The ABINIT Users Mailing List ( CLOSED )

["zhang yongbin" <yongbinzhang@sina.com>]

----- Original Message ----- 
From: "Samantha Jenkins" <Samantha.Jenkins@htu.se>
To: "Petr Casek" <casek@physics.muni.cz>; <forum@abinit.org>
Sent: Friday, December 13, 2002 5:14 PM
Subject: Re: [abinit-forum] AIM and psuedopotentials


> Dear Petr
> 
> 
>  > fhi98PP is the package for generating norm-conserving
>  > pseudopotentials (see abinit page for link, it is well documented).
> 
> OK thankyou.
> 
>  > For the question of point properties - we have performed some study
>  > of these quantities (laplacian, rho, eigenvectors at BCP) in oxides.
>  > The numeric precision is rather good (even in laplacian) and the
>  > results seem to be comparable to the results based on the localised
>  > basis algorithms found in literature (at least for oxides). However,
>  > so far our interest was concetrated mainly to the integrated
>  > quantities.
> 
> That does sound promising, I am interested in nanotubes, so to summarise 
> then I use fhi98PP to generate the carbon psuedopotential, presuming the 
> bonds at the end of the tube need passivating, then I can get the point
> properties from the charge density map that I get out of abinit.
> 
> 
> Best wishes
> 
>   Samantha
>  >
>  >> Dear Petr
>  >>
>  >>
>  >>> I don't know if I have explained well the situation. In our
>  >>> analysis we use the valence electron density (calculated by
>  >>> Abinit) AND the core electron densities (for each type of atom)
>  >>> generated by the fhi98PP (i.e. the core density corresponding to
>  >>> the pseudopotentials used in the calcul=> *.fc). All these files
>  >>> give the density in the real space (valence density =>
>  >>> equidistant grid in elementary cellule, core density => radial
>  >>> logarithmic grid) and whole our analysis is then based on 3D
>  >>> interpolation in real space!!!
>  >>>
>  >>
>  >>
>  >> Ok, I did not know that it was OK to do this adding back in of core
>  >> charge.
>  >>
>  >>
>  >>> All what you need is the output density file *_DEN of abinit
>  >>> (valence) and the files *.fc (for all elements), generated
>  >>> together with your pseudopotentials by fhi98PP. The precision of
>  >>> the charge determined in this way is usually something about
>  >>> 0.003 electrons (even at the surfaces).
>  >>>
>  >>>
>  >>
>  >> What is fhi98PP that sounds like a DFT code ? How good is abinit
>  >> for properties at the BCP?, sounds very good for the integration
>  >> quantities but I am more  interested in the point properties at the
>  >> BCP.
>  >>
>  >> Best wishes
>  >>
>  >> Samantha J.
>  >>
>  >>
>  >>> Best wishes Petr Casek
>  >>>
>  >>>
>  >>>
>  >>>
>  >>>
>  >>>> Dear Petr,
>  >>>>
>  >>>> Thankyou for your prompt reply, I can see why you don't use
>  >>>> psuodopotentials with AIM since many people want to calculate
>  >>>> Bader charges etc.
>  >>>>
>  >>>> Can you calculate G and V at the BCP as well, (using localised
>  >>>> basis sets of course). So I will have to use localised basis
>  >>>> sets, I take it their is a library of these avaliable, are they
>  >>>> optimised for different envíroments. Is the format of these
>  >>>> basis sets similar to any other standard forms.
>  >>>>
>  >>>> Thanks again,
>  >>>>
>  >>>> Best wishes
>  >>>>
>  >>>> Samantha J.
>  >>>>
>  >>>>
>  >>>> Bader analysis is already implemented in 3.4 branche of Abinit
>  >>>> (see Release notes) as the postprocessing tool (aim). The short
>  >>>> help for the latter is in the Infos subdirectory (aimhelp.*).
>  >>>> With this tool it is possible to obtain all basic quantities of
>  >>>> Bader theory, i.e. Bader charge, critical points (with
>  >>>> eigenvalues and eigenvectors of Hessian), the volume of Bader
>  >>>> atom and of course Bader surface. So far we have not touched
>  >>>> the problem with the using of pseudopotentials in this analysis
>  >>>> (but you can always see in the output separate contributions of
>  >>>> core and valence electrons to the Bader charge and so estimate
>  >>>> possible problems)
>  >>>>
>  >>>> Best wishes Petr Casek
>  >>>>
>  >>>>
>  >>>>
>  >>>>
>  >>>>
>  >>>>> Dear abinit users,
>  >>>>>
>  >>>>> I am new to abinit, can you tell me if abinit can be used to
>  >>>>> obtain AIM (atoms in molecules, Bader analysis) using
>  >>>>> psuedopotentials? I am looking for these quantities at the
>  >>>>> BCP (bond critical point) only? Such and rho and laplacian
>  >>>>> (rho) and the eigenvectors. I know that psuedopotentails and
>  >>>>> AIM are in principle OK as long as the psuopotential is not
>  >>>>> too hard core.
>  >>>>>
>  >>>>>
>  >>>>> Thanks for any advice,
>  >>>>>
>  >>>>> Best wishes
>  >>>>>
>  >>>>> Samantha J
>  >>>>>
>  >>>>>
>  >>>>>
>  >>>>
>  >>>> Dr. Samantha Jenkins                 H       Ph. +46 (0)520
>  >>>> 475318 Dept. of Informatics & Mathematics    \      Fax +46
>  >>>> (0)520 475399 University of Trollhattan/Uddevalla    O:
>  >>>> Samantha.Jenkins@htu.se P.O. Box 957, Trollhattan             /
>  >>>> S.Jenkins@physics.org SE 461 29 Sweden                     H
>  >>>>
>  >>>>
>  >>>>
>  >>>>
>  >>>
>  >>
>  >>
>  >>
> 
> 
> -- 
> Dr. Samantha Jenkins                 H       Ph. +46 (0)520 475318
> Dept. of Informatics & Mathematics    \      Fax +46 (0)520 475399
> University of Trollhattan/Uddevalla    O:    Samantha.Jenkins@htu.se
> P.O. Box 957, Trollhattan             /      S.Jenkins@physics.org
> SE 461 29 Sweden                     H
> 
> 
>