VASP document online which states that VOSKOWN = 1 is the default for PBE, but when I check my OUTCAR file it has VOSKOWN set to 0. Can anyone speak to why VOSKOWN = 1 is preferred and what effect it has on the calculations? I haven't been able to find any discussions of this in my searches. Any clarification of this issue would be greatly appreciated.

#2 Post by **admin** » Mon Dec 18, 2006 3:16 pm

VOSKOWN only has to be set for PW91-GGA calculations, it affects the parametrization of the LDA-part of the XC-contribution of the free electron gas.
Setting VOSKOWN=1 leads to the following results:
the magnetization / V curves are (almost) parallel:
--) for the free atom limit (V-> infinity), there is no change in the magnetic moments
BUT
--) the onset of magnetisation in mag/V curve is shifted to lower V
--) at volumes in the region of onsetting magnetism, the magnetic moments are higher
--) in adsorption calculations: adsorption on non-magnetic sites is slightly preferred.

danielk · #3 Post by **danielk** » Tue Dec 19, 2006 4:50 pm

Thank you for that very useful and prompt reply.