Fast methane diffusion at the interface of two clathrate structures

Methane hydrates naturally form on Earth and in the interiors of some icy bodies of the Universe, and are also expected to play a paramount role in future energy and environmental technologies. Here we report experimental observation of an extremely fast methane diffusion at the interface of the two most common clathrate hydrate structures, namely clathrate structures I and II. Methane translational diffusion—measured by quasielastic neutron scattering at 0.8 GPa—is faster than that expected in pure supercritical methane at comparable pressure and temperature. This phenomenon could be an effect of strong confinement or of methane aggregation in the form of micro-nanobubbles at the interface of the two structures. Our results could have implications for understanding the replacement kinetics during sI–sII conversion in gas exchange experiments and for establishing the methane mobility in methane hydrates embedded in the cryosphere of large icy bodies in the Universe.

If we assume that for a cystalline matrix the D 2 O coherent cross section does not contribute to the elastic component of the spectra, the result of the estimation is lowered by 15%. In the scenario of a partial decomposition of the water clathrate structure, the result of the estimation is lowered by 25% at most.

Supplementary Note 2: Estimating the Origin of the Diffusing Extra-Cage Methane
Let us first assume that no sI methane hydrate decomposes. In the starting sI clathrate hydrate cage occupancies are typically 86% for the small cages and 99% for the large cages. If similar occupancies are maintained in the sI hydrate and also characterise the sII hydrate of the sI-sII clathrate sample, then almost no methane can be released during transformation from sI to sII.
However, if occupancies in sI are maintained but the cages of the sII hydrate contain a lower amount of methane, a significant fraction of methane could be released during the transformation.
For example, one can calculate that approximately 10% of the methane in the sample is released if cage occupancies in sII are as low as 65% for the small cages and 85% for the large cages (based on the estimated composition of the sI-sII sample in terms of sI and sII, that is 2/3 and 1/3 respectively).
On the other hand, part of the diffusing extra-cage methane must originate from partial decomposition of the clathrate structure. Though the starting sI methane clathrate hydrate sample is in a stable and equilibrated phase, where all water molecules are part of the crystalline structure and all methane molecules are trapped in the cages of the structure, the compressed sample shows coexistence of stable structure I and metastable structure II and such coexistence in near equilibrium is likely characterised by a continuous dynamical rearrangement of water and methane molecules at phase boundaries. During the sI and sII coexistence, the two structures have been suggested to develop intercalated micrometer-sized thin layers 1 and disordered regions where methane is able to diffuse would form in between them. It must be noted that the liquid-like contribution of such 5 disordered regions to the diffraction patterns would be hardly detectable compared to a bulk amorphous or liquid. The previous estimation of a fraction of one third for the diffusing extra-cage methane suggests that a fraction of approximately 20-25% of the water molecules in the sample could belong to these disordered regions between clathrate sI and sII.

Supplementary Note 3: 2D Diffusion Model
For a particle restricted to move along a single plane, the scattering law is a Lorentzian whose half-width-half-maximum is D 2D (Q)(Q sin θ) 2 , where D 2D (Q) is the Q-dependent 2D translational diffusion coefficient and θ is the angle between the vector − → Q and the normal to the plane 2, 3 .
Then, for a polycrystalline sample where a large number of planes are oriented randomly the experimentally observed scattering law < S 2D (Q, ω) > orient. is the isotropic orientational average: The integral can be calculated analytically and gives the following expression 2, 3 : where In our 2D diffusion data analysis, the expression for < S 2D (Q, ω) > orient. given in equation (2) substituted the simple Lorentzian scattering law used in the 3D diffusion data analysis. Supple-