Time-resolved in situ neutron diffraction studies of gas hydrate: Transformation of structure II (sII) to structure I (sI)

Article abstract of DOI:10.1021/ja010280y

We report the in situ observation from diffraction data of the conversion of a gas hydrate with the structure II (sII) lattice to one with the structure I (sI) lattice. Initially, the in situ formation, dissociation, and reactivity of argon gas clathrate

Full text of DOI:10.1021/ja010280y

12826
J. Am. Chem. Soc. 2001, 123, 12826-12831
Time-Resolved in Situ Neutron Diffraction Studies of Gas Hydrate:
Transformation of Structure II (sII) to Structure I (sI)
,†
†
‡
‡
Yuval Halpern,* Vu Thieu, Robert W. Henning, Xiaoping Wang, and
Arthur J. Schultz*,‡
Contribution from the Energy Systems DiVision, Argonne National Laboratory, Argonne, Illinois 60439,
and Intense Pulsed Neutron Source, Argonne National Laboratory, Argonne, Illinois 60439
ReceiVed January 31, 2001
Abstract: We report the in situ observation from diffraction data of the conversion of a gas hydrate with the
structure II (sII) lattice to one with the structure I (sI) lattice. Initially, the in situ formation, dissociation, and
reactivity of argon gas clathrate hydrate was investigated by time-of-flight neutron powder diffraction at
temperatures ranging from 230 to 263 K and pressures up to 5000 psi (34.5 MPa). These samples were prepared
from deuterated ice crystals and transformed to hydrate by pressurizing the system with argon gas. Complete
transformation from D2O ice to sII Ar hydrate was observed as the sample temperature was slowly increased
through the D2O ice melting point. The transformation of sII argon hydrate to sI hydrate was achieved by
removing excess Ar gas and exposing the hydrate to liquid CO2 by pressurizing the Ar hydrate with CO2.
Results suggest the sI hydrate formed from CO2 exchange in argon sII hydrate is a mixed Ar/CO2 hydrate.
The proposed exchange mechanism is consistent with clathrate hydrate being an equilibrium system in which
guest molecules are exchanging between encapsulated molecules in the solid hydrate and free molecules in
the surrounding gas or liquid phase.
Introduction
The literature contains many experimental reports in which
different ratios of two or more gases are premixed and then
typically reacted with water to map the pressure-temperature
Clathrate hydrates are a general class of compounds composed
of water and gas molecules. They usually form at high pressures
1
phase diagram and to investigate which of two or three structure
and low temperatures although a few are stable at ambient
pressure. The natural occurrence of gas hydrates, as well as
environmental concerns over the increase of carbon dioxide in
the atmosphere, have led to speculation that CO2 could be stored
in hydrate form on the ocean floor.2 This could be achieved
by sequestering the excess CO2 within a hydrate framework.
Moreover, it may be possible to simultaneously extract the
methane from methane hydrate reserves on the ocean floor and
replace the guest molecule with carbon dioxide.5
7-10
types formed.
However, the experimental results described
in this paper are different from others in the literature since
they involve immersing a prepared sample of one hydrate (Ar
hydrate) with one type of hydrate structure (sII) in liquid CO2,
which is a hydrate former of another structure type (sI).
Gas hydrates typically form one of two structures, labeled
-4
structure I (sI) and structure II (sII), although a third lattice
1,11
named structure H (sH) is known.
Structure type I has
primitive cubic unit cell with contents denoted as 6X‚2Y‚48H2O,
where X and Y represent the large, 14-hedra, and small, 12-
hedra, cavities which can accommodate a guest atom or
molecule. Structure type II is face-centered cubic with contents
of 8X‚16Y‚136H2O, where X and Y represent 12-hedra and
At temperatures and pressures found a kilometer or more in
depth in the ocean, carbon dioxide is a liquid. In experiments
in which CO2 is transported to the ocean floor, it is shown to
pool as a liquid and react with the seawater to produce CO2
6
hydrate. However, it could also interact with other existing
1
6-hedra cavities, respectively. For a particular gas or guest
hydrates. Because of this, the environmental and safety issues,
and possible potential benefits, of direct exposure of hydrates
on the ocean floor to liquid carbon dioxide need to be examined.
molecule, the structure type formed appears to be dependent
on the size of the guest molecule, as well as how easily it can
be accommodated by the cavities of the crystal lattices.
1,11
Carbon dioxide is known to form sI, whereas argon forms sII.
In this contribution, we report the observation of the conversion
of a sII hydrate to sI by immersing the sII Ar hydrate in liquid
CO2 (a structure I former) under pressure. It should be noted
that although methane hydrate forms structure I, natural gas
*
Corresponding authors.
Energy Systems Division.
Intense Pulsed Neutron Source.
†
‡
(1) Sloan, E. D., Jr. Clathrate Hydrates of Natural Gases, 2nd ed.; Marcel
Dekker: New York, 1998.
(2) Herzog, H.; Golomb, D.; Zemba, S. EnViron. Prog. 1991, 10, 64-
7
4.
(
(7) Saito, S.; Kobayashi, R. AIChE J. 1965, 11, 96-99.
(8) Handa, Y. P.; Ratcliffe, C. I.; Ripmeester, J. A.; Tse, J. S. J. Phys.
Chem. 1990, 94, 4363-4365.
3) Nishikawa, N.; Morishita, M.; Uchiyama, M.; Yamaguchi, F.;
Ohtsubo, K.; Kimuro, H.; Hiraoka, R. Energy ConVers. Manage. 1992, 33,
51-657.
4) Saji, A.; Yoshida, H.; Sakai, M.; Tanii, T.; Kamata, T.; Kitamura,
H. Energy ConVers. Manage. 1992, 33, 643-649.
5) Ohgaki, K.; Takano, K.; Sangawa, H.; Matsubara, T.; Nakano, S. J.
Chem. Eng. Jpn 1996, 29, 478-483.
6) Brewer, P. G.; Friederich, G.; Peltzer, E. T.; Franklin M. Orr, J.
Science 1999, 294, 943-945.
6
(9) Subramanian, S.; Kini, R. A.; Dec, S. F.; Sloan, E. D., Jr. Ann. N.Y.
Acad. Sci. 2000, 912, 873-886.
(
(10) Montano, P. A.; Linton, J.; Thieu, V.; Halpern, Y. J. Synchrotron
Rad. 2001, 8, 972-974.
(
(11) Jeffrey, G. A. In Inclusion Compounds; Atwood, J. L., Davies, J.
E. D., MacNicol, D. D., Eds.; Academic Press: London, 1984; Vol. 1, pp
135-190.
(
1
0.1021/ja010280y CCC: $20.00 © 2001 American Chemical Society
Published on Web 12/01/2001

Neutron Diffraction Studies of Gas Hydrate
J. Am. Chem. Soc., Vol. 123, No. 51, 2001 12827
(
mixtures of hydrocarbons and other gases) hydrates found in
12
the Gulf of Mexico and possibly other regions are structure II.
We have also examined the kinetics of the argon hydrate system.
These experiments were carried out using the high-intensity
powder diffractometer (HIPD) instrument at the Intense Pulsed
Neutron Source (IPNS) at Argonne National Laboratory.
Experimental Procedures
Formation of Structure II Ar Hydrates. Powdered ice was
prepared by freezing deuterated water (Aldrich, 99.9%) in liquid
nitrogen and then crushing the ice in a mortar and pestle. Large ice
particles were removed with a 250 µm sieve, and the powdered ice
was placed in the pressure cell that was already cooled in liquid
nitrogen. On the basis of the weight of the water at the end of the
experiment, the amount of initial powdered ice in a fully loaded pressure
cell was 1.50 ( 0.05 g. The apparatus has been described in more
detail in an earlier publication.13 The pressure cell was closed, mounted
on the cold stage of a Displex closed-cycle helium refrigerator, and
placed into the sample chamber on the HIPD diffractometer. The
pressure cell was kept in liquid nitrogen throughout this process to
prevent the ice from melting and to minimize condensation buildup on
the outside of the cell. Before argon gas was introduced to the system,
the sample was allowed to stabilize at the working temperature (230,
Figure 1. Conversion of deuterated ice to argon hydrate at 34.5 MPa
(5000 psi) at various temperatures. Each data point represents the mole
fraction of hydrate refined from a 15 min histogram.
Results and Discussion
Data Analysis. Neutron diffraction data were collected in
15 min intervals starting with the initial introduction of the gas
into the system. The total length of the data collection depended
on the initial starting temperature, with longer times required
at lower temperatures. Time-of-flight neutron powder diffraction
data were obtained using the 90° data bank on the HIPD. Data
243, 253, or 263 K). The cell was then charged to 5000 psi (34.5 MPa)
with argon gas. Neutron data were collected in intervals of 15 min to
observe the transformation from ice to sII argon hydrate. Limited
instrument time did not allow for the complete conversion to hydrate
under these conditions, with conversion typically less than 50%.
Conversion of sII Ar Hydrates into sI Hydrates by Gas Exchange
1
4
were analyzed using the GSAS program.
The short data collection times and relatively low resolution
of the HIPD instrument did not allow for a full Rietveld analysis
of each data set. The lattice parameters of the phases observed
in the spectrum (argon hydrate, CO2 hydrate, ice, and aluminum)
were refined in the initial stages but then fixed since the
temperature and pressure of the sample did not change. The
atomic positions and thermal parameters were determined in
separate experiments for each temperature and were not varied
during the refinements. In addition to four background param-
eters, the histogram scale factor, an absorption coefficient, and
the phase fractions were allowed to refine. The weight fractions
were extracted from each refinement and plotted in terms of
mole fractions of hydrate.
2
with CO . In these samples, the argon hydrate was formed at 263 K
and 34.5 MPa in the manner described above. Complete conversion of
ice to hydrate was then obtained by keeping the sample pressurized
with 5000 psi (34.5 MPa) Ar and slowly warming it through the ice
melting point. Since D
was 276.8 K (3.8 °C). The sample temperature was increased from
63 to 278 K (0.8 K/hour) to obtain complete conversion. When the
2 2
O was used instead of H O, the melting point
2
conversion to hydrate was finished, the sample was cooled to a
predetermined, working temperature (230, 243, 253, or 263 K).
Excess argon gas was then released by opening a valve used to isolate
the sample from the atmosphere. This valve was kept open for several
minutes to vent all of the excess argon gas. Complete gas removal
was confirmed by closing the valve and noting the cell pressure
stabilized around 1 atm. Next, carbon dioxide at a pressure of 900 psi
Formation of sII Ar hydrates. The conversion of ice to
argon hydrate (Figure 1) is a temperature-dependent process
with ∼25% conversion occurring in 51, 15, 9, and 3 h at 230,
243, 253, and 263 K, respectively, under 5000 psi (34.5 MPa)
(6.2 MPa) was quickly introduced to the sample at the same temper-
ature. At these pressures and temperatures carbon dioxide is a liquid.
Neutron diffraction data were collected in 15 min intervals during the
initial, rapidly transforming stage, but were extended to 30 min intervals
after approximately 24 h. Data were collected for up to 160 h depending
on the temperature. At the end of this period, around 96% of the sII
Ar hydrates were converted to the sI type.
5
argon gas pressure. We have suggested that after an initial
period of fast conversion to hydrate on the ice particle surface,
the formation process is then controlled by the diffusion rate
of gas molecules through the accumulating hydrate layer. The
following equation describes a conversion process of a particle
from the outside to the inside during its diffusion-controlled
The gases encaged in the hydrate samples at this point were collected
for mass spectroscopy (MS) analysis. This was done for each of the
1
5
stage at a constant temperature:
four reaction temperatures. For this process, at 243 K excess CO
first vented to 1 atm in the manner described above. At 243 K, excess
CO present in the pressure cell was in the liquid form, and the venting
process took about 10 min, as the liquid had to boil off. The venting
temperature of 243 K was used in all trials to avoid solid CO formation
sublimation point of 194.5 K at 1 atm), and to minimize hydrate
2
was
1
/2
-(2k)
1
/3
1/2
1/3
(1 - R)
)
(t - t*) + (1 - R*)
(1)
2
(
)
r0
2
where k and r0 are the diffusion constant and the original radius
of the particles, and R and R* are degrees of reaction at times
t and t*. This equation has recently been used to fit the shrinking
core model of the hydration of cement grains
also be applicable to the growing inward of a hydrate layer on
(
dissociation (∼216 K for CO
2
hydrate decomposition temperature at 1
1
3
atm) which proceeds slowly at 243 K. A 150-cm evacuated gas
collection bottle was then attached to the experimental setup. The
sample was then heated from 243 to 298 K to dissociate the hydrate,
and the evolving gases were collected in the bottle for mass spectros-
copy (MS) analysis. The final pressure in the collection bottle was 30
psi (207 kPa). The mass spectrometer used was a Spectra Multi-Quad,
model LMI, manufactured by Leda-Mass.
1
6,17
and should
(
14) Larson, A. C.; Von Dreele, R. B. GSAS--General Structure Analysis
System. Report No. LAUR 86-748, Los Alamos National Laboratory, Los
Alamos, NM, 2000.
(15) Fujii, K.; Kondo, W. J. Am. Ceram. Soc. 1974, 57, 492-497.
16) FitzGerald, S. A.; Neumann, D. A.; Rush, J. J.; Bentz, D. P.;
(
(
12) Ripmeester, J. A. Ann. N.Y. Acad. Sci. 2000, 912, 1-16.
Livingston, R. A. Chem. Mater. 1998, 10, 397-402.
(17) Berliner, R.; Popovici, M.; Herwig, K. W.; Berliner, M.; Jennings,
H. M.; Thomas, J. J. Cem. Concr. Res. 1998, 28, 231-243.
(13) Henning, R. W.; Schultz, A. J.; Thieu, V.; Halpern, Y. J. Phys.
Chem. A 2000, 104, 5066-5071.

12828 J. Am. Chem. Soc., Vol. 123, No. 51, 2001
Halpern et al.
Figure 2. Plot of the experimental data of the conversion of deuterated
ice to argon hydrate in terms of the shrinking core model.
Figure 4. Neutron scattering patterns of sII Ar hydrate changes into
Figure 3. Plot of ln(k′) as a function of 1/T (K-1) for the conversion
of deuterated ice to argon hydrate. An activation energy of the diffusion
process to form argon hydrate was calculated from the slope of the
straight line.
sI hydrate at 243 K after excess Ar was vented and exchanged with
CO at 6.21 MPa: (a) Initial sII Ar hydrates at 243 K and 34.5 MPa
2
of Ar; (b) approximately 50% of the sII hydrates converted to sI
hydrates within 2 days after the excess Ar was replaced with 6.21 MPa
2
CO at 243 K; (c) 95% of the sII hydrates converted to sI hydrates
within 4 days after the gas exchange.
an ice particle.13 For each temperature, obtaining a straight line
by plotting (1 - R)1 as a function of (t - t*) indicates an
agreement with eq 1. For a given particle size, the diffusion
constant (k) can be calculated from the slope of the straight
line.
/3
1/2
needed to break the hydrogen bonding in ice (12.7 kcal/mol)18
but is greater than the energy needed to break the hydrogen
bonding in liquid water (5 kcal/mol). Analogous to the
1
formation of CO2 hydrate with an activation energy of 6.5 kcal/
The term t* indicates the time where the conversion process
is initially dominated by the diffusion of Ar molecules through
the hydrate layer. In our experiments, we selected the t*
corresponding to ∼5% conversion for each of the temperatures.
This produced the best linear fit of eq 1 to the data. At times
before or after 5% conversion, the fit greatly deviated from
linearity, indicating that diffusion became the dominant factor
in the reaction rate at t* corresponding to ∼5% conversion.
Assuming an average ice particle size of 200 µm, 5% conversion
corresponds to a hydrate layer with a thickness of approximately
13
mol, the argon hydrate-forming reaction is believed to occur
between argon molecules and internal water in the quasi-liquid
layer (QLL) rather than directly with ice molecules. The QLL
is believed to be a thin mobile phase of water molecules with
mobilities intermediate of those of liquid water and crystalline
19
ice.
Conversion of sII Ar Hydrate into sI Hydrate by Gas
Exchange with CO2. Neutron diffraction data in Figure 4 show
the transformation, at 243 K, of the sII Ar hydrate to sI hydrate
after the excess argon was replaced with CO2. Figure 4a shows
the sample of sII Ar hydrate at 243 K and 34.5 MPa. The hydrate
sample, 2 days after the gas exchange from Ar to CO2, is shown
in Figure 4b. Diffraction peaks indicate the presence of both sI
and sII hydrates. This sample was kept at 243 K for about 4
days in an effort to convert all sII hydrate to sI hydrate. After
the 4-day period, neutron diffraction data indicated about 4%
of the sample was sII hydrate and 96% sI hydrate (Figure 4c).
MS results from these trials are discussed in the next section.
The rate of conversion of sII Ar hydrate to the sI hydrate at
various temperatures is shown in Figure 5. The decomposition
1
000 sII unit cells. Figure 2 shows a plot of the data in terms
of eq 1 for each of the four temperatures.
The diffusion constants depend on the particle size and
temperature, but for a given particle size distribution (r0), the
activation energy of the diffusion process is independent of these
variables. The temperature dependence of a thermally dependent
diffusion process should follow Arrhenius behavior described
by k ) A exp[-Ea/(RT)], where R is the gas constant. If k′ is
2
defined as k/r0 , k′ can be calculated for each temperature from
the slopes in Figure 2.
Figure 3 shows the straight line obtained by plotting ln(k′)
against 1/T. An activation energy value of 9.7 kcal/mol was
calculated from the slope of this line. This is less than the energy
(
18) Itagaki, K. J. Phys. Soc. Jpn 1967, 22, 427-431.
(19) Mizuno, Y.; Hanafusa, N. J. Phys., Colloq. C1 Suppl. 1987, 48,
511-517.

Neutron Diffraction Studies of Gas Hydrate
J. Am. Chem. Soc., Vol. 123, No. 51, 2001 12829
Table 1. Results from Mass Spectrometry (MS) Analysisa
mol % CO
2
mol % Ar
total
100
100
100
100
100
2
2
2
2
2
30 K (70% conversion)
43 K (96% conversion)
43 K (96% conversion)
53 K (96% conversion)
63 K (96% conversion)
41.4
53.2
54.7
56.1
66.2
58.6
46.8
45.3
43.9
33.8
gas composition
average of 243 K trials)
calculated mol of gas
243 K trials)
54.0 ( 0.8 46.0 ( 0.8 100
(
0.0069 mol 0.0059 mol 0.0128 mol
(
a
Limited neutron beam time did not allow the 230 K experiment to
reach ∼96% conversion.
Table 2. 1Ratios of Guest Molecular Diameters to Cavity
Diameters
Guest-to-Cavity Diameter Ratio
hydrate
2
Figure 5. Conversion of sII Ar hydrate to sI hydrate after CO gas
guest
structure for guest in for guest in for guest in for guest in
12 12 2 12 12 4
exchange at various temperatures. Upward sloping curves indicate the
sI hydrate formation, whereas the downward sloping curve (for 243 K
trial) indicates the sII Ar hydrate decomposition.
molecule formed
5
of sI
5 6 of sI
5 of sII 5 6 of sII
CO2
Ar
sI
sII
1.003
0.745
0.874
0.648
1.020
0.757
0.769
0.571
the latter two cases the conversion process is not diffusion-
controlled until the ice particles are covered by a hydrate layer.
As mentioned above, we estimate 5% conversion corresponds
to a hydrate layer of about 1000 unit cells. In the conversion of
sII to sI hydrate, the initial phase is already a gas hydrate such
that a diffusion-controlled process through hydrate exists from
the very beginning, which is consistent with t* of zero. However,
we wish to emphasize that the values of t* are not obtained
with high precision, and therefore any inferences based on these
values should not be considered conclusive.
Composition and Occupancy of the sI Hydrates. From
mass spectrometry (MS) analysis of the gases collected from
the dissociating hydrate sample, the relative amounts of CO2
and Ar in the final mixture of hydrates at 243 K were 54.0 (
Figure 6. A plot of the experimental data of the conversion of Ar
hydrate (sII) to Ar/CO hydrate (sI) in terms of the shrinking core model.
2
0
.8 and 46.0 ( 0.8 mol %, respectively. Table 1 lists these
results and shows a trend of increasing CO2 mol % with
increasing temperature, presumably because at any given
pressure CO2 hydrate is more stable than Ar hydrate. The values
used in subsequent calculations were obtained from the average
of two gas exchange experiments at 243 K. On the basis of the
MS results and using the ideal gas equation, (PV ) nRT, where
3
P ) 207 kPa, V ) 155 cm , T ) 298 K, and R is the universal
gas constant) the amounts of CO2 and Ar collected in the
collection bottle were calculated to be 0.0069 and 0.0059 mol,
respectively.
Neutron diffraction data indicated about 4% of the sample
was sII hydrate and 96% sI hydrate. Assuming that the sII
hydrate sample does not contain any CO2, the amount of argon
forming the 4 mol % of argon sII hydrate is 0.0005 mol. This
amount is more than 1 order of magnitude less than the 0.0059
mol Ar determined from MS analysis. It is clear that the residual
4 mol % of the sII hydrate cannot account for the ∼45 mol %
of argon gas in the sample. The presence of the relatively large
quantity of argon in the hydrate sample indicates that the sI
hydrate formed by the introduction of CO2 is not pure CO2
hydrate, but mixed Ar/CO2 hydrate of the type sI lattice.
The case where the sII argon hydrate sample had converted
to 96% sI hydrate is considered below:
Figure 7. A plot of ln(k′) as a function of 1/T (K-1) for the conversion
of Ar hydrate (sII) to Ar/CO hydrate (sI).
2
rate of Ar hydrate (243 K trial) is also shown in the figure.
Similar calculations using eq 1 and a t* corresponding to
conversion of 0% gives an activation energy of 13.9 kcal/mol
for the structure change process (Figures 6 and 7). This value
is higher than the energy needed to break hydrogen bonding in
either water (5 kcal/mol) or ice (12.7 kcal/mol).
It is of interest to compare the t* value of zero for the sII to
sI hydrate conversion to the nonzero t* values of 5% for ice to
Ar hydrate and 20% for ice to CO2 hydrate13 conversions. In
The ideal unit cell formulas for the two hydrate structures
are: structure I [2X‚6Y‚46D2O] and structure II [16X‚8Y‚
136D2O]; where X and Y are the small and large cavities in
the structures, respectively. Table 2 lists the guest-to-cavity-
size ratios for Ar and CO2 in the different cavities of sI and sII

12830 J. Am. Chem. Soc., Vol. 123, No. 51, 2001
Halpern et al.
Table 3. Moles of Cavities Available for Gas Occupancy Based
on 1.5 g D O (0.075 Mol) Converted into a Mixture of 96% sI and
% sII Hydrate
2
4
molecule
or cavity
structure I (96%)
structure II (4%)
[2X.6Y.46D
2
O] [16X.8Y.136D
2
O]
total
D
2
O in structure
0.072
0.003
0.075
cavities (total)
small cavities
large cavities
0.0125
0.0031
0.0094
0.0005
0.00033
0.00017
0.0130
0.00343
0.00957
hydrates. The guest-to-cavity-size ratio can be viewed as a
measure of the fit of the gas in the cavities, as well as a measure
of the extent of cavity stabilization by the guest. Guest-to-cavity-
size ratios between 0.75 and 1.0 are typical, with a ratio larger
than 1.0 implying that the guest is probably too large to stabilize
1
the cavity.
For the case of the 243 K experiments (see Tables 1 and 3),
nearly 100% of the cavities (0.0130 mol) are filled with either
Ar or CO2 (0.0128 mol). Carbon dioxide is known to occupy
both types of cages in structure sI. However, the fit of CO2 in
the 5 6 cage is better, due to size limitations of the molecule
in the 512 cage, as seen with a guest-to-small cavity-size ratio
of slightly greater than 1. If all of the CO2 (0.0069 mol) and
Figure 8. Effect of type of gas and its pressure on the rate of
decomposition of sII Ar hydrate at 243 K. (a) Hydrate is stable under
5
6
.86 MPa (850 psi) of Ar. (b) sII Ar hydrate decomposes slowly under
.21 MPa (900 psi) of He. (c) sII Ar hydrate decomposes at a moderate
2 2
rate under 6.21 MPa of CO to form sI Ar/CO mixed hydrate. (d) sII
Ar hydrate decomposes rapidly under atmospheric pressure to form
ice.
1
2 2
0
.0023 mol of argon were to occupy the large cavities of sI
In addition, the empirically observed value of zero for the time
t* where the sII-to-sI conversion is dominated by diffusion (vide
supra) also supports a one-step process. Although we cannot
entirely rule out the existence of an undetectable intermediate
phase, such as a quasi-liquid layer, with an activation energy
of 13.9 kcal/mol for the sII-to-sI transformation, an intermediate
phase with weaker hydrogen-bonding energies is no longer
required for the conversion to occur.
hydrate, giving a cage occupancy of 98%, then the small cavities
of sI could be completely occupied by the remaining argon gas
(0.0031 mol). Argon is known to form sII hydrate, even though
12
it can easily be accommodated by the 5 cavities of either
hydrate structure, which are of roughly the same size (see ratio
in Table 2). However, it is more likely that there is a mixture
of both gases in both cavities.
Probable Mechanism for Transformation of sII Ar Hy-
drate to Mixed Ar/CO2 sI Hydrate. Two possible mechanisms
can be envisioned for the transformation of the sII Ar hydrate
to the mixed Ar/CO2 type sI hydrate at temperatures below the
ice point: (1) ice as an intermediate: decomposition of the sII
Ar hydrate (at 6.21 MPa and 243 K) to form ice, followed by
the conversion of the ice to the mixed gases product, and (2)
direct conVersion: the sII Ar hydrate converts directly, with no
ice intermediate, into the mixed Ar/CO2 type sI hydrate.
However, this does not rule out the possibility of the involve-
ment of a quasi-liquid layer in the conversion mechanism.
Experiments conducted with Ar hydrate at 243 K and 5.86
MPa (850 psi) of Ar showed the hydrates were stable under
these P-T conditions (Figure 8a). Additionally, experiments
were conducted with Ar hydrate in which excess Ar gas was
replaced with He, a nonhydrate-forming gas, at 243 K at a total
pressure of 900 psi (6.21 MPa, the same as the CO2 pressure in
the exchange experiments). As depicted in Figure 8b, only
Hence, we suggest that gas hydrate is not a stagnant system,
but rather an equilibrium assemblage in which hydrogen bonds
open and close constantly and cavities are continually breaking
and reassembling. During the time that the cavity is open, the
guest molecule is free to leave the open cavity. A molecule of
the same gas or of a different gas might become the next guest.
The nature of the new guest determines the structure that results
from the exchange. If the gas is removed from the system at
such a rate that no other guest is available to stabilize the cavity
(such as in cases where the pressure in the system is reduced)
the cavity will not be reassembled. This destruction of the
structure will continue until enough pressure of a hydrate-
forming gas is rebuilt (e.g., if gas removal is stopped or a
different hydrate-forming gas is introduced) to force guests back
into damaged cavities and reassemble them. If the removal of
gas is continued, the cavities will continue to break until the
whole structure collapses. The rates of these breaking/reas-
sembling processes are temperature-dependent. Hence, exchange
or decomposition rates can be controlled significantly by
changing the temperature of the gas hydrate system.
∼
17% decomposition (of the hydrate to form ice) occurs during
the first 10 h after the exchange. Less than an additional ∼11%
decomposition took place in the following 10 h. In contrast,
the conversion of the sII Ar hydrates to the sI Ar/CO2 mixed
hydrates under 900 psi of CO2 proceeds at a significantly faster
rate under these pressure and temperature conditions (Figure
It should be emphasized that there are many reported studies
of mixed-gas hydrates in which the dependence of the formation
of either sI or sII has been investigated. For instance, mixtures
of Xe (sI former) and Kr (sII former) were shown to form sII
with an initial gas composition of less than 5 mol % of Xe, and
8
c). Approximately 40% conversion occurs within the first 10
h, with additional 20% during the following 10 h. Figure 7d
shows the rapid decomposition of Ar hydrate under atmospheric
pressure to form ice. The slower rate of decomposition (under
pressure of a nonhydrate-forming gas) than the rate of converting
to mixed Ar/CO2 sI hydrate (under the same pressure of CO2)
strongly suggests a direct conversion mechanism in which the
sII Ar hydrate converts into the sI Ar/CO2 mixed hydrate without
first decomposing into ice. The absence of ice peaks in the
neutron scattering pattern supports this proposed mechanism.
8
sI with greater than 5 mol % of Xe. The formation of either sI
or sII from mixtures of methane with other hydrocarbons has
9
,20
also been studied extensively.
These previous studies all
began with an initial composition of gases which were then
reacted with ice or liquid water to form hydrate, which was
then analyzed to determine its structure type. There are also
several studies examining the kinetics of hydrate formation from
(20) Holder, G. D.; Hand, J. H. AIChE J. 1982, 28, 440-447.

Neutron Diffraction Studies of Gas Hydrate
J. Am. Chem. Soc., Vol. 123, No. 51, 2001 12831
either liquid water or ice.13,21,22 However, to the best of our
knowledge, the study in this paper is the first to examine the in
situ transformation of one hydrate structure type to another by
exposing the hydrate to a second gas hydrate former.
with structure sI as shown by neutron powder diffraction. Once
an initial layer of sI hydrate forms around the sII hydrate
particles, the reaction kinetics are dominated by the diffusion
of CO2 through the outer sI hydrate layer. On the basis of mass
spectroscopic data, the sI hydrate contains a mixture of argon
and CO . We previously observed that CO hydrate is stable
Conclusions
2
2
under an applied pressure of helim gase which does not form a
As pointed out previously,1,8 small guests, such as argon, are
sII formers because the relative number of small cavities is
higher, with a ratio of 2:1 of small-to-large cavities. In the case
of sI, the ratio of small-to-large cavities is 1:3, which favors
larger guests that stabilize the larger cavities, such as carbon
dioxide. In this paper we have described the observation by
neutron diffraction of the conversion of a gas hydrate from
structure type sII to sI. Initially, pure argon hydrate is formed
with structure sII and is transformed to mixed argon/CO2 hydrate
1
3
hydrate. Argon hydrate under an applied pressure of helium
gas dissociates at a slower rate than it reacts with CO2 at the
same pressure and temperature. These results are consistent with
an equilibrium exchange between encapsulated guest molecules
in the solid hydrate and free molecules in the surrounding gas
or liquid phase.
Acknowledgment. The work at Argonne National Labora-
tory was supported by the U.S. Department of Energy, Basic
Energy Sciences-Material Sciences, under Contract No. W-31-
109-ENG-38.
(
21) Koh, C. A.; Savidge, J. L.; Tang, C. C. J. Phys. Chem. 1996, 100,
412-6414.
22) Takeya, S.; Hondoh, T.; Uchida, T. Ann. N.Y. Acad. Sci. 2000, 912,
73-982.
6
9
(
JA010280Y