# [Pw_forum] Ultrasoft Pseudopotential of Eu

Tram Bui trambui at u.boisestate.edu
Wed Feb 1 00:13:01 CET 2012

```Dear Nicola, Hanghui and Trinh,

I'm sorry for interupting but I've been trying to find someone who
have had experiences with generating an ultrasoft pseudopotential for QE,
and I saw your email back and forth so I just want to give it a shot to ask
you couple of questions. Please accept my apology for any inconvenience.
By the way, I've tried to generate an US PP for Cs and have not
get much luck with it. To be honest I really don't have any background on
PP, so it was really tough for me to be able to understand what each
variable means and how to determine them (even though I have read the PP
generation notes for QE users, and several papers or information online
about PP). So would you spare me some time and give me some guidance on how
to build a meaningful input file for PP generation?
For Cs (Z=55), I want to include the 5s, 5p and 6s state for my
valence and below is the input file I have build so far (it is not running
correctly though). However, there are several things that I'm still unclear
&input
title='Cs',
zed=55.0,
prefix='Cs',
iswitch=3,
rel=1,
dft='PBE',
config='[Kr] 4d10 5s2 5p6 6s1 ',
till this part, I understand that is to run the all-electron (AE)
calculation

rlderiv=4.20, (here I'm not quite sure how one can come up with the
value for rlderiv, in this case I used the radius at most outpeak, 6s, from
the AE wfc plot), so is it a right approach to obtain rlderiv? and what
effect does it have on pseudizing the wfc process? is this the same as the
eminld=-8.0,
emaxld=3.0,
deld=0.01d0,
nld=3, For this value, I assume that I want to include 3 energy state,
5s 5p 6s, so I need to have 3 number of logarithmic derivative. is that
correct?
/
&inputp
lloc=0, here, from the ld1-input.txt, i know the lloc has to be equal
to the last lls # in the card list below (in this case =0), but what does
it mean by the "local chanel"? (what is a local chanel?) if I have lloc=0
here, does it mean that my local chanel is the 5s ?
pseudotype=3,
rho0=0.01,
file_pseudopw='Cs.uspp-tb.UPF'
/
4
6S 1 0 1.00 0.00 4.20 4.50
6S 1 0 0.00 1.00 4.00 4.50
5P 2 1 6.00 0.00 1.60 1.60
5S 1 0 2.00 0.00 1.40 1.40
For the wfc configuration above,
- what does it indicate when one have difference (or the same) rcuts for
norm-conversing and US PP?
- when rcuts of US and NC are difference, why do one need two
energies/channel?
- In the case of when the second row is needed when rcuts are different,
based on what one can determine the value for the "ener" (energy used to
pseudize correspoding state)? I only know what it ener is zero for bound
state, but most of the examples i looked at, everyone was using a nonzero
value when they have another row for the same wfc( e.g. 6s in this case),
does it mean the second row is the unbound state?

I'm really sorry for the long email and the color code i used if that is
bothering you. I'm doing this all alone without any help of any expert
around here. So things are very confusing to me especially when it comes to
quantum physics in which I have very limited understanding. I really
appreciate any input you have for me.

Best Regards,
Tram Bui

On Wed, Apr 27, 2011 at 12:48 PM, Vo, Trinh (388C) <Trinh.Vo at jpl.nasa.gov>wrote:

> Dear Nicola,
>
> I am working on generating pseudo potentials for a few elements of
> Lanthanides too, but so far I am not successful yet (it is still bad).  I
> appreciate if you could share with me your parameter input.
>
> Thank you very much in advance,
>
> Trinh
>
>
> On 4/27/11 10:30 AM, "Nicola Marzari" <nicola.marzari at materials.ox.ac.uk>
> wrote:
>
> > On 4/27/11 5:45 PM, hanghui chen wrote:
> >> Dear PWscf community,
> >>       I am wondering if any one has succeeded in generating the
> >> ultrasoft pseudopotential of Eu, using uspp-736. Now in the literature,
> >> for most of the work that involves Eu, the calculation is done with VASP
> >> where the pseudopotential is provided in its own library.
> >>       Thank you very much indeed.
> >>
> >> Hanghui Chen
> >> Department of Physics,
> >> Yale University
> >
> >
> > Dear Hanghui,
> >
> > we have a test one normconserving, with f electrons in the valence, and
> > an ultrasoft one, with f electrons in the core. This is part of a
> > project developing and testing lanthanides pseudos, to be released later
> > on, but if you want to give them a try, I can send them to you (would
> > be great to have some test feedback).
> >
> > nicola
> >
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
>

--
Tram Bui

M.S. Materials Science & Engineering
trambui at u.boisestate.edu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20120131/1c210c3f/attachment.html>
```