Water–rock interactions (WRIs) are a topic of interest in geology and geotechnical engineering. Many physical and chemical reactions are involved in the WRI, including lubrication, precipitation, oxidation–reduction, and ion exchange. Many geological hazards and engineering failures, such as slope stability (Zhao et al. 2018c), reservoir dam stability (Ukpai 2021), rock bursting (Chen et al. 2019a; Ma et al 2022), karst collapse (Bai et al. 2013), and water inrush from mines and tunnels (Huang et al. 2016; Li et al. 2019b; Liu et al. 2022), are caused by the WRI. Analysis and interpretation of the influence of water on the mechanical behavior of rocks are based on the above problems. In recent years, many studies have illustrated the effects of water on the mechanical characteristics of rocks. In general, the presence of water reduces the elastic modulus, compressive strength, cohesion, tensile strength, and rock brittleness (Baud et al. 2000; Erguler and Ulusay 2009; Zhou et al. 2017; Talesnick and Shehadeh 2007) and changes the fracture distribution and fragment shape after rock failure (Haberfield and Johnston 1990; Shen et al. 2020; Guha Roy et al. 2017; Kataoka et al. 2015). The water presence increases the pore water pressure in rock and soil mass, reducing the effective stress of rock and soil mass skeleton particles and changing the physical and mechanical parameters of the rock and soil mass. In addition, it is easy to cause the dissolution or change of mineral composition in rock, soil mass, and cement between grains to produce new mineral composition. The WRI effect will be greater if the water contains corrosive mineral components (Luo et al. 2021). For example, rocks with higher clay mineral content are more susceptible to water penetration (Verstrynge et al. 2014).

This short review focuses on the water weakening of rock strength and its influencing factors, such as the water content, immersion time, and wetting–drying cycles. On this basis, the micromechanism of WRI and its corresponding research methods are addressed.

Mechanical experiments combined with corresponding characterization methods are the primary methods used to study the weakening characteristics of the WRI. In conventional mechanical experiments (uniaxial and triaxial compression tests, Brazilian splitting tests, shear tests, point load tests, needle penetration tests, and Hopkins impact tests, as shown in Fig. 1), the shear, tensile, compressive, hardness, and dynamic load strengths can be obtained. In the test process, in addition to obtaining the stress–strain curves, acoustic emission (AE), infrared radiation temperature, digital image correlation (DIC), 3D laser scanning, and other methods are typically used to characterize the influence of the WRI (Fig. 1).

WRI mechanics test and primary analysis

The weakening analysis of the WRI focuses on the influence of water content, immersion time, and wetting–drying cycles on rock strength, as well as the corresponding AE and infrared radiation (IR) characteristics.

2.1 Water content

Water content can be divided into two aspects: (1) the water content state, including the dry state, natural water content state, and saturated water content state, and (2) the actual moisture content of the rock mass. There are two methods for preparing rock samples with different water contents: direct immersion in water and a non-destructive immersion method that places rock samples in a wet and closed environment to avoid the exchange or hydrochemical reaction between rock and water.

The compressive, shear, tensile strength, and elastic modulus decrease to varying degrees with the increasing water content, as summarized in Table 1. Among these, mudstone rocks have the highest water sensitivity. Granite and other hard rocks have the lowest sensitivity to water, and some rocks are unaffected by water. An increase in water content enhances the plasticity of the rock mass and reduces brittleness (Noël et al. 2021). Some quantitative relationships (fitting formula) between the mechanical parameters (σ_t, σ_c, σ_tc, τ, E, v, c φ, etc.) and water content are summarized in Table 1. In addition, the water content reduces the fracture toughness after rock failure (Zhou et al. 2018a, b; Hua et al. 2015).

AE describes the energy release characteristics of rocks during fracturing or cracking. Hard rocks typically accumulate more energy during loading and would release stronger AE signals during failure. As the water content increases, the cumulative and peak AE signals, high-frequency AE signals, AE signal differences, and AE signal distribution uniformity decrease (Li et al. 2019a, 2021b; Ranjith et al. 2008; Guo et al. 2018; Lin et al. 2019; Zhu et al. 2020; Liu et al. 2019). In addition, with an increase in water content, AE is concentrated in the fracture stage, and the fractal dimension decreases (Song et al. 2020; Kong et al. 2017, 2019). Wet rocks generally produce small fragments during failure, which decreases the fractal dimension, and irregular fractures occur during loading.

For the correlation between moisture content and IR, the average IR temperature (AIRT) generally increases with increasing load and depends on the state of the rock samples (wet or dry, damaged or not). Before the damage, wet rocks generally show a faster increase in the AIRT, whereas dry rocks would produce more increments in the AIRT during damage (Deng et al. 1997; Liu et al. 2010). However, in the entire loading process, the higher the water content, the greater the AIRT, and the smaller the AIRT fluctuation (Zhou et al. 2018c). Sun et al. (2021a) further showed that the applied stress controls IR; for example, the IR count (IRC) will simultaneously increase when the stress suddenly drops. The water content reduces the stress due to water weakening. The above conclusions mainly rely on the uniaxial compression tests. Shear tests by Yao et al. (2020a) concluded that the IR characteristics were comparable to those of the uniaxial compression tests.

With increasing water content, the dynamic load strength and dissipated energy decreased, but the elastic modulus increased. The dynamic strength of saturated rocks is more sensitive to the strain rate than that of dry rocks. With an increase in the strain rate (43.9–156.7 per second), the water weakening effect decreases gradually (Cai et al. 2020b). Further, compared with the uniaxial compressive strength of dry rock samples, that of saturated rock increases with the loading rate in two stages: rapidly increases at low loading rates, and then decreases at high loading rates (Zhu et al. 2021).

2.2 Wetting–drying cycles

Many practical engineering problems involve wetting–drying (WD) cycles, such as rocks in exposed slopes, coastlines, and pumping reservoirs. Similar to preparing rock masses with different water contents, there are also two methods (free and pressure immersion method, air and oven (or heater) drying) for preparing rock samples experiencing WD cycles. A vacuum pressure condition is adopted to accelerate rock saturation when preparing rock immersion, whereas the free immersion method is performed under atmospheric pressure conditions. In the drying process, oven drying can accelerate the drying of rocks compared with air drying. Previous laboratory tests have shown that the WD cycle treatment has a significant impact on the mechanical and physical properties (A_w, P, PLI, SDI, K_IC, K_IIC, K_eff, V_p, σ_c, σ_t, E, c, φ, τ, σ_cd, E_d, T_c, R, and H) of the rock, as summarized in Table 2 (Liu and Zhang 2020; Momeni et al. 2017; Zhou et al. 2018b; Chen et al. 2019c; He et al. 2020; Huang et al. 2022).

Although the WD cycle treatment methods were different in these studies, consistent weakening characteristics were observed. With progressive WD cycles, the porosity and water absorption increased monotonically, whereas the other parameters in Table 2 generally decreased. The decrease in mechanical and physical parameters gradually diminished with the progression of WD cycles (Sun and Zhang 2019; Khanlari and Abdilor 2015; Huang et al. 2010; Fu et al. 2017; Zhao et al. 2018a; Gratchev et al. 2019; Yao et al. 2019b; Wu et al. 2020b; Cai et al. 2020a; Li et al. 2021c). For example, it was found that the first 10 WD cycles significantly impacted the rock's strength; there was no change (Zhao et al. 2021; Guo et al. 2021) in the following cycles. The fracture toughness and crack propagation were also affected by the WD cycle. The fracture energy and fraction coefficient decrease with WD cyclic treatment (Zhao et al. 2017b, c; Song et al. 2019; Ma et al. 2018). The rock's mineral composition has been found to be the main factor affecting the weakening of the WD cycle (Zhou et al. 2017; Tang et al. 2021).

2.3 Immersion time

The immersion time is more closely related to the actual field situation than the water content. Underground rock masses are occasionally immersed in water for months or years, and immersion time significantly affects rock strength (Bai et al. 2016). However, in most laboratory tests, the maximum immersion time is generally no longer than 1 year. The immersion time is currently limited because some rocks (especially those with strong hydrophilicity, such as mudstone) disintegrate after immersion for a short period (Azhar et al. 2020; Fujii et al. 2020). In addition, many studies have shown that some rocks do not weaken even after long-term immersion (Ai et al. 2021; Lyu et al. 2022).

As the immersion time increases, the mechanical parameters (σ_c, σ_t, E, c, φ, τ) of some rocks are weakened, their brittleness decreases, the failure mode becomes stable, and the roughness of the fracture planes increases (Zhu et al. 2020; Ma et al. 2021). The time-dependent immersion weakening varies for different rocks. For example, as the immersion time increases, the uniaxial compressive strength of coarse sandstones first decreases rapidly, then increases slightly, and finally decreases (Wu et al. 2020a). The uniaxial compressive strength and elastic modulus of argillaceous slates decrease with increasing immersion time, whereas Poisson's ratio remains roughly unchanged (Huang et al. 2020b).

2.4 Other WRI related factors

The chemical composition and water pressure play significant roles in WRI weakening. The mechanical parameters (σ_c, σ_t, E, c, φ, τ) of chemically treated samples generally demonstrate more significant weakening compared with natural immersion conditions, especially for pre-fractured samples (Zhang et al. 2019; Gong et al. 2021). For example, the salts contained in water can gradually accumulate in the pore networks of rocks under wetting–drying cycles, which can cause rock deterioration (Jiang et al. 2022). The water immersion height also influences rock strength; the strength of partially presoaked specimens is lower than that of wholly presoaked specimens (Chen et al. 2021). Seepage water pressure enhances the deformation resistance of rock and affects rock strength. As seepage pressure increases, the stress thresholds for crack initiation and damage during rock compression decrease (Xiao et al. 2020; Zhong et al. 2019; Li et al. 2020b). Generally, studies on the WRI of rock mass cover various factors, and strength tests investigate the weakening characteristics under various immersion conditions to reflect engineering environments. Then, it is used to predict the degree of influence of the WRI on engineering scales.

Many studies have shown that WRI is mainly the interaction between water and clay-related minerals in rocks, which changes the pore structure and further degrades their strength. Therefore, in addition to characterizing the macroscopic strength, investigating the internal microstructure is a valuable way to uncover WRI mechanisms. Microscopic observations include laser scanning confocal microscopy (LSCM) (An et al. 2020), polarizing microscopy (PM), scanning electron microscopy (SEM) (Dehestani et al. 2020; Zhang et al. 2014; Liu et al. 2018; Yang et al. 2018; Zhou et al. 2018b; Du et al. 2019), neutron radiography (NR), nuclear magnetic resonance (NMR) (Xie et al. 2018; Zhao et al. 2017a, 2018a, b, 2019b), computed tomography (CT), and small angle X-ray scattering (SAXS) and other methods (Liu et al. 2016; Zhao et al. 2014, 2019a; Wang et al. 2021). The main application of the micro-observation techniques is shown in Fig. 2. The XRD pattern shown in Fig. 2 is typically used to investigate the hydrophilic mineral composition of the rocks.

Main characterization means of pore structure in WRI analysis

The microinvestigation methods shown in Fig. 2 can be divided into three categories: SEM, PM, and LSCM. They are primarily used to observe the surface structure of a rock mass. SEM can be used to observe the mineral occurrence morphology, crystal morphology, surface morphology, and composition. However, the tested samples must be sprayed before SEM scanning; therefore, it is a destructive test. LSCM is mainly used for scanning the fracture surface. NMR and NR are mainly employed to judge pore structures and immersed liquids. The NMR is a non-destructive test that is widely used to characterize pore structures. However, NMR cannot be used to reconstruct a three-dimensional pore structure. CT, SAXS, and XRD are used for X-ray fluoroscopy, and the pore structure and mineral composition can be reconstructed with post-processing. CT is a non-destructive technique that can reconstruct pore and fracture structures in real time and is widely used to track pore and fracture evolutions during rock deformation. However, in the process of three-dimensional reconstruction of CT images, the division of pore fracture and mineral composition threshold is generally manually defined, which sometimes affects the accuracy of the reconstruction model.

Thus, various microcharacterization techniques are generally used for complementary analysis (Zhang et al. 2021; Ai et al. 2021). For example, NMR, XRD, and CT can be used to accurately reconstruct pore structures and mineral compositions (Fig. 3). Specifically, the mineral composition in the sample is first identified by XRD, such as clay mineral composition with strong hydrophilicity, and then the threshold is used during the reconstruction of CT images. Similarly, NMR can accurately provide the pore structure division threshold for CT reconstruction.

Reconstruction process of "pore-fracture" dual structure of a coal sample

Based on the CT reconstruction method (Fig. 3), we reconstructed the pore structure and mineral composition model of the coal samples before and after water immersion. Table 3 shows the distribution of the pores with increased connectivity (blue) and reduced clay minerals (red) before and after immersion. The locations where the clay minerals are reduced coincide with the positions where the connectivity pores are increased. Further, the change in the pore structure in the sample during WRI is mainly caused by the dissolution and expansion of hydrophilic mineral components (Azhar et al. 2020; Huanget al. 2020b; Liu et al. 2021).

Currently, five mechanisms for rock strength weakening caused by water immersion have been proposed: (1) expansion and dissolution of clay minerals, (2) reduction of capillary tension, (3) increase in pore pressure, (4) reduction of fracture energy, and (5) weakening of intergranular cohesion and friction (Zhu et al. 2020; Li et al. 2020a). By comparing the CT images before and after immersion (Fig. 4), the internal damage process caused by the WRI can be observed. Figure 4a and b show the CT images at the same position before and after immersion. The darkening of the greyscales in the CT images indicates an increase in damage and a decrease in density. Therefore, pre-existing weaknesses are displayed in darker colors. As shown in Fig. 4a, there are apparent darker parts (weaknesses) in the dry sample because coal is a typical porous medium with a large number of pore structures (connected pores and isolated pores) and fracture structures (Yao et al. 2019a). After immersion, the size and aperture of the internal fractures in the coal sample increased, and cleats between beddings gradually developed with water degradation. Fracture structures have better connectivity and higher permeability, forming the main flow channels of the fluid medium; meanwhile, clay minerals near the fractures are dissolved in water. In addition, water weakens the intergranular bonding of pre-existing weaknesses, and the development of fractures and pores (connected pores) increases the contact area between water and coal, resulting in the appearance of cleats between beds. For the rock samples, water mainly weakened the cementation between the crystals (Fig. 4d). The stress and energy required for failure along the fractures and cleats are significantly lower than those required for the direct penetration of the coal matrix and rock crystals; therefore, the higher the water content, the smaller the rock strength and the weaker the AE signal (Fig. 4) (Deng et al. 2021; Li et al. 2021a, b; Miao et al. 2021).

CT images of a coal sample before and after water immersion. a Dry sample b Saturated sample c Coal-weakening mechanism and d Rock weakening mechanism

Research on the water weakening of rock masses mainly focuses on the effects of water content, water immersion time, and cyclic water immersion. The strength–weakening degree is characterized by uniaxial and triaxial strength, shear strength, tensile strength, point load, penetration, etc. AE and IR methods are typically used to study the fracture and energy characteristics during loading under WRI. The strength of the rock mass decreases to varying degrees with water content, immersion time, and WD cycles and is related to the type of rock mass and mineral composition. Generally, the strength decreases exponentially with an increase in water content. The previous several WD cycles have a significant impact on the strength of the rock mass and have little effect on the progression of WD cycles. The degree of rock weakening gradually decreases with an increase in the immersion time. The corresponding AE count, intensity, and IR intensity also weaken accordingly. The WRI enhances the plasticity of rocks and reduces their brittleness.

Various microscopic methods have been used to study pore characterization and the weakening mechanism of WRI. SEM can qualitatively observe the fracture structure and mineral composition of the rock mass, but it is impossible to directly compare the porosity before and after water immersion owing to gold spraying during observation. NMR can quantitatively determine the porosity and pore size distribution and is a non-destructive test technique, making it possible to compare the pore size changes before and after immersion. CT scanning combined with corresponding reconstruction algorithms can quantitatively compare the changes in pore structure and mineral composition before and after immersion. However, the threshold division in CT image reconstruction is significant and directly affects the accuracy of the analysis. Thus, various micro-methods can be used to study the evolution of pore structure changes under WRI.

In the WRI weakening mechanism, the clay minerals in the rock mass dissolve in water, which expands the pore structures, increases the connected pores, further expands the primary fractures, and consequently increases the porosity and permeability. At the same time, the presence of water weakens the cementation strength near the primary fractures, making it easier to expand the splitting fracture along the joint surfaces. All these weakening processes lead to a decrease in rock strength and AE intensity.

Currently, studies on both the physical and mechanical properties and the weakening mechanism of WRI are relatively mature. Targeted experiments can only be carried out for special rock masses (different mineral compositions) or water environments (different water chemical compositions) in on-site engineering (water influence conditions). On this basis, the corresponding mechanical test modes (tension, compression, shear, or other tests), water influencing factors (moisture content, wetting–drying cycles, immersion time, etc.) and macro (AE, AIRT, etc.) and micro (SEM, NMR, CT, etc.) characterization methods can be considered. Thus, more attention should be paid to how laboratory-scale tests can be applied in engineering scale practice. Therefore, some qualitative conclusions must be transformed into quantitative models, which can then be applied to field engineering problems through numerical simulations. In addition, when investigating the effect of the immersion time, most laboratory tests are applied for less than half a year; however, long-term water immersion (even hundreds of years) problems generally occur. The weakening mechanism of WRI in rocks can also be explained using the changes in water immersion chemical ions and components.