Các định luật cơ bản và ứng dụng của khí Hydrates

Sloan

11/11/03

10:06 am

Page 37

insight review articles

Fundamental principles and
applications of natural gas hydrates
E. Dendy Sloan Jr
Center for Hydrate Research, Colorado School of Mines, Golden, Colorado 80401, USA (e-mail: esloan@.mines.edu)

Natural gas hydrates are solid, non-stoichiometric compounds of small gas molecules and water. They form
when the constituents come into contact at low temperature and high pressure. The physical properties of
these compounds, most notably that they are non-flowing crystalline solids that are denser than typical fluid
hydrocarbons and that the gas molecules they contain are effectively compressed, give rise to numerous
applications in the broad areas of energy and climate effects. In particular, they have an important bearing on
flow assurance and safety issues in oil and gas pipelines, they offer a largely unexploited means of energy
recovery and transportation, and they could play a significant role in past and future climate change.

U

ntil mankind learns how to economically generate hydrogen for fuel cells, natural gas will be the
premium fuel for this century for two reasons.
First, gas burns cleanly, causes few pollution
problems and, relative to oil or coal, produces less
carbon dioxide. And second, liquid fuels are better used as
feedstocks (raw material) for generation of petrochemicals.

Two examples herald this coming change: many power
plants are being converted from coal to natural gas, and fleets
of cars have been converted from petrol to natural gas fuel.
As we deplete the readily accessible reserves, we will need
to obtain natural gas from conditions that are both more
severe and more remote. We will need to explore deep ocean
environments with higher pressures, and permafrost environments with lower temperatures than we presently do.
And gases that were previously thought to be uneconomical,
such as those containing non-combustible components of
nitrogen, carbon dioxide and hydrogen sulphide, will also
be explored. Such unusual conditions also promote the formation of a solid compound of gas and water — namely
clathrates of natural gas — commonly called gas hydrates.
Here, I indicate the motivation for hydrate science and
engineering; that is, the applications where technical workers
find a use for the physics, chemistry and biology that are
associated with science. This is not to indicate that hydrate
engineering is simply an applied science. As indicated
recently in a defining book1 on differences in technical philosophy, engineering frequently cannot afford the luxury of
a thorough scientific foundation, and must proceed at risk.
As only one example, the past decade’s development of the
new ‘low-dosage’ pipeline hydrate inhibitors proceeded in
an Edisonian research mode (a process of intelligent guesswork, intuitive reasoning and testing), and scientific
progress is currently being made to refine trial and error
research gaps in inhibitor developments.
Below, I describe five major applications of hydrate
research: flow assurance, safety, energy recovery, gas storage/transportation and climate change. Before the applications are addressed, an introduction to hydrate structures
and their overall properties is presented. I conclude this
review with an outline of future challenges. For readers who
want a more detailed understanding, several hydrate
reviews2–8 are available.

Hydrate structures
Clathrate hydrates typically form when small (<0.9 nm)
‘guest’ molecules such as methane or carbon dioxide contact

water at ambient temperatures (typically less than 300 K)
and moderate pressures (typically more than 0.6 MPa). On a
molecular scale, single small guest molecules are encaged
(enclathrated) by hydrogen-bonded water cavities in these
non-stoichiometric hydrates. Guest-molecule repulsions
prop open different sizes of water cages, which combine to
form the three well-defined unit crystals shown in Fig. 1,
and structural details and references have been provided in a
recent book2.
Cubic structure I predominates in the Earth’s natural
environments, and contains small (0.4–0.55 nm) guests;
cubic structure II generally occurs with larger (0.6–0.7 nm)
guests in mostly man-made environments; and hexagonal
structure H may occur in either environment, but only with
mixtures of both small and large (0.8–0.9 nm) molecules.
The smallest hydrated molecules (Ar, Kr, O2 and N2), with
diameters less than 0.4 nm, form structure II. Most hydrate
science, and thus most applications, concentrates on structure I and structure II, with structure H in anecdotal occurrence. Although this review will emphasize structure I and
structure II, many analogues occur for structure H hydrates.
Figure 1b lists the properties of the three common unit
crystals. Water molecules form hydrogen bonds in a basic
building block for both structure I and structure II, called
the 512 (pentagonal dodecahedra) because there are 12 faces
of pentagonally bonded water molecules in that cavity.
Within the cavity, small guest molecules are enclathrated,

with limited translation motion but substantially more
rotation and vibration ability. The 512 building blocks are
joined to other 512s either through the vertices (structure I)
or through the 512 faces (structure II).
Yet all structures need to fill space within their cavities to
prevent hydrogen-bond strain and breakage. The 512 basic
building blocks cannot fill space without bond breakage,
and so interstices between the 512 cages are filled with other
cavities that relieve the strain by incorporating hexagonal
faces — two in the 51262 cavity of structure I, and four hexagonal faces in the 51264 of structure II, in addition to the
12 pentagonal faces in each cavity. Thus, the cages can contain larger guest molecules.The cages form basic repeating
unit crystals with ratios of 2•512 +6•51262 in structure I and
16•512+8•51264 in structure II, indicating sixteen 512s and
eight 51264s in the structure II unit crystal. Although both
large and small cages are present in each crystal structure,
sometimes single guests are too large for the smaller cage,
which must go empty so that only the larger cage is filled.
However, smaller molecules can fill both cages.

NATURE | VOL 426 | 20 NOVEMBER 2003 | www.nature.com/nature
©2003 Nature Publishing Group

© 2003 Nature Publishing Group

353


Sloan

11/11/03

10:06 am

Page 38

insight review articles
Cavity types

a

‘Guest molecules’

Hydrate structure

6

51262

2

46 H2O

Methane, ethane,
carbon dioxide
and so on

136 H2O

Propane,
iso-butane

and so on

34 H2O

Methane + neohexane,
methane + cycloheptane,
and so on

Structure I

16

8

Figure 1The three common hydrate unit
crystal structures. Nomenclature: 51264
indicates a water cage composed of 12
pentagonal and four hexagonal faces. The
numbers in squares indicate the number of
cage types. For example, the structure I unit
crystal is composed of two 512 cages, six
51262 cages and 46 water molecules.

51264

512

Structure II

3

2
435663

1
51268
Structure H

b
Hydrate crystal structure

I

II

H

Cavity

Small

Large

Small

Large

Small

Medium

Large

Description

512

51262

512

51264

512

4 35 66 3

51268

6

16

8

3

2

1

4.06†

5.71†

20

36

Number of cavities per unit cell 2
Average cavity radius (Å)

3.95

4.33

3.91

4.73

3.91†

Coordination number*

20

24

20

28

20

Number of waters per unit cell

46

136

34

*Number of oxygens at the periphery of each cavity.
†Estimates of structure H cavities from geometric models.

In all three structures, typically there is only one guest molecule
within each cage. However, at unusual conditions such as very high
pressure, it is possible to have multiple-cage occupancy with unusually small guests, such as hydrogen or noble gases. For example, it was
recently shown9 that hydrogen can form hydrates at very high pressure with as many as two occupants in the small cage and four in the
large cage of hydrate structure II. However, very small guests and
multiple occupancies are considered an aberration.
Remarkably, when all hydrate cavities are filled, the three crystal
types have similar concentrations of components: 85 mol% water
and 15 mol% guest(s). Hydrate formation is most likely to take place
at the interface between the bulk guest and aqueous phases, because
the hydrate component concentrations greatly exceed the mutual
fluid solubilities. The solid hydrate film at the interface acts as a barrier
to prevent further contact of the bulk-fluid phases, and fluid surface
renewal is required for continued clathrate formation.

Guest to cage ratio
Examination of the size ratios of the guest molecules to the cages they
occupy can aid the understanding of hydrates. This heuristic controls
not only the occupancy but also the thermodynamic properties of
these structures, to a first approximation.

Simple hydrate guests
A number of researchers have commented on the fit of the guest molecule within the hydrate cage, beginning with von Stackelberg10,
whose modified original diagram is shown in Fig. 2. Table 1 lists the

size ratios of guests of natural gas in the four common cavities of
structure I and structure II. The discriminating size ratio is not
absolute for each cage, instead it occurs over a molecular size range,
which has a number of important implications. The implication is
that clathrate hydrates have no fixed size ratio of guest to host, as
shown by the ranges in Fig. 2. It is particularly interesting to note the
resulting behaviour of molecules at cage size borders.
Take, for example, the most common natural gas compound,
methane, in the phase diagram for methane and water in Fig. 3 (ref. 11).
The non-stoichiometry of the hydrate causes the hydrate composition area (vertical axis parabola) shown in green in Fig. 3 rather than
the vertical stoichiometric line originally proposed12 for methane
hydrate. Assuming equilibrium, the implication of this hydrate phase
area is that laboratory-made hydrates (from methane-rich systems)
may differ slightly in composition from in situ hydrates, which can
form in methane-lean systems. This methane hydrate composition
difference, although small (4%), when multiplied by the entire
hydrate reserve is sufficient to supply the USA for 600 years at the current energy usage.

Binary hydrate guests
For binary systems, Holder and Manganiello13 indicated that an optimized fit of guests in the cages was sufficient for hydrate azeotropes,

for which the vapour composition is equal to that of the hydrate. This
is particularly unusual because azeotropy in vapour–liquid equilibria is only possible for a non-ideal solution (with an activity coefficient), whereas hydrates typically form ideal solid solutions.

354
©2003 Nature Publishing Group

© 2003 Nature Publishing Group

NATURE | VOL 426 | 20 NOVEMBER 2003 | www.nature.com/nature


Sloan

11/11/03

10:06 am

Page 39

insight review articles
The earlier measurements of von Stackelberg10 showed that some
single guest molecules that are structure I formers will form structure II
in binary guest mixtures — a somewhat counter-intuitive notion.
This idea was later predicted14 and measured15 over a wide composition range for methane and ethane mixtures. However, other structure I simple guest combinations such as methane and carbon dioxide will not form stable structure II as binary mixtures16. The question arises as to why this occurs for methane and ethane mixtures but
not for methane and carbon dioxide mixtures.
The reason for such structural transitions was addressed initially

Size (Å)

Hydrate

former

3

Cavities
occupied

No hydrates

by Ripmeester3 and subsequently by Hester and Sloan17. The structure
II transition occurs when two molecules are at each extreme of the
structure I molecular sizes shown in Fig. 2. That is, small structure I
formers in the 512 cage (which are almost small enough to form structure II) and large structure I formers in the 51262 cage (which are
almost large enough to form structure II) will, when mixed, cause
structure II to form from two structure I formers.
Two other examples serve to indicate that hydrate structure
research has been a minefield for the complacent who think that
everything structural is known. First, for almost 40 years, until the
prediction of structure II formation13 by the smallest guests was validated by Davidson et al.18, it was erroneously thought that they formed
structure I as simple hydrates. Second, in 1987, structure H was discovered19, but on close inspection structure H crystals had been measured much earlier in the diffraction data of von Stackelberg10. Again
for four decades, the hydrate community paradigm forced the data
into structure I and structure II categories, counting structure H as an
aberration.

Physical properties and implications
Several key physical properties of hydrates determine the roles that
they play (or might play in the future) in both industry and the environment. They are solids with densities greater than those of typical
fluid hydrocarbons, and this has practical implications for flow
assurance in pipelines and the safety thereof. Furthermore, the fact
that, in effect, hydrates concentrate their guest molecules results in

three potential applications: that energy can be recovered from in situ
hydrates; that hydrates can be used to transport stranded gas; and
that hydrates may be a factor in climate change. Each of these implications and applications is briefly discussed.

Ar

512 + 51264
4

Kr

Structure II
52/3 H2O

N2
O2
CH4

Xe

H2S

Flow assurance
53/

5

4

H 2O

First, and perhaps most importantly, when hydrates form, they are
solid, non-flowing crystalline structures. This leads to the most
urgent consideration — that of flow assurance in oil and gas
pipelines. Oil and gas wells always produce undesired water along
with hydrocarbons that are in the hydrate guest size range. As the
flowing mixed phases cool, hydrates form and plug transmission
lines, causing costly production stoppages, sometimes for as long as
months, in large pipelines, while the hydrates are dissociated.
The petroleum industry would like to maintain their processes
outside the hydrate stability range. Fortunately, the hydrate stability
temperature and pressure range is predictable to within experimental
accuracy using modern thermodynamic programs usually based
upon a Gibbs energy extension20 of the van der Waals and Platteeuw21
method. Unfortunately, however, low temperatures (such as the deepsea floor temperature of 277 K) and the mandates of high pressure for
economic energy densities place many pipelines well within the
hydrate-formation region. High pressures and low temperature
require hydrate-inhibition methods. During the past decade this phenomenon has initiated a new type of engineer — the flow-assurance
engineer — whose major objective is to prevent pipeline blockages,

512 + 51262
Structure I

CO2

C2H6

51262
72/3 H2O

Structure I

C-C3H6

6
(CH2)3 O
C3H8
iso-C4H10

51264
17 H2O

Structure II

7
n-C4H10

Table 1 Ratios of molecular diameters* to cavity diameters† for some

No SI or SII
hydrates

molecules including natural gas-hydrate formers
Molecule

Guest diameter Structure I
(Å)
512

Figure 2 Hydrate guests versus hydrate cavity size ranges. Along the line are the

size of the guest molecules in hydrates. The broad shaded areas and the numbers on
the right discriminate the number of water molecules in hydrates occupied by single
guest occupants shown on the left. For example, methane has a typical hydration
number of 53⁄4 and occupies both the 512 and 51262 cavities of structure I. However,
propane is so large, it can only fit into the largest structure II hydrate cavity (51264).
Adapted from ref. 43. Copyright Geological Society, London.

NATURE | VOL 426 | 20 NOVEMBER 2003 | www.nature.com/nature

N2
CH4
H2S
CO2
C2H6
C3H8
i-C4H10
n-C4H10

4.1
4.36
4.58
5.12
5.5
6.28
6.5
7.1

0.804
0.855F
0.898F

1.00
1.08
1.23
1.27
1.39

51262
0.700
0.744F
0.782F
0.834
0.939F
1.07
1.11
1.21

Structure II
512

51264
F

0.817
0.868
0.912
1.02
1.10
1.25
1.29
1.41

0.616F
0.652
0.687
0.769
0.826
0.943F
0.976F
1.07

F indicates the cavity occupied by the single hydrate guest.
*Molecular diameters obtained from von Stakelberg12.
†
Cavity radii from Table 2-1 minus 1.4 Å water radii.

355
©2003 Nature Publishing Group

© 2003 Nature Publishing Group


11/11/03

10:06 am

Page 40

insight review articles
primarily of hydrates, but also of waxes and other solids.
To provide flow assurance, the energy industry injects hydrogenbonding fluids (for example, alcohols or glycols) into the flowing

stream at the wellhead to compete with solid hydrates for the available water. Worldwide methanol costs for hydrate inhibition are estimated at US$220 million annually (P. K. Notz, personal communication). In addition, severe financial penalties are paid for large
methanol storage capacity on offshore platforms and for greater than
50 p.p.m. methanol contaminations in refinery feedstocks. During
the production process from hydrocarbon reservoirs, any amount of
water, which generally increases over the life of the well, must be
hydrate inhibited. Although the aqueous phase is the primary place
where hydrate inhibition occurs, a small concentration (typically less
than 0.1 mol%) of methanol partitions into the hydrocarbon vapour
and liquid phases.
Yet the bulk of methanol is lost to the hydrocarbon phases
because, even with low methanol concentrations, the hydrocarbon
phase amount greatly exceeds that of the aqueous phase. Methanol is
a difficult molecule to model partitioning into both hydrocarbon liquid and vapour phases, particularly with high accuracy, at low concentrations. If we apply Hildebrand’s solubility maxim, ‘like dissolves
like’, we conclude that the methyl group makes methanol soluble in
hydrocarbons, whereas the hydroxyl group increases its solubility in
water. Thus any equation-of-state that models methanol partitioning must address the challenge of modelling hydrocarbon and aqueous phases equally well.
In addition to hydrogen-bonding inhibitors, which require high
concentrations, new low-dosage hydrate inhibitors — both kinetic
inhibitors and anti-agglomerants — have been developed in the past
decade to prevent hydrate crystal growth and agglomeration, respectively. These new chemicals are rapidly being adopted in the field22
and provide a fertile research area for molecular modelling.
Safety

Secondly, the hydrate solid specific gravity is typically 0.9 compared
with typical fluid hydrocarbon specific gravities of 0.8 or less. This
higher density leads to the problem of ensuring hydrate safety and
preventing the annual loss of property and lives. When hydrate
blockages dissociate in pipelines, they detach first at the pipe wall;
therefore, any pressure gradient across the high-density hydrate plug
will cause the hydrate to travel rapidly (measured at 300 km hr–1)

down the pipeline. This effect will compress the downstream gas,
either causing pipeline blowouts or causing the plug to erupt though
pipeline bends. Case studies of hydrate safety problems are given in a
monograph on this subject4.
A second safety concern arises when hydrate plugs are locally
heated (for example, using a blowtorch outside a pipeline) to dissociate them. Frequently, the evolving gas from the hydrate is contained
by the ends of the plug until the pipeline bursts owing to the pressure
being too high. This safety concern is a result of the next hydrate
property — the ability of hydrates to concentrate high levels of gas in
a hydrated form.
Energy recovery

When small (Յ0.9 nm) hydrocarbon guest molecules are encaged in
hydrates, with typically one molecule per cavity, the guests are separated approximately 0.5 nm by the water cages. This means that the
energy density in hydrates is approximately the same as that of a compressed gas, that is, less than the energy density of liquefied natural
gas (LNG). For example, if every hydrate cavity were filled with a
guest molecule, one volume of hydrate would dissociate to 180 volumes (STP) of gas. The gas concentration in clathrates is comparable
to that of a highly compressed gas (that is, methane gas at 273 K and
18 MPa).
A large fraction of the Earth’s fossil fuels is stored in clathrate
hydrates. Even the most conservative current estimates23 suggest that
the amount of energy in hydrates is equivalent to twice that of all

V

V
LW
H
M
LM

I

Vapour
Liquid water
Hydrate
Solid methane
Liquid methane
Ice

V–LW
Temperature

Sloan

Previous hydrate line
by Kobayashi and Katz
LW
H–LW
H–V

LW–I

LM–H

H–I

LM–H

H
M–H

H 2O

CH4·?H2O

M–LM
CH4

Figure 3 The isobaric methane and water phase diagram. Compare the vertical
parabolic hydrate area (green) with the previous vertical stoichiometric hydrate line of
Kobayashi and Katz12. The parabolic region is a result of incomplete filling of the
small cages (512) in structure I. Variation in hydrate cage filling and the resulting
hydrate parabola is a function of temperature, overall methane composition and
pressure (not shown here).

other fossil fuels combined. Most of the natural-gas-containing
hydrates are in the ocean bottom, and although production of gas
from such deep-lying hydrates is now too expensive, it is likely that
within the next two decades we will tap that fuel source to meet growing energy demands. Table 2 compares hydrated methane to that in
conventional reserves for 11 arbitrary divisions of the world.
Most of the natural hydrates around the world are biogenic — the
guest gas comes from bio-degraded plant and animal matter that
have been buried in the sea floor at low temperature over long periods. Substantial but anecdotal evidence exists for thermogenic
hydrates from deeper gas sources in places like the Gulf of Mexico24
and the Caspian Sea25. Most of the estimates of gas hydrates have
come from indirect seismic evidence using a bottom simulating
reflector (BSR), which indicates reflections from gas at the base of the
hydrate (see Fig. 4).
BSR indications are not totally reliable, and other more accurate
methods are needed. There are a significant number of cases in which
hydrates occurred without bottom simulating reflections, or when

the BSR did not indicate the presence of hydrates. Notwithstanding
this problem, the resource numbers are so large that they warrant
energy-recovery studies, even if they are in error by as much as two
orders of magnitude.
Much of the available public funding of hydrate research has been
channelled toward industrial field experiments, aimed at the production of energy. Although results from industrial experiments (for
example, in Alaska and in Japan) may not be publicly available, results
from two recent drilling expeditions are soon to be published. These
detailed field experiments will probably serve as design bases for the
foreseeable future, owing to their thorough documentation.
Pilot drilling, characterization and production testing of hydrates
have begun in permafrost regions, which have higher concentrations
of hydrates (for example, 30 vol.% in the 1998 Mallik 2L-38 well in
Canada), to learn how to approach the more dispersed, but much
greater, ocean resource in the future. The Mallik 5L international
field experiment was concluded in March 2002 on Richards island in
the MacKenzie delta of Canada at a cost of US$17 million. This permafrost experiment provided the first direct evidence that hydrates

356
©2003 Nature Publishing Group

© 2003 Nature Publishing Group

NATURE | VOL 426 | 20 NOVEMBER 2003 | www.nature.com/nature


11/11/03

10:06 am

Page 41

insight review articles
could be economically recovered at high concentrations. The information provided by this experiment, when it becomes available in a
Geological Survey of Canada report early in 2004 (ref. 26), will be a
landmark upon which the industry will base designs for energy
recovery.
For ocean hydrates, the benchmark will be Leg 204 of the Ocean
Drilling Program (ODP), completed in September 2002 on Hydrate
ridge of the Cascadia Margin off Oregon. This is one of the first ocean
drillings for hydrates to recover large, pervasive hydrates, other than
anecdotal evidence. Much of the information from Leg 204 was presented in preliminary findings at the combined American Geological
Union, European Geophysical Society, European Union of Geosciences meeting in Nice (April 7–10, 2003). This drilling, the result
of which should appear in print in the final quarter of 2003, provides
the first evidence that ocean hydrates may be present in sufficient
concentration to be economically producible.
Until the publication of the benchmark results from the Mallik 5L
and Leg 204 wells, the best literature for natural hydrates can be
found in summaries and volumes about the Mallik 2L well27, the
Blake Ridge ODP28 Leg 164 and the Gulf of Mexico24.
It should be noted that the amount of energy in ocean hydrates is
several orders of magnitude greater than that in permafrost hydrates.
Put another way, the error in the ocean hydrated energy estimate is
greater than the entire permafrost hydrated energy estimate. Until
ODP Leg 204, however, it was thought that the most concentrated
hydrates were in the permafrost, which provided more accessible
recovery.
Methods for the economic recovery of methane from natural
hydrates are uncertain, and substantial creativity has gone into devising new recovery methods, as well as into applying existing oil and gas
technology to hydrate recovery. All recovery methods apply one or

more of the following three principles: (1), reduction of the pressure
below that of hydrate stability; (2), addition of enough energy to disrupt the water hydrogen bonds; and (3), addition of strong hydrogen-bonding chemicals (such as alcohol or glycol) to disrupt the
hydrate structure at reservoir conditions.
Finite difference reservoir recovery models29 indicate that gas
production is only economical at rates larger than 500,000 standard
cubic metres (0.1 MPa and 289 K) per day. This will require both
depressurization and thermal/inhibitor stimulation. The most producible of the permafrost hydrate deposits are those lying adjacent to

a gas reservoir, because free gas production will dissociate hydrates by
decreasing reservoir pressures below hydrate stability. Heat from the
Earth allows hydrate decomposition to slowly replenish the gas reservoir. Makogon5 indicated that a Siberian permafrost reservoir was
produced in this manner during the 1970s.
Gas production from hydrates close to conventional permafrost
reservoirs will begin in the West during the next decade at incremental costs over normal gas production. Production from stand-alone
hydrates in the permafrost or in the ocean will be much more costly,
but is technically feasible. Both Japanese and American programmes
forecast that stand-alone hydrated energy recovery will begin by
2015.
Storage and transportation

It is estimated that about 70% of the total gas reserve is either too far
from an existing pipeline or too small to justify a liquefaction facility.
Gudmundsson and Borrehaug30 suggested that it is economically feasible to transport stranded gas in hydrated form. In the fourth international hydrate conference, workers from Mitsui Shipbuilding31
showed that work in conjunction with the Japan Maritime Research
Institute32 provides a basis for extending the basic concept of
Gudmundsson et al.33 to transport stranded gas.
Climate change

A recent publication34 thoroughly documents evidence for Late Quaternary climate change caused by hydrates, commonly called ‘the
hydrate gun hypothesis’. The concept is that, as little as 15,000 yr ago,

methane from hydrates caused significant global warming.
The hydrate gun hypothesis seems analogous to another, somewhat less controversial, hypothesis, proposed by Dickens et al.35–37.
They suggested that an ancient (55.5 Myr ago), massive ocean
methane hydrate dissociation might explain a 4–8 °C temperature
rise over a brief geological time interval (103 years) called the Late
Palaeocene Thermal Maximum (LPTM). This is documented in
deep ocean drilling samples as a prominent negative carbon isotope
(Ȏ13C = –2.5‰) in ocean sediments, in fossil tooth enamel, and in
carbonates and organic sediments in terrestrial sequences. This Ȏ13C
reduction in the ocean and the recovery over the ensuing 0.2 million
years (see Fig. 5a) is consistent with pronounced dissolution of calcium carbonate in the deep sea sediment deposited during the LPTM,
as shown in Fig. 5b.

SW

NE
10 km

Line 8
Two-way traveltime (s)

VE =10.0

3.5
4.0
4.5

Figure 4 A seafloor slump in the Blake-Bahama
ridge shown in both seismic (top) and cartoon
(bottom) relief40. Note the bottom simulating

reflector (BSR) parallel to the ocean bottom,
except in the middle section (dotted line) where it
appears that a seafloor eruption has occurred.
Reproduced with permission from ref. 40.
Copyright Geological Society, London.

5.0
SP 2,000

1,500

1,000

500

SW
Two-way traveltime (s)

Sloan

3.5

NE
10 km

Line 8

VE =10.0

4.0

BSR
4.5
5.0
SP 2,000

1,500

NATURE | VOL 426 | 20 NOVEMBER 2003 | www.nature.com/nature

1,000

500

357
©2003 Nature Publishing Group

© 2003 Nature Publishing Group


11/11/03

10:06 am

Page 42

insight review articles
Table 2 Conventional and hydrated gas resources in trillion cubic metres

a
Warm surface δ13C (‰)

3.0
2.0
Shallow Atlantic
1.0
0.0

–1.0
–2.0
50
100
Time after methane release (103 yr)

0

b

150

Shallow Atlantic δ13C (‰)

2.0
ODP site 1001:
Caribbean bulk carbonate

1.0
0.0

–1.0
–2.0

0

c

50
100
Time after the LPTM (103 yr)

150

CH4 release (1014 g yr)

2.0
1.5
1.0

Average flux: 1.12 × 1014 g yr

0.5
0
0

d

50
100
Time after the LPTM (103 yr)

150

1,000
D
12CO

2O2

2

(+2H2O)

1,200
12CH

1,400

B
∆T = +4°C

1,600

12CH

4

4

(+6H2O)

C

6H2O
A

Water
Sediment
Orig

inal

New
geot

herm

geo

the

rm

1,800
CH4-hydrate-pore water
equilibrium curve
2,000

10

15

20

Temperature (°C)

25

Region

Conventional gas
(TCM)

Methanehydrate
(TCM)

North America
Latin America and Caribbean
Western Europe
Central and Eastern Europe
Former Soviet Union
Middle East and North Africa
Sub-Saharan Africa
Central Asia and China
Pacific OECD
Other Pacific Asia
South Asia

32.82
21.1
15.27
2.05
117
77.2

13.9
10.07
2.68
11.18
4.72

6,853
5,139
856
0
4,711
214
429
429
1,713
214
429

Total

310.3

20,987

TCM, trillion cubic metres

3.0

Depth (m below sea level)

Sloan

30

Figure 5 Hypothesized causes of the Late Paleocene Thermal Maximum (LPTM). a,
Carbon isotope reduction and recovery during LPTM. b, Dissolution of calcium
carbonate during LPTM. c, Evolution of methane from hydrates. d, Initial shift in
ocean hydrate equilibrium curve to cause the methane release. In sequence: in d,
the geotherm shifted by 4 °C, causing release of a large quantity of methane from
hydrates, shown in c. The result was a Ȏ13C isotope reduction and recovery, as
shown in a and b through conversion of methane, first to carbon dioxide and then to
calcium carbonate. Adapted from ref. 44. Copyright Société géoglogique de France.

In the LPTM hypothesis, the evolution of a large amount of
methane from hydrates (1.12ǂ1018 g of methane) is the only plausible explanation that has been offered to explain this environmental
perturbation. The abnormal Ȏ13C isotope indicates that the source
was external to the normal ocean–atmospheric–biomass carbon
pool. Figure 5c shows a rapid evolution of methane from hydrates;
methane is hypothesized to be oxidized to carbon dioxide that is
greatly enriched in Ȏ12C.
Figure 5d shows the hydrate equilibrium curve as a function of
depth and temperature in the ocean. Hydrates are only stable
between the equilibrium line and the original geotherm to the left of
the curved line, at depths below the sediment surface, shown by the
small vertical rectangle at A. In the LPTM, if the ocean was warmed by
4 °C, the hydrates between the original geotherm and the equilibrium
curve would melt, as the new geotherm was established. The warming from the original to the new geotherm would result in methane
expulsion to the environment, where it would be oxidized to carbon
dioxide, resulting in significant further warming. It was hypothesized that the resulting carbon dioxide was re-absorbed by the ocean
over the ensuing 0.2 Myr.

The importance of the LPTM perturbation is that it is the first well
documented instance of an explanation for how the global carbon
cycle and other systems relate to a rapid, massive input of fossil fuel,
such as may occur in modern industrial times. The data and summary in the publication by Kennett34 are the most thorough source of
information in support of extending the theory to more modern
times (the Late Quaternary), about 15,000 yr ago. However, there is a
considerable controversy concerning the validity of the hypothesis,
as suggested below.
In the recent meeting, it was suggested38 that ‘the hydrate gun is
firing blanks’, and that the atmospheric methane spike was due to
emissions from wetlands and peat bogs. This new theory requires a
glacial–interglacial vegetation time shift of 1,000 Gt C, which the
proposers of the theory, Maslin and Thomas, admit is difficult.
However, even this counter hypothesis requires some hydratederived methane for a mass balance, along with a shift in time for
the wetlands.
In a review of the hydrate gun monograph34, Dickens39 generally
concurs with the theory, but criticizes it on the grounds that it “perpetuates the common misconception that present-day methane
hydrates are stable. These systems may be in a steady state, but they
must be viewed as dynamic, with large carbon fluxes to and from the
ocean, even at [the] present day”.
In closing the discussion on hydrate-related climate change, it
should be noted that seafloor hydrate dissociation is also directly
related to slumping of sediments on the sea floor. Significant hydrated sediment slumps in the ocean can jeopardize the foundation of
sub-sea structures, such as platforms, manifolds and pipelines. The
single incident off the Carolina coast shown in Fig. 4 took place about
15,000 years ago40 and increased the extant Earth’s atmospheric
methane by as much as 4%. The effect of subsidence on sub-sea structures and foundations represents the initial meeting point for the two

358
©2003 Nature Publishing Group

© 2003 Nature Publishing Group

NATURE | VOL 426 | 20 NOVEMBER 2003 | www.nature.com/nature


Sloan

11/11/03

10:06 am

Page 43

insight review articles
energy communities — the first is concerned with hydrates in the
Earth, and the second with concerns for hydrates in man-made production systems. The interested reader is referred to the recent
monograph41 on this topic.

Future challenges
Currently, it appears that hydrate research has acceptably addressed
the thermodynamic challenge for most conditions. Time-independent hydrate quantification, courtesy of extensions20 of the van der
Waals and Platteeuw21 model, is at the bounds of experimental accuracy, and is the most common industrial exemplar of statistical thermodynamics use.
The most accurate thermodynamics have been provided
through modern spectroscopy for hydrate phase measurements, by
Raman, NMR and diffraction spectroscopy, through the bridge of
statistical thermodynamics to the macroscopic domain. The incorporation of these three spectroscopic methods have enabled more
accurate descriptions of hydrate mixtures, so that mixture thermodynamic predictions no longer depend solely on single hydrate
guest measurements. An overview of such hydrate spectroscopic
methods and results is provided in a recent review6.

Although the central concerns of hydrate thermodynamics have
been addressed, challenges remain at the periphery — for example,
at very high pressures (>1,000 bar), in unusual fluids such as black
oils, hydrate–sediment mixtures, and the methanol-partitioning
challenge indicated earlier. As an example of one such challenge,
hydrate– sediment mixtures have an unexplained thermal diffusivity maximum when plotted against sediment concentration. As we
begin to examine hydrates in nature, such challenges for time-independent properties will require decades to resolve.
However, the largest challenge is to describe the kinetics of
hydrates42. The fact that hydrates are solid compounds makes their
slow, solid-phase kinetics particularly challenging to researchers. An
additional challenge arises from the fact that hydrate solids form
interfacial barriers between the liquid and vapour phases that typically compose them. Hydrate research is most accurate when studying a
time-independent target. Typically, time-dependent (kinetic)
research is much more difficult and at least an order of magnitude of
accuracy is lost, relative to time-independent (thermodynamic)
research.
The use of kinetic-model results to predict data from other laboratories is problematic. Molecular dynamic simulations of hydrate
kinetics have been hindered by stochastic nucleation and the large
number of molecules and time required for growth processes. More
hydrate phase measurements are required to provide a needed
breakthrough — a unified hydrate kinetics model.

Conclusions and outlook
Wherever small molecules contact water, the potential for a hydrate
phase should be considered. The size ratio (guest to cavity) determines hydrate structural stability to a first-order approximation.
Other simple hydrate properties such as solid behaviour, density and
concentration of guest molecules affect the major applications of
hydrate safety, flow assurance, energy production and storage and
climate change.
During the next decade, gas production will begin from permafrost hydrates associated with conventional gas reservoirs. However, efficient production of ocean hydrates is problematic and

requires an engineering breakthrough to be economically feasible.
Yet, the potential to tap the Earth’s largest hydrocarbon energy
resource cannot be ignored.
Although hydrate thermodynamics are understood to an acceptable degree for most engineering applications, the kinetics arena will
represent the largest challenge for advancing the information on
hydrates. Although we know quite a lot about what hydrates are, the
question of how hydrates form is still very much unanswered. FindNATURE | VOL 426 | 20 NOVEMBER 2003 | www.nature.com/nature

ing the answers to such questions provides the intrinsic motivation
for future research.
■
doi:10.1038/nature02135
1. Koen, B. V. Discussion of the Method: Conducting the Engineer’s Approach to Problem Solving (Oxford
Univ. Press, Oxford, 2003).
2. Sloan, E. D. in Clathrate Hydrates of Natural Gases (Marcel Dekker, New York, 1998).
3. Ripmeester, J. A. Hydrate research – from correlations to a knowledge-based discipline: the
importance of structure. Ann. NY Acad Sci. 912, 1–16 (2000).
4. Sloan, E. D. in Hydrate Engineering (Soc. Petrol. Eng., Richardson, TX, 2000).
5. Makogon, Y. F. Hydrates of Hydrocarbons (Pennwell Books, Tulsa, 1997).
6. Sloan, E. D. Clathrate hydrate measurements: microscopic, mesoscopic, and macroscopic. J. Chem.
Thermo. 35, 41–53 (2003).
7. Davidson, D. W. in Water: A Comprehensive Treatise (ed. Franks, F. ) (Plenum, New York, 1973).
8. Englezos, P. Clathrate hydrates. Ind. Eng. Chem. Res. 32, 1251–1274 (1993).
9. Mao, W. L. et al. Hydrogen clusters in clathrate dydrate. Science 297, 2247–2249 (2002).
10. von Stackelberg, M. Solid gas hydrates. Naturwissenschaften 11-12, 1–22 (1949).
11. Huo, Z., Hester, K., Miller, K. T. & Sloan, E. D. Methane hydrate non-stoichiometry and phase
diagram. Am. Ind. Chem. Eng. J. 49, 1300–1306 (2003).
12. Kobayashi, R. & Katz, D. L. Methane hydrate at high pressure. J. Petrol Technol. 2579, 66–70 (1949).
13. Holder, G. D. & Manganiello, D. J. Hydrate dissociation pressure minima in multicomponent
systems. Chem. Eng. Sci. 37, 9–16 (1982).

14. Hendriks, E. M., Edmonds, B., Moorwood, R. A. & Szczepanski, R. Hydrate structure stability in
simple and mixed dydrates. Fluid Phase Equilibria 117, 293–298 (1996).
15. Subramanian, S., Kini, R. A., Dec, S. F. & Sloan, E. D. Evidence of structure II hydrate formation from
methane+ethane mixtures. Chem. Eng. Sci. 55, 1981–1985 (2000).
16. Staykova, D. K., Hansen, T., Salamatin, A. N. & Kuhs, W. F. Kinetic diffraction experiments on the
formation of porous gas hydrates. Proc. 4th Int.Conf. Gas Hydrates (2002).
17. Hester, K. & Sloan, E. D. Structure II hydrates from binary structure I simple guests. Fluid Phase
Equilibria (in the press).
18. Davidson, D. W., Handa, Y. P., Ratcliffe, C. I., Tse, J. S. & Powell, B. M. The ability of small molecules
to form clathrate hydrates of structure II. Nature 311, 142–143 (1984).
19. Ripmeester, J. A., Ratcliffe, C. I. & Powell, B. M. A new clathrate hydrate structure. Nature 325,
135–136 (1987).
20. Ballard, A. & Sloan, E. D. The next generation of hydrate prediction: An overview. Proc. 4th Int.Conf.
Gas Hydrates (2002).
21. van der Waals, J. H. & Platteeuw, J. C. Clathrate solutions. Adv. Chem. Phys. 2, 1–58 (1959).
22. Mehta, A. P., Hebert, P. B., Cadena, E. R. & Weatherman, J. P. Fulfilling the promise of low-dosage
hydrate inhibitors: Journey from academic curiosity to successful field implementation. SPE Prod.
Facil. 73–79 (2003).
23. Kvenvolden, K A. in Methane Hydrates: Resources in the near Future? (Proc. Int. Japan Natl Oil Comp,
Chiba City, Japan, 1998)
24. Milkov, A. V. & Sassen, R. Resource and economic potential of gas hydrates in the northwestern gulf
of Mexico. Proc. 4th Int.Conf. Gas Hydrates 111–114 (2002).
25. Soloviev, V., Ginsburg, G., Telepnevm, E. & Mikhailyk, Y. Cryogeothermy and Natural Gas Hydrates of
the Arctic Ocean Sediments (Ministry of Geology USSR, Leningrad, 1987).
26. Dallimore, S. R. et al. (eds). Scientific Results from the Mallik 2002 Gas Hydrate Production Research
Well Program, Mackenzie Delta, Northwest Territories, Canada. Bull. Geol. Soc. Can. 585 (in the press).
27. Dallimore, S. R., Collett, T. S. & Uchida, T. Summary of Mallik 21–38 Well GSC Bull. 544 1, 1–10 (1999).
28. Paull, C. K. & Matsumoto. R. Leg 164 overview. Proc. ODP Sci. Res. 164, 3–10 (2000).
29. Moridis, G. et al. Numerical simulation studies of gas production scenarios from hydrate
accumulations at the Mallik Site, Mackenzie Delta, Canada. Proc. 4th Int.Conf. Gas Hydrates 239-244

(2002).
30. Gudmundsson, J. & Borrehaug, A. Frozen hydrate for transport of natural gas in Proc. 2nd Int. Conf.
on Natural Gas Hydrates 415–422 (1996).
31. Nakajima, Y., Takaoki, T., Ohgaki, K. & Ota, S. Use of hydrate pellets for transportation of natural gas
– II; Proposition of natural gas transportation in form of hydrate pellets in Proc. 4th Int.Conf. Gas
Hydrates 987–990 (2002).
32. Shirota, H. et al. Measurement of methane hydrate dissociation for application to natural gas storage
and transportation. Proc. 4th Int.Conf. Gas Hydrates (2002).
33. Gudmundsson, J., Andersson, V., Levik, O. I., Mork, M. & Borrehaug, A. Hydrate technology for
capturing stranded gas. Ann. NY Acad. Science 912, 403–410.
34. Kennett, J. P., Cannariato, G., Hendy, I. L. & Behl, R. J. Methane Hydrates in Quaternary Climate
Change: The Clathrate Gun Hypothesis (Am. Geophys. Union, Washington DC, 2003).
35. Dickens, G. R., Castillo, M .M. & Walker, J. C. G. A blast of gas in the latest paleocene: simulating firstorder effects of massive dissociation of ocean methane hydrate. Geology 25, 259–262 (1997).
36. Dickens, G., O’Neil, J., Rea, D. & Owen, R. Dissociation of oceanic methane hydrates as a cause of the
carbon isotope excursion at the end of the Palaeocene. Paleoceanography 10, 965–971 (1995).
37. Kaiho, K. et al. Latest palaeocene benthic foramiferal extinction and environmental changes at
Tawanui, New Zealand. Paleoceanography 11, 447–465 (1996).
38. Maslin, M. & Thomas, E. The clathrate gun is firing blanks: Evidence from balancing the deglacial
global carbon budget. Geophys. Res. Abstr. 5, 12015 (2003).
39. Dickens, G. R. A methane trigger for rapid warming? Science 299, 1017 (2003).
40. Dillon, W. P et al. in Gas Hydrates: Relevance to World Margin Stability and Climate Change (eds
Henriet, J.-P. & Mienert, J.) 293–302 (Geol. Soc., London, 1998).
41. Paull, C. K. & Dillon, W. P. (eds) Natural Gas Hydrates: Occurrence, Distribution and Detection (Am.
Geophys. Union, Washington DC, 2001).
42. Bishoi, P. R. & Natarajan, V. Formation and decomposition of gas hydrates. Fluid Phase Equilibria
117, 168–177 (1996).
43. Sloan, E. D. Jr in Gas Hydrates: Relevance to World Margin Stability and Climate Change (eds Henriet,
J.-P. & Mienart, J.) 31–40 (Spec. Publ. 137, Geol. Soc, London, 1998).
44. Dickens, G. R. Methane oxidation during the Late Palaeocene Thermal Maximum. Bull. Soc. Geol. Fr.
171, 37–49 (2000).

359
©2003 Nature Publishing Group

© 2003 Nature Publishing Group