This thread is locked.Only browsing is available.
Top Page > Browsing
non collinear calculations in openmx
Date: 2022/02/25 13:58
Name: Gurung

Hello OpenMx users,
I have some questions regarding the non collinear calculations

1)The non collinear calculations filename.out file shows spin moment with up and down with sum and differences format. It also shows the orbital decomposed part in the same format of the table. In non collinear calculations, spin is not a good quantum number. So, I am confused with the notations of the up and down. Can anyone please explain why such notation is used? I am using the Euler angles provided to calculate the mx, my and mz components for each orbitals.

2)Is the filename.nc.xsf output file for the final spin projection or initial spin projection?

3) The non collinear calculations seems to be difficult to converge. What will be the suggestions for the better convergence?

Best regards
Gurung
メンテ
Page: [1]

Re: non collinear calculations in openmx ( No.1 )
Date: 2022/02/26 01:40
Name: Gurung

Hello OpenMx users,
I am currently carrying out non collinear calculations without imposing any constrained as the ground state for the system is non collinear in the experiment. I found that the scf steps do not run up to the max iterations with following set up

scf.maxIter 400
scf.criterion 1.0e-8
MD.Type nomd

Following is the part of the calculations which stops before reaching any criteria mentioned above

SCF= 172 NormRD= 0.000096848529 Uele= -70.323742886103
SCF= 173 NormRD= 0.000079490277 Uele= -70.323742611937
SCF= 174 NormRD= 0.000083425122 Uele= -70.323743221517
SCF= 175 NormRD= 0.000108897043 Uele= -70.323742566469
SCF= 176 NormRD= 0.000090733962 Uele= -70.323742307241
SCF= 177 NormRD= 0.000077585938 Uele= -70.323742301006

*******************************************************
Total energy (Hartree) at MD = 1
*******************************************************


I saw in some of the non collinear calculation examples all the scf mixing weights (Initial, maximum, minimum) tags are assigned zero (example in NiO_NC.dat). I am using the openmx3.9 version.



Best regards
Gurung
メンテ
Re: non collinear calculations in openmx ( No.2 )
Date: 2022/02/26 05:39
Name: Malone

> I found that the scf steps do not run up to the max iterations

From the execution log you may see that the scf.criterion was met (last deviation of electron energy was less than 1E-8).

メンテ
Re: non collinear calculations in openmx ( No.3 )
Date: 2022/02/26 12:34
Name: Gurung

Hello Malone,
Thank you.

Best regards
Gurung
メンテ

Page: [1]