timestep and damping for langevin thermostat

Axel,

I determined that the atoms in the CNT are at the equilibrium distance since
we know what the equilibriation distance between the carbon atoms is for our
potential (ie 1.42 A).
So I created the geometry with this CC bond distance of 1.42 for the sp2 CNT
lattice.

have you actually verified this?

are you sure you have the right units flag?

are you sure there is no typo in the parameters?

have you run a minimization on the geometry?

i am always amazed a the trust people put in their input and assume
that things are how they "should be". i almost never have this trust
knowing that even the most experienced people make typos or overlook
something.

Are you suggesting I do this (ie remove the thermostat):

# Relaxation phase

velocity cnt create 0.5 4928459 rot yes dist gaussian
fix 1 cnt nve

timestep 0.0001
run 50000

And you think the temp should be have the initial temp at the end of the
50000 steps?

that is not what i said.

As you see I give my CNT atoms a velocity consistent with a temp of 0.5 K
and then apply a thermostat to try keep this constant.
Am I missing a step or something?

yes. before applying a thermostat, you have to verify that your MD can
run reliably at the chosen parameters. they only way to confirm this,
is to run *without* a thermostat. energy has to be conserved by that,
and if you start of an optimal structure the equilibrium kinetic
energy should be *half* of what you initially assign, for obvious
reasons.

axel.