Abstract
Thermal stimulation is a cutting-edge technology for improving shale matrix permeability and achieving efficient shale gas development, as heating can reshape shale pore structures and enhance reservoir percolation capacity. Currently, numerous studies have revealed basic thermal-action laws and clarified the intrinsic relationship between temperature and shale reservoir structures. Nevertheless, most previous tests adopt several discrete temperature points, and quantitative descriptions of successive pore-fracture evolution in the high-temperature range under equal-interval temperature gradients are still lacking. This restricts theoretical support for optimizing field thermal-simulation parameters. To fill this gap, shale samples gradient-treated at 50–300 °C are investigated. Micro-CT scanning, ImageJ, FracLac, and Avizo are employed for two-dimensional statistics and three-dimensional reconstruction. Pore-fracture parameters, fractal dimension, and lacunarity are extracted. Complementary SEM observations at 100 and 250 °C corroborate CT-derived pore-fracture features at these two temperatures. Permeability curves covering 50–300 °C gradient series are further plotted, whose overall trend agrees well with pore-fracture structural variations captured by CT reconstruction. The results show that pore quantity and porosity decrease initially and then rise with increasing temperature. The maximum porosity appears at 300 °C, and the most complicated pore frameworks are formed at 200 °C. Fractal dimension effectively characterizes the spatial heterogeneity evolution of thermally modified shale pore networks, providing insights for shale gas thermal stimulation applications.


1. Introduction
Thermal stimulation is regarded as a cutting-edge technical route adopted to improve matrix permeability of shale. , Internal pore structures of shale can be reshaped by temperature elevation. New fractures are induced to develop, and gas seepage channels are widened. Reservoir permeability is therefore remarkably enhanced. , The in situ superheated-steam injection technology has shown favorable engineering feasibility in pilot tests carried out at the mine site. Most existing investigations are only carried out for static characterization of shale pores under ambient temperature. Combined quantitative analyses of two-dimensional morphology and three-dimensional topological networks are rarely conducted for pores and fractures in gradient-heated shale. The complete evolution mechanism covering primary pore closure and secondary fracture initiation-propagation during heating has not been fully clarified. Multiple testing approaches, including CT scanning, two-dimensional image quantitative analysis, three-dimensional reconstruction, and fractal characterization, are coupled in this work. Transformation laws of shale pore structures under temperature gradients are revealed using these combined methods. Theoretical support can be supplied for parameter optimization of thermal stimulation production technologies.
A complete multiscale characterization system covering macroscopic, mesoscopic, and microscopic dimensions has been established in international research fields at present. Each scale has differentiated observation objects and analysis methods. − Three-dimensional seismic exploration, well logging interpretation, and field outcrop investigation are adopted as core macroscale testing methods. Overall identification of reservoir spatial distribution, stratum thickness, and large structural fractures can be realized. Lateral heterogeneity of reservoirs is intuitively characterized, and favorable gas production sweet spots are delineated. , Industrial CT, nuclear magnetic resonance, and other nondestructive testing devices are utilized at the mesoscopic scale. A three-dimensional reconstruction of internal pore networks inside intact rock cores is performed. Core parameters such as pore connectivity and fracture propagation morphology are quantitatively acquired. Scanning electron microscopy (SEM) and focused ion beam scanning electron microscopy (FIB-SEM) were applied at the microscopic scale. Visual observation of microscale pores and clay mineral morphologies is achieved. − Laboratory tests, including nitrogen adsorption and high-pressure mercury intrusion, can be supplemented as auxiliary characterization means. Fractal theory is introduced to quantify the complexity of pore-fracture structures. Nevertheless, the whole set of characterization methods and quantitative evaluation models available are mostly constructed based on natural marine shale samples tested at ambient temperature. These approaches are applicable only to intact strata without artificial thermal modification. Special evolutionary behaviors, such as internal pore reconstruction and mineral thermal damage after thermal treatment, cannot be described by these methods. Direct application to the quantitative evaluation of microstructures in thermally modified shale cannot be realized.
Multiscale characterization methods are continuously optimized by scholars to adapt to geological differences of shale formed under different sedimentary environments. At the microscale, electron microscopy sample preparation procedures are refined, and combined adsorption and mercury intrusion tests are performed. A fractal quantitative evaluation system is established to distinguish marine, continental, and marine-continental transitional shales. , At the mesoscale, the scanning resolution of the CT equipment is continuously upgraded. Quantitative correlations between the internal fracture development scale and hydraulic fracturing stimulation effects of rock cores are established. At the macroscale, seismic inversion and three-dimensional geological modeling algorithms are continuously updated. The prediction accuracy of key reservoir parameters, such as formation porosity and in situ gas content, is further improved. , Although current technologies can effectively characterize the original pore features of natural shale, an integrated quantitative scheme that combines two-dimensional planar morphology, three-dimensional spatial topology, and fractal heterogeneity indices is still lacking for gradient thermally modified shale. The coupled evolutionary mechanisms of thermally induced fractures, thermally compacted primary pores, and thermally generated organic pores cannot be fully revealed using conventional single-scale testing methods.
Significant limitations exist in the current domestic and international shale characterization studies. Overall, multiscale data fusion and high-precision quantitative evaluation models remain the key bottlenecks to be addressed in this field. Existing research results mostly focus on the static observation of pores and fractures in intact shale under ambient temperature. The dynamic evolution characteristics of microstructures under thermal effects cannot be reflected. Only two-dimensional imaging or three-dimensional reconstruction has been separately applied in limited studies on thermally treated shale. An integrated quantitative analysis of planar morphology and spatial topology is lacking. Multidimensional characteristics, including the geometry, spatial distribution, and connectivity of pore networks, cannot be fully characterized. Integrated quantitative evaluation of two-dimensional morphological parameters and three-dimensional network topological structures is also insufficient. The entire evolution process of pore spaces cannot be completely restored. Meanwhile, most existing evaluation systems rely on basic indicators, such as porosity and average pore diameter. Parameters including fractal dimension and lacunarity, which can quantitatively characterize pore-fracture complexity and spatial heterogeneity, are rarely introduced. Subtle differences in temperature-induced microstructures cannot be accurately distinguished. More importantly, the closure of primary micropores and the development of thermally induced fractures occur simultaneously during shale thermal treatment. , Systematic quantitative investigations on the competitive coupling mechanism of the two aforementioned behaviors and the critical temperature thresholds for abrupt pore structural variations at different temperature intervals are insufficient. Existing research results cannot effectively support the optimization of key operational parameters, including the target temperature and constant-temperature duration, for field thermal stimulation development. Furthermore, current shale pore characterization studies are predominantly restricted to single-scale analyses at ambient temperature.
In addition, most current characterizations of shale pore systems are restricted to single-scale analysis at ambient temperature. Quantitative heterogeneity indicators, including fractal dimension and lacunarity, are rarely explored for shales subjected to gradient high-temperature treatment. Conventional parameters (e.g., porosity and pore diameter) fail to accurately characterize the thermal modification effects on pore-fracture complexity. The full evolutionary mechanism of primary pore closure and secondary fracture initiation and propagation in shale at different temperature intervals remains unclear. Systematic quantitative experimental findings for guiding the thermal stimulation technology design are insufficient.
On the basis of this, thermally treated high-temperature shale is selected as the research object in this paper. Combined with two-dimensional quantitative image analysis, three-dimensional pore-fracture reconstruction, and fractal geometric characterization, the evolution laws of micropore structures in shale under high-temperature treatment are systematically revealed. The specific research contents are presented as follows: first, CT scanning is performed on shale cores treated at different temperatures to obtain two-dimensional tomographic slices. Second, binarization segmentation of pore networks was conducted via ImageJ software. Typical two-dimensional morphological parameters, including pore-fracture area, perimeter, porosity, aspect ratio, and roundness, are quantitatively extracted to compare the morphological variations of pore structures under different temperatures. Third, the FracLac plugin is employed to calculate the fractal dimension and lacunarity of the pore structures. The development complexity and spatial distribution heterogeneity of shale pores after heating are quantitatively characterized. A three-dimensional reconstruction of CT images is further carried out with Avizo software. Internal pore networks inside the shale are reconstructed in a visualized form. Three-dimensional characteristic parameters such as pore volume, spatial connectivity, pore size distribution, and topological relations are quantitatively acquired. Finally, two-dimensional quantitative analysis from ImageJ and three-dimensional visualization using Avizo are combined. Evolution features of the shale pore and matrix under high temperature are fully presented from planar morphology and spatial network dimensions. Multiscale integrated quantitative evaluation is achieved. Internal mechanisms of high-temperature-induced micropore closure, secondary fracture generation and propagation, as well as improved pore connectivity can be clarified by the research outcomes of this paper. Critical temperature intervals corresponding to abrupt variations in the shale pore structures are identified. Provide experimental and theoretical support for the design of thermally enhanced recovery mining technologies.
2. Rock Sample and Microstructure Measurement Methods
The shale used in this study was collected from outcrops of the Upper Sinian strata in the Jiu-Rui Basin, Lushan City, Jiujiang City, and Jiangxi Province.
Shales distributed in northwest Jiangxi belong to the Lower Paleozoic organic-rich mudstone. The geological resource volume of shale gas in this area is 1.0 trillion cubic meters. Its renewable resource volume is approximately 0.2 trillion m3.
Specimens were processed following laboratory rock sample standards proposed by the International Society for Rock Mechanics. Cylindrical shale specimens were fabricated with a diameter of ∼25 mm and a length of 50 mm. The flatness of both end faces was controlled within 0.02 mm. Parallelism between the top and bottom surfaces was limited to less than 0.05 mm.
Weathered layers and fractured zones on shale surfaces were removed to guarantee clear scanning images. After the surface treatment, six shale specimens were obtained. These samples were labeled Sample A, Sample B, Sample C, Sample D, Sample E, and Sample F in sequence. Relevant details are presented in Figure .
1.

Shale sample.
2.1. Testing Equipment
A high-temperature oven and a nanoVoxel3502E CT scanner were adopted for all tests. The nanoVoxel3502E CT scanner is displayed in Figure . A maximum operating voltage of 300 kV is supported by this equipment. The maximum measurable sample dimension is set at 427 mm in diameter and 427 mm in length. A detail resolution of 500 nm per pixel can be achieved theoretically. Its spatial resolution is calibrated to 2 μm per pixel. Accordingly, the effective detection limit for pore identification is approximately 4–6 μm, as at least two to three pixels are required to reliably resolve a pore. Pores below this scale, particularly nanopores, cannot be directly resolved or quantitatively characterized by the present CT system.
2.

CT testing machine.
Prepared shale specimens were divided into several portions before testing. Specimens were heated in advance and then placed on the scanning stage of the CT device. A full 360° rotation was completed by the scanning stage. X-ray scanning was conducted on the samples during rotation. Density differences exist within the shale matrix. Distinct X-ray absorption rates are produced when X-rays pass through different substances. In general, absorption capacity is positively correlated with material density. Primary CT slices of shale samples can be reconstructed based on the varying X-ray absorption rates of internal components. These raw CT slices are grayscale images with 256 gray levels ranging from 0 to 255. Brighter tones are displayed for regions with higher material density. Darker shades correspond to areas with lower density.
2.2. High-Temperature Heat Treatment
A temperature interval of 50 °C was adopted for this thermal test. Six shale specimens were chosen to receive heat treatment at distinct target temperatures. Heating was carried out inside the high-temperature oven. The temperature was raised from ambient temperature (20 °C) at a constant rate of 5 °C per minute. Specimens A, B, C, D, E, and F were heated to 50, 100, 150, 200, 250, and 300 °C, respectively. Once each preset target temperature was reached, specimens were retained in the oven for 2 h. Natural cooling was allowed until room temperature was restored. The whole temperature rise procedure is illustrated in Figure .
3.

Temperature treatment process.
2.3. Determination of Fractal Dimension and Lacunarity
The fractal dimension serves as a key parameter to describe the irregular distribution of pore sizes and the complexity of pore geometries, whereas lacunarity characterizes the degree of aggregation of spatial pore distribution within porous media. Lacunarity can not only quantify the aggregation of pore distribution but also differentiate CT images with identical fractal dimensions, which acts as a vital supplement for fractal analysis of microstructures in porous media. For CT images, the box-counting method is effective for fractal dimension measurement, and the sliding box-counting method is adopted to calculate the lacunarity. Schematic diagrams of the box-counting method and sliding box-counting method are presented in Figures and . The small squares denote pixels: blue represents pores, white denotes the rock matrix, and red squares denote the counting boxes. In both approaches, boxes of varying sizes are employed to cover the image, and the number of boxes covering the pores is recorded. The difference lies in the movement step of each box. ImageJ software is used to analyze the fractal dimension and lacunarity of the CT images after the plugin of fractal analysis was installed. When an image is opened, we can select the command Plugins → Fractal Analysis to obtain the fractal dimension and lacunarity.
4.

Box-counting method.
5.

Sliding box-counting method.
Square grids with side length δ are used for two-dimensional cases, while cubic grids are adopted for three-dimensional analysis. These grids are laid to fully cover the target images or numerical models. Grids entirely enclosed within pores and fractures are counted. This counting process is repeated after the grid size δ is adjusted. For objects with fractal characteristics, a fixed correlation exists between the total number of grids and the grid side length δ.
| 1 |
where D represents the fractal dimension of the pore structure. Taking the double logarithm of both sides gives us the linear relationship
| 2 |
where C is the constant fitting. Plotting lg R on the x-axis and lg N on the y-axis, the absolute value of the slope of the resulting straight line is the fractal dimension D of the pore structure.
The two-dimensional fractal dimension ranges from 1 to 2. Larger values reflect more complex, rougher planar shapes of pores and fractures, alongside a stronger filling capacity. The three-dimensional fractal dimension varies between 2 and 3. Higher values correspond to more intricate spatial geometries of pores and fractures and a stronger heterogeneity.
Sliding windows with side length δ are adopted for calculation. A single pixel is set as the moving step size. Scanning is carried out across CT images or three-dimensional models. The scanning path covers the entire area from left to right and top to bottom. The number of pore pixels inside each individual window is recorded. Calculate the average and standard deviation of the number of pore pixels in all the windows. The lacunarity is defined as follows
| 3 |
For a fixed window size, larger lacunarity values indicate an uneven spatial distribution of pores and fractures. These pore features tend to cluster, and a stronger heterogeneity is presented accordingly. On the contrary, smaller lacunarity values correspond to a uniform and scattered spatial distribution of pores and fractures. Lacunarity values calculated via multiscale windows are employed in this study. These indexes are used to quantify the spatial heterogeneity of pores and fractures within the shale samples.
Two-dimensional CT slices were segmented using interactive manual grayscale thresholding, which was determined based on grayscale histogram analysis, to separate pore spaces from the rock matrix, and three-dimensional pore reconstruction with quantitative analysis was subsequently performed on the Avizo software platform. For image processing, we applied median filtering available in both ImageJ and Avizo, while ring-artifact correction for the three-dimensional data sets was carried out using the built-in function of Avizo.
Shale pores exhibit complex geometries, and some maintain spatial connections only through weak edge-or-corner contacts. The 26-connectivity rule is therefore well suited for identifying such weakly connected features, as it prevents pores that are inherently connected from being misclassified as isolated. We adopted this criterion to screen connected pore components in 3D, where adjacent voxels sharing faces, edges, or corners were treated as a single continuous pore domain. Correspondingly, the 8-connectivity rule was used for pore identification on 2D slices.
For 3D pore-fracture reconstruction, a cubic region of interest (ROI) of 1 × 1 × 1 mm3 was consistently extracted from the geometric center of each shale core. This central region avoids CT edge artifacts and end effects and corresponds to the thermally most homogeneous zone during heat treatment. It should be noted that this subvolume does not strictly satisfy the representative elementary volume (REV) criterion. The main objective of 3D reconstruction here is to characterize microscale pore-throat topological features rather than to derive absolute macroscopic petrophysical parameters.
3. Results
3.1. Characteristics from CT Images
To clarify the criteria for determining the temperature thresholds in this study, quantitative definitions are provided for three key terms as follows. (1) Porosity turning point. It refers to the temperature at which porosity rebounds and exceeds its initial value (the value at 50 °C) for the first time after an early-stage decrease. It marks the transition of pore-fracture evolution from closure-dominated to generation-dominated patterns. (2) Pore-fracture complexity peak. It is defined as the temperature corresponding to the local maximum of fractal dimension. This term characterizes the dense development of newly formed microfractures. (3) Rapid expansion. It represents the temperature stage where the growth rate of pore-fracture area exceeds the average growth rate across the full temperature range. This interval corresponds to the most remarkable enlargement of pore-fracture dimensions.
Two-dimensional pore slices at various depths of six specimens A–F were acquired using high-precision CT scanning. These slice images are displayed in Figure . The slices are arranged into six rows by their corresponding heating temperatures. One specimen corresponds to each row, which contains three slices captured at distinct scanning depths. Details of each row are listed as follows: Row 1: Slices 134, 267, and 400 of Specimen A; Row 2: Slices 270, 539, and 808 of Specimen B; Row 3: Slices 255, 509, and 763 of Specimen C; Row 4: Slices 259, 517, and 775 of Specimen D; Row 5: Slices 204, 407, and 610 of Specimen E; Row 6: Slices 283, 566, and 848 of Specimen F.
6.

Partial CT images. (a) specimen A, slice 134, (b) specimen A, slice 267, (c) specimen A, slice 400, (d) specimen B, slice 270, (e) specimen B, slice 539, (f) specimen B, slice 80, (g) specimen C, slice 255, (h) specimen C, slice 509, (i) specimen C, slice 763, (j) specimen D, slice 259, (k) specimen D, slice 517, (l) specimen D, slice 775, (m) specimen E, slice 204, (n) specimen E, slice 407, (o) specimen E, slice 610, (p) specimen F, slice 283, (q) specimen F, slice 566, (r) specimen F, slice 848.
Slices from different scanning depths of an identical specimen are placed on the same row. Sound consistency in the internal pore structures is observed within each specimen group along the scanning depth direction. For the three slices of a single specimen, the overall development scale, size range, and distribution pattern of pores and fractures are largely consistent. Minor discrepancies are spotted only in the geometry and area proportion of pores and fractures at a few local positions. Variations across different depth layers remain slight, and relatively uniform microstructures are exhibited by the specimens.
Slices captured at similar scanning depths of distinct specimens are arranged in the same columns. Distinct intergroup differences in pore characteristics are detected among different specimens. Pores and fractures in specimens A, B, C, and D are generally tiny with low development degree and indistinct outlines. Obvious local aggregation of pores and fractures is found in specimen E. Prominent through-going fractures are present in specimen F.
To further verify the reliability of CT-derived observations and obtain microstructural details at a higher resolution, shale specimens treated at 100 and 250 °C were selected for SEM observation. A magnification of 4000× was adopted. The results are presented in Figure . SEM images show that after heat treatment at 100 °C, primary intercrystalline pores are present on shale surfaces. No preferentially oriented microfractures are observed. After treatment at 250 °C, distinct microfractures are observed within the shale matrix. Interconnected microfractures can be observed in localized regions.
7.

SEM micrographs acquired at 4000× magnification: (a) 100 °C, (b) 250 °C.
As a supplementary technique to CT, SEM observations confirm the pore-fracture evolution trends derived from CT slicing. Specifically, obvious fractures identified on CT slices for the 250 °C specimen are well corroborated by the microfractures captured in SEM images.
Pore areas of all parallel specimens in each group show an obvious unimodal skewed distribution. Most pores are distributed within the range of 0–50 μm2. Occurrence probability drops rapidly as pore area rises. Consistent distribution shapes, peak positions, and declining trends are recorded for pores within one identical group. Their cumulative distribution curves largely overlap (Figure ).
8.

Pore size distribution with specimen A, specimen B, specimen C, specimen D, specimen E, and specimen F.
Statistical quantitative parameters of pore structures for shale samples treated at various heating temperatures are listed in Table . Mean values extracted from the table are plotted into a line chart shown in Figure . The heating temperature is set as the horizontal axis of this chart. Six key microscopic pore indicators are adopted as the vertical axis, including pore count, pore area, pore perimeter, porosity, overall roundness, and solid fraction.
1. Pore and Fracture Structure Parameters of Shale Samples.
| Specimen temperature | Group | Count | Area/mm2 | Perimeter/mm | Porosity/% | Overall roundness | | Solid fraction |
|---|---|---|---|---|---|---|---|
| 50 | 1 | 6849 | 0.380 | 0.028 | 6.849 | 0.819 | 0.817 |
| 2 | 7254 | 0.335 | 0.026 | 6.046 | 0.833 | 0.823 | |
| 3 | 7142 | 0.384 | 0.028 | 6.932 | 0.817 | 0.818 | |
| Average | 7082 | 0.366 | 0.027 | 6.609 | 0.823 | 0.819 | |
| 100 | 1 | 4979 | 0.289 | 0.028 | 4.839 | 0.852 | 0.838 |
| 2 | 4874 | 0.338 | 0.030 | 5.661 | 0.838 | 0.830 | |
| 3 | 5316 | 0.335 | 0.030 | 5.611 | 0.837 | 0.830 | |
| Average | 5056 | 0.321 | 0.029 | 5.370 | 0.842 | 0.833 | |
| 150 | 1 | 4225 | 0.413 | 0.038 | 6.995 | 0.831 | 0.829 |
| 2 | 2461 | 0.404 | 0.044 | 6.846 | 0.812 | 0.824 | |
| 3 | 3088 | 0.367 | 0.042 | 6.213 | 0.803 | 0.820 | |
| Average | 3258 | 0.395 | 0.041 | 6.685 | 0.815 | 0.824 | |
| 200 | 1 | 3441 | 0.352 | 0.039 | 6.121 | 0.818 | 0.825 |
| 2 | 3463 | 0.468 | 0.040 | 8.152 | 0.813 | 0.825 | |
| 3 | 5589 | 0.451 | 0.032 | 7.842 | 0.833 | 0.830 | |
| Average | 4164 | 0.424 | 0.037 | 7.372 | 0.821 | 0.827 | |
| 250 | 1 | 3994 | 0.489 | 0.039 | 8.919 | 0.826 | 0.824 |
| 2 | 5194 | 0.508 | 0.038 | 9.272 | 0.809 | 0.817 | |
| 3 | 3072 | 0.494 | 0.048 | 9.004 | 0.795 | 0.815 | |
| Average | 4087 | 0.497 | 0.042 | 9.065 | 0.810 | 0.819 | |
| 300 | 1 | 5520 | 0.635 | 0.038 | 10.851 | 0.826 | 0.825 |
| 2 | 6138 | 0.652 | 0.038 | 11.137 | 0.819 | 0.822 | |
| 3 | 5525 | 0.477 | 0.035 | 8.156 | 0.827 | 0.824 | |
| Average | 5728 | 0.588 | 0.037 | 10.048 | 0.824 | 0.824 |
9.

Relationship between various indicators and temperature changes: (a) pore count, (b) pore area, (c) pore perimeter, (d) porosity, (e) overall roundness, (f) solid fraction.
Pore counts decline first and then rise as temperature rises. The minimum count of 3258 is recorded at 150 °C, followed by a gradual upward trend. Porosity follows a similar variation rule: it drops at first and rebounds afterward. The lowest porosity value of 5.370% is captured at 100 °C, and a rapid growth is observed subsequently. Pore area and perimeter are both generally enlarged. Their values shift from 0.366 mm2 and 0.027 mm at 50 °C to 0.588 mm2 and 0.037 mm at 300 °C. The most prominent gaps are detected between 250 and 300 °C. Slight variations are presented for overall roundness and solid fraction. Overall roundness ranges from 0.810 to 0.842, while solid fraction varies between 0.819 and 0.833. No obvious regular evolutionary trend can be identified for these two parameters.
Positive effects are exerted by temperature on the development of shale pores and fractures as well as reservoir storage space. The most significant promotion of pore propagation is triggered by temperature growth within 250–300 °C. The above findings can be taken as reference standards for high-temperature permeability enhancement modification of shale reservoirs.
In this study, the heterogeneity degree is defined as
| 4 |
| 5 |
where s represents the standard deviation of raw data from three parallel specimens, x̅ denotes the mean value of the raw data for the three parallel specimens, and n is the number of parallel specimens.
As shown in Figure , the nonuniformity values of pore number, pore area, pore perimeter, and porosity stay relatively high overall. Variations and peaks are observed within the 150–250 °C range. In contrast, nonuniformity of overall roundness and solid fraction remains consistently low.
10.

Heterogeneity degree of pore parameters versus temperatures: (a) pore count, (b) pore area, (c) pore perimeter, (d) porosity, (e) overall roundness, (f) solid fraction.
Two primary factors explain the high nonuniformity of pore-related parameters across parallel specimens. First, the 150–250 °C range forms a transition zone with substantial microstructural reorganization. Organic decomposition, gas escape, pore nucleation, random coalescence, and collapse take place during this stage. These processes are stochastic. They respond strongly to local temperature gradients, thermal disturbances, and initial microdefects, which directly increase the absolute standard deviation (s) of measured results. Second, nonuniformity is defined as the relative standard deviation. At each temperature point, the mean and standard deviation are calculated independently from three parallel specimens. Within 150–250 °C, mean values of these pore parameters are relatively low. For equal absolute variation, this denominator effect further amplifies the calculated nonuniformity.
By comparison, overall roundness and solid fraction are less sensitive to the random evolution of local pores. Smaller measurement discrepancies were found between parallel specimens. Their nonuniformity therefore stays at low levels.
3.2. Heterogeneity Analysis
Comparisons of porosity and fractal dimension measured from slice images of the six specimens are listed in Table . These data sets are plotted to form a line chart, which is displayed in Figure .
2. Porosity and Fractal Dimension.
| Specimen temperature | Number | Porosity/% | Fractal dimension |
|---|---|---|---|
| 50 | 1 | 6.849 | 1.381 |
| 2 | 6.046 | 1.351 | |
| 3 | 6.932 | 1.383 | |
| Average | 6.609 | 1.372 | |
| 100 | 1 | 4.839 | 1.252 |
| 2 | 5.661 | 1.290 | |
| 3 | 5.611 | 1.297 | |
| Average | 5.370 | 1.280 | |
| 150 | 1 | 6.995 | 1.349 |
| 2 | 6.846 | 1.348 | |
| 3 | 6.213 | 1.314 | |
| Average | 6.685 | 1.337 | |
| 200 | 1 | 6.121 | 1.567 |
| 2 | 8.152 | 1.585 | |
| 3 | 7.842 | 1.592 | |
| Average | 7.372 | 1.581 | |
| 250 | 1 | 8.919 | 1.420 |
| 2 | 9.272 | 1.562 | |
| 3 | 9.004 | 1.398 | |
| Average | 9.065 | 1.460 | |
| 300 | 1 | 10.851 | 1.475 |
| 2 | 11.137 | 1.488 | |
| 3 | 8.156 | 1.408 | |
| Average | 10.048 | 1.457 |
11.

Temperature and porosity effects on fractal dimension: (a) specimen temperature, (b) porosity.
Minor data fluctuations are observed among the three parallel specimen groups under identical temperatures. No obvious outliers are found for either porosity or fractal dimension. Reliable test results with strong repeatability are thus verified, which can accurately characterize the micropore features of shale at corresponding temperatures.
A general trend of decreasing first and then gradual growth is shown by porosity as heating temperature rises. Porosity falls from 6.609% to 5.370% within 50–100 °C. A slow rise takes place when the temperature exceeds 150 °C. Sharp growth is captured from 200 to 300 °C, and the peak porosity of 10.048% is reached at 300 °C. By contrast, fractal dimension presents a distinct evolutionary pattern. It declines continuously before a mild rebound and finally stabilizes. Fractal dimension drops from 1.372 to 1.280 between 50 and 100 °C, which indicates relatively simple and homogeneous pore frameworks at this stage. A slight growth is recorded at 150 °C, and the maximum fractal dimension of 1.581 is obtained at 200 °C, representing the most intricate and heterogeneous pore structure. Slight drops in fractal dimension are detected at 250 and 300 °C, yet values are kept within 1.45–1.46, and well-developed pores and fractures are still maintained. A positive correlation is displayed between fractal dimension and porosity in Figure . Better-developed pores and fractures with larger area fractions correspond to more complex structures and stronger heterogeneity.
Central-slice images of each specimen are selected as research subjects. Lacunarity analysis of pores and fractures is carried out on the second group, and relevant data are summarized in Table . Eight box sizes are adopted for each specimen, with pixel side lengths of 54, 103, 152, 201, 250, 299, 348, and 397 pixels. Distinct lacunarity values are yielded when different box sizes are applied to one specimen, and the variation of lacunarity is illustrated in Figure .
3. Hole and Fracture Lacunarity under Different Box Sizes.
| Box size | A | B | C | D | E | F |
|---|---|---|---|---|---|---|
| 54 | 1.19 | 2.75 | 4.14 | 2.13 | 1.89 | 1.86 |
| 103 | 1.12 | 2.51 | 4.11 | 1.84 | 1.67 | 1.62 |
| 152 | 1.08 | 2.19 | 3.66 | 1.63 | 1.47 | 1.45 |
| 201 | 1.04 | 1.87 | 3.03 | 1.46 | 1.32 | 1.32 |
| 250 | 1.02 | 1.61 | 2.46 | 1.33 | 1.21 | 1.22 |
| 299 | 1.01 | 1.41 | 1.96 | 1.21 | 1.12 | 1.14 |
| 348 | 1.01 | 1.25 | 1.57 | 1.11 | 1.05 | 1.08 |
| 397 | 1.01 | 1.12 | 1.29 | 1.04 | 1.01 | 1.03 |
| Average | 1.06 | 1.84 | 2.78 | 1.47 | 1.34 | 1.34 |
12.

Box size and specimen temperature effects on lacunarity: (a) box sizes, (b) specimen temperature.
The box size and heating temperature exert notable influences on the lacunarity of pores and fractures within shale. For all specimens, lacunarity is observed to decline gradually as the box size grows. Strong fluctuations in the calculated lacunarity are triggered for small windows. Such deviations are induced by concentrated local pores or a dense matrix nearby. Local heterogeneity is smoothed out when the window range is expanded, and lacunarity converges to a stable level.
Lacunarity rises first and then falls with increasing temperature. The peak lacunarity value of 2.78 is recorded at 150 °C. At this temperature, the highest pore aggregation degree and strongest heterogeneity are presented. Lacunarity is lowered and remains nearly constant after 200 °C. A negative correlation is identified between lacunarity and fractal dimension.
4. Discussion
4.1. Three-Dimensional Characterization of Pore Features of Shale Structures at Different Temperatures
Three-dimensional reconstruction is performed with Avizo software. Ring artifacts generated during CT scanning interfere with analysis, and computational overhead needs to be cut down. To address both issues, raw CT slices of six specimens are split into small 1 × 1 mm2 cubes, as illustrated in Figure . Damage to specimen surfaces during sample preparation should be minimized while reliable reconstruction quality is guaranteed. For this purpose, specific CT frames are selected for each sample: frames 51–450 of specimen A, frames 301–700 of specimen B, frames 301–700 of specimen C, frames 301–700 of specimen D, frames 201–600 of specimen E, and frames 201–600 of specimen F. All reconstructed models are standardized to a uniform dimension of 1 × 1 × 1 mm3.
13.

Schematic diagram of cutting.
The reconstruction workflow is given as follows. Noise reduction and image enhancement were applied to slices to raise image contrast. Threshold segmentation is implemented based on grayscale differences afterward. The rock matrix, high-density minerals, and pore voids are separated using this segmentation process. Three-dimensional connected component analysis is then executed to extract individual pores and connected flow channels. Four distinct visual modules are built for each specimen, covering the 3D morphology of the rock matrix, distribution of high-density minerals, connected pore spaces, and pore-throat networks. Three-dimensional visualization of internal shale microstructures was therefore realized.
4.2. Pore-Fracture Physical Parameters
Four reconstructed structural models of shale specimens A, B, C, D, E, and F are displayed in Figure (a). The 3D distribution of high-density materials is shown in Figure (b). Schematic diagrams of connected pores and pore throats are illustrated in Figure (c,d).
14.

3D structure properties of different cases of shale samples. (a) Reconstructed 3D shale samples, (b) high-density materials, (c) connected pores, (d) pore network model.
Parameters including pore, fracture, and throat sizes of the six specimen groups are calculated based on the three-dimensional reconstructed models. All calculated parameters are compiled and summarized in Tables and . Correlations between each parameter and heating temperature are visualized in Figure .
4. Characterization of Pore and Fracture Parameters (Part 1).
| Specimen A | Specimen B | Specimen C | |
|---|---|---|---|
| Pore number | 129 | 230 | 120 |
| Maximum pore diameter (μm) | 58.15 | 53.97 | 62.85 |
| Minimum pore diameter (μm) | 10.31 | 8.45 | 10.70 |
| Average pore diameter (μm) | 31.15 | 21.82 | 31.56 |
| Maximum pore surface area (μm2) | 3.74 × 105 | 1.63 × 105 | 1.63 × 105 |
| Minimum pore surface area (μm2) | 2.48 × 103 | 1.11 × 103 | 1.86 × 103 |
| Average pore surface area (μm2) | 7.40 × 104 | 2.09 × 104 | 3.50 × 104 |
| Maximum pore volume (μm3) | 8.24 × 105 | 6.59 × 105 | 1.04 × 106 |
| Minimum pore volume (μm3) | 4.59 × 103 | 2.53 × 103 | 5.13 × 103 |
| Average pore volume (μm3) | 1.63 × 105 | 5.54 × 104 | 1.87 × 105 |
| Throats number | 215 | 476 | 290 |
| Maximum throat length (μm) | 264.22 | 216.46 | 239.80 |
| Minimum throat length (μm) | 20.23 | 19.53 | 23.67 |
| Average throat length (μm) | 113.05 | 76.51 | 103.10 |
| Maximum throat diameter (μm) | 18.70 | 20.60 | 48.82 |
| Minimum throat diameter (μm) | 0.86 | 0.76 | 1.56 |
| Average throat diameter (μm) | 6.57 | 6.52 | 15.04 |
| Maximum throat surface area (μm2) | 1098.72 | 1332.91 | 7487.68 |
| Minimum throat surface area (μm2) | 2.31 | 1.79 | 7.69 |
| Average throat surface area (μm2) | 173.48 | 166.71 | 924.28 |
| Maximum coordination number | 13 | 13 | 12 |
| Minimum coordination number | 1 | 1 | 1 |
| Average coordination number | 3.35 | 4.148 | 4.85 |
5. Characterization of Pore and Fracture Parameters (Part 2).
| Specimen D | Specimen E | Specimen F | |
|---|---|---|---|
| Pore number | 265 | 249 | 160 |
| Maximum pore diameter (μm) | 50.22 | 53.80 | 58.13 |
| Minimum pore diameter (μm) | 10.56 | 10.06 | 9.10 |
| Average pore diameter (μm) | 24.11 | 25.20 | 27.67 |
| Maximum pore surface area (μm2) | 9.30 × 104 | 2.10 × 105 | 2.92 × 105 |
| Minimum pore surface area (μm2) | 1.84 × 103 | 1.74 × 103 | 1.49 × 103 |
| Average pore surface area (μm2) | 2.29 × 104 | 3.55 × 104 | 4.51 × 104 |
| Maximum pore volume (μm3) | 5.31 × 105 | 6.52 × 105 | 8.23 × 105 |
| Minimum pore volume (μm3) | 4.94 × 103 | 4.27 × 103 | 3.16 × 103 |
| Average pore volume (μm3) | 7.53 × 104 | 9.08 × 104 | 1.31 × 105 |
| Throats number | 536 | 521 | 375 |
| Maximum throat length (μm) | 251.76 | 283.89 | 310.82 |
| Minimum throat length (μm) | 14.21 | 23.49 | 31.84 |
| Average throat length (μm) | 89.64 | 96.72 | 104.44 |
| Maximum throat diameter (μm) | 26.13 | 31.65 | 30.17 |
| Minimum throat diameter (μm) | 0.76 | 0.76 | 0.95 |
| Average throat diameter (μm) | 8.08 | 7.37 | 9.97 |
| Maximum throat surface area (μm2) | 2145.72 | 3146.08 | 2858.93 |
| Minimum throat surface area (μm2) | 1.79 | 1.79 | 2.83 |
| Average throat surface area (μm2) | 271.57 | 227.98 | 406.52 |
| Maximum coordination number | 12 | 20 | 14 |
| Minimum coordination number | 1 | 1 | 1 |
| Average coordination number | 4.05 | 4.21 | 4.7 |
15.

Relationship between each parameter and temperature. (a) Pore number, (b) average pore diameter, (c) average pore surface area, (d) average pore volume, (e) throat number, (f) average throat length, (g) average throat diameter, (h) average throat surface area, (i) average coordination number.
129 pore voids and 215 throats are identified in specimen A. Pore diameters are measured ranging from 10.31 to 58.15 μm, with an average value of 31.15 μm. Pore volumes are recorded between 4.59 × 103 μm3 and 8.24 × 105 μm3, and their average volume is 1.63 × 105 μm3. Throat lengths are detected from 20.23 to 264.22 μm, averaging 113.05 μm. Coordination numbers fall within 1–13, and the average coordination number is 3.35.
230 pore voids and 476 throats are detected in specimen B. Pore diameters span 8.45 to 53.97 μm, with a mean diameter of 21.82 μm. Measured pore volumes range from 2.53 × 103 μm3 to 6.59 × 105 μm3, and the average volume equals 5.54 × 104 μm3. Throat lengths are captured between 19.53 and 216.46 μm, with a mean length of 76.51 μm. Coordination numbers vary from 1 to 13, and the average value is 4.148.
120 pore voids and 290 throats are extracted from specimen C. Pore diameters range from 10.70 to 62.85 μm, and the average pore diameter is 31.56 μm. Pore volumes are documented between 5.13 × 103 μm3 and 1.04 × 106 μm3, with an average of 1.87 × 105 μm3. Throat lengths are measured ranging from 23.67 to 239.80 μm, averaging 103.10 μm. Coordination numbers range from 1 to 12, and the average coordination number is 4.85.
265 pore voids and 536 throats are quantified for specimen D. Pore diameters are distributed in the range of 10.56–50.22 μm, with a mean diameter of 24.11 μm. Pore volumes are constrained within 4.94 × 103–5.31 × 105 μm3, and their average volume is 7.53 × 104 μm3. Throat lengths are recorded from 14.21 to 251.76 μm, with a mean length of 89.64 μm. Coordination numbers range from 1 to 12, and the average value is 4.05.
249 pore voids and 521 throats are observed inside specimen E. Pore diameters are measured from 10.06 to 53.80 μm, and the average pore diameter is 25.20 μm. Pore volumes are detected spanning 4.27 × 103 μm3 to 6.52 × 105 μm3, with an average volume of 9.08 × 104 μm3. Throat lengths are captured between 23.49 and 283.89 μm, averaging 96.72 μm. Coordination numbers extend from 1 to 20, and the average coordination number reaches 4.21.
160 pore voids and 375 throats are counted in specimen F. Pore diameters are varied within 9.10–58.13 μm, with a mean diameter of 27.67 μm. Pore volumes are documented from 3.16 × 103 μm3 to 8.23 × 105 μm3, and the average volume is 1.31 × 105 μm3. Throat lengths are measured ranging from 31.84 to 310.82 μm, with a mean length of 104.44 μm. Coordination numbers fall between 1 and 14, and the average coordination number is 4.70.
4.3. Heterogeneity Analysis
By calculating the three-dimensional porosity and fractal dimension of each sample group, the complexity and heterogeneity of the three-dimensional pore structure of shale can be quantitatively characterized. All calculated data are summarized in Table . Variation trends of porosity and fractal dimension under different heating temperatures are visualized in Figure .
6. Temperature, Porosity, and Fractal Dimension.
| Temperature (°C) | Porosity (%) | Fractal dimension | |
|---|---|---|---|
| Specimen A | 50 | 6.94 | 2.41 |
| Specimen B | 100 | 5.14 | 2.34 |
| Specimen C | 150 | 6.33 | 2.32 |
| Specimen D | 200 | 7.1 | 2.39 |
| Specimen E | 250 | 9.12 | 2.46 |
| Specimen F | 300 | 9.43 | 2.39 |
16.

Specimen temperature effects on porosity and fractal dimension: (a) porosity, (b) fractal dimension.
A general trend of initial decline followed by gradual growth is exhibited by three-dimensional porosity as heating temperature rises. Porosity drops from 6.94% to 5.14% within the temperature range of 50–100 °C. A steady increase is observed for porosity once the temperature exceeds 100 °C, and the peak value of 9.43% is reached at 300 °C. A distinct three-stage evolution is presented by the three-dimensional fractal dimension: it decreases first, then rises, and declines again in the final stage. Fractal dimension falls from 2.41 to 2.32 between 50 and 150 °C. Relatively simple and homogeneous pore frameworks are indicated within shale at this temperature interval. The maximum fractal dimension of 2.46 is recorded at 250 °C, which represents the most intricate and irregular pore network formed under this thermal condition. Consistent variation patterns are detected between the 3D fractal dimension and the previous two-dimensional analysis results. A decisive influence is exerted by temperature on the complexity of pore structures inside shale.
To further verify the influence of heat treatment on the evolution of pore-fracture structures in shale at different temperatures, we additionally include a plot of measured permeability variations of rock samples after treatment at various temperatures, as shown in Figure . The results reveal consistency between the temperature-dependent evolution of permeability and the change in complexity of pore-fracture structures. The permeability of shale specimens exhibits a three-stage evolution as temperature rises. A slight increase is observed first, followed by a continuous decline to the minimum value, and finally a dramatic surge.
17.

Permeability of shale under different temperatures. Permeability experimental data are adopted from our group’s previous work, and this figure is replotted in the present study.
In the low-temperature range of 50–100 °C, permeability rises slightly from 0.00113 mD to 0.00124 mD. Vapor pressure produced by massive vaporization of adsorbed water opens ultrafine pore throats originally occupied by water films. More effective connected pathways are formed.
In the range of 100–200 °C, permeability decreases continuously from 0.00124 mD at 100 °C to 0.0012 mD at 150 °C. The global minimum of 0.00106 mD is reached at 200 °C. This trend is attributed to dual damage imposed on effective seepage channels by thermal expansion of the rock matrix and softening of organic matter. As temperature increases, the rock matrix expands under heating. Pore space inside existing microfractures is compressed significantly. Cross-sectional areas of seepage pore throats are reduced markedly. Meanwhile, organic matter is softened and plastically deformed. Merging and collapse of local pores are triggered. Fragile connected networks are broken. Numerous voids are transformed into isolated dead-end pores. These pores cannot contribute to macroscopic mass transport. Continuous compression and collapse directly block seepage pathways.
Above 200 °C, permeability rises abruptly from 0.00106 mD at 200 °C to 0.00372 mD at 300 °C. The permeability increases by approximately 3.5 times. Large amounts of high-pressure gas are generated during extensive pyrolysis of organic matter. Strong thermal stress is induced by mismatched thermal expansion coefficients among different minerals. Numerous microfractures are randomly initiated and interconnected. Previously blocked isolated dead-end pores are linked. New macroscopic seepage channels are constructed.
5. Conclusions
Shale specimens treated at heating temperatures ranging from 50 to 300 °C are taken as research objects in this paper. Microscale CT imaging technology is adopted for sample observation. Image binarization is conducted with ImageJ software. Fractal analysis is completed using the FracLac plugin. Three-dimensional reconstruction is performed using Avizo software. These integrated methods are applied to identify and quantitatively characterize pores and fractures inside shale under high-temperature environments. Variation trends of pore parameters, thermal evolution of pore properties, and heterogeneity are obtained from the above tests.
-
(1)
Obvious effects are imposed by heat treatment on the development of shale pores and fractures. A universal evolution law is summarized: Pore closure occurs at low temperatures, while new voids are generated and expanded under moderate and high temperatures. Internal pore spaces within shale are sealed at 50–150 °C, and the minimum pore quantities are recorded in this interval. New pores and fractures are generated once the temperature exceeds 150 °C. Pore count, area, perimeter, and porosity are continuously increased with temperature increase. The maximum porosity of 10.048% is achieved at 300 °C. The most rapid expansion of pores and fractures is detected between 250 and 300 °C.
-
(2)
Minor influences are exerted by high temperature on morphological parameters of pores and fractures, whereas significant impacts are produced on structural complexity. Overall roundness and solid fraction are slightly altered during heating, fluctuating within 0.810–0.842 and 0.819–0.833, respectively. Slight interference with the regularity of pore shapes is therefore verified by temperature. Fractal dimension is first reduced and then elevated as temperature rises. The peak fractal dimension of 1.581 is captured at 200 °C, representing the strongest heterogeneity of pore structures. A mild decline is observed afterward, yet fractal dimension values remain at a high level. A positive correlation is confirmed between fractal dimension and porosity.
-
(3)
Prominent promotion effects on pore-throat connectivity of shale are demonstrated by high temperatures according to three-dimensional pore-throat models. The pore volume and specific surface area are gradually enlarged during temperature elevation. Throat diameters were also expanded continuously. Average coordination numbers are increased from 3.35 to above 4.7. More connecting channels are formed between pores and fractures, and a more favorable seepage network is constructed. Better storage and migration conditions are thus provided for the shale gas.
-
(4)
Consistent results are yielded by both two-dimensional and three-dimensional characterization tests. 150 °C is identified as the turning point for the evolution of shale pore structures. 200 °C is determined as the temperature corresponding to the most complicated pore frameworks. The temperature range of 250–300 °C is proven favorable for the formation of high-quality reservoir spaces. Important references are offered for high-temperature permeability enhancement modification of shale reservoirs.
-
(5)
The SEM observations at 100 and 250 °C provide intuitive morphological evidence for pore-fracture development. Furthermore, the permeability evolution curve shows consistent variation trends with the pore structural parameters derived from CT reconstruction, further supporting the thermally induced pore-fracture evolution law obtained in this study.
Acknowledgments
This work was funded by the financial support by the National Natural Science Foundation of China (52374105, 12572464 and 52204113). The authors would like to express great appreciation to the reviewers for their valuable comments and suggestions that helped to improve the quality of the paper.
The authors declare no competing financial interest.
References
- Jiang C. B., Yang W., Duan M. K.. et al. Pore structure alteration and permeability enhancement of shale under cyclic thermal impacts. Powder Technol. 2022;396:385–393. doi: 10.1016/j.powtec.2021.11.010. [DOI] [Google Scholar]
- Zhao J., Yang L. S., Yang D.. et al. Study on pore and fracture evolution characteristics of oil shale pyrolysed by high-temperature water vapour. Oil Shale. 2022;39:79–95. doi: 10.3176/oil.2022.1.05. [DOI] [Google Scholar]
- Rigby S. P., Jahan H., Stevens L.. et al. Pore structural evolution of shale following thermochemical treatment. Mar. Pet. Geol. 2020;112:104058. doi: 10.1016/j.marpetgeo.2019.104058. [DOI] [Google Scholar]
- Kang Z. Q., Zhao Y. S., Yang D.. et al. A pilot investigation of pyrolysis from oil and gas extraction from oil shale by in-situ superheated steam injection. J. Pet. Sci. Eng. 2020;186:106785. doi: 10.1016/j.petrol.2019.106785. [DOI] [Google Scholar]
- Hasan M. R., Reza M. T.. Hydrothermal deformation of Marcellus shale: Effects of subcritical water temperature and holding time on shale porosity and surface morphology. J. Pet. Sci. Eng. 2019;172:383–390. doi: 10.1016/j.petrol.2018.09.078. [DOI] [Google Scholar]
- Kilic K. I., Abiyev R. H.. Exploiting the synergy between fractal dimension and lacunarity for improved texture recognition. Signal Process. 2011;91:2332–2344. doi: 10.1016/j.sigpro.2011.04.018. [DOI] [Google Scholar]
- Huang X. D., Kang Z. Q., Zhao J.. et al. Experimental investigation on micro-fracture evolution and fracture permeability of oil shale heated by water vapor. Energy. 2023;277:127677. doi: 10.1016/j.energy.2023.127677. [DOI] [Google Scholar]
- Guo X. S., Wang E. Z., Ma X. X.. et al. Lacustrine shale oil systems in China: Advances in characterization methods and resource accumulation models. Earth-Sci. Rev. 2025;270:105256. doi: 10.1016/j.earscirev.2025.105256. [DOI] [Google Scholar]
- Liu Q., Sun M. D., Sun X. D.. et al. Pore network characterization of shale reservoirs through state-of-the-art X-ray computed tomography: A review. Gas Sci. Eng. 2023;113:204967. doi: 10.1016/j.jgsce.2023.204967. [DOI] [Google Scholar]
- Chen S., Zhao W. Z., Ouyang Y. L.. et al. Prediction of sweet spots in shale reservoir based on geophysical well logging and 3D seismic data: A case study of Lower Silurian Longmaxi Formation in W4 block, Sichuan Basin, China. Energy Explor. Exploit. 2017;35:147–171. doi: 10.1177/0144598716679961. [DOI] [Google Scholar]
- Bawazer W., Lashin A., Kinawy M. M.. et al. Characterization of a fractured basement reservoir using high-resolution 3D seismic and logging datasets: A case study of the Sab’atayn Basin, Yemen. PLoS One. 2018;13:e0206079. doi: 10.1371/journal.pone.0206079. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Guan L. J., Xiong T., Jiang Q. T.. Three-dimensional reconstruction and characterization of microscopic pore structure in tight reservoirs. Front. Earth Sci. 2025;13:1563846. doi: 10.3389/feart.2025.1563846. [DOI] [Google Scholar]
- Liu Y. N., Zong P., Wang L.. et al. Quantitative Characterization of Inorganic Pores in Sinian Doushantou Dolomitic Shale Based on FIB-SEM in Western Hubei Province, China. ACS Omega. 2024;9:8151–8161. doi: 10.1021/acsomega.3c08640. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Liang Y. P., Wang S., Feng Q. H.. et al. Ultrahigh-Resolution Reconstruction of Shale Digital Rocks from FIB-SEM Images Using Deep Learning. SPE J. 2024;29:1434–1450. doi: 10.2118/218397-PA. [DOI] [Google Scholar]
- Chen S. Y., Huang Z. K., Zhu G. Y.. et al. Experimental techniques for qualitative and quantitative characterization of shale oil reservoir space, minerals, and fluids. Energy Geosci. 2026;7:100516. doi: 10.1016/j.engeos.2025.100516. [DOI] [Google Scholar]
- Dong F. J., Zhou C., Zhou J. Y.. et al. Study of the pore structure and fractal characteristics of mixed-sediments reservoirs based on low-temperature gas adsorption and high-pressure mercury injection. Carbonates Evaporites. 2025;40:52. doi: 10.1007/s13146-025-01094-0. [DOI] [Google Scholar]
- Chen S., Huang Z., Zhu G.. et al. Quantitative characterization of shale pore structures: A critical review of experimental techniques and numerical methods. Energy Environ. Sci. 2026;19:1234–1267. doi: 10.1039/D5EE03456A. [DOI] [Google Scholar]
- Khan D., Chao L., Qiu L. W.. et al. Depositional Environment and Lithofacies Analyses of Eocene Lacustrine Shale in the Bohai Bay Basin: Insights from Mineralogy and Elemental Geochemistry. Acta Geol. Sin.-Engl. Ed. 2023;97:589–609. doi: 10.1111/1755-6724.14985. [DOI] [Google Scholar]
- Zhao L. F., Cheng Y., Zhang Y. L.. Pore and fracture scale characterization of oil shale at different microwave temperatures. Oil Shale. 2023;40:91–114. doi: 10.3176/oil.2023.2.01. [DOI] [Google Scholar]
- Zhang K., Xi Z. D., Zhang S. H.. et al. Characteristics of Organic Matter Types and Organic Matter Pore Development in Marine-Continental Transitional Shale. Geol. J. 2025;60:2095–2113. doi: 10.1002/gj.5160. [DOI] [Google Scholar]
- Ning S. T., Xia P., Hao F.. et al. Pore Fractal Characteristics between Marine and Marine-Continental Transitional Black Shales: A Case Study of Niutitang Formation and Longtan Formation. Fractal Fract. 2024;8:288. doi: 10.3390/fractalfract8050288. [DOI] [Google Scholar]
- Jiang T. Q., Huang Y., Wu Q. F.. et al. Study on crack propagation and parameter effects of concrete under hydraulic fracturing. AIP Adv. 2026;16:065015. doi: 10.1063/5.0333909. [DOI] [Google Scholar]
- Zhang Y., Yang H.. Direct inversion of 3D seismic reservoir parameters based on dual learning networks. Pet. Sci. 2025;22:4037–4051. doi: 10.1016/j.petsci.2025.05.027. [DOI] [Google Scholar]
- Li J., Dai Y. H., Meng X. B.. et al. Three-dimensional parameter characterization and fracture development prediction of deep tight reservoirs. J. Appl. Geophys. 2026;248:106179. doi: 10.1016/j.jappgeo.2026.106179. [DOI] [Google Scholar]
- Liu M. L., Mukerji T.. Pushing the Boundaries of Porous Material Characterization: Universal Stochastic Data Fusion With Deep Learning. J. Geophys. Res.-Solid Earth. 2025;130:e2025JB031673. doi: 10.1029/2025JB031673. [DOI] [Google Scholar]
- Li L. Y., Tang X., Zhou X. Y.. et al. Experimental study on the improvement of micro-pore structure in medium-low maturity continental oil shale by temperature alteration. Sci. Rep. 2025;15:7706. doi: 10.1038/s41598-025-90784-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gou Q. Y., Xu S., Hao F.. et al. Full-scale pores and micro-fractures characterization using FE-SEM, gas adsorption, nano-CT and micro-CT: A case study of the Silurian Longmaxi Formation shale in the Fuling area, Sichuan Basin, China. Fuel. 2019;253:167–179. doi: 10.1016/j.fuel.2019.04.116. [DOI] [Google Scholar]
- Yuan Y., Yang J. J.. Temperature-Driven Hydrocarbon Generation-Expulsion and Structural Transformation of Organic-Rich Shale Assessed by in situ Heating SEM. Front. Earth Sci. 2021;9:797760. doi: 10.3389/feart.2021.797760. [DOI] [Google Scholar]
- Zheng S., Wang R. F., Li Z. H.. et al. Quantitative Characterization of Microscopic Pore Structure of Tight Sandstone Gas Reservoirs Based on Micron CT Scanning. Chem. Technol. Fuels Oils. 2025;61:408–417. doi: 10.1007/s10553-025-01881-4. [DOI] [Google Scholar]
- Dong Y. T., Song L. M., Lai F. P.. et al. Characterization of the Macroscopic Impact of Diverse Microscale Transport Mechanisms of Gas in Micro-Nano Pores and Fractures. Energies. 2024;17:1145. doi: 10.3390/en17051145. [DOI] [Google Scholar]
- Hu B., Ren Y., Sun R.. et al. Heterogeneity properties and permeability of shale matrix at nano-scale and micron-scale. Phys. Fluids. 2024;36:096613. doi: 10.1063/5.0223200. [DOI] [Google Scholar]
- Kang Z. Q., Zhao Y. S., Meng Q. R.. et al. Micro-CT experimental research of oil shale thermal cracking laws. Chin. J. Geophys. 2009;52(3):842–848. [Google Scholar]
- Wang L., Yang D., Kang Z. Q.. et al. Evolution of permeability and mesostructure of oil shale exposed to high-temperature water vapor. Fuel. 2021;290:119786. doi: 10.1016/j.fuel.2020.119786. [DOI] [Google Scholar]
- Xia X. Z., Xia Y. X., Wang H.. et al. Comparative experimental investigation on pore structure and fractal characteristic of marine and continental shales. Front. Earth Sci. 2025;19:406–422. doi: 10.1007/s11707-025-1155-0. [DOI] [Google Scholar]
- Shang X. J., Wu H., Zhang Z. Z.. et al. A thermally sensitive permeability model for brittle mineral-cracking and fractal microstructure evolution. Phys. Fluids. 2025;37(8):086648. doi: 10.1063/5.0272556. [DOI] [Google Scholar]
