The study area of this paper is the Qianjiadian uranium deposit, Inner Mongolia, China (Figures 1A, B). The Qianjiadian uranium deposit is distributed in the southeast uplift belt of the Songliao Basin (Wang Q. et al., 2022), where the main exposed strata include the Qingshankou Formation (K₂qn), Yaojia Formation (K₂y), and Nenjiang Formation (K₂n) of the Upper Cretaceous and the Quaternary (Q) (Figure 1C). The Yaojia Formation is the main ore-bearing formation, and the Qingshankou Formation is the secondary ore-bearing formation (Figure 1D). The thickness of the Yaojia Formation ranges from 181.20 to 214.20 m, with an average of 197.88 m. The bottom depth of the ore formation ranges from 311.30 to 455.00 m, with an average of 420.18 m. The red and gray fine sandstones and a small amount of medium and coarse sandstones deposited in the fluvial phase are composed of more mudstone and siltstone interlayers with loose structures. The thickness of the mudstone interlayer is highly variable, with poor continuity in strike and tendency, and mostly lenticular. The thin mudstone is mostly gray and grayish, and the thicker mudstone is mostly purplish-red. Sediment grain size is mainly medium-to-fine sand, with coarse sand and a small amount of gravel seen at the bottom and more silt and mud at the top. Sandstone fragments of the Yaojia Formation are relatively rich in silica and aluminum elements, and the rock types are mainly lithic sandstone and feldspathic sandstone. The deposit has four aquifers, of which the lithology of the mineral-bearing aquifer in the upper section of the Yaojia Formation is mainly gray fine sandstone, with a small amount of thin-layered gray and purplish-red mudstone and gray muddy siltstone which are characterized by weak water-richness and permeability. The lithology of the mineral-bearing aquifer at the lower end is mainly light gray and light red fine-medium sandstone, with a small amount of fine siltstone.

The rock samples were collected from the geological-prospecting boreholes drilled by the Beijing Institute of Chemical Metallurgy of Nuclear Industry. The selected samples originated from the middle ore-bearing horizon (Figure 1D). According to the differences in particle size, cementation degree, and mineral composition, the samples were classified as coarse sandstone, medium sandstone, fine sandstone, siltstone, and sandy mudstone (named CS, ZS, XS, FS, and SN, respectively) (Figure 2). The bulk samples were first prepared and dried in a constant temperature and humidity test box for 24 h at 60°C. Then, about 30 g of the processed samples was picked out and placed in the gas–liquid–solid high-temperature/pressure reactor for the CO₂ + O₂–water–rock geochemical reaction experiment. Finally, the deionized water and a certain proportion of CO₂ and O₂ were injected into the glass container in the reactor to conduct the CO₂ + O₂ in situ leaching simulation experiment.

The CO₂ + O₂–water–rock geochemical reaction experiment was conducted in a gas–liquid–solid high-temperature/pressure reactor (Figure 3). This device mainly consists of three subsystems: the gas storage subsystem, the gas pressurization subsystem, and the reaction retort. The gas pressurization subsystem elevates the pressure of CO₂/O₂ from the gas storage subsystem to the design value. The pressurized gas CO₂/O₂ is injected into the reaction retort for participating in the geochemical reaction. The thermocouple, temperature sensor, and pressure sensor are installed on the reaction retort to maintain the required temperature/pressure conditions. The control accuracies of the temperature and pressure are 0.01°C and 0.01 MPa, respectively. The maximum temperature and pressure values of this device that can be reached are 150°C and 60 MPa, respectively.

Each sample was immersed in 30 ml deionized water. During the CO₂ + O₂–water–rock geochemical reaction process, the pressure of CO₂ and O₂ was set as 10 MPa and 6 MPa, respectively, according to the solution proportioning parameter field test (Yang et al., 2020). The ambient temperature of the geochemical reaction is 40°C, and the reaction time is 20 days.

The mineral component identification of the samples was conducted using an X-ray diffractometer (Bruker D8 Advance). Approximately 50-g samples were ground using a fully automatic crusher, and 10 g of each was screened using a sieve to obtain powder samples of less than 300 mesh. Then, the prepared powder sample was placed on the test instrument, and a copper X-ray tube operating at 40 kV and 30 mA was used with counts collected from 3° to 90° in steps of 0.02° and at a speed of 3°/min to obtain the scattering curve of the samples (Li et al., 2022a; Li et al., 2022b). After the test, the mineral compositions were obtained by profile fitting of the scattering curve.

The mineral constituent of samples during the CO₂ + O₂–water–rock geochemical reaction process was measured using the Empyrean X-ray diffractometer (Panako, Netherlands). Original rock blocks collected on site were first ground and sieved to obtain the powder samples with a size of <300 mesh. Then, they were placed on the test instrument, and the copper X-ray tube operating at 40 kV and 30 mA was used with counts collected from 3° to 90° in steps of 0.02° and at a speed of 3°/min to obtain the scattering curve of the samples. Finally, the mineral composition characteristics of samples were acquired by calculating the peak area of the X-ray spectrum.

The micropore structural characteristics of the samples were analyzed by the HPMIP method, which can measure pores of sizes between 3 nm and 950 μm. For HPMIP, the pores are assumed to be cylindrical, and the relationship between the mercury pressure and the pore radius can be determined by the Washburn equation (Kumar and Bhattacharjee, 2003): r = − 2 γ COS θ p c , (1) where p is the mercury pressure, γ is the surface tension of mercury, θ is the wetting angle between mercury and the solid medium, and r is the pore radius.

The internal microstructure of uranium-bearing sandstone is extremely heterogeneous and has been confirmed to have fractal characteristics (Zeng et al., 2019; Zeng et al., 2022). The fractal dimension is usually used to characterize the irregularity of complex structures and can be calculated from the date of HPMIP, which is defined as N r = V H g π r 2 l ∝ r − D f , (2) where N(r) is the number of units with the radius of r required to fill the entire fractal object; D is the fractal dimension.

The relationship between the diameter and length of capillaries follows a similar fractal scaling law (Yu et al., 2009): l = 2 r 1 − D T l 0 D T , (3) where l is the tortuous length, r is the diameter of capillary tubes, l ₀ is the representative length of a straight capillary tube, and D _T is the tortuosity dimension.

Combining Eqs 1–3 leads to V H g ∝ p − 3 − D f − D T . (4)

The mercury saturation S _Hg is defined as S H g = V H g V p , (5) where V _Hg is the mercury inlet volume and V _p is the total pore volume.

By combining Eqs 4, 5 and taking the logarithm, the following equation can be obtained (Su et al., 2018): lg S H g ∝ D f + D T − 3 lg p c . (6)

The fractal dimension can be obtained by the gradient of S _Hg and p _c by using the double-logarithmic coordinate.

The test results of XRD show that the quartz content, the albite content, and the micro-plagioclase content contribute 13.21%–38.23%, 19.66 %–28.40%, and 27.67 %–35.26%, respectively, of the total minerals of samples without a CO₂ + O₂–water–rock geochemical reaction, indicating that the felsic mineral dominates the substance composition in the Qianjiadian uranium deposit. Moreover, a small number of clay minerals are also developed in these samples, including chlorite, illite, and kaolinite. The chlorite content and the illite content of sample SN can reach 17.53% and 21.92%, displaying the typical mineral distribution characteristic of sandy mudstone.

Equations 7–13 show the ion changes after CO₂ + O₂ is injected in water and the geochemical reaction process of typical minerals. The formed solution of CO₂ + O₂ and water creates an acidic-oxidation environment, causing the occurrence of chemical reactions such as dissolution and precipitation. C O 2 + H 2 O ⇌ H 2 C O 3 ⇌ H + + H C O 3 − ⇌ C O 3 2 − + 2 H + , (7) Q u a r t z : S i O 2 + 2 H 2 O ↔ H 4 S i O 4 , (8) M i c r o − p l a g i o c l a s e : 2 K A l S i 3 O 8 + 2 + + 9 H 2 O ↔ 2 K + + A l 2 S i 2 O 5 O H 4 + 4 H 4 S i O 4 , (9) A l b i t e : 2 N a A l S i 3 O 8 + 2 H + + 9 H 2 O ↔ 2 N a + + A l 2 S i 2 O 5 O H 4 + 4 H 4 S i O 4 , (10) C a l c i t e : C a C O 3 + H 2 O + C O 2 ↔ C a 2 + + 2 H C O 3 − , (11) C h l o r i t e : M g , F e 5 A l 2 S i 3 O 10 O H 8 + 5 C O 2 + 5 C a C O 3 ↔ 5 C a F e / M g C O 3 2 + A l 2 S i 2 O 5 O H 4 + H 4 S i O 4 , (12) I l l i t e : K A l 2 S i 3 A l O 10 O H 2 + 10 H + ↔ K + + 3 A l 3 + + 3 H 4 S i O 4 . (13)

The changes in the mineral content of samples before and after the CO₂ + O₂–water–rock geochemical reaction are shown in Figure 4. The mineral composition of the original sample is complex, and the types of geochemical reactions are various; thus, the mineral content demonstrates significantly inconsistent change trends. The samples CS, ZS, XS, FS, and SN represent coarse sandstone, medium sandstone, fine sandstone, siltstone, and sandy mudstone, respectively. After the CO₂ + O₂–water–rock geochemical reaction, the contents of quartz, albite, and micro-plagioclase in the coarse sandstone decrease by 9.05%, 12.68%, and 2.99%, respectively, while the contents of chlorite and illite in the coarse sandstone increase by 69.68% and 97.29%, respectively; the contents of albite, micro-plagioclase, chlorite, and illite in the medium sandstone increase by 4.29%, 8.27%, 77.45%, and 100%, respectively, and only the quartz content in the medium sandstone decreases by 31.44%; the contents of quartz and micro-plagioclase in the fine sandstone decrease by 54.4% and 51.13%, respectively, while the contents of albite, chlorite, and illite in the fine sandstone increase by 2.41%, 37.88%, and 92.07%, respectively; the contents of quartz and albite in the siltstone decrease by 54.43% and 51.13%, while the contents of micro-plagioclase, chlorite, and illite in the siltstone increase by 29.67%, 66.50%, and 68.85%, respectively; the contents of quartz, albite, and micro-plagioclase in the sandy mudstone increase by 81.23%, 49.34%, and 7.99%, while the chlorite and illite in the sandy mudstone decrease by 68.63% and 48.36%, respectively. To sum up, the felsic mineral and clay mineral exhibit an opposed change trend after the CO₂ + O₂–water–rock geochemical reaction. Specifically, the felsic mineral of coarse sandstone, medium sandstone, fine sandstone, and siltstone reduces by 1.95 %–103.15%; the clay mineral of coarse sandstone, medium sandstone, fine sandstone, and siltstone increases by 61.54–88.72%; and the felsic mineral and clay mineral of sandy mudstone increase and decrease by 138.55% and 57.36%, respectively.

According to Eqs 7–13, all the quartz, albite, micro-plagioclase, chlorite, and illite can react with the CO₂ + O₂ aqueous solution. For coarse sandstone, medium sandstone, fine sandstone, and siltstone, the dissolution of felsic mineral dominates the geochemical reaction, while the dissolution reaction of clay mineral in sandy mudstone occupies the main position during the CO₂ + O₂–water–rock geochemical reaction. Thus, the differential geochemical reaction is dramatically influenced by the initial mineral content in the ore bed, the reactions involving clay minerals may be unfavorable for CO₂ + O₂ in situ leaching, and the ore bed with high argillaceous content always acts as the aquiclude in the solution leaching process (Zhang T. et al., 2021).

The relationship between lgS _Hg vs. lgp _c of samples in this work are shown in Figure 8, which are fit by Eq. 6 with high correlation coefficients (R ² > 0.90), indicating that the pores of the samples have fractal properties and the fractal theory can be applied, as described in this work. For the coarse sandstone, medium sandstone, fine sandstone, and siltstone, the curves of lgS _Hg and lgp _c can be divided into two segments, i.e., the low-pressure stage and the high-pressure stage, which, respectively, represent the macropore and the mesopore, transition pore, and micropore. For the sandy mudstone, the curves of lgS _Hg and lgp _c can be divided into four segments, representing the four pore types.

The fractal dimension calculation results are shown in Table 1. The fractal dimension D₁ ranges from 3 to 6 because it contains fractal dimensions for pore space and tortuosity, and the fractal dimension D₂ ranges from 2 to 3. It can be seen that D₁ is greater than D₂, demonstrating that macropores are more complicated than mesopores, transition pores, and micropores (Jiang et al., 2016; Peng et al., 2017; Su et al., 2018). From the fractal dimension curve of samples SN-bef and SN-aft, it is found that there are notable multiple inflection points, and because the pore structure of samples SN-bef and SN-aft are more complex, four segments are divided and compared with samples CS, ZS, XS, and FS. After the CO₂ + O₂–water–rock geochemical reaction, the fractal dimensions of the macropore in coarse sandstone, medium sandstone, fine sandstone, siltstone, and sandy mudstone are in the range of 2.819–3.311, which are lower than those (3.385–5.544) before the CO₂ + O₂–water–rock geochemical reaction; the fractal dimensions of mesopore, transition pore, and micropore in coarse sandstone, medium sandstone, fine sandstone, siltstone, and sandy mudstone after the CO₂ + O₂–water–rock geochemical reaction are in the range of 2.063–2.991, which are larger than those (2.023–2.468) before the CO₂ + O₂–water–rock geochemical reaction; this reflects that the CO₂ + O₂–water–rock geochemical reaction makes large pores more smooth and uniform, promoting the improvement of roughness and heterogeneity of small pores.

As discussed previously, the geochemical reaction types are complex and their reconstruction effects on the pore structure are multifarious. The sample surface of representative coarse sandstone is relatively dense (Figure 9A), which then becomes rougher and bumpier (Figure 9B) by the destruction and reorganization of the pore structure. The intergranular pores are mainly developed in the original sample, while after the CO₂ + O₂–water–rock geochemical reaction, one part of the intergranular pores is dissolved and enlarged, and another part of adjacent intergranular pores is connected and formed to fracture. Moreover, some precipitates generated during the CO₂ + O₂–water–rock geochemical reaction process can block the pore. These reconstruction effects will occur simultaneously and influence the pore characteristics.

To clarify this reconstruction behavior, a schematic diagram describing the different reconstruction modes during the CO₂ + O₂–water–rock geochemical reaction process is shown in Figure 10. Considering the mineral composition and its fabric mode of uranium-bearing reservoirs, the microstructure of samples is shown in Figure 10A; the pores with different sizes are developed between the mineral detritus, including the connected macropores, the narrow mesopore, transition pore, and micropore with pore throats. The geochemical reactions between CO₂ + O₂, water, and rock expressed by Eqs 7–13 appear (Mitiku et al., 2013; Cui et al., 2021; Li S. et al., 2022), showing that 1) the quartz is dissolved weakly and its surface becomes rough and uneven, 2) the feldspar minerals (micro-plagioclase and albite) are resolved into the clay minerals (kaolinite, chlorite, and illite), 3) the carbonate minerals are completely dissolved into ions and preserved in the solution, and 4) the secondary mineral precipitates are generated during the geochemical reactions (Figure 10B). Additionally, the physical effects also appear in this process, such as clay mineral swelling by water absorption and the secondary mineral particle migration driven by the leaching solution.

The aforementioned reactions lead to the differential transformation of the pore structure of uranium-bearing reservoirs, which can be classified as the positive transformation and negative transformation for CO₂ + O₂ in situ leaching. The positive transformation includes the dissolution of carbonate and feldspar minerals, which enlarges the pore space and promotes the connectivity between pores of different sizes and types; it shows a favorable aspect for the leaching of hosted uranium and the migration of the leaching solution. The negative transformation includes the secondary mineral precipitates and clay mineral swelling by water absorption. The formed mineral precipitates and clay mineral swelling can occupy the original space and reduce the pore volume, and simultaneously, these small mineral particles can flow with the leaching solution and block the pore throat, decreasing the connectivity of pores. The macropores always have a larger pore throat and are unaffected by the blocking effect; thus, the positive transformation is stronger than the negative transformation in the macropores. However, the pore throats with a small size (mesopore, transition pore, and micropore) are narrow and susceptible to the blocking effect, and then the positive transformation is weaker than the negative transformation in the mesopore, transition pore, and micropore, which exhibits the disadvantageous aspect for CO ₂+ O₂ in situ leaching. The two transformations are simultaneous and mutually restrictive, depending on the mineral composition, pore size distribution, and pore throat distribution (Li S. et al., 2020). For uranium-bearing reservoirs in the study area, the CO₂ + O₂–water–rock geochemical reaction promotes the expansion of the small pore percentage and reduction of the large pore percentage. This effect may improve the reservoir permeability to a certain degree, while the reservoir siltation effect will gradually reveal with the extension of CO₂ + O₂ in situ leaching (Zhao et al., 2018; Zhao K. et al., 2022); the reduction of the total pore volume of samples during the CO₂ + O₂–water–rock geochemical reaction process has confirmed this phenomenon. Moreover, in the field tests, the clay-blocking seepage had occurred in a part of uranium deposits (Xu, 2014; Wang et al., 2022), and there were no particularly effective solutions because of the weak acid–alkali solubility of clay minerals.

Therefore, only adopting the chemical enhancement permeability can hardly solve the blocking problem of uranium in situ leaching. The CO₂ phase-change blasting method realizes reservoir stimulation using the rapid expansion of CO₂ from the liquid state to gas state, whose technical principles are to 1) promote the fracture network generation induced by blasting shock waves and 2) extend and increase connectivity of the pore-fracture system by the gas wedging effect. This will provide the source power and smooth passage for the discharge of siltation sediment, and CO₂ phase-change blasting and CO₂ + O₂ in situ leaching possess natural technology intersection. The residual CO₂ in uranium reservoirs can also be used as one source of solvent for CO₂ + O₂ in situ leaching. Therefore, combining the CO₂ phase-change blasting and CO₂ + O₂ in situ leaching method may be a potential physical-chemical reservoir stimulation direction for uranium-bearing reservoirs, and the mechanism and method for the dynamic regulation of permeability enhancement and CO₂ + O₂ in situ leaching are the focus of future research.