The No. 3 coal seams of Early Permian Shanxi formation in Sihe mine, in the southern Qinshui basin of China were taken as research objects to study the pore structure characteristics after microbial treatment. These are typical primary mineable coal beds found in northern China. The study area lies in the southeast of Shanxi Province and forms a part to the western margin of North China Platform. The average thickness of No. 3 coal seam is about 6 m and R _{o, Max.} is 2.54%. The research area is a focus of CBM research in China, and is the earliest and most successful commercialization area of the CBM industry in China. To maintain the contact between the coal and microorganisms as much as possible, while avoiding the channels becoming blocked by pulverized coal, the particle size of the granular coal specimens was about 20 × 20 mm and a coal sample of 18 kg was loaded into each experimental device. Proximate analysis, low-temperature nitrogen adsorption (LTNA), mercury porosimetry (MP) and isothermal adsorption measurements, and field emission scanning electron microscopy (FESEM) were applied to pre-test/post-test samples. Two sets of experimental devices were designed (Figure 1). The test data are summarized in Table 1.

The MECBM treatments were performed using specially designed apparatus (Figure 1). The flora and experimental water were taken from CBM wells in the Sihe mine. The floras were cultured in the laboratory. To ensure the enrichment of methanogens in water samples, fluorescence observation and counting were performed before the experiment under a fluorescence microscope with 420 nm incident light. Preparation of 10 L of substrate was carried out, and the components mainly included glucose, peptone, sodium chloride, beef extract, l-cysteine hydrochloride, resazurin, sodium bicarbonate, and sodium sulfide. The experiment lasted for 47 days (until 72 h no gas production was noted). Two experimental devices produced 10,449 and 11,120 ml of gas respectively, of which the methane content was 4,013 and 4,583 ml.

An Automatic Proximate Analysis Instrument was used to perform proximate analysis (the moisture, ash, and volatile and fixed carbon contents) experiments, as per ISO 17246-2010. A Micromeritics ASAP-2020 automated surface area analyzer (United States) was adopted to perform the low-temperature nitrogen adsorption/desorption experiments, as per the ISO 15901.2-2006 test method over a pore diameter analysis range of 0.35–300 nm. An Automated Isothermal Gas Adsorption/Desorption Experiment System TerraTek ISO-300 (United States) was used to conduct the isothermal adsorption/desorption experiments, as per Chinese National Standard GB/T 19560-2008. All the above tests were completed at the State Key Laboratory of Coal and Coalbed Methane Co-Extraction in China. A Micromeritics Autopore IV 9500 Instrument (United States) was used to perform mercury porosimetry experiments, according to the ISO 15901-1-2005 test method, the particle size of the coal sample used was 4.75–3.35 mm over a pore-diameter analysis range of 7.5 nm–360 μm. The pore characteristics of coal and methanobacteria were observed under FEI Quanta 250 FEG (United States) field emission scanning electron microscopy (these two tests were conducted at Henan Polytechnic University in China).

In this study, the following Xoдot criteria developed in the former USSR were adopted: the pores in a coal specimen were divided into: micropores (<10 nm), minipores (10–100 nm), mesopores (100–1,000 nm), and macropores (>1,000 nm) (Hudot, 1966; Cheng et al., 2021).

From the distribution of specific surface area (Figure 2), the specific surface area of coal samples comprised mostly micropores and minipores. The specific surface area of micropores and minipores of coal samples increased from 93.6% to 97.8% and 96.5% after bioconversion under LTNA test conditions, and the specific surface of micropores and micropores of coal samples decreased from 99.7 % to 99.5% under mercury injection test conditions. From Figures 3, 4A,4B, the specific surface area of micropores of coal samples was found to increase significantly after bioconversion, while that of the minipores increased slightly, and the overall specific surface area increased from 1.79 to 4.01 m²/g on average after LTNA testing; however, the overall specific surface area decreased slightly, from 5.398 to 5.246 m²/g under mercury injection test conditions.

Different from the specific surface area, the contribution of pore volume arose mainly from minipores, mesopores, and macropores (Figure 3), the contribution of meso-macropores increased from 72.4 % to 76.7% and 73.1% after bioconversion under MP test conditions (Figure 2). The contribution of micropores to the pore volume increased from 4.34 % to 18.0% and 12.98% under the LTNA test, although the contribution of micropores to the increase in pore volume after the experiment was large, the proportion of micropores to the increase in the pore volume remained small. As can be seen from Figures 4C,D, the pore volume of coal samples increased from 0.0018 cm³/g (LTNA) and 0.0530 cm³/g (MP) before testing to 0.0051 cm³/g (LTNA) and 0.0626 cm³/g (MP) after testing in LTNA and MP conditions, respectively (Table 1).

The differences in the data of specific surface area and pore volume between the LTNA and MP tests methods are mainly based on two reasons: one is that MP focuses on macropores and mesopores, while LTNA is more advantageous for micropores and minipores; the other is that according to the experimental results, microbial modification of mesopores is greater, which reduces the specific surface area of mesopores and increases the pore volume.

The pore structure of coal has strong heterogeneity and anisotropy, with fractal characteristics (Zhang et al., 2017), therefore, the fractal dimension is often used to quantify and characterize the pore structure in coal reservoirs (Wang et al., 2021). In this study, the Frenkel-Halsey-Hill (FHH) fractal model was used to estimate the fractal dimension from the LTNA data. The FHH equation can be expressed as (Lowell et al., 2004; Zhang et al., 2014): ln v v o = C + A [ ln ( ln p o p ) ] (1)   Either  A = ( D − 3 ) / 3 (2) Or  A = D − 3 (3) where v is the sorption capacity; v o denotes the volume of monolayer coverage; D is the fractal dimension; A represents the pre-exponential factor, which is dependent on D; P ₀ is the saturated vapor pressure; C is a constant.

According to Eq. 1, there must be a linear relationship between ln V and ln (ln (P ₀/P)) when the pores have fractal characteristics. As can be seen from Figure 5, the bilogarithmic coordinates show a significant linear relationship, indicating that the pores of coal samples can be characterized by fractal theory. It is noteworthy that there are two different formulae available for calculating D with A based on the different situations: Eq. 2,3 (Pfeifer and Avnir, 1984; Avnir and Jaroniec, 1989; Ismail and Pfeifer, 1994; Yao et al., 2008; Dou et al., 2021). Eq. 2 is applicable to the membrane/gas interface controlled by van der Waals force, and Eq. 3 is applicable to the interface controlled by liquid–gas tension. In this study, many of these D values are less than 2 as calculated using Eq. 2, which is unrealistic, so D is calculated using Eq. 3 (Table 2).

To study the effect of microbial action on pore structure of coal, the data are divided into three regions to calculate the fractal dimension, d ≤ 10 nm, 10 nm < d < 100 nm, d ≥ 100 nm. The fractal results showed that when d ≤ 10 nm, the fractal dimension of coal after microbial treatment was increased from 2.73 to 2.89, when 10 nm < d < 100 nm, the fractal dimension D was increased from 2.60 to 2.81, however, when d ≥ 100 nm, the fractal dimension D was reduced from 2.90 to 2.85, The results indicated that the surface irregularities of mesopores and macropores decreased, the pores became smoother and the pore volume decreased after the bioconversion (Figure 6). This finding was consistent with the aforementioned research (the pore structure evolution model is shown in Figure 7).

From the perspective of industrial analysis (Table 1), compared with pre-test values, the moisture content, ash yield, and clay content of coal samples increased from 2.99%, 13.13 % and 4.13%–5.53%, 15.65 % and 7.37%, respectively, while the volatile matter and fixed carbon contents reduced from 6.68%, 78.63 %–5.09% and 75.63%, respectively. This showed that microorganisms participated in the degradation of coal reservoir and promoted the generation of methane gas.

The curves of the methane adsorption on samples can be drawn according to the Langmuir model Eq. 4: V = v L P P L + P (4) where V is the adsorption volume, V _L denotes the Langmuir volume, P is the equilibrium pressure, P _L is the Langmuir pressure (Zhang et al., 2017; Chen et al., 2018). It can be concluded from Figure 8 that the Langmuir volume was increased after bioconversion, from 34.84 to 36.34 cm³/g (mean average), and the Langmuir pressure was increased from 2.73 to 3.28 MPa (mean average) after bioconversion based on the isothermal methane adsorption experimental data (Table 1). This result showed that transformation of coal by microorganisms is more beneficial to the adsorption and production of CBM.