The ABINIT Users Mailing List ( CLOSED )

[Marc Torrent <marc.torrent@cea.fr>]

Is there a reason why you absolutely want to include f- partial waves in 
your PAW dataset while you do not have any f- valence electrons ?
This is not a typical way to proceed when generating atomic data... 
except if you have particular thermodynamical conditions...

Marc Torrent


Pio Baettig a écrit :
> Hi,
>
> thank you very much for your information. Using your hints, I managed to
> get good fits for the logarithmic derivatives.
>
> Unfortunately the wavefunctions still give me some small problems.
>
> In wfn 1,3,5 and 7, the wavefunctions psi and {psi}~ are quite different
> in magnitude from the projector {p}~. wfn5 looks especially bad, {p} has
> a maxiumum of about 50 and the psi are <1, compare wfn5_8.pdf.
>
> I tried to generate PAW for the neighbouring atoms Pb and Po to see
> whether this behaviour was specific to Bi, but I do not seem to get a
> good fit for wfn5 (and others) for these two elements neither.
> For Bi, I tried to move the reference energies (for 4 and 7 Ry) to 3 and
> 8 Ry as proposed in the usersguide, but the magnitude of the projector
> did not decrease. (The logarithmic derivatives for l=2 remained correct)
> The wavefunctions (see wfn_compare.pdf) seem no longer dependent on the 
> number of gridpoints, a step ahead.
>
> I would be very grateful for any hints on how to make good PAWs for 
> these elements.
>
> Best regards and thank you very much again for your help,
>
> Pio Baettig
>
> P.S. The Pb-scf doesn't seem to converge in less than 1000 iterations, 
> but the logderivs look ok.
>
> attachments:
> Inputs for Pb, Bi, Po
> Analysis for Bi, logderivs and wfn for 1000 gridpoints, wfn-comparison 
> for 750, 1000 and 2000 gridpoints for wfn1
>
>
>
> Marc Torrent wrote:
> > Hi,
> >
> > When you have a materials with electrons up to l=lmax, it's 
> > recommended to generate the pseudopotential with l=lmax+1. So, you 
> > should put "4 0." on the corresponding line in input file. After that, 
> > suppress your third partial wave for s and increase the value of the 
> > remaining reference energy... It should correct your problem.
> >
> >
> > Concerning "hard projectors", you can plot the wfni files and look at
> >  the magnitude of wave functions and projectors; they should have 
> > values around 1...5...10...  ut not 100 or 500...
> >
> >
> > Marc Torrent