Computationally efficient Monte Carlo simulations for polarisable models: multi-particle move method for water and aqueous electrolytes

Abstract

We demonstrate improved computational efficiencies of Monte Carlo (MC) simulations for polarisable force fields by implementing the multi-particle-move MC (MPM-MC) method [Moučka F, Rouha M, Nezbeda I. Efficient multiparticle sampling in Monte Carlo simulations on fluids: application to polarizable models. J Phys Chem. 2007;126:224106]. Force bias and smart alternatives along with the Ewald summation and generalised reaction field have been considered and tested for both pure polarisable water models and polarisable aqueous electrolytes. For water, we consider the recently developed Baranyai–Kiss force field [A transferable classical potential for the water molecule. J Chem Phys. 2010;133:144109] and the SWM4-DP force field of Lamoureux et al. [A simple polarizable model of water based on classical Drude oscillators. J Chem Phys. 2003;119:5185], and for electrolytes the polarisable AH/SWM4-DP force field associated with the SWM4-DP solvent. These force fields incorporate polarisability as a charge-on-spring model, and the Baranyai–Kiss model incorporates charges in the form of a Gaussian distribution instead of point charges. For neat water, in addition to calculating the standard thermodynamic properties and water structure and comparing our results with those obtained from standard molecular dynamics simulations, we also investigate its vapour–liquid equilibrium properties. For aqueous electrolyte solutions, we study NaCl at ambient conditions and calculate its density and chemical potential as a function of concentration, representing apparently the first such simulation results using these polarisable models at finite concentrations. Our results demonstrate that MC simulations can be efficiently performed on polarisable models if the MPM method is incorporated.

Keywords::

• multi-particle move MC
• polarisable water
• polarisable electrolytes
• Baranyai–Kiss model
• vapour–liquid interface

Acknowledgements

Support for this work was provided by the Czech Ministry of Education, Youth and Sport (Project No. LH 12019), by the Czech National Science Foundation (Grant No. 13-35793P/P208) by the Natural Sciences and Engineering Research Council of Canada (Discovery Grant No. OGP1041) and the SHARCNET (Shared Hierarchical Academic Research Computing Network) HPC consortium (http://www.sharcnet.ca).