[CPMD-list] XC in the pseudopotential file

[Ari.P.Seitsonen at iki.fi]

Dear Dr Siupang,

> 	I want to generate the Troullier-Martins type pseudopotential by
> using the fhi98PP package. Then I use the scilab-script from Walter Langel
> (I get it from the mailing list at 2002-02) to convert its output file
> ncpp.cpi to the CPMD format. However, it don't have the &ATOM and &INFO
> section.
> 
> 	In &ATOM section, we should have 4 entries, 
> 
>  Z  =   13
>  ZV =    3
>  XC = 1100        .666667
>  TYPE = NORMCONSERVING BHS
> 
> 	The first one (Z) should be the atomic number, the second one (ZV)
> should be the number of valance electron. Now, I want to ask how to set
> the third and the 4 entry. 
> 
> 	From the is Goedecker's pseudopotential generation package (in the
> file pseudo.f), I know that
> 
> LDA  should be 1100
> PADE should be 0900 
> BONL should be 1110
> BP   should be 1111
> PW   should be 1134
> BLYP should be 1312 
> 
> 	However, in the fhi98PP, we have the following
> exchange-correlation function:
>     1  LDA   Wigner
>     2  LDA   Hedin/Lundqvist
>     3  LDA   Ceperley/Alder Perdew/Zunger (1980)           
>     4  GGA   Perdew/Wang (1991)
>     5  GGA   Becke (1988) X, Perdew (1986) C
>     6  GGA   Perdew/Burke/Ernzerhof (1996)
>     7  LDA   like 8 + MacDonald/Vosko relativistic correction
>     8  LDA   Ceperley/Alder Perdew/Wang (1991)
>     9  GGA   Becke (1988) X, Lee/Yang/Parr (1988) C
>    10  GGA   Perdew/Wang (1991) X, Lee/Yang/Parr (1988) C
> 
> 	How can I relate the one in phi98PP to CPMD? Also, how to set the
> second value (in this example, it is to 0.666667)? 
> 
> 	Finally, how to set the the fourth entry TYPE = xxxx. 

I'll try to give an answer... Please others correct me if I'm
mistaken.

  The second value on the line 'XC = ' should be 2/3, it's the
exponent in the exchange; to be changed only if trying out x\alpha
method.

  'TYPE = NORMCONSERVING BHS' always, the FHI98pp cannot generate
other types of pseudo potentials.

  The correspondence between the different XC functionals should
approximately be as follows:

      FHI98pp   CPMD
PW91     4      1123
BP86     5      1111
PBE      6   ca 0934
LDA CA   3      1100
LDA PW   8   ca 0900
BLYP     9      1312

Actually they are somewhat incorrect in the version of 'psgen' which I
sent to the list. In any case, these are ONLY used to check the
consistency of the XC functional in the pseudo potential file and the
input, if they are inconsistent a warning is given, however the
functional given in the input for CPMD is used in the calculation, the
value in the pseudo file is only for the check, i.e. it is not used
in CPMD.

    Greetings from Zurigo,

       apsi

-- 
-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-
 Ari Paavo Seitsonen / Ari.P.Seitsonen at iki.fi / http://www.iki.fi/~apsi/
 Tel +41 1 635 44 97 / Fax +41 1 635 68 38 / GSM +41 79 719 09 35      
 Anschrift: Physikalisch Chemisches Institut (PCI), Universität Zürich (UniZh)
 Indirizzo: Winterthurerstraße 190, CH-8057 Zürich
 Address:   Schweiz / Svizzera / Suisse / Svizra / Switzerland