Routine scf crashing

[Abdulrahman_Allangawi]

Dear all,

I am running optimization for system containing a graphene sheet and macrocycle anchored on it. For some reason, the calculations crash during the initial scf steps, it seems that the energy fluctuates quite a lot. To solve this issue, I have added the following commands in the INCAR, as suggested to me by another colleague,

Code: Select all

AMIX     = 0.005       # Mixing parameter for charge density (controls how new densities mix with old ones)
AMIX_MAG = 0.05       # Mixing parameter for magnetic moments
BMIX     = 0.0001    # Beta mixing parameter for charge density
BMIX_MAG = 0.0001    # Beta mixing parameter for magnetic moments

I recognize that the values I chose for AMIX and AMIX_MAG are quite small, however, if I choose larger values, such as 0.01 and 0.1, respectively, the scf crashes in the initial steps. Can someone advise on this issue? I have attached the files below.

CoPc_C.zip

Note that I canceled the run of CoPc_C after 1000 scf steps, but I previously let it run for a day and it only finished 3 ionic steps, with each step having 1500+ scf.

[marie-therese.huebsch]

Hi, I see you are using VASPsol

VASPsol++ is a framework for implementing complex continuum fluid models within density functional theory calculations performed using the Vienna Ab initio Simulation Package (VASP). It is being actively developed within the Plaisance group at Louisiana State University and has its origins in the VASPsol code developed in the Hennig group at the University of Florida.

Mind that this is not developed by the VASP team and that any issues that ocure with VASPsol should be reported to the developers of VASPsol. Of course, as VASP is used in these calculations the error may be due to VASP. In that case, please isolate a pure VASP calculatiuon that yields the same error and report it here. We are happy to have a look at it.

That said, the mixing parameters are really odd. Maybe you can try a different ALGO instead. But anyways it rather points to some inconsistency in the electronic update (perhaps caused by VASPsol but I don't want to jump to conclusions). Hope my reply is a good starting point for you to further tackle the issue. Please let me know, how you want to proceed!

[Abdulrahman_Allangawi]

Thanks very much for your reply, I will do as suggested and report back with the results.

[Abdulrahman_Allangawi]

Dear,

Please view the attached file

CoPc_C.zip

, I have recreated the issue on the latest VASP patch.

[marie-therese.huebsch]

Sorry for the long scilence. Unfortunately the forum had a system change and we did not notice that the notification where not send for quite some time. So I totally missed your response. I will look into it this week and get back to you with an update asap.

[marie-therese.huebsch]

I seem to miss something. Both of the calculations you have attached are VASPsol calculations. (See lines SOL: in OSZICAR)

Here is some general advice and comments.

• Did you try to simplify your system and find good parameters for that calculation first? For instance, you are performing a magnetic calculation (ISPIN=2) with all magnetic moments set to zero. Did you try ISPIN=1?

• Did you consider electrostatic corrections?

• Have you tried ALGO=Conjugate with ISEARCH=1?

Please try to create a VASP only example if you want us to look at the issue.