Page 1 of 1

Symmetry and k-point in AIMD

Posted: Fri Sep 14, 2018 7:51 am
by jinylee
Dear VASP users,

When and how do the VASP create an IBZKPT file by KPOINT when calculating AIMD?
I calculated ‘single point energy’ for graphene monolayer with 30 angstrom(c-axis) of vacuum, with Na-ion inside.
The Na-ion moved at everyseconds, of course the system has no symmetry.
The IBZKPT file is below, for the Monkhorst-Pack 2 x 2 x 1 mesh when I used the same ISYM = default option.
Automatically generated mesh (M 221)
2
Reciprocal lattice
0.25000 0.25000 0.00000 2
-0.25000 0.25000 0.00000 2

Does it mean that the program itself recognizes symmetry and determines the grid settings accordingly?
If so, how is it determined in AIMD calculations, where the structure is constantly changing (It means ''molecular dynamics)?

I also wonder that differences of working of ISYM=2 comparing with ISYM=1, 3,
And what is the correct meaning of ‘fastsym’, that is default of PAW?

Plz help me, please!

Re: Symmetry and k-point in AIMD

Posted: Wed Sep 19, 2018 9:26 am
by admin
Hello,

depending on which DFT functional is used ISYM has a default value of 1, 2 or 3.
Please read following wiki-page entry for more information.
https://cms.mpi.univie.ac.at/wiki/index.php/ISYM

Regargding your question.
In general, ISYM > 0 means that symmetry operations are used to reconstruct the wavefunction outside the irreducible Brillouin zone.
The allowed symmetry operations are determined at the very beginning of the calculation based on crystal structure, magnetic alignment and velocities of the ions and are fixed during a VASP run. However, after each ionic step of an molecular dynamics simulation the structure, magnetism and velocities change and with that the allowed symmetry operations. Unfortunately, this change cannot be taken into account currently, which is the reason why we strictly recommend to set ISYM=0 for MD runs (IBRION=0).

with regards,
vasp-support