Introduction

In general, moisture is known to reduce rock strength and alter the failure state of rock. Moisture always has a negative effect on the mechanical properties, long-term behaviour and stability of rock masses. The effects of moisture include weathering, erosion, freeze–thaw cycling, chemical and physicochemical degradation and instigation of biological degradation (Verstrynge et al. 2014). Moisture can affect rock in one or both of two possible manners. The first manner is mechanical in nature: moisture content decreases the effective strength of rock (Masuda 2001; Reviron et al. 2009; Navarro et al. 2010; Duda and Renner 2013; Wasantha and Ranjith 2014). Baud et al. (2000) measured a reduction in brittle strength ranging from 5 to 17 % in the presence of water by conducting triaxial compression experiments. Additionally, Taibi et al. (2009) explored the water–mechanical behaviour of partially saturated soft rock at a low confining pressure through uniaxial and triaxial compressive tests. The second manner is chemical. Hydro-chemical interactions between the rock minerals change the mechanical behaviour of rock specimens (Feng et al. 2009; Shukla et al. 2013; Mathurin et al. 2014). The mineral composition is an important factor that determines the effect of moisture on the strength of rock. The subcritical crack growth of rock is considerably affected by the relative humidity and by the clay content (Nara et al. 2010, 2012). Computerized tomographic scanning (CTS) and scanning electron microscopy (SEM) were used in the real-time detection of the weakening behaviour induced in mudstone by water (Jiang et al. 2014).

Many rock burst problems under large ground stresses have been studied through engineering research and in practice (Feng et al. 2006; Jiang et al. 2010; Li et al. 2012; Chen et al. 2013; Zhang et al. 2012). Underground tunnels can be exposed to a complex geological environment due to the presence of rich groundwater content (Lipponen and Airo 2006; Perello et al. 2014; Yu et al. 2014). Therefore, further study of rock mechanical behaviour in deep hard rock tunnels in the presence of moisture is of great significance and has become a prominent research topic in the field of tunnel engineering.

The effects of groundwater on rock deformation have been studied through various field investigations (Valery and Fridrich 1993). Many methods of rock burst prevention that rely on analyses of the effects of moisture have been discussed (Frid 2000; Huang et al. 2011; Cheng et al. 2012; Song et al. 2014). A considerable amount of research has focused on rock mechanical characteristics as a function of moisture content. However, few experiments have discussed the acting mechanisms of moisture from the perspective of microscopic mineral analysis and acoustic emission (AE) energy. The influence of water on rock bursts is rarely studied.

To understand the mechanical characteristics and rock burst mechanisms of hard rock in a rich groundwater environment, a series of uniaxial mechanical tests with respect to different moisture conditions are performed in this study. Stress–strain relationships are monitored, and AE activity and SEM analysis are incorporated as well. The influence of moisture on the mechanical behaviour and rock burst potential of hard rock are primarily analysed from the perspective of microscopic failure characteristics, macroscopic mechanical behaviour and the evolution of AE energy.

Materials

The background of engineering geology

Rock bursts have occurred many times in the Bamiao Tunnel in Sichuan Province of Southwest China. Post-failure images of rock bursts are shown in Figs. 1 and 2. These figures show that the steel arch underwent twisting deformation, leading to a continuous collapse. They also show that the failure range is wide and that the degree of failure is serious. The stratum and lithology are those of Cretaceous rock. The layers of sandstone are mainly composed of the delta–lacustrine depositional system, and the rock mass quality is very high. The rock mass has almost no structural planes or joints. These conditions increase the likelihood of rock bursts.

Fig. 1
figure 1

Distant view of rock burst hazard

Full size image
Fig. 2
figure 2

Grid steel frame distortion

Full size image

Two aspects are important when selecting rock samples in the study of brittle characteristics as affected by moisture: mechanical properties, including high strength and stiffness, and a strong sensitivity to water. Rock samples were obtained from a complete homogeneous rock mass in the Bamiao Tunnel. The sandstone samples were formed into cylinders based on ISRM recommendations, i.e., the sample is a cylinder with dimensions of \( \phi \) 50 × 100 mm. The coring process was controlled according to Chinese National Standard (2001) (SL264-2001).

Properties of sandstone samples

Sandstone samples were examined by X-ray diffraction to determine their mineral composition; a representative X-ray diffraction spectrum is shown in Fig. 3. Constituent mineralogical components and their compositions in sandstone are listed in Table 1. Figure 3 and Table 1 show that the framework grains of sandstone contain not only coarse particles, such as quartz and feldspar, but also clay minerals such as kaolinite, vermiculite and smectite.

Fig. 3
figure 3

X-ray diffraction spectrum of sandstone (c.p.s. means counts per second)

Full size image
Table 1 Mineralogical composition of the tested material (weight %)
Full size table

The influence of water on rock is mainly via the expansion of clay minerals in water from the perspective of mineral composition. Clay minerals appear as tiny flat particles, and their surfaces can produce strong interactions with water. The sandstone in this study has three main types of clay minerals: kaolinite, vermiculite and smectite. Moreover, the hydrotropism, expansibility and shrinkage of smectite are greater than those of kaolinite and vermiculite. Differences in mineral composition can lead to different water absorption abilities and to different expansion ratios, thus resulting in the occurrence of micro cracks.

In addition, the influence of water on rock can be mainly classified into two types from the perspective of engineering geology: on the one hand, interstitial hydraulic pressure can be produced by water filling the cracks of rock, and then, the effective normal stress of any surface can be decreased. Therefore, the friction resistances of these surfaces will be decreased. On the other hand, the engineering geological characteristics can be changed through the influence of water on the rock structure. Moreover, the influence of water on the mechanical properties of rock can be classified into two types: (1) wedge action—the water molecules will impose a certain pressure on the two walls of a crack, resulting in a considerably decreased rock strength—and (2) lubrication action—a water film will be produced along the crack wall after the water enters the rock crack. The water molecules can separate the solid surface and function as a lubricant when relative shear displacement has occurred on the two walls, resulting in an increased deformation performance and decreased friction angle. In general, the internal friction angle \( \varphi ^{\prime } \) and effective normal stress \( \sigma ^{\prime } \) of any surface are decreased under the action of water. According to the shear strength equation of effective stress,

$$ \tau _{{\text{f}}} = c^\prime + \sigma ^\prime \tan \varphi ^\prime $$
(1)

where \( \tau_{\text{f}} \) is the effective shear stress on the shear failure surface, \( c^{\prime} \) is the effective cohesion intercept between rock particles, \( \sigma^{\prime } \) is the effective normal stress on the shear failure surface, and \( \varphi ^{\prime } \) is the internal friction angle.

Equation (1) shows that \( \sigma ^{\prime } \) and \( \varphi ^{\prime } \) of internal rock gradually decrease with increasing water content. Then, rock particles are more likely to slip past each other, resulting in a decreased \( \tau_{\text{f}} \). Therefore, the rock failure strength gradually decreases.

Test methods

Sample preparation and testing

Moisture conditions were regulated by controlling the moisture content. Five levels were established by selecting the following values of the water saturation coefficient k s: 0 (dried state), 0.19, 0.48, 0.64 and 1 (saturated state). The water saturation coefficient \( k_{\text{s}} \) is calculated as follows:

$$ k_{\text{s}} = {\raise0.7ex\hbox{${\omega_{\text{a}} }$} \!\mathord{\left/ {\vphantom {{\omega_{\text{a}} } {\omega_{\text{sa}} }}}\right.\kern-0pt} \!\lower0.7ex\hbox{${\omega_{\text{sa}} }$}} $$
(2)

where k s is the water saturation coefficient, ω a is the actual water absorption and ω sa is the water-saturated absorption.

The moisture level control procedure was as follows:

  1. 1.

    The natural density and natural moisture content of each sample were similar because all samples originated from the same rock. The mass and size of sandstone under different moisture conditions were tested. The testing results demonstrated that the average dried density was 2.6 g/cm3.

  2. 2.

    The size and natural weight of the samples were measured, and the dried weight was calculated from the average density and volume in the dried state. Then, the saturated weight as a function of the water saturation coefficient of each sample was calculated with the set moisture content and dried weight.

  3. 3.

    A temperature range from 105 to 110 °C was used to achieve the dried condition. First, all samples were dried. Second, the dried samples were exposed to water, whose absorption served to establish different moisture conditions. If the moisture content exceeded the set value, the sample was placed in air to decrease its internal moisture content. The entire process was controlled by monitoring the weight of the sample in real time.

The average mass increase of the water-saturated sandstone in comparison to dried sandstone was determined to be approximately 6.8 g. The mass difference following water absorption between each level was approximately 1.7 g when the water content was divided into five levels. It is difficult to accurately control moisture content during sample preparation; therefore, a certain error exists between the actual water absorption and the set water absorption. The rock samples before the test are shown in Fig. 4. Their physical properties are listed in Table 2.

Fig. 4
figure 4

Prepared sandstone samples for uniaxial compression tests

Full size image
Table 2 Physical properties of sandstone samples for uniaxial compression testing
Full size table

Test principles and methods

Uniaxial compression tests were conducted by imposing an axial load on samples under unconfined conditions at different moisture conditions and by measuring the resulting axial deformation. Wawersik and Fairhurst (1970) divided the failure pattern of rock into two types (I, II) according to the stress–strain curve obtained during uniaxial compression tests (Fig. 5). Type I corresponds to stable failure. The accumulated deformation of rock samples after the peak does not continue to drive crack growth. However, a crack can be further expanded by subjecting the rock to additional loading. Rock characterized by low hardness tends to exhibit type I characteristics. Type II corresponds to unstable failure. Additional external loading is not necessary to promote failure; instead, the sudden outburst of accumulated strain energy can cause rock failure. In general, brittle hard rock often tends to exhibit type II characteristics and is susceptible to rock bursts. The complete stress–strain curve was measured and enabled the calculation of the uniaxial compressive strength, elastic modulus and Poisson’s ratio.

Fig. 5
figure 5

Failure type under uniaxial compression (according to Wawersik and Fairhurst 1970) I stable failure, II unstable failure

Full size image

The loading rate was controlled at an axial displacement rate of 0.1 mm/min in the elastic deformation stage and then at a lateral deformation rate of 0.03 mm/min until ultimate failure when the sample entered the failure stage.

Test results

The effects of water on the mechanical behaviour, AE characteristics and fracturing behaviour of sandstone are discussed below.

Effects of water on mechanical behaviour

The characteristics of the mechanical behaviour under various moisture conditions are discussed in terms of five categories: (1) stress–strain response, (2) uniaxial peak strength, (3) axial peak strain behaviour, (4) elastic modulus and (5) Poisson’s ratio.

Stress–strain response

The stress–strain characteristics obtained from the uniaxial compression tests under different moisture conditions are shown in Fig. 6.

Fig. 6
figure 6

Uniaxial stress–strain curve under different moisture conditions

Full size image
  1. 1.

    The stress–strain curves are in accordance with the type II curve under different moisture conditions. Moreover, the peak strengths were relatively high (up to 90 MPa), and the peak strains were all relatively small (less than 9 × 10−3). The sandstone exhibits high strength—an enabling factor of rock bursts—as well as brittle failure characteristics.

  2. 2.

    Elastic deformation stage: The stress–strain curve slightly increased when the stress level was low, therein presenting a concave shape. This indicates that sandstone features a relatively small number of microscopic cracks, which results in the compression of internal particles in the sandstone. The stress–strain curve transitions to a straight line when the pressure increases to a certain value, producing a convex shape. This feature indicates that the microscopic cracks have entered into the stable extending stage.

  3. 3.

    The stage after peaking: The stress–strain curve decreased after the peak. The loading rate control was converted to the lateral deformation control of 0.03 mm/min until complete failure when the sample entered the failure stage. Hysteresis appeared after the peak as a result of the use of lateral deformation control. Moreover, the cohesion of sandstone is high, and the bond strength between particles is inhomogeneous, resulting in a plastic deformation stage indicated by the part of the curve exhibiting a convex shape. The stress was quickly reduced, and the change in strain also decreased, which indicating a brittle failure.

Uniaxial peak strength

The dependence of the uniaxial peak strength on the moisture content is shown in Fig. 7.

Fig. 7
figure 7

Peak strength as a function of moisture content

Full size image

The moisture content strongly influenced the uniaxial peak strength, and increasing moisture content reduced the peak strength. The peak strength was 171.89 MPa when \( k_{\text{s}} = 0 \). The peak strength rapidly decreased to 132.40 MPa, corresponding to a strength reduction of 23 %, as the moisture content was increased to 0.19. With further increase in moisture, the peak strength decreased in a more gradual manner. Finally, the peak strength reached 103.39 MPa (60.2 % of the peak strength of dried sandstone) when \( k_{\text{s}} = 1.00 \).

Axial peak strain variation

The dependence of the axial peak strain on moisture content is shown in Fig. 8.

Fig. 8
figure 8

Peak strain as a function of moisture content

Full size image

Figure 8 shows that the axial peak strain decreased with increasing moisture content. The maximum axial peak strain was 8.38 × 10−3 when \( k_{\text{s}} = 0 \). However, the peak strain decreased to 6.62 × 10−3, and a strain reduction of 21 % was observed when the moisture content was increased to 0.19. The peak strain continued to decrease with an increasing moisture content, reaching a minimum of 6.12 × 10−3 (corresponding to 73.1 % of the axial peak strain of dried sandstone) when \( k_{\text{s}} = 1.00 \).

Elastic modulus

The elastic modulus of sandstone was calculated as suggested by the Chinese National Standard (2001) (SL264-2001):

$$ E_{50} = {\raise0.7ex\hbox{${\sigma_{50} }$} \!\mathord{\left/ {\vphantom {{\sigma_{50} } {\varepsilon_{\text{L50}} }}}\right.\kern-0pt} \!\lower0.7ex\hbox{${\varepsilon_{\text{L50}} }$}} $$
(3)

where \( E_{50} \) is the elastic modulus of rock (GPa), \( \sigma_{50} \) is the stress value corresponding to 50 % of the peak value, and \( \varepsilon_{\text{L50}} \) is the strain in the lengthwise direction when the stress value is \( \sigma_{50} \).

The dependence of the elastic modulus on the moisture content is shown in Fig. 9.

Fig. 9
figure 9

Elastic modulus as a function of moisture content

Full size image

The elastic modulus decreased with increasing moisture content. The maximum elastic modulus was 19.42 GPa when \( k_{\text{s}} = 0 \). A reduction of 39 % was observed when the moisture content reached 0.64; past this point, the elastic modulus rapidly decreased. The minimum elastic modulus was 15.23 GPa (representing 78.4 % of the elastic modulus of dried sandstone) when \( k_{\text{s}} = 1.00 \).

Poisson’s ratio

Poisson’s ratio was calculated as suggested by Chinese National Standard (2001) (SL264-2001):

$$ \mu_{50} = {\raise0.7ex\hbox{${\varepsilon_{{{\text{d}}50}} }$} \!\mathord{\left/ {\vphantom {{\varepsilon_{{{\text{d}}50}} } {\varepsilon_{{{\text{L}}50}} }}}\right.\kern-0pt} \!\lower0.7ex\hbox{${\varepsilon_{{{\text{L}}50}} }$}} $$
(4)

where \( \mu_{50} \) is Poisson’s ratio, \( \varepsilon_{\text{L50}} \) is the axial strain corresponding to \( \sigma_{50} \), and \( \varepsilon_{\text{d50}} \) is the circumferential strain corresponding to \( \sigma_{50} \). The axial and circumferential strains are assumed to be compressive and tensile, respectively.

Measured values of the Poisson’s ratio as a function of moisture content are shown in Fig. 10.

Fig. 10
figure 10

Poisson’s ratio as a function of moisture content

Full size image

The values generally increased with increasing moisture content. The minimum Poisson’s ratio was 0.076 when \( k_{\text{s}} = 0 \). In addition, the Poisson’s ratio increased to 0.087 and then decreased to 0.078 before decreasing again. The Poisson’s ratio reached 0.155 when \( k_{\text{s}} = 1.00 \), and this value was approximately 2.04 times the Poisson’s ratio of dried sandstone.

Effects of water on AE energy

AE occurs as a result of energy being released from the original microscopic cracks in rock. The observed AE characteristics during failure play an important role in the analysis of brittle failure mechanisms.

Figure 11 shows the AE energy rate as a function of time for all samples. Axial stress measurements over time are plotted in the same figures. It is clear that the internal energy of sandstone is continuously accumulated with time. The energy is suddenly released and the rock sample undergoes failure when the strain energy reaches the maximum threshold of the rock. The process of energy release exhibits the characteristics of brittle failure when the water saturation coefficient is low (see Fig. 11a, b) and the characteristics of plastic failure when the coefficient is high (see Fig. 11c–e). Overall, the AE energy is not easily detectable during the initial period, and it is maximized at the moment of failure. In Fig. 11a, b, the AE energy was detectable during and after failure. In Fig. 11c–e, the AE energy was detectable before and during failure. The AE energy before failure increased and the AE energy after failure decreased with increasing moisture. Moreover, the peak energy rate decreased with increasing moisture.

Fig. 11
figure 11

AE energy rate and \( \sigma_{1} \) as a function of time

Full size image

Maximum instantaneous AE energy

The relationship of maximum instantaneous (MI) AE energy with moisture content is shown in Fig. 12.

Fig. 12
figure 12

MI AE energy as a function of moisture content

Full size image

The MI AE energy decreased with increasing moisture content. The maximum moisture-dependent MI AE energy was 32.65 × 103 aJ when \( k_{\text{s}} = 0 \). The MI AE energy decreased to a minimum value of 18.82 × 103 aJ when \( k_{\text{s}} = 1.00 \). The energy exhibited a sudden reduction when the moisture content was changed from 0.48 to 0.64.

Cumulative AE energy

The cumulative AE energy was calculated as the summation of the AE energy before time t:

$$ E_{\text{c}} = \sum\limits_{t = 0}^{t = n} {E_{t} } $$
(5)

where \( E_{\text{c}} \) is the cumulative AE energy and t is time.

The dependence of the cumulative AE energy on the moisture content is shown in Fig. 13.

Fig. 13
figure 13

Cumulative AE energy as a function of moisture content

Full size image

The cumulative AE energy decreased with increasing moisture content. The cumulative AE energy of dried sandstone was 30.6 × 106 aJ but decreased to a minimum of 16.2 × 106 aJ at the saturation point. Moreover, the cumulative energy exhibited a sudden reduction when the moisture content was changed from 0.48 to 0.64.

Figures 12 and 13 demonstrate that the MI AE energy and cumulative AE energy decreased with increasing moisture content and exhibited sudden decreases when the moisture content was increased from 0.48 to 0.64. Microscopic cracks are more likely to initiate and propagate at the larger end of this range, resulting in the reduction in MI AE energy and cumulative AE energy.

Effects of water on fracture behaviour

Table 3 presents post-failure samples at different moisture levels. Note that the width and length of the failure cracks increase with increasing moisture content. Moreover, the decrease in the failure plane angle with increasing moisture content implies a decrease in the friction angle, possibly as a result of the increasing degree of plastic behaviour exhibited by the sandstone under conditions of higher moisture.

Table 3 Post-failure samples tested at different degrees of moisture
Full size table

In addition, microscopic failure features at the fracture surface at different moisture levels can be identified by SEM, as shown in Fig. 14. In Fig. 14a, the surface is smooth and exhibits a shallow stepped pattern, which is conclusive evidence of brittle deformation. In Fig. 14b, the surface is relatively smooth, and shallow fluvial patterns can be observed, indicating some degree of plastic deformation. In Fig. 14c, the surface is rough, and a large number of river patterns have joined to form a cleavage crack, thus indicating a greater amount of plastic deformation. An obvious feature of Fig. 14d is that the layer structural plane is dilacerated. Figure 14e shows prominent roughness and signs of large-scale failure in the form of local arc-shaped patterns. From the microscopic characteristics described above, it can be concluded that the fracture surface becomes coarser with increasing moisture content.

Fig. 14
figure 14

Microscopic failure features arising from different moisture conditions, as imaged using SEM

Full size image

Discussion

This study reveals important behaviours regarding water-based weakening mechanisms in sandstone. In addition, the moisture effect response and its influence on strength are highlighted. X-ray diffraction and SEM were used to characterize the weakening influence of water. The progressive fracturing behaviour of sandstone was also analysed by monitoring AE events.

The X-ray diffraction results show that the original sandstone contains clay minerals, resulting in particle swelling at different water content levels in water, which increases the inter-particle pore space and promotes inter-particle slip. Then, the internal frictional angle and effective normal stress of internal rock gradually decrease with increasing water content. Consequently, the effective normal stress and the friction force between the particles are affected (see Fig. 15a, b). Therefore, the compressive strength of the rock sample decreases, and the internal energy is released relatively early because of particle slippage under the action of external stress.

Fig. 15
figure 15

Microscopic structure of sandstone under dry and moist conditions

Full size image

The uniaxial test verified the above analysis from a macroscale perspective. The uniaxial compressive strength and elastic modulus decreased considerably, and the peak axial strain decreased moderately, whereas the Poisson’s ratio increased with increasing moisture. The decreasing failure plane angle with increasing moisture implies a decrease in friction angle, which is connected to the increasing amount of plastic behaviour exhibited by the sandstone at higher moisture levels. Moreover, the axial strain of saturated sandstone substantially increased, whereas the axial stress remained constant, thus indicating that the reservoir capacity of the strain energy decreased with increasing moisture. The internal energy was dissipated through the process of plastic strain, which provided an alternative to brittle fracture.

Furthermore, AE tests were also conducted in this study. The AE energy results indicate that the internal released energy increased before failure and that the maximum released energy decreases with increasing water content (see Fig. 11). A mechanical explanation for this is that the energy was previously dissipated through inter-particle slippage promoted by increases in the inter-particle pore space.

Combined with the SEM results, it can be concluded that the fracture surface becomes coarser with increasing moisture content. This is due to the higher internal friction angle and effective normal stress between particles under the condition of low moisture content. Therefore, particles rarely slip, and a crack is directly generated when the sandstone is under the stress effect. Finally, the failure fracture presents brittleness and smoothness. On the contrary, the internal friction angle and effective normal stress decrease, resulting in a higher probability of particles slipping past each other under high moisture content conditions. Therefore, particle slippage and plastic deformation occur primarily when the sandstone is under the stress effect, leading to the generation of cracks. Clearly, the accumulation of internal energy is preferentially dissipated through particle slippage, and the released energy decreases at the moment of ultimate failure. The failure fracture presents plasticity and roughness.

Conclusion

The following conclusions can be drawn from the analysis of the stress–strain curves, mechanical parameters, AE energy, macro failure modes and microscopic failure features under different moisture conditions:

  1. 1.

    With X-ray diffraction technology, the sandstone in this study is found to contain three types of clay minerals: kaolinite, vermiculite and smectite. Differences in mineral composition can lead to different water absorption abilities and to different expansion ratios, thus resulting in the occurrence of microscopic cracks. Moreover, the internal friction angle and effective normal stress of particles decreases because of the water absorption abilities of clay minerals. Sandstone particles are more likely to slip past each other at a higher moisture content.

  2. 2.

    Sandstone has a high strength, which is associated with brittle failure and a greater likelihood of rock bursts. The uniaxial compressive strength and elastic modulus significantly decrease, whereas the Poisson’s ratio and peak axial strain increase with moderately increasing moisture. These mechanical results are in good agreement with the AE results. The AE phenomenon shows that MI energy and cumulative energy decrease with increasing moisture. The dissipative energy before failure increases with increasing moisture, meaning that partial internal energy dissipates in advance. It can be concluded that the brittle failure phenomenon of sandstone becomes less prominent and that plastic failure predominates with increasing moisture, revealing that moisture could inhibit rock bursts.

  3. 3.

    The SEM results show that the fracture surface becomes coarser with increasing moisture content. The microcosmic reasons are that particles rarely slip and that cracks are directly generated under low moisture conditions when the sandstone is under the stress effect, leading the failure fracture to present brittleness and smoothness. On the contrary, particle slippage and plastic deformation occur primarily at a higher moisture content, leading to the generation of cracks. The failure fracture presents plasticity and roughness. In response, the microscopic failure surface becomes rougher, and the characteristics of plastic failure became more prominent. It can be concluded that the brittleness of sandstone decreases and that its plasticity increases with increasing moisture.