My Community

Hello everyone,
I am very new to High pressure AIMD simulations with VASP. I would like to calculate the diffusivity and ionic conductivity at High pressure from the MD simulation outputs. However, I am not very sure if my INCAR file is correct.
Please see the INCAR file used in this simulation and help in setting the tags correctly. Thank you.
SYSTEM = KAlSi3O8
NCORE = 16
ISTART = 0
ISMEAR = 0
SIGMA = 0.05
NSW = 10000
POTIM = 2
#SMASS = 0
NBLOCK = 1
TEBEG = 1000
TEEND = 1000
MAXMIX = 40
IBRION = 0
ENCUT = 500
EDIFF = 1E-4
ALGO = Very_Fast
ISYM = 0
NELMIN = 4
LREAL = Auto
LWAVE = .FALSE.
LCHARG = .FALSE.
#INCAR molecular-dynamics tags NpT ensemble
MDALGO = 3 # using Langevin thermostat
ISIF = 3 # compute stress tensor and change box volume/shape
LANGEVIN_GAMMA = 10.0 10.0 10.0 10.0 # Langevin friction coefficient for the 4 atomic species
LANGEVIN_GAMMA_L = 10.0 # Langevin friction coefficient for lattice degrees of freedom
PMASS = 1000 # the fictitious mass of the lattice degrees of freedom
PSTRESS = 100 # Pressure of 10 GPa

Hello!

Your INCAR tags look reasonable to me, they correspond to an NpT simulation using the Langevin thermostat at high pressure. However, please note that we cannot optimize the settings for you. For example, you still need to verify yourself whether the time step is acceptable and your thermostat/barostat settings correspond to a fitting relaxation time. Also, please ensure the system is properly equilibrated before starting production runs. If the density of the starting structure does not correspond to the barostat pressure expect that the system will take quite some time to accommodate. Note that a dynamic lattice typically requires a 30% increase of ENCUT to avoid Pulay stress. After large volume changes, e.g., after equilibration, the simulation should be restarted because then the basis set needs to be adapted (see this related Wiki article: https://www.vasp.at/wiki/index.php/Ener ... lay_stress).

I hope this gives you a good starting point so you can avoid some common pitfalls.

All the best,
Andreas Singraber