Monte Carlo simulations with LAMMPS

LAMMPS includes a variety of different fix styles that augment a normal molecular dynamics simulation with occasional discrete random changes, such as adding or removing particles. These allow the user to quickly evolve the system in ways that would be either forbidden or prohibitively slow with molecular dynamics alone. One downside of these fixes is the risk of generating far-from-equilibrium configurations of atoms that can cause numerical instabilities and large systematic errors in calculated properties. Two of these fixes (gcmc and swap) use Monte Carlo sampling of the Boltzmann energy distribution at the specified temperature, consistent with the canonical ensemble. Both of these fixes also allow atoms or molecules to be randomly added and removed from the system with probabilities defined by specified chemical potentials, consistent with the grand canonical and semigrand ensembles. In this talk, I will review the different ways that these fixes can used, and I will present results from several representative examples.