ABINIT Discussion Forums

Hi,
I am experiencing problems to calculate the electronic structure of LiC6 in Abinit.
At the beginning the Li states was missing in the band structure. After some readings in the literature I realized that the 2s should be treated as semi-core or I have to use the non-linear core correction to reproduce correctly the band structure of LiC6.
So I downloaded opium and generated the LDA fhi pseudopotential for Li using the non-linear core correction lfc (Louie, Froyen, and Cohen).
The lattice parameter c is always much bigger than the experimental one (of about 6.99 a.u) or the LDA-calculated one (6.98 a.u).
I made many pseudoptentials by varying the core radius from 1.6 to 2.3 a.u and the partial charge radius cut-off from 0.7 to 1.4 but I always have a c lattice parameter of more than 7.60 a.u??
The opium input is a as follow:
[Atom]
Li
4
100 2.00 -
200 1.00 -
210 0.00 -
300 0.00 -

[Pseudo]
3
1.60
1.60
1.60
opt

[Optinfo]
8.0 10
8.0 10
8.0 10
[XC]
lda

[Configs]
2
#
200 0.00 -
210 1.00 -
300 0.00 -
#
200 0.00 -
210 0.00 -
300 1.00 -
[Pcc]
1.4
lfc

and the one for Abinit optimization:

#######################################################################
acell 8.1401508285 8.1401508285 7.0692008242

xred 1/2 1/2 1/2
5/6 1/6 0.0
1/6 5/6 0.0
1/2 1/6 0.0
1/6 1/2 0.0
1/2 5/6 0.0
5/6 1/2 0.0

rprim 1.0 0.0 0.0 # Hex. primitive vectors(to be scaled by acell)
0.5 0.8660254037844386 0.0
0.0 0.0 1.0
#Definition of the atom types
ntypat 2 # There are two types of atoms
znucl 3 6 # The keyword "zatnum" refers to the atomic number of the
# possible type(s) of atom. The pseudopotential(s)
# mentioned in the "files" file must correspond
# to the type(s) of atom.
#Definition of the atoms
natom 7
typat 1 2 2 2 2 2 2
ngkpt 10 10 10
nshiftk 1 # Use one copy of grid only (default)
shiftk 0.0 0.0 0.0 # This choice of origin for the k point grid
# preserves the hexagonal symmetry of the grid,
# which would be broken by the default choice.
nband 20
nbdbuf 2
occopt 3
toldff 1.0E-12
#Definition of the planewave basis set
ecut 50 # Minimal kinetic energy cut-off, in Hartree
#Definition of the SCF procedure
nstep 500 # Maximal number of SCF cycles

Many thanks,
Samir

Hi Samir,

As a preliminary step, did you try to reproduce the lattice of a bulk lithium as for example COD 9008473.cif with your homemade pseudo?

Kind regards,

Alain

Dear Alain,
Thank you so much for your help/suggestions.
Apparently I have some problems with the pseudopotential! The problem is that I don't see what I am missing exactly!
I made a test as you suggested for the Lithium bcc bulk and now the lattice parameter is underestimated not overestimated like in LiC6!.
the starting a is 6.62 a.u and the optimal one is 5.58?? while the LDA parameter I found in the literature is 6.36 a.u.
I think that the pseudo-potential itself and its transferability still far from being optimal!
Thank you again for your help which I appreciate a lot.
Samir

Hi Samir,

Keep in mind when reading the rest that I have zero experience with Opium ...

I assume that your electronic configuration is 1s2 2s1 2p0 3s0 ... why the empty 3s valence state? For my PAW Li pseudo, I have 1s2 2s1 2p0, all three valences. The 1.6 au core radius seems fair. And my PAW pseudo reproduces a bcc Li quite nicely.

Kind regards,

Alain

Hi Alain,
Apparently I have to use the semi-core approximations rather than the non-linear correction. I tried the non-linear correction because it is cheaper than the semi-core approximation. The Ecut is relatively smaller in NLC than in SC.
Especially since I would be interested in big systems (around 50 atoms (Carbon and Lithium)).
Did you use the 1s as semi-core in your PAW pseudopotential?
If so, could you please send me the input for the atompaw pseudopotential generator.
Thanks,
Samir

Hi Samir,

Here we go for atompaw input ...

1s 2s and 2p in valence. Ecut around 20Ha IIRW.

Kind regards,

Alain

Hi Alain,
I am sorry but I took exactly your input and I still have the same problem (the optimal a is 5.87 a.u.). When I run atompaw2abinit I took every thing default!......probably there is some thing special there! or do I have to include charge compensation and/or real space optimization?
The input for abinit is the following:

#The relaxation
optcell 2
ionmov 3
ntime 10
tolmxf 1.0E-15
dilatmx 1.05
ecutsm 0.5
#######################################################################
#Common input variables
#######################################################################
acell 6.6226415090 6.6226415090 6.6226415090
xred 0.0 0.0 0.0
rprim -0.5 0.5 0.5
0.5 -0.5 0.5
0.5 0.5 -0.5
#Definition of the atom types
ntypat 1 # There are two types of atoms
znucl 3 # The keyword "zatnum" refers to the atomic number of the
# possible type(s) of atom. The pseudopotential(s)
# mentioned in the "files" file must correspond
# to the type(s) of atom.
#Definition of the atoms
natom 1
typat 1
ngkpt 10 10 10
nshiftk 1 # Use one copy of grid only (default)
shiftk 0.5 0.5 0.5 # This choice of origin for the k point grid
# preserves the hexagonal symmetry of the grid,
# which would be broken by the default choice.
occopt 3
toldff 1.0E-16
#Definition of the planewave basis set
ecut 30 # Minimal kinetic energy cut-off, in Hartree
pawecutdg 50
#Definition of the SCF procedure
nstep 500 # Maximal number of SCF cycles

Thanks,
Samir

Hi Samir,

I have little time to debug your input. Please try with my test Abinit input file for Li ...

With the LDA PAW pseudo mentioned earlier, it gives a fcc cell size of 8.05 bohr (i.e. 4.26 A) in reasonably good agreement with the experimental value provided in 9008473.cif (i.e. 4.407 A). Have a look on the Etotal vs. celle size plot.

Kind regards,

Alain