Pressure of a liquid enclosed by solid crystalls

Dear All

I’m working with a liquid placed on a solid crystal in which latter substance is totally immobile. I’d like setting pressure of the liquid to achieve density equal to experimental value.

I’ve already searched the mailing list and find out that fix NPT may not work fine while having a group of immobile atoms.

So, could anybody please hint on this issue?

Many thanks in advances.

There’s an example how to do this for (a pair of) rigid walls in contact eith a liquid located with the moltemplate examples, here:

http://moltemplate.org/examples/translocation/run.in.npt

(You don’t have to be using moltemplate to follow the instructions in this example.)

There’s a movie of this system here:
https://m.youtube.com/watch?v=hxAufspeg2s

…and some other related files here:
http://moltemplate.org/visual_examples.html#translocation

…and there’s a similar file for the “nanotube+water” example at:
http://moltemplate.org/examples/nanotube+water/run.in.npt

http://moltemplate.org/visual_examples.html#nanotube+water

These examples may or may not be optimal.

I remember when we debated how to run these simulations (back in 2011), it was not clear whether this was the best approach to use at high pressures. However the tests I ran (in 2011) showed that the approach used in the example seemed to work the best at low pressures (<500 bar). (One issue to consider is that the thermostat sometimes gets confused when you freeze a large percentage of the atoms in your system, but there’s a way of dealing with this. See the run.in.npt example file.)

Either way, you will have to test the behavior with your system carefully.

Here’s a link to the LAMMPS discussions we had back then:
https://sourceforge.net/p/lammps/mailman/message/32396301/

https://sourceforge.net/p/lammps/mailman/message/28151357/

2011 was a long time ago and LAMMPS may have changed since then. If anyone else has experience with this, feel free to post.

Andrew