ecp

[asaf]

Hi,
I want to use ecp for ri-dft calculation, and my questions are:
1)Are the pseudo potentials in Turbomole (e.g "li" ecp-2-sdf) are legal for this purpose?
2)Are the pseudo potentials in TM are functional dependent or good for all DFT functionals (namely gga)? 

[Arnim]

Hi!

0) There is no reason to use an ECP for Li.
1) There is no optimized basis set for ecp-2-sdf available, only an unpublished guess.
2) ECPs have not really something to do with functionals, but rather with basis sets.

All in all: Don't do that

Arnim

[asaf]

It is good to know, but if I want to model a surface which is composed from Li an F atoms by the PEEC method how should I treat the Li atoms outside the QM region ? which ECP is suitable for this purpose ?
Thanks,

[Arnim]

Ok! In this case the basis set doesn't matter, because you just use the ECPs without basis functions in order to model the cationic site.

[asaf]

Just a technical question:
As  I have understood, in the PEEC the cationic sites are described by pure ECPs (without electrons), but when assigning the EPC ("lii" ecp-2-sdf) the li atom remains with one valance electron (2s electron).So how can I described the whole Li by ECP?
Thanks!