Dear Alex,
Regarding your first part of you response, the question I brought up in my
last email had been asked a long time ago by others. Some suggestions were
provided at that time. However, I thought that things have been changed and
there is possibility that if I ask this question again, LAMMPS users and
even LAMMPS developers would like to contribute their version of REBO
potential for C/H/O system. After all, there is a good numbers of systems
which contain C, H, and O atoms and it would be very beneficial if a
forcefield for this kind of system become available to LAMMPS community.
this kind of reasoning always makes me angry, because it is at its
core dishonest and disrespectful. while there may be others that would
benefit, it is primarily *you* that would benefit the most. ...and
massively so, because if it wasn't so difficult to you, you would have
done it already, right? how do you know, it is not difficult and a lot
of effort for *anybody*?
thus with this kind of statement you a) claim that what is primarily
your personal benefit is a large benefit for the community (and the
LAMMPS community is larger and more diverse than what you seem to
think, so the benefit is definitely only for a subset of the LAMMPS
community) and b) imply that it is not worth spending that much of
your own time to learn and figure out how to do this kind of work
yourself. now, i grant you that in this particular case, the effort is
objectively large for somebody with little experience, but that
doesn't completely compensate the implied disrespect of the work of
others.
so before asking the community to work for *your* benefit, you should
first ask yourself what can *you* contribute that is of equal or
greater value to the community. does this statement remind you of
something? well, that is not a coincidence. 
Regarding the second part of your suggestion, I did minimize my system.
After the minimization, the system was equilibrated using NPT time
integrator. I received the following error in the first initial steps of
equilibration.
ERROR on proc 0: Non-numeric box dimensions - simulation unstable
(../pppm.cpp:1882)
I would like to find a clue as to what is wrong with either my simulation or
atomic structure.
impossible to say from remote, but it looks highly suspicious. could
be a problem with parameters or the topology or the geometry or a
combination of those. it definitely looks like something somewhere
between FUBAR and SNAFU is happening between minimization step 150 and
200.
have you visualized what is going on? that might provide a hint.
also, you seem to have a complex system with lots of components. why
not start with something simpler and check that out by itself and then
gradually make it more complex again until you see which component is
giving you troubles.
axel.