Recovering Methane from Solid Methane Hydrate with Carbon Dioxide

CO2 in, CH4 out! From NMR spectroscopic measurements of composition and kinetics, it can be shown that the recovery of methane from methane hydrate is possible through its reaction with CO2 (see diagram), even though the yield is less than that expected from kinetic and thermodynamic arguments. The discovery of vast hydrocarbon resources in hydrate form on the continental margins and in permafrost regions has sparked an interest in the possible recovery of relatively clean-burning methane gas,1 especially in resource-poor national economies. At the same time, it has been proposed that CO2, obtained as a by-product of combustion processes, possibly could be stored either in the deep ocean as liquid, or as a solid hydrate,2 in order to reduce the release of greenhouse gas into the atmosphere. As such, the conversion of methane hydrate to CO2 hydrate with the net recovery of methane seems quite attractive.3 The limiting equilibrium composition of the mixed hydrate and the kinetics of conversion then become of primary interest in assessing this potential technology. In this communication we address these various issues, which have been considered so far only from a macroscopic point of view, by following not only kinetics, but also the distribution of guests over different cages using solid-state NMR methods. This combined approach allows the derivation of thermodynamic and kinetic information and also gives a molecular-level rationale for the observations; thus it is a powerful way of studying solid-state reactions. Both CO2 and CH4 form a structure I (sI) hydrate,4 as do mixtures of these gases,5 with an icelike framework that consists of hydrogen-bonded water molecules. It has eight cagelike guest sites in the unit cell, including two pentagonal dodecahedra (512) and six tetrakaidecahedra (51262) consisting of 12 pentagonal and two hexagonal faces. The ideal unit cell can be written as 2 MS⋅6 ML⋅46 H2O with an ideal hydration number M⋅5H2O if a single type of guest is present. However, sI hydrates are non-stoichiometric, and the actual hydration numbers usually lie somewhere between 6 and 8.4 The latter can be measured by the careful application of direct methods,6 and also from cage occupancies obtained from structures derived by single crystal X-ray diffraction,7 or from spectroscopy8 where the cage occupancy ratio can be linked to the hydration number by using the van der Waals–Platteeuw,9 or related, models. To observe favorable exchange between CO2 and methane hydrate, there must be preferential partitioning of CO2 and CH4 between the gas and the hydrate solid phases. The only way this can arise is if CO2 has a preference for the large cage (51262) in the hydrate, as the larger cages outnumber the smaller by a factor of three. This premise was investigated by examining the distribution of methane over the two cage sites by MAS 13C NMR for hydrate samples prepared from gas mixtures (Figure 1). One can see that for pure methane hydrate θequation/tex2gif-inf-16.gif/θequation/tex2gif-inf-18.gif is ≈1.26, so that the small cage is occupied to a smaller degree than the larger cage, as reported before for synthetic and natural methane hydrates.8, 10 With increasing CO2 in the gas mixture, the ratio declines steadily to a value that shows that now fewer than half as many large as small cages are filled with methane. This behavior can be explained in terms of the molecular sizes of CO2 and CH4. The size of CO2 is almost the same as the cavity diameter of the small 512 cage in the sI hydrate, the molecular diameter to cavity diameter ratio being 1.00 for CO2 and 0.855 for CH4. For CO2 hydrate, NMR and diffraction techniques have indicated that the 512 cages have smaller fractional occupancies than the large cages. Hydration numbers of ≈6.0 for methane and 6.2 for CO2 hydrate are consistent with fractional occupancies of the small cages θS of ≈0.8 and ≈0.7, respectively. Thus, when CO2 competes with CH4 in occupying the small cage of the sI hydrate, CO2 is a relatively poorer guest, so that the occupancy ratio for CH4 (θequation/tex2gif-inf-33.gif/θequation/tex2gif-inf-35.gif) becomes lower at higher CO2 compositions because CO2 preferentially occupies the large 51262 cage in the mixed CH4/CO2 hydrate. Another aspect to be noted is that the equilibrium occupancy ratios ranged from 1.26 (100 % CH4) to 0.23 (very dilute CH4). Thus when the CH4 concentration becomes infinitely dilute, the value of 0.23 for θequation/tex2gif-inf-44.gif/θequation/tex2gif-inf-46.gif shows that there is a limit to the degree of substitution that one can expect under equilibrium conditions, even when the gas composition approaches 100 % CO2. A second plot (Figure 2) shows the composition of the hydrate as a function of the corresponding gas composition at equilibrium. Considerable enrichment in CO2 of the hydrate with respect to the gas occurs for concentrations between ≈10 and 70 % CO2. When considering the data in Figure 1 and Figure 2, along with some assumptions,11 a limiting composition of the mixed hydrate can be estimated showing that at least 64 % of methane should be recoverable from a hydrate of composition CH4⋅6.05 H2O, which after reaction with CO2 gives a product hydrate with a CO2/CH4 ratio of 1.8 or greater. 13C HPDEC MAS NMR spectra of an sI hydrate prepared using CH4/CO2 mixtures with compositions as indicated. Equilibrium composition of the sI hydrate as a function of the composition of the source gas in mol % CO2. The circles show the hydrate phase composition at the corresponding vapor phase composition of the source gas, as represented by squares. AL/AS is the intensity ratio of the 13C MAS NMR resonance lines of methane molecules in large and small cages. θL/θS is the occupancy ratio of methane molecules in large and small cages, as calculated from the intensity ratio. The kinetics of transformation are also a key element in considering the feasibility of the gas-replacement reaction. First of all we follow the reaction of CO2 with powdered ice by measuring the intensity of the 13C CO2 powder pattern as a function of time. This reaction is remarkably slow at 268 K, with the reaction incomplete after 30 h (rate k1≈0.0026 min−1 at 268 K, k1≈0.0025 min−1 at 270 K, P=58 bar).12 Similarly, the reaction of methane with powdered ice is very slow (rate k1≈0.0009 min−1 at 270 K, P=215 bar). On the other hand, when methane hydrate is exposed to CO2 gas, the guest replacement reaction appears to be complete in less than 5 h (rate k1≈0.01 min−1 at 270 K). The amount of methane obtained from the reaction is about half of the total methane present in the initial hydrate (Figure 3), which indicates that the product is a hydrate with a CO2/methane ratio of ≈1. This suggests that the conversion is limited by considerations other than thermodynamics, as this predicted a product with a CO2/methane ratio of ≈1.8.11 It is likely that sample morphology plays an important part, where the conversion reaction proceeds so as to isolate volumes of pure methane hydrate surrounded by the converted hydrate. Sample grinding clearly would allow the reaction to proceed to equilibrium rather than to the pseudo-equilibrium state observed in this study. Formation and conversion of sI methane hydrate: a) Integrated intensity of the 1H NMR signal during formation of sI methane hydrate from CH4 and powdered ice (268 K, 215 bar). The solid line represents an exponential fit with a rate constant of 9×10−4 min−1; b) conversion of the sI methane hydrate with CO2, as obtained from the integrated 13C powder-pattern intensity due to CO2 in the large cage of sI hydrate (60 bar 13CO2 at 270 K). The solid line is an exponential fit with a rate constant of 0.97×10−2 min−1. We note that the reaction of CO2 hydrate with methane, on the other hand, is again very slow (Figure 4), with a steady state established after ≈60 h. We note that this experiment reports only the replacement of CO2 in the large cages, so that just over 10 % of the large cages in the product contain methane. The reaction of CO2 with methane hydrate and the reverse reaction of methane with CO2 hydrate are obviously completely different. One may argue that since CO2 is a much better large-cage guest than methane, two factors that control the kinetics may be identified. From an equilibrium point of view, it appears that CO2 should replace methane rather easily, but this requires the lattice to be disrupted completely, as it involves changing the guests in a majority of the 51262 cages, which are the major constituents of the hydrate lattice. So, the process is likely to lead relatively easily to the pseudo-equilibrium composition, with the replacement of methane by CO2 accomplished as the hydrate particles recrystallize from the outside in. On the other hand, CO2 at the hydrate particle surface apparently is not easily replaced by methane, as it is the favored guest only in the small cages; only ≈10 % of the CO2 in the large cages is replaced. The hydrate crystal likely remains largely intact, so that methane can only find its way to the interior sites by diffusing through a near-intact hydrate lattice; such a process must be very inefficient. Intensity of the 13C NMR signal due to CO2 in the large cages of the sI hydrate shown as a function of time during the formation of sI CO2 hydrate (on the left side of the dotted vertical line; T=268 K, P=58 bar) and after exposure of the material to CH4 (152 bar; points on the right side of the dotted line). The results presented here show that, at least in a laboratory setting, the replacement of CH4 by CO2 is favored both from the points of view of equilibrium thermodynamics and kinetics. Reaction rates and the total yield of the replacement reaction in actual natural gas hydrate deposits will depend on a variety of factors, such as the degree of hydrate dispersion in sediment, hydrate particle size12, and gas transport. For composition studies, pure and mixed-phase hydrates were prepared in porous silica gel (Aldrich, pore size=15 nm) to facilitate the rapid equilibration of hydrates. 13C NMR spectra were recorded at 200 K by placing the hydrate samples in a 4-mm diameter Zr-rotor that was loaded into the variable-temperature (VT) probe on a Bruker 400 MHz solid-state NMR spectrometer. All spectra were recorded at a Larmor frequency of 100.6 MHz with magic-angle spinning (MAS) at about 2–4 kHz under high-power proton decoupling (HPDEC). A pulse length of 2 μs and pulse repetition delay of 20 s with radio-frequency field strengths of 50 kHz corresponding to 90° pulses of 5 μs duration were used. The downfield carbon resonance peak of adamantane, assigned a chemical shift of 38.3 ppm at 300 K, was used as an external chemical shift reference. For the mixed hydrate samples of CO2 compositions over 80 mol %, 13CH4 gas was used to obtain higher intensity CH4 peaks. Kinetic studies of the reaction of powdered solids (deuterated ice or hydrate, particle size 5–50 μm) with gases were carried out in a high-pressure NMR cell connected to a gas handling system. For CO2, the reactions were followed by measuring the integrated 13C powder pattern, characteristic of the guest in the large cage of the sI hydrate, as a function of time. The variation of methane content in the hydrate with time was obtained from the integrated 1H NMR signal. The absolute amount of product was obtained by comparing the signal with that of known samples.