Discussion:
[Wien] scf not converging for slab calculations with spin orbit coupling
Anup Shakya
2018-10-28 07:08:04 UTC
Permalink
Dear All,

I am doing non magnetic GGA+SOC+U calculations for a material containing
Sm. Without SOC the convergence was obtained using MSR1 scheme with a
mixing factor of 0.1. But when SOC is included then the scf is oscillating.
in cycle 119 ETEST: 0.2087 CTEST: .187796
in cycle 149 ETEST: .1798 CTEST: .2784
Then I stopped the calculations and tried to perform calculations by
changing the scheme from MSR1 to PRATT but then I got an error as shown
below.
in cycle 2 ETEST: .1283169350000000 CTEST: .1903329
in cycle 3 ETEST: .3288391250000000 CTEST: .7860095
in cycle 4 ETEST: 12.7007578000000000 CTEST: 6.2762989
Error in LAPW1
'SELECT' - no energy limits found for atom 1 L=
0
'SELECT' - E-bottom -200.00000 E-top -5.32257
This error was not there with MSR1 scheme, but its not getting converged.
Should I increase the mixing factor and keep the MSR1 scheme?
Please suggest me what to do.

Sincerely,
Anup Pradhan Sakhya (Ph.D.)
Visiting Post-Doctoral Fellow
DCMP&MS, TIFR, Mumbai
Laurence Marks
2018-10-28 13:19:56 UTC
Permalink
The energy is a very crude test of convergence and can be misleading. You
should look at the magnetic moments (:MM*), the interstitial charge and
pseudo charge (":CTO ", :CPC) and the orbital moments. They may be walking.
If they are really oscillating (which :ENE won't show), MSEC3 may help.
Don't use PRATT.

Most problems are due to bad technical parameters/physics/chemistry. For
instance a badly setup slab does not have to converge well or even at all.

_____
Professor Laurence Marks
"Research is to see what everybody else has seen, and to think what nobody
else has thought", Albert Szent-Gyorgi
www.numis.northwestern.edu
Post by Anup Shakya
Dear All,
I am doing non magnetic GGA+SOC+U calculations for a material containing
Sm. Without SOC the convergence was obtained using MSR1 scheme with a
mixing factor of 0.1. But when SOC is included then the scf is oscillating.
in cycle 119 ETEST: 0.2087 CTEST: .187796
in cycle 149 ETEST: .1798 CTEST: .2784
Then I stopped the calculations and tried to perform calculations by
changing the scheme from MSR1 to PRATT but then I got an error as shown
below.
in cycle 2 ETEST: .1283169350000000 CTEST: .1903329
in cycle 3 ETEST: .3288391250000000 CTEST: .7860095
in cycle 4 ETEST: 12.7007578000000000 CTEST: 6.2762989
Error in LAPW1
'SELECT' - no energy limits found for atom 1 L=
0
'SELECT' - E-bottom -200.00000 E-top -5.32257
This error was not there with MSR1 scheme, but its not getting converged.
Should I increase the mixing factor and keep the MSR1 scheme?
Please suggest me what to do.
Sincerely,
Anup Pradhan Sakhya (Ph.D.)
Visiting Post-Doctoral Fellow
DCMP&MS, TIFR, Mumbai
Loading...