symmetric slab optimization

Message

apple · #1 Post by **apple** » Mon Oct 15, 2012 11:04 pm

hi,
I am optimizing a surface slab of an inverse spinel structure with two sides prepared to be perfectly symmetric in the POSCAR file. However, when optimizing it with ISYM=0 there are some deviations (less than 0.026 A though) in the bond lengths for the equivalent atom pairs in the two surfaces (two sides). Can someone comment if the observed deviations are too big to be due to numerical noise. Could they be due to the isym flag set to zero, or there is something wrong with my poscar?
Thanks.

kambiz · #2 Post by **kambiz** » Thu Oct 18, 2012 12:36 am

Hi,

Are you using spin-polarized calculations? Are you adding any U term to your system?

apple · #3 Post by **apple** » Thu Oct 18, 2012 6:07 pm

Yes, I am doing spin-polarized calculations with Ueff terms. Why should it matter?

kambiz · #4 Post by **kambiz** » Tue Oct 23, 2012 6:11 am

Optimize your structure with non-spin polarized tag and then check your structure