Page 330 - Handbook Of Multiphase Flow Assurance

Computer modeling of interaction between a hydrate surface and an inhibitor 329

It can be concluded from the monomer study that both inhibitors (PVP and PVCap) and
non-inhibitors (PVA) adsorbed to the hydrate surface. Difference in the strength of adsorption
was not large (under 10%), but the difference in inhibition is great (inhibitor vs. non-inhibitor).

Adsorption of inhibitor polymers on hydrate
Simulating the adsorption of short polymer chains on hydrate called for a serious mod-
ification of the program. The Monte Carlo moves of rotation and translation were already
present in the simulation. In order to obtain realistic conformations of the polymer chain on
hydrate surface, the polymer backbone had to be flexible. Molecular dynamics was ruled
out because the results would not be directly comparable to the previously obtained data for
monomers. Lattice Monte Carlo is a widely used model for studying polymer adsorption on
solids. However, this approach requires prior knowledge or an assumption of the adsorption
strength for each segment, and atomically correct polymer chains could not be simulated.
The pivot Monte Carlo method was used for obtaining a new conformation of a polymer by
making a bold change in backbone structure. Many of the pivot attempts got rejected because of
the higher energy final state, but the accepted move produces an “essentially new” configuration
(Madras and Sokal, 1988). This algorithm allowed us to use the existing Monte Carlo code and
run simulations with atomically correct polymer chains and hydrate surfaces. Extensive valida-
tion of the polymer adsorption program was not done since it included the already validated
monomer adsorption program and a new pivot-move part. Several trial runs were done to ensure
the proper functioning of the added pivot-move part of the simulation. Tests were performed
with 10- segment polymers of PVP and PVA as simple but representative chemicals. The chains
segments reoriented in proper directions and the polymer chain wasn't broken by this motion.
The sII {111} surface was selected for studying the adsorption of short polymers because of
sII hydrates' industrial significance and because our single crystal experiments indicated that
the {111} face dominates the sII THF hydrate.
From the monomer studies it was found that the strength of adsorption on sII {111} face var-
ied with the surface distance (depth) from unit cell origin. A study of hydrate surface stability
2
was done using Cerius in order to determine the most stable hydrate surface for polymer
adsorption. A one-water-molecule-thick 5 Å {111} slice of hydrate was cut from the sII unit cell
at different positions, and the surface potential energy was calculated. We used DREIDING
interaction parameters with hydrogen bonding potential well depth set to 5 kcal/mol and SPC
Lennard-Jones and charge parameters for atoms in water. Two valleys were found at about 3
and 7 Å from the unit cell origin. These locations correspond to the surfaces where the large
and the small cavities get completed (Fig. 10.82). The same study was repeated for 8.5 Å thick
slices of hydrate. The most stable slice was the one with its surface 7 Å from the unit cell origin.
12 4
12
This surface exhibits completed small 5 cavities and open 5 6 cavities. The most stable sur-
face was adopted for studying adsorption of polymers.
The same procedure was used to compare relative stabilities of {100}, {110}, and {111} hy-
drate surfaces of sII. It was found that in each of the three directions the number of water
molecules falling into a 5 Å-thick slab of hydrate varies, so the potential energies were nor-
malized by the number of water molecules. The equilibrium shape of a crystal is that of its
minimum energy. This is called the Wulff condition (Myerson, 1993). {111} planes show the
most stability (Fig. 10.83) which possibly explains why they are observed experimentally;
same result was obtained for non-normalized surface energies.