Thermophysical Response Patterns of Pore Structures in Coals of Different Ranks Authors: Chen Guo, Xinhang Yu, Jinxiao Yang, Junzhe Gao, Jiang Gou, Xiankuo Yang Published: 2026 ABSTRACT Understanding the differential evolution characteristics of pore structures in coals of various ranks under heat treatment conditions is fundamental for developing coal thermal mining technologies. This study systematically collected coal samples from the Zhangjiamao Mine in the Ordos Basin, the Dahebian Mine in western Guizhou, and the Sihe Mine in the Qinshui Basin, representing a series of samples with different coal ranks. Through drying, saturation, and heating processes, combined with low-field nuclear magnetic resonance (NMR) testing and pore fractal geometry theory, the effects of heat treatment at 200 °C, 300 °C, and 400 °C on NMR T 2 spectra, pore size distribution, two-dimensional relaxation distribution, and pore fractal dimensions of coal samples were investigated. The results indicate that after heat treatment, the porosity of coal samples across all coal ranks increased, with low-rank coal exhibiting the greatest increase, suggesting that low-rank coal pore structures possess the strongest temperature sensitivity. When the temperature reached 300 °C, the porosity of high-rank coal began to increase significantly, suggesting that its pore structure requires a higher activation temperature for thermal evolution. The DHB coal sample exhibited caking properties, and significant thermal expansion of the coal matrix after heating led to a smaller porosity at 300 °C compared to ZJM and SH samples. Heat treatment promoted the transformation of micropores into mesopores and macropores. Low-rank coal was characterized by an increase in mesopore volume, the pore connectivity of medium-rank coal was optimized, and the macropore volume of high-rank coal increased markedly with rising temperature. After heat treatment, the fractal dimensions of mesopores and macropores generally decreased, indicating reduced pore surface irregularity and enhanced connectivity, which facilitates fluid seepage. This study reveals the thermophysical differential response patterns of pores in coals of different ranks. Heat treatment affects coal pores though expansion, increase in number, and improved connectivity, creating favorable conditions for the fluid product during thermal extraction. The research findings provide theoretical and parametric support for geological site selection and process control of thermal mining, thereby promoting the development of thermal extraction technologies for coal resources. Introduction Thermal recovery of coal resources has broad application prospects for achieving clean and efficient utilization of coal. The pore structure of coal determines the seepage conditions and production efficiency of pyrolysis fluids. The evolution of pores during pyrolysissuch as expansion, interconnection, merging, or the formation of new poresrestructures the pore system, thereby influencing the production process and efficiency. Pore evolution behaviors vary significantly at different temperatures: at low temperatures, slight expansion of micropores and small pores predominates, while at medium to high temperatures, the massive release of volatiles leads to the formation of more interconnected pores and even macroscopic fractures ultimately affecting thermal extraction efficiency. Furthermore, due to differences in pore structure, mechanical properties, and material composition among coals of different ranks, the evolution of pore structure with temperature during heat treatment follows distinct patterns. Recent investigations of thermally stressed porous media have shown that elevated temperatures significantly modify pore-scale transport properties and connectivity patterns under sustained thermal gradients. This finding further underscores the importance of conducting in-depth investigations into the evolution patterns of coal pore structures under complex thermodynamic conditions. Therefore, uncovering the differential responses of pore structures in coals of varying ranks under thermal treatment conditions has become a core scientific issue in advancing clean and efficient thermal extraction technologies. Traditional methods for characterizing coal pore structure, such as scanning electron microscopy, mercury intrusion porosimetry, and gas adsorption, can achieve static characterization but exhibit significant limitations. Scanning electron microscopy struggles to capture the dynamic evolution of pores, mercury intrusion may cause micropore damage due to high-pressure intrusion, and gas adsorption fails to simultaneously reflect fluid mobility and pore connectivity. , Consequently, these methods are inadequate for accurately characterizing the dynamic evolution of pores under varying thermal conditions. In recent years, low-field nuclear magnetic resonance (NMR) technology has been widely applied in unconventional oil and gas accumulation and development research due to its advantages of being rapid, nondestructive, and capable of real-time monitoring, providing a technical means to reveal the dynamic evolution of coal pores during thermal treatment. Yao et al. identified a strong correlation between the transverse relaxation time T 2 of coal and its pore structure, establishing a pore distribution calculation model based on T 2 spectra; Mandelbrot introduced topological fractal theory, offering an effective tool for describing rock physical properties via NMR methods; , Zhou et al. further revealed the relationship between the fractal dimension of pore structure and seepage space, providing a basis for evaluating permeability; Zhao et al. derived a fractal geometric expression for NMR T 2 values; Lin et al. combined NMR technology with fractal analysis to investigate the impact of thermal treatment on coal pore properties. Recent data-driven approaches have further emphasized the importance of integrating microstructural characterization with predictive modeling frameworks in clay-dominated porous systems. This suggests that for coal, as a porous medium, a single characterization method or static analysis is no longer sufficient to meet the research needs of complex pore evolution. There is an urgent need to integrate multiscale experimental data with intelligent prediction models to deeply reveal the pore structure response mechanisms of different coal ranks under heat treatment conditions. Although existing studies have preliminarily explored the impact of thermal treatment on coal pores, they have predominantly focused on a single coal rank or a single temperature, failing to systematically reveal the differential responses of pore structures in coals of different ranks under varying temperatures. Therefore, systematically investigating the pore evolution patterns of different coal ranks at various pyrolysis temperatures, and uncovering the transformation of pore types, changes in porosity, and evolution of pore fractal dimensions under thermal treatment, are urgent needs for refining the theory of thermal evolution of coal pores and promoting the fluidized extraction of coal resources. Based on this, this study takes low-, medium-, and high-rank coals as research objects, employs thermal treatment as the primary method, and utilizes NMR technology to examine the T 2 spectrum characteristics, pore size distribution, two-dimensional relaxation distribution, and fractal dimension variations of coal samples at 200 °C, 300 °C, and 400 °C. The research aims to reveal the thermophysical response patterns and mechanisms of pore structures in different coal ranks, providing precise parameter support and a theoretical basis for the effective production of pyrolysis fluids, thereby advancing the development of thermal mining technologies for coal resources. Samples and Methods Coal Sample Collection and Preparation The experimental coal samples were collected from the Zhangjiamao Coal Mine in the northern Ordos Basin, the Dahebian Coal Mine in western Guizhou, and the Sihe Coal Mine in the southern Qinshui Basin. The primary reasons for selecting these areas are as follows: (1) The coal ranks in these three regions differ significantly, and their coal characteristics are highly representative. (2) These areas are all coal-rich regions in China with substantial resource reserves and typical occurrence conditions, making them of significant engineering application value for studying the thermal mining of coal resources. The samples from these three regions were designated as ZJM, DHB, and SH, respectively. From each region, three cylindrical plug samples with a diameter of 2.5 cm and a length of 2.5–5 cm were prepared. Taking the ZJM coal sample as an example, the three plug samples were labeled ZJM-1, ZJM-2, and ZJM-3 for heat treatment and nuclear magnetic resonance (NMR) experiments at 200 °C, 300 °C, and 400 °C, respectively. The experimental temperatures were selected based on the following considerations: 400 °C corresponds to the pyrolysis temperature range for medium- to low-rank coals, at which coal tar is abundantly generated, marking the threshold for in situ pyrolysis of tar-rich coal. Higher temperatures could induce significant melting in caking bituminous coals of medium rank, potentially interfering with measurements such as NMR. Additionally, due to the limited scope of this study, it is impractical to simulate all temperature points. Therefore, this study focuses on heat treatment simulations at temperatures between 200 and 400 °C. The pulverized coal generated during sample preparation was sieved to specified particle sizes for proximate analysis, maceral composition identification, and maximum vitrinite reflectance measurement. The test results indicate that the ZJM, DHB, and SH coal samples correspond to low-rank, medium-rank, and high-rank coal, respectively. The ZJM sample exhibits high inertinite and moisture content, along with low mineral content and ash yield. The SH sample shows high vitrinite and fixed carbon content. The DHB coal sample has the highest ash yield and volatile matter yield, which is associated with its high mineral and liptinite content, respectively (Table ). The DHB sample has the highest mineral content and is mainly composed of clay minerals and quartz, The specific experimental procedure is illustrated in Figure . Experimental flowchart. 1Maximum Vitrinite Reflectance, Maceral Composition, and Proximate Analysis Results of Coal Samples proximate analysis/%maceral analysis/%sample origin R o,max/% V daf M ad A d FCd vitriniteinertinitelipidomemineralsZJM0.5238.686.169.1855.6953.4640.090.465.99DHB0.8142.502.5019.4146.3450.9423.367.0118.69SH3.0513.793.1015.9172.4971.0816.670.0012.25aNote: R o,max, maximum reflectance of vitrinite; V daf, volatile matter, dry ash-free basis; M ad, moisture, air-drying basis; A d, ash yield, air-drying basis; FCd, fixed carbon content, air-drying basis. Maximum Vitrinite Reflectance, Maceral Composition, and Proximate Analysis Results of Coal Samples Note: R o,max, maximum reflectance of vitrinite; V daf, volatile matter, dry ash-free basis; M ad, moisture, air-drying basis; A d, ash yield, air-drying basis; FCd, fixed carbon content, air-drying basis. NMR Experiments By analyzing the relaxation characteristics of nuclear magnetic signals, multidimensional information on the pore structure can be obtained. When a core sample is placed in a uniform static magnetic field (B0), the hydrogen-rich nuclei in the fluid within the core interact with the magnetic field, generating a magnetization vector (MZ). At this point, applying a radiofrequency pulse (B 1) at the Larmor frequency perpendicular to the static magnetic field direction induces a nuclear magnetic resonance signal, as illustrated in Figure . This signal contains information such as frequency, phase, and amplitude. The collected nuclear magnetic signals can be inverted to obtain the T 2 spectrum of the sample. Since there is a corresponding relationship between the T 2 spectrum and the pore structure, parameters such as porosity, permeability, movable fluid saturation, irreducible fluid saturation, and pore size distribution can be derived from the T 2 spectrum. (a–d) Evolution mechanism of the macroscopic magnetic field vector in a magnetic field. Adapted with permission from Tong, 2025. T 2 relaxation time is influenced by several factors, and its expression is given by 1 1T2=1T2S+1T2B+1T2D where T 2S is the surface relaxation time; T 2B is the bulk relaxation time; T 2D is the diffusion relaxation time. The transverse relaxation time is typically dominated by the surface relaxation time; therefore, the T 2 relaxation time can be approximately expressed as 2 1T2≈1T2S=ρ2SV=ρ2FSr where ρ2 is the transverse surface relaxivity; S is the pore surface area; V is the pore volume; r is the pore radius; and F S is the geometric shape factor (F S = 3 for spherical pores; F S = 2 for cylindrical pores). As shown in eq , T 2 is proportional to the pore radius r, meaning that a larger pore radius results in a longer relaxation time. The surface relaxivity ρ2 is a critical parameter for converting nuclear magnetic resonance T2 spectra into pore size distributions, and its value directly affects the accuracy of quantitative pore size results. In this study, a combination of mercury intrusion porosimetry (MIP) and NMR technology was employed to determine ρ2. The pore-specific surface area of coal samples was obtained through MIP experiments, while the T 2 distribution was measured via NMR. Based on eq , the surface relaxivity of coal samples at various ranks was inversely calculated to ensure the reliability of pore size conversion. The experimental results indicate that the surface relaxation rates of the coal samples are as follows: 3.960 nm/ms for the ZJM coal sample, 1.233 nm/ms for the DHB coal sample, and 2.563 nm/ms for the SH coal sample. The signal amplitude in the T2 spectrum obtained from NMR testing is positively correlated with the quantity of pores at the corresponding scale, with the peak intensity reflecting the total volume of pores of that size. The area under the T2 spectral peak represents the pore volume. Furthermore, the continuous distribution of the T2 spectrum indicates the presence of multiscale pores within the coal sample, along with favorable connectivity between pores and microfractures. Experimental Protocol and Procedures The experimental procedure is as follows:1.Calibration of experimental equipment: The nuclear magnetic resonance spectrometer was calibrated using standard coal samples of different masses.2.Drying of coal samples: According to GB/T 474–2008 “Preparation method of coal”, the coal samples were placed in a drying oven and dried at 105 °C for 24 h. After removal, they were cooled to room temperature in a desiccator, and the weight M1 of the dried coal samples was measured.3.Saturation of coal samples: According to GB/T 23561.3-2009 “Methods for determining the physical and mechanical properties of coal and rock - Part 3: Methods for determining the density of coal and rock lumps”, the dried coal samples were placed in a vacuum extraction device and subjected to vacuum extraction for 20 to 30 min at a vacuum degree of 0.001 MPa. Then, purified water for the experiment was injected until the water level was 2–3 cm above the coal samples. Vacuum extraction continued until no more gas bubbles emerged from the coal sample surfaces and the extraction duration was no less than 1 h. Subsequently, the chamber was pressurized to 10 MPa for an additional 12 h of pressurized saturation. The saturated samples were removed, surface moisture was wiped off, and the weight M2 of the saturated coal samples was measured. The initial porosity φ0 of the coal samples was obtained according to eq . 3 φ0=M2−M1V where φ0 is the initial porosity of the coal sample (%); M 1 is the weight of the dry coal sample (g); M 2 is the weight of the saturated coal sample (g); and V is the volume of the coal sample (cm3).4. T 2 spectrum measurement of saturated coal samples: In accordance with GB/T 42035-2022 “Determination of pore size distribution in coal and rock by nuclear magnetic resonance method”, nuclear magnetic signal testing was performed on the coal samples to obtain their nuclear magnetic decay signals by using the MacroMR12-150H-I low-field nuclear magnetic resonance analysis system manufactured by Niumag, Suzhou, China. The acquired decay signals were then inverted to generate T 2 spectrum diagrams, thereby obtaining the T 2 spectra of the saturated coal samples prior to heat treatment.5.Heat treatment of coal samples at different temperatures: ZJM-1, DHB-1, and SH-1 coal samples were heat-treated at 200 °C; ZJM-2, DHB-2, and SH-2 at 300 °C; and ZJM-3, DHB-3, and SH-3 at 400 °C. The heat treatment process involved placing the coal samples in a muffle furnace, heating them at the set temperature for 2 h, followed by removal and cooling to room temperature in air.After heat treatment at 400 °C, both the ZJM-3 and DHB-3 coal samples exhibited expansion and deformation, which prevented the conduction of NMR experiments. Consequently, nuclear magnetic resonance data for these two coal samples at the 400 °C temperature are unavailable. However, such structural alteration is a significant thermal response, Figure shows photographs of the coal samples before and after the heat treatment during the experiment.In the DHB-3 coal sample, intense melting and expansion occurred within the vitrain bands, accompanied by the release of volatiles, with well-developed tensile fractures in the coal matrix. In the ZJM-3 coal sample, no obvious melted material was observed. However, tensile fractures resulting from thermal expansion were well developed, indicating predominantly brittle deformation. In contrast, the SH-3 coal sample showed no significant tensile fractures, which is likely related to the high mechanical strength of anthracite. Thus, the DHB sample exhibits melting behavior, is prone to expansion and deformation upon heating, and shows the poorest thermal stability. The ZJM sample does not melt but has relatively low mechanical strength, with moderate thermal stability. The SH sample shows no melting and possesses high mechanical strength, indicating the best thermal stability (Figure ).6.Resaturation and T 2 spectrum measurement of heat-treated coal samples: Steps 3 and 4 were repeated to conduct nuclear magnetic resonance tests on the heat-treated coal samples under saturated conditions. The obtained NMR signals were inverted to derive the T 2 spectrum of the heat-treated coal samples. 1.Calibration of experimental equipment: The nuclear magnetic resonance spectrometer was calibrated using standard coal samples of different masses. Calibration of experimental equipment: The nuclear magnetic resonance spectrometer was calibrated using standard coal samples of different masses. 2.Drying of coal samples: According to GB/T 474–2008 “Preparation method of coal”, the coal samples were placed in a drying oven and dried at 105 °C for 24 h. After removal, they were cooled to room temperature in a desiccator, and the weight M1 of the dried coal samples was measured. Drying of coal samples: According to GB/T 474–2008 “Preparation method of coal”, the coal samples were placed in a drying oven and dried at 105 °C for 24 h. After removal, they were cooled to room temperature in a desiccator, and the weight M1 of the dried coal samples was measured. 3.Saturation of coal samples: According to GB/T 23561.3-2009 “Methods for determining the physical and mechanical properties of coal and rock - Part 3: Methods for determining the density of coal and rock lumps”, the dried coal samples were placed in a vacuum extraction device and subjected to vacuum extraction for 20 to 30 min at a vacuum degree of 0.001 MPa. Then, purified water for the experiment was injected until the water level was 2–3 cm above the coal samples. Vacuum extraction continued until no more gas bubbles emerged from the coal sample surfaces and the extraction duration was no less than 1 h. Subsequently, the chamber was pressurized to 10 MPa for an additional 12 h of pressurized saturation. The saturated samples were removed, surface moisture was wiped off, and the weight M2 of the saturated coal samples was measured. The initial porosity φ0 of the coal samples was obtained according to eq . 3 φ0=M2−M1V where φ0 is the initial porosity of the coal sample (%); M 1 is the weight of the dry coal sample (g); M 2 is the weight of the saturated coal sample (g); and V is the volume of the coal sample (cm3). Saturation of coal samples: According to GB/T 23561.3-2009 “Methods for determining the physical and mechanical properties of coal and rock - Part 3: Methods for determining the density of coal and rock lumps”, the dried coal samples were placed in a vacuum extraction device and subjected to vacuum extraction for 20 to 30 min at a vacuum degree of 0.001 MPa. Then, purified water for the experiment was injected until the water level was 2–3 cm above the coal samples. Vacuum extraction continued until no more gas bubbles emerged from the coal sample surfaces and the extraction duration was no less than 1 h. Subsequently, the chamber was pressurized to 10 MPa for an additional 12 h of pressurized saturation. The saturated samples were removed, surface moisture was wiped off, and the weight M2 of the saturated coal samples was measured. The initial porosity φ0 of the coal samples was obtained according to eq . 3 φ0=M2−M1V where φ0 is the initial porosity of the coal sample (%); M 1 is the weight of the dry coal sample (g); M 2 is the weight of the saturated coal sample (g); and V is the volume of the coal sample (cm3). 4. T 2 spectrum measurement of saturated coal samples: In accordance with GB/T 42035-2022 “Determination of pore size distribution in coal and rock by nuclear magnetic resonance method”, nuclear magnetic signal testing was performed on the coal samples to obtain their nuclear magnetic decay signals by using the MacroMR12-150H-I low-field nuclear magnetic resonance analysis system manufactured by Niumag, Suzhou, China. The acquired decay signals were then inverted to generate T 2 spectrum diagrams, thereby obtaining the T 2 spectra of the saturated coal samples prior to heat treatment. T 2 spectrum measurement of saturated coal samples: In accordance with GB/T 42035-2022 “Determination of pore size distribution in coal and rock by nuclear magnetic resonance method”, nuclear magnetic signal testing was performed on the coal samples to obtain their nuclear magnetic decay signals by using the MacroMR12-150H-I low-field nuclear magnetic resonance analysis system manufactured by Niumag, Suzhou, China. The acquired decay signals were then inverted to generate T 2 spectrum diagrams, thereby obtaining the T 2 spectra of the saturated coal samples prior to heat treatment. 5.Heat treatment of coal samples at different temperatures: ZJM-1, DHB-1, and SH-1 coal samples were heat-treated at 200 °C; ZJM-2, DHB-2, and SH-2 at 300 °C; and ZJM-3, DHB-3, and SH-3 at 400 °C. The heat treatment process involved placing the coal samples in a muffle furnace, heating them at the set temperature for 2 h, followed by removal and cooling to room temperature in air.After heat treatment at 400 °C, both the ZJM-3 and DHB-3 coal samples exhibited expansion and deformation, which prevented the conduction of NMR experiments. Consequently, nuclear magnetic resonance data for these two coal samples at the 400 °C temperature are unavailable. However, such structural alteration is a significant thermal response, Figure shows photographs of the coal samples before and after the heat treatment during the experiment.In the DHB-3 coal sample, intense melting and expansion occurred within the vitrain bands, accompanied by the release of volatiles, with well-developed tensile fractures in the coal matrix. In the ZJM-3 coal sample, no obvious melted material was observed. However, tensile fractures resulting from thermal expansion were well developed, indicating predominantly brittle deformation. In contrast, the SH-3 coal sample showed no significant tensile fractures, which is likely related to the high mechanical strength of anthracite. Thus, the DHB sample exhibits melting behavior, is prone to expansion and deformation upon heating, and shows the poorest thermal stability. The ZJM sample does not melt but has relatively low mechanical strength, with moderate thermal stability. The SH sample shows no melting and possesses high mechanical strength, indicating the best thermal stability (Figure ). Heat treatment of coal samples at different temperatures: ZJM-1, DHB-1, and SH-1 coal samples were heat-treated at 200 °C; ZJM-2, DHB-2, and SH-2 at 300 °C; and ZJM-3, DHB-3, and SH-3 at 400 °C. The heat treatment process involved placing the coal samples in a muffle furnace, heating them at the set temperature for 2 h, followed by removal and cooling to room temperature in air. After heat treatment at 400 °C, both the ZJM-3 and DHB-3 coal samples exhibited expansion and deformation, which prevented the conduction of NMR experiments. Consequently, nuclear magnetic resonance data for these two coal samples at the 400 °C temperature are unavailable. However, such structural alteration is a significant thermal response, Figure shows photographs of the coal samples before and after the heat treatment during the experiment. In the DHB-3 coal sample, intense melting and expansion occurred within the vitrain bands, accompanied by the release of volatiles, with well-developed tensile fractures in the coal matrix. In the ZJM-3 coal sample, no obvious melted material was observed. However, tensile fractures resulting from thermal expansion were well developed, indicating predominantly brittle deformation. In contrast, the SH-3 coal sample showed no significant tensile fractures, which is likely related to the high mechanical strength of anthracite. Thus, the DHB sample exhibits melting behavior, is prone to expansion and deformation upon heating, and shows the poorest thermal stability. The ZJM sample does not melt but has relatively low mechanical strength, with moderate thermal stability. The SH sample shows no melting and possesses high mechanical strength, indicating the best thermal stability (Figure ). 6.Resaturation and T 2 spectrum measurement of heat-treated coal samples: Steps 3 and 4 were repeated to conduct nuclear magnetic resonance tests on the heat-treated coal samples under saturated conditions. The obtained NMR signals were inverted to derive the T 2 spectrum of the heat-treated coal samples. Resaturation and T 2 spectrum measurement of heat-treated coal samples: Steps 3 and 4 were repeated to conduct nuclear magnetic resonance tests on the heat-treated coal samples under saturated conditions. The obtained NMR signals were inverted to derive the T 2 spectrum of the heat-treated coal samples. Photos of coal samples before and after heat treatment The basic physical property parameters of each plunger coal sample before heat treatment are shown in Table . It should be acknowledged that each heat treatment experiment utilized different samples. To a certain extent, heterogeneity exists among different samples of the same coal rank, such as variations in porosity, which may interfere with the analysis of heat treatment experimental results. To address this, our analysis primarily focuses on comparing and examining the differences in pore structure of the same sample before and after heat treatment at specific experimental temperatures, thereby minimizing the adverse effects of sample heterogeneity on the interpretation of experimental results. Even so, it must be recognized that the limitations imposed by sample variability cannot be completely eliminated. 2Fundamental Physical Property Parameters of the Plunger Coal Sample sample weight (g) sample originsample no.diameter (cm)length (cm)dry weight M1 wet weight M2 sample density ρ (g·cm–3)initial porosity φ0 (%)ZJMZJM-12.5024.94328.2831.911.1614.94 ZJM-22.5125.00729.0333.121.1716.48 ZJM-32.5014.95128.6133.321.1819.36DHBDHB-12.5051.98613.4813.911.384.39 DHB-22.5032.03012.9613.471.305.50 DHB-32.5041.89813.2713.661.424.17SHSH-12.5074.38334.1935.121.584.30 SH-22.5084.14429.4630.091.443.08 SH-32.5074.38031.5832.181.462.78 Fundamental Physical Property Parameters of the Plunger Coal Sample Results and Discussion T 2 Spectrum Characteristics and Pore Size Distribution T 2 Spectrum Characteristics The test results, as shown in Figures , , and , indicate that the coal samples exhibit 2–3 main peaks, with the P1 and P2 peak having a relatively large signal amplitude, corresponding to adsorption pores and seepage pores, respectively, while the P3 peaks are comparatively smaller, corresponding to fractures. NMR curves of the ZJM coal sample before and after heat treatment at different temperatures. NMR curves of the DHB coal sample before and after heat treatment at different temperatures. NMR curves of the SH coal sample before and after heat treatment at different temperatures. As shown in Figure , the NMR curve of the ZJM coal sample exhibits a distinct bimodal characteristic. The peak P3 is relatively small, indicating that the fractures are less developed. A comparison of the NMR curves before and after heat treatment reveals a decrease in the peak intensity of P1, suggesting a reduction in the number of adsorption pores. In contrast, the peak intensity of P2 increases, indicating an increase in the number of seepage pores. Additionally, the position of the P1 peak shifts to the left, while that of the P2 peak shifts to the right, demonstrating that after heat treatment, the pore size and total pore volume of adsorption pores in the low-rank coal decrease, while those of seepage pores and fractures increase. As shown in Figure , the NMR curve of the DHB coal sample exhibits a continuous trimodal pattern with well-defined peaks, indicating favorable development of all pore types and good connectivity among adsorption pores, seepage pores and fractures. A comparison of the NMR curves before and after heat treatment reveals that at 200 °C, the P1 peak shows no significant change, while the P2 peak broadens, suggesting that the number of adsorption pores remains relatively stable, whereas the number of seepage pores increases. The intensity of the P3 peak also increases, indicating a rise in the number of fractures. At 300 °C, the development of the P1 peak weakens, the P2 peak strengthens significantly, and the P3 peak shows a slight enhancement. These results demonstrate that after heat treatment, the total volume of adsorption pores in medium-rank coal decreases, while the total volume of seepage pores and fractures increases. As shown in Figure , the NMR curve of the SH coal sample exhibits a distinct bimodal characteristic. The signal amplitude of the P3 peak is very low, indicating that high-rank coal is predominantly characterized by the development of adsorption pores and seepage pores, while fractures are essentially undeveloped. Comparing the NMR curves before and after heat treatment, the P1 peak value decreases after heat treatment, while the peak value of the P2 peak increases significantly, suggesting a reduction in adsorption pores and an increase in seepage pores. As the treatment temperature rises, the peak value of the P3 peak gradually increases. The peak proportions of the P3 peak after treatment at 200 °C, 300 °C, and 400 °C are 0.369%, 0.371%, and 5.52%, respectively, indicating that the fractures volume of the high-rank coal sample increases with higher treatment temperatures. Additionally, the signal amplitude of the trough between the P1 and P2 peaks shows a decreasing trend after heat treatment. When the temperature reaches 300 °C or higher, the separation between the P1 and P2 peaks becomes more pronounced, suggesting that the connectivity between adsorption pores and seepage pores deteriorates after heat treatment. Pore Size Distribution Converting relaxation time to pore radius allows for a more intuitive analysis of the impact of temperature on the pore structure of coal samples with different ranks before and after heat treatment. The T2 relaxation time of the coal samples is converted into pore size distribution according to the following formula: 4 1T2LM=ρ2(SV) In the equation: S represents the specific surface area of the coal sample; V represents the pore volume of the coal sample; ρ2 is the surface relativity rate; and T 2LM is the logarithmic mean value of the T 2 spectrum, which is calculated as follows: 5 T2LM=exp(∑i=1nAiATln⁡T2i) where T 2i is the T 2 value at point i; A i is the amplitude at point i; A T is the total amplitude of the T 2 spectrum. 6 1T2=FS(ρ2rC) where ρ2 is the surface relaxivity; F S is the geometric shape factor (F S = 3 for spherical pores; F S = 2 for cylindrical pores); r C is the pore radius of the coal sample. The obtained coal sample pores were classified according to the coal pore classification method proposed by B. B. Hodot in 1961, namely: micropores, <10 nm; small pores, 10–100 nm; mesopores, 100–1000 nm; and macropores, >1000 nm. The pore size distribution corresponding to the T 2 spectrum of each coal sample was obtained through calculation. Figure illustrates that the ZJM coal sample is predominantly characterized by the development of micropores and small pores, with a moderate development of mesopores and minimal presence of macropores. After heat treatment, the proportion of micropores decreases, while the proportions of small and mesopores increase. When the heat treatment temperature reaches 300 °C, the change in pore distribution becomes more pronounced compared to that at 200 °C, indicating that higher temperatures exert a more significant impact on low-rank coal. Pore size distribution curves of the ZJM coal sample before and after heat treatment at different temperatures. Figure shows that the DHB coal sample is also dominated by the development of micropores and small pores, with moderate development of mesopores and essentially no development of macropores. After heat treatment at 200 °C, although the overall pore size distribution does not change significantly, a reduction in the development of micropores and small pores can still be observed, along with an increase in the development of mesopores and macropores. When the treatment temperature reaches 300 °C, a clear decrease in micropores and small pores and an increase in mesopores and macropores become evident, indicating that higher temperatures lead to more pronounced thermal alteration of the pore structure in medium-rank coal. Pore size distribution curves of the DHB coal sample before and after heat treatment at different temperatures. As shown in Figure , the SH coal sample is predominantly characterized by the development of micropores and small pores, with macropores and mesopores being scarcely developed. After heat treatment, the number of micropores tends to decrease, while the number of small pores increases significantly. As the temperature further rises, the quantity of mesopores and macropores gradually begins to increase. Pore size distribution curves of the SH coal sample before and after heat treatment at different temperatures. Two-Dimensional Relaxation Distribution Density Different hydrogen-containing fluids or different phases of the same fluid do not exhibit independent T 1 and T 2 values but rather possess characteristic T 1/T 2 ratios and two-dimensional distribution patterns. The two-dimensional data set acquired by measuring two independent time variables (T 1 and T 2) integrates the evolution information on the spin system under T 1 and T 2 relaxation mechanisms. Two-dimensional nuclear magnetic resonance relaxation distribution (T 1–T 2) enables rapid identification of multiphase fluids within matrix pores and reflects the distribution of pores at different scales. , Under the influence of diffusion coupling effects exhibited by hydrogen-containing substances in coal under saturated conditions, the T 1 and T 2 spectra of hydrogen atoms in free water within inorganic pores and bound water within organic pores overlap, making it impossible for the T 2 spectrum to distinguish the distribution locations of various hydrogen-containing substances in coal. In contrast, the two-dimensional T 1–T 2 spectrum obtained through inversion can more accurately differentiate between different hydrogen-containing fluids and different phases of the same hydrogen-containing fluid (e.g., bound water, capillary water, and free water). The types of hydrogen-containing fluids represented by signals in different regions are shown in Figure . T 1–T 2 spectra of different hydrogen-bearing fluid signal regions. Figures , , and show the T 1–T 2 two-dimensional relaxation distribution density maps of ZJM, DHB, and SH coal samples under heat treatment at different temperatures. In the density maps, regions with larger T 1 and T 2 values correspond to free fluids and interconnected pores, while regions with smaller T 1 and T 2 values correspond to micropores containing bound water. The brightness of the color in the maps represents the signal intensity, with brighter colors indicating stronger signals. Two-dimensional relaxation distribution density maps (T 1–T 2) of the ZJM coal sample before and after heat treatment. Two-dimensional relaxation distribution density maps (T 1–T 2) of the DHB coal sample before and after heat treatment. Two-dimensional relaxation distribution density maps (T 1–T 2) of the SH coal sample before and after heat treatment. The changes in the two-dimensional spectra of the ZJM sample before and after heat treatment are shown in Figure . The signal clusters are primarily distributed in the upper part of the T 1–T 2 diagonal, exhibiting two peak clusters located in regions III and IV, respectively, indicating well-developed micropores and mesopores. The strong connectivity between regions II and III, as well as between regions III and IV, suggests good pore interconnectivity. Based on the experimental data, a computational analysis of the signal volume in each region of the two-dimensional relaxation spectrum was conducted. The results indicate that after heat treatment, the signal intensity in regions II and III significantly decreases, by 13.71% at 200 °C and 11.99% at 300 °C, while that in region IV increases, by 13.29% at 200 °C and 6.53% at 300 °C, indicating that micropores expanded to form mesopores and macropores following the heat treatment. The changes in the two-dimensional spectra of the DHB sample before and after heat treatment are shown in Figure . From Figure (a) and (c), it can be observed that the fluid distribution in various regions of the coal sample is relatively dispersed, primarily located in the upper part of the T 1–T 2 diagonal line. Multiple peak clusters are evident, mainly distributed in Region IV, with some in Region III, indicating that medium-rank coal is dominated by the development of micropores and mesopores, while macropores are less developed. After heat treatment, the signal intensity in the low T 1 and T 2 value regions significantly decreases, by 9.90% at 200 °C and 11.63% at 300 °C, whereas the signal intensity in Region IV increases, by 6.57% at 200 °C and 7.51% at 300 °C, forming peak clusters. This suggests that after heat treatment, the number of mesopores and macropores in medium-rank coal increases, while the number of micropores decreases. The changes in the two-dimensional spectra of the SH sample before and after heat treatment are shown in Figure . From Figure (a), (c), and (e), it can be observed that the signal clusters of the SH sample are more concentrated, exhibiting only one peak cluster, indicating a simplified pore distribution in high-rank coal. The signal clusters are predominantly located at the junction of regions II, III, and IV, with extremely low signal intensity for macropores, suggesting that micropores overwhelmingly dominate the pore structure. After heat treatment, the signal intensity in regions II and III decreases, by 2.77% at 200 °C, 8.57% at 300 °C and 9.98% at 400 °C, while that in region IV increases, by 2.17% at 200 °C, 5.63% at 300 °C and 6.90% at 400 °C, indicating that some micropores in the high-rank coal have expanded into mesopores or macropores following the heat treatment. The Effect of Heat Treatment on the Porosity of Coal Samples The porosity of coal samples was measured using NMR, and the changes in porosity before and after heat treatment were calculated. The formula for calculating the change in porosity is as follows: 7 Δφ=φ1−φ0 where φ0 represents the porosity of the coal sample before heat treatment,%; φ1 represents the porosity of the coal sample after heat treatment,%; and Δφ represents the variation in porosity before and after heat treatment, %. The measurement error range for porosity is ≤0.1% using MacroMR12-150H-I low-field nuclear magnetic resonance analysis system produced by Niumag, Suzhou, China. The calculation results are shown in Table . 3Changes in Porosity of Coal Samples Before and After Heat Treatment porosity φ (%) sample no.temperature (°C)φ0 φ1 porosity variation Δφ (%)ZJM-120014.9417.492.55ZJM-230016.4829.5013.02ZJM-340019.3635.2815.91DHB-12004.395.421.02DHB-23005.507.942.44DHB-34004.17--SH-12004.304.950.65SH-23003.087.284.20SH-34002.788.515.74aNote: φ0 is the porosity prior to heat treatment. φ1 is the porosity subsequent to heat treatment. The data ″-″ in the table indicates that the coal sample underwent expansion deformation after heat treatment, making NMR experiments impossible. Changes in Porosity of Coal Samples Before and After Heat Treatment Note: φ0 is the porosity prior to heat treatment. φ1 is the porosity subsequent to heat treatment. The data ″-″ in the table indicates that the coal sample underwent expansion deformation after heat treatment, making NMR experiments impossible. A comparative analysis of the porosity evolution in coal samples of different ranks before and after heat treatment reveals that, with the exception of the DHB-3 sample, whose porosity could not be measured due to expansion-induced deformation, all other coal samples exhibited an increase in porosity following heat treatment. Under the same coal rank conditions, the magnitude of porosity change increased with rising treatment temperature. At identical treatment temperatures, the ZJM coal sample demonstrated the greatest change in porosity, indicating that the pore structure of low-rank coal is more sensitive to temperature variations. A significant increase in the porosity of the SH coal sample was observed only when the heat treatment temperature reached 300 °C, suggesting that the thermal evolution of pore structures in high-rank coal requires a higher activation temperature. The DHB coal sample, classified as a high-volatility bituminous coal with caking properties, exhibited significant thermal expansion of the coal matrix upon heating. This expansion likely resulted in the compression and deformation of the pore structure, leading to a lower porosity increase at 300 °C heat treatment compared to the ZJM and SH coal samples. Porosity and its variation in coal samples before and after heat treatment. The Effect of Heat Treatment on the Fractal Dimension of Pore Structure in Coal Samples Fractal dimension is a method for quantitatively characterizing the irregularity of pore surfaces, which can reflect the heterogeneity and complexity of the internal pore structure of coal. Fractal dimension can quantitatively describe pore surface roughness, the complexity of pore distribution, and spatial irregularity, serving as a key parameter reflecting the heterogeneity of pore structures. Traditional pore characterization methods mostly rely on macroscopic statistics or averaged parameters, which struggle to capture the variations in pore distribution and complexity across different scales. In contrast, fractal dimension enables a unified quantification of pore structures across multiple scales, offering a more refined representation of pore evolution characteristics and addressing the limitations of conventional methods in describing pore heterogeneity. Fractal Dimension Calculation Methods According to fractal geometry theory, the cumulative pore volume fraction VC can be expressed as 8 VC=V(