Molecular Dynamics Simulation of Surfactant Flooding Driven Oil

Dec 27, 2018 - A flooding from rear (FFR) phenomenon is revealed that the surfactant ... the oil molecule detachment occurs at the rear bottom of the ...
0 downloads 0 Views 8MB Size
Article Cite This: J. Phys. Chem. B XXXX, XXX, XXX−XXX

pubs.acs.org/JPCB

Molecular Dynamics Simulation of Surfactant Flooding Driven OilDetachment in Nano-Silica Channels Xianqiong Tang,*,†,‡ Shaofei Xiao,† Qun Lei,§ Lingfang Yuan,† Baoliang Peng,*,§,∥ Lipeng He,§,∥ Jianhui Luo,§,∥ and Yong Pei*,† †

J. Phys. Chem. B Downloaded from pubs.acs.org by UNIV OF KANSAS on 01/01/19. For personal use only.

Key Laboratory for Green Organic Synthesis and Application of Hunan Province, Key Laboratory of Environmentally Friendly Chemistry and Applications of Ministry of Education, Department of Chemistry, Xiangtan University, Xiangtan 411105, P. R. China ‡ Department of Civil Engineering and Mechanics, Xiangtan University, Xiangtan 411105, P. R. China § Research Institute of Petroleum Exploration & Development (RIPED), PetroChina, Beijing 100083, P. R. China ∥ Key Laboratory of Nano Chemistry (KLNC), CNPC, Beijing 100083, P. R. China S Supporting Information *

ABSTRACT: Recovery of crude oil in rock nanopores plays an important role in the petroleum industry. In this work, we carried out molecular dynamics (MD) simulations to study the process of ionic surfactant solution driven oil-detachment in model silica (SiO2) nanochannels. Our MD simulation results revealed that the oil-detachment induced by the ionic surfactant flooding can be described by a three-stage process including the formation and delivery of surfactant micelles, the surfactant micelle disintegration-spread and migration on the oil− aggregate surface, and oil molecular aggregate deformation-todetachment. A flooding from rear (FFR) phenomenon is revealed that the surfactant molecules tend to migrate to the rear bottom of the oil molecular aggregate caused by the water flow effect and hydration of polar head groups of surfactants, which facilitate the penetration of water molecules into the oil−rock interface, and the oil molecule detachment occurs at the rear bottom of the oil molecular aggregate. The present MD simulation results also indicate that the dodecyl benzenesulfonate (SDBS) has higher oil-driven efficiency than that of dodecyl trimethylammonium bromide (DTAB). The difference of oil displacement efficiency between the two surfactants is attributed to the hydration property of the polar head groups. Compared with the −N(CH3)3+ headgroup in DTAB, the bare O atom in the −SO3− group has a stronger H bond interaction with the surrounding water molecules. The stronger interaction between the headgroup of SDBS and the adjacent water molecule results in the surfactant migrating to the rear bottom of the oil molecules more quickly, thus accelerating the detachment of oil molecules.

1. INTRODUCTION The deformation and detachment of liquid droplets on solid surfaces and microchannels is of great importance in many engineering fields, including enhanced oil recovery (EOR), surface decontamination, membrane emulsification, etc.1−5 In petroleum industries, after water flooding, the residual oil is trapped in the nanopores of rock reservoirs. The exploitation of trapped residual oil has imposed a grand challenge in the oil recovery process. In recent years, many EOR techniques including the chemical injection and gas injection have been developed.6,7 In these techniques, surfactant flooding is one of most efficient EOR techniques.8−11 From a macroscopic perspective, the ease of oil droplet deformation and movement determines the EOR through the rock structures.4 Surfactant molecules adsorbed on the oil−water interface can significantly reduce the interfacial tension (IFT) between oil and water, resulting in the change of wettability of the liquid/solid interface, which are considered to be the two main factors in the process of enhancing oil recovery. © XXXX American Chemical Society

In order to understand the microscopic mechanism of surfactant flooding, it is urgent to understand the aggregation, displacement, and transportation of oil droplets in nanopores with surfactant solution interactions at the atomic and molecular levels. From the experimental perspective, it is still a grand challenge to directly observe the microscopic details of the detachment of oil molecules in the nanochannel.12 Owing to the rapid development of computer science and technology, computational simulation methods including molecular dynamics (MD),13−17 Monte Carlo (MC),18 and dissipative particle dynamics (DPD)19 have been extensively applied to explore the microscopic details of the conformational change and detachment mechanism of oil molecules.20−25 For instance, Yang et al.26 suggest that oil can imbibe into the organic nanopores of rigid rock pores faster than what classic Received: October 7, 2018 Revised: December 4, 2018

A

DOI: 10.1021/acs.jpcb.8b09777 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

Figure 1. Schematic illustration of the nanoslit model. The dashed line denotes the periodic boundary.

Figure 2. (a) The four components of model crude oil molecular aggregate. C atoms in gray ball, H atoms in white ball. (b) Side and top view of the equilibrium oil molecular adsorption configuration on the silica surface. The heavy components closest to the silica surface are displayed by a ball and stick model. (c) Density profile of the oil molecules along the Y direction (vertical to the silica substrate). Colors for the atom scheme are O - red, Si - yellow, and C - gray, respectively. The H atoms of oil molecules are hidden for clarity.

(NEMD) simulation showed that the continuous fluid dynamics (Poiseuille equation) can reasonably describe the transport behavior of liquid hydrocarbons in large pore size silica nanopores. Moncayo-Riascos et al.28 proposed a method to evaluate and represent the wettability alternation phenomenon caused by the action of organosilane surfactants using MD simulations. The results showed that the presence of surfactant molecules reduced the affinity of the n-heptane with the surface. Zhang et al.29 studied the detachment mechanism of alkane molecules from a hybrid hydrophobic and hydro-

imbibition model prediction and organic nanopores constitute a large portion of pore volume in target tight rocks. Wang et al.27 studied the static properties and pressure-driven flow behavior of liquid n-octane confined in quartz nanopores of shale by MD simulations. The results show that the n-octane molecules are disordered and tend to be equivalent to the bulk fluid in the slit center with an aperture larger than 3.6 nm. However, within the interfacial regions perturbed by the substrate, n-octane molecules diffuse more slowly. The velocity distribution obtained by nonequilibrium molecular dynamics B

DOI: 10.1021/acs.jpcb.8b09777 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

Figure 3. Snapshots of molecular configurations of two models after pre-equilibrium (corresponding to the t = 0 ns configuration). (a) DTAB surfactant system. (b) SDBS surfactant system. Water atoms are represented as red points. From left to right: the push board (made by pseudo atoms in light blue color), DTAB (N atoms in blue and C atoms in gray), oil, silica nanoslit (Si atoms in yellow, O atoms in red, H atoms in white), and pressure holding board (pseudo atoms in light blue). All H atoms (except the silica surface) are hidden for clarity.

cationic dodecyl trimethylammonium bromide (DTAB) surfactant flooding. The molecular structure effects of two ionic surfactants on the performance of oil detachment were discussed.

philic solid surface. The simulation results show that the formation of a water channel between the water phase and the solid surface is the key to the oil molecular detachment. More recently, DPD simulations have been applied to study the nanoscale structure of fluids in nanochannels. Millan and Laradji used the DPD method to understand the structural and transport properties of polymer solutions in nanoscale channels.30 Chen et al. carried out DPD simulations to study the water−oil displacement process in capillaries under an external force.19 They demonstrated that both strengthening the interaction between water and capillaries and weakening the applied force can enhance the oil drop displacement from capillaries. In the study conducted by Sedghi et al., they studied the effect of wall−liquid interaction on the critical capillary pressure of oil/water displacement in a square crosssection nanopore.31 In this work, we report atomic MD simulation of surfactant flooding driven oil molecule detachment in a nanosilica channel. It is well-known that the polyaromatic fractions in crude oil are easy to precipitate and deposit in the process of production and transportation, resulting in reservoir rock and transportation pipeline blockage.32−34 Moreover, the association of polyaromatic molecules such as asphaltenes and resins strongly affected solubility, viscosity, density, and other physical properties of crude oil.33,34 Therefore, it is of great significance to understand the effect of surfactant molecules on the displacement and transportation of crude oil (including asphaltene and resin) in nanochannels, which is of significance for promoting the development and utilization of shale oil. By using NEMD simulation, in this work, we studied the detachment processes of oil molecules in nanosilica channels upon surfactant flooding. The interactions between water/ surfactant molecules and oil molecules during the oil detachment process were analyzed. Our MD simulation results revealed a flooding from rear (FFR) phenomenon that the detachment of oil molecules started from the rear bottom of the oil molecular aggregate. This phenomenon is observed in both anionic sodium dodecyl benzenesulfonate (SDBS) and

2. SIMULATION METHOD AND DETAILS 2.1. Nanoscale Slit Model. The α-quartz (α-SiO2) was used to mimic the rock reservoir. The unit cell of α-SiO2 is derived from the database of the Materials Studio software, and it was expanded into an orthorhombic supercell along three dimensions with x × y × z of 24.6 × 102.1 × 602.5 Å3. Considering the size of nanopores in rock reservoirs,35 a nanoslit with a height of 87.5 Å along the y-axis was constructed by removing the middle part of the silica supercell (see Figure 1 for details). The exposed (001) surface was hydroxylated by −OH groups with a density of 9.6 nm−2, a value which agrees with previous literature.36 The distance between the “push board” and “pressure holding board” is 49.9 nm. 2.2. Crude Oil Model. The model crude oil aggregate is composed of four components including the saturate, aromatic, resin, and asphaltene. The molecular configurations of the four components are displayed in Figure 2a. We build the model systems using the similar procedure as in refs 37 and 38. First, 40 saturated, 30 aromatic, 20 resin, and 10 asphaltene molecules were randomly distributed near to the silica surface. Second, an energy minimization and a subsequent 1.0 ns canonical ensemble MD simulation were performed to obtain the equilibrium structure of the oil molecular aggregate. In order to conveniently observe the surfactant flooding process, the oil molecular aggregate was placed on the left side of the nanoslit. In the oil aggregate, due to the strong π−π interactions, asphaltene molecules formed “face to face” aggregates, which were in agreement with our previous studies.39 Meanwhile, because of the strong interactions between the polar atoms and the hydroxylated surface, it was found that the heavymolecular-weight components tended to preferentially adsorb C

DOI: 10.1021/acs.jpcb.8b09777 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B on the silica surface, as presented in Figure 2b,c. The first maximum peak of density distribution (1.34 g/cm 3 ) corresponded to the closely adsorbed heavy-molecular-weight components on the nearest silica surface. The thickness of the oil molecular aggregate is around 35 Å. Finally, we built the initial model for NEMD modeling by adding water and surfactant molecules (60 SDBS or DTAB surfactant molecules were randomly dissolved in the left part of the aqueous solution containing 31 100 water molecules) and carried out a 1 ns NVT simulation for pre-equilibrium of systems. The molecular configurations after pre-equilibrium were shown in Figure 3. As shown in Figure 3, the nanoslit does not extend to the right end of the simulation box. If the oil drop was placed on the middle or right side of the nanoslit, due to the driving force of the water phase, the oil molecules may slide along the Z direction and detach from the nanoslit, which was not conducive to observe the mutual interaction between the surfactant and the oil drop. 2.3. MD Simulation Parameters. All MD simulations were performed using LAMMPS,40 and all oil molecules were parametrized using the polymer consistent force field (PCFF).41 The PCFF had been validated to be capable of accurately predicting structural and thermodynamic properties of oil components.42−49 The charge set of the silica surface and the modified hydroxyl was taken from Yan’s work.35 The detailed force field parameters were given in the Supporting Information. The short-range van der Waals (vdW) interactions were calculated with a cutoff (12.0 Å); the longrange Coulombic interactions were treated by the particle− particle particle−mesh (PPPM) algorithm51 with a convergence parameter of 10−4. The time step and time interval of collecting data were set to 1 fs and 10 ps, respectively. In order to mimic the oil reservoir condition, a temperature of 330 K was set in our simulation. The velocity of the leftmost “push board”, defined as a rigid body, was set to 1.0 m/s (1.0 × 10−5 Å/fs) along the z direction, and the external force was applied on the rightmost “pressure holding board” to keep the pressure of the water phase being maintained at 20.0 MPa, as presented in Figure 1 and Figure 3. During MD simulations, the atoms in the silica nanoslit were kept frozen except for the modified hydroxyl. The periodic boundary condition was applied in all directions.

Figure 4. Comparison of the RDF of the COM of surfactants before and after the pre-equilibrium stage.

spread of surfactant micelles onto the oil molecular surface, and (III) deformation and detachment of oil molecules. 3.1.1. Micellization of Surfactant Molecules (Stage I). During the pre-equilibrium stage, micellization of surfactant molecules was clearly seen that the randomly distributed SDBS and DTAB surfactant molecules aggregated rapidly (Figure 3). It is well-known that the surfactants tend to aggregate into micelles in aqueous solution; therefore, during the set of the initial configurations, we have placed the 60 surfactants in a relatively narrow region to avoid the long diffusion process of surfactant molecules, as shown in Figure S3 in the Supporting Information. To probe the spatial distribution of the surfactant molecules, the COM radial distribution functions (RDFs) of SDBS and DTAB molecules before and after pre-equilibrium were analyzed and compared in Figure 4. The dominant peaks were seen in the RDF curve of equilibrated configuration, while the peaks were rather flat before the pre-equilibrium, which indicated remarkable changes of spatial distribution of surfactant molecules. For the SDBS and DTAB, their COMRDF curves both showed a maximum at 5.5 Å, which indicated that there were no substantial differences of the spatial structures of two kinds of surfactant micelles. 3.1.2. Disintegration and Spread of Surfactant Micelles (Stage II). As the push board moves, the surfactants gradually approached the oil molecules. Before contacting with oil molecules (0−5 ns), surfactant micelles retained a roughly spherical shape and moved along the water flow direction with a constant speed. At ∼5.5 ns, both DTAB and SDBS micelles began to contact with oil molecules. A micelle disintegrationspread process was distinctly found that, when the surfactant micelles approached the oil aggregate, the surfactant head groups (−N(CH3)3+ and −SO3−) close to the oil molecules were discharged from the water−oil interface. This phenomenon can be understood from the hydrophobic interactions. As the oil molecules were majorly composed of hydrophobic hydrocarbon compounds, they tended to interact more strongly with the hydrocarbon tails of surfactant molecules and hence led to disintegration of surfactant micelles. After surfactant micelle disintegration, surfactant molecules rapidly spread on the oil molecular surface, formed a typical amphiphilic adsorption configuration that the hydrophobic tails inserted into the oil molecules, and the hydrophilic head groups exposed to water. We called these processes the surfactant micelle disintegration-spread stage.

3. RESULTS AND DISCUSSION 3.1. Dynamical Process of Surfactant Flooding. A preliminary simulation is carried out to examine the effect of the water flow effect without addition of surfactant molecules. The simulation model is identical to the model illustrated in Figure 3 except that the surfactant molecules are not added. The snapshots of the trajectory and the center of mass (COM) of the oil aggregate in the Z direction are displayed in Figures S1 and S2, which indicated that, without the presence of surfactant molecules, the oil aggregate is less affected by the pure water flow and it cannot detach from the nanosilica channel. Using the nanoscale slit model displayed in Figure 3, we illustrated the atomic and molecular details of the dynamic process of surfactant flooding driven oil molecule detachment in the nanosilica channel. The whole dynamic process of oildetachment assisted by the surfactant solution flow can be described by a three-stage mechanism, including (I) micellization of surfactant molecules, (II) disintegration and D

DOI: 10.1021/acs.jpcb.8b09777 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

Figure 5. (a) The change of Esurfactant−oil and Esilica−oil along with MD simulations. (b) The snapshots of the surfactant/water/oil at different time stages.

increase of interaction energy between surfactant and oil molecules between 5.5 and 7.0 ns, which corresponded to the surfactant micelle “disintegration-spread” process. After this time stage, the significant increase of Esurfactant−oil and gradual decrease of Esilica−oil are seen. The following stage corresponded to the deformation and detachment of oil molecules from the silica caused by the surfactant interactions. 3.1.3. Deformation and Detachment of Oil Molecules (Stage III). After stage II, caused by the water flow effect, surfactant molecules migrated to the rear side of the oil molecular surface, which led to deformation and detachment of oil molecules, as shown in Figure 5b and Figures S4 and S5 in the Supporting Information. The detachment of oil aggregate (oil layers and droplets) from solid surfaces under either static surfactant solution interactions or shear flow conditions had been studied by both atomic MD simulations and the fluid dynamic model. Using atomic MD simulations, Liu et al.14 and Tang et al.50 revealed that the formation and expansion of water molecule channels was a key to detach the oil layer from the rock surface. Upon

In order to reveal the essence of the above-mentioned phenomena, the interaction energies between the oil, silica surface, and surfactant molecules were calculated and diagrammed in Figure 5a. Here, the interaction energy (Esurfactant−oil) between surfactant molecules and oil was defined as Esurfactant−oil = Etotal − (Esurfactant + Eoil), where Etotal was the total energy of surfactant molecules and oil and Esurfactant and Eoil were the energies of individual surfactant molecules and the oil aggregate, respectively. Taking the surfactant molecule energy (Esurfactant) calculation as the example, we deleted all other atoms in the trajectory frame that do not belong to the surfactant molecules. Then, the single-point energy is calculated using the force field parameters. By taking these steps, we obtained Esurfactant at different time stages. The more negative value of Esurfactant−oil represents stronger interaction between two components. The interaction energy between oil molecules and silica substrate was also calculated, defined as Esilica−oil = Etotal − (Esilica + Eoil), where Esilica−oil and Esilica were the total energy of silica and oil and the silica surface, respectively. As shown in Figure 5b, there was a significant E

DOI: 10.1021/acs.jpcb.8b09777 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

Figure 6. (a) Schematic diagram of the front contact angle θ1 and rear contact angle θ2. (b) Evolution of θ1 and θ2 over the simulation time of the DTAB−water−oil and SDBS−water−oil systems. (c) Illustration of the flooding from rear (FFR) mechanism.

In comparison to the static surfactant solution interactions, the liquid droplet detachment from the solid substrate in shear flow showed a different mechanism. The fluid dynamics models indicated that the dynamic behavior of a macroscopic oil droplet subjected to shear flow can be described by a twostage process where the shear flow of fluid first led to deformation of the droplet first and then the droplet completely detached from the solid surface. Jones et al.52 concluded that, in the equilibrium contact angle range (e.g., θe > 120°), the drop preserved its integrity as it escaped from the boundary, whereas at lower contact angles the drop assumed a distinctly elongated shape prior to its removal, which developed “necks” at subsequent times. Zhang et al.53 studied the effect of surfactants on the deformation of oil droplet in shear flow. The simulation results indicated that, in shear flow, surfactants will be convected by the interfacial velocity to the tip regions of the droplet, leading to a nonuniform surfactant distribution, which further caused the gradient of the interface tension and Marangoni effect. Our MD simulation results showed two effects observed in previous atomic MD and fluid dynamics simulations. In order to track the adsorption configuration change of the oil molecular aggregate, the contact angle was measured at different simulation stages and displayed in Figure 6. We

Figure 7. Change of the COM of the oil aggregate along the waterflow direction under the DTAB and SDBS surfactant flooding.

surfactant interactions, water molecules can penetrate into the oil layer and propagate onto the rock surface, forming a gel layer on the rock surface. Caused by the replacement of surface occupied sites by water molecules, oil molecules were peeled off from the rock surface. F

DOI: 10.1021/acs.jpcb.8b09777 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

Figure 8. Schematic illustration of the oil detachment process under surfactant solution flooding.

defined two contact angles,54 denoted as the front contact angle θ1 and the rear contact angle θ2 (Figure 6a). As shown in Figure 6b and c, the equilibrium contact angles of oil drop after the 1 ns pre-equilibrium process were around 90 and 91°, respectively. With the progress of simulations, an obvious decrease of θ1 and an increase of θ2 were seen, indicating significant structural deformation of the oil aggregate. A larger θ2 meant that the oil aggregate tilted more to the direction of the shear flow, which increased its deformation. According to Figures S4 and S5, in the early stage of the simulation (0−5 ns), the surfactant molecules are not in contact with the oil drop, so the main driving factor for the change of the contact angle is the water flow effect. At the subsequent stage (5−10 ns), the surfactant molecules contact the front side of the oil molecules; hence, a quick deformation of the oil molecular aggregate at the front side can be observed. As a result, θ2 tends to increase, while there is no significant change in θ1. After the simulation time exceeds 10 ns, the surfactant molecules move to the rear bottom of the oil molecular aggregate, the deformation of the whole oil molecules is accelerated, and both θ2 and θ1 change quickly and simultaneously. An interesting “flooding from rear (FFR)” phenomenon was observed from the current MD simulations that the detachment of oil molecules started from the “rear side” of the oil aggregate. By observing the MD trajectories, it was interestingly found that due to the water flow effect the surfactant molecules were pushed to the rear side of the oil aggregate. After then, significant deformation of the oil aggregate occurred that the front contact angle θ1 decreased while the rear contact angle θ2 increased. With the deformation of the oil aggregate, water molecules gradually penetrated into the oil−water interface at the rear bottom of the oil aggregate. At the same time, some surfactant molecules adsorbed on the SiO2 surface, showing an obvious wetting alternation phenomenon, as shown in Figures S4 and S5 in the Supporting Information. These two effects eventually led to complete detachment of oil molecules from the silica nanoslit. From Figure 2, the aromatic and asphaltene molecules have strong interactions with the silica substrate. They formed a tight adsorption layer nearest to the silica surface. The break of this adsorption layer is therefore a key to completely detach the oil aggregates. According to the present simulation results, the water permeation at the rear bottom oil−water interface with the help of the surfactant interaction (FFR mechanism) played an important role in the detachment of strongly

adsorbed oil molecules, as shown in Figure S6 in the Supporting Information. Finally, in order to illustrate the oil deformation-detachment process more clearly, Figure 7 shows the COM change of the oil aggregate along the water flow direction (z-direction). For both systems, the slope of the COM curve was small in the initial stage (0−7.5 ns). In this time stage, surfactant micelles were far away from oil molecules; hence, the main driving force of oil movement came from the impact of water flow. When the simulation time reached and exceeded 7.5 ns, the slope of the curve increased significantly. The increase of the movement speed of the oil aggregate was attributed to the water−surfactant−oil interactions. In this stage, a water− surfactant−oil interface was formed. Because of hydrophilic attractions between polar headgroups of surfactants and water molecules, the water flow sped up the movement of surfactant−oil complexes. From the COM curves and the MD trajectories (Figures S4 and S5), the DTAB surfactant flooding driving oil detachment speed was faster than the SDBS. The complete detachment of the oil aggregate was observed at around t = 27.0 ns and t = 32.0 ns in DTAB and SDBS flooding systems, respectively. However, due to the randomness of micelle size and delivery, the adsorption capacity of surfactant molecules on the oil molecular surface in the two systems was different, so it is difficult to compare the oil driving efficiency of two surfactants. Even though a common oil-driven mechanism was concluded from these simulations, the surfactant flooding induced oildetachment can be described by a three-stage process including the surfactant micelle formation and delivery, the disintegration and spread of surfactant micelles, and then oilmolecular aggregate deformation to detachment, as illustrated in Figure 8. 3.2. Comparison of Oil-Driving Efficiency of SDBS and DTAB. In this part of the discussions, we attempted to compare the oil-driven efficiency of two kinds of ionic surfactants under water flow conditions. To do this, we further designed two sets of models as displayed in Figure 9, denoted as DTAB-II and SDBS-II (model set 1, containing 30 surfactant molecules) and DTAB-III and SDBS-III (model set 2, containing 45 surfactant molecules). The simulation box along three dimensions with x × y × z is 24.6 × 102.1 × 524.1 Å3, and the number of water molecules is 27 000. In addition, the thickness of the silica layer was set to be 19.3 Å, which can isolate the nonbonding interactions along the y-direction due to periodic boundary conditions. In two sets of models, they have the same number of surfactant molecules and the same G

DOI: 10.1021/acs.jpcb.8b09777 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

Figure 9. (a) Schematic illustration of the nanosilica channel model before pre-equilibrium. (b and c) Side views of the configuration of the atomic model. Top and bottom layer: silica nanochannel model (Si, yellow ball; O, red ball; H, white ball); blue CPK ball, N atom of DTAB; yellow CPK ball, S atom of SDBS; gray CPK ball in left, push layer; gray ball in right, pressure layer; red point, O atom of water; the H atoms of water, surfactant, and heavy oil molecules were hidden for clarity. In the models displayed in parts b and c, the number of surfactant molecules was 30 and 45, respectively. The dashed lines denoted the periodic boundary along three directions.

initial spatial and contact configurations on the oil molecular surface. The surfactant micelle formation and delivery processes were omitted in these new models. The initial molecular sets in these new models avoided the randomness of micelle size and contact area between surfactant molecules and oil molecules, which enabled us to compare the oil-driven efficiency of two kinds of surfactant molecules more straightly. On the other hand, the nanoslit model is replaced by a nanochannel model, as shown in Figure 9. In this new model, the push board is in the nanochannel, which does not interact with the silica channel, but only employs forces on the water molecules to drive the water flow in the nanochannel. A 1 ns NVT simulation is carried out for the pre-equilibrium of

systems. The other simulation parameters were the same as those in the previous section. Figures S7−S10 displayed the molecular trajectory snapshots. In these new models, the FFR phenomenon was seen as well. At the initial state (Figure 9b and c), the surfactants fully dissolved in the water phase. After the start of the MD simulations, the surfactants inserted into the oil molecular aggregate and migrated to the rear side of the oil molecular aggregate rapidly caused by the water flow effect. After that, the deformation-to-detachment of the oil aggregate occurred. Our simulation results indicated that the SDBS has better oil-driven performance than that of DTAB under water flow conditions. Figure 10 displayed the change of the COM of the H

DOI: 10.1021/acs.jpcb.8b09777 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

Figure 10. COM displacement of the oil aggregate in the Y and Z direction.

Figure 11. (a) Definition of the ΔZ between the leftmost oil phase atom and the surfactant head atom and (b) the evolution of ΔZ along with the simulation times.

oil aggregate along and perpendicular to the water flow direction, denoted as Z-COM and Y-COM, respectively. The detachment speed of oil molecules can be reflected from the slope of the Y-COM curve. The large slope of the Y-COM curve suggested faster detachment of the oil molecular aggregate. From Figure 10, the oil aggregate moved much faster along both the Y and Z directions under the SDBS interactions. Moreover, the increase of adsorbed SDBS molecules resulted in an obvious increase of oil detachment speed. From Figures S9 and S10 in the Supporting Information, the oil molecules in the SDBS-II system (30 surfactants) were completely detached at about 18 ns, while the oil molecules in the SDBS-III system (45 surfactants) were completely detached at about 16 ns. However, the DTAB showed a reverse tendency. By checking the MD trajectories, it was found that the complete detachment of oil molecules

happened at around t = 44 ns in the DTAB-II system. While in the DTAB-III system, the complete detachment of the oil aggregate was seen at t = 50 ns. Furthermore, it was found that, with the increase of the number of surfactant molecules, the movement of DTAB molecules became slow. Moreover, the oil molecules were developed into a “strip” configuration under the DTAB interactions. To compare the movement speed of surfactant molecules on the oil/water interface, we defined a parameter ΔZ, which was the difference between the minimum z coordinate of the surfactant headgroup (S or N atom) and the minimum z coordinate of oil molecules (C atom), as shown in Figure 11. For the SDBS-II and SDBS-III systems, the ΔZ increased more rapidly than the DTAB-II and DTAB-III systems. This meant that SDBS molecules migrated faster on the oil aggregate surface than DTAB molecules. According to the FFR I

DOI: 10.1021/acs.jpcb.8b09777 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

and migration of surfactant molecules on the oil aggregate, and oil aggregate deformation-to-detachment. In the oil deformation-to-detachment stage, a flooding from rear phenomenon was found. It was found that the detachment of oil molecules started from the rear bottom of the oil aggregate. Under the action of water flow, surfactant molecules migrated to the rear bottom of the oil aggregate, which induced the water molecule to occupy the surface sites and detachment of oil molecules. For two ionic surfactant molecules investigated in this work, the SDBS showed better oil-driven efficiency than DTAB. The difference of the oil-detachment performance of two kinds of surfactants was attributed to the headgroup effect. The SO3− group in SDBS had stronger attractions to water molecules, which induced a faster migration speed of SDBS on the oilaggregate surface and hence a higher deformation-detachment speed of oil molecules. The current MD simulation results may be helpful in understanding the interaction between the surfactant solution and crude oil and give some new insights into the surfactant flooding EOR process.

Figure 12. MSD curves of water molecules around the headgroup of surfactants in the DTAB-II and SDBS-II systems. The curve is an average of 10 ns simulations (in a time period of 10−20 ns).



phenomenon, the effect of surfactant flooding was better if more surfactant molecules were aggregated at the rear bottom of the oil aggregate. As can be seen from Figures S7−S10, compared with the SDBS systems, the DTAB molecules moved to the rear side of the oil aggregate with slower speed, and many of them were dispersed on the upper surface of heavy oil, which resulted in a poorer flooding effect and the oil aggregate had evolved into a “stripe” configuration. Finally, the molecular structure effect of different oil-driven performance of SDBS and DTAB was analyzed. We proposed that the better oil-driven efficiency of SDBS is due to the stronger H-bond interactions between the −SO3− headgroup and water. Compared to the −N(CH3)3+ headgroup in DTAB, the bare O atom in the −SO3− group formed more H-bonds with adjacent water molecules. The stronger interaction between the headgroup of SDBS and the adjacent water molecules results in a larger “friction” force to drive the migration of surfactant molecules on the surface of the oil molecular aggregate and hence accelerated the deformation and detachment of oil molecules. In Figure 12, we displayed the mean square displacements (MSD) of water molecules that were in close proximity to the surfactant headgroups. To be specific, we had performed the calculations for those water molecules which reside within 6 Å from the polar atoms (S or N) of the surfactant headgroups. This distance essentially corresponded to the first hydration layer with respect to the headgroups. The distances were measured by tagging the water molecules at different time origins. It was evidently seen from Figure 12 that the translational mobility of the hydration layer water in the SDBS system is lower than that of the DTAB system, which indicated that the SO3− group has stronger attractions with water molecules compared to the −N(CH3)3+, which explained the faster migration speed of SDBS on the oil aggregate and its higher oil-driven performance.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcb.8b09777.



The snapshots of MD trajectories of various simulation systems are displayed, and the force field parameters used are given (PDF)

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. *E-mail: [email protected]. ORCID

Yong Pei: 0000-0003-0585-2045 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work is supported by PetroChina Scientific Research and Technology Development Project (2014A-1001), Research Funds of Xiangtan University (2017XZX28), and National Natural Science Foundation of China (21773201).



REFERENCES

(1) Jian, C.; Liu, Q.; Zeng, H.; Tang, T. A molecular dynamics study of the effect of asphaltenes on toluene/water interfacial tension: Surfactant or solute. Energy Fuels 2018, 32, 3225−3231. (2) 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, 700−714. (3) Yaseen, S.; Mansoori, G. A. Asphaltene aggregation due to water flooding (A molecular dynamics study). J. Pet. Sci. Eng. 2018, 170, 177−183. (4) Xie, W. K.; Sun, Y. Z.; Liu, H. T. Atomistic investigation on the detachment of oil molecules from defective alumina surface. Appl. Surf. Sci. 2017, 426, 504−513. (5) Jin, Y.; Liu, X.; Liu, Z.; Lu, S.; Xue, Q. Effect of interfacial layer on water flow in nanochannels: Lattice Boltzmann simulations. Phys. B 2016, 487, 18−24. (6) Rezaei, N.; Zendehboudi, S.; Chatzis, I.; Lohi, A. Combined benefits of capillary barrier and injection pressure control to improve

4. CONCLUSION In summary, by employing NEMD simulations, the detachment of the crude oil aggregate in the silica nanoslit and the nanochannel under surfactant flooding were investigated. The simulation results indicated a three-stage process of surfactant flooding driven oil-detachment, including the initial formation of surfactant micelles and delivery, the disintegration-spread J

DOI: 10.1021/acs.jpcb.8b09777 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B fluid recovery at breakthrough upon gas injection: An experimental study. Fuel 2018, 211, 638−648. (7) Escrochi, M.; Mehranbod, N.; Ayatollahi, S. The gas−oil iterfacial behavior during gas injection into an asphaltenic oil reservoir. J. Chem. Eng. Data 2013, 58, 2513−2526. (8) Pei, H.; Zhang, G.; Ge, J.; Tang, M.; Zheng, Y. Comparative effectiveness of alkaline flooding and alkaline−surfactant flooding for improved Heavy-Oil recovery. Energy Fuels 2012, 26, 2911−2919. (9) Zhu, Y.; Hou, Q.; Jian, G.; Ma, D.; Wang, Z. Current development and application of chemical combination flooding technique. Petrol. Explor. Dev. 2013, 40, 96−103. (10) Zhou, X.; Dong, M.; Maini, B. The dominant mechanism of enhanced heavy oil recovery by chemical flooding in a twodimensional physical model. Fuel 2013, 108, 261−268. (11) Zhang, H.; Dong, M.; Zhao, S. Which one is more important in chemical flooding for enhanced court heavy oil recovery, lowering interfacial tension or reducing water mobility? Energy Fuels 2010, 24, 1829−1836. (12) Afsharpoor, A.; Javadpour, F. Liquid slip flow in a network of shale noncircular nanopores. Fuel 2016, 180, 580−590. (13) Yan, H.; Yuan, S. Molecular dynamics simulation of the oil detachment process within silica nanopores. J. Phys. Chem. C 2016, 120, 2667−2674. (14) 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, 2867− 2875. (15) Liu, B.; Wang, C.; Zhang, J.; Xiao, S.; Zhang, Z.; Shen, Y.; Sun, B.; He, J. Displacement mechanism of oil in shale inorganic nanopores by supercritical carbon dioxide from molecular dynamics simulations. Energy Fuels 2017, 31, 738−746. (16) Wang, S.; Javadpour, F.; Feng, Q. Fast mass transport of oil and supercritical carbon dioxide through organic nanopores in shale. Fuel 2016, 181, 741−758. (17) Xie, W. K.; Sun, Y. Z.; Liu, H. T.; Zhang, F. H. Investigations on oil detachment from rough surfaces in an aqueous solution. RSC Adv. 2017, 7, 31049−31060. (18) Zehl, T.; Wahab, M.; Schmidt, R.; Schiller, P.; Mögel, H. J. Monte Carlo studies on self-assembly of surfactants in aqueous solutions. J. Mol. Liq. 2009, 147, 178−181. (19) Chen, C.; Zhuang, L.; Li, X.; Dong, J.; Lu, J. A Many-Body dissipative particle dynamics study of forced water−oil displacement in capillary. Langmuir 2012, 28, 1330−1336. (20) Xiong, Y.; Cao, T.; Chen, Q.; Li, Z.; Yang, Y.; Xu, S.; Yuan, S.; Sjöblom, J.; Xu, Z. Adsorption of a polyaromatic compound on silica surfaces from organic solvents studied by molecular dynamics simulation and AFM imaging. J. Phys. Chem. C 2017, 121, 5020− 5028. (21) Greathouse, J. A.; Cygan, R. T.; Fredrich, J. T.; Jerauld, G. R. Adsorption of aqueous crude oil components on the basal surfaces of clay minerals: Molecular simulations including salinity and temperature effects. J. Phys. Chem. C 2017, 121, 22773−22786. (22) Wang, X.; Zhang, Z.; Torsæter, O.; He, J. Atomistic insights into the nanofluid transport through an ultra-confined capillary. Phys. Chem. Chem. Phys. 2018, 20, 4831−4839. (23) Chen, J.; Yu, H.; Fan, J.; Wang, F.; Lu, D.; Liu, H.; Wu, H. Channel-width dependent pressure-driven flow characteristics of shale gas in nanopores. AIP Adv. 2017, 7, 045217. (24) Xiao, J.; Pan, F.; Xia, H.; Zou, S.; Zhang, H.; George, O. A.; Zhou, F.; Huang, Y. Computational study of single droplet deposition on randomly rough surfaces: Surface morphological effect on droplet impact dynamics. Ind. Eng. Chem. Res. 2018, 57, 7664−7675. (25) Bahramian, Y.; Bahramian, A.; Javadi, A. Confined fluids in clay interlayers: A simple method for density and abnormal pore pressure interpretation. Colloids Surf., A 2017, 521, 260−271. (26) Yang, S.; Dehghanpour, H.; Binazadeh, M.; Dong, P. A molecular dynamics explanation for fast imbibition of oil in organic tight rocks. Fuel 2017, 190, 409−419.

(27) Wang, S.; Javadpour, F.; Feng, Q. Molecular dynamics simulations of oil transport through inorganic nanopores in shale. Fuel 2016, 171, 74−86. (28) Moncayo-Riascos, I.; de León, J.; Hoyos, B. A. Molecular dynamics methodology for the evaluation of the chemical alteration of wettability with organosilanes. Energy Fuels 2016, 30, 3605−3614. (29) 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, 164702. (30) Millan, J. A.; Laradji, M. Cross-Stream migration of driven polymer solutions in nanoscale channels: A numerical study with generalized dissipative particle dynamics. Macromolecules 2009, 42, 803−810. (31) Sedghi, M.; Piri, M.; Goual, L. Atomistic molecular dynamics simulations of crude Oil/Brine displacement in calcite mesopores. Langmuir 2016, 32, 3375−3384. (32) Spiecker, P. M.; Gawrys, K. L.; Trail, C. B.; Kilpatrick, P. K. Effects of petroleum resins on asphaltene aggregation and water-in-oil emulsion formation. Colloids Surf., A 2003, 220, 9−27. (33) Speight, J. G. Petroleum Asphaltenes - Part 1: Asphaltenes, resins and the structure of petroleum. Oil Gas Sci. Technol. 2004, 59, 467−477. (34) Pierre, C.; Barré, L.; Pina, A.; Moan, M. Composition and heavy oil rheology. Oil Gas Sci. Technol. 2004, 59, 489−501. (35) Yan, Y.; Li, C.; Dong, Z.; Fang, T.; Sun, B.; Zhang, J. Enhanced oil recovery mechanism of CO2 water-alternating-gas injection in silica nanochannel. Fuel 2017, 190, 253−259. (36) Koretsky, C. M.; Sverjensky, D. A.; 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, 349−438. (37) Wu, G.; Zhu, X.; Ji, H.; Chen, D. Molecular modeling of interactions between heavy crude oil and the soil organic matter coated quartz surface. Chemosphere 2015, 119, 242−249. (38) Li, X.; Xue, Q.; Zhu, L.; Jin, Y.; Wu, T.; Guo, Q.; Zheng, H.; Lu, S. 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. (39) He, L.; Wang, P.; He, L.; Qu, Z.; Luo, J.; Peng, B.; Tang, X.; Pei, Y. Molecular dynamics simulations of the self-organization of side-chain decorated polyaromatic conjugation molecules: Phase separated lamellar and columnar structures and dispersion behaviors in toluene solvent. RSC Adv. 2018, 8, 11134−11144. (40) Plimpton, S. Fast parallel algorithms for Short-Range molecular dynamics. J. Comput. Phys. 1995, 117, 1−19. (41) Sun, H.; Mumby, S. J.; Maple, J. R.; Hagler, A. T. An ab Initio CFF93 All-atom force field for polycarbonates. J. Am. Chem. Soc. 1994, 116, 2978−2987. (42) Sengupta, S.; Pant, R.; Komarov, P.; Venkatnathan, A.; Lyulin, A. V. Atomistic simulation study of the hydrated structure and transport dynamics of a novel multi acid side chain polyelectrolyte membrane. Int. J. Hydrogen Energy 2017, 42, 27254−27268. (43) Xu, J.; Zhang, Y.; Chen, H.; Wang, P.; Xie, Z.; Yao, Y.; Yan, Y.; Zhang, J. Effect of surfactant headgroups on the oil/water interface: An interfacial tension measurement and simulation study. J. Mol. Struct. 2013, 1052, 50−56. (44) Gamba, M.; Kovár,̌ P.; Pospíšil, M.; Torres Sánchez, R. M. Insight into thiabendazole interaction with montmorillonite and organically modified montmorillonites. Appl. Clay Sci. 2017, 137, 59− 68. (45) Xu, Y.; Liu, Y.; Liu, G. Molecular dynamics simulation of primary ammonium ions with different alkyl chains on the muscovite (001) surface. Int. J. Miner. Process. 2015, 145, 48−56. (46) Chen, Y.; Xu, G.; Yuan, S.; Sun, H. Molecular dynamics simulations of AOT at isooctane/water interface. Colloids Surf., A 2006, 273, 174−178. K

DOI: 10.1021/acs.jpcb.8b09777 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B (47) Tokarský, J.; Andrýsek, T.; Č apková, P. Molecular modeling of gel nanoparticles with cyclosporine a for oral drug delivery. Int. J. Pharm. 2011, 410, 196−205. (48) Samieadel, A.; Oldham, D.; Fini, E. H. Multi-scale characterization of the effect of wax on intermolecular interactions in asphalt binder. Constr. Build. Mater. 2017, 157, 1163−1172. (49) Markina, A.; Ivanov, V.; Komarov, P.; Khokhlov, A.; Tung, S. H. Self-assembly of micelles in organic solutions of lecithin and bile salt: Mesoscale computer simulation. Chem. Phys. Lett. 2016, 664, 16−22. (50) Tang, J.; Qu, Z.; Luo, J.; He, L.; Wang, P.; Zhang, P.; Tang, X.; Pei, Y.; Ding, B.; Peng, B.; Huang, Y. Molecular dynamics simulations of the Oil-Detachment from the hydroxylated silica surface: Effects of surfactants, electrostatic interactions, and water flows on the water molecular channel formation. J. Phys. Chem. B 2018, 122, 1905−1918. (51) Darden, T.; York, D.; Pedersen, L. Particle mesh Ewald: An N· log(N) method for Ewald sums in large systems. J. Chem. Phys. 1993, 98, 10089−10092. (52) Jones, J. L.; Lal, M.; Noel Ruddock, J.; Spenley, N. A. Dynamics of a drop at a liquid/solid interface in simple shear fields: A mesoscopic simulation study. Faraday Discuss. 1999, 112, 129−142. (53) Zhang, Y.; Xu, J.; He, X. Effect of surfactants on the deformation of single droplet in shear flow studied by dissipative particle dynamics. Mol. Phys. 2018, 116, 1851−1861. (54) Sergi, D.; Scocchi, G.; Ortona, A. Molecular dynamics simulations of the contact angle between water droplets and graphite surfaces. Fluid Phase Equilib. 2012, 332, 173−177.

L

DOI: 10.1021/acs.jpcb.8b09777 J. Phys. Chem. B XXXX, XXX, XXX−XXX