Fuel xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Full Length Article

Enhanced oil recovery mechanism of CO2 water-alternating-gas

injection in silica nanochannel
Youguo Yan a, Chuanyong Li a, Zihan Dong a, Timing Fang a, Baojiang Sun b, Jun Zhang a,⇑
a
College of Science, China University of Petroleum, 266580 Qingdao, Shandong, People’s Republic of China
b
School of Petroleum Engineering, China University of Petroleum, 266580 Qingdao, Shandong, People’s Republic of China

h i g h l i g h t s

Molecular dynamics simulation was firstly used to address CO2 WAG in nanochannel.

The different roles of scCO2 slug and water slug in CO2 WAG injection were revealed.

The synergistic effects of scCO2 slug and water slug were proposed.

a r t i c l e i n f o a b s t r a c t

Article history: CO2 water-alternating-gas (WAG) injection has been proven as a promising and effective technology to
Received 16 July 2016 significantly enhance the recovery of the residual oil. However, the detailed displacing process and exqui-
Received in revised form 31 October 2016 site interaction mechanism still remain ambiguous. By using Nonequilibrium Molecular Dynamic
Accepted 3 November 2016
Simulation (NEMD), three comparative displacement processes of adsorbed oil layer in nanoslit using
Available online xxxx
water, supercritical carbon dioxide (scCO2), and CO2 WAG injection were conducted to study the
enhanced oil recovery (EOR) mechanism of CO2 WAG injection. Simulation results indicate that the
Keywords:
scCO2 and water play different roles in CO2 WAG: the scCO2 slug could selectively dissolve apolar oil com-
WAG
scCO2
ponents, and the water slug provides stable water/scCO2 promoting interface, which compels the misci-
Nanoslit ble oil/scCO2 slug out of the nanoslit. Furthermore, the synergistic effects of scCO2 slug and water slug
EOR were found to account for a decreased injecting pressure and an improved dissolution capability of
Molecular dynamics simulation scCO2 slug, which endows the CO2 WAG higher oil recovery efficiency than that of both scCO2 injection
and water injection. Our work here provides a comprehensive interpretation of EOR mechanism of CO2
WAG, and the results could provide insights into the optimization of the engineering process.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction happens and causes an early breakthrough, resulting in low volu-

Enhancing oil recovery (EOR) is an unremitting pursuit in the
petroleum field. After an initial quick extraction, most of the resid-
ual oil stays inside the reservoir on adsorption status [1,2]. In order
to displace the residual oil, numerous oil-displacing agents have
been developed [2–5], such as the surfactant, the polymer,
nanoparticles, and the scCO2. Among them, scCO2 exhibits great
advantages due to its unique properties, such as its large diffusion
rate approximately equal to gas, and high dissolving capability
similar to liquid [6,7]. ScCO2-EOR has been confirmed to be a suc-
cess technique to reduce the residual oil saturation [5,8].
However, due to the extremely low viscosity, the displacement
front of scCO2 is inherently unstable and viscous fingering often
⇑ Corresponding author.
E-mail addresses: yyg@upc.edu.cn (Y. Yan), zhangjun.upc@gmail.com (J. Zhang).
metric sweep efficiency [2,9]. To disentangle this problem, CO2
WAG technique was developed. Experimental researches demon-
strated that, in CO2 WAG injection, the water slug could improve
the volumetric sweep efficiency, and the scCO2 slug could enhance
the oil displacement ratio [10–12]. In current opinions, the CO2
could diffuse into and swell the crude oil to reduce its viscosity
[13]; meanwhile, the CO2 also distributes at the interface between
oil and water to reduce the interfacial tension [14,15], which favor
forming of their miscible phase. In addition, the water dissolved
with CO2 could corrode and broaden the channel. These advantages
will facilitate improving the EOR. However, limited to the detecting
method to observe the displacing process of crude oil in opaque
and nanoscale channel, the underlying mechanism of microscopic
interaction mechanism and the synergistic effect between the
scCO2 and the water slug is not very clear. Therefore, the in-
depth understanding the microscopic mechanism of the WAG is

http://dx.doi.org/10.1016/j.fuel.2016.11.019
0016-2361/Ó 2016 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Yan Y et al. Enhanced oil recovery mechanism of CO2 water-alternating-gas injection in silica nanochannel. Fuel (2016),
http://dx.doi.org/10.1016/j.fuel.2016.11.019
2 Y. Yan et al. / Fuel xxx (2016) xxx–xxx
fundamentally significant for improving EOR. Whereas, due to the
nano- and micro-scale size of the porous reservoirs, conventional
fluid theory (such as Darcy’s law) is no longer suitable in such
small scale [16]. Further, because of the invisibility of rock reser-
voirs and the great difficulty in preparing the synthetic core with
a comparable porous structure to the real core, it is still a formid-
able challenge to obtain the detailed information of the displacing
process.
In the past decades, molecular dynamics (MD) simulation has
been fully developed and exhibits great advantages offering
dynamic atomic-level information. In addition, some studies have
been conducted to investigate the influence of the residual water
[16], surface wettability [17], polarity of oil components [18],
and various additives [4,17] on the oil displacement. And, the
behavior of the scCO2-substituted nanopore oil has drawn dramat-
ically increasing research interest [19–21]. These studies shed light
on the oil displacement behavior in water injection and scCO2
injection. However, to the best of our knowledge, there exist very
limited studies on the CO2 WAG injection in nanopore [22,23].
The detailed displacing process and microscopic interaction is still
unclear.
In this work, under the employment of NEMD simulations,
three parallel displacing processes using scCO2, water and CO2
WAG injections were investigated. The roles of scCO2 and water
slugs in the CO2 WAG injection were analyzed, and some synergis-
tic effects of scCO2 slug and water slug were found to be responsi-
ble for high oil recovery efficiency. Our work provides an atomic-
level understanding of the CO2 WAG process, and the results would
be useful to realize some engineering phenomena in CO2 WAG and
guide the application of scCO2 in the tertiary oil recovery
efficiency.
2. Experimental and simulation
2.1. Nanoscale slit
The initial silica lattice was derived from the database of the
Material Studio software, and a a-quartz silica supercell with
orthorhombic dimensions (xyz) of 24.6  81.1  97.6 Å3 was con-
structed, the detailed building process could be seen in Supporting
information (SI). Considering there are abundant nanoscale pores
in the size range of 30–80 Å in rock reservoir [24], a nanoslit with
a height of 63 Å along y-axis was built through deleting the middle
part of silica supercell, and the exposed surface was (0 0 1). The sil-
ica surface was hydroxylated by hydroxyl with a density of
9.6 nm2, which is consistent with the result of crystal chemistry
calculations (5.9–18.8 nm2) [25].
2.2. Adsorbed oil layer
Here, decane and asphaltene were selected as representative
apolar and polar components of crude oil. The molecular configu-
rations of decane and asphaltene are displayed in Fig. 1a and b,
respectively. A detailed explanation for the selection of these two
oil components could be found in SI. To obtain equivalent adsorbed
oil layer, 10 asphaltene and 40 decane molecules was spread
evenly onto the silica surface with dimensions (xz) of
24.6  68.0 Å2, and an energy minimization and a subsequent
1 ns MD simulation were performed [27]. In order to conveniently
observe the displacement consequence, these two adsorbed oil lay-
ers were placed onto the left side of the nanoslit. In the adsorbed
oil layer, due to the strong interaction between the polar asphal-
tene and the hydroxylated surface, asphaltene molecules could
preferentially adsorbed onto the surface, meanwhile the decane
molecules adsorb onto the unoccupied silica surface and wrap
Please cite this article in press as: Yan Y et al. Enhanced oil recovery mechanism of CO2 water-alternating-gas injection in silica nanochannel. Fuel (2016),
http://dx.doi.org/10.1016/j.fuel.2016.11.019
the adsorbed asphaltene molecules [26–29], as presented in
Fig. 1d–f.
2.3. Fluid
Based on the typical reservoir condition, a temperature of 330 K
and pressure of 20 MPa were selected in our simulation, and the
corresponding density of scCO2 (0.743 g/cm3) and water (0.993 g/
cm3) were adopted in the initial model, according to National Insti-
tute of Standards and Technology’s (NIST) database (2011) [30]. By
adding certain number of CO2 and water molecules into box, the
pressure (20 MPa) at 330 K could be maintained. The amount of
added CO2 and water molecules could be calculated from the
Peng–Robinson equation according to their density [31], and is
listed in Table S1 in SI. As shown in Fig. 1d (scCO2) and e (water),
a fluid box with dimensions (xyz) of 24.6  81.1  130.0 Å3 was
set on the left side of silica nanoslit which was also filled with
the fluid. As to the CO2 WAG model (Fig. 1f), it is composed of a
water slug (left) and a scCO2 slug (middle) with same length of
65 Å, and the nanoslit filled with water. In these three models, a sil-
ica sheet was placed on the left of the fluid as slab to push the fluid
across the channel.
2.4. MD simulation
The NEMD simulation was performed with LAMMPS package
[32] through employing EPM2 CO2 model [33] and a simple point
charge (SPC) water [34]. The CLAYFF force field [35] and CHARMM
force field [36,37] were used for silica and oil molecules, respec-
tively. Detailed parameters could be found in Table S2 in SI.
The MD simulations were conducted in NVT (the number of
molecules, volume and temperature are constant) ensemble at
330 K with the Nosé–Hoover thermostat [38]. During simulation,
the leftmost silica sheet was defined as a rigid body and a constant
velocity 1 m/s along z axis was applied to it to push the fluid across
the channel; and the two silica slabs of nanoslit were kept frozen
except for the modified hydroxyl. The periodic boundary condition
was applied in all directions. The particle-particle particle-mesh
(PPPM) algorithm [39] with a convergence parameter of 104
was used to treat the long-range electrostatic interactions. The
van der Waals interactions were calculated with a cutoff (10 Å).
A time step of 1 fs was used, and the corresponding data was col-
lected in every 5 ps. A 13 ns simulation was conducted for all three
systems. The configuration snapshots were rendered by VMD soft-
ware [40].
3. Results and discussion
In the WAG injection process, the scCO2 slug and water slug will
alternately pass through the reservoir pore. Sometime the pore
was filled with the scCO2 slug, and in the next period, the pore will
be occupied by the water slug. Hence, the displacement behaviors
in the single scCO2 injection and single water injection were ana-
lyzed firstly to unveil their roles in the displacement process.
And then, a detailed CO2 WAG injection process was studied to
unveil the EOR mechanism of the CO2 WAG.
3.1. The displacement behavior of scCO2 and water
3.1.1. Displacement phenomena
Fig. 2 presents several sequential snapshots in the scCO2 injec-
tion. As shown, nearly all decane molecules dissolve into the scCO2
phase, and asphaltene molecules could only slip on but cannot
detach from the silica surface. Detailed observation manifests that
 Y. Yan et al. / Fuel xxx (2016) xxx–xxx 3

Fig. 1. Snapshots of initial system configurations: (a) decane (C10H22); (b) asphaltene (C27H23O2N1); (c) silica surface; (d) scCO2 injection model; (e) water injection model; (f)
CO2 WAG injection model. Atom color code: O, red; Si, yellow; H, white; N, blue. For clarity, CO2 molecules are represented as cyan points; water molecules are represented as
red points; decane, gray; asphaltene, green. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 2. Snapshots of displacement process of adsorbed oil layer in the scO2 injection.

after all scCO2 slug is injected, most of the decane molecules are
pulled out of the nanoslit. However, it is also clearly showed that
there still are some decane molecules resided inside the nanoslit.
As to the water injection, Fig. 3 demonstrates that no oil molecule
was displaced after the injection, and the only notable change was
that the oil layer shrank and displayed a semi-spherical configura-
tion. These results indicate three facts: (1) the scCO2 could selec-
tively dissolve the apolar oil component; (2) the expulsion
outcome of scCO2 to decane molecules is undesirable; and (3)
the water injection is not beneficial to displace the adsorbed oil.
3.1.2. Displacement mechanism
To reveal the essence of above mentioned displacement phe-
nomena, i.e. dissolution of decane and retention of asphaltene in
scCO2 injection, and retention of both decane and asphaltene in
water injection, the interactions between oil components, silica
surface and the fluid were calculated as diagrammed in Fig. 4.
The interaction energy can be estimated by following equation
[41]. Here, the interaction energy between decane and water is
shown as follow.
Edeca-wat ¼ Etotal  ðEdeca þ Ewat Þ ð1Þ
where Edeca-wat is the interaction energy between decane and water,
Etotal is the total energy of the decane and water, and Edeca and Ewat
are the energy of individual decane and the water, respectively.
Increasingly negative energy represents increasing interaction
between the two components.
(1) scCO2 displacement mechanism: In the case of the scCO2
injection, Fig. 4a shows that the interaction energy between decane
and scCO2 is far larger than that between asphaltene and scCO2,

Please cite this article in press as: Yan Y et al. Enhanced oil recovery mechanism of CO2 water-alternating-gas injection in silica nanochannel. Fuel (2016),
http://dx.doi.org/10.1016/j.fuel.2016.11.019
4 Y. Yan et al. / Fuel xxx (2016) xxx–xxx

Fig. 3. Snapshots of displacement process of adsorbed oil layer in the water injection.

Fig. 4. Interaction energy profiles between oil molecules, fluid, silica surface. (a) scCO2 injection, (b) water injection. For clarity, asphaltene, decane, silica surface, scCO2 and
water were denoted with asph, deca, surf, CO2 and wat, respectively.

hence scCO2 could selectively dissolve decane over asphaltene
[42]. And the high attraction between decane molecules and CO2
phase (seen in Fig. 4a) enables the decane molecules dispersing
into the CO2 phase, meanwhile the large diffusion coefficient
(0.83  108 m2/s) of decane in low viscosity scCO2, which is
nearly 4 times of that of water ((2.30 ± 1.5%)  109 m2/s) at the
room temperature and 1 atm [43,44], ensures the quick diffusion
of decane in CO2 phase, allows decane molecules quickly to dis-
perse into the scCO2 phase, and the detailed calculation method
could be found in SI.
Further, in terms of the different adsorption status of decane
and asphaltene, a quantitative analysis of the interaction energy
could provide us a rational explanation. Firstly, as to the dissolu-
tion of decane, the driving force derives from the interaction
between decane and scCO2, while the resistance force comes from
two parts, one is the interaction between decane and silica surface,
the other is the interaction between decane and asphaltene. Fig. 4a
indicates the dissolving ability (Edeca-CO2) is dramatically stronger
than the adsorption ability (Edeca-surf + Edeca-asph). Consequently,
the decane could be quickly dissolved into the scCO2 phase.
As to the dissolution of the asphaltene, after the quick dissolu-
tion of the decane, the driving force (Easph-CO2 + Easph-deca) coming
from the attraction of the miscible CO2/decane phase is slightly lar-
ger than the resistance force (Easph-surf). In principal, the asphaltene
should be slowly dragged into the miscible decane/scCO2 phase
rather than adsorbed onto the silica surface. A detailed explanation
is given below. Generally, if there are hydrogen atoms and nega-
tively charged atoms between two adjacent molecules, they tend
to form H-bonds. In our case, there are three type of strongly neg-
atively charged atoms N (0.6200 e), O1 (0.5700 e) and O2
(0.6500 e) in asphaltene molecule as well as abundant positively
charged hydrogen atoms on the hydroxylated silica surface.
Herein, we ascribed the adsorption of asphaltene molecules to
the formation of hydrogen bond. Fig. 5 shows radical distribution

Please cite this article in press as: Yan Y et al. Enhanced oil recovery mechanism of CO2 water-alternating-gas injection in silica nanochannel. Fuel (2016),
http://dx.doi.org/10.1016/j.fuel.2016.11.019
 Y. Yan et al. / Fuel xxx (2016) xxx–xxx
Fig. 5. Radial distribution functions of H atom of the silica surface with O1, O2 and N
atoms of the asphaltene (shown in Fig. 1b).
functions of H atoms on the silica surface with O1, O2 and N atoms
of asphaltene molecule. It can be observed that the first peak posi-
tions of both H-O1 and H-O2 are approximately same and equal to
2.1 Å, which locate in the regime of typical H-bonds [44]. And,
the high peak value indicates the formation of abundant H-bonds
between asphaltene molecules and silica surface. This point could
be further confirmed by an enlarged interface configuration
between several asphaltene molecules and silica surface (see
Fig. S4 in SI). In general, the H-bond strength is stronger than that
of the non-bond interaction. Hence, the formation of H-bonds
strongly promotes the adsorption of the asphaltene molecules.
However, an interesting phenomenon that some decane mole-
cules still reside inside the nanoslit after injection draws our atten-
tion, as displayed in Fig. 2. Thanks to the high dissolution and quick
diffusion features of scCO2, nearly all decane molecules diffused
into the scCO2 phase to form a miscible slug at initial injecting
stage (3 ns). At this time, it was found that there was a long scCO2
slug at the upstream of the decane/CO2 miscible slug, which would
push this miscible slug across the nanoslit in the next injection.
However, some decane molecules still stayed inside the nanoslit
after injection. This result could be ascribed to the competitive
consequence between the random diffusion of decane and the
expulsing effect of injection fluid. The high diffusion coefficient
of decane in scCO2 induces some decane molecules diffusing
towards the left side; while the injection of scCO2 slug will compel
the dissolved decane molecules moving to right side. So, at low
injection rate (1 m/s) with weak expulsing effect, some decane
molecules could diffuse to left and reside inside the nanoslit (seen
Fig. 2) after injection. In addition, a higher injecting rate (5 m/s)
system was conducted and it was found that the strong expulsing
effect compels most of the decane molecules passing across the
nanoslit and only a few decane molecules residue inside the nano-
slit (see Fig. S5). Nevertheless, a higher injecting rate means a
higher injecting pressure at the expense of more power and an
inevitable high-cost. Meanwhile, it is reasonable to deduce that a
larger volume of scCO2 slug will generate a better displacement
outcome. However, a larger injection volume of scCO2 also implies
a high cost for the unit product.
(2) Water displacement mechanism: As for the water injec-
tion, the comparison of the interaction energies (Fig. 4b) indicates
that there is no adequate driving force for the dissolution of both
the decane and asphaltene. We even tried a larger injecting rate
(5 m/s), and a similar adsorption configuration was also observed
(see Fig. S6). Here, a detailed inspection of the interaction energy
could offer us some insights into the adsorption mechanism.
Please cite this article in press as: Yan Y et al. Enhanced oil recovery mechanism of CO2 water-alternating-gas injection in silica nanochannel. Fuel (2016),
http://dx.doi.org/10.1016/j.fuel.2016.11.019
5
Firstly, owing to the strong interaction and the formed H-bonds
(described above) between asphaltene and silica surface, asphal-
tene could be firmly adsorbed on silica surface. Secondly, decane
has slightly smaller interaction with silica surface than that with
water, which means decane could dissolve into water phase. How-
ever, due to the fact that there is strong interaction between
asphaltene and decane, the adsorbed asphaltene will acts as an
‘‘anchor” for the adsorption of decane. In terms of the dissolution
of decane, the driving force (Edeca-water) is smaller than the resis-
tance force (Edeca-surf + Easph-deca). As a consequence, decane mole-
cules would absorb onto the adsorbed asphaltene molecules,
resulting into the exposure to the water phase. Giving a hypothesis
that there are only decane molecules in the adsorbed oil layer,
based on the value of the interactions, without the ‘‘anchor” effect
of asphaltene molecule, the decane molecules should dissolve into
water phase and is pushed out of the nanoslit [45,46]. Our simula-
tion results indeed confirm this inference (see Fig. S7).
3.2. The displacement behavior in WAG injection
3.2.1. Displacement phenomenon
In the WAG injection, the interaction between water and scCO2
slugs brings out a significantly different displacement perfor-
mance, as shown in Fig. 6. When the scCO2 slug enters into the
nanoslit, most of the apolar decane molecules dissolve into the
scCO2 phase, resulting into the formation of decane/scCO2 miscible
slug. When the frontier interface of water slug reaches the left
entrance of the nanoslit, the whole miscible decane/scCO2 slug is
propelled out of the nanoslit. Compared with that in the case of
single scCO2 (some decane molecules residue inside the nanoslit
after injection), it is unambiguous that the EOR in the CO2 WAG
is improved.
3.2.2. Displacement mechanism
Here, we reveal the displacement mechanism through analyz-
ing the detailed injection process.
First, at the initial stage (0.5 ns), when the scCO2 slug starts to
be injected into the nanoslit which is prefilled with water, the
hydrophobic interaction between the injected scCO2 and adsorbed
oil layer combined with the large viscosity difference between the
scCO2 and water together result in the formation of the backward
meniscus (the right interface between the scCO2 and water slugs at
0.5 and 2 ns in Fig. 6). This induces the injected CO2 entering into
the nanoslit from its two sides; while in scCO2 injection, the scCO2
enters into the nanoslit through its whole channel. As a conse-
quence, for one unit of injected scCO2, these scCO2 in CO2 WAG
injection has higher probability to contact with the adsorbed oil
layer than that in scCO2 injection. Hence, a higher dissolution effi-
ciency of injected scCO2 could be achieved.
Second, after a quick dissolution and diffusion of decane mole-
cules in scCO2 phase, a miscible scCO2/decane frontier slug forms.
The dissolved decane molecules considerably increase the viscosity
of the scCO2 slug, and a semi-flat interface (the right interface
between the scCO2 and water slug at 4 ns in Fig. 6) is induced to
form due to the decrease of viscosity difference between the mis-
cible scCO2/decane slug and the water slug. Compared to the back-
ward meniscus, in which the injected scCO2 preferentially pass
across the channel from two sides of nanoslit and bypass the resid-
ual water, the formation of semi-flat interface will make the
injected scCO2 passing across the nanoslit through the whole chan-
nel, and the residual water will be compelled out. So, once the
semi-flat interface forms, it will lead to an easy discharging of
the residual water. This process is clearly illustrated in the snap-
shots of 2–6 ns in Fig. 6.
In addition, compared with that all oil components shrink into a
semi-spherical configuration in the case of water injection (Fig. 3),
6 Y. Yan et al. / Fuel xxx (2016) xxx–xxx
Fig. 6. Snapshots of displacement process of adsorbed oil layer in CO2 WAG injection.
the dissolution of adsorbed apolar oil components (majority of
crude oil) significantly broadens the fluid channel, which would
largely decrease the injecting pressure [47] and is of the special
significance to save cost for the oil exploration.
Third, we pay attention to the effects of two scCO2/water inter-
faces. As to the front interface, its influence on the scCO2 fluid was
studied. For comparison, the density and velocity distributions of
CO2 inside the nanoslit along y-axis were calculated in scCO2 and
CO2 WAG injections, as depicted in Fig. 7. The configuration at
6 ns, which corresponds to the status that the nanoslit was just
completely filled with injected scCO2 in the case of CO2 WAG,
was selected. The two density peaks near to the silica surface refer
to CO2 adsorption layer, and the corresponding velocity in this
adsorption layer can be found slower than that in the middle nano-
slit. The sharp and high velocity profile in the scCO2 case denotes a
viscous fingering tendency, while the semi-flat and low velocity
Fig. 7. Density and velocity distribution profiles of CO2 in scCO2 and CO2 WAG
injection in nanoslit, which is derived of the statistical average of eleven
configurations in a range of 5.9–6.1 ns. The gray rectangles denote the CO2
adsorption layer.
Please cite this article in press as: Yan Y et al. Enhanced oil recovery mechanism of CO2 water-alternating-gas injection in silica nanochannel. Fuel (2016),
http://dx.doi.org/10.1016/j.fuel.2016.11.019
profile in the WAG case stands for a slow planar flow. The slow pla-
nar scCO2 fluid implies the migration of scCO2 slug is controllable
and results in a prolonged contacting duration between these
scCO2 and adsorbed oil layer. As a result, the dissolution efficiency
of unit volume of injected scCO2 is improved.
As to the back water/scCO2 interface, its function is to push the
scCO2 slug ahead. Due to the quick diffusion velocity of decane
molecules in scCO2 phase, the dissolved decane molecules nearly
diffuse anywhere in the scCO2 slug. But, because of the insolubility
between water and decane, the decane molecules only stay inside
the scCO2 slug to form miscible scCO2/decane slug. Meanwhile, the
stable propelling water/scCO2 interface will push the miscible
scCO2/decane slug forward. Along with the water/scCO2 interface
passing across the nanoslit, all decane molecules merging in the
scCO2 slug will also be pulled out of the nanoslit. However, in
the case of scCO2 injection, there were some decane molecules
merging in the scCO2 slug still stayed inside the nanoslit after
injection (Fig. 2). This displacement consequence exhibits the
advantage of CO2 WAG injection over scCO2 injection.
4. Conclusion
In summary, by employing NEMD simulations, the displacing
processes of adsorbed oil layer in scCO2, water and CO2 WAG injec-
tions were investigated to unveil the EOR mechanism of CO2 WAG.
Simulation results demonstrate that the scCO2 and water play dif-
ferent roles in the displacing process. The scCO2 slug is responsible
for dissolving the apolar oil component, and the water slug pushes
the miscible decane/scCO2 slug out of the nanoslit. Furthermore,
several attracting synergistic phenomena between the scCO2 slug
and water slug, including of high dissolution efficiency, easy dis-
charging of residual water, broadening of the fluid channel, and
desirable displacement consequence, endow the CO2 WAG injec-
tion a low injecting pressure and high oil recovery efficiency. Our
study provides an atomic-level understanding of the CO2 WAG dis-
placing process, and the results may be helpful to recognize the
 Y. Yan et al. / Fuel xxx (2016) xxx–xxx
essence of some engineering phenomena and provide reference for
optimizing CO2 WAG injection engineering.
Acknowledgment
National Basic Research Program of China (2015CB250904),
National Natural Science Foundation of China (U1262202,
51302321), Fundamental Research Funds for the Central Universi-
ties (2015020008, 14CX02222A, 15CX05049A, 15CX05059A).
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.fuel.2016.11.019.
References
[1] Shah A, Fishwick R, Wood J, Leeke G, Rigby S, Greaves M. A review of novel
techniques for heavy oil and bitumen extraction and upgrading. Energy
Environ Sci 2010;3(6):700–14.
[2] Chen Y, He J, Wang Y, Li P. GC-MS used in study on the mechanism of the
viscosity reduction of heavy oil through aquathermolysis catalyzed by
aromatic sulfonic H3PMo12O40. Energy 2010;35(8):3454–60.
[3] Qiao R, Zhang R, Zhu W, Gong P. Lab simulation of profile modification and
enhanced oil recovery with a quaternary ammonium cationic polymer. J Ind
Eng Chem 2012;18(1):111–5.
[4] Wu J, He J, Torsater O, Zhang Z. Effect of nanoparticles on oil-water flow in a
confined nanochannel: a molecular dynamics study. In: SPE international
oilfield nanotechnology conference and exhibition. Society of Petroleum
Engineers; 2012.
[5] Yuan Q, Sun C-Y, Yang X, Ma P-C, Ma Z-W, Liu B, et al. Recovery of methane
from hydrate reservoir with gaseous carbon dioxide using a three-dimensional
middle-size reactor. Energy 2012;40(1):47–58.
[6] Shen Z, Wang H, Li G. Feasibility analysis of coiled tubing drilling with
supercritical carbon dioxide. Petrol Explor Edv 2010;37(6):743–7.
[7] van Bergen F, Gale J, Damen KJ, Wildenborg AFB. Worldwide selection of early
opportunities for CO2-enhanced oil recovery and CO2-enhanced coal bed
methane production. Energy 2004;29(9–10):1611–21.
[8] Hamouda A, Chukwudeme E, Mirza D. Investigating the effect of CO2 flooding
on asphaltenic oil recovery and reservoir wettability. Energy Fuels 2009;23
(2):1118–27.
[9] Orr Jr F, Heller JP, Taber JJ. Carbon dioxide flooding for enhanced oil recovery:
promise and problems. J Am Oil Chem Soc 1982;59(10):810A–7A.
[10] Christensen JR, Stenby EH, Skauge A. Review of WAG field experience. In:
International petroleum conference and exhibition of Mexico. Society of
Petroleum Engineers; 1998.
[11] Kulkarni MM, Rao DN. Experimental investigation of miscible and immiscible
water-alternating-gas (WAG) process performance. J Pet Sci Eng 2005;48
(1):1–20.
[12] Han L, Gu Y. Optimization of miscible CO2 water-alternating-gas injection in
the Bakken formation. Energy Fuels 2014;28(11):6811–9.
[13] Yang Z, Li M, Peng B, Lin M, Dong Z. Dispersion property of CO2 in Oil. 1.
Volume expansion of CO2 + alkane at near critical and supercritical condition
of CO2. J Chem Eng Data 2012;57(3):882–9.
[14] Sun C-Y, Chen G-J. Measurement of interfacial tension for the CO2 injected
crude oil + reservoir water system. J Chem Eng Data 2005;50(3):936–8.
[15] Georgiadis A, Maitland G, Trusler JM, Bismarck A. Interfacial tension
measurements of the (H2O + n-Decane + CO2) ternary system at elevated
pressures and temperatures. J Chem Eng Data 2011;56(12):4900–8.
[16] Botan A, Rotenberg B, Marry V, Turq P, Noetinger B. Hydrodynamics in clay
nanopores. J Phys Chem C 2011;115(32):16109–15.
[17] Zhang L, Lu X, Liu X, Yang K, Zhou H. Surface wettability of basal surfaces of
clay minerals: insights from molecular dynamics simulation. Energy Fuels
2016;30(1):149–60.
[18] Sedghi M, Piri M, Goual L. Atomistic molecular dynamics simulations of crude
oil/brine displacement in calcite mesopores. Langmuir 2016;32(14):3375–84.
[19] Wu H, Chen J, Liu H. Molecular dynamics simulations about adsorption and
displacement of methane in carbon nanochannels. J Phys Chem C 2015;119
(24):13652–7.
[20] Le T, Striolo A, Cole DR. CO2–C4H10 mixtures simulated in silica slit pores:
relation between structure and dynamics. J Phys Chem C 2015;119
(27):15274–84.
Please cite this article in press as: Yan Y et al. Enhanced oil recovery mechanism of CO2 water-alternating-gas injection in silica nanochannel. Fuel (2016),
http://dx.doi.org/10.1016/j.fuel.2016.11.019
7
[21] Liu B, Shi J, Wang M, Zhang J, Sun B, Shen Y, et al. Reduction in interfacial
tension of water–oil interface by supercritical CO2 in enhanced oil recovery
processes studied with molecular dynamics simulation. J Supercrit Fluids
2016;111:171–8.
[22] Liu B, Shi J, Sun B, Shen Y, Zhang J, Chen X, et al. Molecular dynamics
simulation on volume swelling of CO2–alkane system. Fuel
2015;143:194–201.
[23] Zhao L, Tao L, Lin S. Molecular dynamics characterizations of the supercritical
CO2–mediated hexane-brine interface. Ind Eng Chem Res 2015;54
(9):2489–96.
[24] Dong T, Harris NB. Pore size distribution and morphology in the horn river
shale, middle and upper Devonian, northeastern British Columbia, Canada.
AAPG Mem 2013;102:67–79.
[25] Koretsky CM, Sverjensky DA, Sahai N. A model of surface site types on oxide
and silicate minerals based on crystal chemistry; implications for site types
and densities, multi-site adsorption, surface infrared spectroscopy, and
dissolution kinetics. Am J Sci 1998;298(5):349–438.
[26] Zhang Y, Wang J, Morrow NR, Buckley JS. Effect of synthetic drilling fluid base
oils on asphaltene stability and wetting in sandstone cores. Energy Fuels
2005;19(4):1412–6.
[27] Puntervold T, Strand S, Austad T. New method to prepare outcrop chalk cores
for wettability and oil recovery studies at low initial water saturation. Energy
Fuels 2007;21(6):3425–30.
[28] Wu G, He L, Chen D. Sorption and distribution of asphaltene, resin, aromatic
and saturate fractions of heavy crude oil on quartz surface: molecular dynamic
simulation. Chemosphere 2013;92(11):1465–71.
[29] Wang P, Dong Z-j, Tan Yq, Liu Zy. Investigating the interactions of the saturate,
aromatic, resin, and asphaltene four fractions in asphalt binders by molecular
simulations. Energy Fuels 2015;29(1):112–21.
[30] National Institute of Standards and Technology. Thermophysical properties of
fluid systems; 2011. <http://webbook.nist.gov/chemistry/fluid/>.
[31] Phan A, Cole DR, Striolo A. Aqueous methane in slit-shaped silica nanopores:
high solubility and traces of hydrates. J Phys Chem C 2014;118(9):4860–8.
[32] Plimpton S. Fast parallel algorithms for short-range molecular dynamics. J
Comput Phys 1995;117(1):1–19.
[33] Harris JG, Yung KH. Carbon dioxide’s liquid-vapor coexistence curve and
critical properties as predicted by a simple molecular model. J Phys Chem
1995;99(31):12021–4.
[34] Toukan K, Rahman A. Molecular-dynamics study of atomic motions in water.
Phys Rev B 1985;31(5):2643.
[35] Cygan RT, Liang J-J, Kalinichev AG. Molecular models of hydroxide,
oxyhydroxide, and clay phases and the development of a general force field.
J Phys Chem B 2004;108(4):1255–66.
[36] Vorobyov I, Anisimov VM, Greene S, Venable RM, Moser A, Pastor RW, et al.
Additive and classical drude polarizable force fields for linear and cyclic ethers.
J Chem Theory Comput 2007;3(3):1120–33.
[37] Zoete V, Cuendet MA, Grosdidier A, Michielin O. SwissParam: a fast force field
generation tool for small organic molecules. J Comput Chem 2011;32
(11):2359–68.
[38] Nosé S. A molecular dynamics method for simulations in the canonical
ensemble. Mol Phys 1984;52(2):255–68.
[39] Darden T, York D, Pedersen L. Particle mesh Ewald: an Nlog (N) method for
Ewald sums in large systems. J Chem Phys 1993;98(12):10089–92.
[40] Humphrey W, Dalke A, Schulten K. VMD: visual molecular dynamics. J Mol
Graphics 1996;14(1):33–8.
[41] Li X, Xue Q, Zhu L, Jin Y, Wu T, Guo Q, et al. How to select an optimal surfactant
molecule to speed up the oil-detachment from solid surface: a computational
simulation. Chem Eng Sci 2016;147:47–53.
[42] Fang T, Shi J, Sun X, Shen Y, Yan Y, Zhang J, et al. Supercritical CO2 selective
extraction inducing wettability alteration of oil reservoir. J Supercrit Fluids
2016;113:10–5.
[43] Krynicki K, Green CD, Sawyer DW. Pressure and temperature dependence of
self-diffusion in water. Faraday Discuss Chem Soc 1978;66:199–208.
[44] Kunieda M, Nakaoka K, Liang Y, Miranda CR, Ueda A, Takahashi S, et al. Self-
accumulation of aromatics at the oil–water interface through weak hydrogen
bonding. J Am Chem Soc 2010;132(51):18281–6.
[45] Liu Q, Yuan S, Yan H, Zhao X. Mechanism of oil detachment from a silica
surface in aqueous surfactant solutions: molecular dynamics simulations. J
Phys Chem B 2012;116(9):2867–75.
[46] Zhang P, Xu Z, Liu Q, Yuan S. Mechanism of oil detachment from hybrid
hydrophobic and hydrophilic surface in aqueous solution. J Chem Phys
2014;140(16):164702.
[47] Nelson PH. Pore-throat sizes in sandstones, tight sandstones, and shales. AAPG
Bull 2009;93(3):329–40.