I am working on a preparation of representative polymeric structures of some popular polymers, such as polyimide or PTFE for MD simulation, during which a single nitrogen or oxygen atom will be shot towards the surface.
I am preparing the polymeric cubes in the following way: 1) I pack a number of polymeric chains into a given volume to obtain a loose configuration (density of around 0.2 g/cm2), 2) I add a geometry restraint (an elastic wall) a few angstroms above the surface 3) I set the volume regime to decrease of -0.00001 in the a and b directions 4) I run a NVT simulation during which my system gets compressed to a more realistic densities, e.g. 1.4 g/cm3 for polyimide or 2.2 g/cm3 for PTFE.
My problem is that I finish the simulation with a lot of residual stress, the pressures recorded at the end of the simulations can be around 2000MPa, this does not allow me to run another simulation on the compressed structure (the atoms shooting part), because the cube just explodes. Could you suggest a way of post-processing to remove this residual stress? or maybe another approach on how to produce a realistic densities for polymers (to pack a volume to 1.4 g/cm, many polymer chains need to be added, which results in a long computation, often ended with an error that a solution could not be found). I am using the SCM software for my simulations. Thanks!
Why cant you run and NPT simulation to get the realistic density ?
Hi,
Have you tried to move the elastic wall away? A few Ang of separation seems too little, especially if you compress your system along a and b. I suspect that you get high stresses because of that. Also, have you checked for chain overlaps?
You could run a simulation with a box periodic along a and b, and a large (nonperiodic) c. You fill half of the box with your polymers and then run a simulation barostatting along a and b.
- Badges
- Science topic