A mechanistic link between coal pore development and molecular structure evolution

2.1. Coal sampling and analyses

Three coal samples, HM, YM, and WYM, were collected from representative coal basins in China, specifically from the Yanbian Liangshui Coal Mine, Liaoyuan Mining, and Xi’an Tonghua Mining, respectively. The Micro Photometry Version Scanning Photometer (MPV-SP) microphotometer (Germany) instrument (calibrated using Sapphire 0.59 and Gadolinium Gallium Garnet 1.75 standards) was employed to measure the vitrinite reflectance of 1-2 cm coal samples, with measurements taken at a minimum of 50 points per sample (Table 1). To determine the molecular structure parameters of the coal, the samples were ground to a 200-mesh size for subsequent characterization using ¹³C-NMR and FTIR analyses. Elemental composition was assessed using a Thermo Scientific elemental analyzer (Table 1), and all related analyses were conducted at the Research Dog testing platform. Surface morphology of the coal samples was examined using a Quanta 250 FEG field emission scanning electron microscope (FE-SEM). In addition, to assess pore structure and distribution characteristics, low-temperature nitrogen and carbon dioxide adsorption experiments were conducted using an Autosorb iQ Specific Surface Area Analyzer. These experiments were performed at Heilongjiang University of Science and Technology.

2.2. Determination of coal structural parameters

2.2.1. Parameters of coal molecular structure

(1) ¹³C-NMR experiment

¹³C-NMR measurements were carried out using a Bruker Avance III 600 MHz NMR spectrometer. Prior to testing, coal samples were crushed to pass through a 200-mesh sieve. The experiment utilized a 4 mm MAS probe with a rotational speed of 10 kHz, a contact time of 3 ms, and a cycle delay of 3 s.

(2) FTIR experiments

Infrared spectroscopy was performed using a Nicolet iS10 FTIR spectrometer. The potassium bromide (KBr) pellet method was employed; specifically, coal samples were mixed with KBr in specific proportions and pressed into pellets. The spectra were recorded in the 400–4000 cm⁻¹ range with 32 scans at a resolution of 4 cm⁻¹. Prior to analysis, coal powders (<200 mesh) were pre-treated using the piezometric method. These data were used for 3D pore modeling (Figure 1).

Close

Figure 1.

Schematic diagram of building a 3D pore model.

Export to PPT

2.3. Simulation

By fitting the peaks in the ¹³C-NMR and FTIR spectra of each sample and analyzing the subpeak assignments, as well as calculating the area and proportion of each peak component, the relative contents and distributions of carbon and oxygen-containing structural units were determined. The coal macromolecular structure was initially drawn using ACD/ChemSketch software. Subsequently, gNMR software was used to calculate the chemical shifts of carbon atoms. The simulation was iteratively refined and compared with the experimental spectra until the simulated mapping closely matched the experimental results, allowing for the identification of the most representative 2D molecular structure model of coal.

Following this, a 3D ball-and-stick model of the coal molecule was developed using Chem3D software. Energy minimization, geometry optimization, and annealing processes were then performed using molecular mechanics and molecular dynamics approaches within the Materials Studio (MS) software to obtain a stable configuration with the lowest energy. The Condensed-phase optimized molecular potential for atomistic simulation study (COMPASS) III force field was applied during these calculations. Pore corrections were conducted using the AtomVolumes & Surfaces module, and density corrections were performed with the Amorphous Cell Calculation module. Based on the micropore volumes obtained from LTCDA, 3D pore models for the HM, YM, and WYM samples were refined. Finally, the number of molecular units for each sample was established, and a 3D molecular model was constructed for each coal sample.

3.1. Fundamental characteristics of coal samples

The vitrinite reflectance (R_o) and average vitrinite reflectance values for each sample are presented below. For HM, R_o ranged from 0.22% to 0.60%, with an average of 0.52%. For YM, the range was 0.93% to 1.17%, with an average of 1.10%. For WYM, R_o ranged from 2.25% to 2.49%, with an average of 2.41%. The ash content of HM, YM, and WYM was 16.87%, 25.57%, and 12.81%, respectively. The volatile matter content was 35.73%, 28.27%, and 5.69%, and the moisture content was 11.02%, 4.10%, and 3.29%, respectively (Table 1). Based on these values, HM can be classified as a high-volatile, medium ash lignite; YM as a high-volatile, medium ash bituminous coal; and WYM as a low-volatile, low ash anthracite.

Elemental analysis was conducted to determine the basic composition of each coal sample, thereby providing a basis for constructing the coal molecular models. The carbon (C) content for HM, YM, and WYM was 88.91%, 90.26%, and 90.71%, respectively. The contents of hydrogen (H), oxygen (O), nitrogen (N), and sulfur (S_td) were relatively low. In particular, sulfur content was below 1% in all samples, indicating that each was a low-sulfur coal. Thus, the influence of sulfur on coal molecular modeling can be considered negligible. After normalizing the contents of C, H, O, and N, the atomic ratio analysis showed that the H/C ratio initially increased and then decreased with coal rank, while both O/C and N/C ratios generally decreased as coal rank increased (Table 2).

3.2. FTIR analysis

The FTIR spectra can be divided into four distinct regions: the hydroxyl absorption region (3600–2800 cm^-1), the aliphatic hydrocarbons absorption region (3000–2800 cm^-1), the oxygen functional group region (1800–1000 cm^-1), and the aromatic hydrocarbons region (900–700 cm^-1) [14] (Figure 2). PeakFit 4.12 software was used for peak fitting and spectral deconvolution, allowing for the extraction of data corresponding to various functional groups Figure 3(a-c) FTIR spectra at 1800-900 cm^-1 for HM, YM and WYM, respectively; Figure 3(d-f) FTLR spectra at 2800-3600 cm^-1 for HM, YM and WYM, respectively.

Close

Figure 2.

FTIR spectra of coal samples (a) HM, Ro,avg = 0.52%; (b) YM, Ro,avg = 1.1%; (c) WYM Ro,avg = 2.41%. The integrated intensities of the characteristic peaks, quantified as baseline-corrected absorbance (a.u.) on the Y-axis.

Export to PPT

Close

Figure 3.

FTIR split-peak fitting profiles of coal samples (a) HM, Ro,avg = 0.52%; (b) YM, Ro,avg =1.1%; (c) WYM, Ro,avg = 2.41%; (d) HM, Ro,avg = 0.52%; (e) YM, Ro,avg =1.1%; (f) WYM, Ro,avg = 2.41%. The integrated intensities of the characteristic peaks, quantified as baseline-corrected absorbance (a.u) on the Y-axis.

Export to PPT

As coal rank increased, the absorption peaks corresponding to aliphatic hydrocarbons and oxygen-containing functional groups progressively weakened and eventually disappeared, indicating a decrease in aliphatic hydrocarbon content [14]. Prominent absorption bands in the 3600–3100 cm⁻¹ range, observed in HM, YM, and WYM, correspond to hydroxyl groups, confirming their presence in all three samples [15]. The region between 3000 and 2800 cm⁻¹ was attributed to C–H stretching vibrations of aliphatic structures [16]. Notable absorption peaks were detected at 1601.35 cm^-1, 1439.19 cm^-1, and 1030 cm⁻¹, corresponding to the stretching vibrations of aromatic ring C = O bonds, aliphatic hydrocarbons CH₃–CH₂ groups, and alkyl ether C–O groups, respectively [17].

Oxygen-containing functional groups in coal primarily include carboxyl, hydroxyl, and carbonyl groups [18]. The oxygen-containing functionalities in HM, YM, and WYM were predominantly hydroxyl and ether bonds. Ether groups were present at 20.06%, 45.22%, and 16.60% in HM, YM, and WYM, respectively. The hydroxyl content decreased with increasing coal rank, with values of 3.12%, 1.53%, and 0.41%, respectively. This reduction is attributed to dehydrogenation and oxidation processes occurring during coalification [19,20].

3.3. ¹³C-NMR analysis

The ¹³C-NMR spectra were analyzed by classifying peak positions and intensities in conjunction with chemical shift values. The spectra Figure 4(a-c) can be divided into two regions: the aromatic carbon region (90–165 ppm) and the aliphatic carbon region (0–90 ppm) [21]. No carbonyl signals were observed in any of the samples, indicating that the carbonyl content was minimal. HM exhibited distinct peaks only in the aliphatic carbon region, suggesting that its molecular structure is primarily composed of aliphatic components. In contrast, both YM and WYM showed high-intensity signals in the aromatic carbon region. YM and WYM showed high-intensity signals in the aromatic carbon region. Additionally, YM exhibited significant intensity in the aliphatic carbon region as well, indicating that its molecular structure includes both aliphatic and aromatic features. WYM, however, is dominated by aromatic structures. These findings demonstrate that the ¹³C-NMR spectra are consistent with the results of the FTIR analysis.

Close

Figure 4.

¹³C-NMR spectra of coal samples (a) HM, Ro,avg = 0.52%; (b) YM, Ro,avg = 1.1%; (c) WYM, Ro,avg = 2.41%). The Y-axis shows normalized, baseline-corrected intensity (arbitrary units) for comparing spectral shapes and relative changes.

Export to PPT

To obtain further insights into the carbon skeleton structure, twelve structural parameters were calculated for each sample. Additionally, the aromatic bridge-to-periphery carbon ratio (X_BP) was computed using Eq. (4) (Table 3).

To further investigate the carbon skeleton structure of the coal samples, the aromatic size and degree of condensation were quantified using the aromatic bridge-to-periphery carbon ratio (X_BP) [17]. The X_BP values for HM, YM, and WYM were 0.13, 0.24, and 0.14, respectively. A positive correlation (R = 0.50) was observed between R_o,avg, and X_BP, supporting the conclusion that the proportion of aromatic structures in the coal molecular framework increases progressively with coal rank.

When integrated with existing research [21], the relationship between R_o,avg, and various structural parameters was systematically analyzed. The results revealed strong correlations between R_o,avg, and several key structural parameter (Figure 5a-j). Alkyl-substituted aromatic carbon (f_a^S) exhibited a negative correlation with R_o,avg (R = –0.27), indicating that alkyl side chains attached to aromatic rings tend to diminish as coal rank increases [21]. In contrast, positive correlations were observed between R_o,avg, and the fractions of aromatic bridge carbon (f_a^B, R = 0.61), aromatic carbon (f_a′, R = 0.53), and unsaturated carbon (f_a, R = 0.79), the latter referring to carbon atoms with double or triple bonds. These results suggest a progressive enrichment in aromatic content with increasing coal rank. Methylene carbon (f_al^H) displayed a strong negative correlation with R_o,avg (R = –0.85), supporting the view that aromatization during coalification reduces methylene content. Aromatic carbon in coal consists of both protonated and non-protonated types [22]. However, non-protonated aromatic carbon (f_a^N) and methyl carbon (f_al^*) showed no significant correlation with R_o,avg. A strong negative correlation was also found between R_o,avg, and aliphatic carbon (f_al) (R = –0.83), which is consistent with the progressive reduction in aliphatic content and increase in aromatic structures due to arylation and condensation reactions occurring during coalification.

Close

Figure 5.

The three orange points represent coal samples HM, YM, and WYM. The four purple points were adapted from Liu Yu’s study [21], corresponding to coal samples YL-1, TL-1, XZ-1, and JH-1. These seven data points were used to compare their structural parameters with the respective vitrinite reflectance values. Correlation of R_o,avg with (a) f_a^s, (b) f_a^B, (c) f_a^H, (d) f_al^H, (e) f_al^*, (f) f_a^N, (g) f_a^’, (h) f_a, (i) f_al, and (j) X_BP.

Export to PPT

3.4. Pore structure characterization

3.4.1. Microporous pore characteristics

The CO₂ adsorption curves of each sample were characterized by type I isothermal adsorption behavior, in which gas adsorption increased rapidly at low relative pressure (P/P₀) and then gradually plateaued (Figure 6a). All samples exhibited asymmetric “C”-shaped hysteresis loops. Combined with SEM observations, this pattern confirmed the presence of “ink-bottle” type pores in YM. During adsorption, gas molecules diffused into wider pore cavities, whereas during desorption, narrower pore necks required lower pressures for capillary evaporation. This discrepancy led to the delayed closure of the desorption branch, producing the characteristic “C”-shaped hysteresis [23]. At all pressure points, YM exhibited greater gas adsorption than HM and WYM (Figure 6b). The pore volume of the samples ranged from 0.007 to 0.024 cm³g^-1, with an average of 0.013 cm³ g^-1, and the specific surface area ranged from 18.667 to 67.021 m² g^-1, with an average of 36.044 m² g¹. Pore size distributions were multimodal, primarily concentrated in the 0.46–0.65 nm and 0.71–0.94 nm ranges (Figure 6c). More abundant pore volume and number were observed in the 0.46–0.65 nm range. Based on the pore volume and pore size distribution, it can be concluded that YM contains a substantially higher number of micropores than HM and WYM. This was attributed to the release of significant quantities of volatile components (e.g., CH₄, CO₂) and the gas escape path during coalification, forming complex micropore networks consistent with the “ink-bottle” morphology observed in SEM images [24].

Close

Figure 6.

(a) CO₂ adsorplion isotherms, (b and c) pore size distributions; (d) N₂ adsorption isotherms, (e and f) pore size distributions; SEM pore images of (g) HM, (h) YM, and (i) WYM.

Export to PPT

3.5. Modeling the 2D/3D structure of coal molecules

Elemental analysis (Table 2) was used to quantify the relative proportions of C, H, and O in the coal samples. The molecular weight of coal macromolecules was found to range from 2000 to 5000 [26]. Based on the elemental composition of the coal sample, the following equation was derived using atomic masses of constituent elements (C:12, H:1, O:16, N:14): 12a + x₁a + 16×x₂a + 14×x₃a = 2600, where a represents the number of carbon atoms, and x₁ (H/C), x₂ (O/C), and x₃ (N/C) denote the respective atomic ratios. After several iterations and adjustments, the molecular formulas were determined as follows: HM – C₁₈₇H₁₃₃O₇N₃, YM – C₁₉₆H₁₄₅O₄N₃, and WYM – C₁₉₇H₁₃₈O₄N₂. These formulations were selected to generate structurally representative coal molecular models. The predicted ¹³C-NMR spectra showed good agreement with the experimental spectra, validating the accuracy of the molecular models (Figure 7).

Close

Figure 7.

¹³C-NMR fitting spectra, 2D molecular structures, and simulation validation for coal samples (a) HM, Ro,avg = 0.52%; (b) YM, Ro,avg = 1.1%; (c) WYM, Ro,avg = 2.41%). The Y-axis shows normalized, baseline-corrected intensity (arbitrary units) for comparing spectral shapes and relative changes.

Export to PPT

Based on the ¹³C-NMR structural parameters (Table 4), and using Eq. (5-11), six key parameters were determined: the number of protonated aromatic carbon atoms (F_a^H), aromatic bridge carbon atoms (F_a^B), alkyl-substituted aromatic carbon atoms (F_a^S), phenolic hydroxyl or ether oxygen-bonded carbon atoms (F_a^P), total aromatic carbon atoms (F_a′), and carbonyl and carboxyl carbon atoms (F_a^C) [27].

The results from the split-peak fitting of the ¹³C-NMR and FTIR spectra were used to determine the relative contents of aromatic carbon and nitrogen-containing structural units, enabling the identification of aromatic ring types and their quantities. The X_BP values for HM, YM, and WYM were 0.13, 0.24, and 0.14, respectively. For reference, a coal sample with an X_BP of 0.25 typically contains two aromatic rings [17]. The aromatic structures of the coal samples were predominantly composed of benzene and naphthalene rings, with anthracene and pyrene rings also present (Table 5). Among the nitrogen heterocycles, pyridine was dominant in HM and YM, whereas pyrroles were the major nitrogen structures in WYM [28]. The X_BP^⁎ values calculated from the molecular models of HM, YM, and WYM were 0.13, 0.26, and 0.18, respectively.

Table 5. Types and quantities of different structural units.

Structure
Name	Benzene	Napthalene	Anthracene	Pyrene	Pyridine	Pyrrole
Code name	X1	X2	X3	X4	A	B
HM	4	4	1	0	2	1
YM	1	7	2	2	2	1
WYM	9	3	2	2	0	2

The internal pore structure of coal can be investigated through its molecular configuration. Using ACD/ChemSketch and gNMR software, 2D molecular structure models of the coal samples were constructed. Simulated ¹³C-NMR spectra were generated and compared with experimental spectra, showing good agreement. The reliability of the molecular models was confirmed by comparing their calculated structural parameters with experimental data (Figure 7). Figure 7 presents a combined spectral, structural, and analytical validation for each coal sample. Panel a (HM), b(YM), and c (WYM) each contain the experimental and fitted ¹³C-NMR spectrum, the proposed 2D molecular structure, and the corresponding simulation fit validation. The relative errors between simulated and experimental values were all below 15% (Table 6), indicating that the 2D molecular models accurately represented the actual structure and could be used to assess model reliability. The 2D molecular structure models of HM, YM, and WYM (Figure 7) were finally determined by the simulated ¹³C-NMR spectra and experimental error validation analysis, which best fit the real coal molecular model.

3D modeling of the coal molecular structures was conducted using Materials Studio software. Energy minimization, geometry optimization, and annealing were performed using a molecular dynamics approach to identify the most stable configuration with the lowest energy (Figure 8). Figure 8 3D structural models for samples HM (a), YM (b), and WYM (c). Each panel shows the ball-and-stick molecular model, the bulk simulation box model, and the associated microporous features. After density calibration, 1.5 g cm^-3 for HM, 1.1 g cm^-3 for YM, and 1.25 g cm^-3 for WYM, the number of representative molecules in each sample was found to be 3, 6, and 4, respectively. The final 3D molecular structures were constructed for HM, YM, and WYM (Figure 8). The dimensions of the modeling cells were 2.0 × 2.0 × 2.0 nm³ for HM, 2.9 × 2.9 × 2.9 nm³ for YM, and 2.4 × 2.4 × 2.4 nm³ for WYM.

Close

Figure 8.

3D ball-and-stick model of HM, YM, WYM and micropore characteristics in coal macromolecules (a) HM, Ro,avg = 0.52%; (b) YM, Ro,avg = 1.1%; (c) WYM, Ro,avg = 2.41%).

Export to PPT

To further validate the accuracy of the 3D molecular models, internal pore characteristics were evaluated using the probe method, with a probe radius of 0.165 nm (corresponding to the molecular dynamics radius of CO₂). The measured pore volumes for HM, YM, and WYM using this method were 0.007 cm³ g^-1, 0.024 cm³ g^-1, and 0.009 cm³ g^-1, respectively. These values closely matched the experimental data, 0.008 cm³ g^-1, 0.020 cm³ g^-1, and 0.010 cm³ g^-1, demonstrating that the 3D molecular models accurately reflect the real microporous structures of the coal samples. The molecular structure of the three coal samples was optimized to yield the lowest-energy 3D ball-and-stick model (Figure 8). By density correction and pore correction, both show good consistency, and finally determine the box model, using the Connolly surface module in the AtomVolumes & Surfaces module to calculate the micro-pore characteristics of coal macromolecules (Figure 8).

3.6. Mechanisms of pore control by coal thermal evolution

The pore structure characteristics of coal are significantly influenced by coal rank, with pore size distribution and the degree of pore development closely associated with the stage of thermal evolution [29]. Previous studies have demonstrated that coal pores predominantly exist at the nanoscale, comprising micropores and mesopores. Among these, micropores contribute most significantly to both pore volume and specific surface area [30-33]. As shown in Figure 9(a,b), the specific surface area and pore volume distributions of micropores and mesopores in HM, YM, and WYM were measured using LTNA and LTCDA. Across all ranks, micropores accounted for the largest proportion of both specific surface area and pore volume, followed by mesopores. Notably, YM exhibited a significantly higher micropore-specific surface area and pore volume than HM and WYM. In contrast, WYM showed the highest proportion of macropore and mesopore-specific surface area and pore volume. This results in the presence of mesopores and macropores but relatively few micropores [34], consistent with SEM observations. For medium-rank coal (YM), the aromatic lamellae begin to align preferentially, forming a more compact cross-linked structure. Thermal cleavage of residual organic matter at elevated temperatures releases gases, forming a substantial number of new micropores, often in the form of “ink-bottle” pores. Consequently, medium and large pore development is relatively limited in YM [29]. In high-rank coals (WYM), the pore system undergoes significant reorganization during advanced coalification under elevated temperature and pressure. Intense compaction and polycondensation lead to a notable reduction in micropore volume. Concurrently, aromatic layers experience further polycondensation and adopt a highly stacked, preferentially oriented arrangement. This structural evolution promotes the development of numerous slit-shaped pores and interlayer voids, which are primarily governed by the stacking morphology of aromatic clusters. As a result, WYM exhibits the highest proportion of mesopores in terms of both specific surface area and pore volume among the studied coal ranks [12].

Close

Figure 9.

(a) specific surface area distribution of coal sample pores; (b) Pore volume distribution of coal sample pores. (HM, Ro,avg = 0.52%; YM, Ro,avg = 1.1%; WYM, Ro,avg = 2.41%). The X-axis shows the samples investigated.

Export to PPT

3.7. Fractal dimension

Fractal theory provides a framework to describe the structure of complex natural systems [8]. The fractal dimension D is used to characterize the irregularity and heterogeneity of pore networks. For porous rocks, D typically ranges between 0 and 3. A lower fractal dimension indicates greater uniformity (homogeneity), while a value approaching 3 reflects increased structural complexity and heterogeneity in the pore system [13,35]. Figure 10 presents the fractal dimensionsderived from the LgS–Lgr plots for the HM, YM, and WYM samples. All three samples exhibit a distinct two-segment fractal characteristic across the analyzed pore size ranges (0–2 nm and 2–50 nm). Within the 0–2 nm scale, the calculated fractal dimensions are 1.075 for HM, 1.699 for YM, and 1.380 for WYM. This sequence indicates an initial increase followed by a subsequent decrease with advancing coal rank. These results suggest that HM exhibits a relatively uniform microporous structure. Combined with SEM and LTCDA findings, this supports the conclusion that low-rank coal (HM) develops primarily natural micropores with simple, homogeneous structures due to its relatively mild coalification environment. For YM, the moderate fractal dimension corresponds to an intermediate coalification stage, during which gas release and solubilization generate solubilized organic micropores, adding to the overall structural intricacy. In contrast, WYM, a high-rank coal, contains partially sealed organic pores, some of which are linked via microfractures, thereby increasing micropore complexity. In high-rank coal (WYM), polycondensation and structural ordering drive the partial closure of organic pores, which contributes to a more homogeneous and simplified microporous morphology.The fractal dimensions of the 2–50 nm mesopores for HM, YM, and WYM were 2.212, 1.889, and 1.717, respectively—demonstrating a decrease (Figure 10a-c). This trend suggests that mesopore structures become progressively homogenized, especially in high-rank coal. The fractal dimension of WYM and YM samples being less than 2 has limited significance for pore characterization, primarily indicating simpler pore structures compared to HM. Thermal evolution analysis reveals that in WYM coal, molecular structure undergoes further condensation, stacking, and oriented arrangement, forming relatively uniform micro-fractures that result in lower fractal dimensions. The contrasting fractal dimensions between HM and YM coals signify distinct pore evolution mechanisms. The higher value in HM signifies pore heterogeneity and disorder, whereas the lower value in YM stems from the organized formation of regular pore clustersvia thermal decomposition. These mechanisms are corroborated by integrated evidence from SEM imaging and LTNA data.

Close

Figure 10.

Fractal characteristics of coal samples (0–50 nm). (2–50 nm fractal characteristics: (a) HM; (b) YM; (c) WYM; 0–2 nm: (d) HM; (e) YM; (f) WYM).

Export to PPT

3.8. Mechanisms of pore structure formation

3.8.1. Pore wall structure and pore connectivity

The internal pore architecture of coal offers insights into its molecular structure. To investigate the pore formation mechanisms across coal ranks, pore slicing was performed on HM, YM, and WYM samples. As illustrated in Figure 11, the micropore morphology and connectivity varied notably by coal rank. The molecular structures exhibited both relatively isolated micropores and intricate microporous networks formed through pore interconnection. Microstructural analysis revealed that the YM sample exhibited the highest abundance of micropores with favorable interconnectivity among the studied samples. In contrast, WYM showed reduced micropore quantity and interconnectivity compared to YM. HM exhibited the lowest level of micropore development and the poorest connectivity. Further evidence indicates that the HM, high moisture and volatile content, along with a loosely organized structure, led to the development of some mesopores, but few micropores. The high micropore connectivity in YM is attributed to thermal cracking of residual organic matter at elevated temperatures, leading to gas release and formation of well-connected micropores. With increasing coalification, the aromatic rings in the WYM progressively expanded and condensed into larger aromatic structural units. These units became more densely and orderly arranged, thereby restricting the connectivity between micropores and resulting in relatively isolated micropores. Atomic characterization of the HM pore walls revealed that hydrogen atoms were the most common atoms in direct contact with the pore wall, with only a small number of oxygen atoms present. Carbon atoms typically did not contact the pore wall directly. With increasing coal rank, aromatic rings become the principal framework for micropore walls. This further illustrates the mechanism of micropore formation and supports the conclusion that reduced micropore connectivity is associated with increased aromatic condensation in the coal molecular structure.

Close

Figure 11.

Pore wall evolution in coal samples with increasing maturity. (a) HM (Ro,avg = 0.52%), main pore image. (b, c) HM, detailed cross-sectional slices at positions 1 and 2. (d) YM (Ro,avg = 1.1%), main pore image. (e, f) YM, cross-sectional slices at positions 1 and 2. (g) WYM (Ro,avg = 2.41%), main pore image. (h, i) WYM, cross-sectional slices at positions 1 and 2.

Export to PPT

3.8.2. Mechanisms of pore control by structural parameters

The development of micropores in coal is fundamentally governed by its aromatic molecular structure [36,37]. Multiscale characterization methods have revealed a size-dependent relationship between nanopore formation and the size of aromatic ring structures, indicating that different pore size ranges are controlled by specific aromatic configurations [12]. Therefore, the structural dependence of pore formation was further explored by analyzing the correlations between ten structural parameters, vitrinite reflectance, and pore size data (Figure 12a-v).

Close

Figure 12.

The three orange points represent HM, YM, and WYM samples; the four purple points are derived from Liu Yu’s study [21]. Correlation thermograms were generated for the seven data points against vitrinite reflectance and pore size intervals (0–2 nm, 2–50 nm), and the ten structural parameters for coal samples YL-1, TL-1, XZ-1, and JH-1 were determined. Correlation of pore volume with structural parameters: (a–g) Pore volume (0–2 nm) versus (a) f_a^s, (b) f_a^B, (c) f_a^H, (d) f_al^H, (e) f_al^*, (f) f_a^N, (g) f_a^’, (h) f_a, (i) f_al, (j) X_BP, (k) R_o,avg; (l–v) Pore volume (2–50 nm) versus (l) f_a^s, (m) f_a^B, (n) f_a^H, (o) f_al^H, (p) f_al^*, (q) f_a^N, (r) f_a^’, (s) f_a, (t) f_al, (u) X_BP, (v) R_o,avg.

Export to PPT

Micropore volume exhibited a strong positive correlation with aromatic bridge carbon (f_a^B, R = 0.70) and aromatic bridge-to-periphery ratio (X_BP, R = 0.80), indicating that X_BP plays a dominant role in micropore formation (Figure 12b,j). Additionally, microporosity correlated positively with unsaturated carbon (f_a, R = 0.42), suggesting that demethylation and aromatization, transforming saturated into unsaturated carbon, may facilitate micropore development (Figure 12h).

Mesopores also showed a positive correlation with X_BP (R = 0.38), suggesting that aromatic crosslinking has a contributing role in mesopore formation (Figure 12u). A highly significant positive correlation was found between mesopore volume and methyl carbon (f_al*, R = 0.92) (Figure 12p). During pyrolysis, escaping gases generate internal channels, and the limited diffusion due to aromatic skeleton rigidity causes internal pressure buildup. This can result in structural collapse and formation of mesopores [38]. The increase in f_al^* enhances structural disorder, further promoting mesopore development. Conversely, mesopores showed a significant negative correlation with aromatic carbon (f_a′, R = –0.39) (Figure 12r). As aromatization progresses, particularly in YM and WYM, higher f_a′ content leads to tightly stacked aromatic layers with aligned molecular orientation, potentially closing mesopores through extrusion [39,40].