Displacement Mechanism of Oil in Shale Inorganic Nanopores by

Dec 14, 2016 - Supercritical CO2 (scCO2), as an effective displacing agent and clean fracturing fluid, exhibits a great potential in enhanced oil reco...
2 downloads 24 Views 5MB Size
Subscriber access provided by NEW YORK UNIV

Article

Displacement mechanism of oil in shale inorganic nanopores by supercritical carbon dioxide from molecular dynamics simulations Bing Liu, Chao Wang, Jun Zhang, Senbo Xiao, Zhiliang Zhang, Yue Shen, Baojiang Sun, and Jianying He Energy Fuels, Just Accepted Manuscript • DOI: 10.1021/acs.energyfuels.6b02377 • Publication Date (Web): 14 Dec 2016 Downloaded from http://pubs.acs.org on December 19, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Energy & Fuels is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

1

Displacement mechanism of oil in shale inorganic nanopores by

2

supercritical carbon dioxide from molecular dynamics simulations

3

Bing Liu a, b,*, Chao Wang a, Jun Zhang a, Senbo Xiao b, Zhiliang Zhang b, Yue Shen a, Baojiang

4

Sun c, Jianying He b,*

5

a

School of Science, China University of Petroleum, Qingdao 266580, Shandong, China

6

b

NTNU Nanomechanical Lab, Department of Structural Engineering, Norwegian University of

7

Science and Technology (NTNU), Trondheim, 7491, Norway

8

c

9

China

School of Petroleum Engineering, China University of Petroleum, Qingdao 266580, Shandong,

10

* Corresponding authors:

11

Email:[email protected]

12

Email:[email protected]

13

1

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

14

ABSTRACT: Supercritical CO2 (scCO2), as an effective displacing agent and clean fracturing

15

fluid, exhibits a great potential in enhanced oil recovery (EOR) from unconventional reservoirs.

16

However, the microscopic translocation behavior of oil in shale inorganic nanopores, has not been

17

well understood yet in the scCO2 displacement process. Herein, non-equilibrium molecular

18

dynamics (NEMD) simulations were performed to study adsorption and translocation of

19

scCO2/dodecane in shale inorganic nanopores at different scCO2 injection rates. The injected

20

scCO2 preferentially adsorb in proximity of the surface and form layering structures due to

21

hydrogen bonds interactions between CO2 and -OH groups. A part of scCO2 molecules in the

22

adsorption layer retain the mobility, due to the cooperation of slippage, Knudsen diffusion, and

23

imbibition of scCO2. The adsorbed dodecane are separated partly from the surface by scCO2, as a

24

result the competitive adsorption between scCO2 and dodecane, and thus enhancing the mobility

25

of oil and improving oil production. In the scCO2 displacement front, interfacial tension (IFT)

26

reduction and dodecane swelling enhance the mobilization of dodecane molecules, which plays

27

the crucial role in the CO2 EOR process. The downstream dodecane, adjacent to the displacement

28

front, are found to aggregate and pack tightly. The analysis of contact angle, meniscus and

29

interfacial width shows that the small scCO2 injection rate with a large injection volume is

30

favorable for CO2 EOR. The morphology of meniscus changes in the order convex-concave-CO2

31

entrainment with the increase of the injection rate. The microscopic insight provided in this study

32

is helpful to understand and effectively design CO2 exploitation of shale resources.

33

KEYWORDS: shale inorganic nanopore, adsorption and translocation, scCO2 displacement,

34

injection rate

35

2

ACS Paragon Plus Environment

Page 2 of 20

Page 3 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

36 37

1. INTRODUCTION

38

The growing demand for crude oil has inspired the innovative technologies to develop enhanced

39

oil recovery (EOR) methods involving the injection of CO2, N2, hydrocarbon gases or chemicals

40

into oil reservoirs. Among these EOR methods, CO2 EOR has become increasingly important

41

because CO2 flooding can effectively improve the oil recovery factor (RF) and considerably

42

mitigate greenhouse gas emissions. 1 The oil recovery rate can be enhanced by 8-16% of the OOIP

43

through CO2 tertiary processes. 2, 3 CO2 injection, in contrast to water flooding, is recognized as a

44

promising agent to EOR in shale reservoirs due to favorable properties such as low viscosity and

45

high injectivity.4 Gamadi et al. studied cyclic CO2 injection for improving oil recovery and

46

observed a maximum increase of recovery from 33% to 85%.5 CO2, as the clean fracturing fluid,

47

can create better fractured networks compared to water.6, 7

48

The commonly-recognized oil recovery mechanisms include oil viscosity reduction, oil swelling,

49

interfacial tension (IFT) reduction, and light-hydrocarbons extraction.8, 9 The CO2 EOR process is

50

quite different for shale reservoirs because CO2 is expected to flow through natural and produced

51

fractures rather than the non-fractured rock matrix. Hawthorne et al. showed that mobilization of

52

light oil components into CO2 is a dominant recovery process rather than dissolution of CO2 into

53

bulk oil.10 Alharthy et al. indicated that molecular diffusion and advective mass migration across

54

the fracture-matrix interface are the main mechanisms for incremental production of oil in shale

55

nanopore.11 Tovar et al. reported that the main recovery mechanism is the vaporization of the

56

hydrocarbons into CO2.12 Wang et al. found that the dissolution of CO2 induces an increase in

57

oil/gas relative permeability due to the reduced viscosity of CO2 saturated oil phase.13 However, it

58

is difficult to speculate the exact mechanism based on experiments due to the universal existence

59

of nanopores in shale 2-100nm.14-17 In nanopores, microstructural, dynamical and thermophysical

60

behavior of the confined fluids differ dramatically from their bulk counterparts. The transport

61

properties in nanochannel is significantly affected by the width of channel, system temperature,

62

hydrophobicity/hydrophilicity, and roughness of wall.18-20 The stress anisotropy is also observed in

63

a fluid confined in a nanochannel.21 These behaviors result from the molecular asymmetry

64

between fluid-fluid and fluid-solid interactions displaying inhomogeneous density distributions of

65

the local fluid, which can be demonstrated by molecular simulation techniques.22-24 Thu et al. 3

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

66

investigated structural and dynamic properties of pure propane and the relation between structure

67

and dynamics of CO2-C4H10 mixtures in slit-shaped silica pores using molecular dynamics (MD)

68

simulations.25, 26 Wu et al.27 and Yuan et al.28 investigated displacement of methane with CO2 in

69

carbon nanochannels from MD simulations. Jin et al. studied methane flow in shale nanopores by

70

non-equilibrium molecular dynamics (NEMD) simulations.29 Wang et al. investigated octane

71

transport in shale inorganic nanopores from NEMD simulations.30 The results showed that much

72

progress has been achieved in adsorption and transport of oil or gas in nanopores. However, it

73

remains challenging for adsorption, migration, interfacial and confinement properties of scCO2/oil

74

during the CO2 driving oil transport process inside shale nanopores.

75

In this work, non-equilibrium molecular dynamics (NEMD) simulations were performed to

76

study the microscopic mechanism of scCO2 displacing oil inside the shale inorganic nanopore.

77

Firstly, we investigate the adsorption and transport of scCO2 in the nanopore. Then, we examine

78

detachment and translocation of dodecane in the nanopore. In addition, we study the effects of IFT

79

and oil volume swelling on oil migration in the nanopore. Finally, we investigated the effect of

80

injection rate on the transport of scCO2 and dodecane in the nanopore. Our results would help to

81

understand and design effective CO2 exploitation of shale resources as well as provide physical

82

insight into CO2 storage, nanoscale fluid flow, and surface cleaning.

83

2. MODELS and METHODOLOGY

84

To construct MD simulation models, the nanopore with a diameter of 30 Å was created by

85

digging a cylindrical hole in a cristobalite supercell31, 32 in the size of 42.96 × 42.96 × 153 Å. The

86

nanopore with hydroxylated surfaces of silica was fixed in all simulations. Initially, 90 dodecane

87

molecules represent oil which were placed inside the left end of the nanopore. An equilibrium MD

88

simulation was carried out for 2000 ps to obtain a steady-state distribution of oil in the nanopore

89

(supporting information Figure. S1). Then the nanopore was connected with 2197 CO2 molecules,

90

which randomly distributed in a cuboid system with edge dimension of 42.96 × 42.96 × 106 Å3.

91

Periodic boundary conditions (PBC) were applied in three dimensions. A vacuum of 26 Å long

92

was constructed along the z direction to eliminate the boundary effect33. The created model is

93

shown in Figure. 1.

94

All MD simulations were performed using LAMMPS software34 in the NVT ensemble with a

95

time step of 1fs. The Nosé-Hoover thermostat35, 36 was employed to control the temperature at 318 4

ACS Paragon Plus Environment

Page 4 of 20

Page 5 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

96

K. The nanopore was modeled with the consistent valence force field (CVFF)32, 37. CO2 was

97

modeled with the EPM2 model which was successfully demonstrated to reproduce critical point of

98

CO2 in experiment.38, 39 The dodecane representing oil was modeled by the CHARMM force

99

field40, 41. The cutoff of non-bonded interaction was set to be 9.5 Å. Long-range electrostatic

100

interaction was calculated using the Ewald summation method42, 43. The precision of calculation

101

was 0.0001. The nonbond energy Eij between two atoms i and j was the sum of Lenard-Jones (LJ)

102

and electrostatic potential energy:

Eij = 103

qi q j rij

 σ ij + 4ε ij    rij 

12 6   σ ij    −     r     ij  

(1)

104

where qi and qj were the charges of the atoms i and j, respectively; rij represented the separation

105

between two atoms i and j; and εij and σij represented the LJ potential well depth and the

106

zero-potential distance. The interaction between different types of atoms was determined by the

107

Lorentz-Berthelot rule: σij = (σii + σjj)/2 and εij = (εii εjj)1/2. σ and ε values employed in the interaction

108

calculations were given in Table 1 of supporting information.

109

NEMD simulations were performed to simulate the process of dodecane displaced by scCO2 in

110

the nanopore. A rigid graphene sheet was placed at the left side of scCO2, and moved at rates of 2,

111

4, 6, 8 and 10m/s in the z direction, which drove scCO2 to enter into the nanopore and displace

112

dodecane molecules. Corresponding to the injection rates, the simulations time were 4.0, 2.0, 1.5,

113

1.0 and 0.8 ns, respectively.

114 5

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

115

Figure. 1. The MD simulation model. (A) Plane view of atomistic cristobalite cylindrical nanopore with diameter D=3nm. (B) Dodecane

116

molecule (marked in red color) and CO2 molecule (marked in blue color). (C) 90 dodecane molecules inside the nanopore. (D) The model

117

of CO2- dodecane in nanopore.

118

3. RESULTS and DISCUSSIONS

119

3.1 Adsorption and transport of CO2

120

To characterize the adsorption of CO2, in Figure 2(A), we report results for the radial number

121

profiles of CO2 in the nanopore at the injection rate of 2 m/s for 100, 800, 1000 and 1500 ps,

122

respectively. The profile exhibits two well defined peaks at 14 Å and 11 Å, indicating the formation of

123

two adsorbed monolayers of CO2 in the range of 9 Å to the nanopore surface. The ordering

124

construction at 14 and 11Ǻ is a result of the small dimensions of the nanochannel as described in works

125

of Giannakopoulos44 and Holland45. The time evolution of the profile shows the increase in the CO2

126

number in the first layer (14 Å), the second layer (11 Å), and the central area of the pore. This is well

127

understood because the movable graphite plate squeezes more and more CO2 into the nanopore as time

128

is longer. However, the peaks of two adsorbed layers, especially for the first layer, increase rapidly

129

compared to the central area. The increment in these three regions is evidently observed to follow the

130

sequence the first layer > the second layer > the central area. It indicates that CO2 molecules

131

preferentially adsorb in proximity of the surface, form layering structures, and subsequently fill the

132

nanopore. It suggests that two typical stages, before and after pore filling, are involved in the

133

adsorption process.

134

The strong layering of CO2 molecules in the nanopore, clearly depicted as higher peaks in the

135

corresponding radial profiles, depends on the nature of the solid surface and CO2. This behavior results

136

from the favorable hydrogen bonds formed by O atoms in CO2 with H atoms in -OH groups. Moreover,

137

supercritical CO2 becomes more polar albeit CO2 is normally nonpolar. This also enhances the

138

electrostatic interaction between CO2 and the surface, leading to more CO2 adsorbing on the surface. In

139

this work, a hydrogen bond is defined to exist if the O-O distance is less than 3.5 Å and simultaneously

140

the angle of H-O-O is less than 30°46. To visualize the hydrogen bond, a typical snapshot is described

141

by the inserted Figure 2a. It shows the formation of hydrogen bonds between CO2 and the hydroxylated

142

surface located at different positions. Inserted Figure 2b depicts time evolution of the number of

143

hydrogen bonds. It is evident that the number of hydrogen bonds increases with time, demonstrating

144

the strong interaction between CO2 and the surface responsible for the adsorption of CO2, consistent 6

ACS Paragon Plus Environment

Page 6 of 20

Page 7 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

145

with the tendency in Figure 2(A).

146

To visualize transport of CO2 molecules, the typical snapshots for CO2 in the nanopore are provided

147

in Figure 2(B). It shows that CO2 molecules advance faster along the hydroxylated surface than in the

148

nanopore center. To quantify CO2 migration in the nanopore, the velocity profiles of CO2 in the

149

nanopore as a function of the distance from the solid surface are presented in Figure 2(C) at 1000 and

150

2000 ps. Two symmetrical distinct peaks of the velocity profiles occur in the vicinity of the solid

151

surface, showing CO2 molecules move forwards faster along the solid surface than in the center of the

152

nanopore. The steric hindrance of dodecane molecules results in that the CO2 velocity in the central

153

area slows down. However, the results also show that the adsorbed CO2 molecules are not immobile but

154

still retain their mobility. Non-zero CO2 velocity on the surface means the slip effect contributing to the

155

flow enhancement of CO2 near the surface. Knudsen diffusion also improves the migration of adsorbed

156

CO2 due to much smaller nanopore size than the mean free path of CO2 molecules29. In addition, the

157

hydroxylated surface is CO2-philic due to that CO2 can displace water from the surface with silanol

158

groups.

159

injection rate on adsorption, we present the radial profiles of CO2 molecules in the nanopore at different

160

injection rates with same injection volume, as shown in Figure 2(D). Two peaks occur at 14 Å and 11 Å,

161

indicating two adsorbed layer structures for CO2 molecules and no effect of injection rate on the

162

location of the adsorbed layer. Peak heights of the profiles decrease with the increase in the injection

163

rate, suggesting that CO2 adsorption enhances at low injection rate and weakens at high rate. Because

164

of constant injection volume and cross section area of the nanopore, low injection rate means low CO2

165

loading per unit time, low injection pressure, and long interaction time between CO2 and the surface.

166

The influence of temperature on the adsorption number of CO2 is investigated as shown in supporting

167

information Figure S2. It is found that two adsorbed layers of CO2 on the surface occurring at 14 Å and

168

11 Å, but temperature has no evident effect on the location of the adsorbed layer. In addition, the peaks

169

decrease with the increase of temperature. This can be ascribed to the higher mobility of CO2 at high

170

temperature and thus the fewer number of CO2 intend to adsorb on the rock. Similar results have been

171

previously reported by Giannakopoulos et al44. Thus, at low injection rate, more CO2 molecules

172

strongly adsorb on the solid surface. While at high injection rate, more CO2 molecules migrate in the

173

central area of the nanopore, reducing the adsorption on the solid surface.

47

This aids imbibition of CO2 into the hydroxylated nanopore.

7

ACS Paragon Plus Environment

To investigate the effect of

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

174 175

Figure 2. (A) Radial distribution of number density of CO2 at gas injection rate of 2 m/s at 100, 800, 1000, and 1500ps; Inserted Figure (a)

176

depicts the adsorption configuration of CO2 molecules around silanol groups on silica surface in which the yellow, red, blue and white

177

sphere represent silicon, oxygen, carbon and hydrogen atoms, respectively; Inserted Figure (b) shows the number of H-bond between CO2

178

and the surface with time. (B) The snapshots of CO2 transport in shale pore at 400ps and 1000ps. (C) The velocity profiles of CO2

179

transport at 1000ps and 2000ps. Red dash lines represent the position of CO2/rock interface; (D) Radial distribution of number density of

180

CO2 at different gas injection rate.

181

To study the effect of the injection rate on the velocity profile, the velocity profiles of CO2 at the

182

rates of 2, 4 and 8 m/s are calculated, respectively, as shown in Figure S3 of supporting information. It

183

shows that curvature of parabola decreases with the increase of injection rate. Obviously, with the

184

increase in the injection rate, the velocity of CO2 in the central area of the pore increases rapidly while

185

the velocity of CO2 near the surface decreases especial for the larger injection rate of 8 m/s. The similar

186

results have also been reported by Duan48 et al. It indicates that CO2 molecules preferentially adsorb in

187

proximity of the surface and subsequently fill the nanopore. Therefore, two typical stages, before and

188

after pore filling, are involved in the adsorption process. We can infer that the flow in the central area

189

of the nanopore will be dominant when it reaches saturated adsorption and steady flow state. The

190

results reveal that adsorption of fluids on the surface plays a significant influence on their transport in a 8

ACS Paragon Plus Environment

Page 8 of 20

Page 9 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

191

nanopore.

192

3.2 Detachment and transport of dodecane

193

3.2.1 Detachment of dodecane

194

In Figure 3(A), we report the radial profiles of molecular number for dodecane in the nanopore at

195

100, 800, 1000 and 1500 ps, as CO2 injection rate is 2 m/s. There is a distinct peak at 13 Å

196

corresponding to the adsorbed dodecane layer. The peak of the adsorbed layer decreases with CO2

197

injection time, implying parts of adsorbed dodecane molecules are gradually detached from the surface.

198

The detached dodecane molecules transfer to the central area of the nanopore and thus increases the

199

amount of dodecane there, shown by the rising curves of the central area. To visualize the dodecane

200

detachment from the surface due to CO2, the representative snapshots for adsorbed dodecane molecules

201

in green at 1, 30, 50, 150 ps are presented in Figure 3(B). It shows that parts of adsorbed dodecane

202

molecules are separated further away from the surface and migrate towards the central area of the

203

nanopore. With the increase in dodecane molecules detached from the surface, moveable oil in the

204

nanopore improves. This results in the enhancement of oil recovery during the exploiting process.

205

Therefore, CO2 detaching oil from the surface plays a significant role in CO2 EOR.

206

The dodecane detachment is highly linked to the adsorption of CO2 on the surface. Comparing

207

Figure 3(A) with Figure 2(A), we observe the dodecane peak is always farther from the surface than the

208

peak of the first adsorption layer of CO2, also indicating preferential CO2 adsorption on the surface.

209

Then, the average adsorption strength of CO2 to the hydroxylated surface plays a crucial role in

210

dodecane detachment or dodecane transport process. To evaluate the average adsorption strength of

211

CO2 molecules, we calculate the interaction energy between CO2 molecules and the nanopore surface

212

from equation (2)49.

Eco2 / rock =

213

214 215

where E co

2 / rock

Etotal − ( Eco2 + Erock )

(2)

N co2

is the interaction energy between CO2 molecules and the surface, N co is the number

of CO2 molecules,

2

E co2 and Erock represent the energy of CO2 and the rock surface, and

216

E total denotes the total energy of the system containing CO2. Similarly, we can calculate the interaction

217

energy between dodecane and the surface to characterize the average adsorption strength of dodecane

218

molecules. Inserted figure in Figure 3(A) depicts the variation of the interaction energy of surface-CO2 9

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

219

and that of surface-dodecane. It can be found that the interaction energy between CO2 and surface

220

increases with CO2 injection time, in agreement with the trend of increasing hydrogen bonds in inserted

221

Figure 2b. It indicates the enhancement of CO2 adsorption on the surface. However, the interaction

222

energy between dodecane and surface reduces, meaning the extent of dodecane adsorption on the

223

surface decreases. After 2300 ps, the interaction energy between CO2 and surface is larger than that

224

between dodecane and surface. It suggests that the capability of CO2 to adsorb on the surface is

225

stronger than dodecane. It is a favorable effect on the dodecane detachment.

226

3.2.2 Transport of dodecane

227

From above analysis of dodecane detachment, we conclude that one of transport mechanisms for oil

228

in shale is that adsorbed dodecane molecules are detached partly from the surface and migrate into the

229

central area of the nanopore. To further visualize transport of dodecane molecules inside the nanopore,

230

several typical snapshots for the process of CO2 displacing oil at injection rate of 2 m/s are presented in

231

Figure 3(C). It shows that dodecane molecules move to the right of the nanopore and are gradually

232

expelled from the nanopore.

233

When the movable graphite plate moves to the right, it compresses CO2 to generate a driving force

234

inducing the transport of oil droplet inside the nanopore. With the CO2 molecules filling the nanopore

235

and occupying the cavity space, the equilibrium of dodecane breaks. The interactions between CO2 and

236

the solid surface, and between CO2 and dodecane dominate the structure and the transport of dodecane

237

within the nanopore. The structural variation of oil droplet is also observed in Figure 3(C), showing

238

two distinct zones. One is formed in the displacement front, where CO2 is miscible with dodecane. The

239

other is the downstream oil zone adjacent to the displacement front. To evaluate the structure of oil

240

droplet, the axial density distribution of dodecane molecules is plotted in Figure 3(D). It shows that the

241

density peak positions shift to the right of the z axis, indicating movement of the oil droplet toward the

242

right side of the nanopore. For the part of oil droplet in the miscible zone, the density distribution

243

shows an increasing span of the curve and a decreasing peak. It indicates the oil volume swelling and

244

the interfacial variation, which have a significant influence on oil recovery (we will discuss it in section

245

3.3). For the part of oil droplet in the downstream oil zone, the density distribution shows a constant

246

span of the curve and an increasing peak. It indicates that dodecane molecules adjacent to the

247

displacement front aggregate and pack tightly, resulting in the formation of oil column. This is mainly

248

caused by the attractive interaction among dodecane molecules, and the driving force resulting from the 10

ACS Paragon Plus Environment

Page 10 of 20

Page 11 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

249

oil swelling in the miscible zone and CO2 flooding. Therefore, the dodecane molecules adjacent to the

250

displacement front migrate forward in the form of oil column within the nanopore, which is beneficial

251

to the displacement of oil droplet. In addition, the RDF between C atoms in different dodecane

252

molecules in the downstream oil zone, reported in Figure S4 in the supporting information, exhibits the

253

increasing peak and area below the curve with time. The result indicates the higher coordinate number

254

(dodecane) around a dodecane molecule and thus the tighter distribution for dodecane molecules.

255

256 257

Figure 3. (A) Radial distribution of number density of CO2 at gas injection rate of 2 m/s at 100, 800, 1000, and 1500ps; Inserted Figure

258

shows the interaction energy between CO2 and the surface (black line), and that between dodecane and the surface (red line); (B)

259

Snapshots of the process of CO2 detach dodecane in the nanopore at 1, 30, 50, 150ps; (C) Snapshots of CO2 displacing dodecane at gas

260

injection rate 2m/s at (a) 1, (b) 200, (c) 500, and (d) 5000ps; and (D) Density distribution of dodecane molecules (a) in the miscible zone

261

and (b) in the oil zone at 1, 500 and 1000 ps at the injection rate of 2 m/s.

262

3.2.3 Effect of injection rate on detachment and transport of dodecane

263

To quantitatively describe the effect of injection rate on detachment and transport of dodecane, we

264

calculate the radial number density, the axial number density and the transport distance of center of

265

mass (COM) for oil droplet at different injection rates, as shown in Figure 4(A), Figure 4(B) and Figure 11

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

266

4(C), respectively.

267

Figure 4(A) shows that a distinct peak always occurs at 13 Å for the injection rate of 4, 6, 8 or 10

268

m/s, which lowers with the decrease in the injection rate. At the injection rate of 2 m/s, no distinct peak

269

occurs in the vicinity of the surface. The results show that small injection rate is of benefit to the

270

detachment of dodecane molecules adsorbed on the surface, while high injection rate is adverse to

271

separate dodecane molecules from the surface. It implies that more residual oil molecules will remain

272

in the nanopore at the higher injection rate, resulting in the loss of oil during the exploiting process.

273

Combining Figure 4(A) with Figure 2(D), we may expect that entrainment or breakthrough of CO2 will

274

occur at higher injection rate. In this case, CO2 bubble moves steadily and leaves behind a thin oil film

275

close to the surface. The thickening residual oil film in the presence of mass transfer leads to a loss in

276

the oil recovery.

277

278 279

Figure 4. (A) Radial distribution of number density of dodecane at different gas injection rate; (B) Number density distribution of

280

dodecane in the axial direction of the nanopore at different CO2 injection rates; (C) Displacement of COM of dodecane droplet at CO2

281

injection rate from 2 to10 m/s.

12

ACS Paragon Plus Environment

Page 12 of 20

Page 13 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

282

Figure 4(B) shows that the peak of the axial number density decreases and the span of the curve

283

increases when decreasing the injection rate from 10 to 2 m/s. It indicates looser packing and wider

284

distribution region for dodecane molecules within the nanopore at smaller injection rate. The peak

285

position shifts to the right of the z axis, implying that the displacement distance of the oil droplet

286

increases with the decrease in the injection rate. The results demonstrate that smaller injection rate is

287

preferable to oil displacement in CO2 flooding, which is also confirmed by Figure 4(C). The figure

288

shows the biggest displacement of COM of oil droplet at the injection rate of 2 m/s as CO2 injection

289

volume is constant. Therefore, a small injection rate with a large gas volume is favorable to CO2 EOR.

290

3.3 Interfacial behavior of CO2 and dodecane

291

3.3.1 Interface between CO2 and dodecane

292

The interfacial behavior, playing an important role in studying two-phase flow, dominates the

293

efficiency of CO2 EOR lying on the achievement of the miscibility front between CO2 and oil.50 Figure

294

5(A) shows the formation of the miscible zone as well as the variation of its width as a function of time.

295

The formation of the miscible zone consists of three processes: (1) detached dodecane molecules

296

migrate into CO2; (2) bulk dodecane molecules dissolve in CO2; and (3) CO2 penetrate into bulk

297

dodecane. The processes (1) and (2) indicate the translocation of dodecane molecules in the direction

298

opposite to the flow direction, compressing the domain through which CO2 molecules permeate into

299

dodecane at the displacement front. This lowers the penetration of CO2 into dodecane, imparting

300

stability to a displacement process in oil recovery. Thus mobilization of dodecane molecules into CO2

301

has a crucial effect on this zone or interface rather than dissolution of CO2 into bulk dodecane.

302

We calculate the width of the interfacial region to examine the variation of the interface between

303

CO2 and dodecane during CO2 flooding process, as shown in Figure 5(A). The interfacial region is

304

established by the “10-90” interfacial width characterized by the distance L between two surfaces at the

305

10% and 90% of the dodecane density along the z axis.51 It is shown that the interfacial widths at 100

306

ps and 800 ps for the injection rate of 2 m/s are L1 = 20 Å and L2 = 50 Å, respectively. The obvious

307

increase in the interfacial width means the extension of the coexistent region of CO2 and dodecane,

308

leading to the reduction in IFT. The lower IFT improves the transport of dodecane molecules, which is

309

favorable for EOR. We also calculate the receding contact angle of oil droplet to examine the

310

morphology of the interfacial meniscus52, reported in Figure 5(B). The contact angle θ is calculated by

311

equation (3). 13

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

θ = tan −1 ( 312

r 2 − h2 ) 2rh

Page 14 of 20

(3)

313

where r is the radius of the nanopore, and h is the height of the meniscus53, as shown in Figure 5(B).

314

The contact angle is plotted in Figure 5(B), showing a nearly constant value of 175° due to the

315

hydroxylated surface is oleophobic.

316 317

Figure 5. (A) The snapshot of the miscible zone; Density profile of CO2-doecane system at injection rate of 2m/s; The label L1, L2 is the

318

“10-90” interface width at T=100ps and 800ps; (B) Representative meniscus profile: the line shows the circular fit and the contact angle;

319

Contact angle θ between the dodecane droplet and the surface at different CO2 injection rate with time.

320

3.3.2 Oil swelling

321

When CO2 is injected into oil-bearing reservoirs, interactions between CO2 and oil will induce

322

favorable effects on the oil recovery, such as oil volume swelling, oil viscosity decrease and miscible

323

CO2-oil. To investigate oil volume swelling stimulated by CO2, several snapshots are reported in Figure

324

6(A), showing the variation of dodecane volume at injection rate of 2 m/s. It is observed that three

325

processes contribute to the oil volume expansion, including that (1) adsorbed dodecane molecules are

326

separated from the surface migrating into the displacement front; (2) upstream dodecane molecules

327

adjacent to the interface disturbed by CO2 disentangle and dissolve into CO2, offering available space

328

for CO2 diffusion and thus enhancing their miscibility; and (3) adsorbed dodecane molecules in the

329

downstream front are detached from the surface and subsequently move to the central area of the

330

nanopore, providing available space for the back dodecane molecules. Volume swelling offers an

331

additional driving to expel the downstream oil toward the right of the nanopore. On the other hand, it 14

ACS Paragon Plus Environment

Page 15 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

332

reduces oil viscosity and thus improving oil mobility. To characterize the enhanced mobility, we

333

calculate the self-diffusion coefficients of dodecane at different time using the mean-squared

334

displacement approach based on the Einstein relation54, reported in Figure 6(B). The result shows that

335

the self-diffusion coefficient increases as the time, indicating the increased mobility of dodecane, as

336

expected.

337 338

Figure 6. (A) Snapshots of the process of CO2 swelling dodecane in the nanopore at 300 and 800 ps; (B) Diffusion efficient of dodecane

339

at injection rate of 2m/s at different time; RDF of (C) C(dodecane)-C(dodecane) and (D) C(dodecane)-C(CO2) at CO2 injection rate of 2

340

m/s at 500, 1000, 1500 and 2000 ps.

341

The volume swelling embodies oil molecule disentanglement and loose distribution in the presence

342

of CO2, which can be described by the radial distribution function (RDF). Figure 6(C) and (D) depict

343

the RDFs between C atoms in different dodecane molecules and that between CO2 and dodecane,

344

respectively. Figure 6(C) shows that the peak of RDF and the area below the curve decrease with time.

345

It suggests the decrease in the coordinate number (dodecane) around one dodecane molecule or the

346

loose distribution of dodecane molecules. Opposite tendency is observed in Figure 6(D), showing that

347

the coordinate number (CO2) around one dodecane molecule increases with time. The results imply that 15

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

348

more and more CO2 accumulate around dodecane molecules, resulting in the larger average separation

349

between dodecane molecules. The larger separation among dodecane molecules indicates a reduced oil

350

viscosity and enhanced oil mobility.

351

3.3.3 Effect of injection rate on the interface between CO2 and dodecane

352 353

Figure 7. (A) Contact angle θ between the dodecane droplet and the surface at different scCO2 injection rate with time; (B) Snapshots of

354

the meniscus curvature at the same injection volume of CO2 but different injection rates of 2, 4, 6, 8, and 10 m/s; (C) Density profile of

355

CO2-doecane system at injection rate of 2m/s and 8m/s; (D) RDF of C(dodecane)-C(dodecane) at different CO2 injection rate. The

356

inserted snapshots of CO2 displacing oil at the same injection volume but different injection rate (a) 10 and (b) 2 m/s.

357

To study the influence of injection rate on the interface between CO2 and dodecane, we firstly

358

calculate the receding contact angle of oil droplet at different injection rates, reported in Figure 7(A).

359

The corresponding snapshots describing the meniscus morphology are presented in Figure 7(B). In

360

order to investigate whether the vacuum area in the right hand side of the simulation model would

361

influence the morphology of meniscus or not, we have made a complement to the initial model. The

362

results show that the vacuum space has little or no influence on the morphology of meniscus when

363

there is enough amount of oil in the nanopore (supporting information Figure S5). Figure 7(A) shows 16

ACS Paragon Plus Environment

Page 16 of 20

Page 17 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

364

that the contact angle is always larger than 150° due to hydrophilicity of the hydroxylated surface; it is

365

independent of time at constant injection rate; it decreases with the increasing rate due to the

366

weakening adsorption and migration of CO2 along the surface as the injection rate increases. Figure

367

7(B) shows that high injection rate attenuates the convex extent of interfacial meniscus. The results

368

imply a tendency for the morphology of meniscus that it would alter in the order convex-concave-CO2

369

entrainment when the injection rate gradually increases (supporting information Figure S6).

370

Then we calculate the interfacial width at the rate of 2 m/s and 8 m/s with same injection volume,

371

plotted in Figure 7(C). It shows the effect of injection rate on IFT. The interfacial widths at 2 m/s and 8

372

m/s are equal to L1 = 50 Å and L2 = 25 Å, respectively. L1 is wider than L2, indicating the lower IFT for

373

L1 than that for L2. Therefore, IFT increases with the increasing injection rate, unfavorable for the

374

translocation of dodecane molecules within the nanopore. We also calculate the RDF between C atoms

375

in different dodecane molecules to examine the effect of injection rate on oil volume swelling, plotted

376

in Figure 7(D). It shows that a small injection rate is of benefit to the expansion of oil volume. The

377

peak of RDF and the area below the curve decrease with the injection rate, indicating that the

378

coordinate number around a dodecane molecule also decreases with the injection rate. Consequently,

379

dodecane molecules distribute loosely and the average separation between them increases at smaller

380

CO2 injection rate. This can be observed by comparing the configurations at 10 and 2 m/s inserted in

381

Figure 7(D). Because of constant injection volume at different rates, more sufficient interaction

382

between CO2 and dodecane at the small injection rate improves the mobilization of upstream dodecane

383

molecules into CO2, which enhances the volume swelling of dodecane. The volume swelling of

384

upstream dodecane provides additional energy to expel the downstream dodecane towards the right of

385

the nanopore.

386

4. CONCLUSIONS

387

In this study, NEMD simulations were carried out to study the microscopic mechanisms of

388

scCO2 displacing oil confined in the shale nanopore at different injection rates. The injected scCO2

389

preferentially adsorbs on the hydroxylated surface and form layering structures due to hydrogen

390

bonds between CO2 and -OH groups. However, many adsorbed CO2 molecules still retain their

391

mobility, which may be caused by slippage, Knudsen diffusion and imbibition of scCO2. Adsorbed

392

dodecane molecules are partly detached from the surface resulting from the competitive

393

adsorption between scCO2 and dodecane, which leads to more moveable oil and thus increasing 17

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 20

394

oil production. The miscible zone is formed by the dodecane molecules separated from the surface,

395

bulk dodecane molecules dissolved in CO2 phase, and CO2 molecules penetrating into dodecane

396

phase. Thus, mobilization of dodecane molecules into CO2 has an important effect on CO2 EOR.

397

The formation of the miscible zone results in IFT reduction and oil volume expansion, which

398

enhance the mobility of oil and have a dominant role in scCO2 driving process. Adjacent to the

399

displacement front, the downstream dodecane molecules aggregate and pack tightly, migrating in

400

the form of oil column favorable to CO2 EOR. With the increase of CO2 injection rate, the contact

401

angle of dodecane phase decreases, the Meniscus convex reduces, and the interfacial width

402

increases. The results reveal that the small injection rate with a large volume is favorable to CO2

403

EOR. The morphology of meniscus may alter in the order convex-concave-CO2 entrainment when

404

the injection rate gradually increases. The findings are benifitial to understand and design effective

405

CO2 exploitation of shale resources as well as may provide physical insight into CO2 storage,

406

nanoscale fluid flow, biological molecules translocation, and so on.

407

ACKNOWLEDGMENT

408

This

409

(2014D-5006-0211), the National Natural Science Foundation of China (U1262202), the

410

Fundamental Research Funds for the Central Universities (14CX05022A, 15CX08003A), the

411

National Basic Research Program of China (2014CB239204, 2015CB250904), and the Shandong

412

Provincial Natural Science Foundation, China (ZR2014EEM035). The Research Council of

413

Norway, Det Norske Oljeselskap ASA, and Wintershall Norge AS are acknowledged for funding

414

support via WINPA project (Nano2021 and Petromaks2234626/O70).

415

REFERENCES

416 417 418 419 420 421 422 423 424 425 426

work

was

financially

supported

by

the

Petro

China

Innovation

Foundation

(1) Aycaguer, A.-C.; Lev-On, M.; Winer, A. M. Energy Fuels 2001, 15, 303-308. (2) Nobakht, M.; Moghadam, S.; Gu, Y. Energy Fuels 2007, 21, 3469-3476. (3) Nobakht, M.; Moghadam, S.; Gu, Y. Fluid Phase Equilib. 2008, 265, 94-103. (4) Zhang, C.; Oostrom, M.; Grate, J. W.; Wietsma, T. W.; Warner, M. G. Environ. Sci. Technol. 2011, 45, 7581-7588. (5) Gamadi, T. D.; Elldakli, F.; Sheng, J. SPE/AAPG/SEG Unconventional Resources Technology Conference, 2014, 1922690. (6) Ishida, T.; Aoyagi, K.; Niwa, T.; Chen, Y.; Murata, S.; Chen, Q.; Nakayama, Y. Geophys. Res. Lett. 2012, 39. (7) Tudor, R.; Poleschuk, A. J. Can. Pet. Technol. 1996, 35. (8) Wang, S.; Chen, S.; Li, Z. Energy Fuels 2015, 30, 54-62. 18

ACS Paragon Plus Environment

Page 19 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470

(9) Cao, M.; Gu, Y. Fluid Phase Equilib. 2013, 356, 78-89. (10) Hawthorne, S. B.; Gorecki, C. D.; Sorensen, J. A.; Steadman, E. N.; Harju, J. A.; Melzer, S. SPE Unconventional Resources Conference Canada, 2013, 167200. (11) Alharthy, N.; Teklu, T.; Kazemi, H.; Graves, R.; Hawthorne, S.; Braunberger, J.; Kurtoglu, B. SPE Annual Technical Conference and Exhibition, 2015, 175034. (12) Tovar, F. D.; Eide, O.; Graue, A.; Schechter, D. S. SPE Unconventional Resources Conference, 2014, 169022. (13) Wang, R.; Lv, C.; Zhao, S.; Lun, Z.; Wang, H.; Cui, M. SPE Asia Pacific Enhanced Oil Recovery Conference, 2015, 174590. (14) Wang, S.; Feng, Q.; Javadpour, F.; Yang, Y.-B. J. Phys. Chem. C 2016, 120, 14260-14269. (15) Saidian, M.; Prasad, M. Fuel 2015, 161, 197-206. (16) Kuila, U.; Prasad, M. Geophys. Prospect. 2013, 61, 341-362. (17) Hantal, G. r.; Brochard, L.; Dias Soeiro Cordeiro, M. N. l.; Ulm, F. J.; Pellenq, R. J.-M. J. Phys. Chem. C 2014, 118, 2429-2438. (18) Sofos, F.; Karakasidis, T.; Liakopoulos, A. International Journal of Heat and Mass Transfer 2009, 52, 735-743. (19) Giannakopoulos, A.; Sofos, F.; Karakasidis, T.; Liakopoulos, A. International Journal of Heat and Mass Transfer 2012, 55, 5087-5092. (20) Liakopoulos, A.; Sofos, F.; Karakasidis, T. E. Microfluidics and Nanofluidics 2016, 20, 1-7. (21) Hartkamp, R.; Ghosh, A.; Weinhart, T.; Luding, S. J. Chem. Phys. 2012, 137, 044711. (22) Argyris, D.; Cole, D. R.; Striolo, A. Langmuir 2009, 25, 8025-8035. (23) Sofos, F. D.; Karakasidis, T. E.; Liakopoulos, A. Phys. Rev. E 2009, 79, 026305. (24) Malani, A.; Ayappa, K.; Murad, S. J. Phys. Chem. B 2009, 113, 13825-13839. (25) Le, T.; Striolo, A.; Cole, D. R. J. Phys. Chem. C 2015, 119, 15274-15284. (26) Le, T.; Striolo, A.; Cole, D. R. Chemical Engineering Science 2015, 121, 292-299. (27) Wu, H.; Chen, J.; Liu, H. J. Phys. Chem. C 2015, 119, 13652-13657. (28) Yuan, Q.; Zhu, X.; Lin, K.; Zhao, Y.-P. Phys. Chem. Chem. Phys. 2015, 17, 31887-31893. (29) Jin, Z.; Firoozabadi, A. J. Chem. Phys. 2015, 143, 104315. (30) Wang, S.; Javadpour, F.; Feng, Q. Fuel 2016, 171, 74-86. (31) Coasne, B.; Galarneau, A.; Di Renzo, F.; Pellenq, R. Langmuir 2010, 26, 10872-10881. (32) Coasne, B.; Fourkas, J. T. J. Phys. Chem. C 2011, 115, 15471-15479. (33) Collell, J.; Galliero, G.; Vermorel, R.; Ungerer, P.; Yiannourakou, M.; Montel, F.; Pujol, M. J. Phys. Chem. C 2015, 119, 22587-22595. (34) Plimpton, S. J. Comput. Phys. 1995, 117, 1-19. (35) Hoover, W. G. Phys. Rev. A 1985, 31, 1695. (36) Nosé, S. J. Chem. Phys. 1984, 81, 511-519. (37) Qamhieh, K.; Wong, K.-Y.; Lynch, G. C.; Pettitt, B. M. Int. J. Numer. Anal. Mod. 2009, 6, 474. (38) Harris, J. G.; Yung, K. H. J. Phys. Chem. 1995, 99, 12021-12024. (39) Senapati, S.; Berkowitz, M. L. J. Phys. Chem. B 2003, 107, 12906-12916. (40) Goldsmith, J.; Martens, C. C. J. Phys. Chem. Lett. 2009, 1, 528-535. (41) de Lara, L. S.; Michelon, M. F.; Miranda, C. R. J. Phys. Chem. B 2012, 116, 14667-14676. (42) Toukmaji, A. Y.; Board, J. A. Comput. Phys. Commun. 1996, 95, 73-92. (43) Nymand, T. M.; Linse, P. J. Chem. Phys. 2000, 112, 6152-6160. (44) Giannakopoulos, A. E.; Sofos, F.; Karakasidis, T. E.; Liakopoulos, A. Microfluid. Nanofluid. 19

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

471 472 473 474 475 476 477 478 479 480 481 482 483 484 485

2014, 17, 1011-1023. (45) Holland, D. M.; Lockerby, D. A.; Borg, M. K.; Nicholls, W. D.; Reese, J. M. Microfluid. Nanofluid. 2015, 18, 461-474. (46) Luzar, A.; Chandler, D. J. Chem. Phys. 1993, 98, 8160-8173. (47) Dickson, J. L.; Gupta, G.; Horozov, T. S.; Binks, B. P.; Johnston, K. P. Langmuir 2006, 22, 2161-2170. (48) Duan, C.; Zhou, F.; Jiang, K.; Yu, T. Int. J. Heat Mass Transfer 2015, 91, 1088-1100. (49) Fang, T.; Shi, J.; Sun, X.; Shen, Y.; Yan, Y.; Zhang, J.; Liu, B. J. Supercrit. Fluids 2016, 113, 10-15. (50) Green, D.; Willhite, G. Soc. Pet. Eng. 1998. (51) Liu, B.; Shi, J.; Wang, M.; Zhang, J.; Sun, B.; Shen, Y.; Sun, X. J. Supercrit. Fluids 2016, 111, 171-178. (52) Li, X.; Fan, X. Int. J. Greenhouse Gas Control 2015, 36, 106-113. (53) De Coninck, J.; Blake, T. Annu. Rev. Mater. Res. 2008, 38, 1-22. (54) Gerek, Z. N.; Elliott, J. R. Ind. Eng. Chem. Res. 2010, 49, 3411-3423.

486

20

ACS Paragon Plus Environment

Page 20 of 20