[Pw_forum] Re : Re:Pseudopotential for Li

[Lorenzo Paulatto]

In data 27 agosto 2009 alle ore 10:33:06, Bertrand SITAMTZE  
<siyouber at yahoo.fr> ha scritto:
> I wanted to use the Li.pw91-s-van_ak.UPF. But, as we all know, this has  
> to be tested.

Normally pps from the quantum-espresso page have already been tested, and  
often used in some published work. Yet, doing some extra testing is always  
a good idea.

> That is why I decided to generate a pseudopotential by myself. Do you  
> know about this regeneration process?

A good point to start to generate a new pp is an old one. At the begging  
of the UPF file there is a short human-readable section where you can find  
most of the parameters used to generate the file. The reference energy for  
the second projector are actually missing, but you can usually find them  
easily.

> I used the default value

 From the manual:
   If not specified, the matching radius is determined
   by the condition:  rho_core(rcore) = 2*rho_valence(rcore)

in you specific case it is probably too small.

Furthermore, the pp you are trying to imitate includes the 1s electrons in  
valence: it is practically an all-electron pseudopotential. Hence, it has  
no core charge. Generating a pseudopotential for Lithium with only the 2s  
electron in valence, using core-charge correction, can be quite difficult;  
furthermore the result won't probably be very accurate. Nevertheless, it  
will also be much softer (it requires a much lower cutoff).

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/