Compress the system to a desired density

LAMMPS
,
1

I am trying to compress a binnary Lennard-Jones fluid to a higher density and see how the particles behave if they suffer from a sudden volume change.

I am planing to use fix deform to compress the box at different rates to different densities. I am planning to use fix nvt for compressing the box and the use fix nve for equilibration. Is this resonable?

However, I also noticed somewhere saying
If you were to immediately scale the system to the target density, the particle’s hard cores would overlap and the integrator would be become unstable. A more effective strategy gradually resizes the simulation box a little at a time over the course of a simulation, allowing any slight overlaps to relax..
Then how can I achieve a quick big volume change?

2

That is mostly a question of the timestep. If you deform your system quickly that has the same effect as if the potential was extremely steep and in those cases you need to use a far shorter timestep so that that atoms can adjust to the change. They will also run faster and that also requires a higher resolution for the discretization of the time integration of the trajectories.

3

I see. Thank you so much!