Molecular Mechanism of Selective Displacement and Competitive Adsorption of Associated Gas Components by CO₂ in Nanopores During Miscible Flooding

Abstract

Associated gas recovery and CO₂ displacement mechanisms in nanopores are critical for optimizing tight oil reservoir development. To investigate these processes in the Hesui area of the Ordos Basin, molecular dynamics simulations utilizing the COMPASSIII force field and NVT ensemble were conducted on experimental associated gas mixtures within 5 nm polar quartz and non-polar graphite slit-pores. The study evaluated free diffusion, adsorption energies, and displacement behaviors. Results reveal that pore walls generally suppress gas diffusion compared to free states, although CO₂ exhibits anomalously high diffusion on graphite due to surface slip. Additionally, non-polar graphite demonstrates stronger alkane adsorption than quartz, forming stable, low-energy layered structures. During CO₂ injection, displacement on quartz yields an incomplete layered structure governed by Poiseuille flow. Conversely, graphite facilitates displacement primarily in the pore center, characterized by low-friction slip flow and superior transport efficiency. In conclusion, pore wall polarity and chemical composition fundamentally govern associated gas migration and CO₂ displacement efficiency. These microscopic insights provide theoretical guidance for enhancing resource utilization and CO₂ displacement strategies in tight oil reservoirs.

1. Introduction

Associated gas, produced alongside crude oil during oilfield development, is a valuable resource whose yield and composition significantly affect reservoir evaluation, gathering and transportation process design, comprehensive resource utilization, and carbon emission reduction [1,2]. In the context of global energy transition and carbon neutrality, the efficient development of associated gas has become a key direction for the sustainable development of the petroleum industry [3,4,5]. As the world’s second-largest oil consumer and natural gas importer, China possesses abundant associated gas resources [6,7]. However, its utilization rate remains low, and flaring or inefficient use of associated gas not only wastes resources but also exacerbates greenhouse gas emissions. The Ordos Basin, China’s largest continental sedimentary basin, hosts enormous associated gas resources in tight oil reservoirs. Changqing Oilfield has established the country’s first large-scale tight oil development base. Nevertheless, a lack of in-depth understanding of the microscopic occurrence, diffusion, and displacement mechanisms of associated gas in nanopores continues to limit development efficiency and recovery rates [8,9].

Associated gas mainly comprises methane (CH₄), ethane (C₂H₆), propane (C₃H₈), butane (C₄H₁₀), pentane (C₅H₁₂), nitrogen (N₂), and carbon dioxide (CO₂). Its adsorption, diffusion, and CO₂ displacement behaviors vary markedly across different mineral surfaces (e.g., quartz representing inorganic minerals and graphite representing organic matter) [10,11]. Polar quartz surfaces exhibit strong affinity for CO₂, while non-polar graphite surfaces preferentially adsorb alkanes. These surface properties directly regulate the occurrence state and migration efficiency of associated gas [12,13,14]. CO₂ displacement, which simultaneously enhances oil recovery and enables geological storage, holds great potential for associated gas reservoirs [15,16]. However, competitive adsorption and residual effects induced by wall polarity limit displacement efficiency.

Molecular dynamics (MD) simulation has emerged as a powerful tool for investigating fluid microscopic behavior in nanopores, enabling precise capture of parameters such as adsorption layers, density distributions, mean squared displacement (MSD), and diffusion coefficients [17,18,19,20,21,22]. While recent studies have employed MD to examine methane adsorption and diffusion in graphite, kerogen, and quartz nanopores, research on the differential mechanisms of multi-component associated gas across realistic mineral surfaces (quartz vs. graphite) and the competitive adsorption and flow characteristics during CO₂ displacement remains limited. The associated gas from the Hesui area of the Ordos Basin exhibits typical continental tight oil characteristics, with average molar fractions of CH₄ 54.734%, C₂H₆ 13.222%, C₃H₈ 17.052%, C₄H₁₀ 7.381%, C₅H₁₂ 2.932%, N₂ 2.164%, and CO₂ 0.976%. Building on this composition, the present study constructs mixed associated gas models and 5 nm slit-pore models (with quartz and graphite walls) and conducts MD simulations using the COMPASSIII force field under the NVT ensemble to systematically investigate free diffusion, wall adsorption energy distribution, adsorption energies, diffusion coefficients, and CO₂ displacement processes.

The objective of this study is to elucidate the regulatory mechanisms of wall polarity on the microscopic occurrence and migration of associated gas, clarify the competitive adsorption and flow characteristics during CO₂ displacement, and provide theoretical support for efficient associated gas development and CO₂ geological storage in tight oil reservoirs of the Ordos Basin. The findings are expected to guide comprehensive associated gas utilization in Changqing Oilfield and contribute to the optimization of CO₂ displacement processes and the achievement of carbon reduction goals in China’s tight oil reservoirs.

2. Methodology

2.1. Experimental Measurement of Associated Gas Composition

Associated gas, produced alongside crude oil during oilfield development, is the core research object of this work, and its accurate compositional data is the fundamental basis for the construction of the subsequent molecular dynamics simulation system [23,24]. The composition of associated gas samples from the Hesui area of the Ordos Basin was measured following the mature oilfield associated gas separation and gas chromatography analysis method described in previous studies [25,26,27]. The final molar fractions of each component (CH₄, C₂H₆, C₃H₈, C₄H₁₀, C₅H₁₂, CO₂, and N₂) are listed in

Table 1. Compositional analysis of associated gas from different blocks in the Hesui area (molar fractions, %).

Block	Horizon	CH4	C2H6	C3H8	C4H10	C5H12	N2	CO2
Hesui shallow	Chang 3 and above	44.66	13.35	23.93	10.71	2.86	2.93	0.38
Hesui extension	Chang 6	45.00	11.60	21.30	10.70	8.30	3.00	0.25
Hesui extension	Chang 8	62.40	14.20	11.00	3.70	0.00	1.30	3.55
Hesui Chang 6	Chang 6	47.41	15.36	19.23	9.895	3.411	1.49	0.25
Hesui Chang 7	Chang 7	74.20	11.60	9.80	1.90	0.10	2.10	0.45
Average		54.73	13.22	17.05	7.38	2.93	2.16	0.98

, which provides the core input parameters for all subsequent molecular modeling work.

2.2. Molecular Models and Simulation Methods

Based on the associated gas compositional analysis presented in Section 2.1 (Table 1), molecular simulation systems were constructed, including methane (CH₄), ethane (C₂H₆), propane (C₃H₈), butane (C₄H₁₀), pentane (C₅H₁₂), carbon dioxide (CO₂), nitrogen (N₂), and quartz (Quartz) and graphite (Graphite) wall structures. The molecular structures are illustrated in Figure 1. To investigate the occurrence state of associated gas in slit pores, 5 nm slit-pore models were built as shown in Figure 2.

For graphite and quartz crystals, the Cleave Surface tool was used to cut along the (0 0 1) plane with a thickness of 15 Å. Supercells were then created using the Supercell function, resulting in lattice parameters of 44.22 Å × 42.55 Å × 15 Å. The associated gas mixture was constructed in an amorphous cell using the Amorphous Cell tool. Subsequently, the Build Layer tool was employed to assemble the graphite or quartz walls with the gas box into composite systems.

After construction, geometric optimization was performed using the Geometry Optimization task in the Forcite module. Molecular dynamics (MD) simulations were then conducted using the COMPASSIII force field in the NVT ensemble, with medium precision, Ewald summation for electrostatic interactions, and atom-based summation for van der Waals interactions. The total simulation time was 500 ps with a time step of 1 fs [18,28].

Post-simulation analysis utilized the Forcite Analysis tools, extracting equilibrium data from the final 200 ps. Density distributions were analyzed with the Density Distribution tool, and mean squared displacement (MSD) curves and diffusion coefficients were calculated using the MSD tool.

To study the flow behavior of associated gas in nanopores, equilibrium molecular dynamics was first applied to obtain the adsorbed equilibrium microstructure. Subsequently, a Perl script in Materials Studio was used to apply driving forces, enabling non-equilibrium molecular dynamics simulations. A constant force of 100 kcal/mol/Å was applied to each CO₂ molecule in the x-direction to induce gas flow. Details are discussed in Section 3.3.

2.3. Calculation Methods

2.3.1. Density Distribution

To determine the spatial distribution of oil and gas molecules, the shale pore space was divided along the z-direction into equal-volume bins, assuming macroscopic properties are centered in each bin. The position function is defined as [29]:

$$\left\{\begin{array}{c}{H}_n\left(z_{i,j}\right)=1 \left(n-1\right)∆z<z_i<n∆z\\ H_n\left(z_{i,j}\right)=0 \mathrm{otherwise}\end{array}\right.$$

(1)

The number density and mass density of oil/gas molecules in the nth bin from timestep $J_a$ to $J_b$ are then:

$${\rho }_{number}={\displaystyle \frac{1}{A∆z\left(J_a-J_b+1\right)}}{\displaystyle \sum _{J_M}^{{J=J}_N}}{\displaystyle \sum _N^{i=1}}{H}_N\left(z_{i,j}\right)$$

(2)

$${\rho }_{mass}={\displaystyle \frac{{10}^{21}}{N_A}}{\displaystyle \frac{1}{A∆z\left(J_a-J_b+1\right)}}{\displaystyle \sum _{J_M}^{{J=J}_N}}{\displaystyle \sum _N^{i=1}}{H}_N\left(z_{i,j}\right)\ast M_i$$

(3)

where $z$ is the direction normal to flow; J is the timestep; $z_i$ is the coordinate of the ith molecule in the nth bin; A is the XY-plane area; $N_A$ is Avogadro’s constant; and $M_i$ is the molecular weight.

2.3.2. Mean Square Displacement and Diffusion Coefficient

Mean square displacement (MSD) represents the average squared change in particle position from initial to subsequent times, reflecting temporal offsets and inversely correlating with shale adsorption strength. Higher diffusion coefficients indicate greater molecular freedom and weaker adsorption. The diffusion coefficient, which quantifies particle mobility and positional variation over time, is derived from MSD. The self-diffusion coefficient for interlayer sodium ions and water molecules in three dimensions is given by Einstein’s equation [30,31]:

$$D={\displaystyle \frac{1}{6N_{\alpha }}}\underset{t\to \infty }{\lim }{\displaystyle \frac{d}{d_t}}{\displaystyle \sum _{i=1}^{N_{\alpha }}}{[r_i\left(t\right)-r_i\left(0\right)]}^2$$

(4)

where ${[r_i\left(t\right)-r_i\left(0\right)]}^2$ is the MSD of the molecular center of mass, and $N_{\alpha }$ is the total number of α-type atoms. Thus, D is one-sixth the slope of the MSD curve.

2.3.3. Adsorption Energy Calculation

Adsorption behavior is primarily governed by gas–surface interactions, with gas–gas interactions being negligible in the present context. Consequently, the focus is placed on gas–surface effects. The total system energy is dominated by van der Waals forces. The interaction energy ($E_{int}$) is negative, indicating spontaneous adsorption and a reduction in overall system energy; more negative values correspond to stronger adsorption affinity. It has been reported that adsorption energy becomes more negative with increasing pressure [21,32].

The interaction energy is calculated as follows:

$$E_{int}=E_{A-B}-\left(E_A-E_B\right)$$

(5)

where $E_{A-B}$ is the total energy of the A − B complex; $E_A$ is the energy of isolated A; and $E_B$ is the energy of isolated B.

To obtain more accurate diffusion coefficients, this study conducted four independent repetitions for each data set. The average diffusion coefficient for each set was calculated, and the corresponding standard deviation was reported.

2.3.4. Velocity Distribution

Molecular dynamics simulations yield atomic positions, velocities, and forces directly. To obtain fluid velocity distribution in pores, statistical mechanics processes simulation results. Pores are divided into bins parallel to solid walls, with alkane velocity in each bin calculated as [33,34]:

$$\upsilon ={\displaystyle \frac{{\displaystyle {\sum }_{\begin{array}{c}i=1\end{array}}^N} m_i{v}_{i,x}}{{\displaystyle {\sum }_{\begin{array}{c}i=1\end{array}}^N} m_i}}$$

(6)

where $\upsilon$ is velocity (m/s); $i$ indexes atoms in the bin; $N$ is the total atoms in the bin; $m_i$ is mass of atom $i$ (g); and $v_{i,x}$ is the $x$-component of velocity for atom $i$ (m/s).

3. Results and Discussion

3.1. Free Diffusion Characteristics of Associated Gas Components

Based on the average composition of associated gas from the Hesui area obtained in Section 2.1, a molecular model of the mixed associated gas was constructed using the Amorphous Cell tool in Materials Studio 2023 software, generating a simulation box (see Figure 3). Molecular dynamics simulations were then performed at 323 K and 20 MPa to investigate the diffusion behavior of the associated gas. The mean squared displacement (MSD) curves for each component were calculated during the simulation, and diffusion coefficients were obtained by fitting the curves according to the Einstein relation. The MSD curves exhibited excellent linearity with time for all components, confirming that the system had reached a stable diffusive state and that the derived diffusion coefficients are reliable.

The diffusion coefficients (in Ų/ps) were as follows: CH₄ 4.80, C₂H₆ 3.67, C₃H₈ 2.34, C₄H₁₀ 4.03, C₅H₁₂ 5.07, N₂ 5.58, CO₂ 4.50. The overall order was N₂ > C₅H₁₂ > CH₄ > CO₂ > C₄H₁₀ > C₂H₆ > C₃H₈, indicating that N₂ exhibits the strongest diffusion ability while C₃H₈ exhibits the weakest.

To validate the reliability of the molecular dynamics simulation results, this study compared the simulated diffusion coefficients of methane under various conditions with experimental data reported in the literature. According to Chen et al. [35], the experimentally measured diffusion coefficient of methane was approximately 1.90 × 10⁻⁸ m²/s, while the value obtained from the present simulations was 4.80 × 10⁻⁸ m²/s. The two values are of the same order of magnitude and numerically close, indicating that the model constructed in this study is reasonable and that the simulation results exhibit high fidelity and reliability.

These differences arise primarily from the combined effects of molecular mass, size, shape, and intermolecular interactions. N₂ and CH₄ have simple structures, low mass, and weak polarity, resulting in minimal interaction with surrounding molecules and walls and thus lower resistance to motion. As the carbon chain length of alkanes increases, molecular volume grows and van der Waals interactions strengthen, restricting molecular mobility and leading to a general decline in diffusion coefficients. Although CO₂ has a higher molecular weight, its linear structure provides greater conformational flexibility, enabling superior diffusion compared to C₃H₈ at similar molecular weight. In low-pressure mixed systems, factors such as collision frequency, transient spatial arrangement, and surface diffusion effects contribute to the non-monotonic trend observed for C₄H₁₀ and C₅H₁₂ [36]. Furthermore, it is worth noting that the current diffusion analysis assumes an ideal mixture. In a complex seven-component system, the cross effects (the mutual influence between diffusing components) are significant and are likely a major contributor to this non-monotonic behavior. Due to the tremendous computational and theoretical complexity required to decouple multi-component cross-diffusion matrices, this specific phenomenon will be systematically investigated as an independent topic in our future work. Under ambient conditions, small, low-polarity components in the Hesui associated gas display higher diffusion ability, while complex molecules with stronger interactions diffuse more slowly. These findings provide important theoretical insight into the microscopic migration mechanisms of associated gas in tight reservoirs of the Hesui area, Ordos Basin, and support subsequent separation and utilization processes.

3.2. Microscopic Behavior of Associated Gas in Nanopore Slits

3.2.1. Occurrence and Energy Characteristics in Nanopore Slits

To investigate the occurrence state and characteristics of associated gas in quartz and graphite slit pores, molecular dynamics simulations were conducted at 323 K and 20 MPa. The results are shown in Figure 4.

In both quartz and graphite slit pores, the associated gas consists of adsorbed and free gas phases. Within the adsorbed phase, methane exhibits the strongest affinity for the wall surface, displaying the highest density peaks and most pronounced adsorption, followed by propane, ethane, butane, and pentane, with CO₂ and N₂ showing the weakest adsorption. Notably, alkane adsorption is stronger on graphite slit pores than on quartz slit pores. Detailed analysis of adsorption energies is presented below.

As shown in Figure 5 and

Table 2. Energy distribution of associated gas in different slit-pore walls (average values, kcal/mol).

Energy Type	Quartz Slit Pore	Graphite Slit Pore
Potential energy	1615	1149
Kinetic energy	4908	4578
Non-bond energy	−872	−1157
Total energy	6523	5727

, under identical temperature and pressure conditions, the total system energy in the graphite slit pore (5727 kcal/mol) is significantly lower than in the quartz slit pore (6523 kcal/mol), with a difference of approximately 796 kcal/mol. This disparity is primarily attributed to the substantial reduction in non-bonded energy: −1157 kcal/mol in the graphite system versus −872 kcal/mol in the quartz system, indicating stronger adsorption interactions on graphite and greater energy release from van der Waals and electrostatic forces. The potential energy is also lower in the graphite system (1149 kcal/mol) than in the quartz system (1615 kcal/mol), reflecting more stable molecular conformations and lower overall system potential. The kinetic energy is similarly reduced (4578 kcal/mol vs. 4908 kcal/mol), suggesting greater restriction of molecular thermal motion.

These energy differences stem fundamentally from the distinct chemical properties of the wall surfaces. Graphite, as a non-polar hydrophobic surface rich in π-electron clouds, forms strong dispersion and induction forces with alkane molecules (e.g., CH₄, C₂H₆) and the quadrupole moment of CO₂, promoting parallel adsorption and stable layered structures, thereby enhancing adsorption affinity and reducing non-bonded and total energy [37,38,39]. In contrast, quartz (SiO₂) surfaces, rich in polar Si–O bonds and hydroxyl groups, exhibit hydrophilic characteristics and weaker interactions with associated gas molecules, relying mainly on electrostatic forces and hydrogen bonding, resulting in looser adsorption and higher non-bonded and potential energies [40]. Non-hydrocarbon components (N₂ and CO₂) also show stronger affinity for graphite, further amplifying the energy difference. These results demonstrate that wall chemical composition and polarity are the dominant factors controlling the energetic state of associated gas occurrence in nanopore slits. Graphite pores favor stable, low-energy occurrence, while quartz pores exhibit relatively weak adsorption.

Furthermore, adsorption energies for individual alkane components (C₁–C₅) on graphite and quartz slit pores were calculated (Figure 6, units: kcal/mol). The absolute values of adsorption energies on graphite are significantly higher than on quartz for all alkanes. For example, methane adsorption energy is −446.118 kcal/mol on graphite versus −391.609 kcal/mol on quartz (an enhancement of approximately 14%); enhancements for ethane, propane, butane, and pentane are about 12%, 12%, 13%, and 8%, respectively. This trend indicates stronger overall adsorption affinity of graphite for associated gas alkanes, with the advantage more pronounced for shorter carbon chains. Adsorption energy decreases in magnitude with increasing chain length, primarily due to reduced effective adsorption sites as molecular volume grows; however, graphite maintains a consistent advantage across components.

These differences arise from the fundamental contrast in wall chemistry: graphite’s non-polar, π-electron-rich hydrophobic surface enables strong π–π dispersion and induction interactions with alkane C–H bonds, promoting parallel adsorption and stable layered structures with substantial energy release [39]. Quartz’s polar, hydrophilic surface interacts more weakly through electrostatic and induction forces, resulting in uneven adsorption sites and lower affinity [41,42]. These adsorption energy results align closely with the preceding energy distribution and diffusion behaviors: stronger adsorption on graphite stabilizes gas molecules, restricts thermal motion, and significantly suppresses diffusion coefficients (except for CO₂), while weaker adsorption on quartz corresponds to higher diffusion ability. These findings confirm the critical role of wall polarity and chemical composition in differentially regulating adsorption–diffusion behavior in nanopores and provide quantitative insight into the impact of mineral pore types on gas occurrence and migration in tight reservoirs. They also offer valuable guidance for pore-type evaluation and gas flow simulation in oil and gas field development.

3.2.2. Effect of Wall Surfaces on Diffusion of Associated Gas

To investigate the influence of wall properties on the diffusion behavior of associated gas, molecular dynamics simulations were performed under three conditions: free diffusion (no walls), graphite slit pores, and quartz slit pores. The mean squared displacement (MSD) curves and corresponding diffusion coefficients (in Ų/ps) for each component were calculated. The results are presented in Figure 7.

The simulations reveal that the presence of walls significantly suppresses the diffusion of associated gas. Compared to free diffusion, diffusion coefficients of N₂ and light hydrocarbons (average for alkanes) decreased markedly in both slit-pore systems. The suppression was stronger in graphite slit pores (N₂ reduced by ~52%, light hydrocarbons by ~14%) than in quartz slit pores (N₂ by ~31%, light hydrocarbons by ~27%). This effect is primarily attributed to increased molecular migration resistance due to wall adsorption. The non-polar graphite surface promotes strong π–π dispersion and induction interactions, facilitating parallel adsorption of associated gas molecules (especially N₂ and alkanes) and forming dense layered structures that substantially restrict movement along the pore axis. In contrast, although quartz exhibits hydrophilic characteristics, its interactions with non-polar components of associated gas are relatively weaker, resulting in looser adsorption layers and less pronounced diffusion suppression.

Notably, CO₂ diffusion in graphite slit pores (9.45 Ų/ps) is substantially higher than in free diffusion (4.50 Ų/ps), whereas it drops sharply to 0.87 Ų/ps in quartz slit pores. This anomalous behavior arises from CO₂’s linear molecular structure and quadrupole moment. On non-polar graphite, CO₂ readily aligns parallel to the surface and undergoes rapid surface slip, resulting in enhanced diffusion dominated by surface mechanisms [43,44,45]. On polar quartz, however, CO₂ forms strong electrostatic interactions and hydrogen bonds with surface hydroxyl groups, leading to firm adsorption and severely restricted mobility [42,46].

These results demonstrate that wall chemical composition and polarity exert component-specific effects on diffusion in nanopore slits: graphite walls strongly inhibit diffusion for most components (except CO₂), while quartz walls exert a more pronounced inhibitory effect on CO₂.

3.2.3. Effect of Temperature on the Diffusion of Associated Gas

This study employed molecular dynamics simulations to systematically investigate the effect of temperature on the diffusion behavior of associated gas (N₂, CO₂, and light hydrocarbons) within shale nano-quartz slits. As shown in Figure 8 and Figure 9, the average diffusion coefficients of all three gas molecules exhibited a significant positive correlation with increasing temperature from 298 K to 398 K.

Specifically, non-polar N₂ exhibited the highest diffusivity and the strongest temperature response, with its diffusion coefficient increasing sharply from 3.37 Ų/ps at 298 K to 7.68 Ų/ps at 398 K. In contrast, CO₂ displayed a relatively low overall diffusion coefficient due to strong adsorption on the pore walls; nevertheless, its diffusion coefficient still increased significantly from 0.80 Ų/ps to 2.27 Ų/ps with rising temperature, representing nearly a threefold enhancement. The diffusivity of light hydrocarbons fell between the two, rising gradually from 2.28 Ų/ps to 4.88 Ų/ps as temperature increased.

From a microscopic perspective, the positive temperature dependence arises because higher temperatures impart greater thermal kinetic energy to the gas molecules. In the nano-confined space, the intensified molecular thermal motion enables the gas molecules to more effectively overcome the van der Waals attractions and adsorption energy barriers of the solid walls. This results in reduced residence time of molecules on the pore surfaces, increased conversion from the adsorbed state to the free state, and a higher frequency of molecular jumps and shuttling within the pore channels, ultimately manifesting as a significant increase in the macroscopic diffusion coefficient.

3.2.4. Effect of Pressure on the Diffusion of Associated Gas

This study utilized molecular dynamics simulations to systematically examine the influence of pressure on the diffusion behavior of associated gas (N₂, CO₂, and light hydrocarbons) in shale nano-quartz slits. As illustrated in Figure 10 and Figure 11, the diffusion coefficients of the associated gas system showed a significant negative correlation as the system pressure increased from 0.101 MPa to 50 MPa.

As the pressure increased from 0.101 MPa to 50 MPa, the average diffusion coefficients of N₂, CO₂, and light hydrocarbons all exhibited a monotonic decreasing trend. Notably, N₂ was most sensitive to pressure in the low-pressure regime (<6 MPa), with its diffusion coefficient decreasing sharply from 3.76 Ų/ps to 1.05 Ų/ps. The diffusion coefficient of light hydrocarbons showed a relatively gradual decline, decreasing from 3.13 Ų/ps to 1.39 Ų/ps. In comparison, CO₂ maintained the lowest diffusion coefficient overall, remaining relatively stable (approximately 0.75 Ų/ps) in the medium-pressure range (6–15 MPa), before slowly decreasing to 0.47 Ų/ps at higher pressures.

From a microscopic perspective, the increase in pressure significantly raises the fluid density within the confined space and markedly reduces the mean free path of the molecules. The resulting increase in intermolecular collision frequency and the compression of free volume strongly restrict molecular migration and jumping, which macroscopically manifests as a reduction in the diffusion coefficient. This pressure dependence indicates that during the later stages of reservoir depressurization, the microscopic diffusion and transport capacity of associated gas will be effectively enhanced. These findings hold important theoretical significance for accurately evaluating the recovery efficiency and ultimate recovery factor of unconventional oil and gas reservoirs under depletion development.

3.2.5. Effect of Nanopore Size on the Diffusion of Associated Gas

This study systematically investigated the nanoconfinement effect of different pore sizes (1–9 nm) on the diffusion behavior of associated gas (N₂, CO₂, and light hydrocarbons) in shale. As shown in Figure 12 and Figure 13, the average diffusion coefficients of all three gas molecules exhibited a significant monotonic decreasing trend with increasing pore diameter.

Specifically, the diffusion coefficient of N₂ reached 5.23 Ų/ps in the extremely confined 1 nm slit and decreased substantially to 1.74 Ų/ps when the pore size increased to 9 nm. Light hydrocarbons exhibited a similar but milder decline, decreasing from 4.07 Ų/ps to 2.18 Ų/ps. In contrast, CO₂ consistently showed the lowest diffusion coefficient among the three, decreasing slowly from 1.49 Ų/ps at 1 nm to 0.73 Ų/ps at 9 nm.

From a microscopic dynamics perspective, this phenomenon is primarily attributed to the transition in molecular flow regime and the change in wall slip effects within the confined space. In the 1 nm extremely confined pore, gas molecules experience strong overlapping potential fields from both walls, tending to form highly ordered layered arrangements. Intermolecular collisions are minimal, and transport is dominated by specular collisions with the walls. When the wall surface is relatively smooth (e.g., carbon-based organic matter), strong surface slip effects can induce near “frictionless” rapid transport. However, as the pore size increases to 9 nm, the free gas volume in the pore center expands significantly, causing molecular motion to shift from wall-dominated to fluid interior-dominated. The sharp increase in molecule–molecule collisions leads to enhanced internal friction and greater kinetic energy dissipation, which substantially weakens the slip enhancement effect at the nanoscale and results in a macroscopic decrease in the diffusion coefficient.

Additionally, the consistently lowest diffusion coefficient of CO₂ further confirms its strong adsorption affinity with the pore walls. These nanoscale confinement transport characteristics indicate that in extremely small shale pores, N₂ and light hydrocarbons can maintain high transport efficiency through slip effects. This finding is of great theoretical importance for accurately assessing microscopic seepage capacity in unconventional reservoirs and optimizing gas displacement strategies.

3.2.6. Effect of Pore Water on the Diffusion of Associated Gas

This study systematically evaluated the nanoconfinement effect of varying pore wall hydration states (from anhydrous to three-layer water film) on the diffusion behavior of associated gas (N₂, CO₂, and light hydrocarbons) in shale nano-slits. As shown in Figure 14 and Figure 15, non-polar gases (N₂ and light hydrocarbons) and polar gas (CO₂) exhibited distinctly opposite diffusion trends with increasing water content.

Specifically, the average diffusion coefficients of N₂ and light hydrocarbons decreased significantly with increasing water film thickness. The diffusion coefficient of N₂ dropped from 3.90 Ų/ps under anhydrous conditions to 1.90 Ų/ps under three water layers, a reduction of over 50%. Light hydrocarbons showed a more moderate decline, decreasing from 3.18 Ų/ps to 2.23 Ų/ps. From a microscopic perspective, the dense water molecule network not only occupies the effective free volume within the pores and narrows the gas transport pathways but also generates strong repulsive forces and steric hindrance against non-polar gas molecules due to its highly polar surface. This leads to a substantial reduction in gas-phase permeability, known as the typical microscopic “water blocking effect.”

In contrast, the diffusion coefficient of CO₂ exhibited an anomalous slight increasing trend with increasing water film thickness (from 0.88 Ų/ps to 1.07 Ų/ps). This reversal is primarily attributed to the preferential adsorption of water molecules on the inorganic mineral surface. Water molecules occupy high-energy adsorption sites on the pore walls through strong hydrogen bonding, thereby effectively shielding the strong van der Waals interaction between the solid wall and CO₂. After being “desorbed” from the wall by the water film, CO₂ experiences a smoother slip boundary at the water–gas interface, resulting in an increased effective diffusion coefficient.

These results suggest that under water-bearing formation conditions, CO₂ injection can effectively mitigate the water blocking effect through competitive adsorption mechanisms, demonstrating superior sweep efficiency and transport potential compared to N₂. This finding provides important microscopic theoretical support for the application of CO₂ displacement technology in unconventional oil and gas reservoirs.

3.3. Mechanisms of CO₂ Displacement of Associated Gas

3.3.1. Competitive Adsorption Mechanisms Between CO₂ and Associated Gas

To elucidate the displacement of associated gas from slit pores by CO₂, density distributions before and after CO₂ injection were compared in quartz and graphite slit-pore systems. The results indicate that (Figure 16), in both wall types, CO₂ injection leads to the formation of distinct adsorption layers near the walls for both CO₂ and associated gas, though the adsorption behavior and displacement efficiency differ significantly.

In quartz slit pores, CO₂ exhibits a much stronger affinity for the wall than associated gas, preferentially occupying adsorption sites and forming a dense CO₂ layer. Associated gas molecules (primarily alkanes) are partially displaced to the pore center, but a substantial fraction remains near the wall, resulting in a composite “associated gas outer layer + CO₂ inner layer” structure. This reflects the significant competitive advantage of CO₂ on quartz due to its polar Si–O bonds and hydroxyl groups, which enable strong electrostatic induction and hydrogen bonding with CO₂’s quadrupole moment. In contrast, non-polar alkanes in associated gas interact more weakly with quartz, making them more easily displaced from the wall; however, residual associated gas remains adsorbed in the outer layer, limiting complete displacement.

In graphite slit pores, although CO₂ forms a noticeable adsorption layer and exhibits high diffusion, the non-polar π-electron-rich surface of graphite interacts far more strongly with alkanes than with CO₂ (despite CO₂’s quadrupole-induced adsorption). Consequently, associated gas alkanes remain firmly adsorbed, and CO₂ struggles to compete for wall sites. Displacement primarily occurs in the pore center, with more associated gas retained near the wall.

Overall, polar quartz walls confer a competitive adsorption advantage to CO₂ but result in incomplete displacement due to residual associated gas. Non-polar graphite walls favor strong associated gas retention, rendering CO₂ less effective at displacing wall-adsorbed molecules [47,48,49]. These findings highlight the critical regulatory role of wall polarity in CO₂ displacement processes: polar walls (e.g., quartz) promote CO₂ adsorption but limit displacement efficiency, while non-polar walls (e.g., graphite) enhance associated gas stability but weaken CO₂ competition. The results provide important microscopic mechanistic insight for CO₂ flooding and gas displacement design in tight reservoirs, particularly regarding pore mineral type selection.

Adsorption energies for associated gas and CO₂ on different walls were further calculated (Figure 17). Graphite exhibits a much higher absolute adsorption energy for associated gas (−808.583 kcal/mol) than quartz (−323.253 kcal/mol), reflecting stronger affinity driven by π–π dispersion and induction interactions with alkane C–H bonds, which promote stable parallel adsorption layers. For pure CO₂, adsorption energies are similar (quartz −352.775 kcal/mol vs. graphite −347.397 kcal/mol), with quartz slightly stronger due to better electrostatic and hydrogen-bonding interactions.

In mixed CO₂ + associated gas systems, quartz shows a higher absolute total adsorption energy (−526.395 kcal/mol) than graphite (−430.727 kcal/mol). This indicates that quartz preferentially adsorbs CO₂ while partially retaining associated gas, forming a stable composite layer. On graphite, the total adsorption energy is dominated by associated gas, with CO₂ contributing less, confirming its weaker competitive ability.

Further analysis reveals that on quartz, the mixed-system adsorption energy (−526.395 kcal/mol) exceeds that of pure CO₂ (−352.775 kcal/mol) or pure associated gas (−323.253 kcal/mol) but is less than their sum (~−676 kcal/mol), indicating competitive rather than fully synergistic effects. CO₂ occupies inner sites, displacing alkanes to the outer layer. On graphite, the mixed-system value (−430.727 kcal/mol) is much lower than pure associated gas (−808.583 kcal/mol) but only slightly higher than pure CO₂ (−347.397 kcal/mol), demonstrating that CO₂ significantly weakens associated gas adsorption without fully replacing it.

In summary, wall polarity differentially regulates CO₂ displacement of associated gas: strong polarity (quartz) favors CO₂ competition but leaves residual associated gas; strong non-polarity (graphite) enhances associated gas retention, limiting CO₂ effectiveness. These findings offer quantitative mechanistic guidance for CO₂ injection optimization and pore-type evaluation in tight reservoir development [50].

3.3.2. Flow Characteristics of CO₂ Displacement of Associated Gas in Nanopores

To investigate the flow characteristics of associated gas displacement by CO₂ in nanopores, displacement models were constructed using the previously established associated gas molecular models in slit pores with different wall properties (Figure 18). The simulation results show that CO₂ can enter quartz slit pores and displace associated gas; however, residual associated gas molecules remain in the pores after displacement, indicating that complete removal of associated gas is challenging at the nanoscale due to strong adsorption effects that significantly restrict gas desorption and migration.

The velocity distribution profiles of associated gas during CO₂ displacement (Figure 18c) in quartz slit pores exhibit a typical parabolic shape, with low velocities near the walls and higher velocities in the pore center—resembling macroscopic Poiseuille flow. This pattern arises because associated gas molecules near the walls experience strong van der Waals interactions, forming stable adsorption layers that markedly reduce molecular mobility. In contrast, free molecules in the pore center are less constrained by the walls and respond more rapidly to external driving forces, resulting in higher flow velocities. Consequently, the presence of adsorption layers effectively reduces the effective flow space and increases flow resistance, thereby lowering overall permeability.

In graphite slit pores, the surface exhibits ultra-low friction and weak adsorption characteristics, resulting in negligible adsorption lag and a nearly flat velocity profile indicative of plug-like, near-rigid-body slip flow [51,52]. This significantly enhances fluid transport efficiency.

Overall, the stronger adsorption and higher interfacial friction in quartz pores limit CO₂ displacement of associated gas, whereas the weak adsorption and low-friction interface in graphite pores promote superior seepage capacity [53,54]. These findings demonstrate that the efficiency of CO₂ displacement of associated gas in nanopores is primarily governed by pore wall interfacial properties, with adsorption strength and friction characteristics serving as key factors influencing displacement effectiveness and flow behavior [51,53]. This understanding provides important theoretical guidance for enhancing associated gas recovery through interfacial modification.

4. Conclusions

This study utilized molecular dynamics simulations to systematically investigate the microscopic occurrence, diffusion behavior, and CO₂ displacement mechanisms of associated gas from the Hesui area in nanopores, with emphasis on the regulatory role of wall polarity (graphite versus quartz). The results highlight wall chemical composition and polarity as key factors governing associated gas migration and displacement efficiency, providing valuable microscopic theoretical guidance for tight reservoir development.

The main conclusions are:

(1) Diffusion behavior differs markedly between free space and nanopore slits. In free diffusion, coefficients depend on molecular mass, size, shape, and interactions, with N₂ showing the highest and C₃H₈ the lowest diffusivity. Walls suppress diffusion, with graphite exerting stronger inhibition on most components (except CO₂). CO₂ exhibits anomalously enhanced diffusion in graphite pores due to linear structure and surface slip, while it is severely restricted on quartz.

(2) Wall properties control the energetic state and adsorption strength. Non-polar graphite displays stronger alkane affinity than polar quartz, yielding lower system and non-bonded energies and more stable layered structures. This stems from graphite’s strong π-dispersion interactions with alkanes versus quartz’s weaker polar interactions, explaining differential occurrence stability across mineral pores.

(3) CO₂ displacement efficiency depends on wall polarity. On polar quartz, CO₂ competitively adsorbs, forming an “associated gas outer—CO₂ inner” layer, but residuals limit completeness. On non-polar graphite, associated gas binds firmly, restricting CO₂ to pore-center displacement. Quartz pores show Poiseuille flow with high friction and low efficiency; graphite exhibits low-friction slip flow, enhancing transport.