Dear LAMMPS users,
I tried to simulate water molecule (with density = 1) in a flexible CNT. I supposed the water molecules can freely move in the CNT including both ends, so I applied periodic boundary condition. I didn’t want the CNT move with water molecules so I fixed the C atoms at both ends. I use AIREBO to describe C-C interaction and SPC/E model to build H2O molecules. I set an initial temperature of 300K by creating velocities. My in file is in the attachment
However, I find the temperature went up to 1000K as NVT began. Although the temperature fell down after a while, a vacuum space was created in CNT (no H2O molecules can be found in that part of CNT), which is weird.
Can someone help me? Any suggestion will be appreciated.
Thank you
Sincerely,
Yuan
in.wincnt (1.24 KB)
Dear LAMMPS users,
I tried to simulate water molecule (with density = 1) in a flexible CNT. I
supposed the water molecules can freely move in the CNT including both ends,
so I applied periodic boundary condition. I didn't want the CNT move with
water molecules so I fixed the C atoms at both ends. I use AIREBO to
describe C-C interaction and SPC/E model to build H2O molecules. I set an
initial temperature of 300K by creating velocities. My in file is in the
attachment
However, I find the temperature went up to 1000K as NVT began. Although the
temperature fell down after a while, a vacuum space was created in CNT (no
H2O molecules can be found in that part of CNT), which is weird.
have you visualized the trajectory and seen what causes the high
temperature? i.e. where water molecules move fast? this is typically a
side effect of a high potential energy initial structure. your input
doesn't include a minimization before starting the MD, which can help
a lot to remove such excess potential energy from a system.
once point of concern is your use of
thermo_modify lost ignore
this should not been done for any simulation unless you actually want
to have atoms leave your simulations (which is impossible under
periodic boundary conditions as you have set up, so lost atoms are a
certain sign of a problem).
in general, this is really something that you should discuss with your
adviser. this is primarily about building your model and general best
practices in setting up and running MD simulations. this has very
little to do with LAMMPS specifically.
axel.
Thank you Axel, I’ll have a try