ABINIT Discussion Forums

Hi to all.

I study a system that has two phases: ordered and disordered (at the same stoichiometry and unit cell). I have found that the atomic ordering is energetically preferable with the ordering energy of 22 eV. I am not comfortable with such a huge value, since an investigation of disordering with the use of 4x4x4 supercell shows that the ordering energy is about 200~meV.

Can I consider the obtained result as an adequate one when comparing total energies of ordered and disordered phases with the latter considered within the VCA?

Best regards, Igor.

Hi Igor,

Are you saying that the difference between the ordered structure and the VCA structure is 22 eV?

If it is the case, there was probably a problem in the calculation, all the more so since you only have a 220 meV difference using supercells.

For VCA, you need a mixture of pseudopotentials, right?

Hi!

This is part of my input file where I use VCA.

#DATASET 1
prtden1 1
toldfe1 1.0d-10
kptopt1 1
nshiftk1 1
shiftk1 0.0 0.0 0.0
ngkpt1 7 7 7
iscf1 7

#DATASET 2
iscf2 -2
getden2 1
kptopt2 -2
ndivk2 2*120 #!
#prtvol 2
kptbounds2
0.3333 0.3333 0.5000
0.0000 0.0000 0.5000
0.5000 0.0000 0.5000

enunit2 1
tolwfr2 1.0d-18

nspinor 2
intxc 1
nband 40
ixc 11

npsp 3
znucl 1 2 3
ntypat 2
ntypalch 1
mixalch 0.5 0.5 # mixture
natom 3
typat 1 2 2

ecut 20

rprim
…..

xred
…...