The ABINIT Users Mailing List ( CLOSED )

[Geneste Gregory <geneste@cemes.fr>]

Hello !
Two suggestions:

1 -  Check that the ground state you find (in both cases) is
spin-polarized with the appropriate magn.
(Check your value of spinat!)

2 -  the value of the bond energy depends on the result you find for the
ground state of the free atom:
From a general point of view with open-shell atoms, you find a lower total
energy by using NSYM 1, which breaks the spherical symmetry (your
supercell is a cubic box). Try this, you should have a lower total energy
for your atomic reference, and thus a lower bond energy for the
molecule... (tell us if it is enough)

The O2 bond energy I calculate with ABINIT is 5.95 eV, using a non
spherical reference for the free atom (GGA-PBE, TM pseudopotentials).
Other results with other codes in GGA-PW91 are around 6 eV. The calculated
bondlength is always excellent (very close to 1.21 Ang).
Good luck,
Gregory