[Pw_forum] Ge.pz-hgh.UPF

Paolo Giannozzi p.giannozzi at gmail.com
Wed Oct 26 10:22:28 CEST 2016 

Ge.pz-hgh.UPF was, like all HGH pseudopotentials available on the QE site,
converted from HGH pseudopotentials in the CPMD format, available on
Mattias Krack's web site (there is a link on the QE site which however is
no longer valid). Information on pseudo-orbitals  and pseudo-charge
density, useful to start the calculation and needed to compute PDOS and the
like, was separately computed using the "atomic" code of QE and included
into UPF files with a semi-automatic procedure, whose correctness is likely
but far from guaranteed. Not all UPF files contain those pieces of
information, since many HGH PPs gave problems with the atomic code, for
unclear (to me) reasons. I don't remember why lmax in UPF equals lmax-1 in
CPMD format. In the UPF format, lmax=2 means s,p,d channels. I guess, and
you may easily verify, that lmax=3 in CPMD format means the same: three
channels, s,p,d.

Cutoff radii should be in a.u.. I don't remember what they are supposed to
mean in your case.

Paolo

On Mon, Oct 24, 2016 at 7:08 PM, Gabriel Greene <gabriel.greene87 at gmail.com>
wrote:

> Hello,
>
> After converting the pseudo on https://www.cp2k.org/static/potentials/ from
> CPMD format to UPF using cpmd2upf.x from QEv5.2.0, and then running in
> pwscf, I realised it may be the same pseudopotential as Ge.pz-hgh.UPF on
> http://www.quantum-espresso.org/pseudopotentials
> <http://www.quantum-espresso.org/pseudopotentials/about/>.
>
> However, now I am a bit confused because in the CPMD format it says
>
> &END
>  &POTENTIAL
>      GOEDECKER
>  3                                    LMAX
>
>
> Which I thought would indicate that lmax is 3. However after converting to UPF, it says lmax="2" in the PP_HEADER section. Assuming converting between formats doesnt change the actual pseudo, then is the definition of lmax different for CPMD compared to UPF?
>
>
> Also, the number of wavefunctions is 0 in the pseudo I converted from CPMD (Ge_GTH_LDA_fromCP2K.UPF in attachments), and it is 3 in Ge.pz-hgh.UPF... and obviously the entries in PP_PSWFC sections are completely different (missing in Ge_GTH_LDA_fromCP2K.UPF...), as are the charge density PP_RHOATOM sections of the files. In Ge.pz-hgh.UPF, there are orbitals generated with ld1.x, so again I thought the pseudos must be generated differently after all..
>
>
> Yet they yield very similar results (total energies, bandgaps, relaxed lattice constants, see attached inputs and outputs). Comparing the pwscf outputs I see that the starting point for SCF was superposition of atomic orbitals for Ge.pz-hgh.UPF and totally random wfcs for Ge_GTH_LDA_CP2K.UPF. So I am assuming because of the similarity of the potentials, the resulting self-consistent densities and potentials after iterating to convergence end up to be the same, despite the different starting wfcs.
>
>
> Yet I am still unsure about how Ge.pz-hgh.UPF was generated. Looking at the non-local projectors of this file, the cutoff radius of for e.g. 1S is
>
> cutoff_radius="3.893327136974858E+000"
>
> and the cutoff radius of the valence orbitals (PP_CHI), taking 4S as an example
>
> cutoff_radius="1.876811407389333E+000"
>
> I would have thought the cutoff radii of core orbitals which are to be pseudized should be smaller than the r_cutoff of the resulting valence pseudo orbitals? What are the units are these cutoff radii?
>
>
> Please let me know if I am misinterpreting something. I would really appreciate any help on this.
>
>
>
> Gabriel,
>
> Research Associate,
>
> Atomistic Simulation Centre,
>
> School of Mathematics and Physics,
>
> Queen's University Belfast
>
>
>
>
> On Fri, Oct 21, 2016 at 6:45 PM, Gabriel Greene <
> gabriel.greene87 at gmail.com> wrote:
>
>> Dear Ari,
>>
>> Thank you but I am not looking for a new GTH-type pseudopotential for Ge,
>> nor am I trying to generate a new Ge pseudo. I am looking for the
>> generation information of the specific Ge GTH pseudo that is hosted on the
>> QE website, with file name "Ge.pz-hgh.UPF". See first attachment.
>>
>> Thank you for the link, I used your instructions to get the 2nd attached
>> file. Seeing as lmax is different here compared to Ge.pz-hgh.UPF, I am
>> presuming these are different pseudopotentials (not just different file
>> formats).
>>
>> Specifically, I am looking for lmax, r_cut for each l-channel, and
>> generating XC functional for Ge.pz-hgh.UPF. I have lmax (2) and the XC
>> funct (PZ-LDA), so really I just need the cutoff radii that were used to
>> generate this pseudo, if available.
>>
>>
>> On another point, I find that the link to M. Kracks pseudopotentials on
>> the page http://www.quantum-espresso.org/pseudopotentials/about/  is
>> currently broken (leads to a "phonebook" page, no pseudo details anywhere).
>> Thank you again for the correct link to this page.
>>
>>
>> Thanks again,
>>
>> Gabriel
>>
>> Research Associate,
>>
>> Atomistic Simulation Centre,
>>
>> School of Mathematics and Physics,
>>
>> Queen's University Belfast
>>
>>
>> On Fri, Oct 21, 2016 at 5:51 PM, Ari P Seitsonen <Ari.P.Seitsonen at iki.fi>
>> wrote:
>>
>>>
>>> Dear Gabriel,
>>>
>>>   You find many GTH/HTH-type pseudo potential via Matthias Karck's site:
>>>
>>> https://www.cp2k.org/static/potentials/
>>>
>>> I could see pseudo potentials with LDA - in CPMD format, that you would
>>> then convert with the UPF tools of QE into the .UPF files - containing four
>>> or 14 valence electrons (the latter most likely being preferred). The
>>> parametrisation of the LDA is not Perdew-Zunger though, but probably the
>>> form with PADE interpolation is very close (experts please correct me).
>>>
>>>     Greetings from Paris,
>>>
>>>        apsi
>>>
>>> -=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=
>>> -=*=-=*=-=*=-=*=-
>>>   Ari Paavo Seitsonen / Ari.P.Seitsonen at iki.fi /
>>> http://www.iki.fi/~apsi/
>>>     Ecole Normale Supérieure (ENS), Département de Chimie, Paris
>>>     Mobile (F) : +33 789 37 24 25    (CH) : +41 79 71 90 935
>>>
>>>
>>>
>>> On Fri, 21 Oct 2016, Gabriel Greene wrote:
>>>
>>> Dear QE community,
>>>> I am wondering if it is possible to retrieve the pseudopotential
>>>> generation information of the Ge pseudopotential Ge.pz-hgh.UPF which can be
>>>> downloaded on the QE pseudopotential page?
>>>>
>>>> In the PP_INFO section, it states
>>>>
>>>>     Generation configuration: not available.
>>>>     Converted from CPMD format using cpmd2upf v.5.0.1 - PG 10Jul201
>>>>
>>>> Which prompted me to check the CPMD website page
>>>>
>>>> http://cpmd.org/downloadable-files/no-authentication/pseudo-
>>>> potentials/germanium
>>>>
>>>> but there I do not see a Ge Goedecker-Teter-Hutter PP generated with
>>>> the PZ-LDA functional (the closest I can see is a GTH PP generated with the
>>>> B
>>>> LYP functional).
>>>> Thanks very much,
>>>>
>>>> Gabriel Greene-Diniz
>>>>
>>>> Research Associate,
>>>>
>>>> Atomistic Simulation Centre,
>>>>
>>>> School of Mathematics and Physics,
>>>>
>>>> Queen's University Belfast
>>>>
>>>>
>>>>
>>> _______________________________________________
>>> Pw_forum mailing list
>>> Pw_forum at pwscf.org
>>> http://pwscf.org/mailman/listinfo/pw_forum
>>>
>>
>>
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum
>



-- 
Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,
Univ. Udine, via delle Scienze 208, 33100 Udine, Italy
Phone +39-0432-558216, fax +39-0432-558222
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20161026/9898d78b/attachment.html>

More information about the users mailing list