Gcmc in lammps and raspa

LAMMPS LAMMPS General Discussion
1

Dear all,

I have recently started Monte Carlo (MC) simulation, and am confused by the difference between RASPA and LAMMPS for running MC.

As we all know, RASPA was developed specially for MC simulation (just my poor understanding is that RASPA is stronger than LAMMPS for MC), and LAMMPS also supports MC. I wanted to know the main difference between the two programs, or what the limitation is for LAMMPS to run MC, and under what circumstances it can be used for MC.

For example, if I want to simulate the water adsorption in an empty and flexible MOF with the polarizable force field set for water molecules (or a polarizable force field for both water and framework), and the number of water molecules will be accounted for. Also need to simulate the process of the first water molecule being adsorbed by the open metal site. Then use MD to monitor the water behaviour inside the framework. Is it possible to use LAMMPS to simulate this case?

Thanks,

wang

2

I cannot talk about RASPA, you have to contact the RASPA developers for that.

What I can tell you about MC support in LAMMPS is that its parallelization scheme with domain decomposition is a bad match with MC moves as they would require multiple communications except for simple force fields that have only local changes for individual atoms (like lj/cut). That makes MC moves very inefficient and they are usually only applied in hybrid MD/MC applications where you have 100s or 1000s of MD steps before doing the next MC step.

Polarizable force fields is a completely different can of worms. LAMMPS supports Drude oscillators, Core-shell and charge equilibration. They are all advanced techniques and require that you know LAMMPS well enough with non-polarizable force fields first.