Stimulated reservoir volume fracturing is creating artificial fractures in reservoirs by hydraulic fracturing, which connected with natural fractures and rock beddings to form fracture network^[1]. This technology is the key to the large-scale development of shale gas and the core means to improve the shale gas recovery. 99.5% of the fracturing fluid used in hydraulic fracturing is water and quartz sand^[2,3]. During shale gas development, the flow-back of fracturing fluid is 9% to 53%^[4,5,6], and most fracturing fluid is remained in reservoirs for hydration with shale matrix. Hydration refers to water permeating into the mineral crystalline lattice of rock soil mass or water molecules attached to the ions of soluble rock, causing the microscopic and macroscopic changes of rock structure to reduce the cohesion of rock soil mass^[7]. Hydration is able to change shale microstructure and macrostructure, leading to changes in rock mechanical properties and physical parameters, so as to affect the stability of shale reservoir and the recovery of shale gas^[8,9,10]. Therefore, the research on hydration of shale gas is of great significance for shale gas development.

Previous studies have shown that the strength of hydration is related to the types of clay minerals, the duration of action, the composition of fracturing fluid and other factors. Smectite and illite/smectite mixed-layer minerals have large specific surface areas and high imbibition abilities, so the shale with high contents of such two minerals has strong hydration, prone to soften^[11,12]. Shale hydration and expansion are prone to fractures. Ma and Shi et al.^[13,14] used distilled water to soak shale for studying the generation, expansion and connection of secondary micro-fractures in shale at different stages. It is found that over the duration of hydration, micro-fractures are continuously developed to form fractures with mutual connection, so as to damage the macrostructure of rock. Zhang et al.^[15] injected clean water and KCl solutions of different concentrations in shale permeability test, and the results showed that compared with pure water, the KCl solutions with concentrations of 4% and 8% had a slightly weaker hydration effect on shale reservoir, that is, KCl solution could slightly reduce the damage to shale reservoir permeability. Previous studies mainly focused on the effects of different types of clay minerals and solutions on the hydration strength of shale, and the scale of research was mostly limited to macro-fracture and micron micro-structure. There are few studies on the correlation of the submicron and nanometer microstructure changes and those after hydration with physical parameters.

In this paper, shale samples in Wuxi region around Sichuan Basin are selected, and on the basis of previous research on hydration mechanism, FESEM and computed tomography (CT) scanning technology are used to study the effect from hydration on microstructure changes of shale, compare the changes in shale porosity and permeability before and after hydration, and analyze the impacts of shale hydration degree on physical parameters.

The experimental samples in this study were collected from Wuxi region in Chongqing City around Sichuan Basin. The samples were taken from the Lower Silurian Longmaxi Formation at buried depths of 1 900-1 980 m. The total organic carbon (TOC) content and mineral composition of the shale samples are shown in Table 1. The samples have TOC content of 1.4%-5.3%, and are composed of mainly quartz (content of 30%-40%, on average 34%), feldspar (content of 4%-11%), calcite (content of 7%-17%), dolomite (content of 6%-17%), pyrite (content of 3%-9%), illite (content of 8%-17%), chlorite (content of 5%-12%), and the illite/smectite mixed-layer at the mixed ratio of 15% (content of 8%-15%).

Two samples were selected for microscopic characterization experiments before and after hydration by field emission scanning electron microscope (FESEM), to study the effects of hydration on micro-fractures and micro-pores of shale. The main component of fracturing fluid is water. In order to more directly reflect the hydration mechanism of shale, distilled water was used instead of fracturing fluid to conduct the modeling experiment of shale hydration in this study. Five plug samples were selected for micron CT scanning, porosity and permeability testing. After the test, the samples were put in a confined container for vacuum extraction, and then soaked in distilled water for 24 hours. The soaked samples were put in the centrifuge and centrifuged for 1 hour at centrifugal force of 2.1 MPa. Micron CT scanning, porosity and permeability testing were performed on the samples before and after centrifugation. The sample analysis tests were completed at the National Energy Shale Gas Research and Innovation (Experiment) Center and the Key Laboratory of Unconventional Oil and Gas, CNPC.

1.2.1. Field emission electron microscope scanning

The bulk regular sample with a size of about 2 cm×2 cm was selected. The surface of the samples was roughly polished with sandpaper and polishing liquid, and then polished and coated with carbon by argon-ion polishing instrument. The processed samples were placed in FESEM for scanning. The 685C-type argon ion polishing instrument produced by Gatan Company was used to polish the samples at the polishing voltage of 3 kV for 4 h in this study. In the FESEM experiment, the Helios650 FIB/SEM focused ion beam dual-beam scanning electron microscope produced by FEI Company was adopted to do the job at the scanning voltage of about 5 kV and the working distance of about 4 mm. After scanning, the samples were put into a confined container to vacuum, then soaked in distilled water for 24 h to simulate hydration. Subsequently, the samples were taken out and then placed in 60 °C oven for drying till constant weight. FESEM experiment was repeated on the samples after drying. The change laws of micro-fractures and micro-pores in shale before and after hydration were compared.

1.2.2. Micron CT scanning

The equipment used in micron CT scanning was Xradia MicroXCT-400 CT scanner of Zeiss Company. The scanning voltage used in the scanning experiment was 140 kV, the scanning time was 8 h, and the resolution was 34 μm. First, the plug sample was drilled by a drilling prototype with a diameter of 2.55-2.56 cm and a length of 1.93-2.45 cm(see Table 2). In the process of scanning, the positions of X-ray source and detector kept unchanged, and the samples were rotated at a constant speed from -180° to 180° to capture 5-10 pictures per rotation of 1°.

1.2.3. Porosity and permeability tests

The porosity and permeability of shale samples were tested by Coreval700 poroperm tester produced by VINIC Company. The gas medium used in the test was nitrogen with a purity of 99.99%. The porosity was tested at the confining pressure of 10 MPa and the pore pressure of 1.4 MPa. Measurement procedure of porosity is as follows: The apparent density of the sample was measured by means of the rock density meter, the skeletal density of sample was measured by poroperm tester, and the porosity of the sample was calculated by equation (1).

where ρ_b—apparent density, g/cm³; ρ_g—skeletal density, g/cm³; ϕ—porosity, %.

The permeability test method was the pressure decltion method^[16], which belongs to the unsteady method. The principle of gas permeability tester by pressure decltion method is shown in Fig. 1. The test steps are as follows: The plug sample is placed in the core holder under certain confining pressure; the downstream valve is closed to fill the upstream chamber with a certain amount of gas; after the pressure is balanced, the downstream valve and connecting valve are turned on successively. As the gas flows from the upstream chamber to the downstream end through core, the pressure in the upstream chamber decreases continuously. The permeability of plug sample is calculated according to the rate of pressure reduction in upstream chamber. The permeability in this study was tested at the confining pressure of 10 MPa and pore pressure of 1.4 MPa.

The variation of microstructure of samples W-3 and W-5 before and after hydration were compared (Figs. 2-5). FESEM experiments show that the organic matter in shale is mostly distributed in strip and dispersion. The organic matter is a plastic material susceptible to the effect of compaction, so the stripped organic matter has few internal pores (Fig. 2a), the scattered organic matter has more pores from 10 nm to 300 nm in diameter (Fig. 4). Organic matter and minerals are different somewhat in stress sensitivity. When the core is taken out from the stratum, micro-fractures less than 1 μm wide are likely to occur due to stress release (Fig. 2a). The pyrite in shale often appears in strawberry shape from a few microns to dozens of microns in diameter (Fig. 3a).

Comparing the images of two samples before and after hydration, it is found that the main morphology and position of organic matter in shale have no changes after hydration (Figs. 2 and 3), and the pores in dispersed organic matter show no obvious changes either (Fig. 4). The main reason is that the shale organic matter is the kerogen insoluble in non-oxidizing acids, alkali and non-polar organic solvents. Water is adsorbed on the surface of kerogen during soaking, but no dissolution nor chemical reaction occurs with organic matter^[17]. The organic pores are formed after long thermal evolution of organic matter, and are main space for shale gas storage. The pore size of organic matter can also guarantee the flow of water molecules in the pores. Therefore, hydration has no obvious influence on the size and structure of pores in shale organic matter. After hydration, the fractures in shale increase in number and width, mainly shown as the extension of fractures between inorganic minerals and the creation of new fractures (Figs. 2b and 3b), and the width of fractures between stripped organic matter and inorganic minerals increase (Fig. 5b). During the soaking process, self-absorption effect happens to water along the fracture, and the stress of hydration is concentrated at the fracture tip, leading to the extension of fractures between inorganic minerals and the creation of new fractures^[18]. The width of newly-created fracture is basically less than 10 μm. In clay minerals, Na⁺, K⁺, Ca²⁺ and other positive ions in illite and illite/smectite mixed-layer minerals can be hydrated and dissolved^[18]. When water is discharged, these positive ions would gather on the surface of clay minerals, making the fractures between inorganic minerals and organic materials grow wider. After hydration, a large number of inorganic pores are formed in shale with sizes ranging from several microns to dozens of microns. These inorganic pores are mainly caused by the weak cohesion between clay minerals and non-clay minerals and the detachment of non-clay minerals after water soaking (Figs. 2b and 3b). In real stratigraphic conditions, because shale is in a confined state in the stratum, non-clay minerals are not easy to fall off to produce inorganic pores, but would be transformed into micro-fractures between clay minerals and non-clay minerals.

The water saturation experiment of shale samples was used to simulate hydration, and centrifugation experiment was adopted to simulate drainage process during shale development. Micron CT scanning was carried out on samples in three states, i.e., before hydration, after hydration and before centrifugation as well as after centrifugation experiment (Fig. 6). According to the principle of CT imaging, the higher the density of the medium, the brighter the medium image will be, and the lower the density, the darker the medium image will be. In Fig. 6, the white part is pyrite and the black part is fracture or pore. The selected samples have micro-fractures, samples W-1, W-2 and W-5 have more fractures, but no micron pores, and stripped pyrite is seen in some samples (Fig. 6a, 6g and 6m). Under the real stratigraphic pressure conditions, micro-fractures are usually closed, but these regions are the weak surfaces of shale with low cementing strength, and would be broken earlier than shale matrix, so that hydraulic fracturing can extend along these micro-fractures to achieve fracture expansion and volume stimulation^[19].

By comparing the results of micron CT scanning, it is found that hydration can make shale fractures extend and generate new fractures, of which the mechanism is the same as FESEM experiment. Hydration damage mainly extends along the developing direction of bedding or original fractures, and hydration has the most significant impact on the shale samples W-1, W-2 and W-5 with relatively developed fractures (Fig. 6b, 6e and 6n). Therefore, hydration intensity is controlled by the development degree of primary fractures, and the more developed the primary fractures, the stronger the hydration effect will be^[20]. As the hydration time prolongs, these fractures would connect with each other to reduce the strength of the core and cause macroscopic damage of the rock^[14]. After water is discharged from shale samples by centrifugal experiments, the fractures increase in width, to 2-5 times of the original width. Among the samples, samples W-1 and W-5 have the most obvious increase in fracture width (Fig. 6c and 6o).

The results of FESEM and micron CT scanning experiments show that the shale samples selected have organic pores from 10 to 300 nm in size and micro-fractures from tens of nanometers to hundreds of microns wide. The poroperm tester was used to test porosity and permeability of the shale samples in three states, before hydration, after hydration and before centrifugation as well as after centrifugation. The measured porosity is effective porosity, ranging from 0.05% to 2.71% (Fig. 7). The samples have organic matter pores (Fig. 4), providing space for shale gas storage. Therefore, the samples with high TOC content have higher porosity, and the two are positively correlated (Fig. 7).

For conventional sandstone, porosity has a good correlation with permeability. Gamage et al.^[21] found that porosity was positively correlated with the logarithm of permeability. The permeability distribution range of shale samples selected in this study is (0.01-0.30)×10^-3 µm², and the logarithm value of permeability is basically positively correlated with porosity (Fig. 8). In the meantime, the permeability is controlled by the degree of fracture development. Although the content of TOC and porosity of shale sample W-1 are low, its permeability is high due to relatively developed fractures. Apparently, shale permeability is mainly affected by porosity and fracture development degree.

According to the porosity test results of shale samples under different states (Fig. 9), the water saturation of shale samples after hydration basically reached 100%, and the porosity of samples reached the detection limit of the instrument, so the indication of porosity test was basically zero. After the centrifugal experiment, most of movable water was centrifuged out, leading to the decrease in water saturation of the samples and significant increase in porosity, but the porosity increase differs widely in different samples. The porosities of samples W-2 and W-4 after centrifugation were lower than those before hydration. The porosity of sample W-3 after centrifugation was equivalent to that before hydration. The porosities of samples W-1 and W-5 after centrifugation are greater than those before hydration. Combined with the results of micron CT scanning (Fig. 6), it is found that samples W-2 and W-4 have few fractures developed and high bound water content after hydration. After centrifugation experiment, its water saturation doesn’t decrease much, leading to the porosity test result lower than that before hydration. After the hydration, sample W-3 have more fractures developed and high movable water content. After centrifugation experiment, its water saturation drops, so the test result of porosity is basically equivalent to that before hydration. After hydration, samples W-1 and W-5 present increase in number and width of fractures most significantly. After centrifugation experiment, their water saturation decrease and fractures expand, so the test results of their porosities are larger than those before hydration. The porosity of sample W-1 increases from 0.05% before hydration to 0.32%, and that of sample W-5 increases from 2.71% before hydration to 3.10%. Therefore, the degree of fracture development has certain influence on shale porosity.

According to the permeability test results of shale samples in different experiment states (Fig. 10), the permeability of samples after hydration is lower than that before hydration. The main reason is that saturated water occupies the pores, fractures and throats of shale, leading to a decline in the nitrogen relative permeability of shale. After the centrifugation experiment, the rock permeability increased as the fluid was drained out, but the degrees increased are different. The permeabilities of samples W-2, W-3 and W-4 recovered to those before hydration. The permeabilities of samples W-1 and W-5 were greater than those before hydration. Combined with the results of micron CT scanning, it can be seen that the fractures of samples W-1 and W-5 increased significantly in width after centrifugation experiment, leading to the permeabilities greater than those before hydration. Therefore, shale permeability is mainly controlled by the degree of fracture development, and secondarily by porosity.

After hydration, the organic matter and organic pores of shale are basically unchanged, but the fractures are likely to form between organic matter and inorganic minerals, and micro-fractures are also prone to be generated or derived in inorganic minerals. Meanwhile, the cohesive force between mineral particles is weak. After hydration, the non-clay mineral particles fall off to form inorganic pores of several microns to dozens of microns in diameter. In real stratigraphic conditions, these inorganic pores may not be generated, but would be transformed into micro-fractures between clay and non-clay mineral particles.

The intensity of hydration is controlled by the development degree of primary fractures. The more developed the primary fractures, the stronger the hydration degree will be, and the more likely the secondary fractures will be developed. The experimental results of shale hydration show that the content of shale TOC plays a leading role in porosity, and the higher the content of TOC, the larger the porosity will be. Secondly, shale porosity is affected by the degree of fracture development. The permeability is influenced by the degree of fracture development and porosity, of which the degree of fracture development plays a leading role. After hydration, if the width of shale fracture increases significantly, the porosity of shale would increase slightly than the original porosity of the sample, while the permeability would increase multiply. When the fracture width increases little, the porosity of shale is smaller than that of the original sample, and the permeability is equivalent to that of the original sample. After hydration, the degree of fracture development has an influence on both porosity and permeability, but the influence on permeability is bigger than that on porosity.