Dear Yunpeng Wang,
As far as I know, Siesta use valence charge, it couldn't reproduce
all-electron charge, so the vdw-DF in siesta should use valence
charge. Maybe PAW scheme could reproduce all-electron
in vdw-DF, but this is unavailable in Siesta. Maybe you could ask
PWSCF team. They use paw scheme, but I don't know whether they use
all-electron density or not.
Best Wishes.
On 4/28/11, Yunpeng Wang <[email protected]> wrote:
> Hi all:
>
> I am interesting with vdw-DF and its implementation in siesta. I have a
> simple question:
>
> As mentioned by M. Dion, the nonlocal correlation energy contribution is:
>
> E = 1/2 int int dr dr' n(r) n(r') phi(r,r')
>
> where n(r) is charge density. Is n(r) charge density of valence charge
> density or all-electron charge density?
>
> thanks!
>
> Yun-Peng
>
> --
> *Yunpeng Peter Wang*
>
