J. Phys. Chem. A 2010, 114, 247–255
247
Kinetics of Methane-Ethane Gas Replacement in Clathrate-Hydrates Studied by Time-Resolved Neutron Diffraction and Raman Spectroscopy M. Mangir Murshed,† Burkhard C. Schmidt,‡ and Werner F. Kuhs*,† GZG, Abt. Kristallographie, Abt. Mineralogie, UniVersita¨t Go¨ttingen, Goldschmidtstrasse 1, 37077 Go¨ttingen, Germany ReceiVed: August 19, 2009; ReVised Manuscript ReceiVed: October 1, 2009
The kinetics of CH4-C2H6 replacement in gas hydrates has been studied by in situ neutron diffraction and Raman spectroscopy. Deuterated ethane structure type I (C2H6 sI) hydrates were transformed in a closed volume into methane-ethane mixed structure type II (CH4-C2H6 sII) hydrates at 5 MPa and various temperatures in the vicinity of 0 °C while followed by time-resolved neutron powder diffraction on D20 at ILL, Grenoble. The role of available surface area of the sI starting material on the formation kinetics of sII hydrates was studied. Ex situ Raman spectroscopic investigations were carried out to crosscheck the gas composition and the distribution of the gas species over the cages as a function of structure type and compared to the in situ neutron results. Raman micromapping on single hydrate grains showed compositional and structural gradients between the surface and core of the transformed hydrates. Moreover, the observed methane-ethane ratio is very far from the one expected for a formation from a constantly equilibrated gas phase. The results also prove that gas replacement in CH4-C2H6 hydrates is a regrowth process involving the nucleation of new crystallites commencing at the surface of the parent C2H6 sI hydrate with a progressively shrinking core of unreacted material. The time-resolved neutron diffraction results clearly indicate an increasing diffusion limitation of the exchange process. This diffusion limitation leads to a progressive slowing down of the exchange reaction and is likely to be responsible for the incomplete exchange of the gases. 1. Introduction Gas hydrates are nonstoichiometric crystalline inclusion compounds. They are built of a framework of hydrogen-bonded water molecules forming polyhedral cavities occupied by guest molecules such as CH4 and C2H6.1 Hydrates of hydrocarbons occur in a wide range of oceanic and permafrost environments where relevant gases and p-T conditions exist for their formation.1–3 An increasing number of gas hydrate deposits has been found around the world, and geophysical estimates suggest that a recovery of even a fraction of the trapped hydrocarbons would provide a substantial energy resource. Gas hydrates are of particular interest also because of their potential as a separating agent, storage medium, and, because of concern from the oil and gas industry, for their ability to plug oil/gas transportation lines.1 The pioneering X-ray crystallography work of von Stackelberg and Mu¨ller4 revealed two types of gas hydrate structures, namely, structure type I (sI) and structure type II (sII). The sI hydrate unit cell is comprised of 46 water molecules forming two 512 (SCI) and six 51262 (LCI) cages, and the sII hydrate unit cell contains 136 water molecules that form sixteen 512 (SCII) and eight 51264 (LCII) cages. Later in 1987, Ripmeester et al.5 introduced a gas hydrate with a hexagonal structure type (sH); all three types of hydrates occur in nature. CH4 and CO2 are known to usually form a sI hydrate, however, each of their sII phases (CH4-sII6 and CO2-sII7) are also reported to exist metastabily. A gas mixture of CH4 and C2H6 forms both sI and sII hydrates; their structure type is sensitive to p-T-x conditions * To whom correspondence should be addressed. E-mail: wkuhs1@ gwdg.de. Phone: +49 551 393891. Fax +49 551 399521. † GZG, Abt. Kristallographie. ‡ GZG, Abt. Mineralogie.
and, again, metastable formations of the thermodynamically less favored form occur fairly frequently in the vicinity of 0 °C.1,8–10 Because of the predominance of CH4, generally produced from the microbial breakdown of sedimentary organic matters, CH4 represents more than 99% of the filling of natural gas hydrates.11 Within the sediment-water column the lowest temperature is usually observed at the sea floor, and temperature increases downward through the sediments following the geothermal gradient. Gas hydrates persist in the so-called gas hydrate stability zone (GHSZ) with/without underlying free gas,12 which extends from the sea floor down to a certain depth, and free gases often occur just below the lower boundary of the GHSZ. The idea of CO2 sequestration by replacing CH4 from the hydrate sediments is a two-in-one approach of CO2 mitigation and concomitant extraction of CH4 gas.13 Recent works on CH4-CO214–19 and other gas replacements in hydrates were triggered by the pressing demand of energy worldwide along with an increasing need for CO2 sequestration. Ota et al.18 conducted replacement experiments of CH4 hydrate with liquid CO2 using an in situ Raman analysis. Lee et al.15 studied the replacement mechanism of CH4 hydrate with gaseous CO2; their replacement of CO2 with CH4 hydrate appeared to be fairly different from replacement of CH4 with CO2 hydrate. Halpern et al.20 performed in situ neutron powder diffraction to observe a transition of Ar-sII hydrate into CO2-sI hydrate exposing ArsII hydrate to liquid CO2. Recently Yeon et al.21 studied the swapping phenomena observed between externally applied (pure CH4) and internal (CH4-rich + C2H6 or isopentane/methylcyclohexane) guests by NMR spectroscopy in trying to explain the preponderant occurrence of sI over sII, and sH in natural methane hydrate deposits. Molecular details of the swapping mechanisms were neither given for this nor for any other experimental work; however, given the size of the molecules
10.1021/jp908016j 2010 American Chemical Society Published on Web 10/28/2009
248
J. Phys. Chem. A, Vol. 114, No. 1, 2010
any swapping must take place via a destruction and reformation of water cages at least for any substantial exchange rates. Prior to the design any large-scale approach and a study of economic feasibility for replacement processes in natural settings a full understanding of the thermodynamics and kinetics of the replacement process should be obtained since partial failures of the intended replacement may affect the stability of continental margins.22,23 Likewise, marine gas hydrate bearing sediments may be vulnerable and eventually even become unstable during a slow transformation by an ongoing attack of migrating free gases, for example, by an interaction of locally produced biogenic CH4 gas with previously formed thermogenic gas hydrates.24,25 Clearly, the stability of gas hydrates is sensitive to p-T-x conditions at the boundary between marine hydrates and free gases. Compositional changes in crystalline hydrates occurring as a consequence of changing p-T-x conditions may lead to regrowth processes and affect the cementing properties of gas hydrates in a sedimentary matrix. Despite such broad implications, only a limited number of studies were devoted to gas exchange reactions and mostly confined to the CH4-CO2 replacement. The in situ observation of the replacement between the enclathrated guest in the hydrate cage and an external gaseous guest in the CH4-C2H6 system will help to understand such processes taking place in marine environments. Accordingly, we present here laboratory work on the exchange reactions of pure C2H6 sI hydrates exposed to CH4 free gas. In many gas replacement studies, the hydrate samples were exposed to external gas at p-T conditions below the equilibrium line of the starting hydrate, however, within the stability field of the transformed hydrates (e.g., ref 14 and 16). At such conditions, the parent hydrate will decompose into the constituent gas and water, and the externally applied gas, possibly mixed with gas from the decomposed hydrate, will reform from water or ice (resulting from the decomposition). However, for a possible CO2 sequestration with concomitant CH4 recovery as well as for a thermogenic-biogenic gas replacement reactions in marine sediments the most relevant p-T conditions are within the stability field of both the parent and the transformed phases, conditions that are met in the present study. Neutron diffraction experiments have enabled some recent successes in studying guest replacement in hydrates, especially when accompanied by a structural transition, and have provided useful information on the time dependency of hydrate transformations.20 In situ neutron diffraction experiments demonstrated its uniquely powerful access to the fast initial parts of clathration reactions.7,10 Undoubtedly, neutron diffraction provides excellent averaged information;26 yet, this technique cannot deliver direct physical insights into the spatial aspects of the hydrate transformation, nor does it deliver highly accurate cage occupancies when mixed hydrocarbon hydrates10 are concerned due to partial cancellation of hydrogen and carbon scattering contributions as well as due to the time-space averaging inherent to the method. Since Raman spectroscopy has been widely used as a powerful tool for identifying structure type and relative cage occupancy of gas hydrates, it could also deliver salient features on guest replacement processes. The vibrational modes of molecules shift in frequency depending on the local environment of the molecules. The frequency shift can be measured by Raman spectroscopy that allows the discrimination between a molecule in the gas phase and in a discrete form in the hydrate cages.8 Although the O-H stretching modes of water can clearly indicate the presence of solid water phases, they cannot distinguish the hydrate structure types. Instead, the stretching
Murshed et al.
Figure 1. Rietveld plots of the synchrotron X-ray diffraction data of the as-synthesized C2H6 sI hydrates; the wavelength used was 0.1244 Å. The vertical black, dark-gray, and gray solid bars depict symmetry allowed positions of Bragg reflections of C2H6 sI hydrate, ice Ih, and monoclinic C2H6, respectively.
modes of the hydrate guests have mostly been used for the assignment of hydrate structure, as well as for quantitative approaches concerning cage fillings.27,28 For the present study, we use both time-resolved neutron powder diffraction and microfocus Raman spectroscopy to monitor structural and compositional transformations during the CH4-C2H6 replacement in hydrates. The combined investigation will allow a cross-check between spatially resolved (Raman) and bulk (neutron) information of the studied hydrates; the complementarity of the results will help to understand some fundamental aspects of the complex CH4-C2H6 mixed hydrate systems.8–10,29–32 2. Experimental Methods 2.1. Preparation and Characterization of Ethane Hydrates. Eight samples of C2H6 sI hydrates were prepared from deuterated spherical ice grains (D2O ice Ih) in a high-pressure cell33 at 270.4 K (for 7 days) followed by 278.3 K (for 14 days) and 3.8 MPa. Details of the preparation and characterization of the spherical D2O ice Ih is available elsewhere.7,34 Two types of C2H6 sI hydrates were used for in situ neutron diffractions to carry out the replacement kinetics, (I) crushed samples with particles sizes ranging from a few micrometers to a few 100 µm and (II) as-synthesized consolidated samples (cylindrical shape with a diameter of ∼7 mm and length of ∼37 mm). One of the as-synthesized C2H6 sI hydrate samples was investigated by synchrotron X-ray powder diffraction and Raman spectroscopy to check the purity of the samples in terms of hydrate/ice ratio. For the synchrotron measurements at 70 K on the hard X-ray beamline BW5, HASYLAB at DESY, Hamburg, and the subsequent data analysis we followed the procedures described in Murshed and Kuhs.10 In Figure 1 the Rietveld refinement plot revealed that the starting C2H6 sI hydrates were almost pure (with an ice fraction below 1% as can be seen in Table 1). Atmospheric H2O frost formation on hydrate surfaces seems to be a problem and was always found after quenching to liquid nitrogen (LN2) and crushing the hydrate samples at low temperatures despite the fact that this operation was carried out under LN2 in a cold room at -10 °C. For instance, Figure 2 depicts a slight contamination of H2O frost (∼5-10 µm particle size) that likely formed on the surface of deuterated pure CH4 sI hydrates while crushing at LN2 temperature. During recovery of the C2H6 sI hydrates from the pressure cell at temperatures below the C2H6 melting point (89.3 K), a solid monoclinic C2H6 phase was identified, which amounted to about 6% of the crystalline constituents of the sample, as given in Table 1. C2H6
Methane-Ethane Gas Replacement in Clathrate-Hydrates
J. Phys. Chem. A, Vol. 114, No. 1, 2010 249
TABLE 1: Results of Rietveld Refinements Based on Synchrotron X-ray Powder Data, in Particular the Phase Fraction of C2H6 sI Hydrate (r-I), D2O Ice Ih (r-ice), Monoclinic C2H6 (r-C2H6), and Lattice Constant of C2H6 sI hydrate (a0)a a0
R-I
R-ice
R-C2H6
ESCI
ELCI
Rwp
χ2
11.8953(29) 11.8979(16)
0.926(1) 0.932(3)
0.002(2) 0.008(1)
0.071(2) 0.059(1)
4.8(1) 4.8(1)
100 100
3.49 1.78
0.43 0.11
experiment A B
a ESCI and ELCI refer to occupancy (%) of C2H6 in the 512 and 51262 cages of C2H6 sI hydrates, respectively. Rwp (weighted residual, %) and χ2 (goodness of fit) are Rietveld refinement statistics as defined in ref 36. The measurements were performed on two different locations A and B (vertically 10 mm apart) of the same sample at 70 K and ambient pressure. ELCI was refined initially, found to be insignificantly different from 100% and subsequently fixed to 100%.
TABLE 2: Results of Some Selective Rietveld Refinements Based on In Situ Neutron Diffraction Data, in Particular the Phase Fraction of C2H6 sI Hydrate (r-I), CH4-C2H6 sII Hydrate (r-II), and Their Lattice Constants (a0)a R-I
R-II
0.78(3) 0.69(2) 0.64(2) 0.61(2) 0.58(2) 0.57(2) 0.54(2)
0.22(3) 0.31(3) 0.36(3) 0.39(3) 0.42(3) 0.43(3) 0.46(2)
2 4 5 6 8 8.8
0.73(1) 0.64(2) 0.61(2) 0.59(2) 0.55(2) 0.54(2)
2 4 6 7.6
0.99(1) 0.96(2) 0.95(1) 0.94(1)
time (h) 2 4 6 8 10 12 14
ELCIIb (%)
Rwp
χ2
Experiment-1c (total data ) 224) 12.0321(4) 17.2527(25) 12.0317(5) 17.2525(17) 12.0315(5) 17.2529(15) 12.0322(6) 17.2532(14) 12.0318(6) 17.2531(13) 12.0316(6) 17.2535(13) 12.0440(7) 17.2690(12) mean
97(12) 88(9) 91(6) 86(7) 91(7) 91(7) 91(7) 91(8)
1.65 1.60 1.61 1.62 1.60 1.65 1.71
5.44 5.08 5.17 5.24 5.10 5.43 3.48
0.27(3) 0.36(3) 0.39(3) 0.41(3) 0.45(2) 0.46(3)
Experiment-2 (total data ) 134) 12.0389(5) 17.2535(20) 12.0391(5) 17.2541(14) 12.0394(5) 17.2547(13) 12.0390(18) 17.2560(27) 12.0391(5) 17.2557(11) 12.0387(6) 17.2543(11) mean
96(10) 100(7) 96(7) 91(7) 94(7) 91(7) 95(7)
1.71 1.64 1.65 1.65 1.58 1.67
6.06 5.57 5.62 5.63 5.14 5.75
0.01(1) 0.04(1) 0.05(1) 0.06(1)
Experiment-3d (total data ) 129) 12.0333(4) 17.264(40) 12.0403(3) 17.276(12) 12.0407(5) 17.2688(81) 12.0408(3) 17.2647(84) mean
100 100 100 100 100e
2.25 1.65 1.59 1.71
2.27 6.60 6.13 7.04
a0 (sI) (Å)
a0 (sII) (Å)
a Crushed C2H6 sI hydrates were used for exp-1 and exp-2, and as-synthesized consolidated cylinder-shaped C2H6 sI hydrates for exp-3. b A linear occupancy constraint (total occupancy ) CH4 + C2H6) was used for the 51264 cages of C2H6-CH4 sII hydrates (ELCII). c Temperature was raised respectively from 260 to 275 K after ∼2.8 h and 270 to 280 K after ∼13 h. d Temperature was raised respectively from 260 to 275 K after ∼2.8 h and 270 to 280 K after ∼13 h. e Using a similar constrainta could not produce any sensible value due to low conversion of C2H6-CH4 sII hydrates.
is likely to have solidified at the top of the aluminum can, and was allowed to completely degas at p-T conditions inside the C2H6 sI hydrate stability field1 before the samples were used for the in situ neutron diffraction. The Raman investigations were performed at temperatures higher than the C2H6 triple point (90.35 K) under atmospheric pressures thus gas condensation could not occur. 2.2. In Situ Neutron Diffraction. Neutron diffraction is very suitable for the investigation of gas hydrate formation and replacement reactions due to the high penetration power of neutrons even through thick-walled sample environment. To observe the ongoing changes we used the D20 powder diffractometer located at the High-Flux-Reactor of the Institut LaueLangevin (ILL), Grenoble at its highest intensity setting for the wavelength (λ) of 2.4182 Å. The in situ neutron diffraction runs were performed while pure C2H6 sI hydrates were exposed to free CH4 gas in a high strength autofrettaged aluminum gas pressure cell33 at 5 MPa. The typically ∼1 cm3 large sample is seen as a whole by the neutron beam thus an excellent bulk representativity is achieved; the accuracy for the degree of transformation (R) obtained is a few per mille for each of the measured time slices.10 The aluminum cans filled with C2H6 sI hydrates were transported in a dry mover dewar to ILL/Grenoble at temper-
atures close to 77 K. After removal from the LN2 dewar, the C2H6 sI hydrate containing aluminum cans were placed on dry ice for typically 30-50 s allowing a complete degassing in case there was any gas (N2 and/or C2H6) condensed inside the cans. The sample was then put into the precooled (∼240 K) pressure cell already fixed to the sample stick, and the Bridgman seal was closed. This filling operation was performed at an applied CH4 gas pressure of ∼1 MPa to ensure a complete filling of the system. The sample was then inserted into a He-flow cryostat installed on the diffractometer and preset at the desired temperature. The temperature reading of the sample in the cryostat was obtained from a calibrated Pt-sensor fixed to the pressure cell wall. Subsequently, the target gas pressure was applied within a few seconds while the acquisition of data was started simultaneously. The reactions started immediately within the diffractometer time resolution of a few seconds after admission of the exchange gas. It is important to note that the CH4 gas phase was not refreshed by flushing during the replacement reactions (1) to strictly maintain homogeneous p-T conditions and (2) to mimic situations on the sea-floor where a refreshing of the initial composition may well not take place. This procedure allowed some mixing of the gases (∼3.5 vol % C2H6 for a complete decomposition of the starting C2H6 sI hydrate) in the sealed volume (∼90 cm3) during some possible
250
J. Phys. Chem. A, Vol. 114, No. 1, 2010
Figure 2. Typical Raman spectra of spherical ice Ih, crushed CH4 sI hydrate, and C2H6 sI hydrates. The dotted bracket indicate a H2O contamination that was caused by micrometer-sized frost particles formed by condensation from the vapor phase on a deuterated CH4 sI hydrate sample at liquid nitrogen temperature. ESCI and ELCI refer to C2H6 trapped in the 512 and 51262 cages of C2H6 sI hydrate, respectively.
growth/regrowth processes of the hydrates. Each reaction was followed over a period of several hours at different temperatures (260-280 K) and constant pressure (5 MPa). The data were collected with a time resolution of about one minute at the initial stage and with a resolution of about five minutes for the later parts of the reactions. All data were subjected to detector efficiency correction and background subtraction. 2.3. Rietveld Refinements. The powder diffraction data for both synchrotron X-ray and in situ neutron experiments were analyzed using GSAS35 and its graphical user interface (EXPGUI)36 facility, which delivered quantitative information on the crystalline constituents of the samples from multiphase Rietveld refinements. The initial lattice constants of D2O ice Ih were taken from their thermal expansion report.37 The crystal structure parameters for monoclinic C2H6, C2H6 sI hydrate, and CH4-C2H6 mixed structure type II (CH4-C2H6 sII) hydrate were taken from van Nes and Vos,38 Udachin et al.,39 and Rawn el al.,40 respectively. General parameters such as six Chebychev polynomial background parameters, the zero point of the detector, three profile parameters (GU, GV, and GW), lattice constants of the hydrates, and the weight fraction of the constituent phases (R) were refined; the atomic coordinates (xyz), isotropic displacement parameters, and the histogram scale factor were fixed. During trial runs the occupancy of LCI and SCI of pure C2H6 sI hydrates were found to be occupied with C2H6 to 100% and approximately 5%, respectively (Table 1); these values were kept fixed during all neutron Rietveld refinements when modeling the decaying pure C2H6 sI hydrate phase. In contrast, approximately 1.5% of SCI was detected to be filled with C2H6 by Raman spectroscopy, however, this value was below the detection limit for the neutron data and not taken into account also due to the observed local variation (vide infra) of Raman measurements. During early stage trial structure refinements, the occupancy of SCII of CH4-C2H6 sII hydrates with CH4 fluctuated close to full occupancy, a value that was fixed for the subsequent runs. Using a linear constraint for the occupancy of LCII was considered as being shared by both CH4 and C2H6 adding to 100% (i.e., LCII total occupancy ) CH4 + C2H6). Results of some selective refinements are given in Table 2; the mean value of the occupancy of LCII was fixed for each experimental data analysis. Subsequently, using the structural parameters established from selected individual data sets an automated treatment of the data of all time slices (their number is given in Table 2) was performed to obtain the phase fractions of type I and type II hydrates.
Murshed et al.
Figure 3. Measurement of the depth resolution (confocal length) of the laser beam of the Raman spectrometer using a silicon wafer at room temperature and ambient pressure.
2.4. Raman Spectroscopy. Raman spectra were recorded on a Horiba Jobin Yvon HR800 UV Micro-Raman spectrometer equipped with an air-cooled Ar laser working at 488 nm with a laser power of less than 20 mW. The use of a 50× long working distance objective (Olympus) with a numerical aperture of 0.55 gave a focus spot of about 1.1 µm diameter when closing the confocal hole to 100 µm. Raman spectra were collected in the range between 200 and 4000 cm-1 with a spectral resolution of approximately 2.2 cm-1 using a grating of 600 grooves/mm and a Peltier-cooled CCD detector (Andor, 1024 × 256 pixels). The spectral positions were calibrated against the Raman mode of Si before and after the sample measurements. The position of the Si peak was repeatedly measured against the Rayleigh line (0.0 cm-1) yielding a value of 520.4 ( 0.3 cm-1. The linearity of the spectrometer was calibrated against the emission lines of a neon lamp. A minimum depth resolution (confocal length) of about 16 µm was estimated from the z-scanning of a Si wafer. For this purpose, a laser beam was focused slightly above the surface of the Si wafer and the intensity of the peak (520.4 cm-1) was measured as a function of the vertical position, as shown in Figure 3. As discussed by Everall,41 this procedure yields a minimum confocal length in air, and focusing the laser beam into a sample with higher refractive index leads to an increase of the confocal length. The gas hydrate samples were handled under LN2 temperature and quickly transferred into a precooled Linkam THMS600 Heating and Freezing Stage to avoid any clathrate decomposition. Raman spectra were collected after the in situ neutron as well as ex situ synchrotron powder diffraction experiments. A few milligrams of the samples were loaded into the Linkam stage from the top of the aluminum can. At least three spectra were taken from three different grains of a sample loaded each time. Multiple Peak Fitting. Raman intensity cannot be correlated simply with concentration, as it is also a function of the molecule’s relative Raman scattering cross section (σ).42 For guest molecules trapped in a binary or mixed gas hydrate the cross section was reported to change due to the local environment. For instance, in the CH4-CO2 mixed hydrate system the occupancy ratio of CH4 in the small cage to large cage calculated by the Raman spectra varied ∼11% from that obtained by the NMR technique.43 This result may reflect the influence of CO2 in the neighboring cages, which affects the polarizability derivatives and consequently the relative intensities of the respective Raman lines of the CH4 molecules. Accordingly, the importance of this effect in the CH4-C2H6 mixed hydrate systems is certainly not fully clarified at this stage. However, it is reasonable to assume that a small amount of neighboring components may only weakly affect σ and thus could be ignored
Methane-Ethane Gas Replacement in Clathrate-Hydrates
Figure 4. Reaction kinetics of CH4-C2H6 sII gas hydrates formed on two types of deuterated C2H6 sI hydrates (square ) as-synthesized consolidated; circle and triangle ) crushed) were exposed to free CH4 gas while collecting time-resolved neutron diffraction data at 5 MPa. Each symbol shows a value of weight phase fraction of the constituent phases (R) obtained from Rietveld refinements. In the case presented here R gives the degree of transformation in terms of the newly formed phase of CH4-C2H6 sII hydrate.
in a first approach to describe the complex mixed hydrate systems.44 Because of the unavailability of the respective σ data of the guest species in different hydrate cages, it seems to be difficult to accurately calculate a CH4/C2H6 ratio in clathrate hydrates. In the present report, our calculations are based on the following three assumptions: (I) σ of C2H6 (C-C) in all available large cages (LCI and LCII) are similar, (II) σ of C2H6 (σ ) 13; λ ) 488 nm; C-H; 2954 cm-1)42 in all large cages are similar in the range 2800-3200 cm-1, and (III) σ of CH4 (σ ) 7.6; λ ) 488 nm; C-H; 2917 cm-1)42 in all large cages are similar in the range 2800-3200 cm-1. The spectra were baseline corrected and fitted (Gaussian/Lorentzian) with the “dmfit” software45 that allowed multiple peak fitting procedure for a precise determination of the integrated band intensities. We estimated the volume concentration of C2H6 sI hydrates from the relative intensity of C-C stretching mode of C2H6 in the large cages (LCI and LCII). The number density of LCI is approximately 2.21 times more than that of LCII, as calculated from the crystal structures at Raman measuring p-T conditions.46 Thus we estimated the volume concentration of sI hydrate from the relative intensity of C-C stretching spectra of C2H6 from (IELCI/θELCI)/[(IELCI/θELCI) + 2.21(IELCII/θELCII)], where IELCI and IELCII refer to integrated intensity of C2H6 in the LCI and LCII, respectively and θELCI and θELCII refer to cage occupancies of C2H6 in the LCI and LCII, respectively. The overall ratio of C2H6 to CH4 in the LCII (IC2/IC1) was calculated from the integrated intensities of the C-H stretching modes. 3. Results 3.1. Time-Resolved Neutron Experiments. The CH4-C2H6 sII hydrates formed instantaneously when the pure C2H6 sI hydrate was exposed to free CH4 gas at 5 MPa. The phase fraction R of CH4-C2H6 sII hydrates obtained from in situ neutron diffraction data Rietveld refinements is plotted as a function of time in Figure 4. A variation of growth kinetics can be seen as function of both temperature and surface area of the starting materials. For instance, using similar starting samples (crushed C2H6 sI hydrates) at 5 MPa, a temperature (279 K) just above the D2O ice melting point converted 6% more CH4-C2H6 sII hydrates (R ) 0.43) than that was produced at 270 K (R ) 0.37) for a reaction period of 7 h. The surface area dependency of the reaction kinetics seems to be larger, as during the same period (7 h) the as-synthesized consolidated hydrates
J. Phys. Chem. A, Vol. 114, No. 1, 2010 251 converted only to R ) 0.056 at a temperature between 260 and 275 K, which is very considerably less than what was produced at 270 K using crushed hydrates. The reaction curves show a nonlinear fast formation for the initial stages followed by an ever slowing down growth of the new sII phase. This suggests an increasing diffusion limitation similar to the behavior observed for hydrate formation from ice particles and described by shrinking core models.7,34 The in situ neutron diffraction data provided very useful information to follow the structural changes that occurred during the gas replacement reactions in hydrates. A deuterated host structure was used to avoid the very strong incoherent scattering background arising from the H2O water framework;26 the difference between deuterated and hydrogenated samples used for the ice-to-hydrate formation kinetics was found to be insignificant by earlier work.7 It was reported that a CH4 pressure of 2000 psi (13.8 MPa) in the sample container produced significantly higher incoherent scattering background from hydrogen atoms of CH4, making it difficult to observe the diffraction peaks.47 Yet, the instrument D20 has a much higher flux and better signal-to-noise ratio than that of the diffractometer used in ref.47 Because of the applied lower pressure of 5 MPa the diffraction peaks emerged very well from the incoherent background, originating form the hydrogenated guest molecules in hydrates and gas phase, and allowed for very meaningful and conclusive Rietveld refinements. 3.2. Micro Raman Investigations. Typical Raman spectra of the starting C2H6 sI hydrates are shown in Figure 2. At 113 K and ambient pressure, the modes at 2890 and 2945.6 cm-1 have been assigned to the resonance doublets of C2H6 in the LCI of deuterated C2H6 sI hydrates. A band at ∼1000.5 cm-1 refers to C-C stretching mode (ν3)48 of C2H6 in the LCI. A weak band at 1020.5 cm-1 relates to the C-C stretching mode of C2H6 in the 512 cages, which clearly reveals that there is a small amount of C2H6 entrapped into SCI. All the frequency values are close to the ones previously reported for deuterated C2H6 sI hydrates.10 Typical Raman spectra of the recovered samples (i.e., after the in situ neutron diffraction experiments) are given in Figure 5. The resonance doublets of C2H6 at 2887 and 2942.6 cm-1 showed a lower shift of ∼3 cm-1 than those observed for pure C2H6 sI hydrate, and the ν3 band now can be seen at ∼992 cm-1. Clearly, the magnitude of the shifts both in the C-H and C-C regions indicates that the initial C2H6 sI hydrate transformed into a C2H6 containing sII hydrate. Moreover, two new bands appeared at 2903.2 and 2914.5 cm-1, which were respectively assigned to the totally symmetric C-H stretching bands of CH4 (ν1)48 in the LCII and SCII of CH4-containing sII hydrates.10 The noticeable intensity (Fermi resonance effect due to slight interaction of 2ν2 with the strongest Raman line ν1) of a band at ∼3055 cm-1 was assigned to 2ν2 of CH4 in the hydrate cages.10,44 Thus, the Raman results show clear and consistent evidence that the starting pure C2H6 sI hydrate was transformed into a mixed CH4-C2H6 sII hydrate. 4. Discussion During the ice-into-hydrate conversion a thin gas hydrate film was reported to rapidly spread over the ice surface at the initial stage.7,34,47,49–52 The only way to continue the subsequent clathration is the transport of gas molecules through the intervening hydrate layer to the ice-hydrate interface, and/or of water molecules from the ice core to the outer hydrate-gas interface. In comparison, the CH4-C2H6 sII hydrate formed on the surface of C2H6 sI hydrate. As schematically shown in Figure
252
J. Phys. Chem. A, Vol. 114, No. 1, 2010
Murshed et al.
Figure 6. Schematic diagram of the growth of CH4-C2H6 sII hydrate shell around the shrinking C2H6 sI hydrate cores during the CH4-C2H6 replacement that takes place in the replacement front. The in-bound diffusion of CH4 and out-bound diffusion of C2H6 gases through the CH4-C2H6 sII hydrate shell are shown by arrows. The zones with different gray-shading are intended to represent a continuous compositional gradient from outside to the unreacted core (and not homogeneous layers with a certain thickness).
Figure 5. Typical Raman spectra of C-H and C-C stretching bands of CH4-C2H6 sII hydrate samples recovered after the in situ neutron diffraction experiments at 5 MPa. MSCII, MLCII, and ELCII represent CH4 in 512, CH4 in 51264, and C2H6 in 51264 cages of CH4-C2H6 sII hydrates, respectively. ELCI represent C2H6 in 51262 cages of pure C2H6 sI hydrates. 2ν2 refers to the overtone of the respective fundamental of CH4 in the hydrate cages. One of the resonance doublets [(ν8 + ν11) and 2ν8] is shown by an asterisk. (a) Four different grains of the sample formed when crushed C2H6 sI hydrate was exposed to free CH4 gas at 270 to 280 K for ∼14 h. (b) Five different grains of the sample formed when crushed C2H6 sI hydrate was exposed to free CH4 gas at 279 K for ∼9 h. (c) Four different grains of the sample formed when assynthesized consolidated C2H6 sI hydrate was exposed to free CH4 gas at 260 to 275 K for ∼7.5 h.
6, once a mixed CH4-C2H6 sII hydrate forms as surface coverage, a (fictitious) spherical C2H6 sI hydrate grain shrinks and its radius r0 decreases due to the inward growth of a CH4-C2H6 sII hydrate layer (r0 - ri). Accordingly, the structural transformation takes place at the replacement front with a rearranged composition of the constituents, and the CH4-C2H6 sII hydrate layer acts as diffusion barrier both for the in-bound CH4 molecules moving toward the replacement front and the out-bound C2H6 molecules moving toward the gas phase. The subsequent growth of CH4-C2H6 sII hydrate is likely to depend on several factors that are of interest when designing a
phenomenological model to explain the inherent kinetics. Most notably they are the (i) particle size of the starting C2H6 sI hydrate (r0), (ii) the permeation coefficients for inward/outward moving gas molecules, (iii) excess fugacity of the gases involved, and (iv) the exchange rate/mobility of water molecules between C2H6 sI and CH4-C2H6 sII hydrates. The composition of the mixed hydrate and the size of the C2H6 sI kernel (ri) or the thickness of the newly formed CH4-C2H6 sII hydrate layer (r0 - ri) will depend on these parameters. Consistent with the first factor, the fast nonlinear, however, smooth development of the two reactions for the crushed samples can undoubtedly be related to their higher specific surface areas. Because the crushed hydrates were not sieved, a broad grain size distribution down to very fine particles results, in which the smallest particles are easily being fully transformed in the initial part of the reaction. Moreover, the grains of the crushed samples often have an irregular shape (as evidenced from cryo scanning electron microscopy) compared with the almost cylindrical shape of the consolidated samples. Ignoring surface microfractures that may further vary the specific surface area, the specific surface area of the consolidated samples is approximately 2 orders of magnitude less than that of the crushed ones, suggesting considerably slower exchange kinetics (Figure 4). The surface area has a pronounced effect in the initial part of the reaction as there is no permeation hindrance at this stage. In shrinking core models for hydrate formation,34,47 one distinguishes usually between a reaction-limited transformation at the surface and a diffusion-limited reaction in the bulk of the starting material; a similar approach seems to be indicated also for a future analysis of the exchange reaction. Care has to be taken in the analysis not to confuse gasexchanged hydrates from hydrates formed with free water or ice residing in the reaction cell, as they are not easily discerned in the end product. The in situ neutron diffraction technique, however, allows for the identification of unreacted ice before the start of the reaction. Thus, this amount can be accounted for in the analysis of the exchange process. Yet, our initial C2H6 sI hydrate contained almost no ice either from the sample handling (e.g., by a possible H2O frost contamination) or from
Methane-Ethane Gas Replacement in Clathrate-Hydrates any unreacted D2O ice from the hydrate formation reaction; this is evidenced by the synchrotron X-ray diffraction data (Table 1). Thus, the analysis of our diffraction data could be done in a straightforward way. It is interesting to note that none of the individual Rietveld refinements based on neutron data sets taken at different time intervals did show any significant amount of D2O ice. Because of the high sensitivity of the D20 diffractometer (allowing to detect phases at the percent level) the formation of any significant amount of ice during the exchange process can be excluded. The sensitivity to minority phases however is much worse for liquids due to their very broad diffraction features. Thus, we cannot exclude the transient formation of some (supercooled) liquid water during the exchange reaction. Water was reportedly formed during CO2(CH4 sI)14,53 replacement experiments at temperatures higher than ice melting point. Since the water density of the starting C2H6 sI (43.8 kmol/m3) and the transformed CH4-C2H6 sII (44 kmol/m3) hydrates does not significantly differ, water from the C2H6 sI hydrate is likely to rearrange locally in the replacement front forming CH4-C2H6 sII hydrates. This is in contrast to the rate-limiting water transport to the outer surface of the reacting grain in case of ice-into-hydrate conversion.7,34 Altogether, the diffraction evidence strongly suggests that, at least for temperatures below the ice point, the water is reused locally for forming the exchanged hydrate at the reaction site (Figure 6). This agrees fully with the conclusions drawn from earlier neutron diffraction work on the Ar-sII hydrate into CO2-sII hydrate conversion20 but can be stated now with higher confidence due to the much higher sensitivity of the neutron instrument used. As shown in Figure 5, the Raman spectra of different grains of the same sample differ in relative intensities and frequency shifts, both in the C-H and C-C stretch regions. Clearly, the spectra demonstrate a local variation in terms of structure as well as gas composition. It is therefore of particular note that one may hit one or another phase (sI/sII) in each Raman measurement of the same hydrate sample. The hydrate structure types even differ within a single polycrystalline grain. This was found when a Raman laser beam (estimated confocal length of ∼16 µm) was focused on a single hydrate grain, and the spectra were collected going down into the grain from the upper surface to the lower one tuning the z-axis of the microscope. At this stage it is worth noting that upon focusing the laser beam into a sample, the confocal length increases due to the laser light refraction at the air-crystal interface. Estimates based on Everall41 suggest that an additional confocal length of ∼13 µm may have been reached when focusing 100 µm into a sample with refractive index of ∼1.3-1.4 using an objective with a numerical aperture of 0.55. Scanning at 10 µm scale, the hydrates differ in gas composition as can be seen in the plot IC2/IC1 ratio versus z-axis in Figure 7; moving the focus through a partly exchanged and transformed grain shows a clear maximum at the grain center. Likewise, the volume concentration of C2H6 sI hydrate shows a similar peak-maximum almost at the same position of the z-axis. These results can be understood in terms of CH4-C2H6 sII hydrate layers formed with an increasing C2H6/CH4 ratio from surface to replacement front. The most likely reason is the limited permeation of the in-bound CH4 and out-bound C2H6 molecules through the hydrate with a core that still remains as pure C2H6 sI hydrate (Figure 6). Clearly, the advance of the reaction depends on the initial grain size and such an observation may no longer be made if the starting hydrate grain is small, or its shape is highly anisotropic. From the locally resolved Raman measurements
J. Phys. Chem. A, Vol. 114, No. 1, 2010 253
Figure 7. Relative volume fraction of C2H6 sI hydrate calculated from the integrated intensity of the C-C Raman stretching band measured at 113 K and ambient pressure. The remaining value of each square is a measure of the CH4-C2H6 sII hydrate. IC2/IC1 refers to the integrated intensity ratio of the C-H stretching band of C2H6 to CH4 located in available large cages (51262 and 51264). The sample was produced when crushed C2H6 sI hydrate was exposed to free CH4 gas at 270 K - 280 K and 5 MPa during in situ neutron diffraction for about 14 h.
through partially exchanged grains we conclude that the CH4(C2H6 sI) replacement reaction is a regrowth process commencing at the surface of the parent C2H6 sI hydrate with a shrinking core of unreacted material; it is a regrowth (and not a replacement) as the structure type changes upon exchanging gas. Moreover, C2H6 cannot escape from the hydrate cages without a local destruction of the crystal structure. Evidently, we observed a readjustment of the gas composition in the newly built cages of CH4-C2H6 sII hydrates according to the prevailing chemical activities at the replacement front. The LCII are occupied by both CH4 and C2H6, while the LCI of C2H6 sI hydrates are only occupied by C2H6. The relative concentrations vary, however, and must reflect an increasing imbalance of the chemical activities at the replacement front when compared to the (also slowly changing) activities in the gas phase. Knowing the total gas volume of the cell and using CSMGEM1 we can estimate the expected composition if a complete equilibration of the inbound and outbound gases could take place and compare the result with our measurements. From the Raman spectra of the observed single grain we then find locally a 20 to 40-fold of C2H6 (LCII) over the expected value from CSMGEM1 calculation, which translates into an increase of the CH4-C2H6 ratio of similar magnitude. This very considerable deviation from the expected composition is likely to occur due to (1) incomplete mixing of the free gases in the cell and/or (2) low and probably also different permeation coefficients for CH4 and C2H6, leading to an increasingly C2H6-enriched regrown sII hydrate toward the grain center (Figure 6). When a gas mixture reacts with water, thermodynamic models predict that at a given temperature the gas component with lower dissociation pressure has a greater driving force thus a higher degree of inclusion than the other component. The composition of the mixed gas hydrates is often significantly different from the parent gas phase and may result in a formation of hydrates in a structure type different from the one predicted for the parent phase composition. For instance, in the CH4-C3H8-H2O mixed hydrate system54 at the initial stage of reaction (with a preferential inclusion of C3H8 in the formed CH4-C3H8 sII hydrates) the partial pressure of C3H8 in the feed composition dropped lower than the equilibrium pressure of pure C3H8 hydrate, and only a pure CH4 hydrate continued to form afterward. Since our replacement reaction occurs in a closed system, that is, the exchange gas was not refreshed, some change of the gas phase will occur as a result of mixing
254
J. Phys. Chem. A, Vol. 114, No. 1, 2010
with the gas evolving from the initial hydrate (mostly in the early stages of the replacement reaction due to fast and complete transformation of the surface). It is interesting to note that, assuming a complete equilibration of the gas phase, a composition (namely 74-99.3 mol % of CH4)8 that promotes a CH4-C2H6 mixed hydrate of structure type II would prevail. Yet, as we stated above, this complete equilibration is not achieved, and the observation of a sII hydrate may be a result of the initial formation in the surface layer (which can be expected to be very methane-rich) with a continued and possibly increasingly metastable growth of sII when the local methane composition falls below the threshold for a sII mixed hydrate. However, it is also possible that the formation of a sII mixed hydrate has a lower nucleation threshold and the transformation into the thermodynamically stable form does not take place on the rather short time scale explored in the experiment, similar to earlier observations in the CO2 hydrate,7 N2 hydrate,55 or CH4-C2H6 mixed hydrate.10 A complete picture of the gas replacement kinetics has not yet been obtained. The process is complex and one must consider some salient physical parameters such as particle size, subtle local endothermic dissociation, or exothermic formation effects, the hydrate microstructure, permeation constants of the gases involved as well as the ongoing change of composition of the vapor phase. In particular, the microstructure of the initial gas hydrate is likely to play an important role in the replacement kinetics. It is established that hydrate formation on an existing hydrate surface is faster than the formation of new hydrate as the initial hydrate can act as a template, for example, the surface of CH4-C3H8 sII hydrate was found to help nucleate CH4 sI hydrate.54 Moudrakovski et al.56 showed that the starting THF sII hydrate led to a shorter induction period for Xe sI hydrate than it required to form on pure ice. Likewise, to what extent the surface of the C2H6 sI hydrate leads to a faster nucleation followed by a diffusion-limited growth of the mixed CH4-C2H6 sII hydrate would be of future interest in order to model the over all kinetics. 5. Conclusion Pure C2H6 sI gas hydrates were transformed into CH4-C2H6 sII hydrates by exerting CH4 gas pressure while studying the replacement kinetics using time-resolved neutron diffraction at different temperatures. The accessible surface area of the starting C2H6 sI gas hydrates plays an important role in the replacement kinetics. Raman spectra of different grains of the same partly transformed samples showed local variations in composition, which is mainly due to a broad grain size distribution, and their corresponding different degree of replacement/transformation. Spatially resolved Raman spectra demonstrated that gas replacement most likely proceeds forming a particle with compositional gradient, that is, formation of CH4-C2H6 sII mixed hydrate with continuous compositional change around the still unreacted C2H6 sI hydrate core. In this regard, the shrinking core diffusion model34 seems to be well suitable and may eventually provide diffusion constants and activation energies for the rate-limiting step of the exchange reaction. To achieve this goal neutron data sets at several different, constant temperatures and extending toward longer times need to be collected. Finally, one needs to be aware of possible deviations from a straightforward thermodynamic expectation for the hydrates formed, be it (1) because of local compositional deviations due to diffusion limitations or (2) because of a metastable formation of hydrates due to a lower nucleation barrier. Acknowledgment. We thank Dr. Thomas C. Hansen (ILL, Grenoble) for his excellent support during the neutron diffraction
Murshed et al. experiment and Dr. Helmut Klein (GZG, Abt. Kristallographie, Go¨ttingen) for his competent help with the synchrotron data collection. We thank ILL, Grenoble and HASYLAB at DESY for providing beam time and support. This work was funded in part by the German Ministry of Education and Research (BMBF) and the German Research Foundation (DFG) in the programme GEOTECHNOLOGIEN of which it is publication GEOTECH1301. Note Added after ASAP Publication. This article published ASAP on October 28, 2009. In the Discussion section, paragraph 4, sentence 7, a reference number has been changed. The corrected version posted on November 2, 2009. References and Notes (1) Sloan, E. D., Jr.; Koh, C. Clathrate Hydrates of Natural Gases, 3rd ed., Chemical Industries Series; CRC Press: Boca Raton, FL, 2008. (2) Kvenvolden, K. A. Natural gas hydrate: Introduction and history of discovery. In Natural Gas Hydrate in Oceanic and Permafrost EnVironments (Coastal Systems and Continental Margins); Max, M. D, Ed.; Kluwer Academic Publishers: Dordrecht, 2000; pp 9-16. (3) Max, M. D.; Johnson, A. H.; Dillon, W. P. Economic geology of natural gas hydrate (Coastal systems and continental margins); Springer: Dordrecht, The Netherlands, 2006. (4) von Stackelberg, M.; Mu¨ller, H. R. Z. Elektrochem. 1954, 58 (3), 1. (5) Ripmeester, J. A.; Tse, J. S.; Ratcliffe, C. I.; Powell, B. M. Nature 1987, 325, 135. (6) Schicks, J. M.; Ripmeester, J. A. Angew. Chem., Int. Ed. 2004, 43, 3310. (7) Staykova, D. K.; Kuhs, W. F.; Salamatin, A. N.; Hansen, T. C. J. Phys. Chem. B 2003, 107, 10299. (8) Subramanian, S.; Kini, R. A.; Dec, S. F.; Sloan, E. D. Chem. Eng. Sci. 2000, 55, 1981. (9) Ballard, A. L.; Sloan, E. D. Chem. Eng. Sci. 2001, 56, 6883. (10) Murshed, M. M.; Kuhs, W. F. J. Phys. Chem. B 2009, 13, 5172. (11) Kvenvolden, K. A. Org. Geochem. 1995, 23 (11-12), 997. (12) Dickens, G. R. Org. Geochem. 2001, 32, 1179. (13) Brewer, P. G.; Friederich, C.; Peltzer, E. T.; Orr, Jr., F. M. Science 1999, 284 (5416), 943. (14) Yoon, J. H.; Kawamura, T.; Yamamoto, Y; Komai, T. J. Phys. Chem. A 2004, 108, 5057. (15) Lee, H.; Seo, Y.; Seo, Y. T.; Moudrakovski, I. L.; Ripmeester, J. A. Angew. Chem., Int. Ed. 2003, 42, 5048. (16) Komai, T.; Kawamura, T.; Kang, S.; Nagashima, K.; Yamamoto, Y. J. Phys.: Condens. Matter 2002, 14, 11395. (17) Ota, M.; Saito, T.; Aida, T.; Watanabe, M.; Sato, Y.; Smith, R. L., Jr.; Inomata, H. AIChE 2007, 53, 2715. (18) Ota, M.; Abe, Y; Watanabe, M.; Smith, R. L.; Inomata, H. Fluid Phase Equilib. 2005, 228/229, 553. (19) Park, Y.; Kim, D. Y.; Lee, J. W.; Huh, D. G.; Park, K. P.; Lee, J.; Lee, H. Proc. Natl. Acad. Sci. U.S.A. 2006, 103 (34), 12690. (20) Halpern, Y.; Thieu, V.; Henning, R. W.; Wang, X.; Schultz, A. J. J. Am. Chem. Soc. 2001, 123, 12826. (21) Yeon, S. H.; Seol, J.; Lee, H. J. Am. Chem. Soc. 2006, 128, 12388. (22) Kvenvolden, K. A. ReV. Geophys. 1993, 31 (2), 173. (23) Paull, C. K.; Ussler III, W.; Dillon, W. P. In Natural Gas Hydrates in Oceanic and Permafrost EnVironments; Max, M.D, Ed.; Kluwer Academic Publishers: Dordrecht, The Netherlands, 2003; pp 149ff. (24) Brooks, J. M.; Kennicutt, M. C., II; Fay, R. R.; McDonald, T. J.; Sassen, R. Science 1984, 225 (4660), 409. (25) Kida, M; Khlystov, O; Zemskaya, T.; Takahashi, N; Minami, H.; Sakagami, H.; Krylov, A.; Hachikubo, A.; Yamashita, S.; Shoji, H.; Poort, J.; Naudts, L. Geophys. Res. Lett. 2006, 33, doi:10.1029/2006GL028296. (26) Kuhs, W. F.; Hansen, T. C. ReV. Mineral. Geochem. 2006, 63, 171. (27) Subramanian, S.; Sloan, E. D. J. Phys. Chem. B 2002, 106, 4348. (28) Sum, A. K.; Burruss, R. C.; Sloan, E. D. J. Phys. Chem. B 1997, 101, 7371. (29) Hendriks, E. M.; Edmonds, B.; Moorwood, R. A. S.; Szczepanski, R. Fluid Phase Equilib. 1996, 117, 193. (30) Subramanian, S.; Ballard, A. L.; Kini, R. A.; Dec, S. F.; Sloan, E. D. Chem. Eng. Sci. 2000, 55, 5763. (31) Ballard, A. L.; Sloan, E. D. Chem. Eng. Sci. 2000, 55, 5773. (32) Ohno, H.; Strobel, T. A.; Dec, S. F.; Sloan, E. D., Jr.; Koh, C. A. J Phys. Chem. A 2009, 113, 1711. (33) Kuhs, W. F.; Hensel, E.; Bartels, H. J. Phys.: Condens. Matter 2005, 17, 3009. (34) Kuhs, W. F.; Staykova, D. K.; Salamatin, A. N. J. Phys. Chem. B 2006, 110, 13283.
Methane-Ethane Gas Replacement in Clathrate-Hydrates (35) Larson, A. C, Von Dreele, R. B. General Structure Analysis System (GSAS); Los Alamos National Laboratory Report LAUR: Los Alamos, NM, 2000. (36) Toby, B. H. J. Appl. Crystallogr. 2001, 34, 210. (37) Ro¨ttger, K.; Endriss, A.; Ihringer, J.; Doyle, S.; Kuhs, W. F. Acta Cryst. B 1994, 50, 644. (38) Van Nes, G. J. H.; Vos, A. Acta Crystallogr. B 1978, 34, 1947. (39) Udachin, K. A.; Ratcliffe, C. I.; Ripmeester, J. A. J. Supramolecul. Chem. 2002, 2, 405. (40) Rawn, C. J.; Rondinone, A. J.; Chakoumakos, B. C.; Marshall, S. L.; Stern, L. A.; Circone, S.; Kirby, S. H.; Jones, C. Y.; Toby, B. H.; Ishii, Y. Proceedings of the Fourth International Conference on Gas Hydrates, Yokohama, Japan, May 19-23, 2002; p 595. (41) Everall, N. J. Appl. Spectrosc. 2000, 54, 773. (42) Burke, E. A. J. Lithos 2001, 55, 139. (43) Wilson, L. D.; Tulk, C. A.; Ripmeester, J. A. Proceedings of the Fourth International Conference on Gas Hydrates, Yokohama, Japan, May 19-23, 2002; p 614. (44) Chazallon, B.; Focsa, C.; Charlou, J. L.; Bourry, C.; Donval, J. P. Chem. Geol. 2007, 244, 175. (45) Massiot, D.; Fayon, F.; Capron, M.; King, I.; Le Calve, S.; Alonso, B.; Durand, J.; Bujoli, B.; Gan, Z.; Hoatson, G. Magn. Reson. Chem. 2002, 40, 70. (46) Uchida, T.; Takeya, S.; Kamata, Y.; Ikeda, I. Y.; Nagao, J.; Ebinuma, T.; Narita, H.; Zatsepina, O.; Buffett, B. A. J. Phys. Chem. B 2002, 106, 12426.
J. Phys. Chem. A, Vol. 114, No. 1, 2010 255 (47) Wang, X.; Schultz, A. J.; Halpern, Y. J. Phys. Chem. A 2002, 106, 7304. (48) Herzberg, G. Infrared and Raman spectra; D Van Nostrand Company Inc.: New York, 1951. (49) Stern, L. A.; Hogenboom, D. L.; Durham, W. B.; Kirby, S. H.; Chou, I. J. Phys. Chem. B 1998, 102, 2627. (50) Henning, R. W.; Schultz, A. J.; Thien, V.; Halpern, Y. J. Phys. Chem. 2000, 104, 5066. (51) Takeya, S.; Hondoh, T.; Uchida, T. Ann. N.Y. Acad. Sci. 2000, 912, 973. (52) Genov, G.; Kuhs, W. F.; Staykova, D. K.; Goreshnik, E.; Salamatin, A. N. Am. Mineral. 2004, 89, 1228. (53) Ersland, G.; Husebø, J.; Graue, A.; Baldwin, B. A.; Howard, J.; Stevens, J. Chem. Eng. J. [Online early access]. DOI:10.1016/ j.cej.2008.12.028. (54) Uchida, T.; Moriwaki, M.; Takeya, S.; Ikeda, I. Y.; Ohmura, R.; Nagao, J.; Minagawa, H.; Ebinuma, T.; Narita, H.; Gohara, K.; Mae, S. AIChE J. 2004, 50, 518. (55) Chazallon, B.; Kuhs, W. F. J. Chem. Phys. 2002, 117, 308. (56) Moudrakovski, I. L.; Ratcliffe, C. I.; Ripmeester, J. A. Angew. Chem., Int. Ed. 2001, 40, 3890.
JP908016J
