unitcell optimization with exact exchange

bjeon · Post by **bjeon** » Tue Sep 06, 2011 8:01 pm

Hi, I am calculating V2O5 unit-cell using screened exchange (EECE).
For an optimized structure, I am running as:

1. ionmov = 2, optcell=0
After convergence,

2. ionmov = 2, optcell=2

Without EECE, abinit predict the unit-cell close to experimental data. But when EECE option is on,
unit-cell volume increases very high, yielding non-physical size (1.25e3 -> 1.5e3 or even larger).

Is there any option to constrain volume increase rate? Or any comment will be very appreciated.

ByoungSeon