Temperature profile of the two phases system in NVT

Dear Lammps users,

I want to output the temperature profile of a two phases system (along the z axis) .

The simulation system consists of three parts: gaseous argon, liquid argon and metal platinum surface. argon with liquid and vapor phases are placed on the platinum surface.

The simulation process includes two stages: preparation and data acquisition. In the preparation stage, 40 ns simulation is performed to achieve equilibrium at the temperature of 90 K, where both the solid platinum and liquid argon atoms are coupled with the Nosé–Hoover thermostat. The boundary conditions: P P F ( Z axis: reflective wall )

I use the command as:

velocity all create 90 4928459 dist gaussian

Fix argon argon nvt temp 90 90 100

Fix platinum platinum nvt temp 90 90 100

output temperature (sample):

region liquid1 block 0 240 0 7.38 4 8

group li1 dynamic argon region liquid1 every 1

compute liquid1 li1 temp/com

region gas1 block 0 240 0 7.38 40 90

group g1 dynamic argon region gas1 every 1

compute gas1 g1 temp/com

fix 1 argon ave/time 1 100000 100000 c_1 c_2 file tem_profile.average ave running

However, during the equilibrium stage, the temperature of the liquid argon kept decreasing (less than 88K) while the temperature of the vapor phase argon kept increasing (more than 162K). The average temperature of the argon kept constant (AROUND 90k).

My question is why the temperature of the gas and liquid are not constant?

I would very appreciate if anyone could give me any advice.

Best regards,


temperature profile.txt (34.3 KB)

What units are you using and do your results depend on the relexation

time you choose for the thermostats? Have you tried a Langevin

thermostat? Have you tried setting the thermostat target

to different temperatures than your initialization and

seeing how if and how quickly the system responds?

In other words, expt with your protocol, to see where the

problem lies.