The ABINIT Users Mailing List ( CLOSED )

[mmikami@rc.m-kagaku.co.jp]

Hello,

This is the effect of the pseudopotentials, which treat only the 
valence electrons. The difference among the pseudopotentials may come 
from the different style of pseudization of the valence wavefunctions 
(i.e. how to produce nodeless valence wavefunctions for soft/hard   
pseudopotentials).
Kindly note that the database of NIST are produced from all-electron 
calculations; the similarity of the results for the hydrogen may be 
understood, because the hydrogen atom has only one valence electron (no 
core electrons).

But the comparison among "relative energies" may be generally OK (if 
one perform good calculations for easy systems), for example, the 
estimaton of cohesieve energies etc.

You might want to refer to several documents about pseudopotentials, 
for example,
-   M. Fuchs, M. and M. Scheffler, "Ab initio pseudopotentials for 
electronic  structure calculations of poly-atomic systems using 
density-functional theory,"  Comput. Phys. Commun. 119, 67-98 (1999), 
and references therein.

For general references about the Pseudopotential+Planewave method, e.g.
-   M. C. Payne, M. P. Teter, D. C. Allan, T. A. Arias, and J. D. 
Joannopoulos,  Rev. Mod. Phys. 64, 1045-1097 (1992), and references 
therein.
- G. P. Srivastava, "Theoretical Modelling of Semiconductor 
Surfaces"(World  Scientific, 1999), esp. Chap. 3
- (Any other suggestions ? I am thinking of updating FAQ ... )
- In passing, the following book may be expected for a guide book: 
Richard M. Martin , Electronic Structure : Basic Theory and Practical 
Methods (ISBN: 0521782856) Cambridge University Press  (planned in Dec 
2003)

Regards,
Masayoshi

On 2003.10.8, at 06:05 Asia/Tokyo, hongki@physics.utexas.edu wrote:

> Hello ABINIT users,
> I tried to test a really simple case such as a single atom in a big 
> box but it
> seems that the results of energy terms are not consistent except 
> Hydrogen.
> For example, in case of Hydrogen(H),energy terms are approximately
> etotal  -0.46~-0.47 (Ha)
> kinetic  0.40
> xc      -0.23
> for pspgth,pspnc and hgh pseudopotentials and those values agree with 
> the other
> calculations such as LSD from the database of NIST
> (http://physics.nist.gov/PhysRefData/DFTdata/Tables/ptable.html)
> etotal   -0.478671 (Ha)
> kinetic   0.466643
> xc       -0.278072
> However, in case of Lithium(Li)
> 1. pspgth
> etotal   -6.4459953871E+00
> kinetic   4.64052708224207E+00
> xc       -1.39621419943090E+00
> 2. pspnc
> etotal   -2.4918288310E-01
> kinetic   4.92291950692443E-02
> xc       -1.31155787732314E-01
> 3. hgh
> etotal   -2.1898809538E-01
> kinetic   5.53895935658446E-02
> xc       -9.56485913178435E-02
> 4. LSD from Nist
> etotal   -7.343957
> kinetic   7.249892
> xc       -1.665189
> Results of other atoms show similar inconsistency.
> I used the following options but I think even proper choice of options 
> for each
> atom may not correct the huge difference of the results.
>
> nspden 2
> nsppol 2
> nspinor 1
> occopt 7
> acell 3*15
> ecut 15.0
>
> Thank you.
> Hongki Min