Creating Initial Configuration

Could you elaborate on the ‘having an understanding of the molecular model’ part? What information do I need to know, other than details of the molecule I want to model? Specifically, is there something be knowing?

I do know how I want my system to be. In a polystyrene molecule, there are carbons and hydrogens. I want to represent each carbon+hydrogen attached to be one sphere. I want the various CH groups part of the benzene ring to be along the same plane. LAMMPS input file for this section should list these are part of improper.
Such details, how do I generate? This is what I am asking

Regards,
Brahm Prakash
NIT Trichy

Could you elaborate on the 'having an understanding of the molecular
model' part? What information do I need to know, other than details of the
molecule I want to model?

1) atom positions.
2) All of the bond, angle, and dihedral angle lists.

Specifically, is there something be knowing?
I do know how I want my system to be. In a polystyrene molecule, there
are carbons and hydrogens. I want to represent each carbon+hydrogen
attached to be one sphere. I want the various CH groups part of the benzene
ring to be along the same plane. LAMMPS input file for this section should
list these are part of improper.
Such details, how do I generate?

I may add this-

1) Create the initial chain of your system using MC method- random walks.
2) You can use any of the known crystal structures(fcc,bcc etc.--> That's
up to you to decide) super-imposing your simulation cell.
3) Specify a suitable nearest neighbor distance.
4) Now add the molecules to your structure in a random(probabilistic)
manner, taking into account the preferred direction of chain growth based
on the angles and the density of sites which are un-occupied.
5) Now convert this into LAMMPS readable format of the data file and
minimize the structure(use a suitable potential, depending upon what you
want to do).

Sagar

This is what I am asking

Dear Sagar,

This is precisely the information I need, and do not know how to generate. As to your suggestions below -

This I have done, created initial chain using MC - Random Walk method

I have used a fcc.

I have done this too (for creating the chain, I used a nearest neighbor distance.

This is a point I do not understand. My monomer polystyrene molecule, will have certain coordinates to each of it’s atoms. Say I do generate these using Avogadro ( and get initial coordinates of each atom, of the polystyrene from the file generated).

Now how do I use this information, to actually place the molecules, inside the points I do have? How to take care of orientation of the molecules, so that each forms bonds with the next? The carbon that forms the bonds is sp3 hybridized, after bond formation. This is what I do not understand.

If above is done, I can write angles and bonds informations to LAMMPS formats. Still I will face trouble defining dihedrals and impropers, but that comes later.

Dear Sagar,

Could you elaborate on the 'having an understanding of the molecular
model' part? What information do I need to know, other than details of the
molecule I want to model?

1) atom positions.
2) All of the bond, angle, and dihedral angle lists

This is precisely the information I need, and do not know how to generate.
As to your suggestions below -

Specifically, is there something be knowing?

I do know how I want my system to be. In a polystyrene molecule, there
are carbons and hydrogens. I want to represent each carbon+hydrogen
attached to be one sphere. I want the various CH groups part of the benzene
ring to be along the same plane. LAMMPS input file for this section should
list these are part of improper.
Such details, how do I generate?

I may add this-

1) Create the initial chain of your system using MC method- random walks.

This I have done, created initial chain using MC - Random Walk method

  2) You can use any of the known crystal structures(fcc,bcc etc.-->
That's up to you to decide) super-imposing your simulation cell.

I have used a fcc.

3) Specify a suitable nearest neighbor distance.

I have done this too (for creating the chain, I used a nearest neighbor
distance.

4) Now add the molecules to your structure in a random(probabilistic)
manner, taking into account the preferred direction of chain growth based
on the angles and the density of sites which are un-occupied.

This is a point I do not understand. My monomer polystyrene molecule, will
have certain coordinates to each of it's atoms. Say I do generate these
using Avogadro ( and get initial coordinates of each atom, of the
polystyrene from the file generated).

Now how do I use this information, to actually place the molecules, inside
the points I do have? How to take care of orientation of the molecules, so
that each forms bonds with the next? The carbon that forms the bonds is sp3
hybridized, after bond formation. This is what I do not understand.

Please read again the comments from me and Axel. replicate the monomer in
the desired orientation, what else. You need to do more literature survey
regarding this.

Sagar

Alas, I don't have the time to read all of the heated messages in this
thread. I just wanted to add that the "polymer melt" problem has come
up a number of times on the LAMMPS mailing list:

http://lammps.sandia.gov/threads/msg34577.html
http://lammps.sandia.gov/threads/msg34585.html
http://sourceforge.net/mailarchive/message.php?msg_id=30024094
http://sourceforge.net/mailarchive/message.php?msg_id=30018399
http://sourceforge.net/mailarchive/message.php?msg_id=30018436

It is not trivial to build these kinds of systems. I don't know of a
preprocessor which can automatically create a polymer melt for you.
(Certainly moltemplate was not designed to do this.) If your monomers
are simple point-like particles, you may be able to take advantage of
lattice methods, such as these:
http://jcp.aip.org/resource/1/jcpsa6/v77/i3/p1554_s1
http://jcp.aip.org/resource/1/jcpsa6/v133/i4/p044903_s1
Otherwise, my understanding was that people usually run some kind of
simulation to obtain reasonable well-packed, well-equilibrated polymer
melts. In any case, my impression is that a good polymer melt is the
end result of running a careful simulation. This is true regardless
of whatever molecule builder you use to create the initial
coordinates.

Andrew
If the monomers are simple enough may be able to accelerate your
initial simulation by allow the particles to pass through each-other.
(However if your monomer has benzene rings, or other loops, this won't
work because they could end up linked together.) You also have to be
a little bit careful because I think LAMMPS might have a limit to the
number of atoms which lie within the cutoff distance of another atom
(neighbors).