American Mineralogist, Volume82, pages 88-98, 1997
Crystal-growth studiesof natural gas clathrate hydratesusing a pressurizedoptical cell
EucnNr A. Strnlmr,2,* ANDH.E. KrNc Jn.'
rExxon Researchand EngineeringCompany, Route 22 East, Annandale,
New Jersey08801, U.S.A.
'zDepaftmentof Geological and GeophysicalSciences,
Princeton University, Princeton,New Jersey08544, U.S.A.
Arsrnacr
The crystal-growth behavior of structure I (sI), structure II (sII), and structure H (sH)
clathrate hydrates has been studied using a specially designed, pressurizedoptical cell.
Single crystals of each hydrate type, methane sI, methane-propane(957o-5vo)sII, and
methane-methylcyclopentane
sH, were grown in equilibrium with aqueousliquid + vapor
-r liquid hydrocarbon.
Each structure type exhibits characteristic crystal morphology,
which suggeststhat crystal habit in natural settings,such as sea-flooroutcrops,may allow
visual identification of hydrate types. In addition, the relative growth rates for different
Miller planes for each crystal type were determined.The relative growth-rate schemesand
resulting crystal morphology of each structure can be related to the unit-cell density distribution of the small cagesin each structure.Four-phaseP-Z equilibrium data for methanemethylcyclopentanesH data were also measuredusing optical methods.Evaluation of these
and previously published phase-equilibrium data for all three known hydrate strucrures
strongly suggeststhat hydrate assemblagesof coexisting sII and sH should be common in
natural settings.
IxrnooucrroN
Clathrate hydrates are inclusion compounds based on
a three-dimensional ice-like framework of hydrogenbonded water molecules. This framework structure createsseveralpolyhedral cavities ofvarious dimensionsthat
house included, or guest molecules. In the natural environment, the guest moleculestrappedin this water framework are typically hydrocarbon gases. These gases are
mostly methane,ethane,propane, and isobutane;carbon
dioxide; and hydrogen sulfide. These compounds are
commonly referred to as gas hydrates.Clathratehydrates
also form from mixtures of water plus ethers,fluorinated
hydrocarbons,and chlorinated hydrocarbons.In-depth reviews of clathrate crystal chemistry and crystallography
were given by Davidson (1973), Jeffrey (1984), Sloan
(1990a),Englezos(1993), and Ripmeesteret at. (1994).
Naturally occurring gas hydrateshave received considerable attention over the last 15 years, ever since researchersbegan to realize their great abundance.Hydrate
stability in nature is a function of the pressure,temperature, and composition of both the gas and liquid phases.
Becauseof these controls, natural hydrate formations are
somewhatrestricted in location, occurring in deep-ocean
regimes and in polar regions. Deep-oceanhydrates usually occur in sedimentsalong the outer continental slopes
where water is cold enoughfor hydrate stabilization (e.g.,
Brooks et al. 1986, 1984; Davidsonet al. 1986a;Kvenvolden et al. 1993; Kvenvolden and Grantz 1990). In po* Presentaddress:
MoltenMetalTechnology,
lnc., 421Currant
Rd.,Fall River,Massachusetrs
02720,U.S.A.
0003-004x/97l0102--0088$05.00
88
lar regions hydrates are generally associatedwith permafrost in offshore and onshoresediments(e.g., Cherskiy
et al. 1985; Collett 1994; Kvenvoldenand Grantz 7990;
Makogan 1981).Recently,unusualair-clathratehydrates
were reported from about 1000 m depthsin the Antarctic
polar ice (Craig et al. 1993). Natural hydrate occuffences
were reviewedby Kvenvolden(1993, 1988a).
From the environmental and geologic points of view,
the importance of naturally occurring gas hydrates may
be consideredfrom three perspectives.First, becausehydrates contain considerable methane, they represent a
possible energy resource.Though the technology has not
yet been developedto economically extract the methane,
the known and inferred natural hydrate formations represent staggeringpotential energy reserves,estimated at
1027-1034
m3 of methane(e.9.,Holder et al. 1984; Kvenvolden 1994, 1988a;MacDonald 1990a).
Second,it has been suggestedthat hydratesmay influence atmospheric CHo and CO, levels as gas is either
releasedor fixed by hydrates(e.g., Kvenvolden 1988b;
MacDonald 1990b; Nisbet 1992). This atmosphericcycling of methane and CO, is thought to have influenced
climate conditions and the advance and retreat of glaciation throughoutgeologictime (e.g.,Kvenvolden1988b;
MacDonald 1990b; Nisbet 1990; Sloan et al. 1992 Paull
et al. 1991).
Third, the existence and behavior of natural hydrates
may play an important role in geologic processes,such
as turbidity flows of unconsolidatedsubmarinesediments
(e.g., Mclver 1977, 1982), sediment slumping and slope
instability along continental margins (e.g., Carpenter
SMELIK AND KING: CRYSTAL GROWTH OF CLATHRATE HYDRATES
89
TeeLe 1. Crystal-chemical
data for water-clathrate
hydrates
Struc- Spacegroup and
ture
cell parameters
(A)
rype
SI
sll
Pm3n (227)
a: 12.0
Fd3m (223)
a : 17.3
Cage types per unit cell (lettercode)512
(D)
51262
(r)
51264
(H)
435663
(D',)
5126s
lFl
16
I
relmmm (191)
a:123
c: 10.2
Structure references-'
ldeal stoichiometry
2D 6T 46H,O
10
16D.8H.136H,O
3 D . 2 D ' , . 13E4 H , O
(McMullanand Jetfrey1965)
(Mak and McMullan1965,
Davidsonet al 1986b,
McMullanand Kvick 1990)
(Gerkeand Gies 1984,t
Riomeesteret al 1987.
Ripmeesterand Ratcliffe1990)
* i\y'"superscript notation describes a cage with m total faces with N edges each. For example, 5f indicates a cage with 12 five-sided faces (a
pentagonaldodecahedron),
whereas5'26sdenotesa cage with 12 pentagonalfaces and eight hexagonalfaces
". Referencesfor structureretinementsor latticemeasurements
only
with sH hydrate
t Structurerefinementfor dodecasil1H, which is isostructural
1981; Kayen and Lee 1991),and submarinemud volcanoes (e.9.,Ginsburget al. 1992;Reed et al. 1990).
In addition to theseenvironmentaland geologic aspects
of hydrates,there has been considerableinterest and researchin gas hydrates by the oil and gas industry since
about 1940. This stems from the considerableeconomic
impact of gas hydrate formation in both drilling and exploration operations and in pipeline-transportoperations
(e.g., Barker and Gomez 1989; Lingelem and Majeed
1992;Lingelem et al. 1994; Ouar et al. 1992).
In terms of HrO clathrates,threemain crystal structures
have been identified. Theseare known as structuresI (sI),
II (sII), and H (sH). StructuresI and II are cubic structures and were first identified in the early 1950s(Claussen
1951a,l95lb; Miiller and von Stackelberg1952;Pauling
and Marsh 1952; von Stackelberg and Mtiller l95l).
Structure H is hexagonaland was first describedby Ripmeesteret al. (1987). This hydrate structureis thought to
be isostructural with dodecasil lH clathrasil (Gerke and
Gies 1984), a framework-guest structure with SiOo tetrahedral groups forming hydrate-like cages. A singlecrystal structure refinement has yet to be performed for
sH hydrate. Crystallographic and crystal-chemical data for
thesethree hydrate structuresare summarizedin Table 1.
Knowledge of the crystal chemistry, crystallography,
and stability of these three hydrate structures has been
obtained from a variety of experimental and theoretical
techniques,including X-ray and neutron diffraction, various spectroscopies(NMR, IR, FIIR), mathematicaland
thermodynamicmodeling, and phase-equilibriumstudies.
In most of the experimental work on hydrates, the presence of hydrate crystals, or the onset of hydrate formation, has been demonstratedeither by the uptake of the
hydrate former, typically a gas (usually measured as a
decreasein pressure);by the observation of turbidity in
the solution; or by the appearanceof diffraction maxima
in the caseof X-ray or neutron-scattering
experiments.
Optical microscopy has not been fully exploited in hydrate research.Optical studiesprovide an opportunity to
examine several macroscopic properties of gas hydrate
crystals. These include crystal-growth behavior, equilibrium crystal morphologiesof different hydrate structures,
and determination of relative growth rates of specific
crystallographicplanes or directions.In this paper we describe a pressurizedoptical cell designedfor the study of
natural gas hydrates and presentoptical data for each of
the three naturally occurring hydrate structures.
Our main goal is to provide a visual framework to aid
in the identiflcation of unknown clathrate hydrate samples. This holds potential for in situ identification of specific hydrate phasesin their natural environments,complementing the usual characteization method of evolved
gas analysis from samplesreturned to the laboratory. As
shown below, the crystal forms characteristicof each hydrate structure are distinctive. In addition, phase-equilibrium data, obtained from freezing-melting experiments,
are presentedfor sH hydrate in the methane-methylcyclopentane-water system. This hydrate structure, only recently found in a natural environment (Sassenand MacDonald 1994), can scavengelarge hydrocarbonmolecules
not found in sI and sII. The phase-equilibriumdata for
severalsH hydratessuggeststhat their natural occurrence
may be widespread.
ExpnnnunurAl
METHoD
Description of experimentalset-up
For the optical study of gas hydrates,a specialcell was
designed that enabled pressureand temperaturecontrol.
A schematic diagram of the cell is shown in Figure l
The cell consistsof alz in. thick stainlesssteelbody with
a 1/zin. bore in the center, which is the actual sample
chamber.
Cooling was achievedusing a Neslab RTE 140D continuous circulation chiller, which pumped a 50-50 mixture
of ethylene glycol-water through insulated lines to the
optical cell, where it flowed through 7+in. ports that surround the sample chamber. The temperature range encountered in our experiments was from 256 to 290 K.
Becauseof the relatively large volume of coolant in the
chiller bath, about 7.0 L, temperaturechangesin the cell
were gradual, rather than instantaneous;however, excellent temperaturestability was achievable. The temperature was measuredusing a YSI model 44011 precision
90
SMELIK AND KING: CRYSTAL GROWTH OF CLATHRATE HYDRATES
Coolant
In
Coolant
Out
b.
Top retainer
ring
Coolant
Upper
window
Lower
by 1/ruin. flexible, high-pressure,stainless steel tubing
connectedto a pressure-controlgas manifold. The highest
attainable pressureswere those of the bottled gases.In
these experiments we used two commercially available
gases:a 6000 psi (-40 MPa) methane (CP grade) and a
2000 psi (-14 MPa) mixture of 95.4lVomethane-4.59Vo
propane (MenrPrr).For sH we used liquid methylcyclopentane(Fluka AG >99Vo pure). Pressurewas measured
using a Heise model 623 pressure gauge accurate to
+lVo. All the crystal-growth experimentswere conducted
at pressuresbetween about 1 and l0 MPa.
The entire optical cell was encasedinside a 4.5 in.
diameter, horizontally split shell composed of Delrin.
This shell, about 3 in. thick, served to insulate the steel
body of the optical cell for excellenttemperaturestability.
Special ports were built into the insulating shell to allow
N, gas flow over the upper and lower windows to prevent
condensationwhen working at temperaturesbelow about
278 K.
The optical microscope was an Olympus model BH-2
equippedwith a polarizer and analyze4and an r-y translation stage.A special spacerwas installed in the microscope stand to allow clearancefor the optical cell assembly. Normal viewing was accomplishedusing transmitted,
plane-polarized,and crossed-polarizedlight. Becauseof
the thicknessof the optical cell, we usedMitutoyo M Plan
APO long-working-distanceobjective lenses(5x, 20x,
50x). All images were recorded on video tape with a
Sony model DXC-151 color CCD cameraand a VCR.
Selected images were captured using a RasterOps
24MxTY 24-bit video card on a Macintosh computer.
Experimental procedure
ittr"':
Frcunn1. Schematic
diagramof thepressurized
opticalcell.
(a) Top view showingstainlesssteelbody of cell with central,
cylindricalsamplechamber.(b) Sideview showingupperand
lower SiO, glasswindowswith o-ring sealsand steelretainer
rings The sampleholderis a small,cylindricalcup into which
wateris placed.It is toppedby a water-impermeable,
gas-permeablelayerof hydrocarbon
liquid.In bothpartsbolt holeswere
left out for clarity.
thermistor,which was insertedinto a well in the side wall
of the stainlesssteelbody, placing the thermistor tip within a few millimeters of the sample chamber wall. Calibration experimentswith sI methanehydrate showedthat
the temperatureof the cell body matchedthe temperature
within the sample chamber to within +0.4 K.
For optical viewing, the central, cylindrical, sample
chamberhas 3/,uin. thick upper and lower SiO, glass windows. These are sealedwith o-ring sealsand held in place
by stainless steel retaining rings. The upper window is
slightly recessedinto the cell body to allow for focusing
in the bottom of the cell.
Two small ports for gas flow also lead into the sample
chamber.These were connectedto standardbottled sases
Experiments were conducted by simply filling about
one-half of the samplechamberwith deionizedwater.After sealing the cell, gas was flushed through the chamber
to drive out any trapped air. Using the gas-manifoldcontrols, the pressurewas then increasedto the desiredlevel.
Once pressure was attained, the temperature was decreaseduntil hydrate formed in the cell. One characteristic element of all our experiments is that becausethe
systemis essentiallyclean, with only deionizedwater and
gases,and is static, a large degree of undercooling was
necessarybefore any freezing would take place. Depending on the pressure,the thermal stability of gas hydrates
is typically in the 273-293 K range. Undercooling to approximately 264-258 K was necessaryto cause freezing
in the cell. In all cases, this initial freezing event was
rapid and usually consistedof dendritic hydratesforming
at the water-gas interface. In many experiments, ice
would crystallize at the same time as the hydrate. Once
crystallized material formed, the temperature was increasedto >2'73 K, depending on the pressure,so that
any ice in the cell would melt, leaving only the hydrate
phase.
The actual crystal-growth experimentswere performed
by slowly melting the dendritic hydrate material until
only small, scatteredhydrate crystals remainedin the cell.
SMELIK AND KING: CRYSTAL GROWTH OF CLATHRATE HYDRATES
9r
We have termed these "melt nuclei." These melt nuclei
then served as the primary sites for new crystal growth,
once temperaturewas decreased,or pressureincreased.
The best results were obtained by placing the liquid
solutions into a plastic, cylindrical, glass-bottomedcup
(isolating it from the cold side walls) and applying a thin
layer of light hydrocarbonoil (two to three drops of decane, C,oHrr)on top of the water. This oil is not a hydrate
former and serves to prevent the water from migrating
out of the cup toward the coldest regions in the cell during crystallization, which would normally lead to hydrate
formation in undesirable regions of the cell (i.e., along
the cold side walls). In the case of sH, the decane was
not usedbecausethe sH component,methylcyclopropane,
served the same purpose.
ExpnnrnnnNrAr- RESULTS
Surface crystallization
Hydratesform from a mixture of water and natural gas,
plus or minus other complex liquid or gas hydrocarbons.
Inherent in any system of this sort is the existenceof a
water-gasinterface. In a turbulent system,like a pressurized, flowing, natural gas pipeline, these interfaces are
probably manifested as bubble surfaces or as interfaces
between larger gas pockets and water. In either case,the
interface areaand geometry probably changecontinuously with time.
In our optical cell, there is a static, stratifled system of
water, a thin layer of oil, and gas. Once the system is
pressurized,there is minimal movement of the gas and
the liquid. This static arrangementproduces a relatively
large, unchangingwater-gasinterface (-50 mm'?),where
water has its easiest accessto hydrocarbon gas. As expected, hydrate-surface crystallization was a common
process during supercooling. These surface textures are
describedfirst.
As mentioned above, the initial stage of any experiment involved supercoolingthe system until rapid freezing took place. The initial hydratesthat formed were typically branching dendrites or very-fine-grained crystals,
and these were not extensively characterized.Once this
systemmelted back, leaving only small melt nuclei, temperature was slowly decreaseduntil crystal growth was
observed,usually occurring within I or 2 "C of the equilibrium curve. For the pure methane system (sI) and the
Men.Pr, system (sII), surface crystallization was ubiquitous, whereas methane-methylcyclopentanesH did not
exhibit surface crystallization.
Surface crystallization proceededin one of two ways.
Commonly, large surfacecrystals would grow from a single melt nucleus or a nuclei cluster at the water surface.
These crystals usually showed a hexagonalcrystal shape
and grew outward from the central nucleus (Figs. 2a and
2b). The rough-looking surfacesand movementsof these
crystals clearly indicated that they were floating at the
water-gas interface. Eventually these surface crystals
would increasein size until they impinged on one another,
for sI and
Frcunr 2. Opticalimagesof surfacecrystallization
sII hydrates.(a and b) Largeindividualsurfacecrystalsof sII
from smallmelt nucleithat were
hydrate,whichhavenucleated
phase(seea). Theirroughsurfacetexture
floatingin theaqueous
interface.(c) Surfacefrontsof
is due to growthat the water-gas
growingacrossthe water-gas
indrapery-likesII crystallization
view of sI surfacecrystallization,
terface.(d) High-magnification
(e andf)
showinghow the surfacedomainsaresuturedtogether.
Two views of the complexsurfacemosaicfor sII hydratein
hydratecrystalshave bewhich individualhexagonal-shaped
cometrappedin the surfacemosaic.Note the unusualgrowth
patternsfor thesecrystalsbecause
of theproximityof thewatergasinterface.Scalebarsequal50 pm.
forming a surfacemosaic. A secondtype of surfacecrystallization involved several large fronts freezing across
the water-gas interface simultaneously. These surface
fronts commonly had a drapery appearanceand almost
always originated near the side walls of the cell and propagatedrapidly acrossthe cell (Fig. 2c). Thesefronts often
met one another,becoming sutured together (Fig. 2d), or
they encounteredand trappedfloating surfacecrystals until the surface was completely frozen over in a complex
mosaic (Figs. 2e and 2f). The frozen hydrate surfacewas
not flat and appearedto have considerabletopography,
and the thicknessof this crust was always greatertoward
the outer rim of the sample cup. Similar surfacetextures
were reportedby Makogan (1994) for sI methanehydrate.
In large-scalehydrate systems,whether natural or industrial in origin, it is likely that surface crystallization
is a common, and perhaps dominant processbecauseof
the ever-presentwater-gasinterface.However,beyond the
recognizablecrystal shapeswithin the surfacemosaic,the
surface crystals themselvesdo not provide much infor-
92
SMELIK AND KING: CRYSTAL GROWTH OF CLATHRATE HYDRATES
Ftcunn 3. Opticalimagesof methanesI singlecrystals.(a)
Thethreecentralcrystalsare { I l0} rhombicdodecahedrons,
the
mostcommonsI form. The crystalin the upperright appears
to
be a {100} cubevieweddown[1] 1], whereas
in thelowerleft
two cube-shaped
crystalshaveformeda growth twin. (b) An
exampleof a possible{hhl} trapezohedral
form for sI. Notethe
faint crossquaneringthe front face of the crystaland that the
sidefacesappearto bulgevery slightly,consistent
with thisform.
Scalebarsequal25 pm.
crystals. By decreasingthe temperatureto about 6-8 'C
below the equilibrium temperature, single crystals of
methanehydrate would form in the bulk liquid below the
hydrate skin, which was, fortunately, transparentenough
to permit crystal investigation.These single crystals were
small, never exceedingabout 25 pm in diameter,and normally floated up to the underside of the hydrate surface
layer.
Severalcubic forms were observed,the most common
being the rhombic dodecahedron.Figure 3a shows three
such crystals in the central part of the image. This form
consistsof 20 rhombohedral(diamond-shaped)faceswith
{110} indices (identical to common garnet).Two other
crystals are shown in Figure 3a, one is a cube, viewed
down the [11] diagonal,and the other appearsto be a
growth twin of two cube-shapedcrystals.
Another possible crystal form that was more rarely observed is the trapezohedron.This form consistsof 24 trapezium-shapedfaces with {hhl} indices. These forms
were difficult to identify becauseof their close similarity
to cubes (Fig. 3b), especially when the h and I indices
are high, as they appearedto be in this case. The image
in Figure 3b shows a view down the fourfold rotation
axis of a cube-shapedcrystal. Notice the faint cross quartering the (100) face and that the other { 100} faces seem
to bulge somewhat, suggesting deviation from regular
cube shape.
All the sI crystals are isotropic, and the observedforms
are consistent with the point-group symmetry of the sI
spacegroup.
Structure II single crystals
Cubic sII hydrate (Table l) was grown from pure water
and an MenrPr,gas mixture. Single crystals of sII within
the bulk liquid could be grown in the absenceof the hydrate surface skin and below the hydrate skin. Structure
II occurs in two dominant morphologies,regular octahedra and thin platelets (Fig. a). In both cases the single
mation about equilibrium growth rates of different crys- crystals were clearly tumbling through the host liquid and
tallographic directions or about equilibrium crystal mor- not in contact with the water-gasinterface. The platelet
phologies.Despitethe ubiquity of surfacecrystallization crystals usually displayed hexagonal, trigonal, or trianfor sI and sII, it was possibleto grow single crystalsof gular shape(Figs. 4a and 4b). In comparison with the sI
hydrate within the bulk liquid. Becausethe solubility of crystals above, these platelets grew quite large, reaehing
gas in the water is low, thesecrystals tendedto be limited up to 100 pm across,but were generally only a few miin size and number. Moreover. once the water surface crons thick. Occasionally, two or more platelets would
froze over, no additional gas could be made available for become intergrown with each other, such that they were
hydrate growth within the fluid unlessit was able to dif- interpenetrating.Structure II hexagonal-shapedplatelets
fuse along the suture boundariesin the mosaic.
have been previouslydescribedfor 8CCl*.3.5Xe.136D,O
hydrate (McMullan and Kvick 1990). The large faces of
Structure I single crystals
the plateletsare {111} planes.
Cubic sI hydrate (Table 1) was grown from pure water
The other dominant form is the regular octahedron
and methane gas. Of the three structuresstudied, single (Figs. 4c and 4d). These crystals are about 25 pm in
crystals of sI methanehydrate were the most difficult to diameter, like the sI crystals, and are bound by regular
produce.We were able to grow single crystals in the bulk { 111J faces.Thesetypically were found growing in the
liquid only by continuing to cool the cell slowly after the bulk water after the formation of a hydrate surfacelayer.
surfacemosaic had formed. Despite the presenceof many The fact that the octahedraremain small despite further
melt nuclei, surface crystallization would always take decreasesin temperature,atteststo the low solubility of
place first, before any nuclei could grow into well-defined gas in the water. The platelets and octahedrawere nor-
SMELIK AND KING: CRYSTAL GROWTH OF CLATHRATE HYDRATES
93
Frcunn4. Opticalimagesof MerrPr,sII singlecrystals.(a)
Clusterof thin individualplateletcrystals.Thesetypicallyhad
hexagonal,
trigonal,or triangularshape,boundby large{1ll)
faces,andwereclearlydrifting in the bulk liquid. (b) A pair of
plateletsnearthewater-gas
interface,aboutto be enhexagonal
gulfed by the advancingsurfacecrystallizationfront from the
left. (c andd) Two viewsof clustersof perfect{ I I I } octahedra
of sII hydrateexistingin thebulk liquidbelowthefrozenhydrate
crust.Theseoctahedra
wereremarkably
uniformin size,reaching
a maximumdiameterof about25 pm. In d, triangularplatelets
Frcuns 5. Optical imagesof methane-methylcyclopentane
Scalebarsequal50 um.
canalsobe seenwith the octahedra.
sH singlecrystals(a) PlaneJightview of a clusterof hexagonal
prisms,the mostcommoncrystalform for sH. (b) Crossed-polarizedJightview of similarareashowinganisotropiccharacter
crystals.All thecrystalshaveparallelextincmally found growing together,suggestingone form is not of thesehexagonal
(c and d) Exampleof a growthtwin involvingtwo welltion.
preferred over the other.
formedhexagonalprisms Part d is the crossed-polarized-light
view with thecrystalin the (001)orientationfully extinct.Note
Structure H single crystals
that the twin boundaryis not planar,but curved.(e) Mixtureof
Structure H is the most recently discovered hydrate hexagonalprisms,one quitelong, and a differenttypeof complateletcrystals.The
structure (Ripmeesteret al. 1987). From a combination mongrowthtwin, involvingtwo rectangular
twin in f reveals
view of this coarsened
of powder X-ray diffraction data and NMR spectroscopic crossed-polarized-light
data (Ripmeester and Ratcliffe 1990; Ripmeester et al. thatthe twin planemay not be vertical.Scalebarsequal50 pm
1987; Tse 1990), sH appearsto be hexagonalP6lmmm
fiable l) and is isostructuralwith dodecasil lH clathrasil
(Gerke and Gies 1984). Our optical observationsare con- birefringence, and all the prismatic crystals showed parsistent with this model.
allel extinction, consistent with hexagonal symmetry.
On the basis of the work of Ripmeester and Ratcliffe There were many examplesof distorted hexagonalprisms
(1990), we chosemethylcyclopentane,
CuH,r,as the hy- and combinationswith other forms as well.
Growth twins were also common in this system (Figs.
drate former, in addition to deionized water and pure
methane gas. In these experiments the methylcyclopen- 5c-5f). These typically involved intergrown hexagonal
tane was addedto the top of the water in the samplecup, prisms (Figs. 5c and 5d) or intergrown rectangularplateafter which the systemwas pressurizedwith methanegas. lets (Figs. 5e and 5f). The latter style of twinning was
The initial freezing event always resultedin the formation more common among crystals floating at the water surof both sI and sH. Once the sI methanehydrate melted, face. The intergrowth plane for the prismatic twins tended
single crystals of sH could be grown from the stable sH to be irregular in shapeand orientation (Fig. 5d), whereas
the twin plane for the platelet twins always nearly bimelt nuclei scatteredthroughout the cell.
The predominant crystal form for sH is the hexagonal sected the obtuse angle created by the two individuals
prism (Figs. 5a and 5b). These beautiful six-sidedpris- (Fig. 5f).
Becausethe rectangular platelet twins were common,
matic crystals grew in all sizes and orientations within
the cell. The majority of the crystals would lie flat, how- a possible crystallographicdescription of this intergrowth
ever, along the water-hydrocarboninterface. Being hex- can be presented.Becausethe twinned crystalsshow biagonal,the crystals could be studied in crossed-polaized refringence,the large flat face parallel to the plane of the
light (Fig. 5b). The thicker crystals showed strong image cannot be (001). Each crystal in a twinned pair
94
SMELIK AND KING: CRYSTAL GROWTH OF CLATHRATE HYDRATES
showstwo facesat 90" ro eachother (Fig. 5e). It is likely Teele 2. Phase-equilibrium
P-f pointsfor sl and sH hydrates
determinedusingthe opticalcell
that these two faces have indices of the sort (00/) and
(frk0), respectively.On the basis ofparallel extinction criMethanesl
Methane-methylcyclopentane
sH
teria, we assumed that these two faces are (001) and
(K)
(K)
r
P (kPa)
(kPa)
r
P
(100),respectively.Given this assumption,it follows that
273.0
2482
1t c.J
2400
the large flat face, normal to theseother two, is (120) and
276 2
3493
276 4
2730
that the view is down the b axis or [010] zone. If the
278 8
4426
277.4
zalJ
twin plane is vertical, in the [010] zone, then it must be
279.9
5 13 0
278 1
3177
281 I
6205
278 I
3228
of the sort {/20/}.The obtuseanglebetweenthe two twin
284.5
8356
280.5
3900
individuals is about 116'. If the unit cell of Ripmeester
282.1
4990
282.2
and Ratcliffe (1990)is used,the anglebetween(001) and
4800
282.9
5 0 14
(503) is 5192 (-116 + 2), making it the bestcandidate.
283 1
5502
If the twin plane is not in the [010] zone, as the unusual
birefringence in Figure 5f might suggesr,then it probably
has a relatively high k index.
assemblageto sit for up to 45 min to seewhether melting
Phase-equilibriumdata for structure H
or growth occurred. As the equilibrium phase boundary
The phasebehavior of hydrate systemsinvolving light was closely approached,the assemblageremained unhydrocarbons and water is complex. This is largely be- changedfor longer periods of time.
The same procedure was followed for several tempercauseof the low mutual solubility between the water and
the hydrocarbon phase(s) and becauseof the formation atures in the pure methanesystem (sI). Comparisonwith
of the solid hydrate compound itself [see review by Har- previous data for methane sI hydrate (Deaton and Frost
mens and Sloan (1990) describing the important aspects 1946) and methane-methylcyclopentanesH hydrate
of these kinds of systemsl. A thorough understandingof (Thomas and Behar 1994) indicatesthat this techniqueis
the phase behavior of hydrate systems is of critical im- an accurate means of determining phase equilibria. The
portance,howeveq especially in light of the potential en- four-phaseequilibrium (L* + H + V + L"") P-Zpoints
sH hydrate and for pure
vironmental impact representedby the behavior of natural for methane-methylcyclopentane
hydrates and the costly hydrate-related problems that methane sI hydrate (L* + H + V) are summarized in
Table 2.
plague the natural gas and petroleum industry.
Phase-equilibrium data for sI and sII hydrates are
DrscussroN
readily available in the literature (e.g., Deaton and Frost
1946; Parrish and Prausnitz l9l2 and referencestherein). Relative growth rates of crystal forms
Excellent reviews of the natural gas phase equilibria of
The resultspresentedaboveindicate that single crystals
these hydrate types were provided by Sloan (1990b) and of all three known hydrate structurescan be easily grown
Holder et al. (1988).
in the laboratory for optical study. Other resultsof optical
Over the last few years, phase-equilibriumdata for sH studieshave been published,mostly by Russianresearchhydrates have begun to appear in the literature. To date, ers (Makogan 1994), but these results focused more on
experimentalphase-equilibriumdata for sH has been re- surface-crystallizationtextures at the water-gasinterface
ported by Lederhos et al. (1992) for the system water- and whisker crystals forming within the gas medium, and
methane-adamantane,
by Mehta and Sloan (1993) for the more recently some crystal-morphologyresults were presystems methane + 2,2-dimethylbutane,methane *
sentedby Larsen et al. (1996). For crystals coexisting
2-methylbutane,and methane + methylcyclohexane,and with liquid, each hydrate type exhibits different morby Thomas and Behar (1994) for several systems con- phology. This observation suggeststhat optical examitaining methaneand larger hydrocarbons,including meth- nation may be potentially very useful for the rapid idenylcyclopentane,the large hydrocarbonused in this study. tification of natural hydratescollected in the field, instead
To evaluatethe usefulnessof our optical cell for phase- of gas extraction and analysis or other more involved
equilibrium measurements,we determinedequilibrium P- techniques(e.g., Brooks et al. 1984; Davidson et al.
I points for methane-methylcyclopentane
sH. The deter- 1986a).In light of recent excursionsby deep-seasubmination of thesepoints was accomplishedby holding the marines to natural hydrate zones (Sassenand MacDonald
cell at various constant temperaturesand changing the 1994), perhaps the vessels could be outfitted such that
pressurewhile observing either the growth or melting of optical identification could be performed in situ, elimithe hydrate crystals. Because pressure changes are nating the problem of preserving and transportingthe hyachievedby varying line pressure,they occur nearly in- drate samples.
stantaneouslythroughout the small volume of the cell.
The observed crystal morphologies allow the evaluaOne of the difficulties in determining optically the exact tion of relative growth rates for different crystallographic
equilibrium points is that the reaction rates are very slow planes or directions for each hydrate type. In conducting
very near the phaseboundary. Each point was bracketed this analysis,it is important to recall that under ideal conby cycling up and down in pressure and allowing the ditions the bounding faces of any crystal generally cor-
SMELIK AND KING: CRYSTAL GROWTH OF CLATHRATE HYDRATES
TABLE3. Unit-celldistributionof small cages in sl, sll, and sH
Millerolane
Structurel.
Structurell"
{f'00}
{hhol
{hhhl
{004
4
7
8
n.a.
StructureH"'
o
1
5
Nofe.n.a. : not applicable.
'Only 51'zcages
.- lncludesboth small cages:51,and 435663
respond to the slowest growing directions or planes,
whereasthe fastestgrowing directions typically never develop correspondingmorphologic faces and insteadoccur
as apices or edgesof the crystal.
For methane sI hydrate, the common rhombic dodecahedralform, boundedby { I 10} faces(Fig. 3a), suggests
that the {110} planesare the slowestgrowing, whereas
growth along the crystallographic axes ({ft00} planes)
and along [11] (the cubic diagonal)is more rapid. Because {100} cubes and possible {hhll tapezohedrons
were also sometimesobserved,but { 1l I } octahedrawere
not seen, the relative growth rates for the primary crystallographic planes of sI hydrates can be given as { 1I I }
> { 1 0 0 }> { 1 1 0 } .
This is in contrastwith sII hydrates,for which the regular { I l1} octahedronwas a commonform, or hexagonal
platelets with large { I I 1} faces (Fig. a). These observations suggestthat the { 111} planesare the slowestgrowing. Work by Smelik and King (unpublished)on sII THFwater hydrates has provided many examplesof distorted
{111} octahedralforms containing{110} edges.Cubeshaped crystals or rhombic dodecahedrawere not observed. These combined observationsgive the following
relative growth rates for crystallographicplanesin sII hyd r a t e s{:1 0 0 } > [ 1 1 0 ] > { 1 1 1 } .
It should be possible to reconcile these different relative growth-rate schemesand morphologies in terms of
the crystal structuresof sI and sII. Structural data for sI
and sII are given in Table 1, indicating that both hydrates
are cubic. Hydrate structures are usually described in
terms of packing or layering of face-sharingpentagonal
dodecahedra([51'z]cages)formed by the hydrogen-bonded water molecules(e.g.,Jeffrey 1984;Ripmeesteret al.
1994; Smelik and King 1996). The particular way in
which these basic building blocks are packed together
leads to the wide array of additional cavities that are
found in hydrate structures.
One may envision the growth of a hydrate crystal proceeding by the addition, through hydrogen bonding, of
new water molecules to the existing three-dimensional
framework. As the framework is built up in this way,
guest molecules are trappedin the cagesbefore the cages
are fully closed. Becausethe larger cagesform as a result
of the packing together of the smaller,dodecahedralcages, it seemspossiblethat the building of the [5''] cages
is critical to the formation of the crystals. Presumably,a
small crystal of hydrate immersed in water is completely
95
surrbundedby molecules of HrO that can be scavenged
by the growing crystal. In a system like ours, which is
saturatedwith the hydrate former (methaneor MenrPrr),
it is also likely that the guest moleculesare equally available to any region of the crystal. Furthermore, it is reasonableto assumethat the energy required to build a new
[5"] cage (hydrogen bonding) is constant, no matter
where at the hydrate crystal surface the growth is occurring. If the constructionof the [5''?]cagesis a rate-limiting
factor, then there should be a correlation between the
number of these cages in a given crystallographic plane
and the growth rate of that plane, with planes containing
a high density of [51'z]cages growing more slowly than
those with fewer cages.
By representing the [5''z] cages by their centers, we
analyzedthe unit-cell density of [5''] cagesfor {ft00},
{hh}}, and lhhhl planesin both sI and sII. The results
are summarized in Table 3. For sI, the planes with the
highest density of [5''] cages are, in fact, the {hh}}
planes, which appear to be the slowest growing. The
{hhh} planesin sI have the lowest [5''] densityand thus
grow most rapidly, consistent with the optical observations. For sII, the {hhh) planeshave the highest density
of [5"] cages,and the {ft00} planeshavethe lowest,with
{hh}l intermediatebut close to {ft00}. This result predicts the development of octahedral forms or distorted
octahedralforms for the sII hydrates,in exact agreement
with our observations.This simple correlation suggests
that the relative growth ratesof various planesin common
hydrates, under ideal or equilibrium conditions, can be
predicted by determining the density distribution of [5''?]
cages. This relationship is consistent with the periodic
bond-chain (PBC) theory of crystal-habit formation presentedby Hartman (1987). In the context of the PBC
theory, the [5"] cage is the fundamental unit of the periodic bond chain.
The situation for sH is slightly more complicated becausethere are two distinct small cages,the [5''z] and the
f435663f.The structure consists of a basal layer of sixmembered rings of [5"] dodecahedraat z : 0, with a
stacked
layer of six-memberedrings of !4356631polyhedra
on top of this at 7 = t7r.Another [5"] layer at z = I closes
the large, asymmetric [5''6'] polyhedron, completing the
unit cell (see Lederhoset al. 1992). Thus, growth of a
crystal would depend on the formation of both types of
small cages. A similar density-distribution analysis was
performed for sH considering both types of small cages
(Table 3). The results suggestthat the crystals should be
bound by {00/} and {h00} planes.The optical results
indicated that the most common form was indeedthe hexagonal prism. This form can be either a first-order { 100}
prism or a second-order{ 110} prism, which are identical
in appearance.The results in Table 3 suggest that the
prisms are first order becausethe density of small cages
is less on {12ft0}planes,such that they would grow faster.
The aspect ratio of the prismatic sH crystals was quite
variable, suggestingthat the growth rates for the {h00}
and {00/} families of planes are similar.
96
SMELIK AND KING: CRYSTAL GROWTT{ OF CLATHRATE HYDRATES
natural gas is a mixture of methane,larger hydrocarbon
gases,and other species(e.g., Kvenvolden 1993). As is
100
evident from Figure 6, such gas mixtures typically form
sII rather than sI. For example, at 277 K sII forms near
1.0 MPa, an oceandepth of only about 100 m. The widespread occunence of natural sII hydrate has been confirmed by analyses of field samples from the Gulf of
Mexico (Brooks et al. 1984, 1986; Davidson et al.
1986a).A recent report provides evidencefor natural sH
hydrate from the Gulf of Mexico continental slope (Sassen and MacDonald 1994).In the laboratory, Ripmeester
A.
and Ratcliffe (1990) identified about 25 large hydrocarbon guests capable of forming sH. These results, plus
recent reports showing that sH forms from liquid hydrocarbon componentsin petroleum (including components
of gasoline) (Thomas and Behar 1994; Sloan 1994),
strongly suggestthat hydrates with this structuremay be
more common than previously thought.
Our phase-equilibriumresults for sH methane-methylcyclopentanehydrate (Table 2) are plotted in Figure 6.
Also shown in Figure 6 are data points from Deaton and
27t
273 275
28t
28J 285
"',r*rrrn
Frost (1946) for sI methane hydrate and for sII using
Frcuns 6. Plot of P-7 equilibriumdatameasured
for meth- MerrrPro, gas, a mixture nearly identical to that used in
anesI (solid squares)and methane-methylcyclopentane
sH hy- this study. In addition, data for several sH systems are
drate(solidcircles)usingthepressurized
opticalcell.Alsoshown also shown for comparison.The phase boundary for sH
are the datafrom the literaturefor methanesI (opentriangles; methane-methylcyclopentanehydrate plots between the
DeatonandFrost1946),Men,,Pr,,
sII (opencircles;Deatonand curves for sI and sII and is identical to that determined
Frost 1946),andseveralsH hydrates(crossedsquares,
opendi- by Thomas and Behar (1994). The slope for methaneamonds,opensquares,
and invertedopentriangles,all dataof
Mehtaand Sloan1993).Curvesfor sI and sII are three-phase methylcyclopentane sH is nearly identical to that for
equilibriumlines,whereasall the sH linesare four-phase
equi- methane sI hydrate.
The combined results for sH hydrate shown in Figure
librium lines.Also shownas a dashedline is the predictedlocationfor sH hydratebasedon a correlationof sI and sII P-? 6 show that the exact P-T locations of sH equilibrium
datawith meancavitysize,givenby Lederhoset al. (1993).The curves are strongly dependent on the nature of the sH
presentresultsfor methane-methylcyclopentane
sH hydratefall guest molecule. Furthermore,the data show that given a
within theenvelopeof publishedsH curvesandarein nearexact mixture of hydrocarbongasesand liquids, an assemblage
agreement
with the measured
P-Z equilibriumdataof Thomas of sII and sH hydrate would clearly form first, before the
andBehar(1994)for this hydrate.
necessarydepths or temperatureswere reached to stabilize sI hydrate. These data therefore suggestthat sH +
sII assemblagesshould be common natural occurrences.
Some prismatic crystals had much higher aspectratios,
however, like the one shown in Figure 5e. The higher
AcrNowr,nocMENTS
ratios may result becausethe large methylcyclopentane
We are indebtedto ThomasSun and Dudley Herschbachfor many stimmolecules, which are housed in the large, asymmetric ulating discussions,and to Mark Disko for his expenise in image procages,have the easiestaccessto thesecagesalong cessing.We thank Lany Wenzelfrom American Design, Stirling, N J , for
[5''?68]
the c axis as the crystal is growing. This is becausethese his help in designingand manufacturingthe optical cell. Thoroughreviews
oblong cagesare stackedend to end along [001], sharing were provided by Laura Stern,Kathryn Nagy, and an anonymousreviewer,
their upper and lower hexagonal rings. This geometry all of which improved the manuscript.
should enhance growth along c, in comparison with all
RnrnnrNcEs crrED
.
other directions.
o
2
a'T("C)e
8
lo
12
Stability of structure H
The importanceof sH in the geologic environment,and
in the natural gas and petroleum industry, has yet to be
fully assessed.Natural methane derived biogenically is
nearly pure and therefore forms sI. The P-T stability region for sI is limited ro P > 3.5 MPa at211 K (Fig.6),
or ocean depth of about 350 m (using the hydrothermal
gradient of Kvenvolden 1988b). Howeveq thermogenic
Barker,J.W, and Gomez, R K. (1989) Formation of hydratesduring deepwater drilling operations.Journal of Petroleum Technology, 41,29'7301.
Brooks,J.M., Kennicutt,M.C., II, Fay, R R , and Sassen,R (1984)Thermogenic gas hydratesin the Gulf of Mexico Science,225, 409-411.
Brooks,J.M., Cox, H.B., Bryant,W.R., Kennicutt,M.C., il, Mann, R C.,
and McDonald, T.J. (1986) Associationof gas hydratesand oil seepage
in the Gulf of Mexico. Organic Geochemistry,10,221-234.
Carpenter,G (1981) Coincident sediment slump/clathratecomplexeson
the U S. Atlantic continental slope. Geo-Mrine Letters, 1,29-32.
Cherskiy, NV, Tsarev, V.P, and Nikitin, S.P (1985) Investigation and
SMELIK AND KING: CRYSTAL GROWTH OF CLATHRATE HYDRATES
predrctionof conditionsof accumulationof gas resourcesin gas-hydrate
pools. PetroleumGeology, 21, 65-89.
Claussen,W.F (195la) Suggestedstructuresofwater in inert gas hydrates
Journal of Chemical Physics, 19, 259-260
-(1951b)
A second water structure for inert gas hydrates.Journal
of Chemical Physics, 19, 1425-1426
Collett, T.S (1994) Permafroslassociatedgas hydrate accumulations In
ED Sloan Jr, J. Happel,and MA Hnatow, Eds,Intemational conferenceon natural gas hydrates,115,p 241-269 New York Academy
of Sciences,New York
Craig, H, Shoji, H, and Langway,C.C, Jr. (1993) Nonequilibriumair
clathratehydratesin Antarctic ice: A paleopiezometerfor polar ice caps
Proceedings
of the NationalAcademyof ScienceU.S A, 90, 11416I l4l8
Davidson,DW. (1973) Clathratehydrates In F Franks,Ed, Water: A
comprehensivetreatise,2, p II5-225 Plenum Press,New York
Davidson,D W, Garg, S K, Gough, S.R., Handa, YP, Ratcliffe, C I ,
Ripmeester,J A , and Tse, J.S (1986a) Laboratory analysisof a naturally occurring gas hydrate from sedimentof the Gulf of Mexico Geochimrca et CosmochimicaActa, 50, 619-623
Davidson, D W., Handa, YP, Ratcliffe, C I , Ripmeester,J A., Tse, J S.,
Dahn, J R, Lee, F, and Calvert,L D (1986b)Crystallographic
studies
of clathrate hydrates:Part I Molecular Crystals Liquid Crystals, 141,
t4t-t49.
Deaton,W.M, and Frost,E.M, Jr. (1946)Gas hydratesand theirrelation
to the operationof natural-gaspipe lines. U S. Bureau of Mines Monograph, 8, p. 101, Departmentof the Intenor.
Englezos,P (1993) Clathratehydrates Industrial and EngineeringChemistry Research,32, 1251-12'74
structure
Gerke, H , and Gies, H (1984) Studieson clathrasils:Mrystal
of dodecasil lH, a synthetic clathrate compound of silica Zeitschrift
ftir Kristallographle, 166, ll-22.
Ginsburg,G.D., Guseynov,R A., Dadashev,A.A, Ivanov,G.A., Kazantsev, S.A, Soloviev, V.A., Telepnev,E.V., Askeri-Nasirov, R.Y, Yesikov, A D , Mal'tseva,VI , Mashirov,YG, and Shabayeva,
l.Y. (1992)
Gas hydrates of the southernCaspian International Geology Review,
34.76s-182
Harmens,A., and Sloan, E.D, Jr (1990) The phasebehavior of the propane-watersystem:A review CanadianJournal of ChemrcalEngineering,68, 151-158
Ed, Morphology
Hartman,P (1987)ModernPBC theory In I Sunagawa,
ofcrystals, p 271-319 Reidel,Boston
Holder,GD., Kamath,V.A, and Godbole,SP (1984) The potentialof
natural gas hydrate as an energy resource Annual Review of Energy,
9.427-445
Holder,G.D, Zetts,S.P, and Paradhan,N. (1988)Phasebehaviorin systems containing clathratehydrates Reviews in Chemical Engineering,
5, 1-70
Jeffrey, G A (1984) Hydrate inclusion compounds In J L Atwood, J E D.
Davies,andD D MacNicol,Eds , Inclusioncompounds,l, p 135-190
Academic Press.New York.
Kayen, RE., and Lee, H.J. (1991) Pleistoceneslope instabilityof gas
hydrateladen sedimenton the Beaufort Sea margin Marine Geotechnology, 10, 125-141
Kvenvolden, K.A (1988a) Methanehydrate:A major reservoir of carbon
in the shallow geosphere?Chemical Geology, 11,4l-51
-(1988b)
Methane hydrates and global climate Global BiogeochemicalCycles,2, 221-228.
-(1993)
Gas hydrates: Geological perspectiyeand global change.
Reviewsin Geophysics,31, 173-187.
-(1994)
Natural gas hydrate occurrenceand issues In ED. Sloan
Jr, J Happel, and M A Hnatow, Eds., Internationalconferenceon natural gas hydrates,7l5, p 232-246 New York Academyof Sciences,
New York
Kvenvolden, K A, and Grantz, A (1990) Gas hydrates of the Arctic
Ocean region In The geology of North America, L, p 539-549 The
Geological Society of America, Boulder
Kvenvolden,K A., Ginsburg,G D., and Soloviev,V.A (1993)Worldwide
distribution of subaquaticgas hydrates Geo-MarineLeners, 13, 32-40
Larsen,R, Makogan,T.Y, Knight, C A, and Sloan,E D, Jr (1996)The
97
influenceof krnetic inhibitors on the morphology of sI and sII clathrate
hydrates In Proceedingsof the secondinternationalconferenceon natural gas hydrates,Toulouse,France,June 2-6,1996, p 163 170.
Lederhos,J P, Mehta,A P, Nyberg,G.B , Wam, K.J.,and Sloan,E D, Jr
(1992) Structure H clathrate hydrate equilibria of methane and adamantane.American Institute of Chemical EngineersJournal,38, 10451048
R L, and Sloan,E D., Jr. (1993)A first order
Lederhos,J P, Christiansen,
method of hydrate equilibrium estimation and its use with new strucnrres.Fluid PhaseEquilibria, 83,445-454
Lingelem, M.N, and Majeed, A I. (1992) Hydrate formation and control
in long distancesubmarinepipelines.tansactions of the Institution of
Chemical Engineers,70, 38-42.
Lingelem,M.N, Majeed,AI, and Stang,E (1994)Industrialexperience
in evaluationof hydrate formation, inhibition, and dissociationin pipeline design and operation In E D Sloan Jr., J Happel, and M A. Hnatow, Eds , Internationalconferenceon natural gas hydrates,715, p.7593 New York Academy of Sciences,New York.
MacDonald, G.J (1990a) Role of methaneclathratesin past and future
climates Climatic Change, 16, 241-281.
-(1990b)
The future of methaneas an energy resource.Annual Review of Energy,15, 53-83
Mak, TC W., and McMullan, R.K (1965) Polyhedral clathratehydrates:
X Sfixcfure of the double hydrate of tetrahydrofuranand hydrogen
sulfide. Journal of Chemical Physics,42, 2'732-2137
Makogan, YF (1981) Hydrates of natural gas,237 p. Pennwell,Tulsa.
-(1994)
Russia'scontribution to the study of gas hydrates In E D.
Sloan Jr, J. Happel, and M A Hnatow, Eds, Internationalconference
on naturalgas hydrates,715, p 119-160 New York Academyof Sciences,New York.
Mclver, R.D (1977) Hydratesof natural gas: Imponant agentin geologic
processes(abstr) Geological Society of America Abstracts with Prog r a m s , 9 ,1 0 8 9 1 0 9 0
-(1982)
Role of naturally occurring gas hydratesin sedimenttransport American Associationof PetroleumGeologistsBulletin, 66,789192.
McMullan, R.K, and Jeffrey, G A. (1965) Polyhedral clathratehydrates:
IX Structureof ethylene oxide hydrate. Journal of Chemical Physics,
42.2125-2732.
McMullan, R.K, and Kvick, A (1990) Neutron diffraction study of the
structure II clathrate hydrate: 3 5Xe8CCl, l36D,O at 13 and 100 K.
Acta Crystallographica,846, 390-399
Mehta, A P, and Sloan, E D , Jr (1993) Sfucture H hydrate phaseequilibria of methane + liquid hydrocarbonmixtures Journal of Chemical
EngineeringData, 38, 580-582.
Miiller, H R , and von Stackelberg,M ( I 952) Zur strukturder gashydrate
39, 20 -21
Naturwissenschaften,
Nisbet, E G (1990) The end of the Ice-Age CanadianJournal of Earth
Sciences.27. 148-157
-(1992)
Sourcesof atmosphericCH* in euly postglacialtime Journal of GeophysicalResearch,97, 12859-12867.
Ouar, H., Cha, S.B , Wilderman,T.R, and Sloan, E.D., Jr. (1992) The
formation of natural gas hydratesin water-baseddrilling muds. Transactions of the Institution of Chemical Engineers,70,48-54
Parrish, WR., and Prausnitz,J.M. (19'72) Dissociation pressuresof gas
hydratesformed by gas mixtures Industrial EngineeringChemical ProcessDesignand Development,lI,26-35.
Pauling, L, and Marsh, RE. (1952) The structure of chlorine hydrate
Proceedingsof the National Academy of SciencesU.S A., 38, 112-1 18
Paull, CK, Ussler,W., III, and Dillon, W.P (1991)Is the extentof glaciation limited by marine gas-hydrates?GeophysicalResearchLetters,
18,432-434
Reed,D L., Silver,J E, Tagudin,J E, and Shipley,T.H (1990)Relations
between mud volcanoes,thrust deformation, slope sedimentation,and
gas hydrate,offshorenorth Panama Marine Pefoleum Geology,7,44, I . , a n d P o w e l l , BM ( 1 9 8 7 )A n e w
R i f - e e s t e r , J . A .T, s e ,J S , R a t c l i f f eC
clathratehydrate structure Nature, 325, 135-136
Ripmeester,J A , and Ratcliffe, C I. (1990) tzqXeNMR studiesof clathrate
98
SMELIK AND KING: CRYSTAL GROWTH OF CLATHRATE HYDRATES
hydrates:New guestsfor structureII and structureH. Journal of Physicaf Chemistry, 94, 8773-87'76.
Ripmeester,J A , Ratcliffe, C.I., Klug, D D., and Tse, J.S. (1994) Molecular perspectiveson structure and dynamics in clathrate hydrates. In
E D. Sloan Jr., J. Happel, and M A. Hnatow, Eds., Internationalconferenceon natural gas hydrates,l15, p. 16l-116. New York Academy
of Sciences.
Sassen,R., and MacDonald, I R (1994) Evidencefor structureH hydrate,
Gulf of Mexico continental slope Organic Geochemistry,22, 10291032.
Sloan,ED., Jr (1990a)Clathratehydratesofnarural gases,365p. Marcel
Dekker. Inc.. New York
-(1990b)
Natural gas hydrate phaseequilibria and kinetics: Understanding the state-of-the-art.Revue de Llnstitut Francais du Petrole,
45.245-266.
-(199a'1
Conference overview. In ED. Sloan Jr, J Happel, and
M A Hnatow, Eds, Intemational conferenceon natural gas hydrates,
715, p I-23. New York Academy of Sciences,New York
Sloan,L.C., Walker,J C G, Moore,T.C.,Jr.,Rea,D.K., and Zachos,J C
(1992) Possiblemethane-inducedpolar warming in the early Eocene.
Nanre, 357, 320-322
Smelik, E.A, and King, H.E., Jr. (1996) Clathratehydratesas layer structures: Comparisonbetween the structuresof type II and hypothetical
type V Zeitschrift ftjr Kristallographie,2l l, 84-89
Thomas, M., and Behar, E. (1994) Structue H hydrate equilibrium of
methaneand intermediatehydrocarbon molecules.Proceedingsof the
'13.100-10'7
Annual Convention on Gas ProcessesAssociation.
Tse,J.S. (1990) Thermal expansionof structure-Hclathratehydrates Journal of Inclusion Phenomenaand Molecular Recognition in Chemistry,
8,25-32.
von Stackelberg,M, and Milller, H.R. (1951) On the structuresof gas
hydrates.Journal of Chemical Physics, 19, 1319-1320.
Menuscnrpr RECETvED
Aprur- 16, 1996
Mnwscnrpr AccEprEDOcroeen 9. 1996