Keeping the temperature constant with Nose-Hoover thermostat

LAMMPS LAMMPS Mailing List Mirror
1

Hi All,

I tried to heat up the nickel surface by using Nose-Hoover thermostat, and then turn off the thermostat and run 3 more ps to ensure that the microcanonical ensemble has reached steady-state. What i have observed that the temperature is fluctuating from 2144 to 2576 K. My target temperature is 2400. I know that we cannot expect constant temperature and some fluctuation is OK, but i think that 400 K is too high. Below one is my code. I do not if there is any way to decrease the fluctuation. Can you please help me?
Thanks,
Rose

units metal
boundary p p p

atom_style atomic
lattice fcc 3.52
region box block 0 5 0 5 0 5
create_box 1 box
create_atoms 1 box

pair_style eam
pair_coeff * * Ni_u3.eam

#Settings

neighbor 0.3 bin
neigh_modify delay 5

region region1 block 0.0 17.6 0.0 17.6 0.0 17.6 units box

# temp controllers

compute newtemp all temp

# equilibrate

velocity all create 2400.0 5812775 temp newtemp

fix 2 all nvt temp 2400.0 2400.0 0.1 tchain 3
fix_modify 2 temp newtemp

thermo 25
timestep 0.001
run 3000

#thermo_modify temp newtemp

thermo_style custom step temp ke pe etotal time vol

timestep 0.001
run 3000

2

Hi All,

I tried to heat up the nickel surface by using Nose-Hoover thermostat, and then turn off the thermostat and run 3 more ps to ensure that the microcanonical ensemble has reached steady-state. What i have observed that the temperature is fluctuating from 2144 to 2576 K. My target temperature is 2400. I know that we cannot expect constant temperature and some fluctuation is OK, but i think that 400 K is too high. Below one is my code. I do not if there is any way to decrease the fluctuation. Can you please help me?

3000 time steps is a horribly short time for equilibration you need to
run for a much longer timer. also, it would be much better to first
use a dissipative thermostat to ensure equipartitioning of the kinetic
energy. a nose-hoover thermostat does not do that, so it can take a
very long time until any fluctuations subside.

it may be beneficial to first run a minimization, to let your system
relax a bit. also, if you increase the size of your system, the
magnitude of the fluctuations will decrease as well.

in general, these are all topics that you should be discussing with
your adviser or whoever is training you in doing MD. these are basic
MD simulation skills.

axel.