[CPMD-list] About getting total energy

[Ari P Seitsonen]

Dear Hong,

  Usually the "correct" procedure is to run geometry optimisation to get
the "total energy" of the system: When the geometry is converged, the
energy printed out in the output is the final total energy. If you have
previously optimised the geometry already (with exactly the same
calculational parameters!), then a single wave function optimisation would
be enough.

  The "total" energy of atoms is often trivial to calculate: One needs the
keyword "LSD" in the section "&CPMD" and sets the correct multiplicity in
the section "&SYSTEM". However for most materials where the electronic
shells cannot be made full there are problems because the lowest-energy
solution of the Kohn-Sham equations is not spherically symmetric (KS is a
single-determinant method and cannot describe these cases correctly).
Oxygen is a famous example, and it has been discussed in the literature
(Painter & Averill, PRB 1982, http://dx.doi.org/10.1103/PhysRevB.26.1781/).

  Hopefully this helps.

    Greetings,

       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

On Sat, 19 Feb 2005, Hong Won Keon wrote:

> Hi, 
> 
> I'm going to get total energy of benzene(C6H6) molecule and C & H atom
> respectively.
> 
> So I tried wavefuction optimization with two pseudopoetentials - MT and
> USPP.
> 
> I wonder if it is right to get total energy through wavefuction
> optimization. Is it geometry optimization? 
> 
> And actually I'd like to get atom's total energy. But I couldn't find
> out atomic energy calculation example in tutorial.
> 
> So I want to listen some information on this matter.
> 
>  
> 
> Anyway I used simple cubic cell put atom, and run wavefuction
> optimization.  
> 
> For benzene, I put it in hexagonal cell, and run wavefuction
> optimization.
> 
> I also used LANCZOS diagonalization for MT, and PCG and ODIIS for USPP.
> I found Davidson and Lanczos don't work for USPP.
> 
>  
> 
> Thanks in advance.
> 
>  
> 
> Hong wk.
> 
>  
> 
>