[Pw_forum] Could anyone share a stable pseudopotential of lithium?

[Axel Kohlmeyer]

On Mon, Jan 21, 2013 at 11:40 AM, Johannes Moeller
<j.moeller1 at physics.ox.ac.uk> wrote:
> Dear Miao,
>
> I've used Li.pbe-s-van_ak.UPF to calculate phonon frequencies in LiF with
> good results. If you know approximately what you're looking for in your
> material it might still be worth a try.

as the person who initially created some of those pseudopotentials
(many, many moons ago), i should perhaps add a few comments as well.

creating a very transferable pseudopotential for lithium seems
almost impossible. and even the existing ones have to be used
with great care. if you want to run calculations without worries,
you may be better off using some all-electron method.

when creating a psp for Li you have two choices:
you can create a "conventional" psp (keep 1s in core),
but those really only make sense for metallic Li and
then you usually need NLCC as well.

if you look for (ionic) Li compounds, you frequently need
to use a "semi-core" psp (i.e. *no* electrons in the core
just a little smearing out of the core charge as in hydrogen psps),
but for those you *have* to make sure you use an adequate
(i.e. quite high) density cutoff. in addition it took quite a bit
of tweaking to get a decent set of parameters for the "ak"
set of Li psps. for many test cases, there were ghost states.

those potentials *have* been used successfully for multiple
studies, but on-and-off people have been reporting problems
that seemed impossible to resolve through tweaking the
pseudization parameters.

my conclusion from this is experience is, that for Li the
plane-wave psp approach is reaching its limitations easily,
so getting proper and consistent results includes some
amount of chance. if you are not comfortable with this,
you will have to look for some other approach or take the
challenge and demonstrate that it *is* possible to generate
a good Li psp. perhaps using some other pseudization
scheme, that hasn't been tried yet.

ciao,
    axel.

>
> Best wishes,
> Johannes
>
> -------------------------------
> Johannes Moeller
> Department of Physics
> University of Oxford
> Parks Road
> Oxford OX1 3PU, UK
>
> From: 高淼 <miaogao at ruc.edu.cn>
> Date: 21 January 2013 06:37:22 GMT
> To: <pw_forum at pwscf.org>
> Subject: [Pw_forum] Could anyone share a stable pseudopotential of lithium?
> Reply-To: PWSCF Forum <pw_forum at pwscf.org>
>
>
> Dear all:
>
> I'm calculating phonon properties of a hypothetical material, containing
> lithium. In the espresso_pp package, there are many pseudopotentials of
> lithium, but many of them with wrong results for force and phonon
> calculation.
> For example, in Li.pbe-s-van_ak.info, one can find the following sentences.
> "BEWARE: Matteo Calandra reported seriously wrong results with this
> potential
> (notably in forces and phonon frequencies)"
>
> I have tested phonon properities of these pseudopotentials for bcc-lithium.
> For Li.pbe-n-van.UPF, Li.pbe-s-van_ak.UPF, Li.pw91-n-van.UPF,
> Li.pw91-s-van_ak.UPF,
> and Li.pz-s-van_ak.UPF, all of them have phonon problems. For
> Li.pbe-s-mt.UPF,
> Li.pz-s-mt.UPF, they gives wrong lattice constants of bcc-lithium.
>
> I know that one can generate pesudopotentials by himself, but it's a very
> challenging task to get a robust pesudopotential. Could anyone share a
> stable
> pseudopotential of lithium, which have correct result of phonon?
>
> I am greatly indebted to you for your help. Thanks again!
>
> Best wishes
>
> Miao Gao
> ------------------
> Ph.D. student of Department of Physics,
> Renmin University of China.
> ------------------
>
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum



--
Dr. Axel Kohlmeyer  akohlmey at gmail.com  http://goo.gl/1wk0
International Centre for Theoretical Physics, Trieste. Italy.