[QE-users] Query for skip_type and equiv_type for DFT+U+V

[Iurii TIMROV]

Dear K C Bhamu,


If you use find_atpert = 1 then the HP code will determine automatically which atoms must be perturbed. In this case the HP code checks the traces of occupation matrices (Tr[ns(na)] in the output file of pw.x) and then decides what atoms must be perturbed. If e.g. you have Fe atoms with occupations that differ more than some threshold (docc_thr = 5.D-5) then the HP code will consider them as nonequivalent atoms and perturb them.


> skip_type(1)=.true.
> skip_type(2)=.true.
> equiv_type(1)=1
> equiv_type(2)=1


This you can try to use with the option find_atpert = 2 (this is a bit more advanced). But I would recommend to try first find_atpert = 1.


In any case, your setup is not correct here. To have a correct setup, you should provide the full list for  ATOMIC_SPECIES,  ATOMIC_POSITIONS, and the atomic occupations: Tr[ns(na)] printed at the last SCF iteration in the output file of the pw.x run.


> As all Fe will have the same Hubbard parameter (as reported in the literature for the same system)...


In the literature people often use empirical U, so that's why they use a single value of U for all Fe atoms. With the HP code you can do things better, i.e. to find out if there are non-equivalent Fe atoms that require their own first-principles U.


HTH


Iurii


P.S.: "Dr." is used with the last name not with the first name, so in my case this will be "Dear Dr. Timrov". But if you want to address me using my first name then just write "Dear Iurii". Sorry, I don't want to bother you with this, but just to makes things clear 😊


--
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: users <users-bounces at lists.quantum-espresso.org> on behalf of Dr. K. C. Bhamu <kcbhamu85 at gmail.com>
Sent: Wednesday, November 24, 2021 1:00:13 PM
To: Quantum Espresso users Forum
Subject: [QE-users] Query for skip_type and equiv_type for DFT+U+V

Dear Dr. Iurii (😊),

The ATOMIC_POSITIONS cart of my structure (ferrimagnetic system with ntyp = 3 ) is

ATOMIC_POSITIONS (crystal)
Fe1           0.0000000000       -0.0000000000       -0.0000000000 (four times)
Fe2           0.3749993721        0.3749993349        0.3750019210 (two times)
O             0.2454868415        0.2454868399        0.2454949640 (eight times)

Previously, based on ATOMIC_SPECIES cart, I defined my Hubbard tag as below

  Hubbard_V(1,1,1) = 1.d-8
  Hubbard_V(2,2,1) = 1.d-8
  Hubbard_V(3,3,1) = 1.d-8

In the hp output file, I saw that only one  Fe1 atom was perturbed.

Then I defined below Hubbard tags according to ATOMIC_POSITIONS cart as

  Hubbard_V(1,1,1) = 1.d-8
  Hubbard_V(5,5,1) = 1.d-8
  Hubbard_V(7,7,1) = 1.d-8

which perturbed four atoms

        1      Fe1    55.8450   tau( 1) = (  0.00000  0.00000  0.00000  )
        4      Fe1    55.8450   tau( 4) = (  1.75000  1.01036  0.40825  )
        5      Fe2    55.8450   tau( 5) = (  0.75000  0.43301  0.30619  )
        7      O      15.9994   tau( 7) = (  0.49098  0.28347  0.20045  )

As all Fe will have the same Hubbard parameter (as reported in the literature for the same system), I would like to skip the perturbation for atom Fe1(4) and Fe2(5).

Is it okay if I do so?

To skip the perturbation for Fe1(4) and Fe2(5), I have updated below tags in my hp input file:

skip_type(1)=.true.
skip_type(2)=.true.
equiv_type(1)=1
equiv_type(2)=1
disable_type_analysis=.true.,

Could you please correct me if this modification is correct or not?

Sorry for the list of questions. But I am sure, my these queries will be helpful for others too.

Thanks and regards

K C Bhamu

Postdoctoral Fellow
University of Ulsan
South Korea

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20211124/3e891141/attachment.html>