SPEADMD, April 2008

J. Richard Elliott

Introduction

This document describes the implementation of molecular dynamics simulation as a standard engineering method for physical property estimation in ChemStations' chemical process simulator software. The success of this molecular model hinges on three fundamental premises. (1) The influence of repulsive forces dominates the physical properties. For example, the intermolecular distributions and their fluctuations are primarily influenced by how closely the atoms can approach each other. As another example, the entanglements that strongly influence transport properties occur because molecules cannot pass through each other, but must find a viable path for wriggling past each other. (2) Repulsive effects are specific to the 3D structure of a molecule, necessitating molecular simulation of that specific molecule. In other words, accurately predicting the intermolecular distributions and their fluctuations from a generalized equation (e.g. Peng-Robinson or SAFT) or from integral equation theory (e.g. PRISM) is not reliable for molecules that may be composed of rings and branches. (3) Thermodynamic Perturbation Theory (TPT) is sufficiently accurate that a quantitative treatment of the attractive details of the potential can be derived from theory. Since the TPT contributions are directly related to the intermolecular distributions and their fluctuations, and these are accurately determined by the repulsive forces, this means there is no need to repeat the simulation for every possible specification of the attractive part of the potential. The parameterization of the attractive part of the potential can therefore be pursued in the manner of an engineering equation of state. The development of a prototype is progressing continuously. A preliminary demonstration package is available by clicking the link below.

Download SPEAD.zip Demonstration

We refer to this prototype as "SPEADMD" for Step Potential Equilibria And Discontinuous Molecular Dynamics. The developments to date can be best understood by executing a brief demonstration (~10 minutes). The demonstration is divided into four parts: (1) the graphical user interface for defining the ".m3d" file, (2) the initialization interface for specifying the potential model (reference, linear, square well, Yukawa, or Lennard-Jones) and configuring the positions and velocities of all atoms at all densities (3) the simulate interface for performing the molecular simulations, and (4) the analyze interface for translating the simulation results into a customized equation of state and analyzing phase equilibria and transport properties. Instructions for conducting the demonstration are given in the help menus. Just select the “Prepare” menu to get started.

This material is based upon work supported by the National Science Foundation under Grant No. 0226532. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.

Detailed descriptions of the methodology are described in the literature. An overview is given in SpeadIntro. A brief bibliography is given below. These articles review the coverage of molecular types including: n-alkanes, branched alkanes, alkenes, alkynes, aromatics, naphthenics, alcohols, amines, amides, esters, nitrates, phosphates, halocarbons, perfluorocarbons, acids, sulfides, and thiophenes.

The demonstration shows how a single brief simulation is conducted. We have performed many of these simulations to characterize the molecular interactions. Our results for vapor pressure (~10% average absolute deviation, %AAD) are superior to any method known to us. The temperature range over which our results are applicable is important. In general, DMD/TPT is accurate to reduced temperatures of 0.45 whereas other united atom models are inaccurate below reduced temperatures of 0.6. Typical engineering equations of state like the Peng-Robinson equation are accurate to reduced temperatures of 0.45, so it is important to extend to lower reduced temperatures. The group contribution methods that we know are focused on the low pressure ranges (10-760 mmHg) and their accuracy at higher pressures is less known. For liquid density, our accuracy is roughly 3%AAD. This compares to ~1.5%AAD for other molecular simulation models. We are working to improve our accuracy for density. For diffusivity, our accuracy at present is ~25%AAD. We are at an early stage of correlating transport properties for engineering applications.

Reference:

1. ‘Inferring Transferable Potential Models,’ Sinan Ucyigitler, Mehmet C. Camurdan, Metin Turkay, J. Richard Elliott, Molecular Simulation., in press (2008).

2. ‘Butadiene Purification Using Polar Solvents. Analysis of Mixture Nonideality Using Data and Estimation Methods,’ Paul M. Mathias, J. Richard Elliott, Andreas Klamt, Ind. Eng. Chem. Res., in press (2008).

3. ‘Transferable Potentials for Perfluorinated Molecules,’ Amanda D. Sans, J. Richard Elliott, Fluid Phase Equilibria, 263:182-189 (2008).

4. ‘Combining Molecular Dynamics and Chemical Process Simulation: The SPEAD Model’ AsiaPacific J. Chem. Eng., 2:257-271 (2007).

5. ‘Transferable Potentials for Alcohol-Amine Interactions,’ J. Richard Elliott, Amir Vahid, Amanda D. Sans, Fluid Phase Equilibria, 256:4-13 (2007).

6. ‘Molecular dynamic simulations and global equation of state of square-well fluids with well-widths from 1.1 to 2.1,’ Sergei B. Kiselev, James F. Ely, J. Richard Elliott, Mol. Phys, 104:2545-2559 (2006).

7. ‘Asymptotic Trends in Thermodynamic Perturbation Theory,’ J. Richard Elliott and Neil H. Gray, J. Chem. Phys, 123:184902 (2005).

8. ‘Transferable Step Potentials for Amines, Amides, Acetates, and Ketones,’ Suhan Baskaya, Neil Gray, Z. Nevin Gerek, and J. Richard Elliott, Fluid Phase Equilibria, 236:42-52 (2005).

9. ‘Molecular Modeling of Isomer Effects in Naphthenic and Aromatic Hydrocarbons,’ Neil Gray, Z. Nevin Gerek, and J. Richard Elliott, Fluid Phase Equilibria, 228-229C, 147-153 (2005).

10. ‘Transferable Step Potentials for the Straight Chain Alkanes, Alkenes, Alkynes, Ethers, and Alcohols,’ Ozlem Unlu, Neil Gray, Z. Nevin Gerek, and J. Richard Elliott, Ind. Eng. Chem. Res., 43:1788-1793 (2004).

11. ‘Phase Diagrams for Multi-step Potential Models of n-Alkanes by DMD/TPT,' J. Cui and J.R. Elliott, Jr., J. Chem. Phys., 116:8625 (2002).

12. ‘Optimized Step Potential Models for n-Alkanes and Benzene,' J.R. Elliott, Jr., Fluid Phase Equilibria, 194:161 (2002).

13. ‘Phase Envelopes For Variable Width Square Well Chain Fluids,' J. Cui and J.R. Elliott, Jr., J. Chem. Phys., 114:7283 (2001).

14. ‘Vapor Liquid Equilibria of Square-Well Chains,' L. Hu, H. Rangwalla, J. Cui, J.R. Elliott, Jr., J. Chem. Phys., 111:1293 (1999).

15. ‘Vapor Liquid Equilibria of Square-Well Spheres,' J.R. Elliott, Jr. and L. Hu, J. Chem. Phys., 110:3043 (1999).