[Pw_forum] c60

[Lorenzo Paulatto]

In data 19 ottobre 2009 alle ore 08:53:12, Gabriele Sclauzero  
<sclauzer at sissa.it> ha scritto:
> That's right. Anyway with the new PAW implementation you should be able  
> to reconstruct the all-electron total energy (in the frozen core  
> approximation) from the pseudo-wavefunctions . I think this has already  
> been implemented, you simply need to use a PAW pseudopotential in place  
> of the ultrasoft or normconserving one that you've been using.
> For instance in espresso-4.1.1/pseudo there is one: C.lda-paw_kj.UPF.

You indeed can, i.e. if you only use PAW datasets in a calculation the  
code will spit out the "total all-electron energy", aside with the total  
valence energy. Nevertheless, I would like to stress that total energy has  
no meaning for isolated systems, i.e. the zero of energy is arbitrary and  
only energy differences matter.

What you should do is compute the binding energy:
    E_bind = E_c60 - 60 x E_c
where E_c comes from a calculation of an isolated C atom with the same  
cutoffs as the C60. This not only give a meaningful binding energy, but  
can also reduce possible truncation errors thanks to error cancellation.  
Computing the ground state of a single C atom is a bit tricky, but is a  
matter that have been discussed countless times in this mailing list:  
<http://www.google.cn/search?q=isolated+atom+pw_forum+site%3Ademocritos.it>

best regards

-- 
Lorenzo Paulatto
SISSA  &  DEMOCRITOS (Trieste)
phone: +39 040 3787 511
skype: paulatz
www:   http://people.sissa.it/~paulatto/

     *** save italian brains ***
  http://saveitalianbrains.wordpress.com/