Help in molecule flow simulation

User discussions
1

GROMACS version: 2024
GROMACS modification: No

I need to perform a simulation on a certain surface A then there will be a continuous and replenished source of fluid molecules (just few). May Idea is construct a very large box then fill it with the molecule, but this is close to impossible/outrageous as it might require about 0.1mm box size on one dimension which is asking more or less a petabyte of ram!

So is there a way to have this continuous and replenished source of fluid molecules in a 100 nm box apart from using a ridiculously larger box?

I was thinking in the line of splitting the simulation into parts such that when run 1 has consumed the fluids, perform a gmx insert molecule to replenish the space vacated by fluid molecules from run 1, then merge the trajectory via gmx trjcat. But trjcat does not seem to do this for a new system? It might also be wrong to use the continuation function with cpi file as the have an additional molecules for run 2.

Any help?

2

Does the new liquid have to be “purified” (e.g. solute molecules absent at start)? If not, there shouldn’t be any need to replenish the liquid, in periodic boundary conditions all you need is to apply directed force to solvent molecules in the center while fixing the surface, and over time viscosity should produce an equilibrium parabolic velocity profile (that’s at least how I imagine it - you might need to e.g. turn off COM motion removal).

I have never tried this in practice, but it’s possible that some variation of constant-force pulling will work. I’d define the pulling group of solvent molecules as a slab, and restrain them along one dimension to avoid mixing with other solvent molecules.