Forces of previously relaxed geometry, should they be zero?

Message

hosfa · #1 Post by **hosfa** » Mon May 05, 2014 9:17 am

Hi,
I relaxed a geometry with regard to the forces less than 0.0001
I again ran a calculation with the final CONTCAR with every tag in INCAR the same as previous one except for NSW=0!
But I got huge forces: .gt. 0.01
What is behind this?

#2 Post by **admin** » Tue May 06, 2014 7:29 am

did the cell volume / shape change as well during your first run as well?
If so, a different basis set is initialized in the new run, which may lead to
larger forces again, depending on the change of the basis set

hosfa · #3 Post by **hosfa** » Sat May 17, 2014 9:53 am

[quote="admin"]did the cell volume / shape change as well during your first run as well?
If so, a different basis set is initialized in the new run, which may lead to
larger forces again, depending on the change of the basis set
[/quote]

Thanks a lot!
In may relaxation run(NSW=100) the cell shape and volume remained the same! Moreover the position of the atoms remained the same as well. (the initial geometry was relaxed, and I just increased the accuracy, KPOINTS,... to be on the safe side!)
But when I use the same incar with NSW=0 I get forces!
Looking forward for your comment.