Chemical Engineering Journal 406 (2021) 126872

Contents lists available at ScienceDirect

Chemical Engineering Journal

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

Hydrate formation from clay bound water for CO2 storage T

a,b a a,⁎ a a
Youhong Sun , Shuhui Jiang , Shengli Li , Xiaochu Wang , Saiyu Peng
a
Key Laboratory of Drilling and Exploitation Technology in Complex Conditions of Ministry of Natural Resources, College of Construction Engineering, Jilin University,
Changchun 130026, China
b
China University of Geosciences, Beijing 100083, China

HIGHLIGHTS GRAPHICAL ABSTRACT

• CO storage in clay was investigated

2
by using low-field NMR technique.
• Gas hydrate can be formed from the
weakly bound water in clay.
• CO storage transforms from adsorp-
2
tion to hydrate formation in mon-
tmorillonite.
• Gas hydrate formation from clay
bound water was greatly affected by
the pressure.

ARTICLE INFO ABSTRACT

Keywords: Geological storage of CO2 in form of hydrate in marine sediments is an alternative option to mitigate global
CO2 storage warming. Clay is one of the most dominant lithology in marine sediments, which has higher water adsorption
Hydrate formation capacity than sand. In this work, the formation of CO2 hydrate from clay bound water was experimentally
Clay bound water investigated by using low-field NMR. The effects of mineral, water content and pressure on the capacity of CO2
Diffuse double layer
storage were analyzed. With the increase of water content, there was a transition of CO2 storage from adsorption
NMR
on the montmorillonite surface to hydrate formation. The strongly bound water competed with CO2 gas for the
adsorption sites on the surface of montmorillonite when the water content was below 15%. Afterwards, CO2
hydrate formation started due to the availability of weakly bound water. The storage of CO2 in hydrate in illite
increased continuously with the increase of water content in the range of 5% to 20% due to the low level of
strongly bound water. During hydrate formation, the peak position of T2 distribution spectra of the wet clay
shifted to the left, indicating an ordered conversion of weakly bound water to hydrate from the outer layer to the
inner layer of diffuse double layer. The final amount of CO2 hydrate formed from weakly bound water increased
with the elevated pressure due to the variation of activity of the bound water in diffuse double layer. These
results will not only broaden the understanding of gas hydrate in clay bound water systems but also promote a
new evaluation of the safety and potential of CO2 hydrate storage in clayey sediments.

1. Introduction can be attributed to the combustion of fossil fuels for electricity gen-
eration, transportation and heating[1]. Before the cost effective use of
CO2 has been considered as a primary greenhouse gas due to the the renewable energy[2] or non-carbon energy[3], the capture of CO2
huge emission. The overall concentration increase of CO2 in atmosphere from burning power plants and the geological storage of CO2 in

⁎
Corresponding author.
E-mail address: [email protected] (S. Li).

https://doi.org/10.1016/j.cej.2020.126872
Received 6 May 2020; Received in revised form 4 August 2020; Accepted 30 August 2020
Available online 03 September 2020
1385-8947/ © 2020 Elsevier B.V. All rights reserved.
Y. Sun, et al. Chemical Engineering Journal 406 (2021) 126872

subsurface porous formation is considered a feasible strategy in miti- significance to the feasibility and safety of CO2 storage in marine se-
gating global warming[4,5]. The injection of CO2 into deep marine diments.
sediments is one of the most common method for large capacity of CO2 Effective evaluation of CO2 sequestration in marine sediments
storage[6,7]. containing certain contents of clay minerals requires a full under-
For the characterization of CO2 storage in certain marine sediments, standing of the potential CO2 hydrate formation from different types of
a number of critical factors such as storage capacity and the potential water in clayey sediment mentioned above. The formation conditions of
leakage need to be assessed. When CO2 was injected into the sub-seabed gas hydrate from capillary water have been found to be above the
sediment under low temperature and high pressure conditions, CO2 equilibrium line of hydrate formed from free water[26], which is at-
hydrate formed from pore water in sediments, which was a promising tributed to the lower activity of capillary-bound water. Compared with
method for CO2 storage. On one hand, gas hydrates are non-stoichio- water bound in capillary, the water adjacent to the surface of clay is
metric inclusion compounds with large capacity of gas storage[8,9]. Up bound in the diffuse electric double layer[27]. The thickness of the
to 160 volumes of CO2 at standard temperature and pressure can be water layer influenced by the negatively charged surface of clay is up to
stored in 1 vol of CO2 hydrate. Thus, clathrate hydrate could trap CO2 in tens of nanometers[28]. The type of water evolves from strongly bound
its crystal structure with very high density, constituting a way of CO2 water to weakly bound water with the decreasing binding strength in
storage. On the other hand, the formation of CO2 hydrate in porous the diffuse double layer. The conversion of clay bound water to gas
sediments decreases the permeability sharply and create an imperme- hydrate could be quite different from that of capillary bound water and
able cap[10,11] that could impede the upward flow of CO2[12,13]. It free water. Previous studies have studied the formation of CO2 hydrate
was found that CO2 hydrate-bearing sediments could significantly re- in sandy sediments, but none of them investigated gas hydrate forma-
duce the risk of CO2 leakage from sub-seabed formation. As a result, tion from clay bound water. The findings obtained from CO2 hydrate
CO2 storage in the form of gas hydrate in marine sediment could be an formation in sandy sediments could not be fully applied to clayey se-
effective and safe method for CO2 sequestration[6]. diments. Specifically, analysis in the previous studies did not consider
To evaluate the potential of CO2 hydrate storage in sediments be- the interaction between the surface of porous media and the superficial
neath seabed, the kinetic behavior of CO2 hydrate formation in sandy water molecules. However, incorporation of water types, binding
sediments was extensively studied[8,14]. Zheng et al.[15] quantita- strength, water activity and their corresponding effects on hydrate
tively analyzed the formation of CO2 hydrate in packed sand by NMR. formation is important in studying the CO2 storage capacity of clayey
The results indicated a two-stage hydrate formation: quick enclathra- sediment.
tion in the initial stage and slowly continuous occupancy in a longer In this work, the formation of CO2 hydrate in wet clay was experi-
stage. Mekala et al.[16] analyzed the effect of particle size on CO2 mentally investigated. The type of bound water in wet montmorillonite
hydrate formation in silica sandy sediments. They found that the and illite was characterized by using low-field NMR. During CO2 hy-
amount of CO2 hydrate formation decreased with the increase of the drate formation from clay bound water, the characteristic spectra of
sand size. The permeability of sandstone can be effectively reduced transverse relaxation time (T2) of the remaining water in wet mon-
with the growth of hydrate in pores. Sun et al.[17] studied the CO2 tmorillonite and illite was obtained. The dynamic effects of mineral,
hydrate formation in silica sediments saturated with water. The results water content and pressure on CO2 hydrate storage in clayey sediments
showed that the conversion of pore water to hydrate reached 39%-55% were quantitatively analyzed. The characteristics of CO2 hydrate for-
after 24 h. The capacity of CO2 storage in hydrate was evaluated with mation from clay bound water were summarized.
respect to different injection modes. Yu et al.[18] studied CO2 hydrate
formation morphologies and rate in sandy sediments under gas/liquid 2. Experimental section
flow condition. It was found that the formation of CO2 hydrate film
resulted in the increase of hydrate saturation and the permeability re- 2.1. Materials
duction of the sandy sediments. Almenningen et al.[19] investigated
the formation of CO2 hydrate in sandstone filled with water after the CO2 with a purity of 99.99% was supplied by Beifang Special Gas
injection of liquid CO2 by using magnetic resonance imaging. It was Industry Corporation in China. Double-distilled water was provided by
verified that hydrate nucleation and growth occurred both at the in- Haodi Chemical Reagent Distribution Co., Ltd. Montmorillonite and
terface of liquid CO2/water and in water phase due to CO2 dissolution. illite was provided by Yanxi mineral factory of Hebei. The specific
The recent studies above mainly focused on the formation of CO2 hy- surface areas of the montmorillonite and illite are 64.56 m2/g and
drate in sandy sediments. 4.88 m2/g, respectively, which were measured by using a multipoint
Besides sand, clay is one of the most dominant lithology in the BET analyzer with N2 as the adsorbate.
marine sediments[20]. The content of clay minerals in gas hydrate
stability zone in Shenhu area in South China Sea is 26%-30%[21].The 2.2. Experimental apparatus
clay minerals are mainly montmorillonite and illite. Natural gas hydrate
in the Krishna-Godavari Basin of India mainly occur in sediments with Two types of experimental apparatus were used in this work: 1)
clay fraction lower than 30%. Clayey sediments under certain tem- nuclear magnetic resonance (NMR) analysis system for the measure-
perature and pressure conditions can be identified as potential sites for ment of transverse relaxation time and 2) quantitative measurement
CO2 hydrate storage in marine sediments. A dense cap layer with low system for the evaluation of CO2 storage capacity.
permeability is easily created with the formation of CO2 hydrate in
clayey sediments[18]. However, there are few works considering CO2 2.2.1. NMR analysis system
hydrate formation in clayey sediments. Due to the difference in particle A schematic diagram of NMR analysis system is shown in Fig. 1. The
size and minerals, the water in clay sediment can be divided into three system consists of a low-field NMR analyzer, a core holder
types: clay-bound water, capillary bound water and free mobile water, (Ф25 × 60 mm) made of polyetheretherketone (PEEK), a pressure
whereas only capillary water and free mobile water are found in packed control system, a temperature control system and a data acquisition
sand[22]. The content of bound water in clayey sediments is more than system. The main frequency is 21.83 MHz. The magnetic field strength
65% in Shenhu area in South China Sea[21]. The gas hydrate formed in is 0.5 ± 0.05 T. The diameter of the probe coil is 70 mm. A fluorine oil
sediments seals pore channels and serves as a media cementing the that has no effects on the NMR is used as the cooling medium and
sediment particles. Consequently, the conversion of the bound water to confining pressure fluid. The uncertainties of temperature and pressure
gas hydrate has crucial effects on the permeability[23,24] and me- measurements in the system are ± 0.1 K and ± 0.01 MPa, respec-
chanical strength of the clayey sediments[25], which is of great tively.

2
Y. Sun, et al. Chemical Engineering Journal 406 (2021) 126872

Fig. 1. Schematic diagram of NMR analysis system.

2.2.2. Quantitative measurement system 2.3. Experimental procedures

A schematic diagram of the quantitative analysis system is shown in
Fig. 2. The main part of the system is a high-pressure cell made of Firstly, the samples with a series of water contents (0%, 5%, 10%,
stainless steel with an effective volume of 240 mL. The specific pressure 15%, 20%, 25%, 30%) were obtained by placing dry montmorillonite or
in the cell was maintained by a reducing pressure valve connected to illite with thin layer in an enclosed space with saturated steam to ab-
CO2 vessel with the same effective volume as the high-pressure cell. The sorb water uniformly. The wet sample was transferred into the core
operating pressure of the whole system is up to 20 MPa. The tempera- holder in NMR analysis system or into the high-pressure cell in the
ture in the cell was controlled by an air bath. The temperature and quantitative measurement system. Secondly, the whole system was
pressure data were recorded by a data acquisition system in experi- vacuumed and purged three times with CO2 to eliminate the air com-
ments. pletely. The core holder or the high-pressure cell was pressurized to a
desired value using CO2 gas. The system pressure was maintained by

Fig. 2. Schematic diagram of the quantitative measurement system.

3
Y. Sun, et al. Chemical Engineering Journal 406 (2021) 126872

Table 1 experiments, CO2 vessel was initially pressurized to a certain value.

Experimental conditions for CO2 storage in montmorillonite (MT) and illite (IL). Then, CO2 gas was injected into the high-pressure cell from CO2 vessel
Runs Sample T/K P / MPa level of water content / % via a pressure reducing valve. The pressure in the high-pressure cell was
maintained by the pressure reducing valve connected to CO2 vessel
1 MT 275.2 3.40 0 during gas consumption. Hence, there was a mass balance among the
2 MT 275.2 3.40 5
CO2 gas phase in CO2 vessel at room temperature, the CO2 gas phase in
3 MT 275.2 3.40 10
4 MT 275.2 3.40 15
the high-pressure cell at experimental temperature and CO2 stored in
5 MT 275.2 3.40 20 sample. The cumulative moles of CO2 storage in samples can be cal-
6 MT 275.2 3.40 25 culated as follows:
7 MT 275.2 3.40 30
8 MT 275.2 2.60 30
nCO2 = n 0 n2 n1 (1)
9 MT 275.2 1.80 30
10 MT 275.2 1.39 30
P0 V0
n0 =
11 IL 275.2 3.40 0 Z0 RT0 (2)
12 IL 275.2 3.40 5
13 IL 275.2 3.40 10 P1 V1
n1 =
14 IL 275.2 3.40 15 Z1 RT1 (3)
15 IL 275.2 3.40 20
P2 V0
n2 =
Z2 RT0 (4)
pressure reducing valve connected to the gas cylinder in NMR analysis
system and to CO2 vessel in the quantitative measurement system, re- where n0 is the number of moles of CO2 initially injected into CO2
spectively. Then the bath temperature was set to a desired value. Once vessel; nCO2 is the number of moles of CO2 stored in montmorillonite or
the system temperature decreased to the required value, the tempera- illite; n1 and n2 are the number of moles of CO2 gas phase in the high-
ture and pressure conditions would be suitable for CO2 hydrate for- pressure cell and CO2 vessel, respectively; P0 is the initial pressure of
mation. The pressure drop in the CO2 vessel indicated the storage of CO2 vessel, which is lower than the liquefied pressure of CO2 at room
CO2 in the sample. The experiment was stopped when there were es- temperature T0; P1 and P2 are the pressures in the high-pressure cell and
sentially no changes in the pressure of the CO2 vessel (or cylinder). For CO2 vessel, respectively; T1 is the experimental temperature in the high-
NMR analysis, the signals of NMR were quantitatively calibrated by pressure cell; V0 is the volume of CO2 vessel; V1 is the gas phase volume
using standard samples containing H2O. The T2 distribution of the of the high-pressure cell containing samples; Z is the compressibility
sample selected was measured from the injection of CO2 to the end of factor calculated by using the Benedict-Webb-Rubin-Starling equation
experiments. of state.
The conditions of experimental runs for CO2 storage in the samples
with different water contents were listed in Table 1. All the experiments 3. Result and discussion
were performed at 275.2 K. The effects of the pressure on the storage of
CO2 were investigated. The conditions of experimental runs were also NMR was a common technique for the determination of the type of
plotted in the typical phase diagrams of CO2 hydrate equilibrium and water in a variety of porous media[31–33]. It was proved that the
CO2 liquefied line as shown in Fig. 3. The conditions of experimental measurement of transverse relaxation time (T2) was particularly useful
runs (except Run 10) were between the CO2 hydrate phase line and CO2 in the formation of gas hydrate in sandstone[34] and wet coal[35]. In
liquefied line. this work, low-field NMR relaxation measurements were carried out to
study the state of clay-bound water. Fig. 4 shows the typical spectra of
T2 distribution of wet montmorillonite, wet illite and standard samples
2.4. Calculations with different water contents. According to previous studies[32], bound
water in clay can be detected by NMR. Bound water distributed on the
The CO2 storage in wet montmorillonite or illite was measured by large surface areas of clay exhibit T2 values in a sub-millisecond range.
using the quantitative measurement system shown in Fig. 2. In the Illites-bound water have specific T2 values of 1 ms while water bound
by kaolinite with smaller specific surface area exhibit higher T2 values
in the range of 10 ms[36]. Hence, in this work, relaxation time around
1 ms corresponds to the bound water in montmorillonite and illite (see
Fig. 4a and 4b). Relaxation time around 100 ms corresponds to free
water (see Fig. 4c). It was observed that there was no free water in wet
montmorillonite when the water content was below 30% or in wet illite
when the water content was below 20% as the intensity of T2 longer
than 100 ms was almost zero. For montmorillonite, it should be noted
that the peak position of T2 spectrum was nearly unchanged with the
increase of water content from 5% to 15% then shifted to the right
when the water increased from 15% to 30%, which indicated that the
water in montmorillonite with the content below and above 15% could
correspond to different types. For illite, the peak position of T2 spec-
trum slightly shifted to the right with the increase of water content,
indicating the water in sample with content above 3% could correspond
to the same type. As is known, a diffuse double layer existed sur-
rounding clay particles. The thickness of the diffuse double layer fully
developed was up to tens nanometers[28], which mainly varied with
the clay mineral and salinity[37,38]. In the diffuse double layer, there
Fig. 3. Experimental conditions shown in the typical phase diagrams of CO2 is an evolution from strongly bound water to weakly bound water with
hydrate equilibrium and CO2 liquefied line. The CO2 hydrate equilibrium was the increase of distance between water molecules and the clay surface.
calculated by Chen-Guo model[29,30]. Due to different mineral composition and physical structure, the

4
Y. Sun, et al. Chemical Engineering Journal 406 (2021) 126872

Fig. 5. The evolution of system pressure and temperature in cooling process.

montmorillonite could contain larger content of bound water than illite

[39]. At low water contents, the water molecules were bound strongly
to the surface of clay. In this study, the bound water with shorter T2
indicates strongly bound water with a maximum content less than 15%
in montmorillonite. The content of strongly bound water in illite could
be about 3%. With the increase of water content, the layer of bound
water expands. The bound water with longer T2 suggests weakly bound
water. The weakly bound water in montmorillonite (above 15%) and in
illite (above 3%) increased obviously with the elevated water content.
The excess-gas method was applied in CO2 hydrate formation ex-
periments. The gas was sufficiently supplied by a CO2 vessel with
known volume via a pressure reducing valve. Fig. 5 shows the evolution
of system pressure and temperature in the cooling process. The tem-
perature decreased to a specific value in montmorillonite with water
contents ranging from 0% to 15% (Fig. 5a). Differently, a transient rise
was observed on the temperature curve for montmorillonite with water
contents higher than 15% and for illte with water contents higher than
5% (Fig. 5b), which is indicative of exothermic reaction of hydrate
formation that occurred in the cell during the cooling process.
In the CO2 storage process, the evolution of T2 distribution of clay
samples was obtained by using the low-field NMR. Fig. 6 shows the T2
distribution of the wet montmorillonite and illite. For all the samples,
Fig. 4. The typical spectra of transverse relaxation time (T2) of montmor-
the T2 spectrum varied with the time, showing a decreasing trend in
illonite, illite and standard samples with different water contents.
intensity in the cooling process. Besides, the intensity of T2 spectrum
decreased further after the temperature balance for the wet montmor-
illonite with water contents of 15% and 30%, and for illite with a water

5
Y. Sun, et al. Chemical Engineering Journal 406 (2021) 126872

content of 20%, indicating the formation of CO2 hydrate. In compar-

ison, there was only a small drop of T2 spectrum intensity for the
sample with a water content of 15%. Hence, it can be deduced that the
amount of CO2 hydrate formed from bound water was small. The reason
may be attributed to the dominant of the strongly bound water on the
surface of clay when the water content was below 15%. For strongly
bound water, the water molecules are nearly immobile as they are
strongly bound to the surface of clay[40]. Compared with the strongly
bound water, the weakly bound waters with relatively higher mobility
form clathrate hydrate structure much easier. Considering the absence
of hydrate in the sample with a water content of 10%, hydrate forma-
tion started when the water content was around 15% due to availability
of the weakly bound water. This was in agreement with the results
presented in Fig. 5 which showed no temperature rise in samples with
water contents below 15%. In addition, it should be noted that the peak
position of T2 spectrum of the wet montmorillonite with a water con-
tent of 30% shifted to the left during CO2 hydrate formation. It is in-
dicative of that weakly bound water adjacent to the clay particle gra-
dually converted to CO2 hydrate in an ordered process from molecules
with longer T2 to molecules with shorter T2. As showed in Fig. 4, the
water with longer T2 corresponded to bound water with weaker
binding strength. Hence, the weaker bound water molecules easily
formed gas hydrate. The phase transition from the weakly bound water
to CO2 hydrate was illustrated in Fig. 7. The growth of CO2 hydrate into
the diffuse double layer of clay consumed the weakly bound water.
Lastly, the final peak position and intensity of T2 spectrum remained
unchanged under the isobaric pressure. The remained bound water did
not transform to CO2 hydrate under the experimental conditions. The
final peak position of the spectra was very close to that of the wet
montmorillonite with water contents below 15% (see Fig. 4a and 6b),
which indicated that the residual water was dominated by the strongly
bound water due to the consumption of the weakly bound water in CO2
hydrate formation. Fig. 6d shows the T2 distribution of the wet illite.
Differently, the T2 spectrum of the wet illite continuously decreased in
intensity with its peak position slightly shifting to the left in hydrate
formation process, which indicated a deep conversion of the weakly
bound water to CO2 hydrate. Only a small amount of water remained in
the final hydrate-bearing illite. The final peak position of the spectra
was close to that of the wet illite with a water content of 3% (see
Fig. 4b) The reason is that illite has weaker absorbing capacity of the
strongly bound water than montmorillonite due to the differences in
mineral structure and specific surface area[39]. The mole ratio of the

Fig. 6. T2 distribution of the wet clay with different water contents during CO2
hydrate formation under 3.40 MPa.

Fig. 7. Illustration of phase transition from weakly bound water to gas hydrate.

6
Y. Sun, et al. Chemical Engineering Journal 406 (2021) 126872

Fig. 9. Variation of the final amount of CO2 storage in clay with water content.
Fig. 8. The cumulative moles of CO2 storage in clay with the time.
final amount of CO2 storage in montmorillonite and illite (per gram)
strongly bound water to weakly bound water of illite with a water with the water content. For montmorillonite, as shown in Fig. 9a, with
content of 20% was much lower than that of montmorillonite with a the increase of water content, the amount of CO2 firstly decreased and
water content of 30%. then increased. The water content of 15% was the inflection point. It
To analyze CO2 storage in wet clay quantitatively, a series of ex- showed that CO2 molecules was adsorbed on the surface of dry mon-
periments were performed in a high-pressure cell with a larger volume. tmorillonite (0%). However, water molecules can compete with CO2
The cumulative moles of CO2 storage in montmorillonite and illite (per molecules for the adsorption sites on montmorillonite and alter the
gram) with time are shown in Fig. 8. In dry clay, CO2 molecules are properties of the surface. After montmorillonite was pretreated with
preferentially adsorbed on the fresh adsorption sites in the early stage. water vapor, the adsorption sites on the surface of montmorillonite
It showed that the storage of CO2 in dry clay increased rapidly after the were preferentially occupied by water molecules[42]. Hence, the ad-
gas injection. When the uptake increases, more adsorption sites were sorption of CO2 in montmorillonite decreased with the increase of water
occupied by the CO2 molecules and the adsorption decelerated. In wet content. Other studies also confirmed the effects of water on the sorp-
montmorillonite with water contents lower than 15%, the adsorption tion of CO2 on clay surface. Busch et al. [43] investigated the sorption
rate was lower than that in dry sample since the adsorption sites were of CO2 in dry and wet Ca-rich montmorillonite, Na-rich montmor-
partly pre-occupied by water molecules. In wet montmorillonite with illonite and illite with the water content varying from 0.1 to 9.7 wt%. It
water contents higher than 15% and in illite with water contents higher was found that the wet samples had markedly lower sorption capacities.
than 5%, CO2 was mainly stored in gas hydrate. It showed that there Kumar [44] measured CO2 sorption in dry and water saturated illite.
was no obvious nucleation time for hydrate formation. Due to the good The results showed that the sorption of CO2 was strongly suppressed by
dispersion of water on the particle surface, the storage of CO2 in wet water. In this work, the lowest CO2 storage was obtained at a water
montmorillonite and illite can be considered as a fast adsorption or content of 15% since the adsorption sites was fully occupied by water
hydration process at the surface of clay minerals, similar to the rapid molecules. At this point, the availability of the weakly bound water led
formation of CO2 hydrate in water-spraying system[41]. The maximum to the formation of gas hydrate (see Fig. 6b). With the increase of water
amount of CO2 stored in clay was attained in hours, which was parti- content above 15%, the amount of weakly bound water increased
cularly affected by the water content. Fig. 9 shows the variation of the steadily. CO2 molecules were trapped in hydrate cavities made up of the

7
Y. Sun, et al. Chemical Engineering Journal 406 (2021) 126872

additional weakly bound water molecules (Fig. 6c and 6d). In general, the hydrate equilibrium pressure at the studied temperature. Ad-
one volume of clathrate hydrate contains up to 160 volumes of gas[45]. ditionally, in the diffuse layer, there was ions concentration gradient in
Hence, the amount of CO2 storage in montmorillonite increased ob- which ions migrated from the surface of clay particles[27] (see Fig. 7).
viously with the increase of water content in the range of 15% to 30%. The further is the distance from the surface, the less is the net positive
Therefore, with the increase of water content, there was a transition charge of the solution around clay particles. Hence, the binding
from CO2 adsorption to CO2 hydrate formation on the surface of strength on weakly bound water decreases with the distance from the
montmorillonite around a water content of 15%. Then montmorillonite surface of the particles, which led to the increase of the activity of the
sediments with low water contents may not be the repository for CO2 weakly bound water from the inner layer to the outer layer. Therefore,
storage as no thick hydrate top cap was formed that impedes the dif- at a certain temperature, the closer was the distance of the weakly
fusion or leakage of CO2. For illite, a constant increase of CO2 storage bound water from the clay surface, the higher was the pressure needed
with the increase of water content was showed in Fig. 9b. Since the for CO2 hydrate formation. The elevated pressure promoted CO2 hy-
content of the strongly water content was about 3% (see Fig. 4b), the drate formation from the weakly bound water layer with increasing
adsorption sites on illite was fully occupied by water molecules when thickness, leading to the increase of CO2 storage with increasing pres-
the water content was higher than 5%. With the increase of water sure. There was a correlation between the pressure and the conversion
content in the range of 5% to 20%, CO2 was stored in gas hydrate ratio of clay bound water to hydrate. Therefore, in the design of CO2
formed from the weakly bound water in illite, which was different from storage in clayey sediments, the effects of pore pressure on the forma-
that of montmorillonite. As a result, illite sediments even with low tion of CO2 hydrate from clay bound water should be quantitatively
water contents may also be repository for CO2 hydrate storage. considered due to its significant impact on the permeability of the top
At certain temperatures, the formation of gas hydrate was greatly hydrate-bearing cap and CO2 storage capacity of the seabed sediments.
influenced by the pressure conditions[46]. In this study, the effects of In summary, the characteristic of the CO2 hydrate formation from
pressure on CO2 hydrate formation from the weakly bound water was clay bound water can be schematically illustrated in Fig. 11. With the
analyzed. Fig. 10 shows the amounts of CO2 hydrate formed in the wet increase of water content, the type of water adjacent the surface of clay
montmorillonite with a water content of 30% under different pressures evolved from strongly bound water to weakly bound water and then
at 275.2 K. Below CO2 hydrate equilibrium pressure (1.58 MPa), the free water. In this work, the water content of 15% in montmorillonite
small amount of CO2 storage may be attribute to the dissolution of CO2 and 3% in illite could be roughly taken as the boundary of the strongly
in water. Above CO2 hydrate equilibrium pressure, the amount of CO2 bound water and the weakly bound water, respectively. The prepara-
hydrate increased with the elevated pressure. It indicated that the final tion of wet sample by steam adsorption method suggested that the free
conversion of the weakly bound water to hydrate was controlled by the water should appear at a water content higher than 30% for mon-
pressure of CO2, which was quite different from that of free water. The tmorillonite and 20% for illite. Correspondingly, the binding strength
conversion of free water to gas hydrate can proceed to a very large on bound water molecules decreased with the distance from the surface
extent under a specific pressure higher than hydrate equilibrium pres- of clay. The activity of bound water evolved in the opposite direction of
sure when the water was well dispersed in porous media. This distinct the binding strength. In strongly bound water layer, CO2 hydrate for-
dependence of water conversion on CO2 pressure is related to the mation is unfeasible due to the immobile water molecules. For weakly
evolution of the weakly bound water activity. Previous reports[26,47] bound water less mobile than free water, the formation pressure of CO2
showed that the formation of gas hydrate depends strongly on the ac- hydrate decreased with the distance between water molecules and clay
tivity of the water molecules. The hydrate inhibition in porous media surface due to the reduction of water activity. Under proper pressure
was attributed to the depression of water activity in the capillary pores conditions, the formation of CO2 hydrate in diffuse double layer started
[26]. Lower water activity led to higher hydrate formation pressure. In at the outer layer and grew into the inner layer of the weakly bound
wet montmorillonite, the weakly bound water molecules are oriented in water orderly (see Fig. 7). The lower activity of water in the inner layer
the electric field of the diffuse double layer on the surface of clay. The inhibited the deeper formation of hydrate. Consequently, CO2 hydrate
binding strength acting at the solid/liquid interface played an im- formation stopped when water activity was low enough so that the
portant role in lowering the activity of the weakly bound water[48]. In hydrate equilibrium pressure increased to the experimental pressure.
turn, this induced a temperature depression of hydrate formation from The point at which hydrate growth in the weakly bound water layer
water weakly bound on the clay surface[49]. The pressure needed for ceased was determined by the pressure. In the experiments, the pres-
CO2 hydrate formation from the weakly bound water was higher than sures of 1.80 MPa, 2.60 MPa and 3.40 MPa contributed to the formation
of CO2 hydrate from the weakly bound water with elevated thickness in

Fig. 10. The effects of pressure on CO2 hydrate formation in wet montmor- Fig. 11. Schematic illustration of CO2 hydrate formation from clay bound
illonite. water.

8
Y. Sun, et al. Chemical Engineering Journal 406 (2021) 126872

wet montmorillonite. The higher capacity of CO2 storage in hydrate [8] Z. Yin, M. Khurana, H.K. Tan, P. Linga, A review of gas hydrate growth kinetic
form in wet clay required a higher pressure of CO2. The present un- models, Chem. Eng. J. 342 (2018) 9–29.
[9] Y. Sun, G. Zhang, S. Li, S. Jiang, CO2/N2 injection into CH4 + C3H8 hydrates for
derstanding will play an important role in filling the knowledge gap gas recovery and CO2 sequestration, Chem. Eng. J. 375 (2019) 121973.
that exists in gas hydrate formation. [10] Z.-F. Sun, N. Li, S. Jia, J.-L. Cui, Q. Yuan, C.-Y. Sun, G.-J. Chen, A novel method to
enhance methane hydrate exploitation efficiency via forming impermeable over-
lying CO2 hydrate cap, Appl. Energy 240 (2019) 842–850.
4. Conclusion [11] B. Tohidi, J. Yang, M. Salehabadi, R. Anderson, A. Chapoy, CO2 Hydrates Could
Provide Secondary Safety Factor in Subsurface Sequestration of CO2, Environ. Sci.
The formation of CO2 hydrate from bound water in wet montmor- Technol. 44 (2010) 1509–1514.
[12] D. Li, S. Ren, H. Rui, CO2 Leakage Behaviors in Typical Caprock-Aquifer System
illonite and illite was experimentally studied at 275.2 K. The low-field during Geological Storage Process, ACS Omega 4 (2019) 17874–17879.
NMR was employed to determine the type and content of the water. The [13] Y. Sun, H. Lu, C. Lu, S. Li, X. Lv, Hydrate dissociation induced by gas diffusion from
effects of mineral, water content and operating pressure on the amount pore water to drilling fluid in a cold wellbore, Advances in Geo-Energy Research 2
(2018) 410–417.
of CO2 storage was analyzed. The result showed the CO2 storage in wet
[14] S.H.B. Yang, P. Babu, S.F.S. Chua, P. Linga, Carbon dioxide hydrate kinetics in
montmorillonite changed from gas adsorption to hydrate formation porous media with and without salts, Appl. Energy 162 (2016) 1131–1140.
with the increasing water content. The inflection point was around a [15] J.-N. Zheng, L. Yang, S. Ma, Y. Zhao, M. Yang, Quantitative analysis of CO2 hydrate
water content of 15%. The decrease of CO2 storage with water content formation in porous media by proton NMR, AIChE J. 66 (2020) e16820.
[16] P. Mekala, M. Busch, D. Mech, R.S. Patel, J.S. Sangwai, Effect of silica sand size on
ranging from 0% to 15% could be attributed to the competitive occu- the formation kinetics of CO2 hydrate in porous media in the presence of pure water
pancy of water molecules on the adsorption sites. The amount of CO2 and seawater relevant for CO2 sequestration, J. Petrol. Sci. Eng. 122 (2014) 1–9.
storage in montmorillonite increased with the water content ranging [17] D. Sun, P. Englezos, Storage of CO2 in a partially water saturated porous medium at
gas hydrate formation conditions, Int. J. Greenhouse Gas Control 25 (2014) 1–8.
from 15% to 30% due to CO2 hydrate formation from weakly bound [18] T. Yu, T. Sato, T. Nakashima, M. Inui, H. Oyama, An integrated model for CO2
water. Differently, in illite containing lower level of strongly bound hydrate formation in sand sediments for sub-seabed CO2 storage, Int. J. Greenhouse
water than montmorillonite, CO2 was stored in gas hydrate formed Gas Control 52 (2016) 250–269.
[19] S. Almenningen, J. Gauteplass, P. Fotland, G.L. Aastveit, T. Barth, G. Ersland,
from the weakly bound water with the increase of water content from Visualization of hydrate formation during CO2 storage in water-saturated sand-
5% to 20%. The peak position of T2 distribution shifted to the left stone, Int. J. Greenhouse Gas Control 79 (2018) 272–278.
during CO2 hydrate formation in wet clay, indicating the ordered [20] A. Dutkiewicz, S. O'Callaghan, R.D. Müller, Controls on the distribution of deep-sea
sediments, Geochem. Geophys. Geosyst. 17 (2016) 3075–3098.
conversion of weakly bound water from outer layer to the inner layer. [21] J.-F. Li, J.-L. Ye, X.-W. Qin, H.-J. Qiu, N.-Y. Wu, H.-L. Lu, W.-W. Xie, J.-A. Lu,
The residual water that recognized as the strongly bound water with F. Peng, Z.-Q. Xu, C. Lu, Z.-G. Kuang, J.-G. Wei, Q.-Y. Liang, H.-F. Lu, B.-B. Kou, The
extremely low mobility cannot be fully converted to gas hydrate. In first offshore natural gas hydrate production test in South China Sea, China Geology
1 (2018) 5–16.
particular, the final amount of CO2 hydrate formed was dependent of
[22] K. Bauer, J. Kulenkampff, J. Henninges, E. Spangenberg, Lithological control on gas
the system pressure. The reason was that the binding strength in diffuse hydrate saturation as revealed by signal classification of NMR logging data, J.
double layer played an important role in lowering the activity of the Geophys. Res. Solid Earth 120 (2015) 6001–6017.
weakly bound water. The continuous growth of CO2 hydrate into deeply [23] G. Li, D.M. Wu, X.S. Li, Q.N. Lv, C. Li, Y. Zhang, Experimental measurement and
mathematical model of permeability with methane hydrate in quartz sands, Appl.
bound water layer is driven by the elevated pressure of gaseous CO2. Energy 202 (2017) 282–292.
Hence, the pore pressure had a significant influence on the capacity of [24] Z. Wu, Y. Li, X. Sun, M. Li, R. Jia, Experimental study on the gas phase permeability
CO2 hydrate storage in clayey sediments. These finding would be of montmorillonite sediments in the presence of hydrates, Mar. Pet. Geol. 91 (2018)
373–380.
helpful for site selection and design of CO2 storage in clay sediments. [25] J.P. Malizia, A. Shakoor, Effect of water content and density on strength and de-
formation behavior of clay soils, Eng. Geol. 244 (2018) 125–131.
Declaration of Competing Interest [26] L.-T. Chen, C.-Y. Sun, G.-J. Chen, Y.-Q. Nie, Thermodynamics Model of Predicting
Gas Hydrate in Porous Media Based on Reaction-Adsorption Two-Step Formation
Mechanism, Ind. Eng. Chem. Res. 49 (2010) 3936–3943.
The authors declare that they have no known competing financial [27] H. Li, L. Wu, H. Zhu, J. Hou, Ion Diffusion in the Time-Dependent Potential of the
Dynamic Electric Double Layer, J. Phys. Chem. C 113 (2009) 13241–13248.
interests or personal relationships that could have appeared to influ- [28] M.A. Mojid, H. Cho, Estimating the fully developed diffuse double layer thickness
ence the work reported in this paper. from the bulk electrical conductivity in clay, Appl. Clay Sci. 33 (2006) 278–286.
[29] G.J. Chen, T.M. Guo, A new approach to gas hydrate modelling, Chem. Eng. J. 71
(1998) 145–151.
Acknowledgement
[30] G.J. Chen, T.M. Guo, Thermodynamic modeling of hydrate formation based on new
concepts, Fluid Phase Equilib. 122 (1996) 43–65.
The financial support received from National Natural Science [31] J. Li, S. Wang, S. Lu, P. Zhang, J. Cai, J. Zhao, W. Li, Microdistribution and mobility
Foundation of China (No. 41872182 and 51991364) are gratefully ac- of water in gas shale: A theoretical and experimental study, Mar. Pet. Geol. 102
(2019) 496–507.
knowledged. [32] V.L. Mironov, A.A. Sukhovskii, Y.I. Lukin, I.P. Aleksandrova, Measurement of the
maximum content of bound water in bentonitic clay using dielectric and nmr
References methods, Russ. Phys. J. 54 (2011) 71–76.
[33] M.N. Testamanti, R. Rezaee, Determination of NMR T2 cut-off for clay bound water
in shales: A case study of Carynginia Formation, Perth Basin, Western Australia,
[1] P. Ciais, J. Tan, X. Wang, C. Roedenbeck, F. Chevallier, S.L. Piao, R. Moriarty, Journal of Petroleum Science and Engineering 149 (2017) 497–503.
G. Broquet, C. Le Quere, J.G. Canadell, S. Peng, B. Poulter, Z. Liu, P. Tans, Five [34] Y. Ji, J. Hou, G. Cui, N. Lu, E. Zhao, Y. Liu, Q. Du, Experimental study on methane
decades of northern land carbon uptake revealed by the interhemispheric CO2 hydrate formation in a partially saturated sandstone using low-field NMR tech-
gradient, Nature 568 (2019) 221–225. nique, Fuel 251 (2019) 82–90.
[2] M.J. Bos, D.W.F. Brilman, A novel condensation reactor for efficient CO2 to me- [35] A.H. Turakhanov, M.Y. Shumskayte, A.V. Ildyakov, A.Y. Manakov, V.G. Smirnov,
thanol conversion for storage of renewable electric energy, Chem. Eng. J. 278 V.N. Glinskikh, A.D. Duchkov, Formation of methane hydrate from water sorbed by
(2015) 527–532. anthracite: An investgation by low-field NMR relaxation, Fuel 116656 (2019).
[3] O. Hernández-Rivas, A. Martínez, H. Pfeiffer, J.A. Mendoza-Nieto, First insights for [36] M.G. Prammer, E.D. Drack, J.C. Bouton, J.S. Gardner, G.R. Coates, R.N. Chandler,
hydrogen production using an alkaline ceramic through the water-gas shift reac- M.N. Miller, Measurements of Clay-Bound Water and Total Porosity by Magnetic
tion, Chem. Eng. J. 392 (2020) 123740. Resonance Logging, SPE Annual Technical Conference and Exhibition, Society of
[4] K.Q. Zhang, N. Jia, L.R. Liu, CO2 storage in fractured nanopores underground: Petroleum Engineers, Denver, Colorado, 1996, pp. 10.
Phase behaviour study, Appl. Energy 238 (2019) 911–928. [37] A.J. Puppala, A. Pedarla, A. Pino, L.R. Hoyos, Diffused Double-Layer Swell
[5] P.M. Rendel, D. Wolff-Boenisch, I. Gavrieli, J. Ganor, A novel experimental system Prediction Model to Better Characterize Natural Expansive Clays, J. Eng. Mech. 143
for the exploration of CO2-water-rock interactions under conditions relevant to CO2 (2017) 04017069.
geological storage, Chem. Eng. J. 334 (2018) 1206–1213. [38] P. Wersin, E. Curti, C.A.J. Appelo, Modelling bentonite–water interactions at high
[6] Y. Teng, D. Zhang, Long-term viability of carbon sequestration in deep-sea sedi- solid/liquid ratios: swelling and diffuse double layer effects, Appl. Clay Sci. 26
ments, Science Advances 4 (2018) eaao6588. (2004) 249–257.
[7] L. Zhang, L. Yang, J. Wang, J. Zhao, H. Dong, M. Yang, Y. Liu, Y. Song, Enhanced [39] H. Wang, H. Qian, Y. Gao, Y. Li, Classification and physical characteristics of bound
CH4 recovery and CO2 storage via thermal stimulation in the CH4/CO2 replacement water in loess and its main clay minerals, Eng. Geol. 265 (2020) 105394.
of methane hydrate, Chem. Eng. J. 308 (2017) 40–49. [40] J.R. Brooks, Water, bound and mobile, Science 349 (2015) 138.

9
Y. Sun, et al. Chemical Engineering Journal 406 (2021) 126872

[41] G. Li, D. Liu, Y. Xie, Y. Xiao, Study on Effect Factors for CO2 Hydrate Rapid (2018) 262–285.
Formation in a Water-Spraying Apparatus, Energy Fuels 24 (2010) 4590–4597. [46] Y. Zhang, X.-S. Li, Z.-Y. Chen, Z.-M. Xia, Y. Wang, G. Li, Experimental and modeling
[42] L. Chong, E.M. Myshakin, The effect of residual water content on preferential ad- study on controlling factor of methane hydrate formation in silica gels, Appl. Energy
sorption in carbon dioxide – methane – illite clay minerals: A molecular simulation 225 (2018) 827–834.
study, Fluid Phase Equilib. 504 (2020) 112333. [47] A.G. Aregbe, B. Sun, L. Chen, Methane Hydrate Dissociation Conditions in High-
[43] A. Busch, S. Alles, Y. Gensterblum, D. Prinz, D.N. Dewhurst, M.D. Raven, H. Stanjek, Concentration NaCl/KCl/CaCl2 Aqueous Solution: Experiment and Correlation, J.
B.M. Krooss, Carbon dioxide storage potential of shales, Int. J. Greenhouse Gas Chem. Eng. Data 64 (2019) 2929–2939.
Control 2 (2008) 297–308. [48] D.-J. Lee, S.F. Lee, Measurement of bound water content in sludge: The use of
[44] S. Kumar, Rock-fluid interaction and phase properties of fluids in nano- and sub- differential scanning calorimetry (DSC), J. Chem. Technol. Biotechnol. 62 (1995)
nano-pores of shales: Sorption-based studies, Ph.D. Thesis (2016). 359–365.
[45] H.P. Veluswamy, A. Kumar, Y. Seo, J.D. Lee, P. Linga, A review of solidified natural [49] S. Guggenheim, A.F.K. van Groos, New gas-hydrate phase: Synthesis and stability of
gas (SNG) technology for gas storage via clathrate hydrates, Appl. Energy 216 clay-methane hydrate intercalate, Geology 31 (2003) 653–656.

10