Chemical Engineering Journal 425 (2021) 130292

Contents lists available at ScienceDirect

Chemical Engineering Journal

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

Molecular insights into carbon dioxide enhanced multi-component shale

gas recovery and its sequestration in realistic kerogen
Sen Wang a, b, *, Xinyu Yao a, b, Qihong Feng a, b, Farzam Javadpour c, Yuxuan Yang a, b,
Qingzhong Xue d, Xiaofang Li d
a
School of Petroleum Engineering, China University of Petroleum (East China), Qingdao 266580, China
b
Key Laboratory of Unconventional Oil & Gas Development, Ministry of Education, Qingdao 266580, China
c
Bureau of Economic Geology, Jackson School of Geosciences, The University of Texas at Austin, University Station, Box X, Austin, TX 78713, USA
d
State Key Laboratory of Heavy Oil Processing, School of Material Science & Engineering, China University of Petroleum (East China), Qingdao 266580, China

A R T I C L E I N F O A B S T R A C T

Keywords: Understanding the governing processes of CO2 huff-n-puff in shales is essential for enhancing shale gas recovery
Shale gas and CO2 sequestration. However, existing studies have not fully accounted for the chemical composition and
Adsorption geometry of kerogen nanopores and the reality that natural gas is a multi-component mixture. We used grand
Kerogen
canonical Monte Carlo (GCMC) simulations to study the competitive adsorption between CO2 and typical hy­
CO2 sequestration
Molecular simulation
drocarbon components (CH4, C2H6, and C3H8). We further studied the recovery mechanisms of CO2 huff-n-puff
within kerogenic circular nanopores at reservoir conditions. We probed the effects of pressure, pore geometry,
and size on gas recovery and CO2 sequestration efficiency. Although pressure drop readily exploits CH4 in the
adsorption layer, the recovery due to CO2 injection primarily occurs within the kerogen matrix for pure CH4.
Injecting CO2 facilitates the recovery of heavier hydrocarbons, whereas pressure drawdown exhibits better
performance for lighter components. CO2 huff-n-puff may serve as a promising method for gas exploitation in
circular pores, whereas pressure drop favors the production in kerogen slit. The tremendously different gas
adsorption and recovery behavior in distinct pore geometries and compositions necessitate the study using
realistic shale kerogen models. Moreover, enlarging the pore size improves the recovery of each component
during pressure drawdown but restrains the performance of CO2 injection; meanwhile, the total gas recovery and
CO2 sequestration efficiency increase. This study provides a more in-depth understanding of multi-component
gas recovery mechanisms within realistic shale kerogen nanopores and sheds light on the CO2 sequestration in
shale reservoirs.

1. Introduction
Shale gas has emerged as an important alternative clean energy
source to conventional fossil fuels because of its low emissions and high
energy efficiency [1-4]. Recently, the exploration and development of
shale gas have attracted global focus. As reported by the Energy Infor­
mation Administration (EIA), gas production in the Eagle Ford shale of
the United States has surged from 0.119 billion cubic feet per day (Bcf/
d) in 2010 to 4.405 Bcf/d in 2020 [5]. Meanwhile, China, Australia, and
several European countries have started strong programs to exploit hy­
drocarbons in shale [6]. The permeability of shale matrix is 6 to 9 orders
of magnitude smaller than that of conventional reservoirs, leading to the
rapid decrease of well production rates [7-9]. Therefore, how to improve
the shale gas recovery using advanced technologies has been an urgent
issue to be solved [10,11].
Injecting CO2 into shale formations (through huff-n-puff) is recog­
nized as a promising solution because it enhances shale gas productivity
and mitigates global warming via geological CO2 sequestration [12-16].
CO2 huff-n-puff generally contains three procedures: CO2 injection
(huff), soaking, and pressure drawdown (puff). There are recent studies
on the transport of gas mixtures in nanomaterials, such as carbon
nanotube (CNT) and metal–organic framework (MOF) [17-22]. How­
ever, the underlying mechanisms are different in amorphous nano­
porous kerogen that motivates a thorough study of the CO2 huff-n-puff in
shales.
Some experimental studies have been reported on the competitive

* Corresponding author.
E-mail address: fwforest@gmail.com (S. Wang).

https://doi.org/10.1016/j.cej.2021.130292
Received 20 February 2021; Received in revised form 4 May 2021; Accepted 7 May 2021
Available online 21 May 2021
1385-8947/© 2021 Elsevier B.V. All rights reserved.
S. Wang et al. Chemical Engineering Journal 425 (2021) 130292

adsorption of CH4 and CO2 in shales [23,24]. Using a manometric setup,

Khosrokhavar et al. [23] found the preferential adsorption of CO2 over
CH4 on a carboniferous shale from Belgium. The experimental results of
Charoensuppanimit et al. [24] on Woodford and Caney shales samples
suggested that the ratio of adsorption amount between CO2 and CH4
may vary from 2.1 to 8.6 depending on the total organic carbon (TOC).
Note that the produced gas from shale formations contains heavier hy­
drocarbon components, e.g., up to 20% molar fraction of C2H6 and C3H8
[25]. Hence, the adsorption of multi-component alkanes in shale is
becoming a focus of recent studies [26-30]. Cheng and Huang [26]
measured the adsorption selectivity of C1–C6 components in clays, coal,
and shale at pressures below 0.3 MPa. Kong et al. [27] used the ther­
mogravimetric method to study the excess adsorption/desorption iso­
therms of C1 and C2 on shales and estimated the absolute adsorption at
different temperatures. The sorption of light hydrocarbons (C1-C4) in
shale samples and isolated kerogens has also been explored by Fir­
oozabadi and his coworkers [29,30]. However, owing to the heteroge­
neous compositions of shale samples, it is complicated—through
experimental measurements—to reveal the molecular interactions be­
tween hydrocarbon mixtures and CO2 under confinement and elucidate
Fig. 1. Scanning electron microscopy (SEM) image showing the morphology of
the underlying mechanisms of CO2 injection. Moreover, the pressure and
organic nanopores in shale. The experimental sample is adopted from the
temperature of a realistic shale reservoir are too high [31], which re­ Longmaxi shale of Sichuan Province, China. The black and gray areas represent
quires more sophisticated equipment to tackling the challenge. the organic matter (OM) and inorganic minerals, respectively. The darkest spots
Therefore, molecular simulations, such as grand canonical Monte in the OM correspond to organic nanopores.
Carlo (GCMC) and molecular dynamics (MD), have been utilized as a
viable and powerful tool to provide deeper insights into the gas
adsorption behavior at the nanoscale [32-34]. The results can not only Table 1
complement the experimental analyses but also be beneficial for the Comparison of the molecular models used in previous literatures and this work.
equipment design. For example, on the basis of Configurational-biased No. References Surface Fluid Simulation Limitations
GCMC and MD, Falk et al. [33] probed the effect of the molecular method
length (from C1 to C12) and pore accessibility on the adsorption of n- 1 Zhang et al. Pure CH4, C2H6, GCMC Cannot
alkane in kerogen matrix. Tesson and Firoozabadi [32] created kerogen [41] carbon C3H8, and account for the
slits having different surface roughness and analyzed their effect on slit, K- CO2 influences of
Illite slit heteroatoms in
methane adsorption and self-diffusion. Using MD, Wu et al. [34] studied
2 Wu et al. Pure CH4 and MD kerogen
the recovery process of methane and ethane mixtures in nanopores [34] carbon C2H6
under a pressure gradient drive. They suggested that ethane shows su­ slits
perior adsorption capacity over methane, and this phenomenon is more 3 Ho et al. Kerogen CH4, CO2, MD
pronounced in narrow pores. Recently, Wang et al. [35] simulated the [40] matrix, H2O
CNTs,
competitive adsorption of CH4 and C2H6 in clay nanopores and reported Kerogen
that the favorable adsorption of C2H6 fails at elevated pressures. slit
However, most of the existing studies only concentrate on the 4 Kazemi Kerogen CH4, C2H6, GCMC, MD Pores within
adsorption of a pure-component alkane or its competitive adsorption et al. [61] matrix H2O, and the kerogen
CO2 matrix is too
with CO2 in shales. In contrast, the studies on the interactions between
5 Tesson et al. Kerogen CH4 GCMC small to
hydrocarbon mixtures and CO2 are scarcely reported [36-45]. Huang [32] matrix, represent a
et al. [36,37] simulated the competitive adsorption between CH4 and Kerogen realistic shale
CO2 on dry and moist kerogen matrix and highlighted the more slit kerogen, and
remarkable impact of moisture on CO2. Similar results were also re­ 6 Huang et al. Kerogen CH4, CO2, GCMC the organic
[36,37] matrix H2O shale pores are
ported by Jin et al. [38]. GCMC simulations were performed by Vasi­
7 Vasileiadis Kerogen CH4, C2H6, GCMC usually
leiadis et al. [39] to study the adsorption of a quaternary mixture (C1, C2, et al. [39] matrix n-C4H10 and circular
CO2, and N2) in the kerogen matrix at various temperatures and pres­ CO2 instead of slit-
sures. They suggested that the confinement effect plays a dominant role 8 Yang et al. Kerogen Multi- MD shaped.
[62] slit component
in the mixture adsorption. Some other simulations have been conducted
shale oil
to study CO2 sequestration and enhanced gas recovery at the nanoscale. 9 Zhou et al. Kerogen CH4, C2H6, GCMC
Ho et al. [40] used carbon nanotubes (CNTs) to mimic the kerogen [38,42] slit C3H8, and
nanopore and studied the retention and release of CH4 and CO2. They CO2
reported that CO2 could diffuse through the water layer in the kerogen 10 Sun et al. Circular CH4 MD Only studied
[63] kerogen the adsorption
matrix and facilitate methane desorption. Zhang et al. [41] used pure-
nanopore and transport
carbon and K-illite to construct the organic and inorganic slits, respec­ of CH4 within
tively, to probe the recovery mechanisms of multi-component shale gas organic
during CO2 injection. Zhou et al. [42] extended this study to kerogen nanopores
11 This work Circular CH4, C2H6, GCMC /
slits and observed a more favorable performance of enhanced gas re­
kerogen C3H8, and
covery in narrower slits. nanopore CO2
Even though several advances have been achieved to understand the
recovery mechanisms of CO2 huff-n-puff and sequestration in shale gas
reservoirs from the molecular perspective, there still exists three unre­
solved problems. First, the pore diameter (<1.4 nm) of the kerogen

2
S. Wang et al. Chemical Engineering Journal 425 (2021) 130292

Fig. 2. (a) Molecular model of the II-D kerogen unit; (b) equilibrium configuration of the kerogen-based circular nanopores. Color code: black, C; gray, H; red, O;
blue, N; yellow, S. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

matrix models, which were widely employed to study shale gas

Table 2
adsorption and transport [46-50], is much smaller than that of realistic
NVT/NPT relaxation procedure to create the kerogen matrix.
kerogen (2–100 nm). Considering the tremendous impact of pore size on
gas storage and diffusion, more realistic kerogen models are required to No. Ensemble T (K) Time (ps) P (MPa)

clarify the actual behavior of shale gas. Second, recent studies on the
shale pore structure characterization via advanced techniques, such as
atom force microscopy (AFM) [51] and scanning electron microscope
(SEM) [52], suggested that the pores located within inorganic minerals
of shales are primarily slit-shaped [53,54]. In contrast, the organic pores
tend to exhibit circular cross-sectional geometry owing to the thermal
conversion of kerogen into hydrocarbons [55], as shown in Fig. 1.
However, most of the relevant studies focused on kerogen slit
[38,41,56]. A multi-layer adsorption model developed by Qajar et al.
[57] suggested that shale was better characterized using a network of
interconnected cylindrical pores. Moreover, because of the greater
sorption energy and the larger surface area, cylindrical kerogen pore
dominates the adsorption capacity of shale gas [58], which necessitates
the study in kerogen pores having circular geometry. The last problem
lies in the influence of heteroatoms. Carbon nanotubes (CNT) have been
used as the shale skeleton without accounting for the heteroatoms (i.e.,
Oxygen, Nitrogen, and Sulphur) in kerogen [40,59,60]. To manifest the
differences of this work from current literature, we summarize the
characteristics of the molecular models used in relevant studies in
Table 1.
Combining MD and GCMC simulations, we studied the recovery
mechanisms of pure CH4 and hydrocarbon mixture (C1/C2/C3) confined
in kerogen-based circular nanopores through CO2 huff-n-puff and pres­
sure drawdown. We discuss the effects of pressure, pore geometry, and
pore size on hydrocarbon recovery and CO2 sequestration efficiency. To
the best of our knowledge, this work is the first attempt to explore the
physics of pressure drawdown and CO2 injection in shale kerogen with
circular nanopores and multi-component gas.
2. Models and methods
2.1. Construction of Kerogen-based circular nanopore
On the basis of the comprehensive experimental results from Kele­
men et al. [64], Ungerer et al. [65] developed molecular structures of
kerogen units having distinct thermal maturities. We used the II-D
kerogen unit (chemical formula: C175H102O9N4S2) [65] to construct
1 NVT 900 400 –-
2 NPT 900 200 20
3 NPT 700 200 20
4 NPT 500 200 20
5 NPT 300 400 20
the circular organic nanopore (Fig. 2). This type of kerogen corresponds
to typical unconventional reserves with high potential in producing
shale gas such as the Barnett shale [66]. The detailed composition and
structural parameters of the kerogen unit match fairly well with the
experimental data of solid-state 13C nuclear magnetic resonance (NMR)
and X-ray photoelectron spectroscopy [64]. Besides, II-D kerogen has
been employed to study the competitive adsorption of CO2 and CH4 in
shales [32,38,40].
The technique proposed by Collell et al. [67] is utilized to construct
the molecular model of kerogen matrix and then cut out a cylindrical
pore. This process is achieved through a series of MD simulations in the
NVT (canonical) and NPT (isobaric-isothermal) ensembles [37,65].
First, we randomly put 72 relaxed kerogen units (a total of 21,024
atoms) into a box with an initial density of 0.1 g/cm3 [67,68], and
performed 400 ps NVT simulations at 900 K to relax the structure. Then,
we conducted NPT simulations with four stepwise decreasing tempera­
tures from 900 K to 700 K, 500 K and eventually 300 K at 20 MPa. For
each step, the duration of our simulation is long enough to reach the
equilibration.
Consequently, we obtained the stable molecular structures of the
kerogen matrix. We summarized the whole simulation procedure in
Table 2 and showed the variation of kerogen density in Fig. 3. As the
temperature decreases, the kerogen is condensed and approaching the
stratigraphic conditions. In Section 3, we will show that the density of
the kerogen matrix agrees well with the experimental measurements,
which justify the validity of our model. Then we deleted those atoms
located within a cylinder having the predefined diameter and added
hydrogens atoms to the dangling bonds to get the circular kerogen
nanopore, as shown in Fig. 2b.

3
S. Wang et al.
Fig. 3. Variation of the kerogen density during the construction of kerogen matrix. The snapshots at different time are also included to show the compression process.
2.2. Force field
Three light hydrocarbons (CH4, C2H6, and C3H8) are taken into ac­
count because of their high proportions in shale gas reservoirs such as
Barnett and Marcellus [69]. According to the data from the Mississip­
pian Barnett Shales [70], our hydrocarbon mixture is composed of C1,
C2, and C3 with the bulk mole ratio of 80:15:5. A widely used all-atom
force field, COMPASS (Condensed-phase Optimized Molecular Poten­
tials for Atomistic Simulation Studies) [71], is employed to estimate the
intramolecular interactions of kerogen and alkanes. This force field
describes the dispersion-repulsion energy and the Coulombic in­
teractions through Lennard-Jones 6–9 potential and point charges.
COMPASS has been successfully employed by Huang et al. [36] to study
the competitive adsorption of CH4 and CO2 within kerogen matrix. We
maintain the temperature constant using the Nose/Hoover thermostat. A
cutoff distance of 12.5 Å is used when estimating the van der Waals and
electrostatic interactions.
2.3. Simulation details
We explore two different processes of hydrocarbon production
(pressure drawdown and CO2 injection) at constant temperature (T =
353 K) to assess the efficiency of shale gas recovery and CO2 storage
[38]. The initial pore pressure is P0 = 30 MPa, which is given according
to the realistic reservoir condition. First, we reduce the pore pressure
from P0 = 30 MPa to P1 = 15 MPa to mimic the shale gas production
caused by primary pressure drawdown, supposing that the hydrocarbon
mixture in pores is in chemical equilibrium with that in an infinite
volume. Then, we inject CO2 into the kerogen nanopore (huff) until the
mole fraction of CO2 reaches 25%. During this process, because the pore
volume in the external bulk reservoir—which acts like the fractures
connected to the nanopores–maintains constant, the density of each
component in the bulk mixture of C1/C2/C3/CO2 at the equilibrium
pressure P2 is equal to that in C1/C2/C3 mixture at P1. Thus, according to
the molar composition of the mixture after CO2 injection, we assume
different pressures P2 for this system and calculate the alkane densities
using the Peng-Robinson equation of state (PR-EOS) until the density of
each component at P2 is equal to that at P1 [72]. Using this method, we
estimate P2 = 19.1 MPa. Then another pressure drawdown process is
conducted from P2 to P3 = 10 MPa while the bulk composition remains
4
Chemical Engineering Journal 425 (2021) 130292
fixed [41]. After shale gas production, we recorded the composition of
confined fluid at the equilibrium state. In Fig. 4, we show the simulation
snapshots in 2-nm kerogen-based nanopores at different stages.
We estimate the total amount of shale gas molecules in the kerogen
nanopore at each simulation stage through GCMC [73,74]. Molecular
dynamics is not employed here because GCMC is much computationally
cheaper to study the equilibrium properties, from which the ultimate gas
recovery can be readily estimated [41]. MD is more favorable to simu­
late the dynamic process and time-dependent properties because each
new configuration is obtained by solving Newton’s second law. How­
ever, GCMC generates an ensemble of configurations under specific
conditions (fugacity) through random perturbations, from which the
relevant thermophysical properties may be estimated although the time-
evolving information cannot be obtained [75]. Thus, the application of
GCMC saves the computational resources required for gas adsorption
simulation. Random insertion, translation, rotation, and deletion moves
are implemented for the gas molecules in each GCMC cycle [76]. As
suggested by Vasileiadis et al. [39] and Jin et al. [38], the kerogen
structure is kept rigid during our simulations. For each pore pressure, we
first calculate the fugacity and then estimate the number of gas mole­
cules contained in the nanopore using GCMC. In other words, the pore
pressure is controlled by the different numbers of gas molecules, whose
amounts are determined by GCMC. A periodic boundary condition is
imposed in each direction. For each point, our simulation consists of 1 ×
107 steps, in which the first half is performed for equilibrium, while the
other steps are utilized to estimate the gas storage. The absolute
adsorption amount of hydrocarbons in the nanopore Nab is given by
〈Ni 〉
Nab = (1)
V⋅NA
Here, <Ni > is the number of gas molecules in our kerogen model;
NA is the Avogadro constant, i.e., 6.022 × 1023 1/mol; and V denotes the
accessible pore volume of the kerogen. We use helium as the probe to
detect V because its atom is small enough to probe the pores, and it is
hardly adsorbed on kerogen [77,78].
A parameter introduced by Zhou et al. [42], CE, is used to evaluate
the confinement effect on the different types of hydrocarbons,
yi,pore
CE = (2)
yi,bulk
S. Wang et al. Chemical Engineering Journal 425 (2021) 130292

Fig. 4. Simulation snapshots showing the recovery of C1/C2/C3 mixture during CO2 huff-n-puff in a 2 nm kerogen nanopore (T = 353 K): (a) initial reservoir (P0 = 30
MPa), (b) after pressure drawdown (P1 = 15 MPa), (c) CO2 huff (P2 = 19.1 MPa), and (d) CO2 puff (P3 = 15 MPa).

where yi pore and yi bulk denote the molar fraction of component i within 3. Model validation
kerogen nanopores and in the bulk fluid, respectively. A value of CE
greater than 1 means a stronger affinity. Prior to the GCMC simulations, we validate our model of kerogen
To quantify the efficiencies of pressure drawdown and CO2 injection, nanopore. The compositions and functional groups of the kerogen unit
we define the hydrocarbon recovery ηi as the ratio of released gas agree well with the experimental data of Kelemen et al [64]. The density
quantity to the initial adsorption amount during a single-stage of our kerogen matrix is 1.24 ± 0.01 g/cm3, which is consistent with the
[41,79,80], experimental results (1.18–1.25 g/cm3) of Type II kerogen reported by
Okiongbo et al. [81]. To estimate the porosity, we detected the free pore
nreleased volume using a helium molecule with a given Connolly radius as the
ηi = i
× 100% (3)
ninitial
i probe. The probe molecules inserted into the kerogen roll on the van der
Waals surface of the skeleton, from which the area wrapped by the
where nreleased
i is the amount of component i recovered from the kerogen, surface is defined as the free pore volume [82]. The estimated porosity
and ninitial
i is the initial adsorption amount of component i before of the kerogen matrix (19.17 ± 0.1%) also lies in the range of mea­
production. surement values (4.45%–22.50%) in the Barnett shale [55].
Moreover, we also computed the excess adsorption capacity of
methane Г ex using GCMC to make a comparison with experimental data,

5
S. Wang et al.
Fig. 5. Comparison of the simulated excess adsorption isotherms in a 1.2 nm
kerogen-based nanopore (T = 338 K) and the experimental results reported by
Gasparik et al. [51] (Barnett shale) and Zhang et al. [84] (Blakely shale). The
samples from Barnett shale have distinct total organic carbon contents (TOCs)
and vitrinite reflectance values (VRs): VR = 2.2%, TOC = 3.5% (red spheres)
and VR = 1.0%, TOC = 4.5% (blue triangles). (For interpretation of the ref­
erences to color in this figure legend, the reader is referred to the web version of
this article.)
〈Ni 〉 ρbulk V
Γex = − (4)
NA M adsorption of C3H8 and CO2 upon the kerogen surface. Meanwhile, the
In this equation, M represents molar mass (g/mol); and ρbulk (g/cm3)
is the density of the bulk fluid, which is obtained from the National
Institute of Standards and Technology (NIST) database [83].
In Fig. 5, we compare the estimated excess adsorption isotherms of
CH4 within our model and the experimental results. The experimental
data are adopted from the Barnett shales measured by Gasparik et al.
[51] and Blakely shale measured by Zhang et al. [84]. Fig. 5 indicates
that the simulated and measured data exhibit the same trend, which first
increases quickly and then reaches a plateau. Furthermore, the esti­
mated excess adsorption capacity is also reasonably close to the exper­
imental data. The deviations may be attributed to the discrepancy of
sample properties, such as the pore radius. Bae et al. [85,86] suggested
that the impenetrable pore necks existing in the realistic shale kerogen
may also contribute the differences. Thus, the validity of our constructed
kerogen nanopore has been justified and we will use it to study the re­
covery of shale gas during pressure drawdown and CO2 injection.
4. Results and discussion
4.1. Adsorption and recovery of CO2-hydrocarbon mixtures within
kerogen nanopore
To investigate the effect of CO2 injection on the shale gas recovery,
the adsorption behavior of pure and multi-component alkanes in
kerogen nanopores is analyzed first. We probe the adsorption behavior
of pure CH4 within kerogen-based nanopores under the conditions of
either CO2 presence or absence at 353 K using the GCMC method. As
shown in Fig. 6a, the absolute adsorption amount of CH4 increases
gradually with the increasing pressures and tends to approach a con­
stant. This trend is consistent with the adsorption characteristics of CH4
in other nanoporous materials, such as CNTs [87-89] and clays
[35,90,91]. According to the recommended classification of IUPAC (the
International Union of Pure and Applied Chemistry), the isotherm cor­
responds to the Langmuir type—a typical behavior for nanoporous
6
Chemical Engineering Journal 425 (2021) 130292
adsorbent [92].
After CO2 injection, there are both CH4 and CO2 molecules confined
in the kerogen nanopore. Therefore, the selectivity of CO2 against CH4,
SCO2/CH4, is introduced to quantify the competitive adsorption [35],
xCO2 /xCH4
SCO2 /CH4 = (5)
yCO2 /yCH4
where, xi and yi denote the mole fractions of component i (CO2 or CH4) in
the adsorbed and bulk phases, respectively. They can be readily esti­
mated from GCMC simulations. If the selectivity is greater than 1, CO2 is
preferentially adsorbed [93]. A higher selectivity signifies a stronger
adsorption capacity of CO2 with respect to CH4. In Fig. 6c, we show the
selectivity corresponding to the bulk mole fraction of yCO2 = 0.25. The
selectivity is greater than unity but decreases with pressure, indicating a
more considerable amount of CO2 to be adsorbed on the kerogen and the
adsorption capacity differences between CO2 and CH4 become smaller at
elevated pressure. A similar trend has also been reported for CO2/CH4
mixtures within coal and kerogen matrix [36,44]. Because the energy of
kerogen surface is heterogeneous, the adsorption sites with higher en­
ergy are favorably occupied by CO2 at low pressure; however, with the
pressure increasing, once the higher energy positions have been filled,
CO2 and CH4 begins to competitively occupy the lower energy positions,
thus leading to the smaller selectivity at high pressures.
Fig. 6b shows the absolute adsorption isotherms of each component
in CH4/C2H6/C3H8 mixtures at 353 K. As the pressure increases, the
adsorption capacity of CH4 is enhanced, whereas the adsorption amount
of C2H6 and C3H8 remains almost unchanged. To better elucidate the
adsorption behavior, the parameter, CE (Eq. (2)), is used for the simu­
lations involved with gas mixtures. Fig. 6d reveals that the CE of C3H8
and CO2 are always greater than 1, indicating that the preferential
CE value of C3H8 decreases rapidly with pressure, and it is even smaller
than that of CO2 when P greater than 20 MPa, because C3H8 confined in
the nanopore saturates at lower pressure. In contrast, the density of the
bulk phase continually increases. Similar behavior is also observed for
C2H6. The variation of pore pressure shows a negligible effect on the CE
of CO2, revealing that CO2 always exhibits a superior adsorption ca­
pacity in the mixture.
To disclose the underlying mechanisms of competitive adsorption,
we calculate the isosteric heat, Qst, using the formula developed by
Karavias and Myers [94]. The Clausius-Clapeyron equation [95] was not
adopted here because it assumes the adsorbed volume is negligible and
the bulk phase is regarded as ideal gases [96], which contradicts the
high-pressure condition. As shown in Fig. 7a, the isosteric heat of CO2 is
approximately 1.54 times greater than that of CH4, manifesting a
stronger affinity between kerogen and CO2. This result explains the
preferential adsorption of CO2 over CH4 within the kerogen nanopore
(Fig. 6c). The isosteric heat of C3H8 is much larger than that of CH4 and
C2H6, suggesting that longer alkane has much superior interaction with
kerogen (Fig. 7b). This result gets in line with the recent work of Liu
et al. [97], who reported that the heavier component is strongly retained
by kerogen matrix. We also note that the isosteric heat of adsorption for
CO2 is smaller than that of C3H8 but greater than CH4 and C2H6. Because
the isosteric heat extrapolated to the limit of zero coverage characterizes
the interactions between adsorbate and adsorbent, we infer that the
affinity of these components with kerogen decreases in the order: C3H8,
CO2, C2H6, and CH4. This conclusion agrees well with the resulting CEs
at lower pressures (Fig. 6d). However, as the pressure increases, the
interactions among adsorbate molecules become stronger and the larger
molecule sizes of C3H8 cause a greater repulsive force, thus leading to
the restrained adsorption capacity of C3H8 in comparison to CO2
(Fig. 6d).
We further study the recovery process of shale gas using CO2 huff-n-
puff. Both the pure CH4 and the CH4/C2H6/C3H8 mixture are taken into
account. In Fig. 8, we present the number density distributions of each
S. Wang et al. Chemical Engineering Journal 425 (2021) 130292

Fig. 6. (a) Absolute adsorption isotherms, Nab, of pure CH4 in a 2 nm kerogen-based nanopore at T = 353 K; (b) Absolute adsorption isotherms of CH4/C2H6/C3H8
mixture; (c) adsorption selectivity of the CH4/CO2 binary mixture; (d): Confinement effect of each component in the CO2/CH4/C2H6/C3H8/ mixture.

Fig. 7. Variation of isosteric heats (Qst) against the pore pressure: (a) CH4/CO2, (b) CH4/C2H6/C3H8, and (c) CO2/CH4/C2H6/C3H8.

component in 2-nm kerogen nanopores. The green dashed lines repre­
sent the interface boundary between the cylinder pore and kerogen
matrix. A pronounced density peak (also termed adsorption layer
[98,99]) forms at the interfacial area, and some gas molecules absorb
within the kerogen matrix. Dominated by the inherent surface proper­
ties, the location of adsorption layers remains unchanged.
Fig. 8a shows the variation of CH4 number density profiles during the
whole production process. The pressure variations from P0 to P1, P1 to
P2, and P2 to P3 correspond to the pressure drawdown stage, CO2 huff
and puff stages, respectively. The reduction of CH4 adsorption peak
during pressure drawdown is greater than that during CO2 injection,
indicating that decreasing the pore pressure primarily produces adsor­
bed CH4 molecules. This phenomenon should be attributed to the
weaker adsorption capacity at lower pressures (Fig. 6a). Meanwhile, for
the adsorbed gas within the kerogen matrix, the CH4 quantity produced
during CO2 injection is greater than that exploited during pressure
drawdown, demonstrating that the adsorbed gas is favorable to be
extracted using CO2 injection. Because the tiny space located within the

7
S. Wang et al. Chemical Engineering Journal 425 (2021) 130292

Fig. 8. Number density profiles of each hydrocarbon and CO2 in a 2-nm kerogen nanopore under different stages (T = 353 K): (a) pure CH4; (b) CH4/C2H6/
C3H8 mixture.

kerogen matrix is much smaller than the outside (cylinder) pore,
CO2—which shows a much superior adsorption capacity in smaller
pores—preferentially occupies the sorption sites of CH4 and lead to the
gas release [38]. This conclusion is also supported by Huang et al. [36],
who reported that CO2 and CH4 are favorably adsorbed upon the inner
and outer surfaces of the kerogen matrix, respectively. During the CO2
puff stage (P2 → P3), there is a decrease in the CO2 adsorption layer, but
no significant variation is observed for the CO2 density profile in the
kerogen matrix (Fig. 8a, right panel). It reveals that due to the strong
fluid–solid interactions, CO2 stored in the kerogen matrix is difficult to

8
S. Wang et al. Chemical Engineering Journal 425 (2021) 130292

Fig. 9. Absolute adsorption isotherms of pure CH4 in CNT, kerogen slit, and kerogen-based nanopore having the same dimension (2 nm).

molecules tend to move from the kerogen matrix to the surface in the
second pressure drop. As mentioned before, C3H8 shows a much stronger
affinity to kerogen than CH4, C2H6, and CO2. Thus, during this stage,
when shorter alkanes are released from the matrix, some C3H8 molecules
enter the kerogen matrix, which is demonstrated by the higher adsorp­
tion density of C3H8 at P3 than P2 (lower left panel, Fig. 8b). Therefore,
these C3H8 molecules occupy the original adsorption sites of CO2 within
the matrix and drive CO2 molecules into the inner pore.

4.2. Effect of pore structure on shale gas recovery

Existing studies always use CNTs [89], graphene slit [41,100], and
kerogen slit [42] to mimic the shale kerogen to study the adsorption and
transport of hydrocarbons. To examine the validity of these models, in
this section, we study the effect of pore structure on the gas recovery
using CO2 huff-n-puff. Owing to the similarity of chemical composition
between CNT and graphene, we account for three different models:
kerogen slit (Fig. 9a), CNT (Fig. 9b), and kerogen-based nanopore. The
kerogen slit is composed of two kerogen cubic matrices, each of which
consists of 12 kerogen units. The aperture of the kerogen slit is 2 nm,
which is as same as the diameters of CNT and our kerogen nanopore. The
Fig. 10. Molecular structure of (a) kerogen slit and (b) CNT having the same
pore size. estimated pore volumes for kerogen slit, carbon nanotube, and circular
kerogen nanopore are 20.413 nm3, 19.195 nm3, and 28.067 nm3,
respectively. Because our kerogen-based nanopore consists of a 2-nm
exploit; organic-rich shale shows good potential for CO2 sequestration.
cylindrical pore and a kerogen matrix, the available pore volume is
Meanwhile, adsorbed CO2 molecules are released from the surface
greater than that of CNT even they have the same inner diameter.
because of the relatively smaller interactions in the inner pore.
Firstly, we compute the adsorption isotherms of pure CH4 in the
In Fig. 8b, we show the density profile of each component when CO2
three models. As shown in Fig. 10, all the absolute adsorption isotherms
is injected into the CH4/C2H6/C3H8 mixture in kerogen. As the pressure
increase monotonically with growing pressure and approach a plateau at
drops (P0 → P1), the density of CH4 in the adsorption layer decreases, but
higher pressure. However, when the pressure is lower, the adsorption
C2H6 doesn’t show significant variations in the density profiles. How­
isotherms within different geometries show distinct increasing rates.
ever, we observe the enrichment of C3H8 near the interfacial region at
The adsorption capacities in kerogen slit and cylinder pore are almost
lower pressures, which is in contrast to CH4 (Fig. 8b, lower-left panel).
identical when P < 8 MPa but much smaller than that in a carbon
We attribute that the adsorption sites on the kerogen surface are not
nanotube. Besides, the maximum adsorption amount in CNT also ex­
fully occupied at low pressures; therefore, the stronger interactions be­
ceeds that in kerogen nanopores. For example, the adsorption capacity
tween kerogen and C3H8 cause the preferential absorption of more C3H8
of CH4, Nab, in CNT is 0.00842 mmol/m2 at P = 40 MPa, which is ~ 1.35
molecules upon the interface [33], leading to an increasing density.
times as larger as that in kerogen nanopores (0.00623 mmol/m2). Both
However, at elevated pressures, the adsorption sites are almost filled,
the surface roughness and chemical composition contribute to this
and the entropic effect becomes more pronounced, causing a decreased
phenomenon [32]. On the one hand, the CNT surface is ultrasmooth,
density of C3H8 [42]. After CO2 injection, the amount of the adsorbed
whereas the surface of kerogen nanopores is rough. Tesson et al. [32]
hydrocarbons decreases, and more carbon dioxide molecules are
suggested that a smooth slit adsorbs 42% gas molecules more than a
adsorbed in the kerogen because of the greater attractions. Fig. 8b also
rough slit. On the other hand, the carbon nanotube is totally composed
suggests that the CH4 recovery during CO2 injection (P1 → P2) is
of carbon atoms, CH4 can be readily adsorbed anywhere on the surfaces;
significantly lower than that obtained during pressure drawdown (P0 →
however, the composition of kerogen matrix is intricate, including a
P1); however, the variation of C2H6 shows the opposite trend. Further­
large number of hydrogen and carbon atoms and a few nitrogen, sulfur,
more, the number of C3H8 molecules in the adsorption layer and kerogen
and oxygen atoms. These heteroatoms exert a distinct interaction on the
matrix decreases remarkably (P1 → P2), which is contrary to the stages of
gas molecules compared to the pure carbon skeleton [47,100]. Fig. 10
pressure drawdown with CO2 presence (P2 → P3) or absence (P0 → P1).
also indicates that if the pressure is greater than 8 MPa, more CH4
Meanwhile, the density peak of CO2 at P3 (after 2nd pressure down) is
molecules are adsorbed within the kerogen slit than in the kerogen
slightly higher than that at P2 (after CO2 injection), revealing that CO2
nanopore with the same pore size. Although the chemical composition of

9
S. Wang et al.
Fig. 11. Absolute adsorption isotherms (left panel) and CE values (right panel) of (a, d) CH4, (b, e) C2H6, and (c, f) C3H8 in their mixtures confined in different pores.
these two models is similar, they show different pore curvatures. The
curvature of a kerogen slit is eliminated, whereas the kerogen-based
nanopore having a greater curvature leads to the decrease of potential
well depth, causing less adsorption [59]. Similar conclusions have also
been reported in carbon nanotube and graphene slit [101].
We also analyze the effect of pore geometry on the adsorption
behavior of CH4, C2H6, and C3H8 mixtures in the three different models.
In general, the adsorption isotherms of CH4 increase with pressure,
whereas C2H6 adsorption capacity remains almost unchanged when P
greater than 10 MPa, and the adsorption of C3H8 tends to be suppressed
(Fig. 11). Because the molar ratio of CH4, C2H6, and C3H8 is 80:15:5, the
amount of adsorbed CH4 exhibits the largest value in each pore followed
by C2H6 and C3H8. For example, the adsorption amounts of CH4, C2H6,
and C3H8 in the kerogen-based nanopore are 0.0044, 0.00082, and
0.00036 mmol/m2 at 40 MPa, respectively. Moreover, the adsorption
amounts of CH4 in kerogen slit are larger than others, whereas the ca­
pacities of C2H6 and C3H8 in CNT are significantly greater than the other
two models. Although the adsorption isotherms in the cylindrical and
slit-shaped kerogen pore show similar trends because of the same
10
Chemical Engineering Journal 425 (2021) 130292
chemical composition, the specific adsorption amounts are different,
especially for methane. Caused by the pore geometry effect, the
adsorption capacity rises faster in the slit-shaped pores (Fig. 11a). We
also note that the adsorption behavior of each component in carbon
nanotubes are tremendously different from those in kerogen cylindrical
pores. Therefore, in order to elucidate the adsorption and transport
behavior of gas confined in shale kerogen, cylindrical kerogen por­
es—instead of the carbon nanotube and kerogen slit—should be utilized,
because the other models cannot account for the influence of hetero­
atoms, amorphons structure, and circular nanopores (diameter: 2–100
nm) of realistic kerogen. This conclusion sheds light on the constructions
of molecular models involved with shale resources and also suggests that
some results of the existing researches may be questionable. A detailed
comparison between our work and previous studies is summarized in
Table 1.
To further assess the selectivity of hydrocarbons in different pore
structures, we show CE values in Fig. 11b. They exhibit the same trend
that the CE values of C2H6 and C3H8 decrease with pressures but that of
CH4 grow slightly, indicating that the competitive adsorption capacity of
S. Wang et al. Chemical Engineering Journal 425 (2021) 130292

CH4 enhances whereas those of C2H6 and C3H8 become weaker. The
reason is that longer alkanes reach adsorption saturation quickly at
lower pressures; however, at higher pressures, the confined fluids
become denser, and the differences in molecular chains tend to be less
significant [102]. For most of the simulation conditions, the CE values of
both C2H6 and C3H8 are greater than unity, suggesting that heavier
components are more favorable for adsorption.
Moreover, the CE value of each component in kerogen slit and cir­
cular nanopore gets close to unity at P = 40 MPa. As reported by Denayer
et al. [102], the adsorption behavior of hydrocarbons in the nanopore is
controlled by the combination of entropic and enthalpic effects, and the
less significant differences among molecule types can be found in the
densely confined space under higher pressures. Owing to the influences
of pore geometry and chemical composition, the CE value of each
component in CNT is tremendously distinct from that in kerogen slit and
nanopore: for CH4, the CE value is lower in CNT, whereas for C2H6 and
C3H8, the values are greater.
Fig. 12 shows the estimated recovery efficiency of each component in
different pores during CO2 huff and puff. For pure CH4 confined in the
kerogen slit, the highest recovery appears during the first pressure
drawdown process because of the steepest adsorption isotherm (Fig. 10).
However, a comparison with the results in circular pores suggests that
the gas recovery behavior is also influenced by pore geometry: CO2 in­
jection stage produced the greatest recovery ratio in CNT and kerogen
nanopores (Fig. 12a). Owing to the stronger interactions between CO2
and kerogen, adsorbed CO2 molecules primarily occupy the inner sur­
faces of the kerogen matrix, while CH4 is mainly located upon the outer
surface and larger pores [37]. Thus, when CO2 is injected, CH4 mole­
cules tend to be expelled from the matrix surface. Because CNT is
exclusively composed of carbon atoms and the surface is ultra-smooth,
the gas releasement is favorable after CO2 injection, causing a higher
recovery ratio.
Moreover, owing to the decreasing adsorption capacity with
reducing pressure (Fig. 11a), CH4 recovery in the mixtures during each
stage reduces gradually with the production process (P0 → P1 → P2 →
P3), indicating that pressure drawdown is a better way to retrieve CH4 in
the mixtures (Fig. 12b). However, CO2 injection is more effective for
exploiting C2H6 and C3H8 (Fig. 12b). This observation can be explained
using the flat isotherms of C2H6 and steeply decreased isotherms of C3H8
(Fig. 11c). In CNT and kerogen-based nanopore, the recovery ratio of
hydrocarbons during CO2 injection increase with the number of carbon
atoms, whereas the pressure drop stages show the opposite trend. The
recoveries of C2H6 and C3H8 during the pressure drawdown stage are
even negative because the adsorption amount of C2H6 remains almost
unchanged with pressure, but that of C3H8 reduces. The “negative re­
covery ratio” means that no C3H8 modules are released during this stage
because the adsorption capacity of C3H8 decreases with pressure when P
greater than 8 MPa (Fig. 11). Zhou et al. [42] reported a similar behavior
in kerogen slit. Owing to the stronger affinity with kerogen (Fig. 7), C3H8
achieves saturation at lower pressure, whereas an elevated pressure
facilitates the adsorption of lighter alkanes. Therefore, shorter compo­
nents are favorably produced during pressure drop, but C3H8 is more
likely to be left in the pore.

4.3. Effect of pore size on shale gas recovery

In Fig. 13, we present the variation of CH4 absolute adsorption iso­

therms against the sizes of different kerogen cylindrical pores. As ex­
pected, the overall trends of these isotherms are the same, i.e., the
adsorption capacity increases with elevated pressure. However, the
Fig. 12. Shale gas recovery in different pore models during CO2 huff-n-puff: (a)
different units (such as mmol/m2 and mmol/cm3) of adsorption iso­
pure CH4; (b) CH4/C2H6/C3H8 mixture.
therms lead to distinct conclusions on the effect of pore size. If Nab is
characterized using per unit surface area (mmol/m2), the adsorption is
higher in larger pores because the greater size facilitates the formation
of layered structures [80]. Therefore, Nab expressed in mmol/m2 is
positively correlated with the pore radius [73,103,104]. However, the

11
S. Wang et al.
Fig. 13. Absolute adsorption isotherms of pure CH4 within kerogen-based nanopores having different radii. The adsorption capacities are shown in different units:
(a) mmol/m2; (b) mmol/cm3.
fluid–solid adhesion energy becomes greater (deeper) with pore size
decreasing [60,84,97.104]; thus, the adsorption amount of CH4 in per
volume of a narrower pore is larger, leading to the opposite variation
tendencies of the adsorption capacities with respect to radius.
Fig. 14 shows the absolute adsorption isotherms of each component
in the shale gas mixtures confined in kerogen-based nanopores having
different radii (T = 353 K). We represent the adsorption amount using
the per accessible volume. Because of the weaker fluid–solid in­
teractions, their adsorption capacities in a 5.4-nm kerogen pore are
much lower than 2-nm and 1.2-nm nanopores. CH4 in the hydrocarbons
mixtures exhibit a similar trend of adsorption isotherms with pure CH4;
however, the exact values are much smaller because C2H6 and C3H8
would occupy some adsorption sites on the kerogen surface. Fig. 14a
also indicates that the adsorption isotherms of CH4 increase with pres­
sure in each pore. However, the adsorption capacities of C2H6 and C3H8
increase at lower pressures and then decrease in a 1.2-nm kerogen-based
nanopore but remain almost constant in a 2-nm pore and exhibit a slight
increase in a 5.4-nm pore. This observation suggests that the adsorption
behavior of longer alkanes in the mixture is pore size dependent;
enlarging the pore causes the shifts of the saturation pressure to elevated
values.
The estimated CE values are imposed in Fig. 14 (right panel) to study
the selectivity adsorption of hydrocarbons mixture in kerogen nano­
pores with different radii. It is evident that the CE value for each
component decreases in the order of C3H8, C2H6, and CH4. More pre­
cisely, the CE values of CH4 are typically smaller than unity, while those
of C2H6 and C3H8 are greater than unity. This phenomenon further
demonstrates the preferential adsorption of the heavier components in
kerogen-based nanopores, which should be attributed to the stronger
fluid–solid interactions. For the gas mixture, the confinement effect of
C2H6 and C3H8 is higher in the vicinity of the solid surface, and more
CH4 molecules are forced to stay in the central pore where the affinities
from the kerogen are weaker. Fig. 14 also suggests that owing to the
weakened confinement effect, the deviations of the CE values from the
unity decrease when enlarging the pore size. As the pressure increased to
40 MPa, the difference among the CE values of these three components
tend to be negligible because the confined space gets more densely
packed at higher pressures, thus leading to the less remarkable differ­
ences among the molecular types [102].
With the increasing pore size, the recovery ratio of pure CH4 during
the pressure drop process increases while that corresponding to CO2
injection decreases (Fig. 15a). We attribute the reason to the reducing
affinities from the kerogen surface, which leads to the diminishing
12
Chemical Engineering Journal 425 (2021) 130292
preferential adsorption capacity of CO2. Because the pressure drop from
P0 to P1, 15 MPa, is larger than that from P2 to P3, 4.9 MPa, the recovery
ratio in the second pressure drop is much smaller than that in the first
pressure drop stage. As shown in Fig. 15b, for each component in the
hydrocarbon mixture, the recovery ratio during CO2 injection in the 1.2-
nm kerogen pore is higher than the other stages, which means injecting
CO2 can efficiently recover the hydrocarbons confined in much smaller
pores. During pressure drawdown, the recovery ratio of C3H8 presents
negative values, indicating that no C3H8 modules are released out during
this stage because of the decreasing adsorption capacity with pressure
(Fig. 14c). In a 2-nm nanopore, the recovery ratio of CH4 in the first
pressure drop (25.51%) is slightly higher than that of the CO2 huff stage
(18.21%). In contrast, for C2H6 and C3H8, the recovery during CO2 in­
jection is much higher, i.e., 29.08% and 36.70%, respectively. Similar
conclusions can also be reached in the 5.4 nm kerogen nanopore,
revealing that CO2 huff-n-puff serves as a promising technology for
exploiting heavier hydrocarbons.
4.4. Implications for shale gas recovery and CO2 sequestration
In the above sections, we have analyzed the competitive adsorption
behaviors of hydrocarbon within shale kerogen nanopores and explored
the gas recovery mechanisms during each stage influenced by various
factors, e.g., pressure, pore structure, and pore size. To gain a better
insight into the field applications, we show the total gas recovery and
CO2 sequestration ratio corresponding to distinct hydrocarbon compo­
sitions (i.e., pure CH4 and hydrocarbon mixture) in Fig. 16.
The total gas recovery is defined as the ratio between the total
number of gas molecules released from P0 (30 MPa) to P3 (after CO2 huff-
n-puff) and the initial loadings. Fig. 16 indicates that the recovery ratio
of CH4 in pure gas is almost identical to that in the mixture, revealing
that the existence of other components, such as C2H6 and C3H8, has no
significant influences on the CH4 recovery. A comparison among the
total recoveries within distinct pores suggests that, in the carbon
nanotube, the recovery of CH4 is highest (56%), whereas that of C2H6
(27%) and C3H8 (-21%) are lowest. This result agrees well with the
aforementioned adsorption behavior. As shown in Fig. 11, caused by the
differences in surface roughness and thermophysical properties, we
observe much greater adsorption amounts of C2H6 and C3H8 confined in
CNT, implying the stronger adsorption capacity as compared to kerogen
slit and nanopore. Therefore, during CO2 huff and puff, fewer C2H6 and
C3H8 molecules will be released in CNT.
Given that the chemical composition of CNT is distinct from the real
S. Wang et al.
Fig. 14. Absolute adsorption isotherms (left panel) and CE values (right panel) of (a, d) CH4, (b, e) C2H6, and (c, f) C3H8 in the mixtures confined in kerogen-based
nanopores having different radii.
shale kerogen and its ultrasmooth surface may lead to an erroneous
understanding of the fluid flow behavior, we suggest that the carbon
nanotube model should not be utilized to mimic shale kerogen. As ex­
pected, increasing the pore size will diminish the interactions from the
kerogen surface and further facilitate shale gas recovery (Fig. 16b).
Among these components, the highest recovery ratio of CH4 (52%)
manifests that the lighter components are preferentially recovered
because of the weakest adsorption capacity (Fig. 7).
The effect of pore structures and sizes on CO2 sequestration ratio,
ξCO2, for the fluids having different compositions are presented in
Fig. 16b. We define this parameter as,
nsequestrated
ξCO2 = CO2
× 100%
ninitial
CO2
where nsequestrated
CO2 is the number of CO2 molecules sequestrated in the
pore, and ninitial
CO2 is the total number of CO2 molecules contained in the
model at the initial stage of the second pressure drawdown process. We
Chemical Engineering Journal 425 (2021) 130292
observe a greater CO2 sequestration ratio in the hydrocarbon mixture,
revealing that more CO2 can be stored in shale reservoirs with the in­
crease of other components. For pure CH4, the sequestration efficiency
in CNT is much larger than the other two models, which accounts for the
highest recovery ratio of CH4 (39.19%), as shown in Fig. 12a. However,
for the hydrocarbon mixture, the maximum sequestration appears in the
kerogen slit (7.05%). When we inject the same amount of CO2 into a
shale reservoir, a greater hydrocarbon recovery also means a higher
sequestration efficiency of CO2 because the total pore volume of the
reservoir remains almost constant. Fig. 16c also suggests that the
sequestration ratio gets a remarkable increase with pore radius because
of the larger storage space and weaker affinity from the solid surface.
These results indicate that the shale reservoirs having larger pores can
(6)
be a good location for the geological sequestration of CO2, which not
only facilitates the adequate replacement of shale gas, but also ensure a
greater sequestration ratio.
13
S. Wang et al. Chemical Engineering Journal 425 (2021) 130292

Fig. 15. Recovery of shale gas confined in shale nanopores having different radii during CO2 huff-n-puff: (a) pure CH4; (b) recovery of each component in the
hydrocarbon mixture.

14
S. Wang et al.
Fig. 16. (a) Effect of pore structure on the total recovery of each component at 353 K (P0 → P3); (b) Effect of pore size on the total recovery of each component within
the kerogen-based nanopore at 353 K (P0 → P3); (c) CO2 sequestration ratio in kerogen pores having different geometries and sizes. Both the results for pure methane
and CH4/C2H6/C3H8 mixtures (80:15:5) are included..
5. Conclusions
Using GCMC simulations, we studied the hydrocarbon recovery
mechanisms and CO2 sequestration in realistic shale kerogen during CO2
huff-n-puff. We explored the effects of pressure, pore structure, and pore
size on gas recovery and sequestration efficiency, accounting for pure
CH4 and hydrocarbon mixture. A comparative study among carbon
nanotube, kerogen slit, and kerogen cylindrical nanopore suggests that
the different chemical compositions and surface roughness induce the
tremendously different hydrocarbon adsorption and CO2 sequestration
behavior in these pores. CO2 injection may serve as an efficient method
for circular pores, whereas the pressure drawdown technology is more
efficient for slit-shaped kerogen nanopore. We recommend that when
one wants to study the adsorption and transport of fluids through shale
kerogen using molecular modeling, it is better to use the cylindrical
kerogen-based nanopore instead of carbonaceous models (such as car­
bon nanotube and graphene) as well as kerogen slit.
Owing to the stronger interactions with the kerogen matrix, the
adsorption capacity of CO2 surpasses other components (CH4/C2H6/
C3H8). Thus, the injected CO2 drives the adsorbed hydrocarbon mole­
cules from the surface and leads to an enhanced gas recovery. The
density distributions of pure CH4 in the cylindrical kerogen nanopore
suggest that decreasing pressure primarily exploits gas in the adsorption
layer, while CO2 injection releases CH4 absorbed within the kerogen
matrix. For the mixture, CO2 huff-n-puff is more promising for the re­
covery of heavier hydrocarbons, while pressure drawdown favors the
production of lighter hydrocarbons. Therefore, the oil company may
develop a shale gas reservoir through low-cost pressure drop technology
first, which could quickly recover the investment and then change it to
CO2 huff-n-puff to improve the benefits. Moreover, enlarging the pore
size will improve the performance of pressure drawdown while
15
Chemical Engineering Journal 425 (2021) 130292
diminishing the CO2 injection effect; however, both the total gas re­
covery and CO2 sequestration ratio will increase. Our study sheds new
light on understanding the fundamental mechanisms of enhanced shale
gas recovery and CO2 sequestration through CO2 huff-n-puff, which
would benefit the efficient design of shale gas development strategies.
Declaration of Competing Interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
Acknowledgements
This work is supported partly by the National Natural Science
Foundation of China [grant numbers 51704312, U1762213, and
51974340]; the Applied Fundamental Research Project of Qingdao
[grant number No. 19-6-2-21-cg]; and the NanoGeosciences Lab and
Mudrock Systems Research Laboratory at the Bureau of Economic Ge­
ology, Jackson School of Geosciences, The University of Texas at Austin.
We would like to thank Dr. Shansi Tian for providing the SEM image.
References
[1] R.D. Vidic, S.L. Brantley, J.M. Vandenbossche, D. Yoxtheimer, J.D. Abad, Impact
of shale gas development on regional water quality, Science 340 (2013) 1235009-
1235009. https://doi.org/10.1126/science.1235009.
[2] J.D. Hughes, Energy: A reality check on the shale revolution, Nature 494 (7437)
(2013) 307–308, https://doi.org/10.1038/494307a.
[3] A. Paylor, The social–economic impact of shale gas extraction: a global
perspective, Third World Q. 38 (2) (2017) 340–355, https://doi.org/10.1080/
01436597.2016.1153420.
S. Wang et al.
[4] J. Li, Z. Chen, K. Wu, K. Wang, J. Luo, D. Feng, S. Qu, X. Li, A multi-site model to
determine supercritical methane adsorption in energetically heterogeneous
shales, Chem. Eng. J. 349 (2018) 438–455, https://doi.org/10.1016/j.
cej.2018.05.105.
[5] Energy Information Administration, Shale gas production drives world natural
gas production growth. https://www.eia.gov/todayinenergy/detail.php?
id=27512, 2016 (accessed January 15, 2021).
[6] S.R. Etminan, F. Javadpour, B.B. Maini, Z. Chen, Measurement of gas storage
processes in shale and of the molecular diffusion coefficient in kerogen, Int. J.
Coal Geol. 123 (2014) 10–19, https://doi.org/10.1016/j.coal.2013.10.007.
[7] J.G. Speight, Chapter 2 - Shale Gas Resources, in: J.G. Speight (Ed.), Shale Gas
Production Processes, Gulf Professional Publishing, Boston, 2013, pp. 25–68,
https://doi.org/10.1016/C2012-0-00596-0.
[8] Y. He, S. Cheng, Z. Sun, Z. Chai, Z. Rui, Improving Oil Recovery Through Fracture
Injection and Production of Multiple Fractured Horizontal Wells, J. Energ.
Resour. Technol. 142 (2020) 1–19, https://doi.org/10.1115/1.4045957.
[9] R.M. Slatt, Important geological properties of unconventional resource shales,
Cent. Eur. J. Geosci. 3 (2011) 435–448, https://doi.org/10.2478/s13533-011-
0042-2.
[10] M. Rezaveisi, F. Javadpour, K. Sepehrnoori, Modeling chromatographic
separation of produced gas in shale wells, Int. J. Coal Geol. 121 (2014) 110–122,
https://doi.org/10.1016/j.coal.2013.11.005.
[11] J. Lee, K. Perry, Unconventional Gas Reservoirs—Tight Gas, Coal Seams, and
Shales, Unconv. Gas. 29 (2007).
[12] Y. Sun, G. Zhang, S. Li, S. Jiang, CO2/N2 injection into CH4+C3H8 hydrates for gas
recovery and CO2 sequestration, Chem. Eng. J. (2019), 121973, https://doi.org/
10.1016/j.cej.2019.121973.
[13] R. Ershadnia, C.D. Wallace, M.R. Soltanian, CO2 geological sequestration in
heterogeneous binary media: Effects of geological and operational conditions,
Adv. Geo-energ. Res. 4 (4) (2020) 392–405.
[14] J. Jiang, Z. Rui, R. Hazlett, J. Lu, An integrated technical-economic model for
evaluating CO2 enhanced oil recovery development, Appl. Energy, 247, 190-211.
https://doi.org/10.1016/j.apenergy.2019.04.025.
[15] T. Ahmed, H. Nasrabadi, A. Firoozabadi, Complex flow and composition path in
CO2 injection schemes from density effects, Energy Fuels 26 (7) (2012)
4590–4598, https://doi.org/10.1021/ef300502f.
[16] J. Huang, T. Jin, M. Barrufet, J. Killough, Evaluation of CO2 injection into shale
gas reservoirs considering dispersed distribution of kerogen, Appl. Energy 260
(2020) 114285, https://doi.org/10.1016/j.apenergy.2019.114285.
[17] P. Zuloaga, W. Yu, J. Miao, K. Sepehrnoori, Performance evaluation of CO2 Huff-
n-Puff and continuous CO2 injection in tight oil reservoirs, Energy 134 (2017)
181–192, https://doi.org/10.1016/j.energy.2017.06.028.
[18] C. Song, D. Yang, Experimental and numerical evaluation of CO2 huff-n-puff
processes in Bakken formation, Fuel 190 (2017) 145–162, https://doi.org/
10.1016/j.fuel.2016.11.041.
[19] A. Abedini, F. Torabi, Oil recovery performance of immiscible and miscible CO2
huff-and-puff processes, Energy Fuels 28 (2) (2014) 774–784, https://doi.org/
10.1021/ef401363b.
[20] X. Meng, Z. Meng, J. Ma, T. Wang, Performance evaluation of CO2 huff-n-puff gas
injection in shale gas condensate reservoirs, Energies 12 (2019) 42, https://doi.
org/10.3390/en12010042.
[21] F. Torabi, A.Q. Firouz, A. Kavousi, K. Asghari, Comparative evaluation of
immiscible, near miscible and miscible CO2 huff-n-puff to enhance oil recovery
from a single matrix–fracture system (experimental and simulation studies), Fuel
93 (2012) 443–453, https://doi.org/10.1016/j.fuel.2011.08.037.
[22] P. Nguyen, J.W. Carey, H.S. Viswanathan, M. Porter, Effectiveness of supercritical
CO2 and N2 huff-and-puff methods of enhanced oil recovery in shale fracture
networks using microfluidic experiments, Appl. Energy 230 (2018) 160–174,
https://doi.org/10.1016/j.apenergy.2018.08.098.
[23] R. Khosrokhavar, K. Wolf, H. Bruining, Sorption of CH4 and CO2 on a
carboniferous shale from Belgium using a manometric setup, Int. J. Coal Geol.
128 (2014) 153–161, https://doi.org/10.1016/j.coal.2014.04.014.
[24] P. Charoensuppanimit, S.A. Mohammad, K.A.M. Gasem, Measurements and
modeling of gas adsorption on shales, Energy Fuels 30 (3) (2016) 2309–2319,
https://doi.org/10.1021/acs.energyfuels.5b02751.
[25] X. Tang, J. Zhang, X. Wang, B. Yu, W. Ding, J. Xiong, Y. Yang, L. Wang, C. Yang,
Shale characteristics in the southeastern Ordos Basin, China: Implications for
hydrocarbon accumulation conditions and the potential of continental shales, Int.
J. Coal Geol. 128 (2014) 32–46, https://doi.org/10.1016/j.coal.2014.03.005.
[26] A.-L. Cheng, W.-L. Huang, Selective adsorption of hydrocarbon gases on clays and
organic matter, Org. Geochem. 35 (4) (2004) 413–423, https://doi.org/10.1016/
j.orggeochem.2004.01.007.
[27] S. Kong, X. Huang, K. Li, X. Song, Adsorption/desorption isotherms of CH4 and
C2H6 on typical shale samples, Fuel 255 (2019) 115632, https://doi.org/
10.1016/j.fuel.2019.115632.
[28] Y.u. Wang, T.T. Tsotsis, K. Jessen, Competitive sorption of methane/ethane
mixtures on shale: Measurements and modeling, Ind. Eng. Chem. Res. 54 (48)
(2015) 12187–12195, https://doi.org/10.1021/acs.iecr.5b02850.
[29] H. Zhao, T. Wu, A. Firoozabadi, High pressure sorption of various hydrocarbons
and carbon dioxide in Kimmeridge Blackstone and isolated kerogen, Fuel 224
(2018) 412–423, https://doi.org/10.1016/j.fuel.2018.02.186.
[30] T. Wu, H. Zhao, S. Tesson, A. Firoozabadi, Absolute adsorption of light
hydrocarbons and carbon dioxide in shale rock and isolated kerogen, Fuel 235
(2019) 855–867, https://doi.org/10.1016/j.cej.2018.08.067.
16
Chemical Engineering Journal 425 (2021) 130292
[31] D.J.K. Ross, R. Marc Bustin, Impact of mass balance calculations on adsorption
capacities in microporous shale gas reservoirs, Fuel 86 (17-18) (2007)
2696–2706, https://doi.org/10.1016/j.fuel.2007.02.036.
[32] S. Tesson, A. Firoozabadi, Methane adsorption and self-diffusion in shale kerogen
and slit nanopores by molecular simulations, J. Phys. Chem. C. 122 (41) (2018)
23528–23542, https://doi.org/10.1021/acs.jpcc.8b07123.
[33] K. Falk, R. Pellenq, F.J. Ulm, B. Coasne, Effect of chain length and pore
accessibility on alkane adsorption in kerogen, Energy Fuels 29 (12) (2015)
7889–7896, https://doi.org/10.1021/acs.energyfuels.5b02015.
[34] H. Wu, Y. He, R. Qiao, Recovery of multicomponent shale gas from single
nanopores, Energy Fuels 31 (8) (2017) 7932–7940, https://doi.org/10.1021/acs.
energyfuels.7b01013.
[35] S. Wang, Q. Feng, F. Javadpour, Q. Hu, K. Wu, Competitive adsorption of
methane and ethane in montmorillonite nanopores of shale at supercritical
conditions: A grand canonical Monte Carlo simulation study, Chem. Eng. J. 355
(2019) 76–90, https://doi.org/10.1016/j.cej.2018.08.067.
[36] L. Huang, Z. Ning, Q. Wang, R. Qi, Y. Zeng, H. Qin, H. Ye, W. Zhang, Molecular
simulation of adsorption behaviors of methane, carbon dioxide and their mixtures
on kerogen: Effect of kerogen maturity and moisture content, Fuel 211 (2018)
159–172, https://doi.org/10.1016/j.fuel.2017.09.060.
[37] L. Huang, Z. Ning, Q. Wang, W. Zhang, Z. Cheng, X. Wu, H. Qin, Effect of organic
type and moisture on CO2 /CH4 competitive adsorption in kerogen with
implications for CO2 sequestration and enhanced CH4 recovery, Appl. Energy 210
(2018) 28–43, https://doi.org/10.1016/j.apenergy.2017.10.122.
[38] J. Zhou, Z. Jin, K.H. Luo, Effects of Moisture Contents on Shale Gas Recovery and
CO2 Sequestration, Langmuir 35 (26) (2019) 8716–8725, https://doi.org/
10.1021/acs.langmuir.9b00862.
[39] M. Vasileiadis, L.D. Peristeras, K.D. Papavasileiou, I.G. Economou, Transport
properties of shale gas in relation to kerogen porosity, J. Phys. Chem. C. 122 (11)
(2018) 6166–6177, https://doi.org/10.1021/acs.jpcc.8b00162.
[40] T.A. Ho, Y. Wang, Y. Xiong, L.J. Criscenti, Differential retention and release of
CO2 and CH4 in kerogen nanopores: Implications for gas extraction and carbon
sequestration, Fuel 220 (2018) 1–7, https://doi.org/10.1016/j.fuel.2018.01.106.
[41] M. Zhang, S. Zhan, Z. Jin, Recovery mechanisms of hydrocarbon mixtures in
organic and inorganic nanopores during pressure drawdown and CO2 injection
from molecular perspectives, Chem. Eng. J. 382 (2020) 122808, https://doi.org/
10.1016/j.cej.2019.122808.
[42] J. Zhou, Z. Jin, K.H. Luo, Insights into recovery of multi-component shale gas by
CO2 injection: A molecular perspective, Fuel 267 (2020) 117247, https://doi.org/
10.1016/j.fuel.2020.117247.
[43] M. Firouzi, J. Wilcox, Molecular modeling of carbon dioxide transport and storage
in porous carbon-based materials, Microporous Mesoporous Mater. 158 (2012)
195–203, https://doi.org/10.1016/j.micromeso.2012.02.045.
[44] J. Zhang, K. Liu, M.B. Clennell, D.N. Dewhurst, M. Pervukhina, Molecular
simulation of CO2–CH4 competitive adsorption and induced coal swelling, Fuel
160 (2015) 309–317, https://doi.org/10.1016/j.fuel.2015.07.092.
[45] J. Zhang, N. Burke, S. Zhang, K. Liu, M. Pervukhina, Thermodynamic analysis of
molecular simulations of CO2 and CH4 adsorption in FAU zeolites, Chem. Eng. Sci.
113 (2014) 54–61, https://doi.org/10.1016/j.ces.2014.04.001.
[46] J. Collell, G. Galliero, R. Vermorel, P. Ungerer, M. Yiannourakou, F. Montel,
M. Pujol, Transport of multicomponent hydrocarbon mixtures in shale organic
matter by molecular simulations, J. Phys. Chem. C. 119 (39) (2015)
22587–22595, https://doi.org/10.1021/acs.jpcc.5b07242.
[47] T. Zhao, X. Li, H. Zhao, M. Li, Molecular simulation of adsorption and
thermodynamic properties on type II kerogen: Influence of maturity and moisture
content, Fuel 190 (2017) 198–207, https://doi.org/10.1016/j.fuel.2016.11.027.
[48] K. Li, S. Kong, P. Xia, X. Wang, Microstructural characterisation of organic matter
pores in coal-measure shale, Adv. Geo-energ. Res. 4 (4) (2020) 372–391, https://
doi.org/10.46690/ager.2020.04.04.
[49] Y. Tang, C. Hou, Y. He, Y. Wang, Y. Chen, Z. Rui, Review on pore structure
characterization and microscopic flow mechanism of CO2 flooding in porous
media, Energy Technol. 9 (1) (2021) 2000787, https://doi.org/10.1002/
ente.202000787.
[50] T.A. Ho, L.J. Criscenti, Y. Wang, Nanostructural control of methane release in
kerogen and its implications to wellbore production decline, Sci. Rep. 6 (2016)
28053, https://doi.org/10.1038/srep28053.
[51] M. Gasparik, P. Bertier, Y. Gensterblum, A. Ghanizadeh, B.M. Krooss, R. Littke,
Geological controls on the methane storage capacity in organic-rich shales, Int. J.
Coal Geol. 123 (2014) 34–51, https://doi.org/10.1016/j.coal.2013.06.010.
[52] A. Afsharpoor, F. Javadpour, Liquid slip flow in a network of shale noncircular
nanopores, Fuel 180 (2016) 580–590, https://doi.org/10.1016/j.
fuel.2016.04.078.
[53] F.P. Wang, R.M. Reed, A. John, G. Katherine, Pore networks and fluid flow in gas
shales, in: Proc. - SPE Annu. Tech. Conf. Exhib., 2009. https://doi.org/10.2118/
124253-ms.
[54] R.J. Ambrose, R.C. Hartman, M. Diaz-Campos, I.Y. Akkutlu, C.H. Sondergeld,
Shale gas-in-place calculations part I: New pore-scale considerations, Spe J. 17
(2012) 219–229, https://doi.org/10.2118/131772-PA.
[55] R.G. Loucks, R.M. Reed, S.C. Ruppel, D.M. Jarvie, Morphology, genesis, and
distribution of nanometer-scale pores in siliceous mudstones of the mississippian
barnett shale, J. Sediment. Res. 79 (12) (2009) 848–861, https://doi.org/
10.2110/jsr.2009.092.
[56] T. Wang, S. Tian, G. Li, M. Sheng, W. Ren, Q. Liu, S. Zhang, Molecular simulation
of CO2/CH4 competitive adsorption on shale kerogen for CO2 sequestration and
enhanced gas recovery, J. Phys. Chem. C. 122 (30) (2018) 17009–17018, https://
doi.org/10.1021/acs.jpcc.8b02061.
S. Wang et al.
[57] A. Qajar, H. Daigle, M. Prodanovic, The effects of pore geometry on adsorption
equilibrium in shale formations and coal-beds: Lattice density functional theory
study, Fuel 163 (2016) 205–213, https://doi.org/10.1016/j.fuel.2015.09.061.
[58] Y.u. Pang, X. Hu, S. Wang, S. Chen, M.Y. Soliman, H. Deng, Characterization of
adsorption isotherm and density profile in cylindrical nanopores: Modeling and
measurement, Chem. Eng. J. 396 (2020) 125212, https://doi.org/10.1016/j.
cej.2020.125212.
[59] X. Zhu, Y.-P. Zhao, Atomic mechanisms and equation of state of methane
adsorption in carbon nanopores, J. Phys. Chem. C. 118 (31) (2014)
17737–17744, https://doi.org/10.1021/jp5047003.
[60] S. Riewchotisakul, I.Y. Akkutlu, Adsorption-enhanced transport of hydrocarbons
in organic nanopores, Spe J. 21 (2016) 1960–1969, https://doi.org/10.2118/
175107-PA.
[61] M. Kazemi, A. Takbiri-Borujeni, M. Mansouri-Boroujeni, T. Sun, Effect of water in
trasnport and storage of non-aqueous species in kerogen, Proc. - SPE Annu. Tech.
Conf. Exhib. (2017), https://doi.org/10.2118/187087-ms.
[62] Y. Yang, J. Liu, J. Yao, J. Kou, Z. Li, T. Wu, K. Zhang, L. Zhang, H. Sun, Adsorption
behaviors of shale oil in kerogen slit by molecular simulation, Chem. Eng. J. 387
(2020) 124054, https://doi.org/10.1016/j.cej.2020.124054.
[63] Z. Sun, X. Li, W. Liu, T. Zhang, M. He, H. Nasrabadi, Molecular dynamics of
methane flow behavior through realistic organic nanopores under geologic shale
condition: pore size and kerogen types, Chem. Eng. J. 398 (2020) 124341,
https://doi.org/10.1016/j.cej.2020.124341.
[64] S.R. Kelemen, M. Afeworki, M.L. Gorbaty, M. Sansone, P.J. Kwiatek, C.C. Walters,
H. Freund, M. Siskin, A.E. Bence, D.J. Curry, M. Solum, R.J. Pugmire,
M. Vandenbroucke, M. Leblond, F. Behar, Direct characterization of kerogen by x-
ray and solid-state **13C nuclear magnetic resonance methods, Energy Fuels 21
(3) (2007) 1548–1561, https://doi.org/10.1021/ef060321h.
[65] P. Ungerer, J. Collell, M. Yiannourakou, Molecular modeling of the volumetric
and thermodynamic properties of kerogen: Influence of organic type and
maturity, Energy Fuels 29 (1) (2015) 91–105, https://doi.org/10.1021/
ef502154k.
[66] D.M. Jarvie, R.J. Hill, T.E. Ruble, R.M. Pollastro, Unconventional shale-gas
systems: The Mississippian Barnett Shale of north-central Texas as one model for
thermogenic shale-gas assessment, Am. Assoc. Pet. Geol. Bull. 91 (4) (2007)
475–499, https://doi.org/10.1306/12190606068.
[67] J. Collell, P. Ungerer, G. Galliero, M. Yiannourakou, F. Montel, M. Pujol,
Molecular simulation of bulk organic matter in type II shales in the middle of the
oil formation window, Energy Fuels 28 (2014) 7457–7466, https://doi.org/
10.1021/ef5021632.
[68] T. Wang, S. Tian, G. Li, M. Sheng, Selective adsorption of supercritical carbon
dioxide and methane binary mixture in shale kerogen nanopores, J. Nat. Gas Sci.
Eng. 50 (2018) 181–188, https://doi.org/10.1016/j.jngse.2017.12.002.
[69] S.L. Montgomery, D.M. Jarvie, K.A. Bowker, R.M. Pollastro, Mississippian Barnett
Shale, Fort Worth basin, north-central Texas: Gas-shale play with multi–trillion
cubic foot potential, Am. Assoc. Pet. Geol. Bull. 89 (2) (2005) 155–175, https://
doi.org/10.1306/09170404042.
[70] J. Zumberge, K. Ferworn, S. Brown, Isotopic reversal (’rollover’) in shale gases
produced from the Mississippian Barnett and Fayetteville formations, Mar. Pet.
Geol. 31 (1) (2012) 43–52, https://doi.org/10.1016/j.marpetgeo.2011.06.009.
[71] H.J. Sun, COMPASS: An ab initio force-field optimized for condensed-phase
applications overview with details on alkane and benzene compounds, J. Phys.
Chem. B. 102 (1998) 7338–7364, https://doi.org/10.1021/jp980939v.
[72] D.-Y. Peng, D.B. Robinson, A new two-constant equation of state, Ind. Eng. Chem.
Fundam. 15 (1) (1976) 59–64, https://doi.org/10.1021/i160057a011.
[73] A. Sharma, S. Namsani, J.K. Singh, Molecular simulation of shale gas adsorption
and diffusion in inorganic nanopores, Mol. Simul. 41 (5-6) (2015) 414–422,
https://doi.org/10.1080/08927022.2014.968850.
[74] Y. Liu, H.A. Li, Y. Tian, Z. Jin, H. Deng, Determination of the absolute adsorption/
desorption isotherms of CH4 and n-C4H10 on shale from a nano-scale perspective,
Fuel 218 (2018) 67–77, https://doi.org/10.1016/j.fuel.2018.01.012.
[75] E. Paquet, H.L. Viktor, Molecular dynamics, monte carlo simulations, and
langevin dynamics: A computational review, BioMed Res. Int. 2015 (2015) 1–18,
https://doi.org/10.1155/2015/183918.
[76] D. Dubbeldam, A. Torres-Knoop, K.S. Walton, On the inner workings of Monte
Carlo codes, Mol. Simul. 39 (14-15) (2013) 1253–1292, https://doi.org/10.1080/
08927022.2013.819102.
[77] Z. Jin, A. Firoozabadi, Effect of water on methane and carbon dioxide sorption in
clay minerals by Monte Carlo simulations, Fluid Phase Equilib. 382 (2014) 10–20,
https://doi.org/10.1016/j.fluid.2014.07.035.
[78] Y. Tian, C. Yan, Z. Jin, Characterization of methane excess and absolute
adsorption in various clay nanopores from molecular simulation, Sci. Rep. 7
(2017) 12040, https://doi.org/10.1038/s41598-017-12123-x.
[79] H. Sun, H. Zhao, N.a. Qi, Y. Li, Molecular insights into the enhanced shale gas
recovery by carbon dioxide in kerogen slit nanopores, J. Phys. Chem. C 121 (18)
(2017) 10233–10241.
[80] K. Bui, I.Y. Akkutlu, Hydrocarbons recovery from model-kerogen nanopores, Spe
J. (2017) 854–862, https://doi.org/10.2118/185162-pa.
17
Chemical Engineering Journal 425 (2021) 130292
[81] K.S. Okiongbo, A.C. Aplin, S.R. Larter, Changes in type II kerogen density as a
function of maturity: Evidence from the kimmeridge clay formation, Energy Fuels
19 (2005) 2495–2499, https://doi.org/10.1021/ef050194+.
[82] M.L. Connolly, Solvent-accessible surfaces of proteins and nucleic acids, Science
(80-.). 221 (1983) 709-713. https://doi.org/10.1126/science.6879170.
[83] G. Chen, J. Zhang, S. Lu, M. Pervukhina, K. Liu, Q. Xue, H. Tian, S. Tian, J. Li, M.
B. Clennell, D.N. Dewhurst, Adsorption behavior of hydrocarbon on illite, Energy
& Fuel 30 (11) (2016) 9114–9121, https://doi.org/10.1021/acs.
energyfuels.6b01777.
[84] T. Zhang, G.S. Ellis, S.C. Ruppel, K. Milliken, R. Yang, Effect of organic-matter
type and thermal maturity on methane adsoption in shale-gas systems, Org.
Geochem. 47 (2012) 120–131, https://doi.org/10.1016/j.
orggeochem.2012.03.012.
[85] J.-S. Bae, S.K. Bhatia, High-pressure adsorption of methane and carbon dioxide on
coal, Energy Fuels 20 (6) (2006) 2599–2607, https://doi.org/10.1021/
ef060318y.
[86] J.-S. Bae, S.K. Bhatia, V. Rudolph, P. Massarotto, Pore accessibility of methane
and carbon dioxide in coals, Energy Fuels 23 (6) (2009) 3319–3327, https://doi.
org/10.1021/ef900084b.
[87] S. Wang, Q. Feng, F. Javadpour, M. Zha, R. Cui, Multiscale modeling of gas
transport in shale matrix: An integrated study of molecular dynamics and rigid-
pore-network model, Spe J. (2020) 1416–1442, https://doi.org/10.2118/
187286-PA.
[88] H. Wu, J. Chen, H.e. Liu, Molecular dynamics simulations about adsorption and
displacement of methane in carbon nanochannels, J. Phys. Chem. C. 119 (24)
(2015) 13652–13657, https://doi.org/10.1021/acs.jpcc.5b02436.
[89] K. Mosher, J. He, Y. Liu, E. Rupp, J. Wilcox, Molecular simulation of methane
adsorption in micro- and mesoporous carbons with applications to coal and gas
shale systems, Int. J. Coal Geol. 109–110 (2013) 36–44, https://doi.org/10.1016/
j.coal.2013.01.001.
[90] J. Zhang, S.K. Choi, Molecular dynamics simulation of methane in potassium
montmorillonite clay hydrates, J. Phys. B At. Mol. Opt. Phys. 39 (18) (2006)
3839–3848, https://doi.org/10.1088/0953-4075/39/18/013.
[91] L.u. Zhang, C. Liu, Q. Li, Molecular simulations of competitive adsorption
behavior between CH4-C2H6 in K-illite clay at supercritical conditions, Fuel 260
(2020) 116358, https://doi.org/10.1016/j.fuel.2019.116358.
[92] M.D. Donohue, G.L. Aranovich, Classification of Gibbs adsorption isotherms, Adv.
Colloid Interface Sci. 76 (1998) 137–152, https://doi.org/10.1016/S0001-8686
(98)00044-X.
[93] Y. Kurniawan, S.K. Bhatia, V. Rudolph, Simulation of binary mixture adsorption
of methane and CO2 at supercritical conditions in carbons, AIChE J. 52 (3) (2006)
957–967, https://doi.org/10.1002/(ISSN)1547-590510.1002/aic.v52:310.1002/
aic.10687.
[94] F. Karavias, A.L. Myers, Isosteric heats of multicomponent adsorption:
thermodynamics and computer simulations, Langmuir 7 (12) (1991) 3118–3126,
https://doi.org/10.1021/la00060a035.
[95] H. Pan, J.A. Ritter, P.B. Balbuena, Examination of the approximations used in
determining the isosteric heat of adsorption from the Clausius− Clapeyron
equation, Langmuir 14 (21) (1998) 6323–6327, https://doi.org/10.1021/
la9803373.
[96] T. Vuong, P.A. Monson, Monte Carlo simulation studies of heats of adsorption in
heterogeneous solids, Langmuir 12 (22) (1996) 5425–5432, https://doi.org/
10.1021/la960325m.
[97] J. Liu, W.G. Chapman, Thermodynamic modeling of the equilibrium partitioning
of hydrocarbons in nanoporous kerogen particles, Energy Fuels 33 (2) (2019)
891–904, https://doi.org/10.1021/acs.energyfuels.8b03771.
[98] Y. Liu, J. Wilcox, Effects of surface heterogeneity on the adsorption of CO2 in
microporous carbons, Environ. Sci. Technol. 46 (3) (2012) 1940–1947, https://
doi.org/10.1021/es204071g.
[99] S. Wang, F. Javadpour, Q. Feng, Fast mass transport of oil and supercritical
carbon dioxide through organic nanopores in shale, Fuel 181 (2016) 741–758,
https://doi.org/10.1016/j.fuel.2016.05.057.
[100] P. Billemont, B. Coasne, G. De Weireld, Adsorption of carbon dioxide, methane,
and their mixtures in porous carbons: Effect of surface chemistry, water content,
and pore disorder, Langmuir 29 (10) (2013) 3328–3338, https://doi.org/
10.1021/la3048938.
[101] W. Song, J. Yao, J. Ma, A. Li, Y. Li, H. Sun, L. Zhang, Grand canonical Monte Carlo
simulations of pore structure influence on methane adsorption in micro-porous
carbons with applications to coal and shale systems, Fuel 215 (2018) 196–203,
https://doi.org/10.1016/j.fuel.2017.11.016.
[102] J.F. Denayer, I. Daems, G.V. Baron, Adsorption and reaction in confined spaces,
Oil Gas Sci. Technol. 61 (4) (2006) 561–569, https://doi.org/10.2516/ogst:
2006025a.
[103] D. Keffer, H.T. Davis, A.V. McCormick, The effect of nanopore shape on the
structure and isotherms of adsorbed fluids, Adsorption 2 (1) (1996) 9–21, https://
doi.org/10.1007/BF00127094.
[104] D. Cao, J. Wu, Self-diffusion of methane in single-walled carbon nanotubes at
suband supercritical conditions, Langmuir 20 (9) (2004) 3759–3765, https://doi.
org/10.1021/la036375q.