Problem with Langevin thermostat in NPT MD simulation

[praveenkumar_hiremath1]

Hello Users,

I am novice at performing ab initio MD (and VASP) and doing NPT simulations using VASP (5.4.4). Here is my INCAR file.

SYSTEM = Bulk
PREC = High
ISTART = 0
ICHARG = 2
ISMEAR = 0
SIGMA = 0.2
EDIFF = 1E-5
EDIFFG = -0.01
LWAVE = .FALSE.
LCHARG = .FALSE.
LVTOT = .FALSE.
ENCUT = 400.0000
MDALGO = 3
ISIF = 3
ISYM = 0
LREAL = Auto
dynamic:
IBRION = 0
NSW = 10000
POTIM = 0.5
SMASS = 0
TEBEG = 1000
TEEND = 1000

The problem is that the temperature keeps dropping and the decrease is almost linear. I have noticed that SMASS value controls the speed of temperature drop. Reducing the timestep (POTIM) resulted in very very slow decrease in temperature. But using very small POTIM means I have to run the simulation for a lot of timesteps to truly get equilibrium NPT configuration. Moreover, using very small POTIM also doesnt stop the temperature from decreasing. Please can I someone look at my INCAR file and may be guide with regards to things that are wrong with my INCAR?? Thank you very much in advance.

Praveen

[henrique_miranda]

Please read carefully this documentation page:
https://www.vasp.at/wiki/index.php/MDALGO

Since you are using MDALGO = 3, you should at least specify the following tag:
https://www.vasp.at/wiki/index.php/LANGEVIN_GAMMA

EDIT: This page might also be relevant
https://www.vasp.at/wiki/index.php/Langevin_thermostat