The sandstone samples for this study were sourced from Erdos City, Inner Mongolia Province, China, as illustrated in Figure 1. To determine the precise composition and microstructural features of the samples, X-ray diffraction (XRD) and petrographic thin section analyses were performed, as shown in Figure 2. Petrographic thin-section observations indicated that the sandstone primarily consists of quartz, feldspar, calcite, and rock fragments (Table 1). Quartz shows no secondary enlargement, and the grains exhibit clear contacts with good sorting. The rock samples, cut from drill cores at the same depth, were prepared into cylindrical samples with dimensions of Φ25 × 50 mm. These samples were placed in a drying oven for 24 h prior to measuring their density and porosity. Samples with similar densities and porosities were selected for the cyclic tests, and their densities and porosities are detailed in Table 2.

To investigate the effects of water evaporation and salt precipitation on reservoir rocks during intermittent CO₂ injection and sequestration in underground saline aquifers, a laboratory simulation experiment was designed to study the changes in pore structure of reservoir rocks after intermittent CO₂ injection. This study aims to evaluate the long-term safety and effectiveness of sequestration technologies. A reaction setup was constructed (see Figure 3) to simulate geothermal conditions and the characteristics of underground saline aquifers. The samples were categorized into three groups based on saline water type: Na₂SO₄, NaCl, and a control group with deionized water. The saline solutions were prepared using deionized water at a concentration of 30 g/L (30,000 ppm), with specific parameters detailed in Table 3. Research indicates that deep saline aquifers at depths of 800–2,000 m with salinity levels of 20,000–250,000 ppm are optimal for CO₂ geological sequestration (Fang et al., 2016; Boampong et al., 2023). The experimental procedure is outlined in Figure 4. Initially, the solutions and dried rock samples underwent oversaturation treatment (vacuum pressure saturation) to simulate real formation conditions. The saturated samples were then placed in the reaction setup at 40 °C and 5 MPa. After the designated reaction times, the samples were retrieved for nuclear magnetic resonance (NMR) scanning, and the solution was analyzed using inductively coupled plasma-optical emission spectrometry (ICP-OES) to measure changes in pore space and ion concentration in the reacted rock samples. Finally, wave velocity, dry mass, metal ion concentration in the solution, and total dissolved solids (TDS) of the samples were recorded. This process constituted one cycle, and a total of five cycles were performed with reaction durations of 0 h, 12 h, 24 h, 72 h, 120 h, and 168 h, over a span of 16.5 days.

The dynamic variations in the dry mass of rock samples before and after the cyclic treatment process are key indicators for evaluating the impact of the injection cycles. Figure 5 illustrates the changes in the dry mass of the samples over time. The horizontal axis represents time, while the vertical axis displays the ratio of the mass before and after the reaction, where M₀ represents the mass of the original untreated rock sample, and M denotes the mass after the treatment. Figure 5 reveals that changes in sample mass can be divided into two distinct phases: Phase I and Phase II. During Phase I, rock samples treated with different types of saltwater show varying trends in dry mass after CO₂ injection. Specifically, samples treated with Na₂SO₄-type saline water exhibit a notable increase in mass, particularly during the second cycle, with a 0.18% increase compared to the initial state. In contrast, samples treated with NaCl-type saltwater show slight fluctuations but remain relatively stable, while samples treated with deionized water experience mass loss after CO₂ injection, decreasing by 0.12% compared to the initial state. In Phase II, with increasing cycles, all three types of saltwater-treated rock samples generally display a decrease in mass following CO₂ injection. Notably, NaCl-type and deionized water-treated samples show more pronounced reductions, with decreases of 0.78% and 0.67%, respectively, compared to their initial states. Although Na₂SO₄-type samples also experience mass loss, the reduction is the smallest, at only 0.21%.

To systematically assess the impact of cyclic CO₂ injection on pore space evolution in sandstone samples, the P/P₀ index was introduced as a metric. Here, P denotes the pore space size of the rock samples in the current treatment state, while P₀ represents the pore space of the original untreated samples. Based on pore size variations, pores were classified into three categories: macropores, mesopores, and micropores, as indicated on the horizontal axis in Figure 6.

In Figure 6, the color shading visually represents the number of cycles the samples have undergone, with darker colors indicating a higher number of cycles, up to the final count of five cycles. The pore space of the sandstone samples showed significant changes with the increase in the number of cycles. The volume of small pores generally decreased, while the volume of mesopores increased significantly and became the predominant feature of pore space changes. Although the volume of macropores also increased during cycling, it did not return to the initial level. When comparing the effects of three different brines on pore structure, the Na₂SO₄-type brine demonstrated the most notable performance. Specifically, the volume of mesopores increased most significantly during the second cycle, reaching three times the initial volume. In contrast, the NaCl-type brine had a lesser effect on mesopores, with a maximum increase of 2.5 times the initial volume, but by the fifth cycle, the mesopore volume was twice the initial volume, exceeding that of the Na₂SO₄-type brine by a factor of 1.4. The pure water type had a less pronounced effect on pore structure under cyclic CO₂ injection. The micropore volume decreased substantially, and the mesopore volume fluctuated, with a maximum increase of only 1.7 times the initial volume, lower than that observed with the other two brine types. The volume of macropores remained relatively stable throughout the cycling process, with the P/P₀ ratio fluctuating around 1, indicating minimal change. Overall, significant differences were observed in the effects of various brines on the pore structure of sandstone samples during cyclic CO₂ injection.

To examine the alterations in the homogeneity and continuity of rock samples during cyclic CO₂ injection, the wave velocity index was introduced. The evolution trend of the longitudinal wave velocity of the samples with an increasing number of cycles was visualized, as illustrated in Figure 7.

Figure 7 illustrates the relationship between time (horizontal coordinate) and longitudinal wave speed (vertical coordinate). To accurately compare differences among various brine types, the coordinate ranges were standardized across all samples. In Figure 7, the brine types, such as pure water, NaCl, and Na₂SO₄, are arranged from top to bottom. The results indicate that the variation in longitudinal wave velocity is directly influenced by the cyclic injection process, with distinct wave velocity evolution trends observed for different brine types. Specifically, for sandstone samples with pure water, the longitudinal wave velocity initially increased to 1.12 times the initial value at the start of the cycle and subsequently decreased to 0.85 times the initial value, showing a relatively straightforward trend. In contrast, for NaCl brine, a significant increase in wave velocity was observed initially, reaching 1.11 times the initial value, followed by a decrease to 0.92 times the initial value before recovering. For Na₂SO₄ brine, there was a notable increase in longitudinal wave velocity during the first three cycles, reaching 1.10 times the initial value, followed by a decrease to 0.89 times the initial value, and a subsequent rise to 1.02 times the initial value in the final phase. This pattern reflects the impact of cyclic injection on wave velocity.

In Figure 8, curves a, b, and c represent the changes in the concentrations of Ca²⁺, Mg²⁺, and K+ ions in solution over the cycling treatment time, respectively, while curve d depicts the TDS values of the solution properties. These curves illustrate the variability in solution properties during the cycling process. Initially, the concentrations of all metal ions were below 10 ppm. As the number of cycles increased, the concentration of Ca²⁺ ions rose significantly to 210.9 ppm, which is 47.1 times the initial concentration, indicating a substantial effect of cycling treatment on Ca²⁺ ions. In contrast, the concentration of Mg²⁺ ions showed an initial rapid increase followed by a decrease, eventually stabilizing at a relatively low level. The concentration of K+ ions exhibited less variation compared to Ca²⁺, though it still showed some fluctuations relative to Mg²⁺, highlighting its distinct trend.

The ion concentration changes in various solution systems were further analyzed. In the pure water system, the concentration of Ca²⁺ ions showed a tendency to stabilize with an increasing number of cycles. Although minor fluctuations occurred, the overall change remained limited. Conversely, the concentration of Mg²⁺ ions peaked quickly during the initial cycles and then gradually decreased, while the K+ ion concentration exhibited a slight increase initially, followed by stabilization and a small overall rise. For NaCl-type brine, the Ca²⁺ concentration peaked after four cycles, then began to decrease and eventually stabilized around 100 ppm. The Mg²⁺ ion concentration generally increased before decreasing. K⁺ ions, after an initial increase, maintained stability for a period and later exhibited a more noticeable increase in concentration. In Na₂SO₄-type brines, the concentration of Ca²⁺ ions displayed more complex behavior, initially rising, then stabilizing, and finally increasing sharply. The concentration of Mg²⁺ ions increased initially, decreased, and then stabilized at double the initial concentration. The K⁺ ion concentration varied more significantly, first increasing, then decreasing, and finally experiencing a rapid increase, reflecting a distinct fluctuation pattern.

In terms of total dissolved solids (TDS), the TDS values of the NaCl- and Na₂SO₄-type solutions were notably higher compared to the pure water-type solutions. Both NaCl and Na₂SO₄ solutions exhibited similar trends during the recirculation process, characterized by an initial rapid decline followed by a gradual decrease. In contrast, the TDS values for pure water-type solutions increased with the number of cycles and eventually stabilized, with the final values remaining higher than the initial values.

In the geological storage of CO₂, the saline aquifer serves as the crucial medium, with its primary components (sulfate, chloride, and carbonate) significantly affecting the storage efficiency (Chen et al., 2023; Wen et al., 2024). Following CO₂ injection, it rapidly dissolves in the brine, leading to quick saturation at the interface and decreasing concentration further away. This dissolution process generates carbonic acid, which substantially lowers the groundwater pH, a phenomenon known as ‘acidification’ (Equations 1–3). The acidification accelerates the dissolution of carbonate and some silicate minerals, releases cations such as Ca²⁺, Mg²⁺, and Si⁺, and increases the concentration of HCO3- ions, thereby further enhancing the acidification process (Md Yusof et al., 2022). Additionally, the dissolved minerals and CO₂ facilitate the formation of new mineral phases (e.g., carbonate and silicate minerals), which help in CO₂ fixation but also alter the pore space. Initially, the dissolution of carbonate minerals raises the concentration of Ca²⁺ and Mg²⁺ ions, which is later reduced by reprecipitation, while the concentration of K⁺ ions remains relatively stable. Prolonged CO₂ injection lowers the solution pH, increasing the environmental acidity. C O 2 + H 2 O → H 2 C O 3 (1) H 2 C O 3 → H + + HC O 3 − (2) HC O 3 − → H + + C O 3 2 − (3)

In sandstone minerals, the dissolution of feldspars is predominant. Feldspars undergo hydrolysis in the presence of water, releasing alkaline metal ions such as Na⁺ and K+. This process typically occurs in acidic conditions, which promote the dissolution and leaching of these metal ions from feldspars (Equations 4–6) (Luquot et al., 2016; Jayasekara et al., 2020). Clay minerals may also dissolve to some extent, forming silicate and aluminum complexes (Kweon and Deo, 2017). Although H₂SiO₄ is usually unstable, it tends to form colloidal or polymeric silicic acid in aqueous environments. This process increases the cation concentration in the solution and also leads to changes in the pore space due to the dissolution and precipitation of minerals. 2 KAlS i 3 O 8 + 9 H 2 O + 2 H + → 2 K + + 2 A l 2 S i 2 O 5 OH 4 + 4 H 2 Si O 4 (4) 2 NaAlS i 3 O 8 + H 2 O + C O 2 → 2 N a + + 2 A l 2 S i 2 O 5 OH 4 + 4 Si O 4 (5) CaA l 2 S i 2 O 8 s + 2 H + + H 2 O l → C a 2 + aq + A l 2 S i 2 O 5 OH 4 s (6)

The solubility of CO₂ is significantly affected by various ionic species and their concentrations, which are primarily determined by the number of electron layers around different ions and their hydration radii (Hassanzadeh et al., 2009; Hu et al., 2013; Ali et al., 2022). As the concentration of ions such as Na⁺ and K⁺ increases in the solution, CO₂ solubility decreases (Chen et al., 2023). Ionized CO₃ ^2– and HCO3– are more likely to react with Ca²⁺ and Mg²⁺ ions (Equations 7, 8). SO₄ ^2– reacts with Ca²⁺ ions to form a CaSO₄ precipitate, a typical precipitation reaction in sulfate-type brines (Equation 9) (Wang et al., 2016). Due to the low solubility of CaSO₄, this reaction typically results in the formation of solid CaSO₄. On the other hand, the solubility of CaCO₃ is influenced by carbonate ion concentrations, leading to a supersaturated solution without immediate precipitation. C a 2 + aq + HC O 3 − aq → CaC O 3 s + H + (7) M g 2 + aq + C O 3 2 − aq → MgC O 3 s (8) C a 2 + aq + S O 4 2 − aq → CaS O 4 s (9)

Throughout the reaction, pore complexity increases and eventually stabilizes as minerals precipitate and dissolve. This alteration in pore structure enhances CO₂ flow ability, increases the contact area for chemical reactions, and further drives dissolution and precipitation processes. During the test, pore space complexity did not increase uniformly but fluctuated, confirming the dynamic competitive relationship between dissolution and precipitation in CO₂ geological sequestration.

To characterize the distribution pattern of pore radius in samples subjected to cyclic injection, Figure 9 illustrates the pore radius on the horizontal axis and the pore volume ratio on the vertical axis. A vertical comparison of the plot indicates that the pore volume ratio of the bottom layer represents the initial state of the sample, with additional layers reflecting an increase in the number of cycles, up to five cycles. Based on established criteria (IUPAC), pore sizes have been categorized into three groups: macropores, mesopores (medium pores), and micropores (Thommes et al., 2015; Medina et al., 2023). Figure 9 demonstrates a clear trend in the change of pore volume share, with the primary peak (representing the most common pore size in the sample) shifting along the pore radius axis as the number of cycles increases, indicating the remodeling of the pore structure during the injection process. The shift of the primary peak toward smaller pore sizes generally signifies significant adjustment in the pore system after multiple cycles. Besides the shift in the primary peak, Figure 9 also reveals increased peak differentiation, where the boundaries between micro-, meso-, and macropores become more distinct, and the area differences between peaks progressively widen. Among the brine types tested, the Na₂SO₄ type had the most pronounced impact on pore volume ratio changes, as evidenced by the maximum shift in the primary peak. The NaCl type also significantly affected the pore structure, though to a lesser extent, while the changes observed with pure water were minimal. Following CO₂ injection, mesopore apertures became the predominant region in the pore distribution of the samples.

After ionization of Na₂SO₄ in water, SO₄ ^2– ions chemically react with mineral components of sandstone (particularly calcite, CaCO₃) to form insoluble sulfate precipitates (e.g., CaSO₄). These precipitates occlude some pores, resulting in reduced pore volume and altered pore size distribution. This high chemical reactivity is a primary factor contributing to the significant impact of Na₂SO₄-type brines on pore space. In contrast, NaCl ionizes in water to produce Na⁺ and Cl⁻ions, which are relatively chemically inert and engage in fewer direct reactions with sandstone minerals. Nonetheless, NaCl solutions can indirectly influence pore structure through physicochemical mechanisms such as osmotic pressure and dissolution-precipitation equilibrium. This effect, while present, is less pronounced compared to Na₂SO₄-type brines. Pure water primarily serves as a carrier for dissolved CO₂ during injection; however, CO₂ solubility is higher in pure water than in brine (Islam et al., 2016; Chen et al., 2023), which activates H+ ions in pure water control, leading to increased acidity. Additionally, the acidic environment promotes the dissolution of soluble minerals, causing changes in pore structure, as depicted in Figure 9C, where the main peak of the pore volume shifts to the left.

Figure 10 illustrates a schematic of the intricate mechanism involving salt precipitation and the shedding and dissolution of mineral crystals in the pore space, driven by differences in chemical properties under cyclic injection conditions. This process has a profound impact on pore structure evolution, making it both complex and dynamic. In the fluid transport system of saline aquifers, macropores serve as the primary pathways for fluid infiltration due to their excellent connectivity. During the transport process, detached mineral and precipitated salt crystals tend to accumulate in macropores, effectively reducing their volume and shifting the proportion of microporous space toward the mesopore range in the samples (Han et al., 2020). Furthermore, mineral crystals in small pores undergo dissolution and detachment, which not only enlarges the small pores but also facilitates their transition to the mesopore range, thus increasing pore distribution diversity (Wang et al., 2021). This explains the significant development of mesoporous regions as the number of cycles increases, resulting in a shift in the dominant pore structure feature. Significant changes in pore structure were observed within just 12 h of initial injection, with mesopores expanding more rapidly than other pore sizes, highlighting the efficiency and rapidity of the process. Overall, the interplay between dissolution and precipitation affects both the morphology and distribution of pore space, leading to notable fluctuations in the reservoir rock’s pore structure during CO₂ injection. This understanding is essential for evaluating geological storage mechanisms, assessing storage capacity, and predicting long-term storage stability.