-
ARTICLE
Carbon dioxide-induced liberation of methane
fromlaboratory-formed methane hydratesKristine Horvat and Devinder
Mahajan
Abstract: This paper reports a laboratory mimic study that
focused on the extraction of methane (CH4) from hydrates
coupledwith sequestration of carbon dioxide (CO2) as hydrates, by
taking advantage of preferential thermodynamic stability of
hydratesof CO2 over CH4. Five hydrate formation-decomposition runs
focused on CH4CO2 exchange, two baselines and three with
hostsediments, were performed in a 200 mL high-pressure Jerguson
cell tted with two glass windows that allowed visualization ofthe
time-resolved hydrate phenomenon. The baseline pure hydrates formed
from articial seawater (75 mL) under 64006600 kPa CH4 or 28003200
kPa CO2 (hydrate forming regime), when the bath temperature was
maintained within 46 C andthe gas/liquid volumetric ratio was 1.7:1
in the water-excess systems. The data show that the induction time
for hydrateappearance was largest at 96 h with CH4, while with CO2
the time shortened by a factor of four. However, when the
secondarygas (CO2 or CH4) was injected into the system containing
preformed hydrates, the entering gas formed the hydrate
phaseinstantly (within minutes) and no lag was observed. In a
system containing host Ottawa sand (104 g) and articial seawater(38
mL), the induction period reduced to 24 h. In runs withmultiple
charges, the extent of hydrate formation reached 44% of
thetheoretical value in thewater-excess system,whereas the
valuemaximizedat 23% in thegas-excess system.TheCO2hydrate
formationin a system that already contained CH4 hydrates was facile
and they remained stable, whereas CH4 hydrate formation in a
systemconsisting of CO2 hydrates as hosts were initially stable,
but CH4 gas in hydrates quickly exchanged with free CO2 gas to
formmorestable CO2 hydrates. In all ve runs, even though the
systemwas depressurized, left for over aweek at room temperature,
andushedwith nitrogen gas in between runs, hydrates exhibited the
memory effect, irrespective of the gas used, a result in
contradictionwiththat reported previously in the literature. The
facile CH4CO2 exchange observed under temperature and pressure
conditions thatmimic naturally occurring CH4 hydrates show promise
to develop a commercial carbon sequestration system.
Key words: sediment hosted hydrates, gas exchange in hydrates,
methane hydrate, carbon dioxide hydrate, carbon sequestration.
Rsum : Cet article prsente les rsultats dune tude visant a`
reproduire en laboratoire les conditions dextraction dumthane(CH4)
a` partir dhydrates et, simultanment, la squestration du dioxyde de
carbone (CO2) sous forme dhydrates, en tirant partide la plus
grande stabilit thermodynamique des hydrates de CO2 par rapport aux
hydrates de CH4. Nous avonsmen cinq essaisde formation-dcomposition
dhydrates centrs sur lchange entre le CH4 et le CO2, dont deux
essais de rfrence et trois,dans des sdiments contenant des
hydrates. Les essais ont t raliss dans une chambre Jerguson a`
haute pression de 200 mlmunie de deux fentres de verre permettant
lobservation du phnomne de transformation des hydrates en fonction
du temps.Les hydrates purs de rfrence se sont forms a` partir de
leau de mer articielle (75 ml) sous une pression de CH4 de 6400
a`6600 kPa ou de CO2 de 2800 a` 3200 kPa (conditions de formation
dhydrates) lorsque la temprature du bain tait maintenueentre 4 C et
6 C et que le rapport volumtrique gaz/liquide tait denviron 1,7 : 1
dans les systmes o leau tait en excs. Lesdonnes rvlent que la
priode dinduction requise pour lapparition de lhydrate tait la plus
longue, soit 96 h, dans le cas duCH4, tandis quelle tait quatre
fois plus courte dans le cas du CO2. Toutefois, lorsque le gaz
secondaire (le CO2 ou le CH4) taitinject dans le systme contenant
des hydrates dja` forms, le gaz entrant sest tout de suite
transform en hydrate (en quelquesminutes); aucun dlai na t observ.
Dans un systme contenant le substrat dans du sable dOttawa (104 g)
et de leau de merarticielle (38 ml), la priode dinduction tait
rduite a` 24 h. Dans les essais a` charges multiples, lhydrate sest
form dans uneproportion atteignant 44 % de la valeur thorique en
systme o leau tait en excs, tandis que la proportion a atteint une
valeurmaximale de 23 % en systme o le gaz tait en excs. La
formation dhydrates de CO2 dans un systme qui contient dja`
deshydrates de CH4 tait aise, et les hydrates forms sont demeurs
stables. linverse, les hydrates de CH4 forms dans le systmecompos
pralablement dhydrates de CO2 taient initialement stables, mais ont
rapidement vu leur CH4 gazeux schangercontre le CO2 gazeux libre
pour former des hydrates de CO2 plus stables. Dans les cinq essais,
mme dans le cas dun systme taitdpressuris, laiss a` temprature
ambiante pendant plus dune semaine et purg avec de lazote gazeux
entre les essais, leshydrates ont manifest l' effet mmoire , quel
que soit le gaz employ, un rsultat qui est en contradiction avec
les rsultatspublis auparavant. Lchange ais entre le CH4 et le CO2
observ en conditions de pression et de temprature imitant
lesconditions naturelles de formation des hydrates de CH4 semble un
moyen prometteur de raliser un systme commercial desquestration du
carbone. [Traduit par la Rdaction]
Mots-cls : hydrates dorigine sdimentaire, changes gazeux dans
les hydrates, hydrate de mthane, hydrate de dioxyde decarbone,
squestration du carbone.
Received 7 December 2014. Accepted 15 March 2015.
K. Horvat. Materials Science and Engineering, Stony Brook
University, Stony Brook, NY 11794, USA.D. Mahajan. Sustainable
Energy Technologies Department, Brookhaven National Laboratory,
Upton, NY 11973, USA.Corresponding author: Devinder Mahajan
(e-mail: [email protected]).This article is part of a Special Issue
in honour of Dr. John Ripmeester and his outstanding contributions
to science.
Pagination not final (cite DOI) / Pagination provisoire (citer
le DOI)1
Can. J. Chem. 93: 19 (2015) dx.doi.org/10.1139/cjc-2014-0562
Published at www.nrcresearchpress.com/cjc on 15 April 2015.
Can.
J. C
hem
. Dow
nloa
ded
from
ww
w.n
rcre
sear
chpr
ess.c
om b
y SU
NY
AT
STO
NY
BRO
OK
on
06/1
4/15
For p
erso
nal u
se o
nly.
-
IntroductionWith the world population expected to reach over 10
billion by
2050,1 the world is facing an ever increasing need for new
energysources. Increasing greenhouse gas levels in the atmosphere
ne-cessitate the use of carbon neutral energy sources. Current
esti-mates show that over 30 trillionm3 ofmethane (CH4) is trapped
ashydrates in marine and permafrost environments.2 The presenceof
vast CH4 hydrate reserves have been known for decades, buttheir
extraction, especially from CH4-rich but dispersed marinehydrates,
remains a challenge. Gas production by external stimu-lation using
techniques such as steam injection, gas depressuriza-tion, or
hydrate-inhibitor injection are now well-established.3 Inmarine
systems, CH4 harvesting from hydrate may result inseaoor
instability.45 At temperatures below 10 C, carbon diox-ide (CO2)
hydrates are more thermodynamically stable at lowerpressures than
CH4 hydrates,6 and some work has been reportedon the CO2CH4
exchange. Both CH4 and CO2 gases form structuresI hydrates,7 and
the exchange is a fascinating approach by itself:CH4 is liberated
while CO2 is sequestered. However, large-scaleapplication of the
exchange concept requires a thorough under-standing of immediate
surroundings (sediments) of natural hy-drate and other
environmental concerns.3
Laboratory studies to test the feasibility of CH4 release
fromhydrates by CO2 gas injection are mostly limited to
analyticalinvestigation tools, such as NMR, Raman, and MRI, to
provide abasic understanding of the exchange process.8 Raman
spectros-copy established that upon liquid CO2 injection into
preformedCH4 hydrates, there was essentially mol to mol correlation
be-tween liberated CH4 and CO2 loss due to CO2 hydrate formation.In
one experiment, 70 mmol CH4 was recovered from
hydratedecomposition, while 71 mmol CO2 was consumed to form
hy-drates, indicating that CH4 guest molecules were replaced by
CO2in the hydrate structure during the gas exchange.6 Another
study9
measured CH4 recovery to be 8.3 mol% in 206 h using gas
chroma-tography (GC). An in situ magnetic resonance imaging (MRI)
mea-surement of a Bentheim sandstone hydrate core sample
yielded50%85% CH4 gas recovery after ushing three times with
CO2using GC.10 Another MRI study of brine solutions in
Bentheimsandstone showed that CH4 hydrate spontaneously converted
toCO2 hydrate when exposed to liquid CO2, and the hydrates werenot
found to dissociate to liquid water during the exchange.11
Since CH4 hydrates located in permafrost regions are easier
toaccess than those in the marine environment, the technology
forcarbon sequestration has reached larger-scale eld testing in
per-mafrost regions. Studies have shown that there are an
estimated2.4 trillion m3 of recoverable gas from accessible hydrate
accu-mulations in the North Slope of Alaska alone.12 In 2012,
aConocoPhilips-led test showed success when a 23 mol% CO2 inN2
mixture was injected to release over 24 210 m3 of CH4 from ahydrate
reservoir in Alaska. CO2 was preferred over N2 in thehydrate phase,
as about 70% of the N2 gas injected was recovered,while only 40% of
the 1376 m3 of CO2 injected was retrieved.13
Recently, a marine eld test was performed off the coast of
Cali-fornia to exchange CO2 in CH4 hydrate. In the experiment,
pureCH4 hydrate was brought to the seaoor from a depth of 690 mand
then enclosed in a cylinder lled with a 25% CO2 75% N2 gasmixture.
The CH4 hydrate was found to dissociate to form amixedgas phase,
though no CO2 hydrate was found to form under thesesimulated
conditions.14 Thus, the CH4CO2 exchange reactionmerits further
research to develop a feasible method to mine CH4from natural
hydrates. In this paper, we describe several laboratory-scale
experiments to understand the CH4CO2 hydrate exchangekinetics.
Materials and methods
GasesPressurized cylinders of CH4 and CO2 gases were ordered
from
Scott Specialty Gases and Praxair, respectively.
Articial seawaterArticial seawater used to simulate natural
medium was pre-
pared according to Kester et al.15
Hydrate former unit descriptionThese experiments were performed
in a 200 mL Jerguson cell
tted with two 12 inch long borosilicate windows in the
FlexibleIntegrated Study of Hydrates (FISH) unit at Brookhaven
NationalLaboratory (BNL) (Fig. 1) and described in detail
elsewhere.16 Thepresence of a pressure transducer at the top of the
cell and twothermocouples, located in the gas and liquid phases
inside thecell, allowed for accurate pressuretemperature monitoring
dur-ing runs. The cell capabilities allowed us to: (i) measure the
per-centage of hydrates formed in the system, (ii) visualize
hydrateformation and observe their morphology, and (iii) measure
com-position of free gas above the liquid phase to quantify the
mixedCO2CH4 system. We completed ve runs, as noted in Table 1,
inthe Jerguson cell without stirring the cell. The rst two
experi-ments used only 75 mL of articial seawater while the last
threeused 38 mL articial seawater and 104 g of 110 m Ottawa
sand,17
which was 99.0%99.9% silicon dioxide.18 The cell was
pressurizedwith gas owing through the bottom such that that the
enteringgas bubbled through the articial seawater. The cell was
initiallycharged with either CO2 or CH4 gas, and after the hydrates
startedto form, the second gas was injected into the cell. For
example, byinitially forming CH4 hydrates, and then injecting CO2
gas, theCH4CO2 exchange in hydrates could be monitored. In
experi-ments with preformed CH4 hydrate, the pressure in the cell
wasreduced by depressurizing the cell below 4000 kPa prior to
CO2injection to insure that CO2 would be injected as a gas rather
thana liquid. The phase diagram for CO2 and CH4 hydrates in
articialseawater (Fig. 2) served as a guide to follow
theoreticalmonitoringof the experimental system. We also conducted
two runs thatinvolved the formation of CO2 hydrate followed by a
charge withCH4 gas to allow monitoring the stability of CO2
hydrates if theywere to come into contact with CH4 gas, as could
potentially occurin a free CH4 gas zone or if CO2 hydrates formed
near a CH4 plume.During the depressurization cycle, the system
pressure was low-ered in multisteps, where in each step the
pressure was reducedseveral hundred kPa and the system was then
allowed to stabilizefor several minutes to an hour prior to
subsequent pressure re-duction. This process was repeated until the
system was totallydepressurized.
Modes of unit operationThe knownmole stoichiometry of water/gas
in CH4 hydrate and
CO2 hydrate are 5.75:1 and 6:1 respectively. The unit was
operatedin either water-excess or gas-excess mode, depending on
themoleratio of water/gas added to the system. These data were used
tocalculate theoretical maximum hydrate from which the extent
ofhydrate saturation was calculated by the gas consumed during
agiven run.
Gas analysisGC samples were taken periodically during runs and
analyzed
for CH4 and CO2 on a Gow Mac series 580 gas chromatographtted
with a Supelco Carbonex 1000 45/60 (1.5 m 0.32 mm)packed column
using helium as carrier gas.
Pagination not final (cite DOI) / Pagination provisoire (citer
le DOI)2 Can. J. Chem. Vol. 93, 2015
Published by NRC Research Press
Can.
J. C
hem
. Dow
nloa
ded
from
ww
w.n
rcre
sear
chpr
ess.c
om b
y SU
NY
AT
STO
NY
BRO
OK
on
06/1
4/15
For p
erso
nal u
se o
nly.
-
Results and discussion
Run 1: Baseline run with no host, CH4 followed byCO2
chargeFreshly prepared articial seawater (75 mL) was syringed
into
the Jerguson cell, and the cell was pressurized to 7065 kPa
withCH4 gas at room temperature and then cooled till the bath
tem-perature was about 4 C. After several days, an observed
pressuredrop of over 2000 kPa was attributed solely to hydrate
formation.
After repressurization to 6114 kPa with CH4, the pressure
droppedto 5252 kPa. Together, total gas consumed corresponded to
44% ofthe theoretical hydrate value in this excess-water system
after twogas charges. At this point, the cell was switched to the
CH4CO2exchange mode by depressurization to 2866 kPa CH4 and
thenpressurized with CO2 gas three times until the cell pressure
in-creased to 4652 kPa, resulting in instant gas hydrate
formation.This was necessitated by the fact that at partial
pressures above4000 kPa, CO2 is in liquid phase. GC samples taken
over 16 hshowed >99% CH4 (
-
were seen in the cell, and the cell was charged with
additionalCO2 gas. Upon repressurization to 3997 kPa at 5.2 C,
liquid CO2wasobserved above the aqueous layer in the cell. After
two hours,transparent needle-like CO2 hydrateswere observed at the
bottomof the cell, and after an additional hour CO2 hydrates were
seen atthe aqueousliquid CO2 meniscus. These hydrates appeared
morelike solid ice, unlike the needle-like hydrates still at the
bottom ofthe reactor. One hour later, the gaswas discharged from
the cell toreduce cell pressure to a point where liquid CO2 was no
longerstable, and many of the hydrates present dissociated. The
cell wasagain repressurized to 3273 kPa with CO2 gas that resulted
ininstant hydrate formation during charging. No further
pressuredrop was noted after the cell was sealed. Shortly
thereafter, thecell was operated in the CO2CH4 exchange mode by rst
depres-surization to 1301 kPa and repressurized to 6265 kPa with
CH4,wherein instantaneous CH4 hydrate formation was observed
dur-ing CH4 charging. Figure 4 shows recorded images of mixed
CO2CH4 hydrates, and the hydrates that formed instantly after
theCH4 injection are clearly seen as small spheres beneath the
gasliquid interface. Though the physical appearance of hydrates
ofCH4 and CO2 is indistinguishable, the time resolved analysis
offree gas above the aqueous phase by GC was used to
quantifyhydrates in themixed gas system. Gas samples taken from the
cellone hour after the CH4 charge established the CH4/CO2 ratio to
be35%:65% that changed to 50%:50% over the next 26 h. At this
point,a stepwise depressurization of the system was initiated. The
GCanalysis indicated that initially CO2 dominated the gas phase,
butas the cell slowly depressurized, the amount of CH4 in the
gasphase increased. The stability of CH4 hydrates was noted,
evenwhen the partial pressure of CH4 was below the hydrate
equilib-rium curve (Fig. 2), indicating that complete CH4 hydrate
decom-position may be a slow phenomenon.
Run 3: CH4 followed by CO2 charge hosted in Ottawa sandRun 3 was
conducted with Ottawa sand (104 g) as the host that
was fully saturatedwith articial seawater (38mL) and to ll
spaceabove the sand pack. After cooling to 4 C, the cell was
pressurizedwith 6617 kPa CH4 gas, when hydrates were observed at
the gasliquid interface in about 24 h. Over time, the hydrate grew
down-wards into the solution until they reached the top of the sand
pack(Fig. 5). The pressure drop corresponded to 11% conversion in
48 hin this excess-gas system. At this point, the cell was
depressurizedbelow the hydrate equilibrium conditions to allow
complete hy-drate dissociation. The cell was repressurized to 6596
kPa andwithin 10min hydrates were observed as a thin lm on part of
theglass window and at the gasliquid interface. The pressure
de-creased continuously as hydrates grew from the gasliquid
inter-face downwards into the solution, until the pressure
stabilized at5810 kPa, which corresponded to 12% gas conversion
into hydratesover 2 days. The observed relatively fast hydrate
formation is at-tributed to the memory effect.21
Similar phenomenon of instantaneous hydrate formation
wasobserved during subsequent CO2 charges in the partial
depressur-ization/repressurization cycle. It is notable that in
these runs, thehydrate formed above the sand pack; few if any,
hydrates formedin the sand pack most likely due to capillary
inhibition.22 GCanalysis indicated pure CH4 for several hours but 3
days after theCO2 injection, the gas phase CH4/CO2 ratio was
86.5%:13.5%. Sub-sequently, hydrate dissociation was induced by
depressurization.A nal GC samplewas taken at 1287 kPa pressure and
3.7 C (belowthe hydrate stability zone) that established the
gasphase CH4/CO2ratio of 15%:85%, suggesting that the cell was
mostly composed ofCO2 hydrates prior to dissociation.
Fig. 3. Image of mixed CH4CO2 hydrates captured during Run 1.The
images were taken at 312 h when the cell pressure was 4135 kPaat
3.2 C. The corresponding gas phase composition in the cell was99%
CH4:1% CO2. Note that that mixed gas hydrates have lled theentire
cell viewing region.
Fig. 4. Image taken at 52.8 h, at cell pressure of 6038 kPa at 4
C.The corresponding gas phase composition was 41% CH4:59% CO2.
Gashydrates can be clearly seen. The spherical hydrates in the
lowerportion of the image formed instantly when CH4 was added to
thecell.
Pagination not final (cite DOI) / Pagination provisoire (citer
le DOI)4 Can. J. Chem. Vol. 93, 2015
Published by NRC Research Press
Can.
J. C
hem
. Dow
nloa
ded
from
ww
w.n
rcre
sear
chpr
ess.c
om b
y SU
NY
AT
STO
NY
BRO
OK
on
06/1
4/15
For p
erso
nal u
se o
nly.
-
Run 4: CO2 followed by CH4 charge hosted in Ottawa sandIn this
run, the Ottawa sand/articial seawater from previous
run were reused after ushing with N2 gas, except that the
addi-tion of the two gases was reversed to establish the effect of
addedCH4 gas to preformedCO2 hydrates. After pre-cooling to 6 C,
the cellwas pressurized to 3211 kPawith CO2, when transparent,
needle-likeCO2 hydrates were observed on top of the sand pack in
less than 7 h.After 24h, thehydrates had spread to the gasliquid
interface, and inthe next 24 h hydrates were visible within the
host sand. The cellpressure stabilized and corresponded to 11%
conversion of CO2 intohydrates in this excess-water system. Another
CO2 charge resulted ininstantaneous CO2 hydrate formation. Shortly
thereafter, CH4 wasinjected to increase the cell pressure to 5314
kPa at 3.1 C to observethe CO2CH4 exchange. The GC analysis
established the CH4/CO2ratio as follows: 6.4%:93.6% (5min);
30.4%:69.6% (36min). These data
are consistent with the liberation of CH4 from the hydrate
phaseshortly after CO2 was introduced in to the system and
displaced CH4to form CO2 hydrates. A second charge of CH4 increased
the cellpressure to 6134 kPa and the measured CH4/CO2 ratio was
34.9%:65.1% (5min) that increased to 47.9%:52.1% over 3 h. It is
known that,initially, CH4 molecules ll both the small and large
hydrate cages,but as time elapses themajority of large hydrate
cages are lledwithCO2 that results in increased CH4 concentration
in the gas phase.23 Astep-wise depressurization of the cell to
bring about hydrate dissoci-ation initially resulted in an increase
of CH4 in the gas phase, but asthe pressure decreased the gas phase
became richer in CO2. Whenthepressurewas reduced to2859kPa,
thegasphase compositionwas53.7% CH4:46.3% CO2. However, as the
pressure further decreased to812 kPa, the gas phase became richer
in CO2 (65.7%), as the CO2hydrates began to dissociate. Figure 6 is
an image of mixed gas hy-drates seen above and in the sand
pack.
Fig. 5. Time-resolved images of hydrate formation during Run 3.
(1) Taken 20.1 h after the pre-cooled cell was pressurized with
6382 kPa CH4at 3.5 C. Prior to (1), the cell was charged with 6617
kPa with CH4 gas. (2) At 194.8 h under 2728 kPa at 3.6 C. Prior to
(2), after CO2 hydratesformed for several days, the cell was
partially depressurized, left cooling for 3 days, and then
partially depressurized and repressurized severaltimes. (3) At 236
h under 2542 kPa at 3.6 C. Prior to (3), the CO2 gas was charged
twice. Shortly after (3), the cell was depressurized in steps
tobring about complete dissociation of hydrates.
Pagination not final (cite DOI) / Pagination provisoire (citer
le DOI)Horvat and Mahajan 5
Published by NRC Research Press
Can.
J. C
hem
. Dow
nloa
ded
from
ww
w.n
rcre
sear
chpr
ess.c
om b
y SU
NY
AT
STO
NY
BRO
OK
on
06/1
4/15
For p
erso
nal u
se o
nly.
-
Run 5: CH4 followed by CO2 charge hosted in Ottawa sandIn Run 5,
the Ottawa sand and articial seawater from Run 4
were re-used to establish the exchange of CO2 gas with
preformedhydrates of CH4. The cell was ushed with N2 gas,
pre-cooled to3 C, and then pressurized to 6651 kPa with CH4 gas.
Within min-utes, a thin coating of ice-like hydrates was seen on
the cell glasswindows above the gasliquid interface that spread to
the gasliquid interface after 1 h and then moved downwards into
thesolution. The noted pressure drop corresponded to 6%
hydratesaturation in this excess-gas system. After 26 h, the second
chargewith CH4 corresponded to an additional 2% gas conversion of
CH4into hydrates. The cell was partially depressurized, then
repres-surized with CO2 up to 4149 kPa, and then 1 h later, the
cell wasagain partially depressurized and charged with CO2.
Multiple GCsamples over 3 h established the presence of a pure CH4
gas phase.Upon quick depressurization to 1791 kPa (below the
hydrate sta-bility zone), GC analysis established the gas phase
rich in CO2(69.1%), indicating the presence of CO2 trapped as
hydrate becameunstable during depressurization. Further
depressurization to405 kPa increased CO2 (81.8%) and decreased CH4
(18.2%) in the gasphase. Figure 7 shows images of the hydrate
growth throughoutrun 5.
Comparison of data from runsThis study consisted of 5
experimental runs to understand a
CH4CO2 system in which gas phase CO2 is pumped into a
natural
CH4 hydrate reservoir that results in sequestered CO2 as
hydratewith concomitant liberation of CH4. Specically, the
completed5 runs focused on understanding the extent and rates of
the CH4CO2 exchange phenomenon. The exact conditions of all
experi-ments are listed in Tables 1 and 2. Runs 1 and 2 were
baseline(without host sediments), while runs 35 included Ottawa
sand asa host. In runs 1, 3, and 5, the cell was initially charged
with CH4,and once CH4 hydrate formed, CO2 was added to the cell.
Theopposite sequence was repeated in runs 2 and 4, wherein CO2
gaswas initially added to the cell, and once CO2 hydrates were
ob-served, CH4 was added to the system.We extracted induction time
data for hydrate formation (the
time when these could be seen by the naked eye) from
completedruns to better understand its dependence on the nature of
hydrateforming gas (CH4 and CO2), host sediment, and memory
effect.The data in Table 2 show that the induction time for CH4
hydrateappearance was the largest at 96 h (Run 1), while with CO2
thetime shortened to 24 h (Run 2). Runs 1, 3, and 5 followed the
samesequence; initially, CH4 gas was added to the system, and
oncehydrates were visible in the reactor, CO2 gas was added. CH4
hy-drates formedmuchmore quickly (24 h) in the presence of
Ottawasand (Run 3) than without it (4 days for Run 1). Similarly,
Run 4, inwhich CO2 hydrates were formed and then CH4 gas was
injectedwithin the host sand, had a shorter hydrate induction time
(7 h)than Run 2 (24 h), in which no porous media was present.
The
Fig. 6. Images of hydrate formation taken at 30.4 h under 6010
kPa at 3.1 C. The corresponding gas phase composition was 48%
CH4:52% CO2.Mixed hydrates lled much of the reactor above the same
pack though few hydrates were seen in the sand pack.
Pagination not final (cite DOI) / Pagination provisoire (citer
le DOI)6 Can. J. Chem. Vol. 93, 2015
Published by NRC Research Press
Can.
J. C
hem
. Dow
nloa
ded
from
ww
w.n
rcre
sear
chpr
ess.c
om b
y SU
NY
AT
STO
NY
BRO
OK
on
06/1
4/15
For p
erso
nal u
se o
nly.
-
presence of porous media is known to affect the stability of
gashydrates. Higher pressures and (or) lower temperatures areneeded
to form hydrates in sand systems.22 In one instance,Handa and
Stupin24 found that systems containing CH4 or pro-
pane hydrates in silica gel required equilibrium pressures
20%100% higher than systems without sediments to form hydrates.The
lack of hydrate formationwithin theOttawa sand pack hereinis in
agreement with these studies. The shortened induction time
Fig. 7. Time-resolved images of hydrate formation during Run 5.
(1) Taken 25 h after the pre-cooled cell was initially pressurized
with6382 kPa CH4 at 3.5 C. Prior to (1), the reactor was charged
with 6651 kPa with CH4 gas. After CH4 hydrates formed in 26 h, the
cell wasrepressurized with CH4 gas to 6686 kPa. (2) At 46.7 h under
6548 kPa of CH4 at 2.9 C. After this image was taken, the cell was
partiallydepressurized and repressurized with CO2 twice. (3). Taken
at 52.2 h under 3797 kPa at 3.0 C. Shortly after (3), the cell was
depressurizedquickly to induce hydrate dissociation.
Pagination not final (cite DOI) / Pagination provisoire (citer
le DOI)Horvat and Mahajan 7
Published by NRC Research Press
Can.
J. C
hem
. Dow
nloa
ded
from
ww
w.n
rcre
sear
chpr
ess.c
om b
y SU
NY
AT
STO
NY
BRO
OK
on
06/1
4/15
For p
erso
nal u
se o
nly.
-
observed in runs with Ottawa sand are likely due to sand
particlesacting as the nucleation sites for hydrate formation. It
is likelythat sand particles became stuck to the walls near the
gasliquidinterface in the cell during gas bubbling through the sand
packduring charges, and these impurities affected hydrate
nucleationand induction times.25
A comparison of pure hydrates of CH4 and CO2 in studies
foundthat CO2 hydrates had shorter induction times than CH4
hy-drates.2627 The gas hydrate phase diagram (Fig. 1) shows that
inarticial seawater, CO2 hydrates are stable at higher
temperaturesand lower pressures than CH4 hydrates. In agreement
with previ-ous studies2627 and thermodynamics, the induction times
inRuns 2 and 4, wherein CO2 hydrates were initially formed,
wereshorter than the induction times for CH4 hydrates (Runs 1 and
3).Interestingly, the shortest induction time was noted in Run
5,
wherein CH4 hydrates formed nearly instantly after the cell
waspressurized. This speedy formation is likely due to the
hydratememory effect. Studies21,28 have shown that hydrate-forming
so-lutions canmaintain a memory of their hydrate structure
whenwarmed up slightly above the hydrate stability region.29 There
aretwo theorieswhy this occurs: (i) a hydrate frame assembly
remainsintact in the solution, either as a partial or ordered
congurationonce hydrates had formed and (ii) some gas is left
dissolved in thesolution after dissociation.29 For Runs 35 in which
the sameOttawa sand and articial seawater were utilized for hydrate
for-mation, the induction timewas signicantly reduced even
thoughdifferent gases (CH4 or CO2) were initially used to form
hydrates.In Run 3, CH4 hydrates formed after about 24 h, while in
Run 4,CO2 hydrates formed in 7 h, and in Run 5, CH4 hydrates
formedwithin minutes. In between each of these runs, the cell was
com-pletely depressurized and ushed with nitrogen gas. Mazloumet
al.30 found that the depressurization of a system of natural
gashydrates, followed by repressurization with fresh natural gas,
didnot result in the hydratememory effect for gas hydrate
formation.It was found that the hydrate memory effect was destroyed
bydepressurization to atmospheric pressure, but reducing the
pres-sure of the system even slightly would still result in the
hydratememory effect.30 In addition, if the solution is warmed
above 2521
or 28C29 or for too long a time period (several hours29),
thisretention effect will not occur. In between each of these
experi-ments the cell was left uncooled for extended periods of
timeranging from 7 days to 2 months. All of these factors are
contraryto the decreased hydrate formation time observed hereinwith
thesystems composed of CH4 and CO2.The sequence of gas addition to
the cell affected hydrate stabil-
ity. In all experiments, once gas hydrate had formed, any
addi-tional gas charge, independent of the gas (CO2 or CH4),
resulted innear instant hydrate formation. This fast formation
indicates thatif CO2 is to be sequestered in a CH4 hydrate
reservoir, it is likelythat CO2 hydrates will form instantly. In
addition, whether CO2 orCH4 was the initial gas for hydrate
formation, when a second gaswas charged into the reactor, initially
the gas phase of the cell wasmostly, if not entirely, composed of
the initial system gas, whilethe newly injected gas entered the
hydrate phase. For systems
where the cell was initially lled with CH4 and then charged
withCO2 gas, the gas phase of the cell composed of pure CH4 for
severalhours after the CO2 injection.Whereas for systems originally
pres-surized with CO2 and then injected with CH4, the gas phase
wasinitially mostly composed of CO2, but over a much shorter
timeperiod, the percentage of CH4 in the gas phase increased,
indicat-ing that CO2 gas exchanged with CH4 gas in the hydrate
structure.Uchida et al.23 reported similar results with the
following expla-nation: during initial stages of the hydrate
formation, CH4 mole-cules are able to occupy both small and large
hydrate cages, butover time the larger cages mostly trap CO2 gas,
resulting in anincreased CH4 in the gas phase.23 The phase eld
simulationmod-els have yielded similar results.31 Overall, whether
CH4 or CO2 gasis introduced into a system of gas hydrates via
bubbling, an in-stant formation of hydrates of injected gas is
likely to occur.
ConclusionsThis study examined preliminary characteristics of
the CH4
CO2 exchange phenomenon. Several noteworthy observations areas
follows: (i) the induction times of hydrate formation were
gen-erally shorter for CO2 than CH4, by as much as by a factor of
four,with or without sediments; (ii) the hydrate formation data
showthat when a secondary gas was injected into a system
containingpreformed hydrates, the entering gas formed the hydrate
phaseinstantly (within minutes); (iii) CO2 hydrates formed in a
systemthat already contained CH4 hydrates were found to be
morestable, whereas CH4 hydrates formed in a system consisting
ofCO2 hydrates as hosts were initially stable, but CH4 gas in
hydratesquickly exchanged with free CO2 gas to form more stableCO2
hydrates; (iv) the hydratememory effect was noted during the3 runs
performed to form sand-hosted gas hydrates. In all 5 runs,even
though the system was depressurized, left over a week atroom
temperature, and ushed with nitrogen gas in betweenruns, the system
still exhibited the memory effect. These resultscontradict those
previously reported in the literature. In sum-mary, the observed
fast CO2 hydrate formation from free CO2 gasin the presence of
preformed CH4 hydrate indicate the feasibilityof developing a CO2
sequestration scheme using natural CH4 hy-drate reservoirs. Our
ongoing and planned work includes runsthat will quantify the extent
of uptake and liberation of gasesthrough the CH4CO2 exchange as a
function of key hydrate pa-rameters and at the micrometer scale
using X-ray computed mi-crotomography.
AcknowledgementsThe authors thank the Ofce of Vice-President of
Research
(OVPR) at Stony BrookUniversity for providing funds for
thework.The work was partially supported by the Program
Developmentfunds at Brookhaven National Laboratory.
References(1) Sulston, J.; Rumsby, M.; Green, N. Environ.
Resour. Econ. 2013, 55, 469. doi:10.
1007/s10640-013-9681-8.(2) Boswell, R.; Collett, T. S. Energy
Environ. Sci. 2011, 4 (4), 1206. doi:10.1039/
C0EE00203H.(3) Collett, T.; Bahk, J.-J.; Baker, R.; Boswell, R.;
Divins, D.; Frye,M.; Goldberg, D.;
Huseb, J.; Koh, C.; Malone, M.; Morell, M.; Myers, G.; Shipp,
C.; Torres, M.J. Chem. Eng. Data 2015, 60 (2), 319.
doi:10.1021/je500604h.
(4) Collett, T.; Bahk, J.-J.; Frye, M.; Goldberg, D.; Husebo,
J.; Koh, C.; Malone, M.;Shipp, C.; Torres, M.; Myers, G.; Divins,
D.; Morell, M. 2013; p Medium: ED.
(5) Priest, J. A.; Clayton, C. R. I.; Rees, E. V. L. Mar.
Petrol. Geol. 2014, 58,
187.doi:10.1016/j.marpetgeo.2014.05.008.
(6) Ota, M.; Morohashi, K.; Abe, Y.; Watanabe, M.; Smith, R. L.;
Inomata, H.Energy Convers. Manage. 2005, 46 (1112), 1680.
(7) Ohgaki, K.; Takano, K.; Moritoki, M. Kagaku Kogaku Ronbunshu
1994, 20 (1),121. doi:10.1252/kakoronbunshu.20.121.
(8) Komatsu, H.; Ota, M.; Smith, R. L., Jr.; Inomata, H. J.
Taiwan Inst. Chem. Eng.2013, 44 (4), 517.
doi:10.1016/j.jtice.2013.03.010.
(9) Jadhawar, P.; Mohammadi, A. H.; Yang, J.; Tohidi, B.
Subsurface carbon dioxidestorage through clathrate hydrate
formation; In Advances in the Geological Storage
Table 2. Induction timemeasuredwhen hydrates were rst sighted
inthe Jerguson cell.
Induction time (h)
Run #Hostsediments
Gas additionsequence Charge 1 Charge 2
1 No CH4/CO2 96 Instant2 No CO2/CH4 22a 23 Ottawa sand CH4/CO2
24
-
of Carbon Dioxide. Nato Science Series: IV: Earth and
Environmental Sciences, 2006;Vol. 65, pp. 111126.
(10) Graue, A.; Kvamme, B.; Baldwin, B. A.; Stevens, J.; Howard,
J.; Aspenes, E.;Ersland, G.; Husebo, J.; Zornes, D. SPE J. 2008,
146.
(11) Baldwin, B. A.; Stevens, J.; Howard, J. J.; Graue, A.;
Kvamme, B.; Aspenes, E.;Ersland, G.; Husebo, J.; Zornes, D. R.
Magn. Reson. Imaging 2009, 27,
720.doi:10.1016/j.mri.2008.11.011.
(12) White, R.; Kumagai, T. Platts Gas Daily 2011, 28, 206.(13)
Schoderbek, D.; Farrell, H.; Hester, K.; Howard, J.; Raterman,
K.;
Silpngarmlert, S.; Martin, K. L.; Smith, B.; Klein, P.
ConocoPhilips Company,2013.
(14) Brewer, P. G.; Peltzer, E. T.; Walz, P. M.; Coward, E. K.;
Stern, L. A.;Kirby, S. H.; Pinkston, J. Energy Fuels 2014, 28 (11),
7061. doi:10.1021/ef501430h.
(15) Kester, D. R.; Duedall, I. W.; Connors, D. N.; Pytkowic, R.
M. Limnol. Oceanogr.1967, 12 (1), 176.
doi:10.4319/lo.1967.12.1.0176.
(16) Horvat, K.; Kerkar, P.; Jones, K.; Mahajan, D. Energies
2012, 5, 2248. doi:10.3390/en5072248.
(17) Santamarina, J. C.; Ruppel, C. In Proceedings of the 6th
International ConferenceonGasHydrates (ICGH 2008), Vancouver,
British Columbia, Canada, July 610, 2008;Vancouver, British
Columbia, Canada, 2008.
(18) U.S. Silica Company. Material Safety Data Sheet Silica Sand
and Ground Silica.(19) Colorado School of Mines. Software;
Available from http://hydrates.mines.
edu/CHR/Software.html.
(20) Bishnoi, P. R.; Natarajan, V. Fluid Phase Equilib. 1996,
117 (12), 168. doi:10.1016/0378-3812(95)02950-8.
(21) Wu, Q.; Zhang, B. J. Nat. Gas Chem. 2010, 19 (4), 446.
doi:10.1016/S1003-9953(09)60086-4.
(22) Clennell, M. B.; Hovland, M.; Booth, J. S.; Henry, P.;
Winters, W. J. J. Geophys.Res.: Solid Earth 1999, 104 (B10), 22985.
doi:10.1029/1999JB900175.
(23) Uchida, T.; Ikeda, I. Y.; Takeya, S.; Kamata, Y.; Ohmura,
R.; Nagao, J.;Zatsepina, O. Y.; Buffett, B. A. ChemPhysChem 2005, 6
(4), 646. doi:10.1002/cphc.200400364.
(24) Handa, Y. P.; Stupin, D. Y. J. Phys. Chem. 1992, 96 (21),
8599. doi:10.1021/j100200a071.
(25) Jensen, L.; Thomsen, K.; von Solms, N. Chem. Eng. Sci.
2008, 63 (12), 3069.doi:10.1016/j.ces.2008.03.006.
(26) He, Y.; Rudolph, E. S. J.; Zitha, P. L. J.; Golombok, M.
Fuel 2011, 90 (1), 272.doi:10.1016/j.fuel.2010.09.032.
(27) Golombok, M.; Ineke, E.; Luzardo, J. C. R.; He, Y. Y.;
Zitha, P. Environ. Chem.Lett. 2009, 7 (4), 325.
doi:10.1007/s10311-008-0173-y.
(28) Zhang, B. Y.; Wu, Q.; Gao, X.; Liu, C. L.; Ye, Y. G. Int.
J. Environ. Pollut. 2013, 53(34), 201.
doi:10.1504/IJEP.2013.059913.
(29) Sloan, E. D.; Koh, C. A. Clathrate hydrates of natural
gases, 3rd ed.; CRC Press:Boca Raton, FL, 2008.
(30) Mazloum, S.; Yang, J.; Chapoy, A.; Tohidi, B. In
Proceedings of the 7th InternationalConference on Gas Hydrates,
Edinburgh, Scothland, United Kingdom, Edinburgh,Scothland, United
Kingdom, 2011.
(31) Qasim, M.; Kvamme, B.; Baig, K. Int. J. Geol. 2011, 5 (2),
48.
Pagination not final (cite DOI) / Pagination provisoire (citer
le DOI)Horvat and Mahajan 9
Published by NRC Research Press
Can.
J. C
hem
. Dow
nloa
ded
from
ww
w.n
rcre
sear
chpr
ess.c
om b
y SU
NY
AT
STO
NY
BRO
OK
on
06/1
4/15
For p
erso
nal u
se o
nly.
ArticleIntroductionMaterials and methodsGasesArtificial
seawaterHydrate former unit descriptionModes of unit operationGas
analysis
Results and discussionRun 1: Baseline run with no host, CH4
followed by CO2 chargeRun 2: Baseline run with no host, CO2
followed by CH4 chargeRun 3: CH4 followed by CO2 charge hosted in
Ottawa sandRun 4: CO2 followed by CH4 charge hosted in Ottawa
sandRun 5: CH4 followed by CO2 charge hosted in Ottawa
sandComparison of data from runs
Conclusions
AcknowledgementsReferences
/GrayImageMinResolutionPolicy /OK/ConvertImagesToIndexed
true/MaxSubsetPct 99/Binding /Left/PreserveDICMYKValues
false/GrayImageMinDownsampleDepth 2/MonoImageMinResolution
1200/sRGBProfile (sRGB IEC61966-2.1)/AntiAliasColorImages
false/GrayImageDepth -1/PreserveFlatness true/CompressPages
true/GrayImageMinResolution 150/CalCMYKProfile (U.S. Web Coated
\050SWOP\051 v2)/PDFXBleedBoxToTrimBoxOffset
[0.00.00.00.0]/AutoFilterGrayImages true/EncodeColorImages
true/AlwaysEmbed []/EndPage -1/DownsampleColorImages
true/ASCII85EncodePages false/PreserveEPSInfo
false/PDFXTrimBoxToMediaBoxOffset [0.00.00.00.0]/CompatibilityLevel
1.3/MonoImageResolution 600/NeverEmbed
[/Arial-Black/Arial-BlackItalic/Arial-BoldItalicMT/Arial-BoldMT/Arial-ItalicMT/ArialMT/ArialNarrow/ArialNarrow-Bold/ArialNarrow-BoldItalic/ArialNarrow-Italic/ArialUnicodeMS/CenturyGothic/CenturyGothic-Bold/CenturyGothic-BoldItalic/CenturyGothic-Italic/CourierNewPS-BoldItalicMT/CourierNewPS-BoldMT/CourierNewPS-ItalicMT/CourierNewPSMT/Georgia/Georgia-Bold/Georgia-BoldItalic/Georgia-Italic/Impact/LucidaConsole/Tahoma/Tahoma-Bold/TimesNewRomanMT-ExtraBold/TimesNewRomanPS-BoldItalicMT/TimesNewRomanPS-BoldMT/TimesNewRomanPS-ItalicMT/TimesNewRomanPSMT/Trebuchet-BoldItalic/TrebuchetMS/TrebuchetMS-Bold/TrebuchetMS-Italic/Verdana/Verdana-Bold/Verdana-BoldItalic/Verdana-Italic]/CannotEmbedFontPolicy
/Warning/PreserveOPIComments false/AutoPositionEPSFiles
true/JPEG2000GrayACSImageDict >/PDFXOutputIntentProfile
()/EmbedJobOptions true/JPEG2000ColorACSImageDict
>/MonoImageDownsampleType /Average/DetectBlends
true/EmitDSCWarnings false/ColorImageDownsampleType
/Average/EncodeGrayImages true/AutoFilterColorImages
true/DownsampleGrayImages true/GrayImageDict
>/AntiAliasMonoImages false/GrayImageAutoFilterStrategy
/JPEG/GrayACSImageDict >/ColorImageAutoFilterStrategy
/JPEG/ColorImageMinResolutionPolicy /OK/ColorImageResolution
300/PDFXRegistryName ()/MonoImageFilter
/CCITTFaxEncode/CalGrayProfile (Gray Gamma
2.2)/ColorImageMinDownsampleDepth 1/PDFXTrapped /False/DetectCurves
0.1/ColorImageDepth -1/JPEG2000GrayImageDict
>/TransferFunctionInfo /Preserve/ColorImageFilter
/DCTEncode/PDFX3Check false/ParseICCProfilesInComments
true/DSCReportingLevel 0/ColorACSImageDict
>/PDFXOutputConditionIdentifier ()/PDFXCompliantPDFOnly
false/AllowTransparency false/UsePrologue false/PreserveCopyPage
true/StartPage 1/MonoImageDownsampleThreshold
1.0/GrayImageDownsampleThreshold 1.0/CheckCompliance
[/None]/CreateJDFFile false/PDFXSetBleedBoxToMediaBox
true/EmbedOpenType false/OPM 0/PreserveOverprintSettings
false/UCRandBGInfo /Remove/ColorImageDownsampleThreshold
1.0/MonoImageDict >/GrayImageDownsampleType /Average/Description
>/CropMonoImages true/DefaultRenderingIntent
/RelativeColorimeteric/PreserveHalftoneInfo false/ColorImageDict
>/CropGrayImages true/PDFXOutputCondition ()/SubsetFonts
true/EncodeMonoImages true/CropColorImages true/PDFXNoTrimBoxError
true>>setdistillerparams>setpagedevice