Hi all,
I am a very beginner in lammps and I have some basic questions to ask.
1. I am trying to run a simulation on 3 atoms of Argon. Could anybody help
me with what pair_coeff command arguments, value of lj cutoff potential,
temperature and etol ftol values for minimize and neb commands should I use
for proper simulation. I am using real units and it is difficult to analyse
what values of various parameters should be used. Is there any database for
lammps where all this kind of data is available for use as reference? and
what should be a correct set of values for initial atom coordinates?
potential parameters you can find in the published literature and text
books. since doing simulations on argon using a lennard-jones potential is
a common assignment in molecular dynamics classes and tutorials, a simple
web search should find you a suitable collection of known to work
parameters.
temperature is irrelevant for neb, since it is - like all minimization
based methods - operating at zero kelvin.
please familiarize yourself with the fundamental theory about NEB and
physics. again, textbooks and the published literature should be your
primariy sources for that.
I have to run neb calculations for this and some other simple model
systems and I have to cross check the minimum energy path obtained from the
neb output files from the one obtained manually on paper, but I am not
quite clear will the values of different parameters and units I am taking
affect this or not.
since you are new to LAMMPS and since NEB is based on minimizations, you
should *first* learn how to do minimizations properly and make certain that
you understand the meaning of all relevant parameters. it is a bad idea to
start with a complex method before understanding the (simpler) method it
is based on.
2. I saw the example input files of neb and nowhere there was velocity
command used because as far as my understanding goes there should be some
velocity required to be given to all atoms to move them. Is there no need
of it because fix command tells the atoms to move or did I misunderstood
something ?
you misunderstood how energy minimizations work. they do on factor in
thermal effects. if you read up on statistical mechanics, as you should
have done, you should learn, that only at zero kelvin there is a single
point of minimum energy on the potential energy hypersurface. for finite
temperature, there always is an ensemble of configurations at the same
total energy (i.e. state) and thus you have to average over all of them (or
rather a suitably large number of relevant samples) to obtain information.
3. The final coordinates which I specified in my final replica coordinates
file, they do not match with my the atom coordinates in my last dump file.
I think they should match.
no necessarily. they would only match, if both endpoints that you provide
are already at their respective minimum. the current NEB implementation
will by default minimize each end-point to its next (local) minimum. if you
want to immobilize atoms at the endpoints or throughout all replica, you
have to adapt your input file accordingly. please study the LAMMPS
documentation on the neb command and related commands.
axel.