Molecular modeling of hydrate-clathrates via ab initio, cell potential, and dynamic methods

Abstract

High level ab initio quantum mechanical calculations were used to determine the intermolecular potential energy surface between argon and water, corrected for many- body interactions, to predict monovariant and invariant phase equilibria for the argon hydrate and mixed methane-argon hydrate systems. A consistent set of reference parameters for the van der Waals and Platteeuw model, ... and ..., were developed for Structure II hydrates and are not dependent on any fitted parameters. Our previous methane-water ab initio energy surface has been recast onto a site-site potential model that predicts guest occupancy experiments with improved accuracy compared to previous studies. This methane-water potential is verified via ab initio many-body calculations and thus should be generally applicable to dense methane-water systems. New reference parameters, ... and ..., for Structure I hydrates using the van der Waals and Platteeuw model were also determined. Equilibrium predictions with an average absolute deviation of 3.4% for the mixed hydrate of argon and methane were made. These accurate predictions of the mixed hydrate system provide an independent test of the accuracy of the intermolecular potentials.

(cont.) Finally, for the mixed argon-methane hydrate, conditions for structural changes from the Structure I hydrate of methane to the Structure II hydrate of argon were predicted and await experimental confirmation. We present the application of a mathematical method reported earlier' by which the van der Waals-Platteeuw statistical mechanical model with the Lennard-Jones and Devonshire approximation can be posed as an integral equation with the unknown function being the intermolecular potential between the guest molecules and the host molecules. This method allows us to solve for the potential directly for hydrates for which the Langmuir constants are computed, either from experimental data or from ab initio data. Given the assumptions made in the van der Waals-Platteeuw model with the spherical-cell approximation, there are an infinite number of solutions; however, the only solution without cusps is a unique central-well solution in which the potential is at a finite minimum at the center to the cage.

(cont.) From this central-well solution, we have found the potential well depths and volumes of negative energy for sixteen single-component hydrate systems: ethane (C₂H₆), cyclopropane (C₃H₆), methane (CH₄), argon (Ar), and chlorodifluoromethane (R-22) in structure I; and ethane (C₂H₆), cyclopropane (C₃H₆), propane (C₃H₈), isobutane (C₄H₁₀), methane (CH₄), argon (Ar), trichlorofluoromethane (R-1 1), dichlorodifluoromethane (R-12), bromotrifluoromethane (R-1 3B 1), chloroform (CHC1₃), and 1,1,1,2-Tetrafluoroethane (R-134a) in structure II. This method and the calculated cell potentials were validated by predicting existing mixed hydrate phase equilibrium data without any fitting parameters and calculating mixture phase diagrams for methane, ethane, isobutane, and cyclopropane mixtures. Several structural transitions that have been determined experimentally as well as some structural transitions that have not been examined experimentally were also predicted. In the methane-cyclopropane hydrate system, a structural transition from structure I to structure II and back to structure I is predicted to occur outside of the known structure II range for the cyclopropane hydrate.

(cont.) Quintuple (Lw-SI-SII-Lho-V) points have been predicted for the ethane-propane-water (277.3 K, 12.28 bar, and Xeth,waterfree = 0.676) and ethane-isobutane-water (274.7 K, 7.18 bar, and Xeth,waterfree = 0.81) systems. A two-fold mechanism for hydrate inhibition has been proposed and tested using molecular dynamic simulations for PEO, PVP, PVCap, and VIMA. This mechanism hypothesizes that (1) as potential guest molecules become coordinated by water, form nuclei, and begin to grow, nearby inhibitor molecules disrupt the organization of the forming clathrate and (2) inhibitor molecules bind to the surface of the hydrate crystal precursor and retards further growth along the bound growth plane resulting in a modified planar morphology. This mechanism is supported by the results of our molecular dynamic simulations for the four inhibitor molecules studied. PVCap and VIMA, the more effective inhibitors, shows strong interactions with the liquid water phase under hydrate forming conditions, while PVP and PEO appear relatively neutral to the surrounding water.