- Email: [email protected]
Adsorption behaviors of shale oil in kerogen slit by molecular simulation
Journal Pre-proofs Adsorption Behaviors of Shale Oil in Kerogen Slit by Molecular Simulation Yongfei Yang, Jie Liu, Jun Yao, Jianlong Kou, Zheng Li, Tianhao Wu, Kai Zhang, Lei Zhang, Hai Sun PII: DOI: Reference:
S1385-8947(20)30045-0 https://doi.org/10.1016/j.cej.2020.124054 CEJ 124054
To appear in:
Chemical Engineering Journal
Received Date: Revised Date: Accepted Date:
23 October 2019 31 December 2019 4 January 2020
Please cite this article as: Y. Yang, J. Liu, J. Yao, J. Kou, Z. Li, T. Wu, K. Zhang, L. Zhang, H. Sun, Adsorption Behaviors of Shale Oil in Kerogen Slit by Molecular Simulation, Chemical Engineering Journal (2020), doi: https:// doi.org/10.1016/j.cej.2020.124054
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2020 Published by Elsevier B.V.
Adsorption Behaviors of Shale Oil in Kerogen Slit by Molecular Simulation
Yongfei Yang a *, Jie Liu a, Jun Yao a, Jianlong Kou b, Zheng Li a, Tianhao Wu c, Kai Zhang a, Lei Zhanga, Hai Suna
a Key
Laboratory of Unconventional Oil & Gas Development (China University of
Petroleum (East China)), Ministry of Education, Qingdao 266580, P.R. China & Research Center of Multiphase Flow in Porous Media, School of Petroleum Engineering, China University of Petroleum (East China), Qingdao, 266580, P.R. China b Institute of Condensed Matter Physics, Zhejiang Normal University, Jinhua 321004 , P.R. China c Reservoir Engineering Research Institute, 595 Lytton Avenue Suite B, Palo Alto, CA 94301, USA *Address all correspondence to Yongfei Yang, E-mail: [email protected] ABSTRACT
Shale oil is widely distributed in organic nanopores, while kerogen plays a complex and key role for adsorption behavior of shale oil, and thus it is crucial to identify the associated storage mechanisms. In this paper, molecular dynamic simulation had been performed to quantify the adsorption behavior of shale oil in kerogen slits. Both the distribution of shale oil properties and potential of the mean force (PMF) were used to identify the interaction mechanisms between the light and heavy components respectively represented by methane and asphaltene. We also examined the effects of different temperatures and apertures on the adsorption behavior. Owning to the extremely strong adsorption capacity between the
asphaltene and kerogen, the adsorbed asphaltene layers reduce the slit width, preventing the light components from adsorbing on the kerogen slits due to the energy barrier formed by heavy components. It is found that, with an increase in temperature, the distribution of hydrocarbons performs more homogeneously. In addition, the adsorption quantity of medium components displays a reduction in kerogen slit, while the heavy component shows a rising as its greater competitive, suggesting that the medium components are the most potential fraction in thermal exploitation, and the light components keep a steady quantity with the combined action of medium and heavy components. The small slit (aperture < 2 nm) can be blocked by asphaltene molecules, and the adsorption density of hydrocarbons reaches the maximum at 2 nm aperture.
Keywords: Adsorption behavior, Shale oil, Kerogen, Molecular simulation
1 Introduction The depletion of the world’s resources is offsetting the declines in conventional oil and gas production, and exploiting the shale oil and gas is changing the global energy system [1]. The shale oil boom has stimulated the oil and gas industry to start the corresponding programs to exploit shale oil and gas [2, 3]; however, the abundant nanopores and extremely low permeability of shale complicate the exploitation and understanding of the adsorption behavior within it [4-7]. In the unconventional shale reservoirs, the pore size is usually in the range of nanometer [8]. By simulating the distribution of hydrocarbons in kerogen nanopores, Perez et al. [9] distinguished the micropores (aperture < 2 nm) and mesopores (2 nm< aperture < 50 nm) in kerogen matrix. The experiments using atomic force microscopy (AFM) confirm the existence of extremely small pores in a range of 2-200 nm [10, 11]; but it is extremely difficult to experimentally characterize the adsorption behavior because such nanopores are below the
optical resolution, and molecular simulation is an effective way to study the molecular behavior in nanopores [12-14]. Shale matrix is a complex porous medium that is composed of quartz minerals and organic matters, among which the latter is mainly the kerogen, i.e., the main source of oil [15]. The hydrocarbon in shale is mostly in the micropore of organic matter generated over geological time [16]. Therefore, it is of practical and fundamental importance to quantify the hydrocarbon adsorption within the kerogen channel in porous media, and the hydrocarbons adsorption in nanopores is also common in effluent treatment [17] and the coating corrosion [18]. Physically, the molecular structure of kerogen is complex, while its physicochemical properties rely on the origin and depositional environments [19, 20]. Previous efforts have been made to build the molecular nanostructure for kerogen by using the diffraction measurements and the generalized phonon densities of states, and the kerogen can be classified into four different types with its maturity indicator [21, 22]. In previous studies, the rigid carbon nanotubes (CNTs) and graphene slits are usually used as the channel models, which make the quantitative analysis of fluid properties more precise [11, 13, 23]. But the extremely smooth wall and pure carbon atoms could overestimate the adsorbing capacity of hydrocarbon mixtures. The adsorbed phases perform different physical and chemical behaviors in such small pore space, due to the affinity for the organic wall [24]. Previous studies found that the source oils contained more aromatic hydrocarbons than reservoir oils, such as methylbenzene and asphaltene [25-27]. Perez et al. [28] determined the adsorption of hydrocarbon mixtures in pore space of kerogen matrix, and the adsorption layer of heavy components reduced the adsorption space of other components, such as methane. The intermediate components (C2 - C12 and methylbenzene) tend to show more adsorption on the pore surface with the increase of carbon chain [29], while the shorter chain alkanes facilitate the exploitation [30]. With the help of molecular dynamics and statistical mechanics to elucidate the diffusion and adsorption of
hydrocarbon in kerogen matrix, Obliged et al. [31] found that the permeance was decreased with an increase in the alkane length. Zhang et al. [32] calculated the interaction energy between kerogen wall and oil molecules using the umbrella sampling method, and they found that the clustered oil molecules inclined to desorb from the kerogen surface compared with single oil molecule. By using equilibrium molecular dynamic (EMD) method, we constructed a slit by two kerogen matrixes and studied the adsorption behavior of shale oil in kerogen slit for the first time, revealing the underlying mechanisms of adsorption and the heterogeneous distribution between shale oil and kerogen by calculating the PMF. Not only did we quantify the density distribution of shale oil within kerogen slit in one direction, but generated the two-dimensional density contour of whole system because of the complex structure of kerogen matrix. Our results show that the heavy component is tightly adsorbed on the surface of kerogen matrix, and the adsorbed layer weakens the interaction force between kerogen and light component, leading to the heterogeneous distribution of mixture density. We also examined the effects of temperature and slit width on the density distribution of shale oil.
2 Methodology 2.1 Molecular models In this work, we used the kerogen monomers and hydrocarbon mixtures to construct a three-dimensional rough kerogen slit as shown in Fig. 1. We selected the type Ⅱ-C kerogen, as it was oil-prone that corresponded to the kerogen in organic-rich shales [22]. The kerogen chemical formula is C242H219O13N5S2 and its molecular structure is also presented in Fig. 1. We set the hydrocarbon mixtures in the outside of slit to obtain the selectivity of kerogen slit for hydrocarbons. The composition of the shale oil was modified and given in Table 1. The hydrocarbon mixture model was simplified and followed the work of Felipe et al. [9], and we
selected the single generic asphaltene model with chemical formula C26OH32 [28, 33]. We ignored some trivial components in that the elaborate components would make the molecular simulation intractable. For a slit of 2 nm aperture, the dimensions of the simulation box are 3.5 × 11.0 × 14.6 nm3. The kerogen matrixes were packed in the both side of the box along y direction, and we constructed the slit models with apertures from 1 nm to 3 nm. We did not simulate the adsorption in slit with larger aperture. Because the larger slit volume weakens the interaction of kerogen, and the hydrocarbon mixtures show more randomness which reduces the accuracy of statistics. Each matrix was constructed with fourteen kerogen monomers and its average density was controlled at 0.98 ± 0.01 g·cm-3 [34]. The models and snapshots were visualized by visual molecular dynamics (VMD) package [35]. The polymer consistent force field plus (PCFF+) was used for all molecules [22, 36, 37]. Periodic boundary conditions were applied in three directions. The Van der Waals interactions were described by using Lennard-Jones (LJ) potentials with Waldman-Hagler combining rules [38], and the cutoff distance was 1.2 nm. We adopted the Ewald method to compute the electrostatic potential [39].
Fig. 1. Kerogen slit was constructed by kerogen matrix and hydrocarbon mixtures, and each component was represented in one color to clearly show the different components. Components color scheme: black, kerogen matrixes; cyan, asphaltene; gray, methylbenzene; purple, n-dodecane; mauve, noctane; pink, n-butane; green, propane; yellow, ethane; blue, methane. Elements color scheme of kerogen and asphaltene molecules: black, C; blue, N; red, O; yellow, S; white, H.
Table 1 Molecular fraction of shale oil components
Species
Mole fraction [%]
Number of molecules
methane ethane propane n-butane n-octane n-dodecane methylbenzene asphaltene
63.5 7.8 4.8 3.2 6.3 4.8 4.8 4.8
1378 169 104 69 137 104 104 103
Molar mass [g·mol-1] 16.535 30.433 44.097 58.124 144.230 170.335 92.140 360.538
2.2 Simulation methods The molecular simulations were performed with the large-scale atomic/molecular massively parallel simulator (LAMMPS) package [40]. We took the MD simulation in the NVT ensemble with a timestep of 0.1 fs on the simulation box for 100 ps which reduced the energy state and got the relaxed configuration of system. Then, we put the simulation system into the NPT ensemble at T = 300 K and stabilized system pressure at 27.5 MPa which consumed 1 ns with a timestep of 1 fs. For the EMD simulations, the NVT ensemble was taken at 300 K during 11 ns, and the last 8 ns of trajectories were collected for analysis. During the simulation, we freezed the kerogen matrix and controlled the temperature of hydrocarbon mixtures by the method of Nosé-Hoover thermostat [41, 42]. The density profiles were calculated by averaging the number of atoms in the bins that parallel to the kerogen wall [43]. The PMF was calculated with colvars package in LAMMPS, which was used to reveal the free energy distribution and the adsorption relation of asphaltene and methane [44, 45], and this method was also called steered and targeted MD due to its spring force on the operated molecule [46]. The result data were collected and analyzed by the umbrella sampling method and the weighted histogram analysis method (WHAM) [44]. Due to the complexity of shale oil, we only selected the heaviest component and lightest component as the operated molecules, i.e., the asphaltene and methane, to compute the PMF between kerogen wall and the operated molecule in the mixed components system. Firstly, we applied a harmonic potential on the operated molecule, and the spring force constant of operated methane and asphaltene molecules were 8.37 kJ·mol-1·Å-2 and 29.29 kJ·mol-1·Å-2 respectively. It took 3.5 ns to reach equilibrium before the umbrella sampling with a timestep of 1 fs. In this case, each window was set as a width of 0.1 nm which took 0.5 ns to reach the equilibrium state, and the whole process included 35 windows that cost 17.5 ns. We fixed the operated molecule on z direction to make sure that they could move on the straight path.
3 Results and discussions 3.1 Density distribution of hydrocarbons in kerogen slit We performed EMD simulations to investigate the adsorption behaviors of hydrocarbon mixtures in kerogen slits. The computed density profiles were depicted in Fig. 2 (T = 300 K, aperture = 2 nm). We run the simulations three times independently to improve the accuracy. The rough surface of kerogen matrix made it difficult to define boundary [47, 48], thus we defined the crosspoint of shale oil and kerogen density profiles as the slit boundary [49]. We made statistics of the number density of hydrocarbon atoms in slit region to show the distribution features. In Fig. 2(a), the whole hydrocarbon mixtures shows the two density peaks in the vicinity of kerogen matrix, which correspond with the simulation results in graphene slits [36]. The methane exhibits a gentle enrichment in the bulk space where the number of kerogen atoms equal to zero. The asphaltene density profile displays two peaks, implying a stronger affinity between asphaltene and kerogen, and the randomness of molecular thermal motion leads to the different values of two peaks. The heterogeneous phenomenon is revealed in the section of PMF analysis. We notice that the kerogen density profiles shake along coordinate axis, as a result of the complex molecular structure of kerogen and porous matrix. The kerogen density profile keeps almost constant over 17.5 Å from the center of slit, then declining to zero, suggesting the roughness surface of kerogen wall has more spatial pores for adsorbate. And the significant characteristic of mixtures density profile is the invaded hydrocarbons on the rough kerogen matrix surface. A few of asphaltene molecules insert in the porous structure of slits surface, in addition, the further padding of smaller molecules reduces the roughness of slit surface. On the base of the preliminary adsorption in the pores of the surface, the hydrocarbon mixtures tend to form adsorbed layers. It is interesting that the peaks of asphaltene and hydrocarbon mixtures
profiles are always corresponding, and they always located in the vicinity of kerogen matrix. To explain this corresponding phenomenon of peaks, as shown in Fig. 2(b), we merge the specific components into light (C1 - C4), medium (C8, C12 and methylbenzene) and heavy (asphaltene) components respectively according to the difference of molecular weight. The light component profile shows a gently single peak, displaying the similar distribution with methane. And the medium component profile presents a higher single peak in the bulk area of slits, revealing that the asphaltene molecules contribute to the double density peaks characteristic of mixture mostly in the slit of 2 nm.
(a)
(b) Fig. 2. Number density profile of kerogen and mixtures in slit region, (a) methane and asphaltene are presented as the lightest and heaviest components, (b) light, medium and heavy components are presented. The shale oil is the sum of whole hydrocarbon mixtures.
Fig. 3. Two-dimensional number density contour of mixtures, to make the profile clearly, the kerogen matrix is the black area.
For a better understanding of the density distribution in such heterogeneous structure of kerogen slit, we calculated the two-dimensional density contour of hydrocarbon mixtures for whole system. As shown in Fig. 3, the high-density adsorbed layers surround the kerogen matrixes, reflecting the strong lipophilicity of kerogen. The density values of mixtures diminish gradually as away from kerogen matrix. According to the previous analysis of different components, the light components tend to distribute in the free volume away from kerogen matrixes. Such heterogeneous density distribution affects the produced fluid compositions in oilfield. The light components in free volume can move with less constraints formed by kerogen, thus the lighter fluid is produced firstly in the oil field. With the depletion of lighter hydrocarbons, the heavier hydrocarbons contribute the production increasingly [9, 50, 51]. In the two-dimensional density contour, the insertion phenomenon of hydrocarbons displays clearly on the porous surface of kerogen matrixes. We also notice that some pores in kerogen matrix trap a few of hydrocarbons. In the slit region, apart from the adsorbed layer on the surface of slit, the high-density values present in the center of slit that is affected by the interaction force of two kerogen walls, and they are mainly composed of medium component as shown in Fig. 2(b). In this case, the enrichment of hydrocarbons clogs the nanochannel, and reduces the connectivity of porous medium.
3.2 Potential of the mean force analysis
(a)
(b) Fig. 4. (a) Potential of the mean force between asphaltene and kerogen (PMFa, the blue oval is the initial asphaltene and the orange oval is the asphaltene barrier), the arrow is the calculation direction. (b) Potential of the mean force between methane and kerogen (PMFm).
In order to get insight into the underlying mechanisms responsible for the heterogeneous distribution between the heavy component and the light component, as shown in Fig. 4, we calculated the PMF of asphaltene and methane on kerogen wall in hydrocarbon mixtures respectively, because the asphaltene and methane exhibited similar distribution properties to heavy and light components in Fig. 2. The overall trend of PMF curves were corresponding to the previous work [32], in addition, we took the effect of adsorbed layer into consideration. The results are shown form the midplane of the slit. In this case, Fig. 4(a) presents a slow decline of PMFa from bulk region to the kerogen wall before the first valley value. When the asphaltene molecule gets closer to the kerogen wall, the PMFa down to 0 kJ·mol-1 at 26 Å, where the asphaltene is most stable due to its lowest energy state, indicating that the asphaltene molecules tend to distribute in the vicinity of kerogen rather than the bulk region. After that, the PMFa keeps a sharp increase because of the strong repulsive interaction in the condition of extremely short range between atoms. It is very interesting that the second valley value occurs in the increasing process of PMFa, which is 28.24 kJ·mol-1 at 31 Å. And this PMFa peak implies the potential energy barrier formed by the adsorption layer. Therefore, the first valley facilitates the adsorption on the adsorption layer, effected by both adsorbed layer and kerogen surface, and the second valley reveals the initial adsorption between asphaltene and kerogen. Between the two valleys, the barrier width is 5 Å, which approaches to the thickness of asphaltene molecule. After 31 Å, the value of PMFa increases to infinity which means the operated asphaltene cannot be closer to kerogen. For the PMFm profiles in Fig. 4(b), it is always zero before 23.5 Å, which illustrates that the methane is more stable in bulk phase. The reduction of PMFa value is 8.37 kJ·mol-1 after the second valley, while the decline of PMFm is only 0.95 kJ·mol-1, thus the asphaltene has more affinity for kerogen. In addition, it also demonstrates that the methane is less affected by adsorption layer as a result of its small size, which causes that some methane molecules can
insert into the kerogen matrix. By comparing PMFa with PMFm, the adsorption layer has more influence distance on asphaltene than methane, and the influence distance includes not only the adsorption thickness, but also the energy barrier which is formed by adsorption layer. The adsorption of asphaltene can reduce the effective width of nanochannels, thereby enriching the methane in bulk phase. 3.3 Effect of temperature on hydrocarbons distribution in kerogen slit
(a)
(b) Fig. 5. Two-dimensional number density contour of mixture with various temperature, (a) T = 360 K, (b) T = 380 K.
To investigate the relationship between temperature and hydrocarbon adsorption, we calculated the two-dimensional density contour and counted the number density of hydrocarbons in the slit region from 280 K to 380 K. We kept the same quantity of hydrocarbon molecules to make the results comparable at different temperature. The Fig. 3 presented the density distribution at 300 K, and we chose representative graphs at 360 K and 380 K as shown in Fig. 5. The density distributions in vertical and horizontal directions are similar at three temperatures, and it is also hard to get accurate representations from the one-dimensional distribution. In the condition of 300 K, a large area of high-density cluster surrounds the kerogen matrixes. From the kerogen to the bulk region, the density contour shows a descending in steps. When the temperature rises to 360 K, the density distribution performs more homogeneously, as a result of the more intense molecular thermal motion. The higher temperatures reduce the area of high density, and we cannot observe the high-density clumps
any more at 380 K except for the adsorbed layer. As shown in the two-dimensional density contour, with the increase of temperature, the hydrocarbons tend to distribute on the middle of slit in vertical direction. The stronger thermal motions of mixture molecules counteract the confined effects of kerogen slit, making the molecules getting out of the slit more easily. Nevertheless, for the center of the slit, the kerogen interaction still plays a major role for the trapped molecules. Hence the hydrocarbons concentrate in the center of slit at high temperature, and fewer of them in the opening of slit.
Fig. 6. The adsorption number of different component atoms in the slit region with various temperature, aperture = 2 nm.
By counting the number of hydrocarbon atoms in slit, we quantified the influence of temperature on adsorption. In contrast with the chain alkane models and the smooth wall, the asphaltene models and rough kerogen wall have more complex topological structures, which introduce more instability into the results. Therefore, the results of multicomponent system always perform larger fluctuations in such complex kerogen slit. As shown in Fig. 6, the adsorption number of whole hydrocarbons performs a linear decline on account of the
increasingly thermal motion. As temperature rises, the adsorption competition of hydrocarbons with higher molecular weight tends to be stronger [52], leading to a rise of heavy component content, thereby the medium components decline, which reveals that the medium component is the major contributor for the thermal recovery and the heavy component can accumulate in nanoporous shale. The adsorption of light components performs a slight drop with various temperatures. Due to the higher kinetic energy at high temperature, the adsorption quantity of light components should be reduced, but the decrease of the sum of medium and heavy components leaves more free space for the light components. With these two opposite effects, the quantity of light hydrocarbons keeps stable. The complementary trend of medium and heavy components, we summed them together and found that it had a similar slope to the adsorption number of shale oil. Therefore, the sum of medium and heavy components reflects the whole developing trend in thermal recovery. 3.4 Effect of slit aperture on hydrocarbons distribution in kerogen slit
(a)
(b) Fig. 7. The two-dimensional number density distribution of mixtures. (a) Aperture = 1 nm. (b) Aperture = 3 nm. (T = 300 K)
By calculating the two-dimensional density contours with different slit apertures, we examined the dependence of hydrocarbon mixture adsorption. To get accurate data, we kept the pressure (27.5 MPa) of shale oil consistent in whole system of different width. As shown in Fig. 7(a), the density profile performs a single peak in 1 nm slit, which corresponds to the previous study, indicating that the mixtures perform the similar distribution with pure alkane components within 1 nm slit [53]. For the complex molecular structure, the asphaltene is hard to enter the 1 nm slit. In the vertical direction, caused by the clogging of heavy components in the slit opening, the density curve shows a concave in the middle of slit, and we can consider that the hydrocarbons block the nanochannel. With the increase of slit width, the phenomenon of clogging becomes less apparent, because the increase of aperture weakens the dominant adsorption effects of kerogen wall. In the 3 nm slit, the profile still depicts the double slight density peaks in kerogen slit
region, and more hydrocarbons enter into the slit and increase the bulk density of slit. Besides, we notice that the two-dimensional density distribution shows more randomness. The Fig. 7(b) shows that, apart from the adsorption layer around the kerogen matrixes, the remaining asphaltene molecules form several high-density clusters in the bulk region out of the slit. Once the hydrocarbons get into the bulk region, the intermolecular interactions of mixtures will play a bigger role for the distribution behaviors. Thus, the asphaltene molecules tend to aggregate as clusters, which attract the lighter components, and then influence the shape of low-density area, such as the convex and concave of density contour line.
(a)
(b) Fig. 8. The adsorption number of component atoms (a) in the slit region and (b) its adsorption per unit of slit aperture (nm) with different apertures, T = 300 K.
To quantify the relation between slit aperture and hydrocarbon density, we counted the number of hydrocarbons in the slit region. The adsorption number of integral hydrocarbons exhibits a continuous rise by widening the slit in Fig. 8(a), which is generally accepted. We are curious about the phenomenon that the quantity of heavy component continues arising and then maintains steady, which is demarcated by 2 nm aperture. The asphaltene molecule has a length of 1.8 nm and a width of 0.5 nm, thus the asphaltene molecules have hard access to the smaller slits (aperture < 2 nm). But asphaltene molecules still can enter the channel at sloping or horizontal angles. With the increase of slit width, the asphaltene can get into the slit more easily. Thus, the heavy component shows a rise with a bigger slope for a slit of above 1.5 nm. The asphaltenes form the clusters with the effect of self-association, but it is uneasy to get into the slit with the form of cluster. In the 2 nm slit, some free asphaltene monomers which are not trapped by clusters can enter the slit with the geometric constraint, driven by the adsorption,
and then form the adsorbed layer on the two kerogen walls. According to PMF analysis, the asphaltene clusters are unstable in the bulk region of slit, and the adsorbed layer weakens the interaction between the kerogen wall and asphaltene molecules in bulk phase of slit. Therefore, the density of asphaltene keeps steady above 2 nm slits, and the asphaltene cluster may get into the slit and cause a rising adsorption when the apertures are wider that the hydrocarbons are unaffected by kerogen. As oppose to the asphaltene, the medium and light components increase at a steeper slope. Because of the increasing bulk volume of slit, more light and medium molecules can get into the slit region. To investigate the adsorptive contribution among three components with increasing aperture, we summed adsorption quantity per unit aperture of the light and medium components in Fig. 8(b). At aperture of 2 nm, both the adsorption of heavy component and whole hydrocarbons depicts a maximum. The wider slit (aperture > 2 nm) weakens the interaction between the kerogen wall and hydrocarbons, leading to the decrease of adsorption per unit slit aperture, and revealing that the maximum adsorption density of hydrocarbon mixtures can be achieved in the aperture of close to heavy molecular size. For the increase of adsorption quantity, the heavy component plays a major role in the aperture of less than 2 nm, although the light component performs a slightly increase at 1.5 nm due to the small molecular structure. The adsorption quantity of light and medium components tends to increase and dominates the increase of adsorption quantity over 2 nm, while the content of heavy component shows a decline.
4 Conclusions In this work, the adsorption behaviors of shale oil in kerogen slit are investigated by using molecular dynamic simulations. We find that the heavy component represented by asphaltene performs a tremendous adsorption capacity on kerogen, while the light component, such as methane, enriches in the bulk region, and the medium component is adsorbed as the
intermediate states. The insertion of hydrocarbons smooths the porous surface of kerogen matrixes relatively, and then the asphaltene molecules facilitate the formation of first density peak on the smoothed surface. Apart from the dense adsorbed layer on the surface of slit, the hydrocarbons present an extra density cluster in the middle of slit, and reduce the connectivity of porous medium. Caused by the strong lipophilicity of kerogen wall and the difference of molecular weight, the shale oil exhibits heterogeneous distribution, which leads to the variation from light oil to heavy oil in production. The interactions between asphaltene molecules and kerogen surfaces take the leading role in the adsorption of hydrocarbons. As temperature rise, the density distribution of shale oil tends to be more homogeneous with less high-density clusters. The adsorption of shale oil in kerogen slit decreases with the increase of temperature, while the adsorption quantity of heavy component shows a rise, as the high temperature makes asphaltene more competitive. The adsorption quantity of medium components displays a downward trend, implying that they are the most potential fraction in thermal exploitation. The positive effects of medium component overweight the impeditive effects of heavy component, which contribute the effectiveness of thermal recovery in the nanochannels. It is hard to enter the small channel (aperture < 2 nm) for an asphaltene molecule caused by its complex structure. But a few of asphaltene molecules still can get into the slit with various angles, and block the opening of slit. Because of the weakened effect of kerogen by widening the slit, the distribution of high-density clusters performs more randomness. In addition, the adsorption quantity of shale oil keeps rising and reach the maximum of adsorption density in 2 nm slit, revealing that the aperture of close to heavy molecular size facilitates the maximum adsorption density of shale oil. The adsorption quantity of heavy component presents a maximum and keeps steady over 2 nm slit, suggesting that the adsorbed layer weakens the adsorption between kerogen wall and asphaltenes which in bulk phase. And the sum of light
and medium components dominates the increase of adsorption quantity in the nanochannel over 2 nm slit. This work provides the comprehensive descriptions of shale oil adsorption behaviors and builds up a good foundation for further studies on the transport of shale oil in kerogen slit. Acknowledgments We acknowledge the following financial support: National Natural Science Foundation of China (No. 51674280, 51722406, 51950410591), Shandong Provincial Natural Science Foundation (ZR2019JQ21, JQ201808), and Program for Changjiang Scholars and Innovative Research Team in University (IRT_16R69). References [1] J.D. Hughes, Energy: A reality check on the shale revolution, Nature 494 (2013) 307-308. [2] P.J.M. Monteiro, C.H. Rycroft, B. Grigory Isaakovich, A mathematical model of fluid and gas flow in nanoporous media, Proc Natl Acad Sci U S A 109 (2012) 20309-20313. [3] L. Maugeri, The shale oil boom: a US phenomenon, Harvard Kennedy School, Belfer Center for Science and International Affairs2013. [4] Y. Yang, J. Yao, C. Wang, Y. Gao, Q. Zhang, S. An, W. Song, New pore space characterization method of shale matrix formation by considering organic and inorganic pores, Journal of Natural Gas Science and Engineering 27 (2015) 496-503. [5] Y. Yang, K. Wang, L. Zhang, H. Sun, K. Zhang, J. Ma, Pore-scale simulation of shale oil flow based on pore network model, Fuel 251 (2019) 683-692. [6] R.A. Kerr, Natural Gas From Shale Bursts Onto the Scene, Science 328 (2010) 1624-1626. [7] Y. Wang, Nanogeochemistry: Nanostructures, emergent properties and their control on geochemical reactions and mass transfers, Chemical Geology 378-379 (2014) 1-23. [8] C.R. Clarkson, N.R. Solano, R.M. Bustin, A.M.M. Bustin, T.P. Blach, Pore structure characterization of North American shale gas reservoirs using USANS/SANS, gas adsorption, and mercury intrusion, Fuel 103 (2013) 606616. [9] P. Felipe, D. Deepak, Spatial distribution of reservoir fluids in mature kerogen using molecular simulations., Fuel 235:448-459. (2019).
[10] F. Javadpour, Nanopores and Apparent Permeability of Gas Flow in Mudrocks (Shales and Siltstone), Journal of Canadian Petroleum Technology 48 (2009) 16-21. [11] S. Wang, F. Javadpour, Q. Feng, Molecular dynamics simulations of oil transport through inorganic nanopores in shale, Fuel 171 (2016) 74-86. [12] W. Song, B. Yao, J. Yao, Y. Li, H. Sun, Y. Yang, L. Zhang, Methane surface diffusion capacity in carbonbased capillary with application to organic-rich shale gas reservoir, Chemical Engineering Journal 352 (2018) 644654. [13] Z. Li, J. Yao, Z. Ren, H. Sun, L. Zhang, Y. Yang, J. Fan, J. Kou, Accumulation behaviors of methane in the aqueous environment with organic matters, Fuel 236 (2019) 836-842. [14] Z. Li, J. Yao, J. Kou, Mixture Composition Effect on Hydrocarbon–Water Transport in Shale Organic Nanochannels, The journal of physical chemistry letters 10 (2019) 4291-4296. [15] H. JM, J. GW, Oil and organic matter in source rocks of petroleum, AAPG Bull 40(3):477–88 (1956). [16] T.F. Rexer, E.J. Mathia, A.C. Aplin, K.M. Thomas, High-Pressure Methane Adsorption and Characterization of Pores in Posidonia Shales and Isolated Kerogens, Energy Fuels 28 (2014) 2886-2901. [17] D. Robati, B. Mirza, M. Rajabi, O. Moradi, I. Tyagi, S. Agarwal, V. Gupta, Removal of hazardous dyes-BR 12 and methyl orange using graphene oxide as an adsorbent from aqueous phase, Chemical Engineering Journal 284 (2016) 687-697. [18] M. Motamedi, M. Ramezanzadeh, B. Ramezanzadeh, M. Mahdavian, One-pot Synthesis and Construction of a High Performance Metal-Organic Structured Nano Pigment Based on Nanoceria Decorated Cerium (III)Imidazole Network (NC/CIN) for Effective Epoxy Composite Coating Anti-corrosion and Thermo-Mechanical Properties Improvement, Chemical Engineering Journal (2019) 122820. [19] B. Durand, Kerogen, insoluble organic matter from sedimentary rocks, (1980). [20] H. Sui, J. Yao, Effect of surface chemistry for CH 4 /CO 2 adsorption in kerogen: A molecular simulation study, Journal of Natural Gas Science & Engineering 31 (2016) 738-746. [21] C. Bousige, C.M. Ghimbeu, C. Vix-Guterl, A.E. Pomerantz, A. Suleimenova, G. Vaughan, G. Garbarino, M. Feygenson, C. Wildgruber, F.J. Ulm, Realistic molecular model of kerogen's nanostructure, Nature Materials 15 (2016) 576. [22] P. Ungerer, J. Collell, M. Yiannourakou, Molecular Modeling of the Volumetric and Thermodynamic Properties of Kerogen : In fluence of Organic Type and Maturity, Energy & Fuels 29 (2015) 91-105.
[23] J. Kou, J. Yao, H. Lu, B. Zhang, A. Li, Z. Sun, J. Zhang, Y. Fang, F. Wu, J. Fan, Electromanipulating Water Flow in Nanochannels, Nanochannels (2014). [24] S.D. Patwardhan, F. Famoori, R.G. Gunaji, S.K. Govindarajan, Simulation and Mathematical Modeling of Stimulated Shale Gas Reservoirs, Industrial & Engineering Chemistry Research 53 (2014) 19788-19805. [25] J.M. Hunt, G.W. Jamieson, Oil and organic matter in source rocks of petroleum, AAPG Bulletin 40 (1956) 477-488. [26] M.C. Brenneman, P. Smith Jr, The Chemical Relationships Between Crude Oils and Their Source Rocks: Topical Papers, (1958). [27] J.M. Hunt, Distribution of hydrocarbons in sedimentary rocks, Geochimica et Cosmochimica Acta 22 (1961) 37-49. [28] F. P, D. D, Spatial distribution of reservoir fluids in mature kerogen using molecular simulations., Fuel 235:448-459. (2019). [29] K. Falk, J.M. Pellenq, F.J. Ulm, B. Coasne, Effect of Chain Length and Pore Accessibility on Alkane Adsorption in Kerogen, Energy & Fuels 29 (2015) acs.energyfuels.5b02015. [30] H. Wang, X. Wang, X. Jin, D. Cao, Molecular Dynamics Simulation of Diffusion of Shale Oils in Montmorillonite, Journal of Physical Chemistry C 120 (2016) acs.jpcc.6b01660. [31] A. Obliger, R. Pellenq, F.J. Ulm, B. Coasne, Free Volume Theory of Hydrocarbon Mixture Transport in Nanoporous Materials, Journal of Physical Chemistry Letters 7 (2017) 3712-3717. [32] Z. Zhang, H. Liu, J. Wang, Free Energy Changes During Fluid-rock Interactions at Oil-shale Interfaces by Molecular Dynamics Simulation, The Journal of Physical Chemistry C (2019). [33] J. Collell, P. Ungerer, G. Galliero, M. Yiannourakou, F. Montel, M. Pujol, Molecular Simulation of Bulk Organic Matter in Type II Shales in the Middle of the Oil Formation Window, Energy & Fuels 28 (2015) 74577466. [34] S. Tesson, A. Firoozabadi, Methane Adsorption and Self-Diffusion in Shale Kerogen and Slit Nanopores by Molecular Simulations, The Journal of Physical Chemistry C (2018). [35] W. Humphrey, A. Dalke, K. Schulten, VMD: visual molecular dynamics, J Mol Graph 14 (1996) 33-38. [36] S. Wang, F. Javadpour, Q. Feng, Fast mass transport of oil and supercritical carbon dioxide through organic nanopores in shale, Fuel 181 (2016) 741-758.
[37] J. Collell, P. Ungerer, G. Guillaume, Y. Marianna, M. François, P. Magali, Molecular Simulation of Bulk Organic Matter in Type II Shales in the Middle of the Oil Formation Window, Energy & Fuels 28 (2014) 74577466. [38] M. Waldman, A.T. Hagler, New combining rules for rare gas van der Waals parameters, Journal of Computational Chemistry 14 (1993) 1077-1084. [39] D.M. York, T.A. Darden, L.G. Pedersen, The effect of long‐range electrostatic interactions in simulations of macromolecular crystals: A comparison of the Ewald and truncated list methods, Journal of Chemical Physics 99 (1993) 8345-8348. [40] S.G. Srinivasan, I. Ashok, H. Jônsson, G. Kalonji, J. Zahorjan, Parallel short-range molecular dynamics using the Ādhāra runtime system, Computer Physics Communications 102 (1997) 28-43. [41] D.S. Sundaram, P. Puri, V. Yang, Thermochemical Behavior of Nickel-Coated Nanoaluminum Particles, Journal of Physical Chemistry C 117 (2013) 7858–7869. [42] S. Nose, M.L. Klein, A study of solid and liquid carbon tetrafluoride using the constant pressure molecular dynamics technique, Journal of Chemical Physics 78 (1983) 6928-6939. [43] A. Martini, A. Roxin, R.Q. Snurr, Q. Wang, S. Lichter, Molecular mechanisms of liquid slip, Journal of Fluid Mechanics 600 (2008) 257-269. [44] R. Kjellander, H. Greberg, Mechanisms behind concentration profiles illustrated by charge and concentration distributions around ions in double layers, Journal of Electroanalytical Chemistry 450 (1998) 233-251. [45] G. Fiorin, M.L. Klein, J. Hénin, Using collective variables to drive molecular dynamics simulations, Molecular Physics 111 (2013) 3345-3362. [46] G. Fiorin, J. Hénin, New Developments in the Collective Variables Module: More Flexible, More Interactive, Biophysical Journal 108 (2015) 160a-160a. [47] T. Lee, E. Charrault, C. Neto, Interfacial slip on rough, patterned and soft surfaces: A review of experiments and simulations, Advances in Colloid & Interface Science 210 (2014) 21-38. [48] C. Neto, D.R. Evans, E. Bonaccurso, H.J. Butt, V.S.J. Craig, Boundary slip in Newtonian liquids: a review of experimental studies, Reports on Progress in Physics 68 (2005) 2859. [49] Okamoto, Naoki, Slip velocity of methane flow in nanopores with kerogen and quartz surfaces, Spe Journal 23 (2018) 102-116.
[50] C. Freeman, G.J. Moridis, G.E. Michael, T.A. Blasingame, Measurement, modeling, and diagnostics of flowing gas composition changes in shale gas wells, SPE Latin America and Caribbean petroleum engineering conference, Society of Petroleum Engineers, 2012. [51] A. Tinni, C.H. Sondergeld, C.S. Rai, Hydrocarbon storage mechanism in shale reservoirs and impact on hydrocarbon production, Unconventional Resources Technology Conference, Austin, Texas, 24-26 July 2017, Society of Exploration Geophysicists, American Association of Petroleum …, 2017, pp. 3598-3605. [52] S. Wang, Q. Feng, F. Javadpour, Q. Hu, K. Wu, Competitive adsorption of methane and ethane in montmorillonite nanopores of shale at supercritical conditions: A grand canonical Monte Carlo simulation study, Chemical Engineering Journal 355 (2019) 76-90. [53] G. Chen, J. Zhang, S. Lu, M. Pervukhina, D.N. Dewhurst, Adsorption Behavior of Hydrocarbon on Illite, Energy & Fuels 30 (2016).
Notes
The authors declare no competing financial interest.
Highlights
We study the adsorption behaviors of hydrocarbon mixtures in kerogen slit.
The heterogeneous adsorption behaviors are illustrated by the PMF in mechanism of energy.
The effects of temperature and slit aperture are analyzed.
The medium component is the most potential fraction in thermal exploitation.
The aperture close to heavy molecular size facilitate the maximum adsorption density of hydrocarbons.
S1385-8947(20)30045-0 https://doi.org/10.1016/j.cej.2020.124054 CEJ 124054
To appear in:
Chemical Engineering Journal
Received Date: Revised Date: Accepted Date:
23 October 2019 31 December 2019 4 January 2020
Please cite this article as: Y. Yang, J. Liu, J. Yao, J. Kou, Z. Li, T. Wu, K. Zhang, L. Zhang, H. Sun, Adsorption Behaviors of Shale Oil in Kerogen Slit by Molecular Simulation, Chemical Engineering Journal (2020), doi: https:// doi.org/10.1016/j.cej.2020.124054
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2020 Published by Elsevier B.V.
Adsorption Behaviors of Shale Oil in Kerogen Slit by Molecular Simulation
Yongfei Yang a *, Jie Liu a, Jun Yao a, Jianlong Kou b, Zheng Li a, Tianhao Wu c, Kai Zhang a, Lei Zhanga, Hai Suna
a Key
Laboratory of Unconventional Oil & Gas Development (China University of
Petroleum (East China)), Ministry of Education, Qingdao 266580, P.R. China & Research Center of Multiphase Flow in Porous Media, School of Petroleum Engineering, China University of Petroleum (East China), Qingdao, 266580, P.R. China b Institute of Condensed Matter Physics, Zhejiang Normal University, Jinhua 321004 , P.R. China c Reservoir Engineering Research Institute, 595 Lytton Avenue Suite B, Palo Alto, CA 94301, USA *Address all correspondence to Yongfei Yang, E-mail: [email protected] ABSTRACT
Shale oil is widely distributed in organic nanopores, while kerogen plays a complex and key role for adsorption behavior of shale oil, and thus it is crucial to identify the associated storage mechanisms. In this paper, molecular dynamic simulation had been performed to quantify the adsorption behavior of shale oil in kerogen slits. Both the distribution of shale oil properties and potential of the mean force (PMF) were used to identify the interaction mechanisms between the light and heavy components respectively represented by methane and asphaltene. We also examined the effects of different temperatures and apertures on the adsorption behavior. Owning to the extremely strong adsorption capacity between the
asphaltene and kerogen, the adsorbed asphaltene layers reduce the slit width, preventing the light components from adsorbing on the kerogen slits due to the energy barrier formed by heavy components. It is found that, with an increase in temperature, the distribution of hydrocarbons performs more homogeneously. In addition, the adsorption quantity of medium components displays a reduction in kerogen slit, while the heavy component shows a rising as its greater competitive, suggesting that the medium components are the most potential fraction in thermal exploitation, and the light components keep a steady quantity with the combined action of medium and heavy components. The small slit (aperture < 2 nm) can be blocked by asphaltene molecules, and the adsorption density of hydrocarbons reaches the maximum at 2 nm aperture.
Keywords: Adsorption behavior, Shale oil, Kerogen, Molecular simulation
1 Introduction The depletion of the world’s resources is offsetting the declines in conventional oil and gas production, and exploiting the shale oil and gas is changing the global energy system [1]. The shale oil boom has stimulated the oil and gas industry to start the corresponding programs to exploit shale oil and gas [2, 3]; however, the abundant nanopores and extremely low permeability of shale complicate the exploitation and understanding of the adsorption behavior within it [4-7]. In the unconventional shale reservoirs, the pore size is usually in the range of nanometer [8]. By simulating the distribution of hydrocarbons in kerogen nanopores, Perez et al. [9] distinguished the micropores (aperture < 2 nm) and mesopores (2 nm< aperture < 50 nm) in kerogen matrix. The experiments using atomic force microscopy (AFM) confirm the existence of extremely small pores in a range of 2-200 nm [10, 11]; but it is extremely difficult to experimentally characterize the adsorption behavior because such nanopores are below the
optical resolution, and molecular simulation is an effective way to study the molecular behavior in nanopores [12-14]. Shale matrix is a complex porous medium that is composed of quartz minerals and organic matters, among which the latter is mainly the kerogen, i.e., the main source of oil [15]. The hydrocarbon in shale is mostly in the micropore of organic matter generated over geological time [16]. Therefore, it is of practical and fundamental importance to quantify the hydrocarbon adsorption within the kerogen channel in porous media, and the hydrocarbons adsorption in nanopores is also common in effluent treatment [17] and the coating corrosion [18]. Physically, the molecular structure of kerogen is complex, while its physicochemical properties rely on the origin and depositional environments [19, 20]. Previous efforts have been made to build the molecular nanostructure for kerogen by using the diffraction measurements and the generalized phonon densities of states, and the kerogen can be classified into four different types with its maturity indicator [21, 22]. In previous studies, the rigid carbon nanotubes (CNTs) and graphene slits are usually used as the channel models, which make the quantitative analysis of fluid properties more precise [11, 13, 23]. But the extremely smooth wall and pure carbon atoms could overestimate the adsorbing capacity of hydrocarbon mixtures. The adsorbed phases perform different physical and chemical behaviors in such small pore space, due to the affinity for the organic wall [24]. Previous studies found that the source oils contained more aromatic hydrocarbons than reservoir oils, such as methylbenzene and asphaltene [25-27]. Perez et al. [28] determined the adsorption of hydrocarbon mixtures in pore space of kerogen matrix, and the adsorption layer of heavy components reduced the adsorption space of other components, such as methane. The intermediate components (C2 - C12 and methylbenzene) tend to show more adsorption on the pore surface with the increase of carbon chain [29], while the shorter chain alkanes facilitate the exploitation [30]. With the help of molecular dynamics and statistical mechanics to elucidate the diffusion and adsorption of
hydrocarbon in kerogen matrix, Obliged et al. [31] found that the permeance was decreased with an increase in the alkane length. Zhang et al. [32] calculated the interaction energy between kerogen wall and oil molecules using the umbrella sampling method, and they found that the clustered oil molecules inclined to desorb from the kerogen surface compared with single oil molecule. By using equilibrium molecular dynamic (EMD) method, we constructed a slit by two kerogen matrixes and studied the adsorption behavior of shale oil in kerogen slit for the first time, revealing the underlying mechanisms of adsorption and the heterogeneous distribution between shale oil and kerogen by calculating the PMF. Not only did we quantify the density distribution of shale oil within kerogen slit in one direction, but generated the two-dimensional density contour of whole system because of the complex structure of kerogen matrix. Our results show that the heavy component is tightly adsorbed on the surface of kerogen matrix, and the adsorbed layer weakens the interaction force between kerogen and light component, leading to the heterogeneous distribution of mixture density. We also examined the effects of temperature and slit width on the density distribution of shale oil.
2 Methodology 2.1 Molecular models In this work, we used the kerogen monomers and hydrocarbon mixtures to construct a three-dimensional rough kerogen slit as shown in Fig. 1. We selected the type Ⅱ-C kerogen, as it was oil-prone that corresponded to the kerogen in organic-rich shales [22]. The kerogen chemical formula is C242H219O13N5S2 and its molecular structure is also presented in Fig. 1. We set the hydrocarbon mixtures in the outside of slit to obtain the selectivity of kerogen slit for hydrocarbons. The composition of the shale oil was modified and given in Table 1. The hydrocarbon mixture model was simplified and followed the work of Felipe et al. [9], and we
selected the single generic asphaltene model with chemical formula C26OH32 [28, 33]. We ignored some trivial components in that the elaborate components would make the molecular simulation intractable. For a slit of 2 nm aperture, the dimensions of the simulation box are 3.5 × 11.0 × 14.6 nm3. The kerogen matrixes were packed in the both side of the box along y direction, and we constructed the slit models with apertures from 1 nm to 3 nm. We did not simulate the adsorption in slit with larger aperture. Because the larger slit volume weakens the interaction of kerogen, and the hydrocarbon mixtures show more randomness which reduces the accuracy of statistics. Each matrix was constructed with fourteen kerogen monomers and its average density was controlled at 0.98 ± 0.01 g·cm-3 [34]. The models and snapshots were visualized by visual molecular dynamics (VMD) package [35]. The polymer consistent force field plus (PCFF+) was used for all molecules [22, 36, 37]. Periodic boundary conditions were applied in three directions. The Van der Waals interactions were described by using Lennard-Jones (LJ) potentials with Waldman-Hagler combining rules [38], and the cutoff distance was 1.2 nm. We adopted the Ewald method to compute the electrostatic potential [39].
Fig. 1. Kerogen slit was constructed by kerogen matrix and hydrocarbon mixtures, and each component was represented in one color to clearly show the different components. Components color scheme: black, kerogen matrixes; cyan, asphaltene; gray, methylbenzene; purple, n-dodecane; mauve, noctane; pink, n-butane; green, propane; yellow, ethane; blue, methane. Elements color scheme of kerogen and asphaltene molecules: black, C; blue, N; red, O; yellow, S; white, H.
Table 1 Molecular fraction of shale oil components
Species
Mole fraction [%]
Number of molecules
methane ethane propane n-butane n-octane n-dodecane methylbenzene asphaltene
63.5 7.8 4.8 3.2 6.3 4.8 4.8 4.8
1378 169 104 69 137 104 104 103
Molar mass [g·mol-1] 16.535 30.433 44.097 58.124 144.230 170.335 92.140 360.538
2.2 Simulation methods The molecular simulations were performed with the large-scale atomic/molecular massively parallel simulator (LAMMPS) package [40]. We took the MD simulation in the NVT ensemble with a timestep of 0.1 fs on the simulation box for 100 ps which reduced the energy state and got the relaxed configuration of system. Then, we put the simulation system into the NPT ensemble at T = 300 K and stabilized system pressure at 27.5 MPa which consumed 1 ns with a timestep of 1 fs. For the EMD simulations, the NVT ensemble was taken at 300 K during 11 ns, and the last 8 ns of trajectories were collected for analysis. During the simulation, we freezed the kerogen matrix and controlled the temperature of hydrocarbon mixtures by the method of Nosé-Hoover thermostat [41, 42]. The density profiles were calculated by averaging the number of atoms in the bins that parallel to the kerogen wall [43]. The PMF was calculated with colvars package in LAMMPS, which was used to reveal the free energy distribution and the adsorption relation of asphaltene and methane [44, 45], and this method was also called steered and targeted MD due to its spring force on the operated molecule [46]. The result data were collected and analyzed by the umbrella sampling method and the weighted histogram analysis method (WHAM) [44]. Due to the complexity of shale oil, we only selected the heaviest component and lightest component as the operated molecules, i.e., the asphaltene and methane, to compute the PMF between kerogen wall and the operated molecule in the mixed components system. Firstly, we applied a harmonic potential on the operated molecule, and the spring force constant of operated methane and asphaltene molecules were 8.37 kJ·mol-1·Å-2 and 29.29 kJ·mol-1·Å-2 respectively. It took 3.5 ns to reach equilibrium before the umbrella sampling with a timestep of 1 fs. In this case, each window was set as a width of 0.1 nm which took 0.5 ns to reach the equilibrium state, and the whole process included 35 windows that cost 17.5 ns. We fixed the operated molecule on z direction to make sure that they could move on the straight path.
3 Results and discussions 3.1 Density distribution of hydrocarbons in kerogen slit We performed EMD simulations to investigate the adsorption behaviors of hydrocarbon mixtures in kerogen slits. The computed density profiles were depicted in Fig. 2 (T = 300 K, aperture = 2 nm). We run the simulations three times independently to improve the accuracy. The rough surface of kerogen matrix made it difficult to define boundary [47, 48], thus we defined the crosspoint of shale oil and kerogen density profiles as the slit boundary [49]. We made statistics of the number density of hydrocarbon atoms in slit region to show the distribution features. In Fig. 2(a), the whole hydrocarbon mixtures shows the two density peaks in the vicinity of kerogen matrix, which correspond with the simulation results in graphene slits [36]. The methane exhibits a gentle enrichment in the bulk space where the number of kerogen atoms equal to zero. The asphaltene density profile displays two peaks, implying a stronger affinity between asphaltene and kerogen, and the randomness of molecular thermal motion leads to the different values of two peaks. The heterogeneous phenomenon is revealed in the section of PMF analysis. We notice that the kerogen density profiles shake along coordinate axis, as a result of the complex molecular structure of kerogen and porous matrix. The kerogen density profile keeps almost constant over 17.5 Å from the center of slit, then declining to zero, suggesting the roughness surface of kerogen wall has more spatial pores for adsorbate. And the significant characteristic of mixtures density profile is the invaded hydrocarbons on the rough kerogen matrix surface. A few of asphaltene molecules insert in the porous structure of slits surface, in addition, the further padding of smaller molecules reduces the roughness of slit surface. On the base of the preliminary adsorption in the pores of the surface, the hydrocarbon mixtures tend to form adsorbed layers. It is interesting that the peaks of asphaltene and hydrocarbon mixtures
profiles are always corresponding, and they always located in the vicinity of kerogen matrix. To explain this corresponding phenomenon of peaks, as shown in Fig. 2(b), we merge the specific components into light (C1 - C4), medium (C8, C12 and methylbenzene) and heavy (asphaltene) components respectively according to the difference of molecular weight. The light component profile shows a gently single peak, displaying the similar distribution with methane. And the medium component profile presents a higher single peak in the bulk area of slits, revealing that the asphaltene molecules contribute to the double density peaks characteristic of mixture mostly in the slit of 2 nm.
(a)
(b) Fig. 2. Number density profile of kerogen and mixtures in slit region, (a) methane and asphaltene are presented as the lightest and heaviest components, (b) light, medium and heavy components are presented. The shale oil is the sum of whole hydrocarbon mixtures.
Fig. 3. Two-dimensional number density contour of mixtures, to make the profile clearly, the kerogen matrix is the black area.
For a better understanding of the density distribution in such heterogeneous structure of kerogen slit, we calculated the two-dimensional density contour of hydrocarbon mixtures for whole system. As shown in Fig. 3, the high-density adsorbed layers surround the kerogen matrixes, reflecting the strong lipophilicity of kerogen. The density values of mixtures diminish gradually as away from kerogen matrix. According to the previous analysis of different components, the light components tend to distribute in the free volume away from kerogen matrixes. Such heterogeneous density distribution affects the produced fluid compositions in oilfield. The light components in free volume can move with less constraints formed by kerogen, thus the lighter fluid is produced firstly in the oil field. With the depletion of lighter hydrocarbons, the heavier hydrocarbons contribute the production increasingly [9, 50, 51]. In the two-dimensional density contour, the insertion phenomenon of hydrocarbons displays clearly on the porous surface of kerogen matrixes. We also notice that some pores in kerogen matrix trap a few of hydrocarbons. In the slit region, apart from the adsorbed layer on the surface of slit, the high-density values present in the center of slit that is affected by the interaction force of two kerogen walls, and they are mainly composed of medium component as shown in Fig. 2(b). In this case, the enrichment of hydrocarbons clogs the nanochannel, and reduces the connectivity of porous medium.
3.2 Potential of the mean force analysis
(a)
(b) Fig. 4. (a) Potential of the mean force between asphaltene and kerogen (PMFa, the blue oval is the initial asphaltene and the orange oval is the asphaltene barrier), the arrow is the calculation direction. (b) Potential of the mean force between methane and kerogen (PMFm).
In order to get insight into the underlying mechanisms responsible for the heterogeneous distribution between the heavy component and the light component, as shown in Fig. 4, we calculated the PMF of asphaltene and methane on kerogen wall in hydrocarbon mixtures respectively, because the asphaltene and methane exhibited similar distribution properties to heavy and light components in Fig. 2. The overall trend of PMF curves were corresponding to the previous work [32], in addition, we took the effect of adsorbed layer into consideration. The results are shown form the midplane of the slit. In this case, Fig. 4(a) presents a slow decline of PMFa from bulk region to the kerogen wall before the first valley value. When the asphaltene molecule gets closer to the kerogen wall, the PMFa down to 0 kJ·mol-1 at 26 Å, where the asphaltene is most stable due to its lowest energy state, indicating that the asphaltene molecules tend to distribute in the vicinity of kerogen rather than the bulk region. After that, the PMFa keeps a sharp increase because of the strong repulsive interaction in the condition of extremely short range between atoms. It is very interesting that the second valley value occurs in the increasing process of PMFa, which is 28.24 kJ·mol-1 at 31 Å. And this PMFa peak implies the potential energy barrier formed by the adsorption layer. Therefore, the first valley facilitates the adsorption on the adsorption layer, effected by both adsorbed layer and kerogen surface, and the second valley reveals the initial adsorption between asphaltene and kerogen. Between the two valleys, the barrier width is 5 Å, which approaches to the thickness of asphaltene molecule. After 31 Å, the value of PMFa increases to infinity which means the operated asphaltene cannot be closer to kerogen. For the PMFm profiles in Fig. 4(b), it is always zero before 23.5 Å, which illustrates that the methane is more stable in bulk phase. The reduction of PMFa value is 8.37 kJ·mol-1 after the second valley, while the decline of PMFm is only 0.95 kJ·mol-1, thus the asphaltene has more affinity for kerogen. In addition, it also demonstrates that the methane is less affected by adsorption layer as a result of its small size, which causes that some methane molecules can
insert into the kerogen matrix. By comparing PMFa with PMFm, the adsorption layer has more influence distance on asphaltene than methane, and the influence distance includes not only the adsorption thickness, but also the energy barrier which is formed by adsorption layer. The adsorption of asphaltene can reduce the effective width of nanochannels, thereby enriching the methane in bulk phase. 3.3 Effect of temperature on hydrocarbons distribution in kerogen slit
(a)
(b) Fig. 5. Two-dimensional number density contour of mixture with various temperature, (a) T = 360 K, (b) T = 380 K.
To investigate the relationship between temperature and hydrocarbon adsorption, we calculated the two-dimensional density contour and counted the number density of hydrocarbons in the slit region from 280 K to 380 K. We kept the same quantity of hydrocarbon molecules to make the results comparable at different temperature. The Fig. 3 presented the density distribution at 300 K, and we chose representative graphs at 360 K and 380 K as shown in Fig. 5. The density distributions in vertical and horizontal directions are similar at three temperatures, and it is also hard to get accurate representations from the one-dimensional distribution. In the condition of 300 K, a large area of high-density cluster surrounds the kerogen matrixes. From the kerogen to the bulk region, the density contour shows a descending in steps. When the temperature rises to 360 K, the density distribution performs more homogeneously, as a result of the more intense molecular thermal motion. The higher temperatures reduce the area of high density, and we cannot observe the high-density clumps
any more at 380 K except for the adsorbed layer. As shown in the two-dimensional density contour, with the increase of temperature, the hydrocarbons tend to distribute on the middle of slit in vertical direction. The stronger thermal motions of mixture molecules counteract the confined effects of kerogen slit, making the molecules getting out of the slit more easily. Nevertheless, for the center of the slit, the kerogen interaction still plays a major role for the trapped molecules. Hence the hydrocarbons concentrate in the center of slit at high temperature, and fewer of them in the opening of slit.
Fig. 6. The adsorption number of different component atoms in the slit region with various temperature, aperture = 2 nm.
By counting the number of hydrocarbon atoms in slit, we quantified the influence of temperature on adsorption. In contrast with the chain alkane models and the smooth wall, the asphaltene models and rough kerogen wall have more complex topological structures, which introduce more instability into the results. Therefore, the results of multicomponent system always perform larger fluctuations in such complex kerogen slit. As shown in Fig. 6, the adsorption number of whole hydrocarbons performs a linear decline on account of the
increasingly thermal motion. As temperature rises, the adsorption competition of hydrocarbons with higher molecular weight tends to be stronger [52], leading to a rise of heavy component content, thereby the medium components decline, which reveals that the medium component is the major contributor for the thermal recovery and the heavy component can accumulate in nanoporous shale. The adsorption of light components performs a slight drop with various temperatures. Due to the higher kinetic energy at high temperature, the adsorption quantity of light components should be reduced, but the decrease of the sum of medium and heavy components leaves more free space for the light components. With these two opposite effects, the quantity of light hydrocarbons keeps stable. The complementary trend of medium and heavy components, we summed them together and found that it had a similar slope to the adsorption number of shale oil. Therefore, the sum of medium and heavy components reflects the whole developing trend in thermal recovery. 3.4 Effect of slit aperture on hydrocarbons distribution in kerogen slit
(a)
(b) Fig. 7. The two-dimensional number density distribution of mixtures. (a) Aperture = 1 nm. (b) Aperture = 3 nm. (T = 300 K)
By calculating the two-dimensional density contours with different slit apertures, we examined the dependence of hydrocarbon mixture adsorption. To get accurate data, we kept the pressure (27.5 MPa) of shale oil consistent in whole system of different width. As shown in Fig. 7(a), the density profile performs a single peak in 1 nm slit, which corresponds to the previous study, indicating that the mixtures perform the similar distribution with pure alkane components within 1 nm slit [53]. For the complex molecular structure, the asphaltene is hard to enter the 1 nm slit. In the vertical direction, caused by the clogging of heavy components in the slit opening, the density curve shows a concave in the middle of slit, and we can consider that the hydrocarbons block the nanochannel. With the increase of slit width, the phenomenon of clogging becomes less apparent, because the increase of aperture weakens the dominant adsorption effects of kerogen wall. In the 3 nm slit, the profile still depicts the double slight density peaks in kerogen slit
region, and more hydrocarbons enter into the slit and increase the bulk density of slit. Besides, we notice that the two-dimensional density distribution shows more randomness. The Fig. 7(b) shows that, apart from the adsorption layer around the kerogen matrixes, the remaining asphaltene molecules form several high-density clusters in the bulk region out of the slit. Once the hydrocarbons get into the bulk region, the intermolecular interactions of mixtures will play a bigger role for the distribution behaviors. Thus, the asphaltene molecules tend to aggregate as clusters, which attract the lighter components, and then influence the shape of low-density area, such as the convex and concave of density contour line.
(a)
(b) Fig. 8. The adsorption number of component atoms (a) in the slit region and (b) its adsorption per unit of slit aperture (nm) with different apertures, T = 300 K.
To quantify the relation between slit aperture and hydrocarbon density, we counted the number of hydrocarbons in the slit region. The adsorption number of integral hydrocarbons exhibits a continuous rise by widening the slit in Fig. 8(a), which is generally accepted. We are curious about the phenomenon that the quantity of heavy component continues arising and then maintains steady, which is demarcated by 2 nm aperture. The asphaltene molecule has a length of 1.8 nm and a width of 0.5 nm, thus the asphaltene molecules have hard access to the smaller slits (aperture < 2 nm). But asphaltene molecules still can enter the channel at sloping or horizontal angles. With the increase of slit width, the asphaltene can get into the slit more easily. Thus, the heavy component shows a rise with a bigger slope for a slit of above 1.5 nm. The asphaltenes form the clusters with the effect of self-association, but it is uneasy to get into the slit with the form of cluster. In the 2 nm slit, some free asphaltene monomers which are not trapped by clusters can enter the slit with the geometric constraint, driven by the adsorption,
and then form the adsorbed layer on the two kerogen walls. According to PMF analysis, the asphaltene clusters are unstable in the bulk region of slit, and the adsorbed layer weakens the interaction between the kerogen wall and asphaltene molecules in bulk phase of slit. Therefore, the density of asphaltene keeps steady above 2 nm slits, and the asphaltene cluster may get into the slit and cause a rising adsorption when the apertures are wider that the hydrocarbons are unaffected by kerogen. As oppose to the asphaltene, the medium and light components increase at a steeper slope. Because of the increasing bulk volume of slit, more light and medium molecules can get into the slit region. To investigate the adsorptive contribution among three components with increasing aperture, we summed adsorption quantity per unit aperture of the light and medium components in Fig. 8(b). At aperture of 2 nm, both the adsorption of heavy component and whole hydrocarbons depicts a maximum. The wider slit (aperture > 2 nm) weakens the interaction between the kerogen wall and hydrocarbons, leading to the decrease of adsorption per unit slit aperture, and revealing that the maximum adsorption density of hydrocarbon mixtures can be achieved in the aperture of close to heavy molecular size. For the increase of adsorption quantity, the heavy component plays a major role in the aperture of less than 2 nm, although the light component performs a slightly increase at 1.5 nm due to the small molecular structure. The adsorption quantity of light and medium components tends to increase and dominates the increase of adsorption quantity over 2 nm, while the content of heavy component shows a decline.
4 Conclusions In this work, the adsorption behaviors of shale oil in kerogen slit are investigated by using molecular dynamic simulations. We find that the heavy component represented by asphaltene performs a tremendous adsorption capacity on kerogen, while the light component, such as methane, enriches in the bulk region, and the medium component is adsorbed as the
intermediate states. The insertion of hydrocarbons smooths the porous surface of kerogen matrixes relatively, and then the asphaltene molecules facilitate the formation of first density peak on the smoothed surface. Apart from the dense adsorbed layer on the surface of slit, the hydrocarbons present an extra density cluster in the middle of slit, and reduce the connectivity of porous medium. Caused by the strong lipophilicity of kerogen wall and the difference of molecular weight, the shale oil exhibits heterogeneous distribution, which leads to the variation from light oil to heavy oil in production. The interactions between asphaltene molecules and kerogen surfaces take the leading role in the adsorption of hydrocarbons. As temperature rise, the density distribution of shale oil tends to be more homogeneous with less high-density clusters. The adsorption of shale oil in kerogen slit decreases with the increase of temperature, while the adsorption quantity of heavy component shows a rise, as the high temperature makes asphaltene more competitive. The adsorption quantity of medium components displays a downward trend, implying that they are the most potential fraction in thermal exploitation. The positive effects of medium component overweight the impeditive effects of heavy component, which contribute the effectiveness of thermal recovery in the nanochannels. It is hard to enter the small channel (aperture < 2 nm) for an asphaltene molecule caused by its complex structure. But a few of asphaltene molecules still can get into the slit with various angles, and block the opening of slit. Because of the weakened effect of kerogen by widening the slit, the distribution of high-density clusters performs more randomness. In addition, the adsorption quantity of shale oil keeps rising and reach the maximum of adsorption density in 2 nm slit, revealing that the aperture of close to heavy molecular size facilitates the maximum adsorption density of shale oil. The adsorption quantity of heavy component presents a maximum and keeps steady over 2 nm slit, suggesting that the adsorbed layer weakens the adsorption between kerogen wall and asphaltenes which in bulk phase. And the sum of light
and medium components dominates the increase of adsorption quantity in the nanochannel over 2 nm slit. This work provides the comprehensive descriptions of shale oil adsorption behaviors and builds up a good foundation for further studies on the transport of shale oil in kerogen slit. Acknowledgments We acknowledge the following financial support: National Natural Science Foundation of China (No. 51674280, 51722406, 51950410591), Shandong Provincial Natural Science Foundation (ZR2019JQ21, JQ201808), and Program for Changjiang Scholars and Innovative Research Team in University (IRT_16R69). References [1] J.D. Hughes, Energy: A reality check on the shale revolution, Nature 494 (2013) 307-308. [2] P.J.M. Monteiro, C.H. Rycroft, B. Grigory Isaakovich, A mathematical model of fluid and gas flow in nanoporous media, Proc Natl Acad Sci U S A 109 (2012) 20309-20313. [3] L. Maugeri, The shale oil boom: a US phenomenon, Harvard Kennedy School, Belfer Center for Science and International Affairs2013. [4] Y. Yang, J. Yao, C. Wang, Y. Gao, Q. Zhang, S. An, W. Song, New pore space characterization method of shale matrix formation by considering organic and inorganic pores, Journal of Natural Gas Science and Engineering 27 (2015) 496-503. [5] Y. Yang, K. Wang, L. Zhang, H. Sun, K. Zhang, J. Ma, Pore-scale simulation of shale oil flow based on pore network model, Fuel 251 (2019) 683-692. [6] R.A. Kerr, Natural Gas From Shale Bursts Onto the Scene, Science 328 (2010) 1624-1626. [7] Y. Wang, Nanogeochemistry: Nanostructures, emergent properties and their control on geochemical reactions and mass transfers, Chemical Geology 378-379 (2014) 1-23. [8] C.R. Clarkson, N.R. Solano, R.M. Bustin, A.M.M. Bustin, T.P. Blach, Pore structure characterization of North American shale gas reservoirs using USANS/SANS, gas adsorption, and mercury intrusion, Fuel 103 (2013) 606616. [9] P. Felipe, D. Deepak, Spatial distribution of reservoir fluids in mature kerogen using molecular simulations., Fuel 235:448-459. (2019).
[10] F. Javadpour, Nanopores and Apparent Permeability of Gas Flow in Mudrocks (Shales and Siltstone), Journal of Canadian Petroleum Technology 48 (2009) 16-21. [11] S. Wang, F. Javadpour, Q. Feng, Molecular dynamics simulations of oil transport through inorganic nanopores in shale, Fuel 171 (2016) 74-86. [12] W. Song, B. Yao, J. Yao, Y. Li, H. Sun, Y. Yang, L. Zhang, Methane surface diffusion capacity in carbonbased capillary with application to organic-rich shale gas reservoir, Chemical Engineering Journal 352 (2018) 644654. [13] Z. Li, J. Yao, Z. Ren, H. Sun, L. Zhang, Y. Yang, J. Fan, J. Kou, Accumulation behaviors of methane in the aqueous environment with organic matters, Fuel 236 (2019) 836-842. [14] Z. Li, J. Yao, J. Kou, Mixture Composition Effect on Hydrocarbon–Water Transport in Shale Organic Nanochannels, The journal of physical chemistry letters 10 (2019) 4291-4296. [15] H. JM, J. GW, Oil and organic matter in source rocks of petroleum, AAPG Bull 40(3):477–88 (1956). [16] T.F. Rexer, E.J. Mathia, A.C. Aplin, K.M. Thomas, High-Pressure Methane Adsorption and Characterization of Pores in Posidonia Shales and Isolated Kerogens, Energy Fuels 28 (2014) 2886-2901. [17] D. Robati, B. Mirza, M. Rajabi, O. Moradi, I. Tyagi, S. Agarwal, V. Gupta, Removal of hazardous dyes-BR 12 and methyl orange using graphene oxide as an adsorbent from aqueous phase, Chemical Engineering Journal 284 (2016) 687-697. [18] M. Motamedi, M. Ramezanzadeh, B. Ramezanzadeh, M. Mahdavian, One-pot Synthesis and Construction of a High Performance Metal-Organic Structured Nano Pigment Based on Nanoceria Decorated Cerium (III)Imidazole Network (NC/CIN) for Effective Epoxy Composite Coating Anti-corrosion and Thermo-Mechanical Properties Improvement, Chemical Engineering Journal (2019) 122820. [19] B. Durand, Kerogen, insoluble organic matter from sedimentary rocks, (1980). [20] H. Sui, J. Yao, Effect of surface chemistry for CH 4 /CO 2 adsorption in kerogen: A molecular simulation study, Journal of Natural Gas Science & Engineering 31 (2016) 738-746. [21] C. Bousige, C.M. Ghimbeu, C. Vix-Guterl, A.E. Pomerantz, A. Suleimenova, G. Vaughan, G. Garbarino, M. Feygenson, C. Wildgruber, F.J. Ulm, Realistic molecular model of kerogen's nanostructure, Nature Materials 15 (2016) 576. [22] P. Ungerer, J. Collell, M. Yiannourakou, Molecular Modeling of the Volumetric and Thermodynamic Properties of Kerogen : In fluence of Organic Type and Maturity, Energy & Fuels 29 (2015) 91-105.
[23] J. Kou, J. Yao, H. Lu, B. Zhang, A. Li, Z. Sun, J. Zhang, Y. Fang, F. Wu, J. Fan, Electromanipulating Water Flow in Nanochannels, Nanochannels (2014). [24] S.D. Patwardhan, F. Famoori, R.G. Gunaji, S.K. Govindarajan, Simulation and Mathematical Modeling of Stimulated Shale Gas Reservoirs, Industrial & Engineering Chemistry Research 53 (2014) 19788-19805. [25] J.M. Hunt, G.W. Jamieson, Oil and organic matter in source rocks of petroleum, AAPG Bulletin 40 (1956) 477-488. [26] M.C. Brenneman, P. Smith Jr, The Chemical Relationships Between Crude Oils and Their Source Rocks: Topical Papers, (1958). [27] J.M. Hunt, Distribution of hydrocarbons in sedimentary rocks, Geochimica et Cosmochimica Acta 22 (1961) 37-49. [28] F. P, D. D, Spatial distribution of reservoir fluids in mature kerogen using molecular simulations., Fuel 235:448-459. (2019). [29] K. Falk, J.M. Pellenq, F.J. Ulm, B. Coasne, Effect of Chain Length and Pore Accessibility on Alkane Adsorption in Kerogen, Energy & Fuels 29 (2015) acs.energyfuels.5b02015. [30] H. Wang, X. Wang, X. Jin, D. Cao, Molecular Dynamics Simulation of Diffusion of Shale Oils in Montmorillonite, Journal of Physical Chemistry C 120 (2016) acs.jpcc.6b01660. [31] A. Obliger, R. Pellenq, F.J. Ulm, B. Coasne, Free Volume Theory of Hydrocarbon Mixture Transport in Nanoporous Materials, Journal of Physical Chemistry Letters 7 (2017) 3712-3717. [32] Z. Zhang, H. Liu, J. Wang, Free Energy Changes During Fluid-rock Interactions at Oil-shale Interfaces by Molecular Dynamics Simulation, The Journal of Physical Chemistry C (2019). [33] J. Collell, P. Ungerer, G. Galliero, M. Yiannourakou, F. Montel, M. Pujol, Molecular Simulation of Bulk Organic Matter in Type II Shales in the Middle of the Oil Formation Window, Energy & Fuels 28 (2015) 74577466. [34] S. Tesson, A. Firoozabadi, Methane Adsorption and Self-Diffusion in Shale Kerogen and Slit Nanopores by Molecular Simulations, The Journal of Physical Chemistry C (2018). [35] W. Humphrey, A. Dalke, K. Schulten, VMD: visual molecular dynamics, J Mol Graph 14 (1996) 33-38. [36] S. Wang, F. Javadpour, Q. Feng, Fast mass transport of oil and supercritical carbon dioxide through organic nanopores in shale, Fuel 181 (2016) 741-758.
[37] J. Collell, P. Ungerer, G. Guillaume, Y. Marianna, M. François, P. Magali, Molecular Simulation of Bulk Organic Matter in Type II Shales in the Middle of the Oil Formation Window, Energy & Fuels 28 (2014) 74577466. [38] M. Waldman, A.T. Hagler, New combining rules for rare gas van der Waals parameters, Journal of Computational Chemistry 14 (1993) 1077-1084. [39] D.M. York, T.A. Darden, L.G. Pedersen, The effect of long‐range electrostatic interactions in simulations of macromolecular crystals: A comparison of the Ewald and truncated list methods, Journal of Chemical Physics 99 (1993) 8345-8348. [40] S.G. Srinivasan, I. Ashok, H. Jônsson, G. Kalonji, J. Zahorjan, Parallel short-range molecular dynamics using the Ādhāra runtime system, Computer Physics Communications 102 (1997) 28-43. [41] D.S. Sundaram, P. Puri, V. Yang, Thermochemical Behavior of Nickel-Coated Nanoaluminum Particles, Journal of Physical Chemistry C 117 (2013) 7858–7869. [42] S. Nose, M.L. Klein, A study of solid and liquid carbon tetrafluoride using the constant pressure molecular dynamics technique, Journal of Chemical Physics 78 (1983) 6928-6939. [43] A. Martini, A. Roxin, R.Q. Snurr, Q. Wang, S. Lichter, Molecular mechanisms of liquid slip, Journal of Fluid Mechanics 600 (2008) 257-269. [44] R. Kjellander, H. Greberg, Mechanisms behind concentration profiles illustrated by charge and concentration distributions around ions in double layers, Journal of Electroanalytical Chemistry 450 (1998) 233-251. [45] G. Fiorin, M.L. Klein, J. Hénin, Using collective variables to drive molecular dynamics simulations, Molecular Physics 111 (2013) 3345-3362. [46] G. Fiorin, J. Hénin, New Developments in the Collective Variables Module: More Flexible, More Interactive, Biophysical Journal 108 (2015) 160a-160a. [47] T. Lee, E. Charrault, C. Neto, Interfacial slip on rough, patterned and soft surfaces: A review of experiments and simulations, Advances in Colloid & Interface Science 210 (2014) 21-38. [48] C. Neto, D.R. Evans, E. Bonaccurso, H.J. Butt, V.S.J. Craig, Boundary slip in Newtonian liquids: a review of experimental studies, Reports on Progress in Physics 68 (2005) 2859. [49] Okamoto, Naoki, Slip velocity of methane flow in nanopores with kerogen and quartz surfaces, Spe Journal 23 (2018) 102-116.
[50] C. Freeman, G.J. Moridis, G.E. Michael, T.A. Blasingame, Measurement, modeling, and diagnostics of flowing gas composition changes in shale gas wells, SPE Latin America and Caribbean petroleum engineering conference, Society of Petroleum Engineers, 2012. [51] A. Tinni, C.H. Sondergeld, C.S. Rai, Hydrocarbon storage mechanism in shale reservoirs and impact on hydrocarbon production, Unconventional Resources Technology Conference, Austin, Texas, 24-26 July 2017, Society of Exploration Geophysicists, American Association of Petroleum …, 2017, pp. 3598-3605. [52] S. Wang, Q. Feng, F. Javadpour, Q. Hu, K. Wu, Competitive adsorption of methane and ethane in montmorillonite nanopores of shale at supercritical conditions: A grand canonical Monte Carlo simulation study, Chemical Engineering Journal 355 (2019) 76-90. [53] G. Chen, J. Zhang, S. Lu, M. Pervukhina, D.N. Dewhurst, Adsorption Behavior of Hydrocarbon on Illite, Energy & Fuels 30 (2016).
Notes
The authors declare no competing financial interest.
Highlights
We study the adsorption behaviors of hydrocarbon mixtures in kerogen slit.
The heterogeneous adsorption behaviors are illustrated by the PMF in mechanism of energy.
The effects of temperature and slit aperture are analyzed.
The medium component is the most potential fraction in thermal exploitation.
The aperture close to heavy molecular size facilitate the maximum adsorption density of hydrocarbons.