How to enhance the symmetry constraints?

Top Page > Browsing

How to enhance the symmetry constraints?

Date: 2024/12/05 20:34
Name: shiyc <shiyc814@163.com>: Dear Contributors,

I am encountering an issue while performing SCF calculations using OPENMX. Specifically, for a given structure, the magnetic moments of symmetry-equivalent atoms exhibit a difference of approximately 0.007 μB. However, when I use the same structure in VASP, the magnetic moments of these symmetry-equivalent atoms are identical, as expected.

I suspect this discrepancy might be related to how symmetry constraints are applied during the SCF process in OPENMX. Is there any way to enhance the symmetry constraints in OPENMX to ensure more consistent results for symmetry-equivalent atoms?

Below are the key parameters I used for my OPENMX calculation:
Species.Number 3
<Definition.of.Atomic.Species
B B7.0-s2p2d1 B_PBE19
Fe Fe5.5H-s2p2d1 Fe_PBE19H
La La8.0-s3p2d2f1 La_PBE19
Definition.of.Atomic.Species>
scf.XcType GGA-PBE
scf.SpinPolarization on
scf.ElectronicTemperature 300.0
scf.energycutoff 400.0
scf.maxIter 200
scf.EigenvalueSolver band
scf.Kgrid 18 18 3
scf.Mixing.Type rmm-diisk # Simple|Rmm-Diis|Gr-Pulay
scf.Init.Mixing.Weight 0.0100 # default=0.30
scf.Min.Mixing.Weight 0.0001 # default=0.001
scf.Max.Mixing.Weight 0.3000 # default=0.40
scf.Mixing.History 30 # default=5
scf.Mixing.StartPulay 40 # default=6
scf.criterion 1.0e-7 # default=1.0e-6 (Hartree)
scf.restart on

Page: [1]

Re: How to enhance the symmetry constraints? ( No.1 )

Date: 2024/12/06 14:24
Name: A Terasawa <a.terasawa@rist.or.jp>

In openmx, it is possible to control the real space grid by scf.Ngrid instead of scf.energycutoff.

https://www.openmx-square.org/openmx_man3.9/node37.html

If you can adjust this parameter so that the positions of real space grids with respect to atoms are symmetric, spacial symmetry may be improved.
Another option is to enlarge scf.energycutoff to make the discrepancy of relative positions small enough.

But sorry, I'm not a specialist for this subject, and still needs some verification.

Re: How to enhance the symmetry constraints? ( No.2 )

Date: 2024/12/12 22:55
Name: shiyc <shiyc814@163.com>

Dear Professor Terasawa,

I hope this email finds you well. Thank you very much for your valuable suggestions regarding the adjustment of the scf.Ngrid parameter in OpenMX to improve spatial symmetry. I will carefully test this method and also explore tuning scf.energycutoff as you recommended to optimize the calculation setup for better symmetry.

Additionally, I have recently been reading your article published in Journal of the Physical Society of Japan 88, 114706 (2019) and have been trying to compute magnetic exchange coupling constants (Jij) using OpenMX. During my calculations, I noticed that the magnetic moments and Jij values vary significantly depending on the choice of atomic orbital basis sets and pseudopotentials. I would like to seek your advice on how to select an appropriate basis set. Are there specific guidelines or principles to follow to ensure accurate results, particularly for magnetic properties?

Thank you again for your insightful work and for taking the time to provide guidance. I look forward to your advice.

Re: How to enhance the symmetry constraints? ( No.3 )

Date: 2024/12/13 10:11
Name: A Terasawa <a.terasawa@rist.or.jp>

Thank you very much for your reply!

We have found that the choice of basis set in SCF calculation affect the calculated result of Jij, and we are now working on this problem.
But in the current implementation, one needs to set the size of basis set not so large.
Recommended basis sets are s2p2d2 for Fe, s2p1d1 for Nd.

The related note is found in this page.
https://www.openmx-square.org/openmx_man3.9/node122.html

Sorry for your inconvenience.

Re: How to enhance the symmetry constraints? ( No.4 )

Date: 2024/12/13 15:30
Name: shiyc <shiyc814@163.com>

Dear Professor Terasawa,

Thank you so much for sharing your insights regarding the choice of basis sets!

I had come across the manual link you kindly shared earlier, and it was very helpful in understanding the basis set selection. However, I couldn’t find specific guidance for La in the documentation. Could you kindly share your advice or recommendations for selecting an appropriate basis set for La?

Thank you again for your invaluable guidance. I greatly appreciate your support and look forward to your advice.

Re: How to enhance the symmetry constraints? ( No.5 )

Date: 2024/12/13 16:01
Name: A Terasawa <a.terasawa@rist.or.jp>

Roughly speaking, choice of the basis set should be small for lanthanoids, and I would recommend minimal basis set,
because PAOs for those elements tend to have long tail functions.
However, it is necessary to say that this feature is not carefully examined for La itself.
It is desirable to compare the calculated results for a known material containing La with those of previous studies.

Sorry this may be a tentative answer, and we are now working on further development of Jij calculation.
And I will let users know when the development is done as quickly as possible.

Re: How to enhance the symmetry constraints? ( No.6 )

Date: 2024/12/13 20:20
Name: shiyc <shiyc814@163.com>

Thank you very much for your advice regarding the basis set selection for La and the explanation about PAOs. I truly appreciate the efforts you are putting into improving Jij calculations and look forward to the results of your development！

Page: [1]