Porosity and permeability measurements of the selected core samples were conducted using a CMS-300 porosity and permeability analysis system at the State Key Laboratory of China University of Petroleum (Beijing), following the industry standard SY/T 5336-2006. X-ray diffraction (XRD) analyses were performed using a D8 DISCOVER diffractometer, and the experimental procedures adhered to the national standard SY/T 5163-2018.

Twelve thin sections (0.03 mm in thickness) of tight sandstone samples were prepared as polished casting slices. After impregnation with blue epoxy resin, petrographic observations were conducted using a polarizing microscope to characterize lithofacies features and pore–throat structures, following the industry standard SY/T 5368-2016. Subsequently, scanning electron microscopy (SEM) was employed to analyze the microscopic characteristics of clay-mineral-related pore–throat structures, cements, and intercrystalline micropores, with a maximum resolution of up to 1.2 nm.

The gas displacement water nuclear magnetic resonance (GDWNMR) experiments were conducted using an integrated testing system consisting of a core holder, NMR detection unit, gas monitoring device, pressure pump, nitrogen cylinder, pressure stabilizer, and data acquisition system (Figure 2). The experiments were performed on a MARAN-DRX/2 NMR spectrometer under the following parameters: resonance frequency of 2.38 MHz, echo spacing of 0.1 ms, 128 scans, waiting time of 3 s, and 17,500 echoes, with the temperature maintained at 25 °C throughout the process. The experimental procedure was as follows: First, the tight sandstone samples were dried at 105 °C for 72 h to remove residual moisture and hydrocarbons. The samples were then vacuumed and saturated with simulated formation water (salinity = 22.357 mg/L) for 7 days to ensure full pore saturation and minimize salinity effects. The mass of the water-saturated samples was recorded, and an initial NMR measurement was performed. Subsequently, the saturated samples were placed in the core holder, and a confining pressure of 12 MPa was applied to simulate in situ reservoir stress conditions. After confirming the system’s seal integrity, nitrogen gas was injected in stages at displacement pressures of 2, 4, 6, 8, and 10 MPa until the irreducible water saturation state was reached. After each displacement stage, the T₂ relaxation spectrum of the sample was measured at 25 °C. Upon completion of the displacement experiments, the samples were centrifuged at 400 psi to remove all remaining movable fluids. After confirming the absence of movable fluids, a final T₂ measurement was conducted. In this study, the T₂ distributions of the samples were measured under three conditions: brine-saturated, after gas displacement at different pressures, and after complete centrifugation.

High-pressure mercury intrusion (HPMI) experiments were conducted using an AutoPore IV9505 porosimeter, with a maximum injection pressure of up to 200 MPa, to characterize the pore system under high-pressure conditions and quantitatively reveal the distribution and structural differences of pore-throat radii. As the mercury injection pressure gradually increased, mercury was forced into the connected pore system. Upon reaching the maximum pressure, the pressure was gradually released to allow mercury withdrawal, thereby obtaining complete mercury intrusion–extrusion curves.

The relationship between capillary pressure and pore-throat radius is described by the Washburn equation (Washburn, 1921): P = 2 σ ⁡ cos ⁡ θ r (1) where P is the capillary pressure (MPa), r is the pore-throat radius (μm), θ is the contact angle (140°), and σ is the mercury–air interfacial tension (0.48 N/m). Based on Equation 1, the pore-throat size distribution and other key parameters characterizing the pore structure can be derived, thereby providing a quantitative basis for evaluating pore-throat structure characteristics and connectivity in tight sandstone reservoirs.

The distribution characteristics of the T₂ spectrum intuitively reflect the fluid occupancy state within pores and are closely related to the pore–throat structure of the reservoir. Therefore, in this study, high-pressure mercury intrusion (HPMI) data and nuclear magnetic resonance (NMR) measurements under water-saturated conditions from the same samples were integrated to comprehensively characterize the pore–throat structure distribution. Specifically, NMR measurements were first conducted under fully water-saturated conditions to obtain the T₂ spectrum, and its cumulative amplitude percentage curve was calculated to represent the cumulative pore-volume fraction associated with different relaxation times. In parallel, HPMI experiments were performed on the same samples to derive the cumulative mercury saturation curve as a function of pore–throat radius, which characterizes the cumulative distribution of pore–throat sizes controlling mercury intrusion (Figure 3a). The transverse relaxation time (T₂) of fluids in pores can be expressed in simplified form as (Kleinberg and Horsfield, 1990): 1 T 2 ≈ ρ 2 S V = 1 ρ 2 F s r (2) where F _s is the pore shape factor and r is the pore–throat radius. Experimental evidence indicates that, for a given core sample, ρ ₂ and F _s can be considered constants, leading to a power-law relationship between T ₂ and r. Accordingly, Equation 2 can be rewritten as: r = C   T 2 1 / n (3) where C and n are fitting parameters representing the surface relaxation properties and the structural complexity of the pore–throat system, respectively.

This method effectively integrates NMR relaxation characteristics with pore–throat structure parameters, providing a theoretical foundation for the quantitative identification of pore–throats and multiscale pore structure characterization. Ultimately, the water-saturated T₂ spectra were converted into pore–throat size distribution curves. Based on pore–throat radius, the pores were classified into three categories: mesopores (0.1–1 μm), micropores (0.01–0.1 μm), and nanopores (<0.01 μm).

At different displacement pressures, gas entering the pore-throat causes changes in the fluid volume, resulting in variations in the NMR T₂ spectrum measured after gas displacement. Since the total volume of the sample’s pore-throat remains constant, the variation occurs in the distribution of fluids within the pore-throat. Therefore, by dividing the values of each point of the T₂ spectrum obtained after gas displacement at different pressures by the cumulative value of the NMR T₂ spectrum of water saturated samples and summing the results, the cumulative probability curve can be obtained after water displacement by gas. The difference between this curve and the cumulative probability curve of water-saturated samples represent the cumulative probability of gas distribution within the pore-throat.

The Lpc primarily occurs above the erosional surfaces at the base of channel-fill deposits. It is mainly composed of thick-bedded gray to brownish-gray coarse-grained sandstones, occasionally interbedded with conglomeratic layers (Figures 5a1–a7). The sandstone exhibit a grain-supported texture, and the dominant pore type is residual intergranular pores. X-ray diffraction (XRD) analysis shows that this lithofacies contains the highest quartz content (average 63.7%), relatively low clay content (15.1%), and minimal carbonate minerals (6.6%), indicating excellent reservoir quality (Table 1). The average porosity and permeability are 13.7% and 7.4 mD, respectively, making this lithofacies the most favorable reservoir type in the study area.

The Lmc lithofacies is widely developed in the middle parts of mid-channel bars and consists of gray to brownish-gray medium–coarse-grained sandstones with well-developed parallel lamination, indicative of relatively stable hydrodynamic conditions and a moderately to highly energetic depositional environment (Figures 5b1–b7). Compared with the Lpc lithofacies, its reservoir quality is slightly inferior. Quartz content remains high (62.3%), but feldspar and clay contents are slightly increased, with carbonates averaging 10.1% (Table 1). Under the microscope, a few relatively large intergranular pores are observed, suggesting moderately good pore preservation.

The Lm lithofacies mainly consists of gray, medium-grained sandstones. It exhibits parallel or ripple lamination and occurs primarily along the margins of mid-channel bars, representing a lower-energy depositional setting (Figures 5c1–c7). This lithofacies has an average porosity of 8.5% and permeability of 4.0 mD, reflecting moderate reservoir properties. The mineral composition includes relatively high quartz content (61.2%) and increased feldspar content (21.9%), with clay minerals accounting for 11.6% (Table 1). A few dissolution pores are visible under the microscope; however, the overall pore–throat structure is significantly influenced by clay minerals.

The Lmf consists of gray, medium to fine grained sandstones with thin horizontal bedding, representing a low-energy overbank depositional environment (Figures 5d1–d7). This lithofacies exhibits the poorest reservoir quality, with an average porosity of 3.0% and permeability of 1.9 mD. It has the highest clay (23.7%) and carbonate (9.0%) contents, whereas quartz and feldspar contents are relatively low (Table 1). The high proportion of argillaceous material severely restricts pore development; only a few residual pores are visible under the microscope. This lithofacies represents the densest and least permeable reservoir type in the study area.

The NMR T₂ spectrum primarily reflects the volumetric distribution characteristics of fluids within pore–throat systems, while its converted distribution curve represents the spatial occurrence of these fluids within the pores and throats. Therefore, the variation in fluid volume induced by gas displacement under different pressures can be directly correlated with the gas-occupied porosity (Pog). In this study, the T₂ spectra of 12 samples were converted into pore–throat radius distribution curves. Combined with the results of the gas displacement water NMR (GDWNMR) experiments (Figure 9), the variations in fluid volume under different displacement pressures were quantified. By comparing the pore–throat distribution curves between the water-saturated and gas-displaced states, the Pog of full size pore-throats corresponding to each displacement pressure were determined.

The results indicate that P_og gradually increases with rising displacement pressure (Table 3). When the displacement pressure reaches 10 MPa, the P_og values for all pore–throat types attain their maxima. Overall, Type I samples exhibit the highest gas charging capacity, with the average gas-occupied porosity (P_og) increasing from 1.66% to 4.17% and the gas saturation (S_g) rising from 12.0% to 30.1%. Among them, Sample 6 is the most representative, showing a P_og of 5.65% and an Sg of 38.8% at 10 MPa, indicating well-connected pore–throat structures, sufficient storage space, and the highest gas displacement efficiency. Type II samples show moderate gas charging capacity, with average P_og and S_g increasing to 2.48% and 27.4%, respectively; for instance, Sample 11 reaches an S_g of 28.0% at 10 MPa. Type III samples display smaller increments, with a maximum S_g of only 18.5%, suggesting limited gas expansion. Type IV samples exhibit the lowest values, with average Sg increasing from 9.8% to 18.0%, and only Sample 2 showing slight improvement under high pressure (S_g = 28.4%). Overall, gas displacement efficiency is closely related to pore–throat structure—the more complex the pore–throat system, the poorer the displacement performance and gas-bearing capacity.

The NMR T₂ spectrum primarily reflects the fluid volume within the pore-throat. The transformed distribution curve represents the fluid distribution characteristics. Therefore, the changes in fluid volume after gas displacement at different pressures directly correspond to P_og. Based on Equation 3 and the conversion process depicted in Figures 4, 5, the T₂ spectrum of the 12 samples was first transformed into pore-throat size distribution curves using the multi-stage fitting method. Then, the changes in fluid volume after gas displacement at different pressures are obtained by combining the results of GDWNMR (Figure 10). By analyzing the difference between the pore-throat distribution curves under water saturated conditions and after water displacement by gas, the P_og under different displacement pressures can be determined. As the displacement pressure increases, the P_og gradually increases (Table 4). Under a displacement pressure of 10 MPa, the P_og reaches its maximum value for all pore-throat sizes, with the highest P_og observed in sample 6 with pore-throat sizes ranging from 0.1 to 1 μm, reaching 3.46%. Significant differences also exist in the distribution of S_g and S_pt within pore-throat of different radius (Table 5). According to the classification based on pore-throat size, at gas displacement pressure of 2 MPa, the S_g in pore-throat ranging from 0.001 to 0.01 μm is relatively low, ranging from 0.17% to 1.85%. The corresponding distribution range of S_pt within pore-throat varies from 0.17% to 1.85%. Conversely, at a displacement pressure of 10MPa, the S_g within pores ranging from 0.1 to 1 μm is the highest, indicating the best displacement efficiency, ranging from 3.38% to 23.76%. The distribution range of S_pt within pore-throats ranges from 52.99% to 86.84%.

Gas-occupied porosity (P_og) shows a strong positive correlation with lithofacies and specifically the quartz content, a weak negative correlation with carbonate cement content, and a strong negative correlation with clay mineral content, although the correlation with feldspar content is not significant (Figure 11a). The high-energy hydrodynamic conditions ensured the purity of the sandstone, providing space for gas accumulation. In Lpc formed under high-energy hydrodynamic deposition, samples 4 and 6 exhibit higher quartz content, resulting in relatively high P_mv, which offers favorable conditions for gas preservation. This phenomenon reflects that fact that the quartz and clay mineral content in diagenetic minerals primarily influence displacement results. Brittle mineral particles are less affected by lithostatic pressure, favoring the retention of intergranular pores (Wang et al., 2020c; Wang et al., 2023; Zhao S. et al., 2022). Consequently, as gas displacement pressure increases, P_og gradually increases, ultimately approaching the movable pore volume (Figure 11a). Previous studies have suggested that feldspar dissolution facilitates the formation of secondary porosity (Ma et al., 2017). However, this study reveals a weak negative correlation between feldspar content and gas distribution. This is because clay minerals generated during the kaolinitization of feldspar reduce reservoir connectivity, which is unfavorable for gas displacement. Secondly, although feldspar may develop dissolution pores internally, its poor connectivity has a minimal overall impact on P_og after sample displacement. Clay minerals mainly fill dissolution pores and residual intergranular pores, enhancing the heterogeneity of tight sandstone reservoir pore-throat structure and reducing reservoir connectivity, thereby decreasing the gas content after gas displacement water.

It is noteworthy that, apart from chlorite, other single clay mineral contents have minimal impact on reservoir fluid flow (Figure 11a). This indicates that the formation of chlorite films plays a positive role in preserving intergranular pores, improving reservoir quality, and facilitating gas storage in the reservoir. Well-developed chlorite can retain up to 20% porosity, as similarly demonstrated in the Cretaceous sandstones of the Sawan Gas Field in Pakistan ((Berger et al., 2009). However, the distribution of gas within the reservoir is influenced by multiple clay minerals collectively, and overall clay mineral development can deteriorate reservoir properties. A higher clay mineral content can lead to reduced reservoir mobility. In Lmf, which develops at the channel overbank under weaker hydrodynamic conditions, the clay mineral content is relatively high, resulting in limited movable pore-throat space and, consequently, poorer gas-bearing capacity of such reservoirs. Additionally, with increasing displacement pressure, the migration of clay minerals can block small-sized throats, similarly leading to a deterioration in displacement effects, thereby strengthening the negative correlation (Figure 11a).

The impact of diagenetic minerals on P_og within different pore-throat sizes is substantial (Figure 12). Previous studies suggested that quartz content contributes to the overall preservation of reservoir pores, thereby increasing gas saturation in gas reservoirs (Dong et al., 2023; Jiang M. et al., 2023; Wang et al., 2024). However, this study found that quartz content has the most significant impact on gas distribution within mesopores and micropores. This indicates that the compressive strength of quartz primarily preserves intergranular pores, particularly mesopores and micropores. The good connectivity of these pore-throat types allows gas displacement within them and facilitating gas storage. Feldspar, on the other hand, significantly affects P_og within micropores, showing an overall weak negative correlation (Figure 12).

This indicates that the dissolution pores within feldspar are primarily micropores, but their poor connectivity reduces the proportion of gas-filled pores in these micropores. Clay minerals, acting as inhibitors of fluid flow, increase the specific surface area of rocks, enhancing the interaction between rock particle surfaces and fluids. On the other hand, clay minerals are key factors contributing to reservoir sensitivity, leading to reservoir damage and restricting fluid mobility (Wang et al., 2020c; Zhao S. et al., 2022). Studies have shown that the development of clay minerals has adverse effects on pore-throat structures of various sizes (Figure 12). Carbonate minerals are generally considered to fill intergranular spaces through cementation, occupying fluid flow pathways and inhibiting fluid mobility. This study demonstrates that carbonate cementation primarily affects the gas saturation of mesopores. The development of carbonate cement reduces mesopore connectivity, thereby decreasing gas distribution in these pores. Additionally, dissolution pores within carbonate cementation are predominantly found in micropores and nanopores, increasing the proportion of these pore types. As a result, there is a weak positive correlation between carbonate cement content and gas saturation in micropores and nanopores.

Many studies have shown that permeability correlates better with reservoir gas-bearing capacity evaluation indicators than porosity (Yang et al., 2020; Zhang et al., 2022). However, this study reveals that the influence of porosity and permeability on reservoir gas-bearing capacity is pressure-dependent. As displacement pressure increases, the correlation between P_og and porosity strengthens (Figure 11a). This suggests that higher displacement pressure allows gas to displace a larger volume of fluid, ultimately approaching the volume of movable fluid, resulting in higher gas saturation. In contrast, as displacement pressure increases, the correlation between permeability and P_og gradually weakens. This indicates that with higher gas displacement pressure, more fluid in micropores and nanopores is displaced by gas. The influence of narrow throats and hydrophilic mineral particles on fluid retention diminishes, leading to reduced effects on pore-throat connectivity and permeability. Therefore, with increasing displacement pressure, the correlation between porosity and reservoir gas-bearing capacity gradually improves, while the influence of permeability on reservoir gas-bearing capacity diminishes.

Under different displacement pressures, P_og in mesopores exhibits a strong positive correlation with both porosity and permeability (Figure 12). However, as pore-throat size decreases, the influence of physical properties on gas saturation within pore-throats gradually diminishes. Specifically, the average correlation coefficient of porosity decreases from 0.88 to 0.29, and that of permeability drops from 0.94 to 0.21. This is because mesopores and micropores contribute the most to overall porosity and permeability, making them the primary targets for gas injection during the charging process.

The type, shape, size, distribution, and various combinations of the pore-throat structure, along with their degree of development, all influence the complexity of reservoir pore-throat structure, influencing reservoir connectivity and directly affecting the distribution characteristics of gas within a reservoir (Liu et al., 2023; Zhang et al., 2022). The correlation between pore-throat structure and heterogeneity parameters with the distribution of gas within full-size pore-throat is seen in Figure 11b. The results show that R_m is the most reliable parameter in evaluating gas-bearing characteristics, followed by R_a, with R_max being the least reliable. The R_max exhibits a weak positive correlation with Pog on the whole, whereas R_a demonstrates a strong positive correlation, and both correlations tend to weaken with increasing displacement pressure. The difference lies in that R_max reflects the maximum value of pore-throat within the reservoir, which has limitations in reflecting pore-throat size (Figure 11b). With increasing displacement pressure, fluids within pore-throat of different sizes are gradually displaced by gas, enhancing reservoir connectivity and directly leading to a further reduction in the correlation of this parameter. On the other hand, the R_a can reflect the concentration trend of pore-throat sizes but poorly reflects the distribution characteristics of pore-throat. Therefore, as more pore-throat become gas-connected, the correlation between P_og and R_a gradually weakens. It can be observed that P_og within mesopores is significantly influenced by R_max and R_a, with a stronger correlation with R_a than R_max (Figure 12). Conversely, R_m, which reflects the distribution characteristics of pore-throats sizes well, affects P_og within all three pore-throat size categories. This viewpoint is clearly supported by the parameters of Lm and Lmc. Although Lm, deposited along the edge of bars, exhibits well-developed mesopores—leading to higher Rmax and Ra values than Lmc—comparative gas displacement results demonstrate that Lm possesses superior gas-bearing properties, with P_og values of 2.48% and 1.56%, respectively. These findings suggest that more homogeneous pore-throats structure is more conducive to gas storage. The R_m reflects the distribution characteristics of pore-throat sizes within the reservoir, and P_og distribution is controlled by the size distribution of pore-throat. Thus, as displacement pressure increases, the correlation between P_og and R_m increases. Considering these three pore-throat size parameters together, P_og is primarily controlled by R_m. Gas distribution within pore-throat of different sizes validates these points.

The P_t obtained from mercury injection exhibit a weak negative correlation with P_og within full-size pore-throat (Figure 11b). A smaller parameter value indicates better reservoir connectivity, making it easier for gas to enter pore-throat. However, as displacement pressure increases, more nanopore pore-throat become connected, weakening the influence of drainage pressure on P_og. Comparing P_og within pore-throats of different sizes post-displacement reveals that P_t is only negatively correlated with P_og within mesopores, with little correlation to the displacement results in other pore-throat sizes (Figure 12). This is because gas preferentially displaces fluid within mesopores' movable volume, whereas fluid displacement in micropores and nanopores is less (Figure 12). When dominant gas displacement pathways form within mesopores, gas accumulation within micropores and nanopores ceases to increase, minimizing the impact of P_t on gas distribution within these pore sizes.

Gas-occupied porosity (P_og) within full pore-throat sizes exhibits a weak negative correlation with the sorting coefficient (S_c) (Figure 11b). A smaller S_c indicates better pore-throat sorting performance, more uniform pore-throat size distribution, better connectivity, and overall improved gas displacement efficiency within the reservoir. This characteristic is similarly demonstrated in the homogeneity coefficient (H_c) of the reservoir, where a larger H_c favors better gas displacement of movable fluids within reservoir pore-throat. The correlation between P_og and these two parameters increases with pressure, indicating that better connectivity and homogeneity often yield better gas displacement effects with increasing displacement pressure, facilitating gas storage within the reservoir. The strong positive correlation between P_og and S_max and W_e validates this viewpoint (Figure 11b). Among different pore-throat sizes, the S_c and H_c primarily exhibit a strong correlation with P_og within micropores, as micropores constitute the main component of pore-throat at an average proportion of 51%. The higher the proportion of micropores, the smaller the selection coefficient, indicating better mineral selectivity, improved reservoir connectivity, and more favorable conditions for gas storage. Hydrodynamic stability and facies play a critical role in controlling the sorting characteristics of reservoirs. The Lm, deposited under unstable flow conditions at the margins of bars, exhibits significant grain size variation and poorly sorted pore-throat structures. In contrast, the Lmc, formed under more stable hydrodynamic conditions, displays improved sorting. Although Lm contains a higher proportion of mesopores, the overall gas-bearing performance of Lmc is more favorable.

Displacement pressure is another factor influencing the distribution characteristics of gas within the reservoir (Yang et al., 2022). Under the same pore-throat structure, a higher pressure leads to better displacement effectiveness and a more significant S_g. The P_og and the S_g within different pore-throat sizes under different displacement pressures can be seen in Figure 13. At the same pressure, gas is primarily distributed within mesopores, resulting in the best displacement effectiveness for movable volume. As the pore-throat size decreases, the P_og gradually decreases. This indicates that larger pore-throat radius and better pore-throat connectivity are conducive to enhancing gas displacement effectiveness and promoting gas storage within the reservoir after displacement (Figure 13a). Micropores and nanopores can also store some gas within the reservoir, but due to limitations in connectivity and pore-throat volume, only a tiny amount of movable volume is displaced by gas.

Comparing the displacement effects of pores-throat of the same size under different pressures reveals that the S_g within all size pore-throat categories gradually increases. Among them, the S_g of mesopores increases the fastest. However, as the pore size decreases, the displacement effectiveness gradually worsens (Figure 13b). The underlying difference causing this phenomenon lies in the varying pore-throat sizes, resulting in different reservoir connectivity and storage capacities. As the pore-throat size decreases, the pore-throat structure becomes more complex, connectivity deteriorates, and porosity rapidly decreases. Consequently, the ability of gas to penetrate pore-throat diminishes under the same displacement pressure, resulting in poorer displacement effectiveness.

The experimental results indicate that the gas accumulation in the tight sandstone of the J58 block exhibits a progressive characteristic. From the relationship between P_og, S_g, and displacement pressure (Figure 14), the gas charging process can be divided into two stages: a rapid growth stage and a slow growth stage. In the rapid growth stage, the charging pressure is generally less than 6 MPa, characterized by a rapid increase with rising charging pressure. For example, at 6 MPa, the gas-occupied porosity of Sample 6 in of Lpc (Type I) reached 4.57% (Figure 14a), and the corresponding cumulative gas saturation reached 31.4% (Figure 14b), accounting for 81% of the final gas saturation. Similarly, Sample 3 of Lmf (Type IV) achieved 83% of the final gas saturation. The slow growth stage is characterized by a relatively small increase in gas saturation with increasing displacement pressure. For instance, during the 6–10 MPa stage, the average increase in gas saturation for the four classes of samples was only 20%. In summary, when the pressure is below 6MPa, the gas saturation in tight sandstones can reach 80% of the total gas saturation. In the later stage, high-pressure gas charging primarily enters smaller pore-throats, leading to a slower increase in P_og and S_g. Therefore, the low-pressure charging during the rapid charging stage is key to achieving gas saturation. Under actual geological conditions, low-pressure displacement and larger reservoir pore-throat space can also form tight gas reservoirs with low gas saturation.