Is there any way to 'shift' molecules from one part of simulation to another?

I am trying to simulate 2 parallel rigid infinite sheets in a solvent.
Now I want to bring them close together slowly while measuring force
on each ( for PMF). But as solvent between the 2 is 'trapped'. It will
increase the pressure.
Here is what am trying to do currently:
Fix nvt -> move both sheets close to each other by small increment
using displace_atoms -> turn portion of my sheet transparent by
setting LJ terms to 0 0, allowing solvent to pass through and
equilibriate -> turn LJ parameters back again slowly by first giving
small negative values (to repel solvent molecules for ensuring nothing
is trapped in sheet) -> turning LJ parameters back to normal -> rinse
and repeat

Is there more clever/simple way of doing it?

I was hoping something on the lines of : measure pressure on solvent
molecules between sheets and outside it by compute group group etc ->
if both values are not equal move certain molecules to bulk, without
crashing lammps due to overlap -> minimize structure, run and repeat

Use displac atoms command

Hi,

Some of the methods for moving an object depending of your purpose:
fix move
fix spring
fix smd
fix displace
fix addforce

Please check the lammps document for each.

Cheers
X

I am trying to simulate 2 parallel rigid infinite sheets in a solvent.
Now I want to bring them close together slowly while measuring force
on each ( for PMF). But as solvent between the 2 is 'trapped'. It will
increase the pressure.
Here is what am trying to do currently:
Fix nvt -> move both sheets close to each other by small increment
using displace_atoms -> turn portion of my sheet transparent by
setting LJ terms to 0 0, allowing solvent to pass through and
equilibriate -> turn LJ parameters back again slowly by first giving
small negative values (to repel solvent molecules for ensuring nothing
is trapped in sheet) -> turning LJ parameters back to normal -> rinse
and repeat

Is there more clever/simple way of doing it?

i don't think that there will be a clever way at all, since it sounds
like a very flawed model.
when you compute a PMF, you cannot just stop, arbitrarily shuffle some
atoms/molecules around and then continue. what kind of a collective
variable would that be and what would your learn from it?
either your atoms/molecules are trapped due to the infinite
replication or they are not due to a finite size system (then, of
course, the size matters).

I was hoping something on the lines of : measure pressure on solvent
molecules between sheets and outside it by compute group group etc ->
if both values are not equal move certain molecules to bulk, without
crashing lammps due to overlap -> minimize structure, run and repeat

that pressure is the same as the force (divided by the area) that you
are sampling. so manipulating your system to reduce the pressure is in
essence deliberately changing the force. that doesn't seem right to
me.

axel.