Possibility of constant pressure and energy with two non-periodic boundaries

LAMMPS LAMMPS Mailing List Mirror
1

Hello,

I am working on a simulation of a small cylinder of silicon (I’m essentially creating a silicon nanotube).

Due to the nature of the cylinder, I have been asked to create a simulation where the dimensions of the cylinder are non periodic. It’s also necessary for the pressure and energy of the cylinder to be constant.

I have tried an nph, as well as an nve simulation with a wall/reflect fix, and neither have been the complete solution that I’m looking for. I have been through the command document and the LAMMPS forum, and I have not found a complete answer.

I appreciate any guidance or suggestions that you have.
Thank you very much for your time.
Casey

Casey Hansen
casey.e.hansen@…24…

2

Hello,

I am working on a simulation of a small cylinder of silicon (I'm essentially
creating a silicon nanotube).

Due to the nature of the cylinder, I have been asked to create a simulation

it not a matter of asking, but it is simply not possible to do this
differently with lammps.

where the dimensions of the cylinder are non periodic. It's also necessary

what do you mean by "dimensions of the cylinder". the entire cylinder
or only the outer wall.

for the pressure and energy of the cylinder to be constant.

what "pressure". for a non-homogeneous object, there is not "the
pressure". it only makes sense to look at components in different
directions.

I have tried an nph, as well as an nve simulation with a wall/reflect fix,
and neither have been the complete solution that I'm looking for. I have
been through the command document and the LAMMPS forum, and I have not found
a complete answer.

unless you explain, *why* you could not get what you want and what
*exactly* you tried, there is little chance to go beyond what you
already have looked at.

if the cylinder is of finitely size, why do you need to do a
"constant" pressure in the first place, it will just relax to where it
wants to be and then it is done.

I appreciate any guidance or suggestions that you have.

before anybody can make more specific suggestions, you have to be more
specific about what your goals are and the motivations behind the
choices.

axel.

3

Thank you very much for your response. I’m sorry my question was not detailed enough.

The axis of the cylinder is along the x-axis. The y- and z-axes (the circumference) of the cylinder are non periodic, but the faces of the cylinder may be periodic, if I am considering an infinitely long cylinder.

The pressure that I am looking to keep constant is that of the silicon atoms on the inside “walls” of the cylinder.

The theory behind the simulation is that eventually the volume will oscillate around an average value. I ran an nph simulation with a barostat, and the resultant volume vs. time graph was a lovely exponential decay, easily fitted. However, I had the boundaries set to be periodic, and I did not fix the circumference of the cylinder, which is necessary for a nanotube simulation.
With the nve simulation, I was able to have the correct boundary and constraints, but the resulting volume is monotonically increasing, and I have lost the fixed pressure.

Again, thank you for your response, and for your time.
Casey

4

Thank you very much for your response. I'm sorry my question was not
detailed enough.

The axis of the cylinder is along the x-axis. The y- and z-axes (the
circumference) of the cylinder are non periodic, but the faces of the
cylinder may be periodic, if I am considering an infinitely long cylinder.

The pressure that I am looking to keep constant is that of the silicon atoms
on the inside "walls" of the cylinder.

i don't understand this. if this is supposed to be a nano object,
there is no physical reason to limit its expansion or contraction. it
will equilibrate and that is that.

The theory behind the simulation is that eventually the volume will
oscillate around an average value. I ran an nph simulation with a barostat,
and the resultant volume vs. time graph was a lovely exponential decay,
easily fitted. However, I had the boundaries set to be periodic, and I did

periodicity would not be a problem for as long as there is sufficient
distance between the images of the cylinders, so they don't interact.

not fix the circumference of the cylinder, which is necessary for a nanotube
simulation.

here i disagree completely. why should there be a need to immobilize
the circumference of that cylinder. if that object cannot maintain its
structure, then you have a bigger problem. forcing it to be in a
specific geometry will only hide that problem.

With the nve simulation, I was able to have the correct boundary and
constraints, but the resulting volume is monotonically increasing, and I
have lost the fixed pressure.

i don't understand what you are trying to say here at all. again, this
hints at something else being out of whack. e.g. a bad structure, a
bad choice of potential or simulation parameters.

and overall, this is still far to vague for anything more specific.

axel.

5

The LAMMPS simulation box can be periodic
is x, but not in y,z, if you specify the boundary
command as p s s.

Axel is correct, the cylinder will simply expand to what
it wants to be in y,z.

Steve