We have used 1 H magnetic resonance microimaging to probe both methane and carbon dioxide hydrate formation processes inside dispersed water droplets. When bulk techniques such as gas uptake measurements are used for determining the kinetics of hydrate formation, these show a gradual conversion to hydrate, suggesting a relatively homogeneous process that might be modeled using a set of intrinsic kinetic parameters. The spatially resolved microimaging measurements show that in fact the conversion to hydrate is quite inhomogeneous, some drops converting quickly, others requiring hours or days. This indicates that the observation of gradual conversion in bulk samples only arises as a result of averaging over many local environments. Quantitative measurements of kinetic processes in subvolumes of a larger sample suggests that the smaller the volume observed, the more inhomogeneous the process appears to be. When hydratecoated water droplets in 3,5,5-trimethylpentane are converted to hydrate, there is evidence that nucleation can take place well away from the hydrate coating, with the hydrate sometimes growing in discrete steps before drops are completely converted. The results obtained indicate that in the quiescent systems studied here the definition of intrinsic kinetic parameters will be difficult, if possible at all, because of a stochastic component that competes with more gradual conversion processes.

Contrary to the thermodynamic inhibiting effect of methanol on methane hydrate formation from aqueous phases, hydrate forms quickly at high yield by exposing frozen water-methanol mixtures with methanol concentrations ranging from 0.6-10 wt% to methane gas at pressures from 125 bars at 253 K. Formation rates are some two orders of magnitude greater than those obtained for samples without methanol and conversion of ice is essentially complete. Ammonia has a similar catalytic effect when used in concentrations of 0.3-2.7 wt%. The structure I methane hydrate formed in this manner was characterized by powder X-ray diffraction and Raman spectroscopy. Steps in the possible mechanism of action of methanol were studied with molecular dynamics simulations of the Ih (0001) basal plane exposed to methanol and methane gas. Simulations show that methanol from a surface aqueous layer slowly migrates into the ice lattice. Methane gas is preferentially adsorbed into the aqueous methanol surface layer. Possible consequences of the catalytic methane hydrate formation on hydrate plug formation in gas pipelines, on large scale energy-efficient gas hydrate formation, and in planetary science are discussed.

Recent observations on the interaction of methane gas with ice surfaces have led to the suggestion that the resulting hydrate layer prevents the encapsulated ice from melting at its usual temperature. This would require ice to exist in a "superheated" state. 1 We have examined the product of the gas-solid reaction with 1 H NMR imaging. The imaging experiments show that the hydrate-encapsulated ice is able to melt at its usual melting point. As a possible alternative model, we suggest that a considerable amount of ice inside the hydrate layer can be converted to hydrate and liquid water under isothermal and constant volume conditions, the hydrate layer acting as an insulating, semipermeable layer that insulates processes inside the hydrate layer from external bulk temperature and pressure measurements.

The second and third virial coefficients of methanol have been determined from 150 to 300°C. At most temperatures the estimated standard error of the second is less than 1%. These values of the second virial coefficient are less negative than results in which the third and higher virial coefficients were neglected. The virial coefficients are interpreted in terms of equilibria between monomer, dimer, and trimer; the enthalpies and entropies of dimerization and trimerization are in line with those found for steam.

The equation of state of steam has been determined at 25° intervals from 150° to 450°C by measuring the mass of water injected into a high-temperature vessel as a function of pressure. The observations have been analyzed by a new method that does not require an independent volume of the vessel and are reported as the second and third virial coefficients and as empirical equations for them. The standard error of the second virial coefficient, taking account of the effects of experimental scatter and of truncation of the virial equation, which are comparable at 450°C, is 3% at 150° and 1% at 450°C. The Stockmayer potential fits the second virial coefficient fairly well, but predicts the wrong form for the third.