(379d) Combined Use of Molecular Dynamics and Monte Carlo for the Prediction of Thermodynamic Properties

The industrial importance of a wide variety of compounds (from simple molecules to more complex structures) and the constant industrial need for synthesizing new and improved materials to be used in a range of different pressure and temperature conditions raises the need for accurate, efficient, systematic and rapid property prediction. Current advances in computational power, molecular simulation methodologies and forcefields (all atom and united atom ) have rendered the use of molecular simulation as an increasingly powerful tool for the prediction of the thermodynamic properties of materials.

Since combined application of molecular dynamics (MD) and Monte Carlo (MC) techniques to compute a range of thermodynamic properties remains relatively unexplored (especially using the same force field), we have used these methods for the calculation of thermodynamic properties of pure compounds ranging from hydrocarbons (alkanes, olefins, aromatic hydrocarbons) to oxygenated molecules (alcohols, aldehydes, ketones, ethers), amines, amides and halogens. We accordingly demonstrate the use of the LAMMPS MD [1] and Gibbs MC [2] simulation programs in a complementary manner to probe a wide variety of properties; some can be uniquely calculated from MD (transport properties like viscosity, diffusion coefficient, thermal conductivity), while others can be uniquely calculated from MC (boiling point and critical point temperatures, chemical potential) and certain properties can be calculated using both techniques (pressure, density, heat of vaporization, specific heat, joule-thomson coefficient). Comparison of the calculated properties with experimental data demonstrates the efficiency of the combined use of MD and MC to characterize pure compounds. Our study illustrates the level of accuracy that is achievable with all-atom (extended PCFF [3], OPLS-AA [4]) and united-atom (TraPPE [5], AUA [6]) forcefields, for each property of interest. For certain functional groups, such as brominated aromatic compounds, it was necessary to re-paramaterize existing force fields; accordingly we will present a detailed description of improved new parameters and property predictions for these materials.

The ability of property prediction for pure compounds is also a necessary step towards simulation and property prediction for binary and multi-component mixtures. Test cases demonstrate that treatment of mixtures via molecular simulation is an effective means of study of the properties of the mixtures and also provides insight to the microscopic details that lead to the macroscopic behavior of such systems.

[1] S. Plimpton, J .Comp. Phys., 117, 1 (1995).
[2] MedeA-Gibbs v.2.6, Materials Design (Gibbs License IFP Energies Nouvelles-CNRS-Université Paris-Sud)
[3] H. Sun, S.J. Mumby, J.R. Maple and A.T. Hagler, J. Am. Chem. Soc.. 116, 2978 (1994)
[4] W.L. Jorgensen, D.S. Maxwell and J. Tirado-Rives, JACS, 118, 11225-11236 (1996)
[5] J.M. Stubbs, J.J. Potoff and J.I. Siepmann, J. Phys. Chem. B, 108, 17596-17605 (2004) and references therein
[6]  N. Ferrando, V. Lachet, J-M. Teuler and A. Boutin, J. Phys. Chem. B, 113, 5985-5995 (2009); J. Perez-Pellitero, PhD Thesis, ISBN: 978-84-691-0377-7/DL: T.2195-2007 and references therein