Long-time-scale molecular dynamics simulations are presented of the spontaneous formation of methane hydrate at a methane/liquid water interface. The water film was prepared at 300 K, 30 bar and showed no significant hydrate order. On crash cooling to 250 K, 300 bar (about 20 K subcooling), the system showed a rapid growth of hydrate clusters. Contrary to popular models for hydrate nucleation, the clusters formed first as two-dimensional arrangements and only later into three-dimensional cage structures; the results are, however, consistent with the local order model proposed recently. The hydrate clusters showed clear signatures of the type II hydrate structure even though the type I structure is the thermodynamically stable form for methane hydrate; this is in accord with the results of recent diffraction experiments.

Clathrate hydrates are important in both industrial and geological settings. They give rise to many technological and environmental applications, including energy production, gas transport, global warming and CO2 capture and sequestration. In all of these applications there is a need to exert a high degree of control on the crystallisation process, either to promote or inhibit it according to the application. This crystallisation process involves the formation of a tetrahedral hydrogen bonding network (as occurs with ice), but is complicated by mass transport limitations due to the poor mixing of the common guest molecules, such as methane, and the water that forms the host lattice. The net effect is that the mechanisms for hydrate formation and growth are still poorly understood, with the consequence that development of additives to control nucleation and growth is still largely governed by trial-and-error approaches. In this paper we show how classical molecular dynamics simulations can be used to provide a direct simulation of the nucleation process for methane hydrate and consequently to allow direct simulation of the effect of additives on the nucleation and growth process. Data are presented for oligomers of PVP and compared with existing data for PDMAEMA. The results show that the two additives work by very different mechanisms, with PVP increasing the surface energy of the interfacial region and PDMAEMA adsorbing to the surface of hydrate nanocrystals. The surface energy effect is a mechanism that has not previously been considered for hydrate inhibitors.

We investigate the fundamental factors controlling polymorphism in 5-fluorouracil by performing molecular dynamics simulations of solutions of the compound in water, nitromethane, and wet nitromethane. Analysis of the effect of solvent on the initial aggregation of 5-fluorouracil molecules shows that the strong binding of water to the 5-fluorouracil molecule hinders the formation of the doubly hydrogen-bonded dimer and, by default, promotes close hydrophobic F...F interactions that are a feature of the unusual (Z' = 4) structure of form I. In contrast, doubly hydrogen-bonded dimers are observed to form readily in solution in dry nitromethane, consistent with the crystallization of the doubly hydrogen-bonded ribbon structure of form II from this solvent. When nitromethane is doped with water, the water forms hydrogen bonds to the solute, interfering with the formation of the doubly hydrogen-bonded dimers, which is consistent with the crystallization of form I from this hygroscopic solvent when it is not dried. Overall, the molecular dynamics simulations provide an atomistic picture of how solvent-solute interactions can significantly affect the initial association of 5-fluorouracil molecules to the extent that they determine the polymorphic outcome of the crystallization.