Page 1 of 1

Md simulation with SOC

Posted: Tue Oct 25, 2022 5:20 pm
by alok_shukla1
Dear Experts,
I have one query about MD simulation. I am trying MD+SOC calculation, but after the first complete cycle calculation stops automatically, while without SOC, it is working fine. Is this possible that we can use SOC with MD or not?

Thanks.

Re: Md simulation with SOC

Posted: Thu Nov 03, 2022 1:33 pm
by jonathan_lahnsteiner2
Dear alok_shukla1,

In principle it should be possible. If you need
further help I would need you to upload your
POSCAR, INCAR, KPOINTS, POTCAR
OUTCAR as it is explained in the
forum guidelines
forum/viewtopic.php?f=4&t=17928


All the best Jonathan

Re: Md simulation with SOC

Posted: Sun Nov 06, 2022 1:08 am
by alok_shukla1
Dear Expert,
Here I am attaching above mentioned files. I am trying to run MD in two steps. First, I considered the NVE ensemble (files are attached). After this, I'll consider the NVT ensemble. In the first steps, calculations are not proceeding. Please check it.

Re: Md simulation with SOC

Posted: Thu Nov 10, 2022 4:05 pm
by jonathan_lahnsteiner2
Dear alok_shukla1,

I was checking your input files. From that I see you want to do
a molecular dynamics run with spin orbit coupling.
Therefore you should remove the ISPIN=2 tag from
your INCAR file. This is only necessary for the collinear spin treatment which your are not planning to do.
Additionally you have to set the flag IWAVPR = 0 in your INCAR file. This will controls the update of the wavefunctions.
Currently this is the only update scheme which can be used in your case.
The EDIFFG flag has to be removed because otherwise your calculation will stop
after the first step.

Last thing I recognized, you are setting up a micro canonical MD simulation by not specifying a MDALGO tag
and setting SMASS = -1. Additionally you are setting a temperature gradient by setting TEBEG=0 and TEEND=300.
This does not make sense because it is not possible to apply a temperature gradient in a microcanonical MD.
Also it is not advisable to initialize an MD at 0K, because all velocities are set to zero.
I would recommend to set MDALGO to either 2 or 3 in your case. See here for more information:
https://www.vasp.at/wiki/index.php/MDALGO.
Or you do a pre-equibrilation with MDALGO=2 or 3 and switch to the microcanonical MD later on.


When setting/removing the before mentioned flags in your INCAR file the MD was working for me as expected.
I hope this helps and if you have further questions just ask.

All the best Jonathan

Re: Md simulation with SOC

Posted: Sat Nov 12, 2022 10:02 am
by alok_shukla1
Dear Expert,

Thank you so much for checking my input files. I implemented your suggestions in my input file. I'll update here if it works fine.