[Pw_forum] Li pseudopotential

[Lorenzo Paulatto]

Replying to message "Re: [Pw_forum] Li pseudopotential" from Elena Redondo
(16/11/09):
> I think that it is because the 2s orbital has a node before rc=1.00.
> But, Yao et al say they used a cutoff radii of 1.0 for the 2s orbital.

Dear Elena,
you are using the Toullier-Martin recipe to pseudize the atomic
wavefunctions (tm=.true.). This method always generates node-less
wavefunctions hence it cannot be used to generate norm-conserving
pseudopotentials that include two orbitals with the same angular momentum
(1s and 2s, in your case). In theory it should be possible to generate
this kind of pseudo using some other pseudization scheme, but ld1 does not
do it, or at least it thinks it does not ;-).

An exception would be possible if rcut was "inside" the node of the 2s
wavefunction, but this is not your case, as the node is at ~0.84 bohr.
Indeed, if you put the rcut of 2s orbital to 0.80 you get a good (but very
hard) pseudopotential (you also have to remove of the tm=.true. switch,
otherwise it will have a ghost state).

Considering what Paolo Giannozzi said, I think you could try some other
pseudopotential generatio software (such as Opium) for this specific
purpose.

cheers

-- 
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/


----------------------------------------------------------------
  SISSA Webmail https://webmail.sissa.it/
  Powered by SquirrelMail http://www.squirrelmail.org/