I started playing with non-collinear magnetization and the bfcmt variable, and as a first example I tested bulk Silicon. I just wanted to test that starting from an antiferromagnetic configuration the system would converge to a non-magnetic configuration.
When using collinear magnetism by turning off the spin-orbit coupling, the calculation converges rapidly to magnetization values of ~E-17. However, when making spinorb=true, the calculation does converge, but giving magnetic moments of ~E-05 mantaining the antiferromagnetic configuration. Increasing calculation parameters like nkgrid or nemtpy lowers this value, but never gives less than E-07. (input file below)
Is this value low enough (I would say that it is not compared to ~E-17), or am I missing something very basic?
did you try to turn Si into a magnetic state? I would recommend for playing around with non-colinear magnetism a more stable system like Fe bcc/fcc for tryout, since its not expected physically that Si becomes magentic (see Stoner criteria for FM for example). Anyway, just for clarity you should be careful with magnetic moments in the order of 1E-5 mub and below provided by elk since these have to be careful converged or a just some kind of noise!!
best regards
Michael
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
The calculations with magnetic systems make perfect sense. I just wanted to test accuracy in convergence with respect to initial magnetization in non-magnetic systems. Probably I should check more carefully the calculation parameters, specially for such low values of magnetic moments, as you said.
All the best,
Jon
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Dear ELK users and developers,
I started playing with non-collinear magnetization and the bfcmt variable, and as a first example I tested bulk Silicon. I just wanted to test that starting from an antiferromagnetic configuration the system would converge to a non-magnetic configuration.
When using collinear magnetism by turning off the spin-orbit coupling, the calculation converges rapidly to magnetization values of ~E-17. However, when making spinorb=true, the calculation does converge, but giving magnetic moments of ~E-05 mantaining the antiferromagnetic configuration. Increasing calculation parameters like nkgrid or nemtpy lowers this value, but never gives less than E-07. (input file below)
Is this value low enough (I would say that it is not compared to ~E-17), or am I missing something very basic?
Thanks in advance,
Jon
tasks
0
xctype
20
nempty
20
reducebf
0.5
spinpol
.true.
spinorb
.true.
avec
5.13 5.13 0.00
5.13 0.00 5.13
0.00 5.13 5.13
sppath
''
atoms
1 : nspecies
'Si.in' : spfname
2 : natoms; atposl below
0.0 0.0 0.0 0.0 0.0 0.01
0.25 0.25 0.25 0.0 0.0 -0.01
ngridk
12 12 12
epspot
0.1000000000E-06
epsengy
0.1000000000E-04
Last edit: JW 2017-05-31
Dear Jon,
did you try to turn Si into a magnetic state? I would recommend for playing around with non-colinear magnetism a more stable system like Fe bcc/fcc for tryout, since its not expected physically that Si becomes magentic (see Stoner criteria for FM for example). Anyway, just for clarity you should be careful with magnetic moments in the order of 1E-5 mub and below provided by elk since these have to be careful converged or a just some kind of noise!!
best regards
Michael
Dear Michael,
Thank you for the response.
The calculations with magnetic systems make perfect sense. I just wanted to test accuracy in convergence with respect to initial magnetization in non-magnetic systems. Probably I should check more carefully the calculation parameters, specially for such low values of magnetic moments, as you said.
All the best,
Jon