The ABINIT Users Mailing List ( CLOSED )

[Martin Fuchs <fuchs@fhi-berlin.mpg.de>]

Dear Murat,

first a general remark: recall the two chief approximations 
that underly *DFT* calculations in a *pseudopotential* approach
i) use of an approximate exchange-correlation functional,
in your case LDA, I suppose
ii) use of pseudopotentials for the electron-ion
interaction (which implies frozen ion cores, linearized
core-valence exchange-correlation, and pseudoization of
wavefunctions - all approximations)

Both approximations incur systematic errors in the 
results compared experimental data, both errors are
different in nature and need to be distinguished.
This always needs to be checked carefully - while
the errors are often noticeable they are also often
acceptable (and accepted). 

ad i) compare with a full-potential calculation (i.e.
without any pseudopotential related error). For
instance S Kurth et al. Int J Quantum Chem (1998/9)
"Quantum Chemical and Solid State tests of DFT
approximations ..." - within the LDA they obtained
the bulk moldulus of NaCl of 32.2 GPa at a crystal 
volume of 276.1 bohr^3. 
These are the values a perfect pseudopotential 
calculation within the LDA should reproduce! 
And this is what the LDA gives. If you can't live
with it, try a "better" functional. Sometimes GGA 
functionals can improve over the LDA ...

ad ii) compare with previous pseudopotential results,
for instance M Fuchs et al PRB (1999) "Pseudopotential
study of binding properties of solids within generalized-
gradient approximations ..." - within the LDA they
obtained the bulk modulus of Nacl of 32 GPa at a
lattice constant of 5.19 Angstrom for (Troullier-Martins)
pseudopotentials without nonlinear core corrections,
and 32 GPa at 5.43 GPa with nonlinear core corrections.

Your value for the bulk modulus is 10% larger than the
full potential result and disagrees with previous
pseudopotential data. Your pw cutoff and kpoint no.
look suffiecient to me. So possibly, your pseudopotentials
are not good enough. Have you tried a pseudopotential
for Na that includes nonlinear core corrections? How
did you generate your pseudopotentials (what are it's
cutoff radii etc.)?  

I also suggest to check a recent paper of Simunek et al
on FP-LAPW derived pseudopotentials, they presented data
for NaCl too.   

Best regards,
Martin

durandur@engin.umich.edu wrote:
> 
> Dear Abinit users,
> 
> I am a new user. I am trying to calculate bulk modulus of NaCl.
> I used the Troullier-Martins pseudopotentials, a cutoff of
> from 20 to 75 Ha and a grid of from 8x8x8 to 15x15x15 kpoints.
> The calculations, regardless of used the cutoff and grid, give
> the same value of 35 GPa, which is larger than experimental value
> of 26 GPa.
> 
> I will be glad if you can help me.
> 
> Thanks,
> 
> Murat Durandurdu
> 
> Here one of the inputfiles is:
> 
> # NaCl  optimization
> #
> 
> #Definition of the unit cell
> acell 5.264 5.264 5.264 angstrom  # This is equivalent to  1 1 1
> angdeg  90 90 90
> 
> #Definition of the atom types
> ntype 2           # There are three types of atom
> znucl 11  17       # The keyword "znucl" refers to the atomic number
> 
> pspcod=1
> #Definition of the atoms
> natom 8           # There are two atoms
> type 1 1 1 1 2 2 2 2
> 
> xangst              # This keyword indicate that the location of the atoms
> 2.632 2.632 0
> 2.632 0 2.632
> 0 2.632 2.632
> 0 0 0
> 0 0 2.632
> 0 2.632 0
> 2.632 0 0
> 2.632 2.632 2.632
> #
> #Definition of the k-point grids
>                   # into account the symmetry
> kptopt   1              # K-points will be provided
> ngkpt 8 8 8
> 
> #Definition of the planewave basis set
> ecut  50.0         # Maximal kinetic energy cut-off, in Hartree
> 
> #Definition of the SCF procedure
> nstep 100          # Maximal number of SCF cycles
> toldfe 1.0d-6     # Will stop when, twice in a row, the difference
>                   # between two consecutive evaluations of total energy
>                   # differ by less than toldfe (in Hartree)
> 
> enunit 1
> ntime  1000
> ecotsm=0.5
> ionmov 3

-- 
Martin Fuchs   
Fritz-Haber-Institut der MPG
Faradayweg 4-6
14195 Berlin
Germany
Phone: +30 8413 4802
Fax:     +30 8413 4701
E-Mail: fuchs@fhi-berlin.mpg.de