Page 314 - Handbook Of Multiphase Flow Assurance

Computer modeling of interaction between a hydrate surface and an inhibitor 313

Early Modeling of clathrate hydrates at CSM
Computer modeling of clathrate hydrates was performed in our lab for the last 5 years.
Long (1994) determined that the coordination number of water molecules around apolar
guest molecules in liquid water is quantized in a series (20, 24, 28, 36, 40) as a function of the
apolar guest size. The first four numbers in this series correspond to the numbers of water
12 8
12 4
12
12 2
molecules in 5 , 5 6 , 5 6 , and 5 6 hydrate cavities.
Pratt (1994) studied hydrate-liquid interface properties. He found that water molecules in
the liquid near the interface align themselves to resemble the hydrate lattice. He also used
cluster-cluster aggregation and diffusion-limited aggregation (DLA) fractal models to sim-
ulate hydrate growth. The DLA growth resembled the hydrate growth observed in brine
(Burruss, 1993) as spikes of hydrate grew to the source of hydrate forming molecules.
Molecular simulation studies performed in CSM laboratory aimed at revealing the mech-
anism of kinetic inhibition of hydrates were reported by Makogon (1994). Adsorption of in-
hibitor species on a surface of hydrate and changing the hydrogen bond structure of water
were identified as possible mechanisms of inhibition. The 1994 work was performed using
the commercial software package SYBYL made by Tripos, Inc. The present work studies the
adsorption mechanism more in-depth.

Studies of monomers adsorption on hydrate with Cerius2

2
In our initial simulation work the Cerius package was used for modeling the inhibitor
adsorption on {100} and {111} planes of sI and sII hydrates.
The DREIDING forcefield (Mayo et al., 1990) used in simulations was developed and vali-
dated by Mayo, Olafson and Goddard for predicting structures and dynamics of organic com-
pounds involving H, C, N, O, F, P, S, Cl, and Br. The authors found that the forcefield produced
excellent structural results for 76 tested organic compounds, rotational barriers of a number of
molecules, and relative conformational energies of a number of molecules. This forcefield was
2
adopted for use in a commercial molecular modeling package Cerius by MSI-Biosym.
The {111} plane was identified as the preferred THF hydrate growth site by our single crys-
tal experiments described in prior sections of this chapter. All hexagonal rings of hydrogen-
bonded water molecules present in the structure II hydrate lie in the {111} planes (Fig. 10.71).
Since H 2 O in hexagonal rings have bonds stretched at 120° (more than in pentagons (108°) or
the normal water angle (104.5°) or tetragonal angles (109.5°)) it is thought that hexagonal ring
formation may be difficult and therefore limiting.
Adsorption of the inhibitor monomers of PVP and PVCap and of a non-inhibitor monomer of
2
PVA was modeled using the “sorption” Monte-Carlo algorithm in the Cerius program at con-
stant volume and temperature. The interactions between the hydrate surface and an inhibitor
monomer were modeled with the DREIDING forcefield which involved Lennard-Jones and elec-
trostatic interaction. No adsorption simulations were done using molecular dynamics (MD).
The hydrate surface was generated from a hydrate unit cell information using the geometric
rules. The sI and sII unit cell data with positions of oxygens and hydrogens of water molecules
in the hydrate lattice were taken from Sparks (1991). During the simulation positions of water
molecules in the hydrate surface were fixed to represent an empty solid hydrate lattice.
Preparation of the monomer for simulation started with its energy minimization and charge
2
equilibration in the Cerius program. The total charge on the monomer was neutral. After the