[lammps-users] npt simulation with a high aspect ratio box?

LAMMPS LAMMPS Mailing List Mirror
1

Hello Lammps users,

I am trying to set up a Muller-Plathe (MP) thermal conductivity calculation of an alkane and need a periodic box with a high aspect ratio. I want to equilibrate the system using npt at 1 atm first before running the MP method but am having some trouble with the behavior of the liquid. If I run npt with the iso option, the liquid cavitates (forms a liquid-vapor interface) along the long axis of the periodic cell. I think this is because the iso command treats each side of the box equally and the long dimension of the box can’t adjust enough to compensate. I guess potentially you could have a negative pressure in one direction and positive pressure in the other two and still have an average positive pressure. Is this correct?

I came up with a solution by alternating between controlling the pressure in the long direction and then the two transverse directions, like so:

fix 1 all npt temp 300. 300. 0.1 x 1.013 1.013 0.1
run 10000
unfix 1
fix 1 all npt temp 300.0 300.0 0.1 y 1.013 1.013 0.1 z 1.013 1.013 0.1 couple yz
run 10000
unfix 1
repeat a few times.

I was just wondering if there was a less clumsy or better a way to do this?

Thanks,

Dave

2

Hello Lammps users,

I am trying to set up a Muller-Plathe (MP) thermal conductivity
calculation of an alkane and need a periodic box with a high aspect
ratio. I want to equilibrate the system using npt at 1 atm first
before running the MP method but am having some trouble with the
behavior of the liquid. If I run npt with the iso option, the liquid
cavitates (forms a liquid-vapor interface) along the long axis of the
periodic cell. I think this is because the iso command treats each
side of the box equally and the long dimension of the box can't adjust
enough to compensate. I guess potentially you could have a negative
pressure in one direction and positive pressure in the other two and
still have an average positive pressure. Is this correct?

I came up with a solution by alternating between controlling the
pressure in the long direction and then the two transverse directions,
like so:

fix 1 all npt temp 300. 300. 0.1 x 1.013 1.013 0.1
run 10000
unfix 1
fix 1 all npt temp 300.0 300.0 0.1 y 1.013 1.013 0.1 z
1.013 1.013 0.1 couple yz
run 10000
unfix 1
repeat a few times.

I was just wondering if there was a less clumsy or better a way to do
this?

how about (pre-)equilibrating a more cubic box first
and then using replicate to turn it into a final
aspect ratio and finish equilibration?

axel.

3

That was my second thought actually. I was just wondering if there was some method within lammps I was missing.

Thanks for the advice.

Dave

4

Would fix npt aniso work any better, i.e. letting the 3 dims be independent,
but all with the same target pressure?

Steve