Chemical Engineering Science 141 (2016) 125–132

Contents lists available at ScienceDirect

Chemical Engineering Science

journal homepage: www.elsevier.com/locate/ces

Formation of cyclopentane methane binary clathrate hydrate

in brine solutions
Lichao Cai, Brian A. Pethica, Pablo G. Debenedetti, Sankaran Sundaresan n
Princeton University, Department of Chemical and Biological Engineering, NJ 08544, USA

H I G H L I G H T S

Formation rate of cyclopentane–methane hydrate is measured in brine solutions.

Methane occupancies are estimated at different operating conditions.

Evaluated the hydrate growth rate reduction caused by salt.

art ic l e i nf o a b s t r a c t

Article history: Building on a our previous study on the kinetics of formation of cyclopentane (CP)-methane binary sII-
Received 2 July 2015 type hydrate in well-agitated fresh water, analogous experiments have been performed in brine solutions
Received in revised form at two salinities (3.5% and 7.0% NaCl, w/w) to investigate how salt affects this formation process ther-
17 September 2015
modynamically and kinetically. The hydrate formation rate in salt water is found to be  2–3 times
Accepted 1 November 2015
smaller than that in fresh water at small subcooling levels ΔT o 4 K but at higher subcooling levels, the
Available online 7 November 2015
two become comparable. The methane occupancy in the small cavities of the sII hydrate is also estimated
Keywords: to be 0.5 for methane pressures between 1 and 3 MPa under equilibrium conditions, as well as during
Binary clathrate hydrate growth at subcooled conditions ðΔT o  7 KÞ.
Growth kinetics
& 2015 Elsevier Ltd. All rights reserved.
Mass transfer
Brine solution
Cyclopentane methane hydrate

1. Introduction
A wide range of guest species methane, ethane, carbon dioxide
etc. form clathrate hydrates of various structures (sI, sII, and sH)
with water (Sloan and Koh, 2007). Clathrate hydrates feature cages
of various sizes formed by water molecules connected via hydro-
gen bonding. These cages are partially occupied by guest mole-
cules that stabilize the structure, allowing hydrates to remain
thermodynamically stable (often at elevated pressures) over a
range of temperatures well above the ice melting point. Hydrates
in oil pipelines can lead to flow assurance problems that motivated
early extensive study to better understand and suppress their
formation (Sloan and Koh, 2007; Sloan et al., 2011). More recently,
clathrate hydrates have been proposed in several applications,
including natural gas storage/transportation (Gudmundsson et al.,
1998; Thomas and Dawe, 2003; Khokhar et al., 1998; Kumar et al.,
2008; Sloan, 2003), carbon dioxide capture and sequestration
(Yamasaki et al., 2000; Saji et al., 1992; Mondal et al., 2012; Ho
n
Corresponding author. Tel.: þ 1 609258 4583.
E-mail address: sundar@princeton.edu (S. Sundaresan).
et al., 2013; Lee et al., 2010; Babu et al., 2013; Linga et al., 2007,
2007; Babu et al., 2015) and desalination (Bradshaw et al., 2007;
McCormack and Niblock, 1998, 2000; McCormack and Ripmeester,
2014; Cha and Seol, 2013; Cai et al., 2014; Park et al., 2011; Kang
and Linga, 2014; Han et al., 2014), and several of them could
benefit from hydrate being produced at higher temperatures and/
or lower pressures. A continuous hydrate-based desalination pro-
cess would involve a crystallizer in which brine or seawater is
mixed with appropriate hydrate formers – usually at sea surface
level or at depth to take advantage of greater pressures (Max,
2003), washing unit (Wiegandt et al., 1972, 1978; McCormack and
Niblock, 1998, 2000; Han et al., 2014) to separate the salt-free
hydrate particles from concentrated brine solutions, and a device
to melt the hydrate to produce fresh water and recover hydrate
former(s) as well as a small amount of seed hydrate crystals to be
returned to the crystallizer. These could be configured for both
continuous and semi-continuous operation. The formation of
hydrate is usually favored by lower temperatures and higher
pressures; single component hydrates, as discussed in our earlier
study (Cai et al., 2014), usually require either high pressures (such
as methane hydrate) or low temperatures (such as cyclopentane
hydrate and CH3CCl2F hydrate), and the cost of refrigeration and/or

http://dx.doi.org/10.1016/j.ces.2015.11.001
0009-2509/& 2015 Elsevier Ltd. All rights reserved.
126 L. Cai et al. / Chemical Engineering Science 141 (2016) 125–132
pressurization is unfavorable to hydrate-based desaliantion. The
use of binary hydrate could be a potential solution to reduce such
cost significantly.
With their large and small cavities filled with different guest
species of appropriate sizes, binary clathrate hydrates can form at
combinations of higher temperatures and lower pressures, where
neither of the corresponding single component hydrates is stable.
Cyclopentane (CP)–methane binary hydrate, thermodynamically
stable up to 22 °C at a moderately elevated pressure of 2 MPa,
conditions where neither methane nor CP hydrates is stable, is a
typical candidate for desalination processes. However, the kinetics
of its formation are less well understood. With four phases
involved – a CH4-rich vapor phase, a hydrocarbon-rich liquid, an
aqueous phase and the hydrate solid phase – the formation and
growth of this hydrate is likely to occur at the interfaces between
the latter three. This suggests two possible mass transfer barriers:
1. Methane must be transferred from the vapor phase into and
across one or both liquid phases to reach the vicinity of the
hydrate growing sites.
2. If the aqueous phase is saline, the salt rejected upon hydrate
formation must diffuse away. Its accumulation in the vicinity of
the growing solid hydrate could lower the driving force for
hydrate growth, the so-called concentration polarization effect.
In our previous paper (Cai et al., 2014) the kinetics of the CP–
methane binary hydrate growth with fresh water were studied in a
well-stirred vessel, and analyzed to show that (CH4) transfer lar-
gely defines the overall hydrate growth rate, and that facilitating
CH4 transfer would promote the hydrate growth rate. In order to
assess the viability of hydrate-based applications such as desali-
nation, which involve forming mixed hydrates from saline water,
one must also quantify the effect of salt on hydrate growth rate.
With this in mind, the present study examines the kinetics of
formation of CP–CH4 hydrates in saline water. Salt dissolved in the
aqueous phase will affect both thermodynamic and kinetic aspects
of hydrate growth:
1. Thermodynamically, the introduction of dissolved salt lowers
the chemical potential of water in the aqueous phase, making it
less favorable to form the hydrate. As a result, a lower tem-
perature (T) and/or a higher pressure (P) for the helper gas is
needed to form the hydrate. In this study, the hydrate equili-
brium temperatures corresponding to various methane pres-
sures are measured at salinities of 3.5% and 7% (NaCl, w/w).
2. Kinetically, as noted above, the concentration polarization effect
associated with NaCl rejection by the growing hydrates may
limit the hydrate formation rate, even if hydrate formation in
saline is allowed to occur at the same subcooling level as in the
fresh water case. In this study, hydrate formation kinetics are
measured over a range of T, P and salinities to assess the
severity of concentration polarization.
2. Experimental setup
The equipment and schematics employed in this work are the
same as in our previous study (Cai et al., 2014). Briefly, the
hydrates are formed in a 450 mL stainless steel reactor equipped
with an annular jacket through which a coolant is circulated. The
coolant temperature is controlled by a Julabo thermostat. Multiple
ports on the reactor head allow injection/discharge of selected
hydrate formers (gas/liquid); a Resistance Temperature Detector
(RTD) probe (1/10 DIN, OMEGA Eng. Inc., accuracy  0.04 °C
between 0 and 20 °C) and a pressure transducer (Viatran 345,
accuracy of 0.6% of full range 0–4 MPa) are also installed and
connected to multi-meters (Agilent 34410A) to allow continuous T
and P measurements. Mechanical stirring is achieved via a mag-
netic drive without a rotating seal. The controllers/detectors for T,
P and stirrer speed are managed by a LabVIEW program, which
performs a series of pre-scripted experiments in a reliable and
reproducible fashion. Additional details can be found in our earlier
article (Cai et al., 2014).
2.1. Procedure for measuring hydrate equilibrium conditions
The hydrate equilibrium conditions are determined experi-
mentally by monitoring their formation/dissociation (Sloan and
Koh, 2007), which is sometimes achieved through visual obser-
vation (Deaton et al., 1946) or using more sensitive apparatus such
as a Quartz Crystal Microbalance (Mohammadi et al., 2003). As the
vessel used in our experiments is not transparent, we followed an
isochoric procedure based on monitoring P while varying T, which
is analogous to that used by Ouar et al. (1992). The reactor is first
loaded with water or brine solution and liquid CP; after purging
the air in the headspace, the helper gas CH4 is introduced up to the
desired P. Starting from a temperature above the melting point
(which is dictated by the selected P), the mixture is cooled down
stepwise with very slow stirring, allowing a fixed time interval
(30 min) following each step. The pressure at the end of each step
is recorded (see the blue points in Fig. 1); the modest decline in P
with decreasing T seen in the figure reflects isochoric cooling of
the vapor and higher methane solubility at lower T. One could
easily reach up to  2 K below the melting point without forming
hydrates, as both subcooling and active stirring are typically nee-
ded to trigger primary nucleation in our experiments. Next, a
much lower coolant T is used, and intense agitation (600 RPM) is
applied, in order to trigger hydrate formation, signaled by a sharp
pressure drop, before raising the temperature back to  2 K below
the melting point. Since part of methane is now clathrated, the
pressure is significantly lower than that observed in the cooling
process at the same T. In the melting sequence, represented by the
red points in Fig. 1, T is raised stepwise with a fixed time interval
(120 min) between steps, and the ending pressures are recorded as
shown in Fig. 1. Before the system temperature reaches the
melting point, each temperature rise results in partial melting of
the existing hydrate, which releases methane into the headspace.
2100
2050 (1) Cooling
(2) Induction Period
2000 ociation
o Hydrate Diss
(5) Heating w/
P/kPa
(3) Hydrate Forms
1950
tes ate
cia ydr
1900
sso /H
Equilibrium T, P
Di ating
He
1850
(4)
1800
18.50 19.00 19.50 20.00 20.50 21.00 21.50 22.00
T/ Co
Fig. 1. Determination of CP–CH4 hydrate four-phase equilibrium. Stage (1): the
mixture is cooled down stepwise with very slow stirring, allowing a fixed time
interval (30 min) following each step; the final system pressure is then recorded.
Stage (2): the mixture could be cooled a few degrees below the equilibrium con-
ditions without forming hydrates because of kinetic limitations. Stage (3): hydrate
formation consumes methane and leads to substantial pressure decline. Stage (4):
the mixture is heated stepwise while maintaining intense stirring, allowing a fixed
time interval (120 min) following each step; the system pressure is then recorded.
Stage (5): heating hydrate-free mixture in steps above the melting point. (For
interpretation of the references to color in this figure caption, the reader is referred
to the web version of this paper.)
 L. Cai et al. / Chemical Engineering Science 141 (2016) 125–132 127

As a result, pressure increases as indicated by stage (4) in Fig. 1.
Once T is above the melting temperature, all the hydrate crystals
have dissociated completely and the P vs. T results for both cooling
and heating experiments essentially overlap, as demonstrated by
stages (1) and (5) in Fig. 1. The intersection of stages (4) and (5) is
taken as the hydrate equilibrium (melting) condition. After each
cooling and heating cycle, additional methane is loaded to repeat
the measurement at different pressures.
Since it takes a long time for the hydrate to stabilize its com-
position by absorbing or releasing the guest molecules in response
to the changing P and T, the ending pressures in stage (4) of Fig. 1
do not represent equilibrium values (see Appendix A), even
though the intersection of stages (4) and (5) does correspond to
the equilibrium melting temperature at the given pressure (see
below). In fact, the pressure would continue to change slowly even
18 h after the temperature step, and the red points in this section
would move up slightly if a longer waiting period was allowed
after each step change in temperature. We have ascertained
through a series of tests that 120 minutes of conditioning time is
adequate to get a good estimate of the melting condition (i.e. the
intersection of stages 4 and 5) with an uncertainty no worse than
that inherent in the experimental measurements of P and T.
2.2. Procedure for hydrate formation kinetic experiments
A series of experiments were conducted over a range of oper-
ating conditions following the procedure developed in our earlier
work (Cai et al., 2014). A mixture with known amounts of water
(usually 200 mL) and CP (22.0 g), plus methane at pressures ran-
ging from 1 MPa to 3 MPa, is first loaded into the reactor. Then, the
mixture is cooled to the desired T (typically below the hydrate
melting point) by circulating the coolant through the reactor
jacket, while maintaining no or very slow stirring. Vigorous stir-
ring is then initiated to trigger hydrate formation and to provide
active mixing. The temperature rises as heat is released by hydrate
formation and the pressure declines as methane transfers into
hydrate. The T and P are automatically logged by a computer for
subsequent analysis. Fig. 2 illustrates the data acquired during two
hydrate formation experiments performed at the same coolant
temperature (15.3 °C), nearly the same starting pressure (  2 MPa)
and the same stirrer speed (600 RPM) but with different salinities
(fresh water vs. 7.0% (w/w) NaCl solution). The differences
between their T and P trajectories demonstrate that salt affects the
hydrate formation rates at nearly identical operating conditions.
15.70
Through an energy balance analysis, the instantaneous hydrate
growth rate on a water basis is recovered from the temperature–
time profiles. Analysis of the pressure data yields methane con-
sumption rate. Aside from the necessary physical parameters
adjusted for experiments conducted with brine solutions (e.g. the
heat capacity of the aqueous phase; see Appendix B), the data
analysis procedure in this work is the same as that applied in the
previous work (Cai et al., 2014) and is therefore not repeated here.
3. Results and discussions
3.1. Hydrate equilibrium in brine solution
Table 1 lists the four-phase CP–CH4 hydrate equilibrium con-
ditions measured in this study over the pressure range from 1 MPa
to 3 MPa. These results are also shown in Fig. 3, and compared to
literature values. The fresh-water data in this work agree well with
those reported by Tohidi et al. (1997), Sun et al. (2002) and Chen
et al. (2010). However, the equilibrium data for NaCl solutions
(with both 3.5% and 7% salinity) acquired in this work differ sig-
nificantly from those reported by Chen et al. (2010) (3.5% and 7%),
which appears to be the only published article that reports such
data. For instance, the equilibrium pressure for 289.2 K in 7% NaCl
solution is 1.49 MPa in our experiments (see Table 1), in contrast to
the 1.78 MPa at 289.14 K reported by Chen et al. (7%); the equili-
brium pressure for 3.5% NaCl solution at 295.4 K is 3.03 MPa in
Table 1, compared to 4.27 MPa at 295.8 K provided by Chen et al.
(3.5%). The discrepancy is beyond the margin of experimental
uncertainty. We point out that Chen et al. determined the equili-
brium conditions through visual observation; at a constant tem-
perature, the methane pressure is raised/lowered and hydrate
formation/dissociation is detected solely by visual observation.
Nevertheless, it is unclear how the two methods could yield closer
results with fresh water but not for the brine solution. Fig. 3a and b
illustrates how the four phase CP–CH4 hydrate equilibrium curve
changes with salinity. In the pressure range of  1–3 MPa, the
melting point declines by  1 K from fresh water to 3.5% in NaCl,
and  3 K from fresh water to 7% in NaCl. In comparison, Chen
et al. (3.5%) find the shift introduced by 3.5% NaCl to be  3 K.
Overall, the magnitudes of the melting point shifts due to salinity
seen in our experiments are similar to those observed in the same
pressure range for other hydrates formed with methane, ethane,
propane, CO2, etc. (Mohammadi et al., 2008). Since seawater has a
salinity close to a 3.5%(w/w) NaCl solution, approximately half of
the feed seawater needs to be converted into hydrate before the
residual aqueous phase salinity rises to 7%. Therefore, for a CP–CH4
2000 hydrate-based desalination process that (1) uses seawater as

15.60 1950 Table 1

Pressure (kPa)

Four phase CP–CH4 hydrate equilibrium conditions measured in our experiments

with (1) fresh water (2) 3.5% (w/w) NaCl solution (3) 7% (w/w) NaCl solution.
15.50 1900
Accuracy: 7 0:1 K for temperature and 7 0:02 MPa for pressure.
T/ C

15.40 1850 Fresh water 3.5% NaCl 7% NaCl

T/K P/MPa T/K P/MPa T/K P/MPa

15.30 Fresh Water 1800
7.0 %(w/w) NaCl Solution 290.6 1.09 290.0 1.26 288.5 1.33
15.20 1750 292.0 1.32 291.0 1.48 289.2 1.49
0 500 1000 1500 292.7 1.54 291.9 1.73 290.3 1.76
Time/s 293.6 1.79 292.8 2.01 291.2 2.04
294.5 2.07 294.0 2.39 291.8 2.24
Fig. 2. Temperature and pressure data for two hydrate experiments, one performed 295.1 2.32 294.8 2.72 292.3 2.38
with fresh water (black) and one with 7.0 % (w/w) NaCl solution (blue). The same 296.0 2.67 295.4 3.03 292.8 2.62
stirrer speed (600 RPM), same amount of cyclopentane (22.0 g) and the same 296.9 3.08 293.2 2.80
volume of aqueous phase (200 mL of either fresh water or NaCl solution) are used 293.6 3.00
in both cases. (For interpretation of the references to color in this figure caption, the 294.2 3.26
reader is referred to the web version of this paper.)
128 L. Cai et al. / Chemical Engineering Science 141 (2016) 125–132

experiments, and Δdis H mðCH4 Þ is readily found from the slopes and
3.50
Fresh water, this work estimated values of Z. These enthalpy values can be used to esti-
3.00 Fresh water fitted mate the methane occupancy level in the small cavities of the
Sun binary hydrate, by assuming that a constant ice-like hydrate dis-
Tohidi
2.50 sociation heat (6.0 kJ/mol water) applies for the hydrates and that
Chen
3.5% this work no methane goes into the large cavities of this sII binary hydrate
P/MPa

2.00 7.0% this work which are fully occupied by CP. Noting the stoichiometric ratio for
3.5 % fitted sII hydrates of 2 small cavities per 17 water molecules, the implied
1.50 7.0 % fitted methane occupancy in small cavities along the equilibrium curve,
denoted by θs;CH4 , is
im;eq
1.00
17 Δdis H mðH2 OÞ
0.50 θim;eq
s;CH4 ¼  ð2Þ
2 Δdis HmðCH4 Þ
0.00
5.00 10.00 15.00 20.00 25.00 Table 2 indicates that addition of NaCl up to 7.0%(w/w) results
in a modest decrease in Δdis H mðCH4 Þ and a small increase in θs;CH4 at
im;eq
T/oC
four-phase equilibrium in the temperature range studied. A similar
trend in the enthalpy is reported by Chen et al. (2010). The trend is
16.00
qualitatively as expected since the formation of clathrate hydrate
is thermodynamically less favorable. The occupancy changes
15.00
reported here are quite small, which supports the use of Clausius–
Clapeyron equation over the temperature and salinity ranges
14.00
studied.
ln(P/Pa)

13.00 Fresh water, this work 3.2. Kinetics of binary hydrate formation in brine solutions
Fresh water fitted
12.00 Fresh water, partial linear fit
Sun The process of forming CP–CH4 hydrate is complicated due to
Tohidi the number of phases and species involved, and the coupling of
11.00 Chen multiple mass transfer processes. Our previous work (Cai et al.,
3.5% NaCl, this work
7.0% NaCl, this work 2014) gave several insights concerning the kinetics of formation of
10.00 CP–CH4 hydrate in fresh water, including the following:
0.00335 0.00340 0.00345 0.00350 0.00355
1/(T/K)
1. At a constant stirring speed, the CP–CH4 hydrate formation rate
Fig. 3. (a) Four phase CP–CH4 hydrate equilibrium conditions for fresh water evolves over time as hydrate builds up. The instantaneous rate
(black), 3.5% (w/w) NaCl solution (red) and 7.0% (w/w) NaCl solution (blue), com- depends not only on the operating conditions (T, P and stirring
pared to those reported by Sun et al. (2002), Tohidi et al. (1997) and Chen et al.
speed) but also on the amount of hydrate already formed.
(2010). (b) Equilibrium conditions presented as ln(P/Pa) vs. 1=T graphs. Solid lines
are derived via linear regressions of lnðP eq =PaÞ with 1=ðT=KÞ. The dashed curve 2. Subcooling level is a useful index for correlating growth rates
corresponds to a quadratic fit of ln(Peq/Pa) as a function of 1/(T/K) that represents obtained across a wide range of T and P, superior to alternatives
all the fresh water data. While the higher-order regression is necessary to capture such as excess pressure, and is a good measure of the thermo-
the data trend across the wider range in the pure water case, the brine data are well
dynamic driving force for rates observed under different oper-
represented by a linear regression. (For interpretation of the references to color in
this figure caption, the reader is referred to the web version of this paper.)
ating conditions.

feedstock and (2) does not require very high water conversions, Here, subcooling level ΔT refers to the difference between the
the effect of NaCl on hydrate formation is a reduction of sub- system temperature and the hydrate equilibrium temperature at
cooling level of no more than 3 K. the prevailing bulk methane pressure. Similarly the excess
Given the four phase equilibrium conditions, the dissociation methane pressure is the difference between the prevailing
enthalpy on a methane basis can be estimated from the Clausius– methane pressure in the vessel and the equilibrium pressure
Clapeyron equation: corresponding to the system temperature. In general, both these
are only apparent quantities as the actual conditions at the
dln P Δdis HmðCH4 Þ growing hydrate crystals are different from the bulk conditions as
¼ ð1Þ
ð1=TÞ ZR a result of mass transfer resistance. As this mass transfer resistance
Here Z is the compressibility factor of CH4 vapor at the equili- is difficult to quantify precisely, the apparent values were used in
brium condition, which is estimated using the Peng–Robinson our earlier study as metrics to compare growth rate at different
equation of state (Peng and Robinson, 1976), and R is the gas conditions. The same procedure is followed in this study, as
constant. Strictly speaking, Eq. (1) requires the hydrate and the detailed in Appendix B (Table 3).
methane-rich vapor phases to maintain fixed compositions along
the equilibrium curve. However, if it is assumed that the hydrate Table 2
composition varies negligibly within the temperature range of our Four phase CP–CH4 hydrate equilibrium conditions measured in our experiments
with (1) fresh water (2) 3.5% (w/w) NaCl solution (3) 7% (w/w) NaCl solution.
interest, Eq. (1) gives a first approximation of the hydrate forma- Accuracy: 7 0:1 K for temperature and 7 0:02 MPa for pressure.
tion enthalpy. Likewise, over the narrow temperature ranges
explored (see Fig. 3 for the 3.5% and 7% NaCl curves), this is a very Salinity Range of T Range of Z Δdis HmðCH4 Þ θim;eq
s;CH4
reasonable assumption for the vapor phase. As indicated by Fig. 3b, (w/w) (K) (kJ/mol)
while a higher order regression will better represent the equili-
brium data over a larger temperature range (the dashed curve in 0 290.6–296.9 0.969–0.926 111.7–116.9 0.437–0.457
3.5% 290.0–295.4 0.963–0.924 105.6–110.1 0.464–0.484
Fig. 3), ln P measured in this work exhibits good linear depen- 7.0% 288.5–294.2 0.962–0.918 102.0–106.9 0.478–0.501
dence on 1=T in the temperature range covered in our
 L. Cai et al. / Chemical Engineering Science 141 (2016) 125–132 129

Following the methods applied in the fresh water study, the
percent water conversion over the first 20 min (Fig. 4a) or 30 min
(Fig. 4b) after starting hydrate formation is used to characterize
the hydrate growth rates at various ΔT levels. The filled and hol-
low symbols correspond to experiments performed at coolant
temperatures of 15.3 °C and 16.3 °C, respectively. The figures
summarize results for hydrate growth in fresh water (black), 3.5%
NaCl solution (red) and 7% NaCl solution (blue). As readily seen
fresh water. For ΔT 4 4 K, the difference is less significant and the
rates in NaCl solutions could sometimes even be higher than those
in fresh water. To better understand this behavior, a closer look at
the temporal evolution of the hydrate formation rates under dif-
ferent conditions is required. The energy balance analysis proce-
dure is described in our previous work (Cai et al., 2014)
C V dT K
rw ¼ þ ðT T 0 Þ ð3Þ
Lf dt Lf

Here rw is the water consumption rate, Lf is the latent heat

5.0
associated with the melting of hydrate (on a water basis), CV is the
Percentage Water Conversion

isochoric heat capacity of the hydrate pressure cell and contents,

4.0 T0 is the temperature of the circulating coolant plus a compensa-
tion term accounting for heat generated by agitation. K is (with
after 20min

3.0 dimension J/(K s) is the product of wall heat transfer coefficient

and the heat transfer area.
2.0 Fig. 5 presents results obtained in hydrate formation experi-
Fresh water, 15.3ºC
Fresh water, 16.3ºC ments in fresh water and 7.0% (w/w) NaCl solution with ΔT  3 K.
1.0
3.5%(w/w) NaCl, 15.3ºC The difference between reactor temperature and T0 at various
3.5%(w/w) NaCl, 16.3ºC
7.0%(w/w) NaCl, 15.3ºC times since the initiation of hydrate formation is shown in Fig. 5a,
7.0%(w/w) NaCl, 16.3ºC along with the instantaneous apparent subcooling level. Clearly,
0.0
0.00 2.00 4.00 6.00 8.00 the apparent subcooling levels are comparable throughout the
Degree of Subcooling ΔT/K course of the experiments, with a difference no greater than 0.5 K.
The initial temperature increase and the rate of increase in the
experiment with NaCl solution are generally lower than those with
7.0
fresh water. According to Eq. (3), that translates to a lower hydrate
Percentage Water Conversion

6.0 formation rate in NaCl solution. As seen in Fig. 5a, ðT T 0 Þ in the

experiment with NaCl solution is larger than that in fresh water
5.0
after 30min

4.0 4.00
0.25 3.50
3.0

Subcooling Level ΔT/K

Fresh water, 15.3ºC 3.00
Fresh water, 16.3ºC 0.20
2.0 2.50
3.5%(w/w) NaCl, 15.3ºC Fresh Water
(T-T0)/°C

3.5%(w/w) NaCl, 16.3ºC 0.15 7% NaCl Solution

1.0 7.0%(w/w) NaCl, 15.3ºC
2.00
7.0%(w/w) NaCl, 16.3ºC 1.50
0.0 0.10
0.00 2.00 4.00 6.00 8.00 1.00
Degree of Subcooling ΔT/K 0.05
0.50
Fig. 4. Percent water conversion to CP–CH4 hydrate at various subcooling levels 0.00 0.00
after (a) 20 min; (b) 30 min. The figures show results from experiments conducted 0 500 1000 1500 2000 2500
at 15.3 °C (solid symbols) and 16.3 °C (empty symbols) at various pressures and Time/s
salinity levels (black: fresh water; red: 3.5% in NaCl; blue: 7.0% in NaCl). The error
bar accounts for the fact that pressure and salinity change during the course of an
experiment, leading to a gradual shift in the subcooling level. The same stirrer 0.70 4.00
speed (600 RPM), the same amount of cyclopentane (22.0 g) and the same volume
0.60 3.50
of aqueous phase (200 mL of either fresh water, 3.5% or 7.0% NaCl solution) apply to
all experiments involved, but the pressures are higher in the cases of NaCl solutions
0.50
3.00 Subcooling Level ΔT/K
at the same coolant temperature and ΔT. The pressure ranges covered by these
r W/(mmol/s)

Fresh Water 2.50

tests are 1–2 MPa for fresh water, 1.5–3 MPa for 3.5% NaCl solution and 1.7–3.3 MPa 0.40
for 7.0% NaCl solution. (For interpretation of the references to color in this figure 7% NaCl Solution 2.00
caption, the reader is referred to the web version of this paper.) 0.30
1.50
0.20
from Fig. 3, at a specified temperature, the methane pressure 1.00

required to achieve a desired ΔT increases with salt concentra- 0.10 0.50

tion; as a result, experiments involving brine solutions have been 0.00 0.00
0.00 0.05 0.10 0.15 0.20 0.25 0.30
done over higher ranges of pressures, as summarized in the
caption of Fig. 4. As the methane pressure decreases during the nW/mol

course of an experiment (e.g., see Fig. 2) and the concentration of
salt in the brine increases, ΔT decreases gradually with time; this
variation is indicated by the error bars in Fig. 4, with the points
themselves denoting an average level.
Comparison of the curves at different salinities indicates that at
the same T and ΔT, there is only a modest change in the hydrate
formation rate as the salinity increases from 0 to 7%, with the
difference never greater than a factor of 2. For ΔT o 4 K, the
hydrate formation rate in NaCl solution is smaller than that in
Fig. 5. (a) Temperature excursion and the evolution of subcooling level in two
hydrate formation experiments done at similar nominal subcooling level of 3 K. The
coolant temperature¼15.3 °C. The stirrer speed (600 RPM), the amount of cyclo-
pentane (22.0 g) and the volume of aquaous phase (200 mL) are kept the same in
the two experiments. During the course of the fresh water experiments (black
lines), the pressure drops from 1.5 MPa to 1.3 MPa. In the experiment with 7.0%
NaCl solution (blue lines), pressure drops from 2.3 MPa to 2.1 MPa. (b) The
instantaneous hydrate growth rates and subcooling levels of the same experiments
are now plotted against the amount of hydrates formed (quantified by the number
of water moles converted to hydrate). (For interpretation of the references to color
in this figure caption, the reader is referred to the web version of this paper.)
130 L. Cai et al. / Chemical Engineering Science 141 (2016) 125–132
0.30
0.20
(T-T 0)/°C
0.10
Fresh Water
3.5% NaCl Solution
0.00
0 200 400 600 800 1000 1200
Time/s
0.80
0.60
r W/(mmol/s)
0.40
0.20 Fresh Water
3.5% NaCl Solution
0.00
0 0.05 0.10 0.15 0.20 0.25
nW/mol
Fig. 6. (a) Temperature excursion and evolution of subcooling level in two hydrate
formation experiments done at similar subcooling level of  6 K. The coolant
temperature is 15.3 °C, the stirrer speed (600 RPM) and the amount of cyclo-
pentane (22.0 g) and the volume of aqueous phase (200 mL) are kept the same in
the two experiments. During the course of the fresh water experiments (black
lines), the pressure drops from 2.3 MPa to 2.1 MPa. In the experiment with 3.5%
NaCl solution (red lines), pressure drops from 2.6 MPa to 2.4 MPa. (b) The instan-
taneous apparent hydrate growth rates and subcooling levels in the same experi-
ments are now plotted against the amount of hydrates formed (quantified by the
number of water moles converted to hydrate). (For interpretation of the references
to color in this figure caption, the reader is referred to the web version of this
paper.)
experiment at later times, suggesting a higher hydrate formation
rate. However, this comparison is deceptive as the cumulative
amounts of hydrates formed in the two experiments are different.
When the instantaneous growth rates (deduced from Eq. (3)) are
plotted against the amount of hydrate formed (found by inte-
grating the instantaneous hydrate growth rate), see Fig. 5b, one
finds that the hydrate growth rate in NaCl solution is system-
atically lower than that in fresh water at every conversion level.
When experiments are repeated at higher subcooling levels,
the difference between the rates in fresh water and NaCl solution
becomes less pronounced. Fig. 6 is similar to Fig. 5, but compares
the results obtained with fresh water and 3.5% NaCl solution at
ΔT  6 K. The subcooling levels are very similar at all times (with a
difference being less than 0.2 K); furthermore, the temperature
profiles are very similar in the two experiments, indicating similar
hydrate growth rates. At the very early stage, the hydrate forms at
a slightly greater rate in NaCl solution than in fresh water.
In order to rationalize these results, multiple factors need be
recognized:
1. As hydrates form and grow in saline solutions, they reject salt
which results in an accumulation of salt near the growing
crystals; the extent of salt accumulation depends on the hydrate
growth rate and the rate at which the salt is transported away
by diffusion and mixing.
2. The apparent excess pressures may be taken as a measure of the
driving force for methane transfer from the bulk gas
6.00 phase to the growing hydrates crystals (defined as ΔP ¼ f ðT; P;
caq Þ ¼ P  P 4  phase equil: ðT; caq Þ). They are generally not the
Subcooling Level ΔT/K
NaCl NaCl
same between experiments conducted with fresh water and
4.00 brine solutions, respectively, when fixing other factors to be the
same or comparable (i.e. operating temperature T and sub-
cooling level ΔT).
2.00 3. The amount of CH4 pre-dissolved in the liquid phases can be
different. For instance, at the same temperature the methane
pressure needed to attain the same degree of subcooling is
0.00 higher in the experiments with saline solutions, while the
1400
methane solubility in water is not significantly affected by salt
concentration. Thus, more methane is pre-dissolved (in both the
aqueous and CP phases) in the experiments with saline solution,
which is readily available for hydrate growth at the early stage
6.00 without need for immediate methane transfer from the vapor
Subcooling Level ΔT/K
phase. This could give rise to faster initial growth rates in saline
solutions when compared to fresh water.
4.00
For the low ΔT cases represented by Fig. 5, evolutions of ΔT are
similar for both the fresh water and salt water experiments, and
2.00 the excess pressure ΔP – driving force of methane transfer – is
larger for the 7%(w/w) NaCl solution case than the fresh water case
(1.0 MPa vs. 0.75 MPa at the beginning of hydrate formation).
0.00 Similarly, the pressure is also higher for the 7%(w/w) NaCl solution
0.30
case (2.3 MPa vs. 1.5 MPa at the beginning of hydrate formation),
and that leads to a favorable greater amount of pre-dissolved
methane. Nonetheless, the growth rate in the salt–water experi-
ment is less than that in the fresh water experiment (see Fig. 5b).
Taken together, these observations suggest that the mass transfer
resistance for the removal of NaCl rejected by the growing
hydrates is likely to play a role in hindering the hydrate
growth rate.
On the other hand, our experiments at higher apparent sub-
cooling levels suggest that, as presented in Fig. 6, the growth rates
in fresh water and salt water can become similar for the most part
at comparable apparent ΔT levels. As in the low subcooling
experiment in Fig. 5, the apparent excess pressure is significantly
higher in the salt–water case than in the fresh water one; however,
the differential between the excess pressures is less significant
during the two experiments shown in Fig. 6. Furthermore, the
growth rates at higher subcooling levels are larger than those at
the lower subcooling level (indicated by the comparison between
Figs. 5b and 6b), which suggests that the salt concentration
polarization associated with removal of NaCl from the hydrate
surface must be greater in the salt–water experiments at higher
ΔT levels. Yet, the growth rate in salt–water experiment is com-
parable to that in fresh water. Therefore, some other factors not
resolved in the present study must be at stake as well. For
instance, prior studies have indicated different morphologies of
hydrate formed at various subcooling levels (Ishida et al., 2012;
Ohmura et al., 2005, 2000), which could potentially contribute to
different mass transfer resistance. However, these are only
hypotheses that could be probed in future studies.
3.3. Occupancy derived from kinetic data
The actual hydrate formation rate on methane basis cannot be
determined solely from T and P data of the kind shown in Fig. 2, as
data on evolution of methane concentrations in the liquid phases
are also needed. However, by introducing a quasi-steady state
assumption, the methane consumption rate can estimated from
the T and P data, following the procedure used in our earlier study
(Cai et al., 2014). Further assuming that all the large cavities of the
hydrate are occupied by cyclopentane and that methane can only
 L. Cai et al. / Chemical Engineering Science 141 (2016) 125–132 131

Apparent Methane Occupancy 1.00

Water 1.05 MPa Water 1.25 MPa Water 1.47 MPa equilibrium methane occupancy of the small hydrate cavities in
Water 1.95 MPa
3.5% 1.47 MPa
Water 2.19 MPa
3.5% 1.71 MPa
Water 2.50 MPa
3.5% 1.97 MPa
the cyclopentane–methane binary hydrate is estimated to be  0.5
0.80 3.5% 2.65 MPa
7.0% 2.57 MPa
7.0% 2.18 MPa
7.0% 2.73 MPa
7.0% 2.34 MPa in the 15–20 °C range. The kinetics of hydrate formation were
followed by measuring the evolution of temperature and pressure.
0.60 Such measurements were carried out for different salinity and
subcooling levels. The occupancy levels estimated from our kinetic
0.40 experiments are also  0:5. The amounts of hydrate formed in 20
and 30 min were used as metrics to estimate the effect of salinity
0.20 on the hydrate growth rates at different subcooling levels. At low
subcooling levels ðΔT o 4 KÞ, the growth rates were smaller in
0.00 saline solutions than in fresh water by a factor of  2–3, which is
0 500 1000 1500 2000
attributed to concentration polarization associated with salt
Duration (s)
rejection by the hydrate. At higher subcooling levels, the growth
Fig. 7. Evolution of apparent methane occupancy in hydrate formation experi- rates in saline solutions were found to be comparable to or
ments conducted at various salinity levels (black symbols: fresh water; red: 3.5%
sometimes larger than those in fresh water. This is likely due to the
(w/w) NaCl solution; blue: 7% (w/w) NaCl solution) and initial pressures (specified
in the legend). The same coolant temperature (15.3 °C), stirrer speed (600 RPM), complex interplay between concentration polarization stemming
amount of cyclopentane (22 g) and volume of aqueous phase (fresh water or NaCl from slow removal of NaCl from the growing hydrate and mass
solutions) are used in every experiment. The apparent occupancy is estimated
transfer associated with the supply of methane to the hydrate.
using Eq. (4). (For interpretation of the references to color in this figure caption, the
reader is referred to the web version of this paper.)

occupy small cavities, the ratio of the rates inferred on water and
methane bases indicates the fraction of small cavities being Acknowledgements
occupied by methane:
 Financial support through the Grand Challenges Program and
rðCH4 Þ 2
θap
s;CH4 ¼ ð4Þ Project X Innovation Fund at Princeton University is greatly
rðH2 OÞ 17
appreciated.
Fig. 7 shows θs;CH4 estimated in this manner for several differ-
ap

ent hydrate formation experiments performed at the same coolant

temperature of 15.3 °C. All the trajectories reveal low apparent
occupancy levels at the beginning of hydrate formation, gradually Appendix
rising to 0:5 7 0:1; however, as discussed in our earlier study (Cai
et al., 2014), only the values estimated for the later times rea- A. The use of dissociation point as the equilibrium point
sonably characterize the composition of the hydrate. At the early
stage, the hydrate formation mainly consumes the excess methane When performing the equilibrium measurement via the pro-
pre-dissolved in liquid phases, at a faster rate than methane being cedure delineated in Section 2.1, the ending pressure for each
supplied from the vapor phase; as a result, our analysis based on temperature step below the melting point is not considered the
pressure response and quasi-steady state assumption under- equilibrium pressure, because the conditioning time (120 in this
estimates both rðCH4 Þ and θs;CH4 at early times. It is clear from
ap
work) is not sufficient for the remaining hydrate to adjust its
Fig. 7 that, at the later stages of hydrate formation, θs;CH4  0:5 in
ap
composition (guest occupancy) and release/absorb the necessary
all the experiments at this temperature. A salinity up to 7% does
methane to reach a complete equilibrium. These intermediate
not appear to have any systematic (or significant) effect on
values are presented as the green circles in Fig. 8, along with the
methane occupancy of the smaller hydrate cavities. Moreover,
these values are commensurate with those listed in Table 2 (0.4– derived equilibrium curves. The comparison indicates that these
0.5), which were estimated from the enthalpy of hydrate forma- values do deviate from the equilibrium conditions, though slightly.
tion on a methane basis (see Eqs. (1) and (2)). The analysis in this
section, taken together, sheds some light on the key question that 3,500
motivated this study, namely, to what extent does the salt impact Fresh Water
Fresh Water Fitted
the kinetics of binary hydrate formation and processes based on 3,000 3.5% NaCl
them. It is clear from our experiments that, at any given operating 7% NaCl
condition, the presence of salt in the range relevant for seawater 2,500 3.5 % Fitted
desalination decreases the subcooling level by up to  3 K; this can 7.0 % Fitted
2,000 Non-Equilibrium Pressures
be compensated by a modest increase in methane pressure. Fur-
P/kPa

ther, at comparable subcooling levels, the overall hydrate growth

1,500
rate is of the same order of magnitude as in fresh water, smaller by
a factor of 2–3 at low subcooling levels, and comparable at higher
1,000
subcooling levels. These results encourage further efforts to
develop desalination processes based on binary hydrates. 500

0
4. Conclusions 5.00 10.00 15.00 20.00 25.00
T/ C
o

Experimental data on four-phase equilibrium conditions for Fig. 8. Non-equilibrium ending pressures obtained at the melting stage of hydrate
CP–CH4 hydrate at three salinity levels (0, 3.5% and 7.0% (w/w) in equilibrium measurement. Each determined equilibrium pressure (solid marks) is
NaCl) and within the pressure range of 1–3 MPa are reported. The accompanied by the corresponding non-equilibrium pressures.
132 L. Cai et al. / Chemical Engineering Science 141 (2016) 125–132
B. Physical properties of fresh water/brine solutions used in the data
analysis
Different parameters for the fresh water and brine solutions are
needed when deriving rw via Eq. (3), a procedure developed in the
earlier paper (Cai et al., 2014), and they are summarized below.
While these properties vary slightly within the pressure and
temperature ranges in this studied, the deviation is neglected
without introducing significant error.
Table 3
Physical properties for fresh water and brine solutions at 285 K and 1 atm. (Archer
and Carter, 2000).
Salinity (w/w) ρ/(103 kg/m3) CV/(103 J/kg K)
0 0.9998 4.19
3.5% 1.0232 4.01
7.0% 1.0486 3.84
References
Archer, D.G., Carter, R.W., 2000. Thermodynamic properties of the NaCl þ H2O
system. 4. Heat capacities of H2O and NaCl(aq) in cold-stable and supercooled
states? J. Phys. Chem. B 104 (35), 8563–8584. http://dx.doi.org/10.1021/
jp0003914.
Babu, P., Kumar, R., Linga, P., 2013. Pre-combustion capture of carbon dioxide in a
fixed bed reactor using the clathrate hydrate process. Energy 50 (0), 364–373.
http://dx.doi.org/10.1016/j.energy.2012.10.046 〈http://www.sciencedirect.com/
science/article/pii/S0360544212008262〉.
Babu, P., Linga, P., Kumar, R., Englezos, P., 2015. A review of the hydrate based gas
separation (hbgs) process for carbon dioxide pre-combustion capture. Energy
85, 261–279. http://dx.doi.org/10.1016/j.energy.2015.03.103 〈http://www.scien
cedirect.com/science/article/pii/S0360544215004120〉.
Bradshaw, R.W., Greathouse, J.A., Cygan, R.T., Simmons, B.A., Dedrick, D.E., Majzoub,
E.H., 2007. Desalination Utilizing Clathrate Hydrates. Report, Sandia National
Laboratories.
Cai, L., Pethica, B.A., Debenedetti, P.G., Sundaresan, S., 2014. Formation kinetics of
cyclopentane–methane binary clathrate hydrate. Chem. Eng. Sci. 119 (0),
147–157. http://dx.doi.org/10.1016/j.ces.2014.08.025 〈http://www.sciencedirect.
com/science/article/pii/S0009250914004436〉.
Cha, J.-H., Seol, Y., 2013. Increasing gas hydrate formation temperature for desali-
nation of high salinity produced water with secondary guests. ACS Sustain.
Chem. Eng. 1 (10), 1218–1224.
Chen, Z.-Y., Li, Q.-P., Yan, Z.-Y., Yan, K.-F., Zeng, Z.-Y., Li, X.-S., 2010. Phase equilibrium
and dissociation enthalpies for cyclopentane þ methane hydrates in NaCl
aqueous solutions. J. Chem. Eng. Data 55 (10), 4444–4449. http://dx.doi.org/
10.1021/je100597e.
Deaton, W.M., Frost, E.M., 1946. Hydrates and Their Relation to the Operation of
Natural-Gas Pipe Lines, U.S. Bureau of Mines Monograph 8.
Gudmundsson, J.S., Andersson, V., Levik, O.I., Parlaktuna, M., 1998. Hydrate Concept
for Capturing Associated Gas. doi:10.2118/50598-MS.
Han, S., Shin, J.-Y., Rhee, Y.-W., Kang, S.-P., 2014. Enhanced efficiency of salt removal
from brine for cyclopentane hydrates by washing, centrifuging, and sweating.
Desalination 354, 17–22. http://dx.doi.org/10.1016/j.desal.2014.09.023 〈http://
www.sciencedirect.com/science/article/pii/S0011916414004986〉.
Ho, L.C., Babu, P., Kumar, R., Linga, P., 2013. Hbgs (hydrate based gas separation)
process for carbon dioxide capture employing an unstirred reactor with
cyclopentane. Energy 63 (0), 252–259. http://dx.doi.org/10.1016/j.
energy.2013.10.031 〈http://www.sciencedirect.com/science/article/pii/
S0360544213008827〉.
Ishida, Y., Takahashi, Y., Ohmura, R., 2012. Dynamic behavior of clathrate hydrate
growth in gas/liquid/liquid system. Cryst. Growth Des. 12 (6), 3271–3277. http:
//dx.doi.org/10.1021/cg300416x.
Kang, K.C., Linga, P., nam Park, K., Choi, S.-J., Lee, J.D., 2014. Seawater desalination by
gas hydrate process and removal characteristics of dissolved ions (Na þ , K þ ,
Mg2 þ , Ca2 þ , B3 þ , Cl  , S2O42  ). Desalination 353, 84–90. http://dx.doi.org/
10.1016/j.desal.2014.09.007 〈http://www.sciencedirect.com/science/article/pii/
S0011916414004731〉.
Khokhar, A., Gudmundsson, J., Sloan, E., 1998. Gas storage in structure H hydrates.
Fluid Phase Equilib. 150 (0), 383–392. http://dx.doi.org/10.1016/S0378-3812
(98)00338-0 〈http://www.sciencedirect.com/science/article/pii/
S0378381298003380〉.
Kumar, R., Linga, P., Moudrakovski, I., Ripmeester, J.A., Englezos, P., 2008. Structure
and kinetics of gas hydrates from methane/ethane/propane mixtures relevant
to the design of natural gas hydrate storage and transport facilities. AIChE J. 54
(8), 2132–2144. http://dx.doi.org/10.1002/aic.11527.
Lee, H.J., Lee, J.D., Linga, P., Englezos, P., Kim, Y.S., Lee, M.S., Kim, Y.D., 2010. Gas
hydrate formation process for pre-combustion capture of carbon dioxide.
Energy 35 (6), 2729–2733. http://dx.doi.org/10.1016/j.energy.2009.05.026
〈http://www.sciencedirect.com/science/article/pii/S0360544209001960〉.
Linga, P., Kumar, R., Englezos, P., 2007. Gas hydrate formation from hydrogen/car-
bon dioxide and nitrogen/carbon dioxide gas mixtures. Chem. Eng. Sci. 62 (16),
4268–4276. http://dx.doi.org/10.1016/j.ces.2007.04.033.
Linga, P., Kumar, R., Englezos, P., 2007. The clathrate hydrate process for post and
pre-combustion capture of carbon dioxide. J. Hazard. Mater. 149 (3), 625–629.
http://dx.doi.org/10.1016/j.jhazmat.2007.06.086 〈http://www.sciencedirect.
com/science/article/pii/S0304389407009557〉.
Max, M., 2003. Land-Based Desalination Using Positively Buoyant or Negatively
Buoyant/Assisted Buoyancy Hydrate. US Patent 6,531,034 (March 11). 〈https://
www.google.com/patents/US6531034〉.
McCormack, R.A., Niblock, G.A., 1998. Build and Operate a Clathrate Desalination
Pilot Plant. Report 31. Thermal Energy Storage, Inc., San Diego, CA, USA.
McCormack, R.A., Niblock, G.A., 2000. Investigation of High Freezing Temperature,
Zero Ozone, and Zero Global Warming Potential Clathrate Formers for Desali-
nation. Report 59. Thermcal Energy Storage, Inc.
McCormack, R., Ripmeester, J., 2014. Clathrate Desalination Process Using an
Ultrasonic Actuator. US Patent App. 14/347,228 (August 14). 〈http://www.goo
gle.com/patents/US20140223958〉.
Mohammadi, A.H., Tohidi, B., Burgass, R.W., 2003. Equilibrium data and thermo-
dynamic modeling of nitrogen, oxygen, and air clathrate hydrates. J. Chem. Eng.
Data 48 (3), 612–616. http://dx.doi.org/10.1021/je025608x.
Mohammadi, A.H., Afzal, W., Richon, D., 2008. Gas hydraytes of methane, ethane,
propane, and carbon dioxide in the presence of single NaCl, KCl, and CaCl2
aqueous solutions: experimental measurements and predictions of dissociation
conditions. J. Chem. Thermodyn. 40 (12), 1693–1697. http://dx.doi.org/10.1016/
j.jct.2008.06.015, URL 〈http://www.sciencedirect.com/science/article/pii/
S0021961408001432〉.
Mondal, M.K., Balsora, H.K., Varshney, P., 2012. Progress and trends in CO2 capture/
separation technologies: a review. Energy 46 (1), 431–441. http://dx.doi.org/
10.1016/j.energy.2012.08.006 〈http://www.sciencedirect.com/science/article/
pii/S0360544212006184〉.
Ohmura, R., Kashiwazaki, S., Mori, Y.H., 2000. Measurements of clathrate–hydrate
film thickness using laser interferometry. J. Cryst. Growth 218 (2), 372–380.
Ohmura, R., Matsuda, S., Uchida, T., Ebinuma, T., Narita, H., 2005. Clathrate hydrate
crystal growth in liquid water saturated with a guest substance: observations in
a methane þ water system. Cryst. Growth Des. 5 (3), 953–957. http://dx.doi.
org/10.1021/cg049675u.
Ouar, H., Cha, S., Wildeman, T., Sloan, E., 1992. The formation of natural gas
hydrates in water-based drilling fluids. Chem. Eng. Res. Des. 70 (A1), 48–54.
nam Park, K., Hong, S.Y., Lee, J.W., Kang, K.C., Lee, Y.C., Ha, M.-G., Lee, J.D., 2011. A
new apparatus for seawater desalination by gas hydrate process and removal
characteristics of dissolved minerals (Na þ , Mg2 þ , Ca2 þ , K þ , B3 þ ). Desalination
274, 91–96. http://dx.doi.org/10.1016/j.desal.2011.01.084 〈http://www.science
direct.com/science/article/pii/S0011916411001111〉.
Peng, D.-Y., Robinson, D.B., 1976. A new two-constant equation of state. Ind. Eng.
Chem. Fundam. 15 (1), 59–64.
Saji, A., Yoshida, H., Sakai, M., Tanii, T., Kamata, T., Kitamura, H., 1992. Fixation of
carbon dioxide by clathrate–hydrate. Energy Convers. Manag. 33 (5), 643–649.
Sloan Jr, E.D., Koh, C., 2007. Clathrate Hydrates of Natural Gases. CRC Press, Boca
Raton.
Sloan, D., Koh, C., Sum, A.K., Ballard, A.L., Creek, J., Eaton, M., Lachance, J., Mcmullen,
N., Palermo, T., Shoup, G., Talley, L., 2011. Natural Gas Hydrates in Flow
Assurance. Gulf Professional Publishing, Bostonhttp://dx.doi.org/10.1016/
B978-1-85617-945-4.00013-3 〈http://www.sciencedirect.com/science/article/
pii/B9781856179454000133〉.
Sloan, E.D., 2003. Fundamental principles and applications of natural gas hydrates.
Nature 426 (6964), 353–363.
Sun, Z.G., Fan, S.S., Guo, K.H., Shi, L., Guo, Y.K., Wang, R.Z., 2002. Gas hydrate phase
equilibrium data of cyclohexane and cyclopentane. J. Chem. Eng. Data 47 (2),
313–315. http://dx.doi.org/10.1021/je0102199.
Thomas, S., Dawe, R., 2003. Review of ways to transport natural gas energy from
countries which do not need the gas for domestic use. Energy 28 (14),
1461–1477. http://dx.doi.org/10.1016/S0360-5442(03)00124-5, Cited By (since
1996) 77 〈http://www.scopus.com/inward/record.url?eid ¼ 2-s2.0-
0141674779&partnerID ¼40&md5 ¼89bd967e8d1ccb5c12342eb323b089d1〉.
Tohidi, B., Danesh, A., Todd, A., Burgass, R., Østergaard, K., 1997. Equilibrium data
and thermodynamic modelling of cyclopentane and neopentane hydrates. Fluid
Phase Equilib. 138 (1–2), 241–250. http://dx.doi.org/10.1016/S0378-3812(97)
00164-7 〈http://www.sciencedirect.com/science/article/pii/
S0378381297001647〉.
Wiegandt, H.F., Von Berg, R.L., Leinroth, J.P., 1972. Piston bed and its design. Ind.
Eng. Chem. Process Des. Dev. 11 (3), 404–414. http://dx.doi.org/10.1021/
i260043a013.
Wiegandt, H., Berg, R.V., Patel, P., 1978. The crossflow piston bed. Desalination 25
(3), 303–324. http://dx.doi.org/10.1016/S0011-9164(00)80328-8 〈http://www.
sciencedirect.com/science/article/pii/S0011916400803288〉.
Yamasaki, A., Wakatsuki, M., Teng, H., Yanagisawa, Y., Yamada, K., 2000. A new
ocean disposal scenario for anthropogenic CO2: CO2 hydrate formation in a
submerged crystallizer and its disposal. Energy 25 (1), 85–96. http://dx.doi.org/
10.1016/S0360-5442(99)00036-5.