Permeability Evolution of Shale during High-Ionic-Strength Water Sequential Imbibition

Abstract

It is widely accepted in the oil and gas industry that high-ionic-strength water (HISW) can improve oil and gas recovery in unconventional shale reservoirs by limiting shale hydration. Despite numerous supporting studies, there is a lack of a systematic analysis exploring the effect of HISW on shale permeability evolution, particularly considering varying chemical compositions. In this work, we investigated the impact of different concentrations of NaCl and CaCl₂ on shale permeability through sequential HISW imbibition experiments, beginning with the highest NaCl and lowest CaCl₂ concentrations. After maintaining the highest effective stress for an extended period, significant permeability reduction and potential fracture generation were observed, as indicated by periodic fluctuations in differential pressure. These effects were further intensified by displacements with HISW solutions. Advanced post-experimental analyses using micro-CT scans and SEM-EDS analysis revealed microstructural changes within the sample. Our findings offer initial insight into how HISW-shale interactions influence shale permeability, using innovative approaches to simulate reservoir conditions. The findings indicate that discrepancies in the chemical composition between injected solutions and shale may lead to shale disintegration during hydraulic fracturing processes.

1. Introduction

Shale oil and gas exploration has drawn attention worldwide due to its abundant potential to fulfill the growing demands in global industries and transportation. Compared to conventional reservoirs, unconventional shale plays are buried at greater depths of thousands of meters beneath the Earth’s surface in elevated geothermal and high-pressure environments. As a result, shale has a lower porosity and is considered near-impermeable for fluids to flow through without pathways being formed within the rock. For such circumstances, advanced drilling and stimulation technologies have been developed over the past few decades to enhance the ability to produce from these tight rocks. Hydraulic fracturing (or fracking) has been in the public eye since the 1950s and has been developed along the way. By injecting fracturing fluid at high pressure, fractures are created in the rock, extending away from the wellbore. The composition of the fracturing fluid can significantly affect the mechanical [1,2] and petrophysical properties [3,4,5] of shales; therefore, it needs to be carefully examined to achieve the desired production outcome. Meanwhile, the salinity of the produced water increases as larger proportions of native formation water are being depleted over time. The disposal of the high-salinity mixture comprised of native formation water and engineered fracturing fluids can be costly. To accommodate these problems, reinjection of the produced water has become popular and has been implemented with some success in shale plays. However, the interaction between the injection fluid composition and shale is still poorly understood.

It is worth noting that ‘shale swelling’ and ‘clay swelling’ are different concepts in the context of fluid-shale interaction. Clay swelling is a result of the exchange of ions in clay minerals with water molecules, while shale swelling can be influenced by various factors related to the interactions between the bulk shale matrix and fluids. Liu et al. stated that clay swelling is a key phenomenon occurring in water-shale interaction [6]. Clay swelling triggers the generation of fractures, hence causing detachment of particles, weakening of the mechanical properties of shales, and affecting the evolution of permeability. Important reactions involving clay include: (1) a tendency of clays to react and attempt to equilibrate water/sediment systems; (2) adsorption of cations by clays because of the high cation exchange capacity; and (3) colloidal dispersion of clays [7]. Clay swelling consists of intracrystalline swelling and osmotic swelling, which refer to the phenomenon of water entering the interlayer resulting in the expansion of the space between the pallets, and a large difference in the ion concentration between the clay surfaces and the fluid. The degree of swelling depends on the type and quantity of clay minerals encountered [8]. For smectite group clays (montmorillonite) composed of stacking several layers of a sheet structure of tetrahedra arranged in an octahedral network, the spacing between the layers can be easily separated by adsorbed water molecules and cations, thus giving it the potential to expand more than the other multilayer phyllosilicates [7].

On the other hand, Dehghanpour et al. reported that shale samples with illite and a negligible amount of smectite still exhibit significant brine intake [9]. Zeng et al. pointed out that the disjoining force in organic matter increases with the decrease in NaCl concentration, leading to reduced electrostatic repulsion between organic matter surfaces [10]. Consequently, more water molecules penetrate the smaller pore spaces within the shale, hence increasing the internal pore pressure, causing shale sample expansion, and propagation of existing microfractures. According to Yang et al., organic matter, being the softer component, creates space for the propagation of fractures under reservoir conditions [11]. They also noted that the reduced capillary pressure in organic matter promotes water imbibition, which contributes to fracture growth in the shale matrix. A study by Zeng et al. found that reducing the brine salinity, especially the salinity of monovalent cations such as potassium, constantly increases the electrical double layer expansion, which in turn increases the repulsive force between organic matter surfaces and triggers an increase in the hydrophilicity of the system [12].

Takeda et al. conducted a chemical osmosis experiment to examine fluid imbibition in siliceous shale samples by introducing HISW solutions. They observed that the differential pressure across the core increases rapidly in less than a day due to the osmotic flow from low-concentration to high-concentration solutions [13]. Roshan et al. suggested that ion exchange reactions are the main driving force behind the water uptake [14]. Zolfaghari et al. measured the ion concentrations during water-shale imbibition experiments and observed leaching of exchangeable sodium/potassium from shale samples due to their high cation exchange capacity [15]. Fritz et al. proposed that shale can be treated as a semipermeable membrane since it partially restricts the flux of ions [16]. Al-Bazali et al. observed in their studies that the amount of associated water overtakes the amount of osmotic water that moved into the shale over time. They summarized three distinct stages of shale hydration when exposed to water-based fluids: (1) water movement out of the shale (chemical osmosis); (2) water movement into the shale (diffusion osmosis); and (3) stationary state (balanced ionic concentration between the shale and salt solutions) [17]. Since there is a considerable salinity difference between the injecting fluid and the shale formation brine during the treatment of hydraulic fracturing, Wang et al. emphasized the importance of chemical osmosis and its consideration, since it intensifies fluid retention in the shale [18].

Previously, studies have primarily focused on water-shale interactions. Morsy and Sheng found that shale samples from highly saline depositional environments tend to generate fractures when coming into contact with fresh water [19]. The large osmotic potential due to the high salinity contrast between the fracturing fluid and connate water contributes to spontaneous imbibition, thereby altering shale mechanical properties. Liu and Sheng conducted water-flooding tests on fractured shale cores, revealing a significant decrease in fracture conductivity due to shale water uptake [20]. They concluded that clay swelling, temperature changes, and stress anisotropy control fracture generation. Existing fractures within the shale tend to intersect with pores, forming a more complex network under atmospheric pressure [18]. Fracture generation under confining pressure takes a longer time, as demonstrated by Roshan et al., Qian et al. and Wang et al. [14,21,22]. Confining pressure negatively impacts clay swelling and fracture propagation, as observed by Zhang and Sheng, and Ewy and Stankovic [23,24]. However, under pore pressure anisotropy, the higher deviatoric stress from confining pressure can enhance fracture generation [20,25,26]. Fractures generated in shale without forming effective flow channels do not necessarily improve permeability. Factors such as particle detachment from fracture surfaces [27], clay particle dispersion [28], and water blocking [29] can impair fracture conductivity. In the realm of HISW-shale interactions, there are fewer studies available. Xu and Dehghanpour conducted shale imbibition tests using NaCl solutions [30]. They found that increasing the salinity of the tested solutions decreases the osmotic effect and reduces shale hydration. Russell et al. indicated that high-salinity fluid can potentially enhance rock permeability by reducing electrostatic barriers and causing the dispersed clay fines to reattach to the rock surface [31]. Zeng et al. identified two key contributions of HISW-shale interactions: (1) increased water uptake in calcite-rich samples with lower content of organic matter, and (2) a minor effect of mineral dissolution due to the salinity increment [32]. Conversely, Mehana et al. observed a greater impact of HISW on fluid-shale imbibition, suggesting that lower salinity does not always lead to greater water uptake [33]. This finding challenges the general belief that HISW can hinder or prevent shale degradation. In addition to the salinity difference between the treating fluids and shale, imbibition can also be driven by capillary pressure and hydraulic pressure [34]. As most shale imbibition experiments have been conducted without fluid pressurization and continuous fluid flow, these results may not accurately predict field application outcomes.

Initially, we subjected a fractured shale sample to loading-unloading cycles under varying effective stresses with nitrogen flooding, resulting in irrecoverable permeability reduction due to plastic deformation and partial compressibility recovery, as presented in Bai et al. [35]. Subsequently, in this study, we conducted sequential HISW imbibition experiments to explore the impact of HISW-shale interactions as influenced by pore pressure, effective stress, and fluid composition. Upon completing these experiments, micro-CT scans and SEM-EDS analysis were performed to comprehensively assess potential changes in the sample induced by HISW-shale interactions.

2. Materials and Experimental

The compositions of the shale sample and applied solutions are presented in Section 2.1 and Section 2.2 below. Fracture surface roughness measurement, micro-CT scan, SEM-EDS analysis, and core-flooding experimental setup are detailed in Section 2.3, Section 2.4 and Section 2.5, respectively. The experimental system and experimental method of the sequential HISW imbibition are specified in Section 2.6 and Section 2.7, respectively.

2.1. Shale Sample

The shale sample used in this study originates from the Gulf Coast Basin in South Texas at a depth of 1914.45 m beneath the ground surface. This sample is light gray, fine-grained, and well-cemented with thinly laminated bedding. A few faint cleats, angled at 40–45 degrees relative to the bedding orientation, are visible on the surface. The sample exhibits low porosity, recorded at 5.1%, derived from the average of eight measurements using a helium porosimeter. The core was extracted from a horizontal well, chosen for its proximity to the targeted field, with the bedding plane aligned parallel to the core’s long axis. The mineralogical composition of the shale sample was determined by a quantitative X-ray diffraction (XRD) analysis using a D8 ADVANCE Powder X-ray Diffractometer sourced from Bruker Co., Billerica, MA, USA, with a Cu-radiation source. Mineral phases were identified through data processing by the Bruker DIFFRAC.EVA software (version 5.2) and Crystallography Open Database reference patterns and quantified by the TOPAZ (version 5) profile-fitting based software. As indicated in

Table 1. Mineral compositions of the selected shale sample.

Compositions	Percentage (%)
Quartz	51.70
Calcite	26.72
Muscovite	4.80
Montmorillonite	3.91
Illite	9.78
Pyrite	0.72
TOC	2.37

, the primary components are quartz and calcite, with minor amounts of clay and organic matter.

For preparation, a cylindrical plug with a 24.02 mm diameter and 31.55 mm length was cored from a shale block. This plug was cut perpendicular to its short axis and at a 30-degree angle to the bedding plane using a diamond saw. Water was used as a coolant to reduce frictional heating during cutting. To reduce the effect of clay mineral-water interaction, cutting was executed at a high speed, typically at 3000–3500 rpm, and the sample was immediately dried and wrapped in lint-free Kimwipes science wipers after cutting to minimize exposure time. The end surfaces were gently abraded using 1200-grit sandpaper (3.8 µm particle size) without the use of coolant fluids, specifically to eliminate the surface particles that had been exposed to water. Subsequently, the sample was oven-dried at 60 °C for 72 h and then subjected to vacuum treatment for 96 h to remove original pore water. During this period, the sample was weighed every 24 h to monitor water mass changes. The oven temperature was kept below the geothermal conditions of the sample’s environment to prevent the collapse of clay minerals.

2.2. Composition of Aqueous Solutions

The selected saline compositions were determined based on existing bore-log data from South Texas. Most of the data were sourced from the U.S. Geological Survey (USGS) National Produced Waters Geochemical Database [36], while minor data on brine in the aquifers were gathered from Nicot et al. and Macpherson [37,38]. The existing information indicates that the concentration of Na⁺ and Cl⁻ ranges from 10,000 ppm to 30,000 ppm and from 50,000 ppm to 70,000 ppm, respectively. The concentration of Ca²⁺ varies significantly across the region from 1000 ppm to 18,000 ppm. Such high-ionic concentrations are likely due to the geological settings of the Texas Gulf Coast Basin, where deep formation waters are formed from surrounding seawater and have undergone substantial subsurface rock-fluid interactions, utilizing mechanisms and processes that include non-marine pore waters, dolomitization of carbonates, diagenetic alteration, and bacterial sulfate reduction [39]. Alternative hypotheses include the burial of evapo-concentrated seawater [40] and ultrafiltration [41,42].

To ensure the aqueous solutions have the same ionic strength but different concentrations of NaCl and CaCl₂, it is crucial to understand how each ion contributes to the overall ionic strength of the solutions. The ionic strength (I) is the measure of the total concentration of ions in a solution multiplied by the square of their respective charges:

$$\mathrm{I}={\displaystyle \frac{1}{2}}{\sum }_{\mathrm{i}}{\mathrm{C}}_{\mathrm{i}}{\mathrm{Z}}_{\mathrm{i}}^2$$

(1)

where C_i is the molar concentration of ion I and Z_i is the charge of ion i. The molarity applies to the solute formula. For instance, if the amount of NaCl dissolved is 1 mole, there are 1 mole of Na⁺ ions and 1 mole of Cl⁻ ions present in the solution. Similarly, a 1 M (mol/L) CaCl₂ solution contains 1 mole of Ca²⁺ ions and 2 moles of Cl⁻ ions per liter of the solution. The calculations of individual contributions can be expressed as:

$${\mathrm{I}}_{\mathrm{N}\mathrm{a}\mathrm{C}\mathrm{l}}={\displaystyle \frac{1}{2}}(\mathrm{C}({\mathrm{N}\mathrm{a}}^{+})\times 1^2+\mathrm{C}({\mathrm{C}\mathrm{l}}^{-})\times 1^2)$$

(2)

$${\mathrm{I}}_{\mathrm{C}\mathrm{a}{\mathrm{C}\mathrm{l}}_2}={\displaystyle \frac{1}{2}}(\mathrm{C}({\mathrm{C}\mathrm{a}}^{2+})\times 2^2+2\mathrm{C}({\mathrm{C}\mathrm{l}}^{-})\times 1^2)$$

(3)

$${\mathrm{I}}_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}={\mathrm{I}}_{\mathrm{N}\mathrm{a}\mathrm{C}\mathrm{l}}+{\mathrm{I}}_{\mathrm{C}\mathrm{a}\mathrm{C}{\mathrm{l}}_2}$$

(4)

These calculations provide the flexibility to compute the Na-Ca-Cl ratio while maintaining a constant ionic strength in each solution and keeping the concentrations within the desired range. An ionic balance check was then conducted to ensure that all solutions are electrically neutral. The aqueous solutions were prepared by dissolving a predetermined amount of NaCl (Chemsupply, 99.7% purity) and CaCl₂ (Chemsupply, 99% purity) in Milli-Q deionized water. All solutions to be injected were degassed by vacuuming for one hour to remove dissolved air and were filtered using a 0.2 µm Nuclepore Track-Etched Polycarbonate membrane filter to separate impurities such as atmospheric dust. As shown in

Table 2. Compositions of the aqueous solutions.

	Solution #1	Solution #2	Solution #3
NaCl (wt. %)	11.25	7.3	2.91
CaCl2 H2O (wt. %)	0.37	3.67	7.34
TDS (ppm)	115,247	100,734	84,426
Ionic strength (M)	2.0	2.0	2.0
Electrolytic conductivity (mS/cm)	149	132.4	116.4
pH	7.22	7.11	6.75

, all injected solutions had the same ionic strength of 2.0 M. The electrolytic conductivity and pH of the solutions were measured after preparation to ensure adequate quality. The decided fluid compositions and specifications are summarized in Table 2 below:

2.3. Fracture Roughness Measurement

The fracture topographies of the sample were measured using an OLS5100 Laser Microscope sourced from Olympus Co., Hachioji, Tokyo, Japan, with a 0.4 $\mathsf{\mu }$m laser beam. This confocal microscope captures multiple images based on focus position and light intensity. The images were formed by joining pixels defined by the selected objective lens. A 10× objective was used for a panoramic view, stitching together the captured images. Each image contains 4096 × 4096 pixels to optimize measurement performance. The fracture surfaces were positioned 10.4 mm away from the lens, and their slope was fixed using a small piece of polymer clay between the sample and the stage (Figure 1). Each fracture surface asperity height was measured with a high resolution of 0.5 nm. The target area was identified on the panoramic view by stitching the images in a linear direction, matching the flow direction from the core inlet to the outlet. For each fracture surface, three narrow rectangular areas with a 0.8 cm sampling interval were measured by drawing a line on the surface. The surface roughness profile was processed through noise elimination and inclination correction to ensure data accuracy. The main horizontal plane was detected at the zero height position to enhance the reliability of the data.

2.4. Micro-CT Scan Analysis

To capture fractures generated under stress anisotropy, a micro-CT scan was performed using a 1276 Scanner sourced from Bruker Co., Billerica, MA, USA, featuring a high spatial resolution down to a 2.8 µm pixel size. This scanner’s variable X-ray energy, combined with a range of filters, ensures optimal image quality for various research applications. The system conducted the scanning with a continuous gantry rotation in a step-and-shoot mode, achieving the fastest scanning cycle of 3.9 s. A voltage of 90 kV and a spatial resolution of 15 µm were selected for the scan. These settings were chosen to shorten the imaging duration and minimize radioactive decay due to the partial absorption of X-rays by the rock minerals. The sample was positioned between an X-ray source and a detector, with a rotation step of 0.3 degrees totaling 180 degrees over three hours. Throughout the scanning process, the sample was rotated, repositioned, and resliced using different orientations to enhance visualization through the DATAVIEWER software (version 1.6.0). For quantitative analysis, the acquired orthogonal slices were imported into the CTAN software (version 1.23.02), which includes identification of fractures and selection of regions of interest. To reduce the data size, the size of the orthogonal slices was adjusted to match the cross-sectional area of the sample. The 2D projection images were subsequently reconstructed using the CTVOL software (version 2.3.2.0), enabling a flexible 3D viewing environment.

2.5. SEM-EDS Analysis

The SU7000 FE-SEM sourced from Hitachi Co., Ltd., Chiyoda City, Tokyo, Japan, was used to perform three-dimensional imaging. To mitigate charging effects, a thin carbon coating was applied to the fracture and end surfaces of the specimen. It was then mounted on a stub using adhesive conductive double-sided tape. A small area at the edge of the surfaces was also covered with conductive tape to establish a conduction path to the stub, ensuring clear image production. Subsequently, the stub was secured on a sample holder and inserted into the sample chamber. An operating voltage of 15 kV and a working distance of 6 mm from the electron beam were set to achieve optimal resolution. The machine is equipped with secondary electron (SE) and backscattered electron (BSE) detectors for topographic imaging and composite analysis, respectively. Both detectors were managed via a console, which allowed for precise adjustment of the beam position and correction of astigmatism by turning the corresponding knobs for alignment and refocusing. Multiple spots on the core inlet and fracture surfaces were selected to observe the post-experiment microstructure of the sample. For surface mineral identification, SEM-EDS (Energy Dispersive X-ray Spectroscopy) analysis was performed using the BSE detector, enabling the generation of elemental maps over the areas of interest.

2.6. Experimental Setup (Former Core-Flooding System)

The experimental setup is shown schematically in Figure 2. The fractured shale core (1) was wrapped in a heat-shrinking sleeve to prevent any leakage between its outer surface and the Viton sleeve (2). The shale core, along with the Viton sleeve and two stainless-steel flow distributors (3), were placed inside the Hassler-type hydrostatic core holder (4). The confining fluid/distilled water (6) within the core holder was pumped by a manual pressure generator (5) compressing water and measured by an absolute pressure transducer (7) with a 0 to 4351 psi measuring range. The injection of liquid was carried out by a Prep-36 HPLC pump (10). The separation of the liquid injection system from that for gas injection was carried out by a two-way shut-off valve (12). Two pressure transmitters (13, 14) were used to measure upstream and downstream pressures on the shale core. A compressed gas cylinder (15) delivered the gaseous nitrogen to the core inlet via a pressure regulator (16), two-way mass flow controller (17), and two-way valves (18, 19). The differential pressure across the sample was measured by differential pressure transmitters (20–23) with ranges from 0–0.58 psi to 0–2030 psi and above for precise measurements. Pressures higher than 2030 psi could be measured via absolute pressure transducers (13, 14). These differential pressure transmitters were connected to the core holder and the remaining tubing by three-way valves (24–27). Output signals from the pressure transmitters were directed to a data acquisition system comprising a signal conditioner (28), analog input module (29), and a computer (30). Pore pressure was controlled by a back-pressure regulator (31) using compressed air from the gas cylinder via a pressure regulator and two-way valve. The compressed nitrogen gas from a cylinder (32) was delivered to the core holder via a gas pressure regulator (33) and a two-way valve (34). Electrolytic conductance of effluents was measured by an electrolytic conductivity sensor (35) with a signal conditioner (36). Effluent from the core holder saline solution stream was delivered to the test tubes (38) sitting in a sampler carousel (37).

2.7. Sequential HISW Imbibition Experimental Procedure

Initially, the sample was loaded step-by-step with effective stresses at 1000 psi, 1500 psi, 2000 psi, and 2500 psi. A pump flow rate of 0.1 mL/min was used to allow for the solutions to have more interaction time with the shale. Each stress stage lasted for 24 h, mirroring nitrogen flooding experiments for a comparative study of permeability changes with effective stress. Once the sample was loaded to the highest effective stress of 3000 psi, the experiment was extended as the differential pressure kept increasing substantially without signs of stabilization. The confining pressure was adjusted to maintain a constant 3000 psi effective stress. However, due to system limitations, the confining pressure could only be manually adjusted through a hand pump. The experiments were conducted at a room temperature ranging from 23 °C to 24 °C during the day. However, at night, the temperature dropped to a minimum of 21 °C. This caused the water used as a confining fluid to shrink and subsequently the confining pressure decreased. Therefore, sometimes the change rate of the confining pressure lagged behind the pore pressure but still remained within an acceptable effective stress range (±50 psi).

At this point, saline displacement tests were performed, monitoring differential pressure data to reflect the changes in shale permeability. Pore pressure was allowed to increase as a function of upstream pressure while maintaining constant downstream pressure. Considering the unpredictable nature of the shale response, the reserved duration for each solution depended on the prompt feedback and the capacity of the experimental system. The injection periods for Solution #1, Solution #2, and Solution #3 were 257 h, 214 h, and 285 h, respectively. Sequential displacements of the solutions were conducted to study permeability changes related to interactions between the solutions and the shale sample. As the pore pressure increased with the upstream pressure, at the beginning of each displacement phase it was reduced to the initial value of 435 psi by adjusting the downstream pressure through a back-pressure regulator. During the second and third injection phases, the downstream pressure was manually reduced to 300 psi and 130 psi, respectively, due to upstream pressure buildup. This alteration had minimal impact on the differential pressure, as both upstream and downstream pressures decreased similarly without causing disparities.

3. Results and Discussion

This section presents the measurements of fracture roughness (Section 3.1) and the sequential HISW imbibition (Section 3.2). The discussion on HISW-shale interactions is divided into Section 3.2.1 and Section 3.2.2.

3.1. Fracture Roughness Profile

The fracture surface roughness can be evaluated using the roughness parameters Sq, Sa, Sz, Sp, and Sv. The height parameters are determined solely by the distribution of asperity height information. Sq (root mean square height) is the root mean square value of ordinate values within the definition area A. It corresponds to the standard deviation of the height distribution and is calculated by the following equation:

$${\mathrm{S}}_{\mathrm{q}}=\sqrt{{\displaystyle \frac{1}{\mathrm{A}}}\underset{\mathrm{A}}{\iint }{\mathrm{Z}}^2\left(\mathrm{x},\mathrm{y}\right)\mathrm{d}\mathrm{x}\mathrm{d}\mathrm{y}}$$

(5)

Sa (arithmetic mean height) is the mean difference in height from the mean plane, given by:

$${\mathrm{S}}_{\mathrm{a}}={\displaystyle \frac{1}{\mathrm{A}}}\underset{\mathrm{A}}{\iint }\left|\mathrm{Z}(\mathrm{x},\mathrm{y})\right|\mathrm{d}\mathrm{x}\mathrm{d}\mathrm{y}$$

(6)

When the height distribution is normal, the relationship between the parameters Sq and Sa becomes Sa ≒ 0.8 × Sq. Sz (equal to Sp + Sv), Sp, and Sv are maximum height, maximum peak height, and maximum pit depth, respectively.

The digitized data from the fracture surface were analyzed using associated processing software to compute height parameters following the reconstruction of the 3D surface map. A Gaussian filter was applied to suppress noise while enhancing height differences, with adjustments made to the color bar scale to distinctly illustrate topographic features within the map. The 3D profile of a fracture surface comprises numerous cross-sectional 2D profiles in the flow direction, which can be averaged to determine the line roughness profile Z (Figure 3). It is important to note that converting the roughness profile to Joint Roughness Coefficient (JRC) values is not applicable due to the smoothness of the fracture surfaces, rendering the JRC roughness scale ineffective. The JRC scale is more attuned to variations in roughness within specific ranges, particularly in moderate to high roughness categories. Assigning a JRC value to a highly polished surface may lead to a misinterpretation of its actual roughness characteristics. Furthermore, employing sampling intervals such as 0.5 mm, 1 mm, and 2 mm proposed in previous studies [43,44,45] can result in an insufficient number of data points for accurately estimating the JRC values of the 2D profiles of the surfaces. Instead, the outlined roughness parameters are calculated by assigning submicron sampling intervals, allowing for a more precise analysis and comparison. The averaged values of these parameters from three evaluation areas on each surface are presented in

Table 3. Average fracture roughness parameters of six scanning lines on both fracture surfaces.

Sq (μm)	Sp (μm)	Sv (μm)	Sz (μm)	Sa (μm)
3.56	50.64	54.53	105.66	2.63

. It is worth noting that the values of the displayed parameters represent the average value of both fracture surfaces, a total of six scanning lines, as measured by the profilometer. The equations mentioned above apply only to the roughness of a single scanning line.

3.2. HISW Sequential Imbibition Experiments

Prior to the HISW sequential imbibition experiments, three loading-unloading cycles under nitrogen flooding were conducted with effective stresses ranging from 100 psi to 3000 psi, along with corresponding confining pressures from 535 psi to 3435 psi [35]. Nitrogen, as a non-reactive gas, was selected to develop a baseline for the shale mechanical behavior and to simulate in-situ cyclic scenarios without introducing additional variables as shales are sensitive to interactions with fluid. This allowed for direct comparison between the results obtained during the cyclic loading-unloading experiments and the subsequent HISW sequential imbibition experiments.

3.2.1. Solution #1 Imbibition Experiments

During the Solution #1 imbibition, the effective stress was incrementally increased from 1000 psi to 3000 psi, matching the loading steps of the second and third cycles under nitrogen injection. Considering the disparity between gas and fluid permeability, a modified permeability ratio was implemented to allow for direct comparison of permeability change rates, as shown in Figure 4. During the fourth loading cycle under Solution #1 injection, permeability behavior began to diverge from the patterns observed under nitrogen injection early on, with this deviation becoming more pronounced at higher effective stresses. Typically, increasing loading cycles should lead to a decrease in the rate of permeability change due to sample consolidation. However, the observed acceleration in permeability reduction was likely caused by shale swelling, suggesting spontaneous imbibition upon exposure to the solution.

Figure 5 shows that the differential pressure stabilized at all effective stress stages lower than 3000 psi after 24 h under the injection of Solution #1. Upon the effective stress reaching 3000 psi (confining pressure of 3435 psi), the differential pressure did not stabilize, indicating the existence of a threshold confining pressure. Under high confining pressure, the artificial fracture continued to close, thus increasing the pressure difference between the core inlet and outlet. Differential pressure was adopted as a secondary indicator to demonstrate the significant permeability changes. Quantifying the permeability became impractical after 150 h of injection unless the pressure-pulse decay method with chemically inert gas injection was applied [47,48], but this exceeds the capacity of the system. The pressure curves, representing upstream pore pressure, downstream pore pressure, and differential pressure, are presented to elucidate the process. Initially, during the imbibition of Solution #1 at 3000 psi effective stress, the gradual increase in differential pressure from 15 psi to 22 psi mirrored trends observed at lower effective stresses, stabilizing after approximately 50 h of injection, and indicating negligible changes in shale mechanical properties at this stage. Subsequently, the differential pressure continued to increase, rapidly escalating until reaching 188 psi, punctuated by two minor sudden drops before stabilizing around 190 psi upon cessation of injection. The limited increase in differential pressure during the initial 122 h of Solution #1 injection can be attributed to the penetration of the injected solution into the surrounding pores near shale surfaces, facilitated by the high threshold capillary pressure of shale due to its silt-like pores and low interfacial tension with high-salinity water. The threshold capillary pressure is the minimum entry pressure required for an invading phase to displace a defending phase from the center of an element, like a pore or throat [49]. It has been established that the geometry of shale pores is so complex that assuming pore channels as simple circular tubes inherently overestimates the capillary pressure [50]. This irregular pore geometry can be divided into two subregions: bulk regions and constricting regions, which contain corners and necks. When the capillary pressure increases, the invading fluids move towards these constricting regions [51]. This explains the phenomenon of the differential pressure initially increasing slowly, then spiking to a much higher level during Solution #1 imbibition. Higher effective stress enhances fluid inflow by reducing pore space and microfractures, subsequently elevating capillary pressure post-breakthrough, and increasing the growth rate of differential pressure.

The recorded sudden drops in pressure towards the end of the first injection cycle indicate the initiation of fracture generation and allow for evaluating injection-induced instability effects, including fluid pressure heterogeneity and the effect of stress imbalance on permeability reduction. This evaluation facilitates discussion on how the migration of pore pressure from the core center to the inlet side influences fracture generation. Previous studies by Roshan et al., Wang et al. and Zhang and Sheng employing confining pressure to investigate shale water uptake observed markedly unfavorable fracture generation under confining pressure compared to the case without it, with isolated fractures exhibiting uncommon extensions and tendencies to close rather than propagate [14,22,23]. Moreover, Roshan et al. noted no changes in upstream pressure when using 5 wt.% and 10 wt.% NaCl solutions compared to deionized water, with axial displacement measurements indicating sample shrinkage rather than swelling during injections, possibly attributed to a greater chemistry mismatch between the deionized water and the shale sample [14]. However, many previous studies using the core flooding technique maintained initial pore pressure at zero, while downstream pressure was kept at atmospheric pressure levels. This could explain the contradictory results observed in this study. Without injection fluid pressurization, HISW struggles to penetrate the core matrix due to its lower wettability and interfacial tension compared to low-ionic-strength water (LISW). Wettability-related capillary pressure is one of the key factors in water uptake by shale [11]. It was reported that up to 90% of hydraulic fluids disappeared in shale play as lowering ionic strength triggers increased hydrophilicity of the system [52,53]. Although LISW has a spontaneous imbibition advantage, fluid motion will stop once the increasing viscous force equals the capillary force unless other forces are applied [54]. On the other hand, reduced surface friction between HISW and shale allows for the creation of additional flow paths throughout the core under high injection pressure. This extends the imbibition time and intensifies shale swelling over time.

Stable effluent pH indicates that the observed changes in permeability are unlikely to result from significant chemical interactions between the shale and the fluid, such as mineral dissolution or other alterations that would normally manifest through pH fluctuations. This steadiness suggests that the reduction in permeability might be predominantly due to physical phenomena, such as injection fluid fingering. Other physical processes like fines migration and clay swelling could also be contributing factors. To better understand these dynamics, further analyses could be beneficial. For instance, trace element analysis of the effluents could reveal subtle changes that do not influence the pH.

3.2.2. Solution #2 and Solution #3 Sequential Imbibition Experiments

Though the ionic strength of all three solutions is the same, the total ionic concentration gradually decreases from Solution #1 to Solution #3 by increasing the concentration of CaCl₂ while reducing the concentration of NaCl. For the second and third injection cycles, the initial pore pressure was brought back to 435 psi and quickly increased with several sudden changes in between. The observed dramatic drop in the differential pressure from 436 psi to 250 psi corresponding to 438 h of injection indicates the generation of a larger hydration-induced fracture. Upon switching to Solution #3, this phenomenon was even more obvious as a series of significant sudden drops occurred non-stop until the end of the experiments. It is worth noting that each of these significant drops was always followed by minor drops, suggesting that the generated fractures were connecting to each other to form a network. However, it seems that the generated fractures had no benefits for the sample’s permeability as the differential pressure kept increasing. Similar results were obtained by Liu and Sheng when they employed axial and confining pressures while monitoring fracture generation through CT scanning to investigate the effect of water uptake [20]. They suggested that isolated fractures are not beneficial for shale permeability unless they connect with each other to form a fracture network throughout the core. They also observed that the fractures were generated more efficiently under stress anisotropy.

The injection of Solution #2 and Solution #3 displaced the pre-existing fluids. During the first injection phase (Solution #1 injection), the high concentration of Na⁺ ions in the solution could result in partial adsorption on the clay interlayer surface. These adsorbed Na+ ions could be desorbed by Ca²⁺ during Solution #2 and Solution #3 injections due to higher charge density. Substitution of calcite for sodium and potassium in the interlayer of illite might lead to clay swelling. However, verifying this exchange of calcium for sodium and potassium in the interlayer illite is challenging for our experiments due to the high concentrations of aqueous sodium in the solutions. Additionally, the stronger binding affinity of Ca²⁺ can result in a higher concentration of divalent cations near the surface, reducing the concentration of other monovalent ions. Al-Bazali et al. experimentally assessed the effect of ion concentration on shale osmotic pressure and found that the osmotic pressure induced by CaCl₂ solutions was higher than that obtained with NaCl solution [55]. The increased salinity gradient led to increased osmotic water flow, and the flux of Ca²⁺ carried more associated water into the shale, leading to increased shale swelling. It can be observed that compared to the first injection phase, the injections of Solution #2 and Solution #3 resulted in growing changes in differential pressure, possibly indicating a greater chemical potential difference between the displacing solutions and the shale.

As shown in Figure 6, the differential pressure curve moved closer to the upstream pressure curve while the downstream pressure was stepwise reduced to maintain the same initial average pore pressure. Simultaneously, the upstream pressure increased. This increasing deviatoric stress, dependent on the applied confining pressure and the differential pressure, weakened the mechanical strength of the shale and enhanced fracture generation. In our studied case, the compressive stress anisotropy as a byproduct of shale swelling cannot be ignored, especially as the upstream pressure was elevated to a high level. The confining pressure was adjusted in response to the increasing average pore pressure to demonstrate the effect of compressive strength anisotropy along the core due to fluid pressure heterogeneity. The fracture closure resulted in a gradual decrease in fluid pressure from the inlet tip to the outlet tip of the fracture. This decrease occurred as the fluid pressure decreased with increasing distance from the injection hole. Consequently, the effective stress near the injection hole more easily exceeded the shear failure along the long axis of the core. Ji and Geehan studied stress changes due to shale hydration and stated that shale hydration can create deviatoric stress that reaches its yield strength [56]. Similar conclusions have been drawn by Wang et al. who state that the local hydration stress concentration causes the fracture to propagate and generate secondary fractures [22].

The gradual decrease in pH during the second injection phase supports the argument that hydrogen ions are desorbed during this stage. Otherwise, H⁺ ions would replace them on the clay surfaces, resulting in a rise in pH observed at the outlet. This phenomenon becomes notable after the last displacement between Solution #3 and Solution #2. The pH initially dropped to a low of 3.32 and subsequently increased to around 7.15 for the remainder of the experiment, which is higher than the pH of Solution #3. This recovery in pH is likely caused by the continuous injection and ready imbibition of previous solutions. Combined with the SEM image of partially dissolved calcite in Figure 7a, it is possible that the decrease in pH facilitated calcite dissolution, which weakened the calcite cement bonding. Jordan et al. validated the hypothesis that progressively greater acidity enhances carbonate dissolution, and it is more important than ionic strength for carbonate solubility [57].

According to the pictures from the past experiment and CT scan of the sample (Figure 8c and Figure 9), there are coherent fractures throughout the sample that became new pathways. Therefore, in our case, the consistent permeability impairment was likely because of fracture closure. This was understood to be the mechanism behind the constant permeability enhancement-impairment, including shale swelling and stress anisotropy. Figure 9b and Figure 10 demonstrates the breakage located near the injection end of the fracture surface after the experiments, providing evidence that the failure region occurred around the injection hole. The fracture formed within the failure region had extended along the weak plane over time. Several studies have proposed wellbore stability analysis models with hydraulic pressure and osmotic pressure acting as driving forces. These studies found that failure regions developed around the wellbore circumference due to pore pressure propagation, which aligns with our findings [58,59,60,61]. Figure 8a,b illustrate the difference between the core outlet surface before and after the experiments. The high deviatoric stress had broken the cement bonds of the shale, which were formed by minerals deposited along thin laminae, creating a fracture network. The shear bond strength due to the bonding of cement to the shale interfaces is commonly used to evaluate the force required to delaminate this composite material at the interfaces [62]. The failure consistently occurred at the cement/shale interface due to the lower shear bond strength compared to the shale bulk strength [63]. This is confirmed by the study by Zhang and Sheng. In their study, the created fractures preferentially developed along the bedding laminations and tended to close under confining pressure at the end of the tests [23]. Tensile fractures developed across the artificial fracture and connected with other secondary fractures (microfractures/macrofractures) throughout the core, creating new pathways that improved permeability. This initiation and expansion of the cracks in turn promoted fluid-shale interactions. Therefore, the observed permeability enhancement was temporary as the fractures continued to close over time.

4. Conclusions

Formation damage in shale gas reservoirs is difficult to quantify in many cases. Often the problem is ignored due to a common rationale that the disappointing production outcome can be overcome by a simplistic ‘fracture-it-through’ method [64]. This strategy may be effective in some scenarios; however, considering the low recovery rates of injected water post-fracturing, it raises both technical and environmental concerns [65,66,67]. Although previous studies have suggested that increasing the salinity of injection fluid may prevent shale disintegration [19,68], it has been observed that shale can become saturated even when exposed to HISW solutions. For example, an earlier study demonstrated that increased concentrations of NaCl, KCl, MgCl₂, and CaCl₂ can enhance water uptake in partially saturated shale [69]. Furthermore, the degree of fluid adsorption into the shale is influenced by various factors, including loading stress, duration of fluid-shale contact, temperature, initial saturation, and pH [2,14,70,71].

This study investigated the impact of sequential HISW imbibition on shale permeability. Our findings reveal that shale permeability reduction occurs in the initial phases of Solution #1 injection. Upon extending the imbibition period under 3000 psi effective stress, a significant increase in differential pressure was observed, followed by abrupt declines. These trends were more pronounced with the introduction of Solution #2 and Solution #3. The displacement between these injected solutions enhanced the weakening effects on the shale’s mechanical properties. Key observations from this study include: (1) periodic fluctuations in permeability were intensified by the displacement of the HISW solutions; (2) a substantial decrease in pH was noted when Solution #2 was displaced with Solution #3; and (3) coherent fractures within the core developed during the experiments, possibly due to interactions between HISW and shale. These findings provide initial insights into the effects of HISW on shale permeability, considering various fluid chemistries and reservoir conditions. Our robust and innovative experimental approach offers valuable guidance for future research in this field. Moreover, the findings challenge the belief that HISW can impede or even prevent shale disintegration. The high salinity of the injection fluids created a significant contrast with the rock, making it challenging to conduct ICP-MS trace element analysis to determine the chemical reactions during the experiments. Consequently, we were unable to draw definitive conclusions about the ions contributing to shale disintegration, and further studies are required to confirm that.

Based on these insights, we recommend optimizing fluid compositions and managing flowback water to mitigate adverse effects in field applications. Conducting comprehensive pre-operative tests over an extended period of time under reservoir conditions can prove to be highly advantageous. This study strongly recommends further investigation into the imbibition process using consistent CT imaging techniques and analyses of the composition of effluents. Moreover, we suggest that other major processes, including adsorption into organic matter and dissolution of carbonates, must be considered when evaluating fluid-shale interactions.