Dear all,
I want to equilibrate a population (1000) of CH3 radicals using nvt fix
in order to generate a list of thermal initial conditions of CH3 for
subsequent use in MD calculation of interaction with diamond substrate.
I chosed the REBO potential for C and H interactions.
I had two problems to solve.
First, to prevent atoms of the CH3 from crossing the boundaries, thus
giving inconsistent spatial configuration of CH3, I chosed the
reflective boundary which keeps integrity of all CH3s.
It works but I would like to know if there is a possibility in Lammps
when using periodic boundaries to keep the spatial consistence of the
CH3 when writing spatial position in dump file ?
I am not sure why you want to prevent CH3 molecules from crossing the
periodic boundaries. Are you trying to avoid this in order to make it
easier to make pictures of your system? If so, then using reflective
boundaries is not necessary.
I have been using this command to write my dump files:
dump 1 all custom 500 dumpfile.lammpstrj id mol type x y z ix iy iz
The dump file has 9 columns..The x y z columns (4th,5th,6th columns)
contain atom coordinates in real units (not scaled coordinates). If I
remember correctly, each atom's coordinates lies in the periodic
boundary box (each x coordinate is between xlo and xhi, for example).
The ix,iy,iz "image" columns (7th,8th,9th) columns indicate in which
unit cell the atom is located. If you don't care about this
information, ten you can ignore these columns. (However keep in mind
that the x y z coordinates for atoms in that molecule will be split
discontinuously at the boundary. Of coarse, this does not effect the
simulation. LAMMPS will look for the minimum image distance between
atom pairs.)
If you are using VMD to view the trajectory (see attached
instructions), you can use this command to unwrap the coordinates of
each molecule
pbc wrap -compound res -all
This will move the center of each molecule into the main unit cell
(without splitting molecules in half). Use the dump command above to
save your trajectory in that format (containing the "mol" column).
This is very useful for making pictures.
Second, as the REBO potential is a reactive one, during equilibration,
as expected, the CH3 react together, forming C2H6, which is unwanted. To
solve the problem, the initial distance between CH3 was chosen high
enough to prevent the CH3 from meeting (or at least in order the
probability of meeting is as low as possible). The question is : is
there a possibility to prevent formation of the C-C bond keeping
reactivity for C-H bond ?
I am unfamiliar with the details of the REBO potential, but during the
initial equilibration, you can use artificial repulsive Lennard-Jones
interactions between carbon atoms to prevent them from getting too
close together.
http://lammps.sandia.gov/doc/pair_airebo.html
I imagine that this will effect the resulting structure at the end
of equilibration. But hopefully, in spite of this, the structure
would still be reasonable.
Later on, after equilibration, when you want to enable the reaction
to proceed, you can use the "pair_coeff" input script command to reset
the Lennard-Jones interactions back to their desired values.
Cheers
Andrew
README_visualize.txt (2.42 KB)