Hello all,
I’m working on a system containing epoxy and graphene layers. The simulation box consists of five layers of graphene at the bottom and mixture of epoxy resin and hardener on the top. I performed some MD procedures (based on distance criteria) to crosslink resin and hardener and create the epoxy network. I did all the above-mentioned steps in the Material Studio software. After constructing the epoxy network, I converted the data file, using the msi2lamp function to continue the simulation in LAMMPS.
Now here is the problem:
I need to have (p p s) and (p s s) boundary conditions in my simulations. There is no problem for the (p p s) boundary since there are no bonds in the z-direction of the periodic box. Furthermore, using “change_box scale” and “fix oneway” commands (as Dr. Kohlmeyer explained in one of the mail-list threads) can solve any probable error.
But for the (p s s) boundary condition, since there are some bonds which expand through y-direction of the boundary, I faced errors like missing dihedrals, etc. which stops the simulation. The “change_box scale” and “fix oneway” trick also do not work here!
Is there any way to solve this problem?
Thank you very much for the time and consideration.
Regards,
Abolfazl
Hello all,
I’m working on a system containing epoxy and graphene layers. The simulation box consists of five layers of graphene at the bottom and mixture of epoxy resin and hardener on the top. I performed some MD procedures (based on distance criteria) to crosslink resin and hardener and create the epoxy network. I did all the above-mentioned steps in the Material Studio software. After constructing the epoxy network, I converted the data file, using the msi2lamp function to continue the simulation in LAMMPS.
Now here is the problem:
I need to have (p p s) and (p s s) boundary conditions in my simulations. There is no problem for the (p p s) boundary since there are no bonds in the z-direction of the periodic box. Furthermore, using “change_box scale” and “fix oneway” commands (as Dr. Kohlmeyer explained in one of the mail-list threads) can solve any probable error.
But for the (p s s) boundary condition, since there are some bonds which expand through y-direction of the boundary, I faced errors like missing dihedrals, etc. which stops the simulation. The “change_box scale” and “fix oneway” trick also do not work here!
Is there any way to solve this problem?
you cannot simply switch from a periodic structure to a non-periodic
structure if you have bonds crossing that periodic boundary. you have
to build a different starting geometry, that is only 1-d periodic in
terms of bonds.
axel.
Thank you very much for the response.
You mean I have to create the data file in a way that I apply 1-d periodic boundary condition from the initial step, before applying crosslinking procedure in the Materials Studio? Since I know, there isn't any option in the materials studio to make one or more boundaries non-periodic!
Furthermore, I have problem in the Image Flags consistency, after converting data to LAMMPS data file! Is it any way to break (and delete) the bonds which have been expanded through the (for example) y-direction?
Regards,
Abolfazl
Thank you very much for the response.
You mean I have to create the data file in a way that I apply 1-d periodic boundary condition from the initial step, before applying crosslinking procedure in the Materials Studio? Since I know, there isn't any option in the materials studio to make one or more boundaries non-periodic!
please read my response again. LAMMPS doesn't care what material's
studio can or cannot do. but it should be *obvious* that in your 3-d
to 2-d case you have no bond connectivity wrapping around periodic
boundaries, so you can "compress" your system and thus remove any
conflicts. in the 2-d to 1-d case, you have to break bonds and remove
atoms or else you cannot "compress" the system. but the best solution
is to create a suitable initial structure. you just have to be a bit
creative, e.g. by including some compound as a "filler" that you then
remove before using the structure in LAMMPS.
Furthermore, I have problem in the Image Flags consistency, after converting data to LAMMPS data file! Is it any way to break (and delete) the bonds which have been expanded through the (for example) y-direction?
what LAMMPS has available as options is documented in the manual. i
don't have the time to check/read it for you.
axel.
Thank you very much Dr. Kohlmeyer. Understood.
I used the same trick for preventing the creation of bonds in z-direction, by employing 5 layers of graphene on the bottom of the cell.
As you mentioned, I need to use the same trick on the other direction too.
For breaking the bonds which have been expanded through the box, I just wanted to know your comments, whether it is ok or not, however, I know that the implementation is not your concern.
Thank you again for the time and considerations.
Regards,
Abolfazl