Pore Structure and Heterogeneity Characteristics of Deep Coal Reservoirs: A Case Study of the Daning–Jixian Block on the Southeastern Margin of the Ordos Basin

Abstract

To clarify the micropore structure and fractal characteristics of the Danning–Jixian block on the eastern margin of the Ordos Basin, this study focuses on the deep coal rock of the Benxi Formation in that area. On the basis of an analysis of coal quality and physical properties, qualitative and quantitative studies of pore structures with different pore diameters were conducted via techniques such as field emission scanning electron microscopy (FE-SEM), low-pressure CO₂ adsorption (LP-CO₂A), low-temperature N₂ adsorption (LT-N₂A), and high-pressure mercury intrusion (HPMI). By applying fractal theory and integrating the results from the LP-CO₂A, LT-N₂A, and HPMI experiments, the fractal dimensions of pores with different diameters were obtained to characterize the complexity and heterogeneity of the pore structures of the coal samples. The results indicate that the deep coal reservoirs in the Danning–Jixian block have abundant nanometer-scale organic matter gas pores, tissue pores, and a small number of intergranular pores, showing strong heterogeneity influenced by the microscopic components and forms of distribution of organic matter. The pore structure of the Benxi Formation exhibits significant cross-scale effects and strong heterogeneity and is predominantly composed of micropores that account for more than 90% of the total pore volume; the pore structure is affected mainly by the degree of coalification, the vitrinite group, and the ash yield. Fractal analysis reveals that the heterogeneity of macropores is greater than that of mesopores and micropores. This may be attributed to the smaller pore sizes and concentrated distributions of micropores, which are less influenced by diagenesis, resulting in simpler pore structures with lower fractal dimensions. In contrast, mesopores and macropores, with larger diameters and broader distributions, exhibit diverse origins and are more affected by diagenesis, reflecting strong heterogeneity. The abundant storage space and strong self-similarity of micropores in deep coal facilitate the occurrence, flow, and extraction of deep coalbed methane.

1. Introduction

China’s coalbed methane (CBM) industry is undergoing large-scale development. On the basis of the two major shallow CBM industrial bases in the southern Qinshui Basin and the eastern edge of the Ordos Basin, significant breakthroughs in deep CBM development have been achieved in recent years [1,2,3]. Since 2018, exploratory evaluations of deep CBM have been completed in various blocks, such as the Yanan South, Daning–Jixian, and Linxing blocks. In the Yanan South block, the daily stable gas production of a single well in the Yan-3 well area exceeded 13,000 m³/d, with horizontal wells producing between 25,000 and 60,000 m³/d. The Jishen 6–7 Ping 01 well in the Daning–Jixian block yielded an output of 101,000 m³/d. The first deep CBM horizontal well, namely, “Deep Coal No. 1” in the Linxing block ultimately yielded a production rate of 60,000 m³/d [2,4]. Additionally, high-yield gas flows were obtained from some CBM wells at depths of 1000 to 2500 m in regions such as Linfen, Sanjiao, Yulin, Jiaxian, and Shenfu in the Ordos Basin and Zhengzhuang, Shizhuang, and Mabi in the Qinshui Basin, as well as in the Cainan area of the Junggar Basin and the Fukang West block [2,5]. These flows indicate that China has enormous potential for deep CBM and that it is crucial for ensuring national energy security [5,6].

The pore size distribution in coal is extensive, encompassing pores and cracks observable to the naked eye or directly observed at the nanometer scale via optical microscopy, micro- and nanopores detectable by other methods, and closed pores that are challenging to observe with conventional methods [7,8]. Owing to the varying structural characteristics of differently sized pores, a single method for characterizing pores is often inadequate for comprehensively capturing all the pore information in coal [8]. Therefore, it is necessary to select appropriate methods for characterizing pores on the basis of the specific research requirements. Several methods have been established for studying the pore structure of coal, and these methods can be categorized into three main types on the basis of experimental principles: image analysis, fluid intrusion, and radiation detection methods [7,8]. Image analysis methods include optical microscopy (OM) [9], scanning electron microscopy (SEM) [10,11], helium ion microscopy (HIM) [12,13], and atomic force microscopy (AFM) [14,15,16]. Fluid intrusion methods consist of mainly low-pressure CO₂ adsorption (LP-CO₂A) [17], low-temperature N₂ adsorption (LT-N₂A) [18], and high-pressure mercury intrusion (HPMI) [19,20]. Radiation detection methods include computed tomography (CT) [19,21,22], nuclear magnetic resonance (NMR) [23,24], small-angle X-ray scattering (SAXS) [25], and small-angle neutron scattering (SANS) [26]. Additionally, fractal theory is the most common method for assessing the heterogeneity of coal reservoirs and determining complex pore characteristics [7,27]. However, previous studies have primarily combined fractal theory with the aforementioned testing methods, using it as a supplementary approach for effective pore structure research rather than solely for exploring and characterizing pores. Fluid intrusion methods provide precise physical characterization at the nanometer scale but also have limitations [17,20]. For example, in HPMI experiments, high pressures and the corrosive nature of mercury can damage coal samples and alter the characteristics of the distribution of internal pores. Radiation detection methods, represented by low-field NMR technology, offer advantages such as rapid, nondestructive, and repeatable measurements [28,29]. These methods surpass traditional methods in characterizing porosity and pore size distribution and can effectively differentiate between movable water porosity and bound water porosity in coal, as well as evaluate the movable water saturation and bound water saturation of coal reservoirs [29].

The fractal dimension is an important indicator that is used to characterize the complexity and heterogeneity of pore structures. To acquire this parameter, basic pore parameters are obtained primarily from other parameters for measuring pores (such as the pore volume (PV), specific surface area (SSA), and pore size distribution (PSD)). On the basis of these parameters, the fractal dimension can be converted via mathematical models [30,31,32,33]. Studies have shown that pore data obtained from methods such as LP-CO₂A, LT-N₂A, HPMI, SAXS, NMR, and image analysis can all be used to calculate the fractal dimension [34,35]. Among these methods, HPMI and LT-N₂A experiments have straightforward data processing and are commonly used for calculating the fractal dimension [32,36,37,38]. The fractal dimension calculation for HPMI can be solved via the Washburn equation in conjunction with similarity laws or via the Menger model [39], thermodynamic model, and Sierpinski model. The conversion models for the LT-N₂A experimental data include the Frenkel−Halsey−Hill (FHH), Avnir, Neimark, Sierpinski [31], and Menger models [32], which each have different applicability ranges [31,32,40]. According to research by Lu et al. [31], the Sierpinski model can be used to describe the heterogeneity of seepage pores in structural coal, whereas the FHH model is used to characterize the heterogeneity of adsorption pores and can be compared with thermodynamic models. The FHH model can be used to characterize the heterogeneity of adsorption pores, and the results from the Sierpinski model can serve as a complement to the heterogeneity of adsorption pores; however, the fractal dimension calculated via the Menger model is often >3, with a large standard deviation, making it unsuitable for describing the heterogeneity of pores in coal. There are also significant differences in the applicable pore size ranges for specific models. Zhu et al. [41] analysed the pore structure of coal via LT-N₂A and the FHH model and calculated the fractal dimensions D₁ and D₂ at relative pressures (P/P₀) of 0–0.45 and 0.45–1.0, respectively. The study revealed that D₁ and D₂ are closely related to the pore structure of coal. D₁ is influenced primarily by the specific surface area of pores in the 10–220 nm size range, which can be used to characterize the surface roughness of pores in that range; D₂ is influenced mainly by the PV in the 2–10 nm range. The applicability of other models still requires further investigation. Additionally, the pore structure is closely related to the fractal dimension. The results indicate that as the degree of structural deformation increases, the surfaces of coal pores gradually become rougher, the pore structure becomes more complex, and the fractal dimension of the pores increases accordingly [42].

Previous research has shown that using gas adsorption methods or mercury intrusion to calculate the fractal dimension within a certain pore size range has limitations for analysing the pore structure characteristics of coal and their relationships with fractal features [7,31,32]. Considering the different sizes of pore structures developed in coal, complexity in the pore structure can lead to variations in fractal characteristics across different scales. However, studies quantifying the fractal characteristics of coal pore structures at different scales, as well as the differences in and factors influencing pore heterogeneity within various pore size ranges, are relatively rare. This study focuses on the Benxi Formation coal in the Daning–Jixian block on the southeastern margin of the Ordos Basin. On the basis of methods such as microscopic composition analysis, proximate parameter analysis, and helium porosity and permeability tests, combined with LP-CO₂A, LT-N₂A, HPMI, and FE-SEM, the multiscale pore structure of deep coal reservoirs in the Benxi Formation was comprehensively characterized via a combination of qualitative descriptions and quantitative calculations. Additionally, targeted fractal models were selected to determine the fractal characteristics of coal pores of different sizes on the basis of differences in the pore structure. Finally, the relationships between the pore structure parameters of the Benxi Formation deep coal reservoir and the characteristics of the coal quality with fractal dimensions were explored, aiming to provide references for the selection of favorable areas and the evaluation of the resource potential of deep coal reservoirs in the Daning–Jixian block, yielding significant theoretical and practical value.

2. Geological Setting

The study area is located at the southern end of the Jinxian Fault–Fold Zone in the eastern margin of the Ordos Basin, southeast of the Yishan slope zone, and adjacent to the Yanan South block to the south; the study area is within the administrative divisions of Shanxi Province and located in the regions of Daning County and Jixian County (Figure 1) [43,44]. Tectonically, it lies within the Jinxi fold belt and the eastern part of the Yishan slope and is characterized by an NNE trend and a gently NW-tilted monocline [45]. The study area exhibits clear structural zonation from east to west and is divided into the eastern slope zone, Taoyuan anticline zone, central steep slope zone, and western buffer zone, with few faults and a generally small dip of less than 2° [45].

In the Permian system, the area is characterized by marine–terrestrial transitional sedimentary strata, which can be divided into the Shanxi Formation, Taiyuan Formation, and Benxi Formation (Figure 1). Among these formations, the Shanxi Formation features predominantly deltaic deposits, whereas the Taiyuan and Benxi Formations contain delta, lagoon, tidal flat, and shallow shelf sedimentary strata from north to south [46]. Vertically, mudstones alternate with coal and sandstones, and pure mudstones are rare. The main target coal seams are found in the Shanxi, Taiyuan, and Benxi Formations [46]. The Shanxi Formation coal seams are relatively thin, ranging from 1~3 m, whereas the Taiyuan Formation coal seams range from 1 to 6 m in thickness. The Benxi Formation coal seams are relatively thick, ranging from 5 to 12 m, and constitute the principal coal seams in the Ordos Basin. The Benxi Formation coal seams are characterized by extensive distributions; significant thicknesses; and continuous, stable development. The area at depths of 2000 to 3500 m has a planar distribution of approximately 6.9 × 10³ m², making it the most favorable target for deep CBM exploration in the Benxi Formation (Figure 1).

3. Samples and Methods

3.1. Sample Selection

In this study, coal samples were collected from three typical deep CBM core wells, namely, Well P, DJ1, and DJ2, with sampling depths ranging from 1800 to 2400 m; these wells were located primarily near Daning County and Jixian County (Figure 1). The experimental samples studied in this paper are all from the Benxi Formation coal seam of 3 core Wells in the western deep area of Daning–Jixian Block. The fresh samples from the Benxi Formation were composed primarily of coal rocks; four core pieces were collected from each well, totaling 12 pieces, making them representative of the study area. After gas from the coal was desorbed onsite, the fresh coal samples were wrapped in cling film and sealed with wax before being sent to the laboratory for related experimental tests. To eliminate the effects of heterogeneity in the coal on the experimental data, a wire-cutting device was used to cut the cores, producing cylindrical coal samples for micro-CT scanning, helium porosity tests, and permeability tests. The cylindrical coal samples were further cut and crushed for experiments on the maximum reflectance of vitrinite (R_o,max) and maceral composition, as well as proximate parameter analysis, LP-CO₂A, LT-N₂A, HPMI, and FE-SEM (Figure 2).

3.2. Experimental Methods

To obtain the characteristics of the quality of the coal samples, a Leica DM4 P photometric microscope was used for observations. Random measurements of the maximum vitrinite reflectance (R_o,max) (100 points) and maceral composition analysis (500 points) were conducted under oil immersion reflected light, following the GB/T 6948—2008 and SY/T 6414—2014 standards, respectively. Proximate parameter analysis of the moisture via air-dried tests, ash, volatile matter, and fixed carbon contents of the coal samples strictly adhered to the GB/T 30732—2014 standard. The helium porosity and pulse permeability measurements of the cylindrical coal samples were carried out according to the GB/T 34533—2017 standard.

Scanning electron microscope observations of the pores were conducted via an FEI Quanta 200 FEG FE-SEM. Sample preparation involved argon ion polishing with a Gatan 691.CS argon ion thinning device, and sample preparation included steps such as sample pregrinding, high-energy argon ion thinning pretreatment, deposition of a conductive film on the sample surface, gold coating, and electron microscope observation. Scanning electron microscopy revealed minerals distinguishable by their color and morphology, and energy-dispersive spectroscopy (EDS) was used to accurately determine minerals on the basis of their elemental composition and content. The argon ion-polished sample surfaces were smooth, making backscattered electron imaging more favorable for observing the microstructure and morphology of the nanoporosity.

The HPMI experiments utilized an AUTOPORE 9505 mercury intrusion instrument, which provided a maximum mercury injection pressure of up to 413 MPa during testing, with a corresponding lower limit for the tested pore diameter of 3 nm. Before testing, the coal samples were prepared into 1 cm³ cubes, and their surfaces had to be kept smooth during testing to avoid the “orange peel effect”. Additionally, the samples were dried at 105 °C for 8 h to ensure that the original porosity was not destroyed by high temperatures and to effectively remove internal impurities and gases. During the test, the instrument’s internal vacuum was maintained, and the data from the mercury intrusion and withdrawal processes were automatically collected. These data, combined with the Washburn equation [47], allowed for the determination of the PV, SSA, and PSD parameters. The experimental process strictly followed the GB/T 21650.1—2008 standard.

Gas adsorption experiments were conducted via an Autosorb-iQ–MP–C fully automated physical-chemical adsorption instrument, with sample particle sizes ranging from 60 to 80 mesh (0.25 to 0.18 mm). Prior to the LT-N₂A experiments, the coal samples were placed in a degassing station and degassed at 105 °C for 12 h to remove moisture and volatile substances. The degassed coal samples were then moved to the analysis station, where high-purity N₂ was used as the adsorbate for adsorption-desorption tests at 77 K. The Brunauer–Emmett–Teller (BET) and nonlocal density functional theory (NLDFT) models were employed to obtain pore structure parameters, including the PV, SSA, and PSD [48]. The experimental process strictly adhered to the GB/T 21650.2—2008 standard. The LP-CO₂A experiments had a sample pretreatment similar to that of the LT-N₂A experiments, where 1 to 2 g of 60 to 80 mesh (0.25 to 0.18 mm) powdered samples were degassed for 16 h, and high-purity CO₂ was used as the adsorbate for adsorption tests at 273 K. Pore structure parameters, including PV, SSA, and PSD, were obtained on the basis of the NLDFT theoretical model [48]. The experimental process strictly followed the GB/T 21650.3—2011 standard.

In this study, the IUPAC classification, which is based on pore size, was used to categorize pores into three main types: micropores (<2 nm), mesopores (2–50 nm), and macropores (50–10 μm) [49]. This classification scheme is widely used for quantitatively describing and rating pore structures in coal reservoirs.

4. Results

4.1. Coal Characteristics

The R_o,max test results for 12 coal samples from three typical wells in the study area are shown in

Table 1. Results of R_o,max, maceral, and proximate analyses of deep coal samples from the Benxi Formation in the study area.

Sample ID	Depth (m)	Ro,max (%)	σ	Macerals (%)				Proximate Parameter Analysis (%)
				V	I	L	M	Mad	Ad	Vdaf	FCad
DJ1-1	1884.19	2.15	0.09	77.36	19.09	0.20	3.35	0.70	15.15	8.45	77.14
DJ1-2	1884.75	2.02	0.06	67.00	14.00	0.00	19.00	1.10	12.77	7.40	79.88
DJ1-3	1885.57	2.08	0.06	85.33	7.43	0.00	7.24	1.48	15.86	8.05	76.21
DJ1-4	1886.94	2.20	0.05	76.48	4.78	0.00	18.74	0.86	24.95	10.83	66.34
DJ2-1	1961.79	2.52	0.08	82.01	10.25	0.00	7.74	1.16	12.84	8.11	79.16
DJ2-2	1963.29	2.13	0.05	75.05	5.50	0.00	19.45	0.76	12.70	11.68	76.51
DJ2-3	1965.60	2.19	0.08	70.33	13.87	1.93	13.87	1.08	19.82	8.81	72.32
DJ2-4	1965.90	2.21	0.07	80.19	9.71	0.00	10.10	1.16	14.25	8.16	77.84
P-1	2275.99	3.05	0.08	86.00	10.00	0.00	4.00	1.25	5.50	6.29	87.45
P-2	2276.44	3.04	0.04	78.00	14.00	0.00	8.00	1.01	9.72	6.75	83.34
P-3	2276.96	2.97	0.08	80.00	16.00	0.00	4.00	1.47	10.21	6.94	82.33
P-4	2279.00	2.98	0.09	79.00	13.00	0.00	8.00	1.06	8.40	8.05	83.33

R_o,max is the maximum reflectance of vitrinite; σ is the standard deviation of the maximum reflectance of vitrinite; V is the vitrinite; I is the inertinite; L is the liptinite; M is the minerals; M_ad is the moisture content; A_d is the ash content; V_daf is the volatile matter content; FC_ad is the fixed carbon content.

. The R_o,max values range from 2.02% to 3.05% (mean of 2.46%), which are considered indicative of high-rank coal, according to the ISO 11760:2018 standard. This indicates a high degree of the thermal evolution of organic matter, resulting in predominantly bright coal and semi-bright coal, while semi-dull and dull coals are not present. The maceral composition is mainly vitrinite (67.00%–86.00%), which is classified as medium vitrinite coal, according to the ISO 11760:2018 standard. The inertinite group comprises 4.78%–19.09%, whereas the liptinite group is not visible. The mineral content is relatively low, with values ranging from 3.35% to 19.45%, and the minerals consist of primarily inorganic minerals, such as kaolinite, calcite, and pyrite.

The M_ad of the coal samples from the three typical wells range from 0.70% to 1.48% (mean of 1.09%). The A_d values range from 5.50% to 24.95% (mean of 13.51%), which is considered indicative of low-ash coal, according to the ISO 11760:2018 standard. The V_daf values range from 6.29% to 11.68% (mean of 8.29%), indicating a low volatile content in the coal. The FC_ad range from 66.34% to 87.45% (mean of 78.49%), indicating a very high fixed carbon content. When coal is ranked at a high stage, FC_ad is usually negatively correlated with A_d.

The physical parameters, such as the porosity and permeability, of the deep coal reservoirs in the Benxi Formation in the study area were obtained through conventional core testing analysis techniques (Figure 3). The test results indicate that the helium porosities of the deep coal reservoirs in the Benxi Formation range from 1.44% to 8.61% (mean of 6.03%). The permeabilities range mainly from 0.0061 to 9.44 mD (mean of 0.79 mD), indicating a low-porosity, low-permeability reservoir.

4.2. FE-SEM Image Analysis

Deep coal rocks are generally dual-porosity media consisting of matrix pores and fractures [4,7,44]. FE-SEM reveals the complex and diverse pore morphology of deep coal samples, with significant cross-scale effects. The primary pore types are organic matter gas pores, tissue pores, and intergranular pores of inorganic matter; microfractures are mainly cleat fractures and structural fractures.

FE-SEM image analysis reveals that, owing to the high degree of metamorphism in deep coal samples from the study area, gas pores formed in large numbers after significant CBM generation. These pores are predominantly circular and elliptical in shape (Figure 4a–c), with diameters ranging from a few nanometers to several hundred nanometers, and are mostly present in the vitrinite component [4]. Tissue pores are remnants of plant tissue structures that were inherited during coalification. During the coalification process, some plant tissue cells were preserved, forming tissue pores, which are found mainly in vitrinite and semivitrinite. After deep burial and compaction, these pores often appear flattened, elliptical, or irregularly triangular and polygonal in shape (Figure 4d), with diameters ranging from tens of nanometers to several micrometers, and may be filled with clay minerals (Figure 4e). Intergranular pores are found between minerals and the coal matrix, as well as within the minerals themselves (Figure 4e,f). These pores often develop along mechanically weak planes in the microscopic structure of the coal body [44], and these pores may evolve into microfractures under external conditions during CBM extraction, facilitating the formation of gas permeation channels.

Numerous microfractures are observable within clay minerals, especially in flaky and accordion-like kaolinite minerals, where many narrow plate-like structural fractures open at one end (Figure 4i), with apertures ranging from 100 to 300 nm and lengths mostly between 2 to 5 μm. Additionally, some cleat fractures are present locally (Figure 4g,h), with cleat lengths at the microscale ranging from tens of micrometers to several millimeters and widths ranging from several hundred nanometers to a few micrometers. Some of the larger pores and fractures in the coal samples are completely or partially filled (Figure 4d), with the filling materials including clay minerals, pyrite, and carbonate minerals, which reduce the connectivity of the pore-fracture system. On the other hand, the open microfractures in clay minerals not only exhibit good connectivity but also connect pores and fractures. The presence of these microfractures can locally increase the permeability of deep coal reservoirs while also providing effective space and channels for the conversion of adsorbed gas to free gas and for gas migration.

4.3. Quantitative Analyses of the Pore Structure

4.3.1. HPMI Measurements

The mercury intrusion-extrusion curve of coal can clearly indicate the pore structure characteristics [44,45]. The HPMI curves of deep coal samples from the Benxi Formation in the study area are shown in Figure 5a. This curve features a gentle beginning and a steep end [44]. In the initial stage (pressures < 30 MPa), as the pressure increases, the gradual rise at the beginning corresponds to a small increase in mercury saturation, indicating that large pores or microfractures are relatively undeveloped. As the pressure continues to increase (pressures > 30 MPa), the amount of mercury intrusion increases rapidly, reflecting the development of numerous nanopores in the coal samples. On the basis of the degree of closure of the mercury intrusion-extrusion curves, most of the Benxi Formation coal samples in the study area have small openings, and for some samples, the intrusion-extrusion curves almost overlap (Figure 5a), indicating that the pores in these coal samples are predominantly open pores.

The PV and SSA distribution curves of the Benxi Formation coal samples are shown in Figure 6a and Figure 7a. There are significant differences in the trends across different pore size ranges. For pores larger than 30 nm, the increases in the PV and SSA change smoothly, indicating a relatively small number of macropores. However, when the pore size is less than 30 nm, there is a marked change in both the PV stage and the SSA stage, with multiple peaks appearing, indicating a sharp increase in the number of mesopores. The coal samples in the study area are characterized mainly by nanopores. According to the results of the mercury intrusion-extrusion analysis shown in

Table 2. Pore structure parameters of deep coal samples from the Benxi Formation in the study area.

Sample ID	LP-CO2A			LT-N2A				HPMI
	VDFT	SDFT	DCO2	SBET	VDFT	SDFT	DN2	VHPMI	SHPMI
DJ1-1	0.072	247.046	0.501	1.607	0.003	2.298	1.273	0.050	30.650
DJ1-2	0.048	163.766	0.501	3.911	0.007	3.879	3.537	0.027	14.400
DJ1-3	0.071	242.289	0.524	2.291	0.004	2.684	1.22	0.035	17.650
DJ1-4	0.061	207.372	0.501	2.976	0.005	3.57	4.411	0.026	16.310
DJ2-1	0.078	263.869	0.524	0.389	0.001	0.453	6.794	0.059	32.371
DJ2-2	0.072	243.78	0.501	0.386	0.001	0.468	6.079	0.030	21.095
DJ2-3	0.066	221.812	0.524	1.225	0.003	1.189	4.887	0.018	12.687
DJ2-4	0.068	227.762	0.479	0.435	0.002	0.519	6.079	0.017	9.598
P-1	0.077	260.323	0.501	0.515	0.002	0.592	7.032	0.038	19.270
P-2	0.075	252.787	0.479	0.376	0.001	0.449	6.079	0.037	19.240
P-3	0.072	239.423	0.524	0.716	0.002	0.686	6.079	0.032	16.534
P-4	0.079	267.185	0.524	0.454	0.001	0.414	6.556	0.038	20.490

, the pore structure parameters (PV and SSA) and PSD curves in the 3 nm to 10 μm range were calculated from the HPMI experimental data via the Washburn equation (Table 2). The PV values of the Benxi Formation coal rocks in the study area range from 0.017 to 0.059 cm³/g (mean of 0.034 cm³/g). The SSA values of the Benxi Formation coal sample range from 9.598 to 32.371 m²/g (mean of 19.191 m²/g).

4.3.2. LT-N₂A Experiments

The N₂ adsorption-desorption isotherm curves of deep coal samples from the Benxi Formation in the study area are shown in Figure 5b. While the shapes of the curves vary slightly, they generally exhibit an inverse S shape (Figure 5b). On the basis of the isotherm adsorption curve classification proposed by the IUPAC [49], the N₂ adsorption isotherm curve is classified as Type IV, reflecting the presence of mesopores and macropores in the coal samples. The adsorption process can be roughly divided into three stages according to the patterns of fluctuation in the N₂ adsorption isotherm curve: (1) at a P/P₀ value between 0 and 0.1, the adsorption curve increases slowly with a slightly convex shape. When the relative pressure reaches 0.3, monolayer adsorption on the pore surface is nearly saturated, and as the relative pressure increases, it gradually transitions to multilayer adsorption. (2) At P/P₀ values between 0.1 and 0.9, the amount of gas adsorbed increases steadily with increasing relative pressure. (3) At P/P₀ values between 0.9 and 1, the adsorption curve becomes steeper, and adsorption saturation is not achieved, even near the saturation vapor pressure (P/P₀ close to 1), primarily because of gas condensation, indicating the presence of macropores or microfractures in the coal samples. When P/P₀ is between 0.45 and 0.90, the adsorption and desorption curves diverge, forming a hysteresis loop between the two curves, which is due to capillary condensation occurring in this relative pressure range, where the adsorption and desorption processes are not completely identical or reversible.

The shape of the hysteresis loop provides important information about the pore structure [48]. On the basis of the IUPAC classification of hysteresis loops, the hysteresis loops of coal samples from the study area are mainly of the H3 type. The hysteresis loops of all the coal samples resemble the H3 type. When the relative pressure is relatively low (P/P₀ < 0.5), the adsorption and desorption curves overlap, indicating that the pore types are mainly cylindrical, conical, and parallel plate pores. When P/P₀ is high, slight hysteresis loops appear, indicating the presence of open pores.

The pore structure parameters and PSD curves for coal samples in the 1.06–77.7 nm pore size range were obtained via the BET method and the NLDFT model, which were used to calculate the LT-N₂A data (Table 2). The BET SSA values of the Benxi Formation coal samples in the study area range from 0.376 to 3.911 m²/g (mean of 1.273 m²/g). The DFT PV values range from 0.001 to 0.007 cm³/g (mean of 0.003 cm³/g). The average pore diameters range from 1.220 to 7.032 nm (mean of 5.002 nm). Among them, the DJ1 well coal sample has a larger PV and SSA and a smaller pore diameter, indicating a lower degree of coal metamorphism and more developed mesopores, resulting in the largest SSA.

Figure 6b shows the variation in the PV with increasing pore size for the coal samples in the study area, revealing clear multipeak characteristics. The main peak regions are in the 1.06–1.5 nm and 5–10 nm pore size ranges, with the highest peak located in the 5–10 nm range. When the pore size is greater than 10 nm, the increase in the PV becomes very small. Figure 7b shows the variation in the SSA with increasing pore size for the coal samples in the study area. Like the PV distribution, the distribution of the SSA of some coal samples also shows clear multipeak characteristics, with the main peak regions appearing in the 1.06–1.5 nm and 5–10 nm pore size ranges.

4.3.3. LP-CO₂A Experiments

The LP-CO₂A isotherm curves of deep coal samples from the Benxi Formation in the study area are shown in Figure 5c. The CO₂ adsorption isotherm curves resemble the Type I isotherms classified by the IUPAC [49], reflecting micropore-filling phenomena, indicating that a certain number of micropores have developed in the coal reservoirs. Although the CO₂ adsorption curves of different coal samples are quite similar in shape, there are differences in the maximum adsorption capacity of CO₂, which reflects certain differences in the micropore structures of the various coal samples. The maturity of different coal samples also affects their adsorption capacity. For example, the CO₂ adsorption capacity of the coal sample from the P well is significantly greater than that of the DJ1 and DJ2 wells, suggesting that the number and development of micropores in the coal samples in the study area may be related to the degree of coal metamorphism.

The NLDFT model was used to calculate the LP-CO₂A data, and the PSD curves and pore structure parameters in the 0.3–1.5 nm pore size range were obtained (Table 2). The DFT SSA values range from 163.766 to 267.185 m²/g (mean of 236.451 m²/g), which are much higher than the results from the LT-N₂A analysis. This indicates that micropores make up a relatively large proportion of the coal samples. The DFT PV values obtained from the CO₂ adsorption isotherms range from 0.048 to 0.079 cm³/g (mean of 0.070 cm³/g). The average pore diameters range from 0.479 to 0.524 nm (mean of 0.507 nm). On the basis of the PSD characteristics derived from the CO₂ isotherm adsorption data (Figure 6c and Figure 7c), the overall micropore development of the samples shows a multipeak pattern, with the main peaks at approximately 0.3–0.4 nm, 0.4–0.7 nm, and 0.7–0.9 nm. This indicates a wide range of distribution of micropores in the coal rock samples, with the main peak pore diameters corresponding to approximately 0.36 nm, 0.52 nm, and 0.82 nm in the 0.3–0.9 nm range.

4.3.4. Fractal Dimension Analysis

Fractal theory is a geometric method used to study the structural characteristics of irregular objects with self-similarity. It can be applied to quantitatively analyze the complexity and similarity or heterogeneity of coal pore structures [31,40]. The parameter used to quantitatively describe this self-similarity is called the fractal dimension (D). The fractal dimension can characterize the roughness and structural irregularity of solid surfaces. A higher fractal dimension indicates a more complex pore surface and pore structure, as well as greater heterogeneity [35,38].

The fractal dimensions of coal pore structures can be calculated via various methods and models, such as mercury intrusion, gas adsorption, and image analysis. Given the different scales of pore spaces developed in coal and the limitations of commonly used experimental methods for representing different pore sizes, accurately and quantitatively characterizing the complexity and heterogeneity of coal pore structures via fractal theory from a single method is difficult [50,51,52,53,54]. Therefore, different fractal models should be selected on the basis of experimental methods to target different pore size structures to determine the fractal characteristics of coal pores. In this study, the data from the LP-CO₂A experiment for pores smaller than 1.5 nm were used to calculate the fractal dimension of micropores via the micropore filling model. Data from the LT-N₂A experiment for pores between 1.5 and 50 nm in size were used to calculate the fractal dimension of mesopores via the FHH model. Data from the HPMI experiment for pores larger than 50 nm were used to calculate the fractal dimension of the macropores via the thermodynamic fractal model.

The thermodynamic fractal model based on HPMI data is given by the following formula [51]:

$$ln\left({\displaystyle \frac{W_n}{r_n^2}}\right)=D_{1}ln\left(Q_n\right)+C$$

(1)

where $W_n$ represents the surface energy of the pores, J; $Q_n$ represents the total mercury intrusion volume, which is the total PV, mL; $r_n$ represents the pore radius corresponding to the n-th mercury intrusion volume, μm; C represents a constant; and $D_1$ represents the fractal dimension of the macropores, which is obtained from the slope of the linear regression fitting curve of $ln(W_n/r_n^2)$ and $ln\left(Q_n\right)$.

The FHH model, which is based on LT-N₂A data [52], is used to calculate the fractal dimension, as follows:

$$ln\left(V\right)=\left(D_2-3\right)ln\left[ln\left({\displaystyle \frac{p_0}{p}}\right)\right]+C$$

(2)

where $V$ is the gas adsorption volume at equilibrium pressure p, cm³; $p_0$ is the saturation vapor pressure of the gas, MPa; p is the equilibrium pressure, MPa; C is a constant; and $D_2$ is the fractal dimension of mesopores, which is obtained from the slope of the linear regression fitting curve of $ln\left(V\right)$ and $ln\left[ln\right(p_0/p\left)\right]$.

Previous studies have shown that adsorbate molecules in micropores undergo mainly micropore filling. Jaroniec [53], through adsorption experiments on activated carbon micropores, reported that the PSD of micropores is a key factor affecting their heterogeneity and proposed a micropore fractal model based on the PSD function $J\left(x\right)$ and pore size $x$, as follows:

$$lnJ\left(x\right)=(2-D_3)lnx+C$$

(3)

where $J\left(x\right)$ is the pore distribution density function; $x$ is the pore size in nanometers, nm; and $D_3$ is the fractal dimension of micropores.

Using different fractal models, the fractal dimensions of pores in different size ranges were calculated from the HPMI, LT-N₂A, and LP-CO₂A experimental data [51]. According to Equations (1)–(3), calculations and fitting of the experimental data for deep coal samples from the Benxi Formation were performed. The resulting fitting curves reflected different fractal characteristics in the coal pores, and the fitting accuracy of each curve was relatively high (Figure 8). A comparative analysis of the fractal dimensions obtained from different pore size ranges via various testing methods yielded the following results (

Table 3. Fractal dimensions of deep coal samples from the Benxi Formation in the study area.

Sample ID	HPMI		LT-N2A		LP-CO2A
	D1	R2	D2	R2	D3	R2
DJ1-1	2.702	0.999	2.736	0.984	2.421	0.988
DJ1-2	2.802	0.995	2.740	0.983	2.466	0.988
DJ1-3	2.825	1.000	2.751	0.978	2.405	0.988
DJ1-4	2.624	0.998	2.756	0.971	2.433	0.989
DJ2-1	2.384	0.998	2.506	0.964	2.433	0.987
DJ2-2	2.626	0.999	2.529	0.956	2.484	0.987
DJ2-3	2.501	0.998	2.663	0.985	2.466	0.988
DJ2-4	2.564	0.997	2.591	0.958	2.478	0.986
P-1	2.807	1.000	2.522	0.995	2.336	0.986
P-2	2.728	1.000	2.511	0.990	2.347	0.986
P-3	2.654	1.000	2.540	0.992	2.382	0.984
P-4	2.787	0.999	2.572	0.964	2.380	0.986

): the fractal dimension values of macropores (D₁) in deep coal samples from the Benxi Formation range from 2.384 to 2.825, with an average of 2.667 and a correlation coefficient greater than 0.99; the fractal dimension values of mesopores (D₂) range from 2.506 to 2.756, with an average of 2.618 and a correlation coefficient greater than 0.9; and the fractal dimension values of micropores (D₃) range from 2.336 to 2.484, with an average of 2.419 and a correlation coefficient greater than 0.9. The pattern of the distribution of the overall fractal dimension of deep coal samples from the Benxi Formation reveals that D₃ values are more concentrated and small, whereas D₁ and D₂ values are more dispersed and greater.

5. Discussion

5.1. Full-Scale Pore Structure Characterization

On the basis of the testing principles and analysis of calculation models, the HPMI, LT-N₂A, and LP-CO₂A methods all have advantages within different pore size ranges and can be used to accurately characterize the pore and fracture structures within their respective dominant pore size ranges [55,56]. In this study, previous research results are integrated, and different testing methods are used to measure the optimal pore size range, via applicable calculation models (Figure 9).

According to the results of the joint characterization of the PV distribution, along with the pore size from the Benxi Formation coal samples from three typical wells in the study area (Figure 10), the PV distribution type is primarily unimodal, with a dominance of micropores, and the PSD shows a unimodal pattern, with values concentrated in the range of 0.3 to 1.5 nm. As shown in Figure 10 and

Table 4. Distribution of the PV and SSA of deep coal samples from the Benxi Formation in the study area.

Sample ID	PV (cm3/g)				SSA (m2/g)
	Micropores	Mesopores	Macropores	Total	Micropores	Mesopores	Macropores	Total
DJ1-1	0.072	0.002	0.002	0.077	247.096	0.928	0.015	248.039
DJ1-2	0.048	0.006	0.001	0.055	163.773	2.718	0.029	166.519
DJ1-3	0.071	0.003	0.004	0.079	242.362	1.502	0.096	243.960
DJ1-4	0.061	0.004	0.003	0.067	207.370	1.893	0.010	209.274
DJ2-1	0.078	0.001	0.001	0.081	263.870	0.448	0.006	264.324
DJ2-2	0.072	0.001	0.001	0.075	243.780	0.461	0.003	244.244
DJ2-3	0.066	0.003	0.001	0.069	221.810	1.073	0.003	222.886
DJ2-4	0.068	0.001	0.001	0.071	227.760	0.499	0.004	228.263
P-1	0.077	0.002	0.005	0.083	260.342	0.462	0.121	260.925
P-2	0.075	0.001	0.005	0.081	252.790	0.442	0.093	253.324
P-3	0.072	0.002	0.004	0.078	239.420	0.666	0.042	240.128
P-4	0.079	0.001	0.004	0.084	267.180	0.407	0.092	267.679

, the total PV values of the Benxi Formation coal samples in the study area range from 0.055 to 0.084 cm³/g (mean of 0.075 cm³/g), where the PV is contributed mainly by micropores, followed by mesopores and macropores, both of which contribute relatively little. The micropores PVs range from 0.048 to 0.079 cm³/g (mean of 0.070 cm³/g), accounting for 87.6%–96.7% (average of 93.2%) of the total PV. The mesopores PVs range from 0.001 to 0.006 cm³/g (mean of 0.002 cm³/g), accounting for 1.3%–10.6% (average of 3.4%) of the total PV. The macropore PVs range from 0.001 to 0.005 cm³/g (mean of 0.003 cm³/g), accounting for 1.2%–6.0% (average of 3.4%) of the total PV.

According to the results of the joint characterization of the pore SSA distribution, along with the pore size from the Benxi Formation coal samples of three typical wells (Figure 11), the distribution type of the pore SSA is mainly unimodal, with a dominance of micropores, and the PSD shows a unimodal pattern, with values concentrated in the range of 0.3 to 1.5 nm. As shown in Figure 12 and Table 4, the total SSA values of the Benxi Formation coal samples range from 166.519 to 267.679 m²/g (mean of 237.464 m²/g), where the SSA is contributed mainly by micropores, with negligible contributions from mesopores and macropores. The SSA values of micropores range from 163.773 to 267.180 m²/g (mean of 236.463 m²/g), accounting for 98.4%–99.8% (average of 99.5%) of the total SSA. The SSA values of mesopores range from 0.407 to 2.718 m²/g (mean of 0.958 m²/g), whereas those of macropores range from 0.003 to 0.121 m²/g (mean of 0.043 m²/g), and the contribution of mesopores and macropores to the total SSA is less than 1%. The results indicate that the curves of the PV and SSA distribution of the Benxi Formation coal samples are similar, where the pore size range corresponding to the developed PV also corresponds to that of the SSA. Furthermore, in the Benxi Formation coal reservoir in the study area, micropores dominate the changes in the SSA, whereas mesopores and macropores contribute relatively less, with micropores smaller than 1.5 nm playing a critical role. Therefore, it can be inferred that micropores in coal provide many adsorption sites for storing adsorbed CBM.

5.2. Factors Influencing Pore Development

The genesis of coal pore types is complex, with a wide range of size distributions and strong heterogeneity. The characteristics of development are the result of the combined effects of multiple factors, including coalification, metamorphism, tectonic evolution, coal composition, and subsurface fluids [57,58]. Coalification is the internal factor, whereas tectonic stress is the primary external factor. The interaction between these internal and external factors has led to the current characteristics of pores and fractures in coal seams with different coal ranks, coal-bearing basins, and structural locations. Additionally, coal-forming materials, depositional environments, the basin burial history, and the thermal history play important controlling roles [7,59,60].

5.2.1. Coalification Level and Maceral Composition

An analysis of the relationships between the degree of coalification and the pore structure parameters of deep coal samples from the Benxi Formation in the study area reveals that the degree of coalification is the most significant factor influencing pore development in the Benxi Formation coal reservoir (Figure 13a). FE-SEM observations reveal that organic matter pores constitute the primary storage space in the Benxi Formation coal. Further analysis indicates that in the Benxi Formation coal reservoir, Ro, max is positively correlated with the micropore PV and SSA of the coal samples (Figure 13a), whereas R_o,max is weakly negatively correlated with the mesopore PV and SSA (Figure 13a). R_o,max is negatively correlated with the macropore PV when R_o,max is less than 2.4% and weakly negatively correlated when R_o,max is greater than 2.4% (Figure 13a). Among these correlations, micropores in coal samples show the best correlation with R_o,max, indicating that the deep coal reservoir in the Benxi Formation is dominated by micropores, This domination stems from the proportion of the micropore PV in the total PV. This may be because, during the high to overmature stages, especially during the gas generation stage, the aromatization of organic matter increases, and the ordered stacking of benzene rings becomes more pronounced, leading to the extensive development of micropores, which is consistent with previous research findings [60].

Different macerals often lead to differences in the characteristics of pore development of coal, making the impact of macerals on coal pores crucial [58]. Research has shown that the vitrinite content in the Benxi Formation coal reservoir is positively correlated with micropore PV and SSA and weakly negatively correlated with mesopore PV and SSA, whereas it is weakly positively correlated with macropore PV and SSA (Figure 13b). FE-SEM observations reveal that vitrinite is the main carrier of tissue pores, stomata, and microscopic endogenous fractures, followed by fusinite, which develops primarily into plant tissue pores. In contrast, pores are poorly developed in liptinite, resulting in increases in the PV and SSA with increasing vitrinite content. Furthermore, generally, researchers suggest that the stronger the hydrocarbon generation potential is, the more conducive it is to the development of gas pores [56,59]. Vitrinite, which has the highest content in coal, combined with its strong thermoplasticity and brittleness, produces more gas, leading to the development of gas pores within the vitrinite [59,60].

Although the mineral content is well correlated with the PV and SSA in coal samples (Figure 13c), FE-SEM observations and maceral analysis reveal that the mineral content in Benxi Formation coal samples is relatively low, generally less than 30%, and that associated inorganic micropores are also scarce, indicating that the contribution of inorganic pores in minerals to the PV in Benxi Formation coal samples is limited. Further analysis reveals that the mineral content in the Benxi Formation coal reservoir is negatively correlated with the micropore PV and SSA, weakly positively correlated with the mesopore PV and SSA, and weakly negatively correlated with the macropore PV and SSA (Figure 13c). This is primarily due to the development of numerous mineral pores, including dissolution pores in carbonate minerals, intercrystalline pores in clay and carbonate minerals, and shrinkage pores associated with differential shrinkage, which are closely related to minerals [57]. These pores reflect the differences in thermoplastic and mechanical properties between minerals and organic matter, which are used to determine the morphology of different shrinkage pores [58]. Additionally, many minerals often fill pores or fractures, sometimes forming fracture veins, which affects the permeability of pores and fractures [57,58]. The minerals that fill these pores and fractures are mainly calcite and other carbonate minerals, along with pyrite and clay minerals.

5.2.2. Proximate Parameter Analysis

The relationships between the pore structure parameters and approximate parameters for deep coal samples from the Benxi Formation in the study area are shown in Figure 14. Since the ash content is a derivative of minerals in coal and is formed through complex reactions, such as decomposition and combination [45], the correlation between pore structure parameters and ash yield (A_d) is consistent with the correlation with the mineral content. In the Benxi Formation coal reservoir, A_d is negatively correlated with the micropore PV and SSA, weakly positively correlated with the mesopore PV and SSA, and weakly negatively correlated with the macropore PV and SSA (Figure 14a). This is primarily because ash in coal reflects the mineral content, which is influenced by the supply of terrigenous material. The impact of ash on a reservoir’s physical properties is manifested as the partial filling of pores and throats, leading to poorer reservoir quality. As the ash yield increases, the degree of pore development in the coal reservoir decreases.

The study also revealed that in the Benxi Formation coal reservoir, fixed carbon on an ash-free basis (FC_ad) is positively correlated with the micropore PV and SSA, weakly negatively correlated with the mesopore PV and SSA, and weakly positively correlated with the macropore PV and SSA (Figure 14b). The reason is that fixed carbon is the organic matter that remains in solid form [45]. As the FC_ad content increases, various small molecular functional groups and side chains in the coal gradually decrease, whereas the large molecular benzene structures become denser, thereby forming more micropores and significantly increasing the micropore PV and SSA.

5.3. Factors Influencing the Fractal Dimension

In this study, the correlation between the fractal dimensions of pores of different sizes in deep coal samples from the Benxi Formation and the R_o,max, vitrinite content, and ash yield was analyzed. The results show that these factors are correlated with the pore structure development to some extent (Figure 15). The D₃ values tend to decrease as the R_o,max and vitrinite contents increase, whereas the values tend to increase with increasing A_d (Figure 15c). Similarly, the D₂ values decrease with increasing R_o,max and vitrinite contents and increase with increasing A_d (Figure 15b), whereas the D₁ value shows no obvious correlation with the R_o,max, vitrinite content, or A_d (Figure 15a). As the coalification and vitrinite contents increase, the D₃ and D₂ values decrease, indicating a simpler pore structure. This is the result of more homogeneous pore types and enhanced pore uniformity, with micropores being particularly prominent (Figure 15b). The ash yield has a suppressing effect on the fractal dimensions of micropores and mesopores, primarily because the reduction in the mineral content is accompanied by organic matter enrichment, leading to the development of numerous micro- and nanopores during thermal evolution [61,62]. The diverse forms of micropores complicate the structure of small pores in the coal pore network, thereby increasing the D₃ value. Conversely, as the mineral content increases, more interlayer pores develop between clay minerals in the coal, and the plasticity of the clay minerals makes them more prone to deformation, which further complicates the interlayer fracture network of clay minerals, resulting in an increase in the D₂ value. Compared with the macropores in the Benxi Formation coal samples, the D₁ value is not significantly correlated with the R_o,max, vitrinite content, or A_d because the proportion of macropores in the deep coal reservoirs of the Benxi Formation is lower, leading to a weaker influence of coalification and coal quality on macropore development. The development of macropores is controlled mainly by other factors, which may include the reservoir temperature, pressure, and burial time.

5.4. Relationships Between the Pore Structure and Pore Fractal Characteristics and Their Geological Significance

The correlations between the fractal dimension and pore structure parameters of coal samples from the Benxi Formation in the study area are shown in Figure 16. In the correlation diagrams between the fractal dimension and pore parameters, significant differences are observed among the macropores, mesopores, and micropores. The micropores have lower D₃ values and higher PV and SSA values, with D₃ showing a weak negative correlation with the PV and SSA (Figure 16c). In contrast, mesopores and macropores have smaller PVs and SSAs, but their fractal dimensions (D₁ and D₂) are more scattered, showing a positive correlation with the PV and SSA (Figure 16a,b). This may be due to the strong heterogeneity of micropores, mesopores, and macropores in the deep coal reservoirs of the Benxi Formation. Deep coal reservoirs are complex geological bodies, and under the combined effects of various organic and inorganic components, a diverse and complex pore structure network is formed. During their geological history, the organic matter in coal accumulates, is buried, matures, and generates hydrocarbons. The higher the organic matter content is, the more nanoscale organic pores are produced during thermal maturation and hydrocarbon generation, and these pores tend to be simple in shape and mostly round, with smooth edges. This simplicity of the pore space is reflected in a decrease in D₃. The compression and deformation of organic pores during organic matter evolution and geological reconstruction also exacerbate the complexity of the pore structure [51], whereas mesopores and macropores tend to be more complex and are composed mainly of intergranular pores and interlayer pores of clay minerals, with smaller PVs and SSAs, resulting in increases in D₁ and D₂.

Comparative analysis reveals that pore size has the most significant influence on the fractal dimension of pores. The average results of the calculation reveal D₁ > D₂ > D₃, indicating that larger pores have greater heterogeneity and more irregular pore sizes, with a more dispersed distribution, whereas micropores have relatively uniform and regular pore distributions (Figure 17). Since the micropores in the deep coal reservoirs of the Benxi Formation are primarily organic pores of organic origin, they exhibit strong self-similarity. Owing to their small size, they are less affected by diagenesis, showing good pore sorting, similar shapes, and weak heterogeneity, thus resulting in smaller fractal dimensions with stronger fractal characteristics (R² = 0.99). In contrast, macropores in coal samples have larger pore diameters, a wider distribution range, and complex origins. Influenced by diagenesis, macropores exhibit diverse shapes, significant deformation, and strong heterogeneity, leading to generally higher fractal dimensions with varying degrees of fractal strength (R² = 0.9~0.99).

The study of the fractal characteristics of coal pores not only reflects the complexity and variability of the reservoir space but also has geological significance for the evaluation of deep CBM [63,64]. In the coal reservoirs of the Benxi Formation, micropores in coal have a high proportion of PV and large SSA values, providing abundant gas adsorption sites and accumulation spaces, which are beneficial for the presence of both adsorbed and free gas in coal. Macropores, with more complex pore structures, also facilitate the accumulation of free shale gas because of their greater reservoir space. However, previous studies have noted that the higher the fractal dimension is, the more complex the pore structure of coal, which, while increasing the gas adsorption capacity, also reduces permeability, making shale gas diffusion and desorption more difficult and reducing the flow capacity [51,64]. In the deep coal of the Benxi Formation, the high organic carbon content and the large proportion of organic pores lead to a more homogeneous micropore pore type, enhanced self-similarity, reduced fractal dimensions, and a lower pore structure complexity, which is conducive to the flow and subsequent extraction of deep CBM.

6. Conclusions

(1)

The Benxi Formation in the study area has relatively high overall porosity in its deep coal reservoirs, with an average porosity of 6.03%. The pore types in the Benxi Formation coal are complex and diverse and consist mainly of organic matter gas pores, deformed plant tissue pores, and intergranular pores, with microfractures that are mainly cleat fractures and structural fractures.

(2)

The cross-scale effects on the pore structure of the Benxi Formation coal reservoirs in the study area are evident, and the heterogeneity is strong. The curve showing the PV distribution is unimodal, with its main peak corresponding to the micropores. Micropores are the primary contributors to the total PV, followed by mesopores, with macropores contributing the least. The SSA distribution curve also exhibits an unimodal shape, with the main peak corresponding to the micropore section, where micropores account for more than 99% of the SSA, providing more gas adsorption sites and enrichment space.

(3)

Different dominant factors in the Benxi Formation coal in the study area control the structural characteristics of pores at different scales. Among these factors, the coalification level and vitrinite content promote the development of pores, which mainly control the development of micropores, whereas the ash content inhibits pore development, which mainly suppresses the development of micropores and macropores.

(4)

The fractal dimensions of the deep coal rocks in the Benxi Formation reveal that D₁ > D₂ > D₃, indicating that the heterogeneity of the macropores in the deep coal reservoirs is greater than that of the micropores and mesopores. Since micropores are predominantly organic pores with smaller pore diameters and concentrated distributions and are less affected by diagenesis, they exhibit smaller fractal dimensions and strong fractal characteristics, indicating self-similarity in their genesis and distribution. In contrast, macropores, owing to their larger diameters, broader distribution ranges, diverse genesis mechanisms, and susceptibility to diagenesis, have larger fractal dimensions, indicating strong heterogeneity and complexity in their genesis.

(5)

The micropores in the deep coal rocks of the Benxi Formation in the study area, owing to their large PV and abundant SSA, are conducive to the existence of free gas and adsorbed gas in the coal; their significant proportion of organic matter pores and strong self-similarity favor the flow and extraction of gas from deep coal seams.