I am using eDPD model to study heat transfer from CNT to water. CNT has its hexagonal lattice shape, and since the size of the beads for water and carbon should be the same, water beads may pass through the hexagonals. (I tried to build a fake CNT with a different lattice shape which exploded in the simulation). My question is, is there any suggestions from lammps users, how to overcome this problem? is there any command the forbids the liquid beads passing through the material if the material has holes in it?
It is not so much the “holes”, but it is the nature of DPD models that they are just an approximation of a particle density less than a particles, so DPD atoms/molecules can actually move “through” each other, while LJ based potentials are far too repulsive.
LAMMPS has pair style srp to prevent polymer chains from crossing through each other, but I am not certain how well you can apply this to your CNT: pair_style srp command — LAMMPS documentation
You could probably use pair style hybrid/overlay to add a more repulsive interaction between the CNT atoms and the water, e.g. using pair style harmonic/cut, but the challenge for that will be to find a suitable parameterization to represents the material properly.
The other alternative would be to avoid DPD and just run all-atom or MARTINI/SPICA style coarse grain simulations (i.e. with more atoms requiring more CPUs, but causing less troubles and using well understood and tested force field parameters).