[QE-users] DFT+U - error when using the pseudopotential projectors

[Iurii TIMROV]

Dear Thomas,


> especially the parenthesis - I can somehow understand the problem; I guess it has to do with the possibility to properly project onto the "localized" subspace, but do you know if there is a more rigorous explanation, maybe a publication? Or Iurii? Maybe in one of your papers?


No, I am not aware of any publication where this is discussed. I do not think this is discussed anywhere, it is just a technicality (how things are implemented).


Regarding the parenthesis: In the older versions of QE (before the new DFT+Hubbard syntax introduced in v7.1), the code was considering in some routines the last "L channel" (where L is the orbital quantum number) as the Hubbard channel. But since v7.1 there is no this logic for DFT+U and DFT+U+V ("atomic" and "ortho-atomic" projectors), because now the user has to specify in the input N and L (where N is the principal quantum number), so there is no ambiguity. I presume (but I am not sure) that this might be the reason why there is a problem for the "pseudo" type of Hubbard projectors (i.e. the same projectors as in VASP) when there are also semicore states (so you have the same L twice).


Greetings,

Iurii


--
Dr. Iurii TIMROV
Senior Research Scientist
Theory and Simulation of Materials (THEOS)
Swiss Federal Institute of Technology Lausanne (EPFL)
CH-1015 Lausanne, Switzerland
+41 21 69 34 881
http://people.epfl.ch/265334
________________________________
From: Thomas Brumme <thomas.brumme at tu-dresden.de>
Sent: Friday, December 9, 2022 11:18:47 PM
To: Paolo Giannozzi; users at lists.quantum-espresso.org
Cc: Iurii TIMROV
Subject: AW: [QE-users] DFT+U - error when using the pseudopotential projectors


Thanks Paolo for this important information - I haven't had the time yet to look at the code itself.


Concerning this small info:

one might run into troubles when using a PP with semicore states with same l as valence states
(also otherwhere for DFT+Hubbard)


especially the parenthesis - I can somehow understand the problem; I guess it has to do with the
possibility to properly project onto the "localized" subspace, but do you know if there is a more

rigorous explanation, maybe a publication? Or Iurii? Maybe in one of your papers?


Kind regards


Thomas

--
Dr. rer. nat. Thomas Brumme
Theoretical chemistry
TU Dresden - KOE / 103
Bergstr. 66c
01069 Dresden

Tel:  +49 (0)351 463 40844

email: thomas.brumme at tu-dresden.de


________________________________
Von: Paolo Giannozzi <paolo.giannozzi at uniud.it>
Gesendet: Freitag, 9. Dezember 2022 17:06
An: users at lists.quantum-espresso.org
Cc: Brumme, Thomas; Iurii TIMROV
Betreff: Re: [QE-users] DFT+U - error when using the pseudopotential projectors

On 07/12/2022 16:26, Iurii TIMROV via users wrote:

> it might be that something has been broken over the years

it might be, given the complete absence of tests and examples for that
feature. It isn't, though: with a different pseudopotential
(Co.pbesol-n-kjpaw_psl.1.0.0.UPF) not containing semi-core states,
neither the current code nor the first version (5.0.3) implementing
DFT+U with pseudopotential projectors crash. The results seem to be the
same (convergence is nasty as usual so it is difficult to get exactly
the same numbers).

A comment in PW/src/init_q_aeps.f90 seems to confirm that the problem is
in the logic used to choose the projectors:

   ! NOTE: one might run into troubles when using a PP with semicore
   ! states with same l as valence states (also otherwhere for DFT+Hubbard

Paolo
--
Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,
Univ. Udine, via delle Scienze 206, 33100 Udine Italy, +39-0432-558216
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20221210/4037dde6/attachment.html>