The ABINIT Users Mailing List ( CLOSED )

[MANUEL PEREZ JIGATO <mpj@fyslab.hut.fi>]

hi,

your point about the functionals is probably a very fundamental one;
i think that not much literature exists about that point:  combinations 
of core and valence Vxc and different kernels (fxc) for tddft with 
pseudopotentials.
i only know of one paper, and that is by Rubio and co-workers (jcp, last 
year i think), not with planewaves but with real space methods, and with the 
time representation (not linear response, rather the full response).
 
I suppose whenever you try any of these functional Vxc and 
fxc combinations, if you think something nasty could 
happen, you could compute excitations of molecules (not periodic systems)  
with one of the many quantum chemical packages available, most of which 
have implemented also the linear response TDDFT, and they are ALL-ELECTRON
(however i am not sure you could find OEP Vxc-fxc), i mean just for 
comparison (usually they use gaussian or slater bases, not planewaves).

in case you are intereseted in the optimised effective potential method
(or exact exchange) there are two codes for generation of OPM pseudos
(as far as i know), one by Engel (also a paper in PRB), and the other one 
by a group in Munich (Seidl, Gorling, etc), they have also a PRB on their 
method of generating OPM pseudos. Probably many people would say that for 
OPM excitations, it is fine to use Hartree-Fock pseudopotentials 
(Kleynman has at least one paper on Hartree-Fock pseudoptentials), 
they say that you get exactly the same thing as Hartree-Fock 
(see also a recent paper by Picket on response in screened-exchange LDA).

sorry i do not have the references at hand, but i could dig them out 
if you were interested.

manolo




PS apparently when you calculate tddft excitations on molecules on 
supercells (molecule in a box), some people have found that the 
interaction between molecules from image cells do modify the excitation 
spectrum as compared to the free molecule, but this is a different story.
(i do not know whether the ABINIT implementation includes any form of
decoupling those).



On Tue, 25 Feb 2003 luppi.eleonora@unimore.it wrote:

> Dear All,
>  I'm trying to make some calculations with TDDFT.
>  I'd like to have some information about TDDFT implementation in ABINIT. 
>  I have read that I can use: ixc 22=>  the Burke-Petersilka-Gross 
>  hybrid xc-kernel, for TDDFT tests.
>  Is it for tests only or for production too?
>  My problem is that I have performed some tests and I have obteined 
> incoherent results,
>  in particular wrong for ixc 20, 21,22, with respect to the case with ixc 7 
> or 1.
>  (as in test :Test_v3/t55.in). 
>  Another question is: have I to generate ad hoc pseudopotentials to use 
> this special xc-kernels 
>  or can I use simple LDA (or GGA-PBE) pseudopotentials.
> 
>  Any Help will be appreciated.
> 
>   Thank you 
> 
>    Best Regards Leo
> 

-- 
###########################################################################
 Manuel Perez Jigato
 Helsinki University of Technology
 Laboratory of Physics
 Room U203
 Otakaari 1M
 Otaniemi, Espoo
 P.O. Box 1100
 FIN-02015 HUT
 Finland

 Tel: +358-(0)9-451 3107
 Fax: +358-(0)9-451 3116
 E-mail: mpj@fyslab.hut.fi
###########################################################################
Quiere entrar alegre en Sevilla, donde un dia fue feliz, donde late el 
corazon del mundo (J. Eslava Galan)
###########################################################################