[Pw_forum] GIPAW pseudopotentials

Tue Oct 20 18:29:49 CEST 2009

Hello again,
After using your script Dr. Seitsonen to generate the GIPAW part of the .UPF file it seems that there is always a strange discontinuity in the projections above 11 bohr. However, examining another GIPAW pseudopotential on your website, http://www-int.impmc.upmc.fr/~software/gipaw/, these discontinuities are normal or expected? If that is the case then the pseudopotential generated from the following input into the ld1.x module should work?

&input
	title = 'Mg'
	prefix = 'Mg'
	zed = 12.0
	rel = 0
	nld = 3
	rlderiv = 5.0
	eminld = -6.0
	emaxld = -1.0
	deld = 0.01
	config = '1s2 2s2 2p6 3s2 3p-1 3d-1 4s-1'
	iswitch = 3
	dft = 'PBE'
/
&inputp
	pseudotype = 1
	lloc = 2
	file_pseudopw = 'Mg.pbe-gipaw-oct09-02.UPF'
	file_recon = 'Mg.pbe-gipaw-oct09-02.recon'
	lgipaw_reconstruction = .true.
	tm = .true.
/
2
3S 1 0  2.00  0.00 1.800 1.800
3D 3 2  0.00 -0.30 1.800 1.800

&test
/
2
3S 1 0  2.00  0.00 1.800 1.800
4S 2 0  0.00 -0.40 1.800 1.800

As you can see the reference energy for the unbound states is certainly set as non-zero. However, the pseudopotential generated from this input, when used in a simple lattice, MgO, will produce an error in the scf calculation:
 - two of the Mg atoms overlap
I am sure the crystal parameters used are accurate so it must be a problem with the pseudopotential?

I greatly appreciate any help or assistance anyone might have to offer!

Peter Pallister

-----Original Message-----
From: pw_forum-bounces at pwscf.org [mailto:pw_forum-bounces at pwscf.org] On Behalf Of Ari P Seitsonen
Sent: October 13, 2009 11:31 AM
To: PWSCF Forum
Subject: Re: [Pw_forum] GIPAW pseudopotentials

Dear Peter et al,

   Sorry for the long time it took me to answer!! :(

   The *.wfc files are generated by default, as far as I know.

   There is a special issue concerning the wave functions for the GIPAW 
reconstruction: They are _not_ generated in the normal procedure how the 
ld1.x/atomic code works, but for GIPAW they are calculated _after_ the 
*.wfc files have been written: the standard part of the code calculates 
them once, with some value (possibly zero) as the reference energy; for 
GIPAW one then recalculates them, but with the reference energies taken 
from the '&test' section. This is why the wave functions for the GIPAW 
reconstruction can be different from the ones in the *.wfc files. And this 
is also why one _has to_ define a reference energy different from zero, at 
least for unbound states (ie. whose eigenvalue would be positive) in the 
section '&test'.

   So, in practise it doesn't matter if the wave functions in the *.wfc 
corresponding to empty orbitals do not make sense, but you can check that 
they look ok in the GIPAW section at the end of the UPF file. I attach an 
unofficial script for writing out the GIPAW data in the UPF file (it is a 
very small script file). You can probably modify the output to show the 
properties you want to have.

   If you still have problems with the wave functions coming out from this 
script please check that you use a non-zero reference energy for 
unbound states, otherwise please contact me again.

     Greetings from Zurich,

        apsi

-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-
   Ari P Seitsonen / Ari.P.Seitsonen at iki.fi / http://www.iki.fi/~apsi/
   Physikalisch-Chemisches Institut der Universität Zürich
   Tel: +41 44 63 55 44 97  /  Mobile: +41 79 71 90 935

On Tue, 13 Oct 2009, Pallister, Peter wrote:

> Hi,
> Actually, one of the tutorials on the QE site has a step-by-step
> pseudopotential generation tutorial showing how to generate the AE
> calculation, examine the logarithmic derivatives, etc. This tutorial
> suggests looking at the file (well, in my case) 'Mg-oct09ps.wfc' to look
> at the pseudo-generated orbitals. Is this incorrect? What do the
> individual file/s mean in the "file_wfcncgen=" input? Thanks!
>
> Peter Pallister
>
> -----Original Message-----
> From: pw_forum-bounces at pwscf.org [mailto:pw_forum-bounces at pwscf.org] On
> Behalf Of Lorenzo Paulatto
> Sent: October 9, 2009 4:20 AM
> To: PWSCF Forum
> Subject: Re: [Pw_forum] GIPAW pseudopotentials
>
>
> On Wed, October 7, 2009 17:10, Pallister, Peter wrote:
>>  Hi all,
>>
>>  I am attempting to generate a gipaw pseudopotentials, in this case
> for
>>  Mg, but I am having tremendous difficulty. The wavefunctions produced
>>  with the input below (or minute variations thereof) continuously show
>>  discontinuities, specifically at a radius of 15bohrs.
>
> Dear Peter,
> the wavefunctions are generated on a logarithmic radial grid, around 15
> bohr the points are very coarse, btw how are you examining the
> wavfunctions? (there is not file_wfcncgen='..' in your input).
>
> best regards
>
>
> -- 
> Lorenzo Paulatto
> SISSA  &  DEMOCRITOS (Trieste)
> phone: +39 040 3787 511
> skype: paulatz
> www:   http://people.sissa.it/~paulatto/
>
>
>
> ----------------------------------------------------------------
>  SISSA Webmail https://webmail.sissa.it/
>  Powered by SquirrelMail http://www.squirrelmail.org/
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
>