1 Introduction

Every object exposed to a temperature above absolute zero will emit electromagnetic radiation with a large range of wave lengths (Fig. 1), such as radiofrequency radiation, microwave radiation, infrared radiation (IRR), and light radiation (Brady and Rowell 1986; Martelli et al. 1989; Wu et al. 2000). Among them, the IRR waves fall in the electromagnetic spectrum within wavelengths ranging between 0.7 and 1000 μm, in which the portions within the length of 2.5–15 μm possess the thermal effect. Therefore, the IRR is also popularly known as “heat radiation”. The IRR waves are invisible to human eyes but visually traceable by specific infrared detectors. Due to its non-destructive and non-contact nature, high sensitivity, and space visibility, the IRR detection is widely employed in industrial, engineering, military, and medical sectors (Yao et al. 2013; Cong et al. 2016; Ren et al. 2019). Particularly in the applications of rock engineering, it proves to be an effective approach for remote sensing, such as landslides (Pappalardo et al. 2018), rock burst (Sun et al. 2017b), pillar degradation (Seo et al. 2017), tunnel excavation (He et al. 2010; Sun et al. 2017b), and earthquakes (Saraf et al. 2008). An in-depth and comprehensive understanding of the IRR features of rock materials during deformation and fracturing processes is indispensable to monitor and forecast the natural and human-induced hazards.

Fig. 1
figure 1

Electromagnetic spectrum of infrared region and applications of infrared radiation technology

Full size image

To date, the investigations on the thermal infrared radiation characteristics have been conducted on various types of rocks, including sandstone (Wu and Wang 1998; Wu et al. 2006a, b; Du et al. 2015; Sun et al. 2017b; Cheng et al. 2018; Ma et al. 2019c), marble (Wu et al. 2002, 2004, 2006a; b), granite (Wu et al. 2002, 2006a, , b) Luong and Emami 2013; Du et al. 2015; Salami et al. 2017; Liu et al. 2018), limestone (Liu et al. 2006; Wu et al. 2006a, b; Wang et al. 2016; Salami et al. 2017), coal (Wu and Wang 1998; Wu et al. 2006a, b; Zhao and Jiang 2010; Ma et al. 2013, 2016; Sun et al. 2017a; Li et al. 2018; Ma and Sun 2018; Xiao et al. 2019), shale (Salami et al. 2017), gabbro (Zheng et al. 2017), basalt (Mineo and Pappalardo 2016, 2019), and concrete (Minh Phono Luong 1990; Zhang et al. 2012; Lou and He 2018). Concretely speaking, Minh Phono Luong (1990) first used the infrared thermographic camera to detect the initiation and development of micro-cracks in rock and concrete samples under uniaxial and fatigue loadings. In the first decade of this century, Wu et al. (2000) made significant contributions to the remote-sensing rock mechanics based on thermal IRR thermographic technique. They detected the temperature both on surface and inside the rock using an IRR thermographic camera and thermal-sensitive sensors, respectively. Their test results demonstrates that the temperature measured inside the rock shows the similar trend with respect to the surface IRR temperature. In addition, Wu et al. (2002) investigated the variations of the average infrared radiation temperature (AIRT) for uniaxial loaded marble and granite samples. They reported that the AIRT rises with the increase of applied stress. They also documented that both of rock type and failure pattern can impact the maximum value of AIRT. Afterwards, Wu et al. (2006a, b) conducted a bunch of tests on various types of rock to study the IRR properties of rock. They obtained the spatiotemporal evolution laws of rocks, respectively, subjected to uniaxial compression, and shear and tension stresses. It is suggested that the abrupt changes of the AIRT-time curve and the IRR image can be used as the precursors for rock fracturing. Sun et al. (2017a) proposed an innovative approach to eliminate the background thermal noise using two controlled samples. Sun et al. (2017b) also observed the evolution of IRR temperature field in rock burst tests. It is documented that the AIRT of sample drops obviously and abnormal bands appear before the occurrence of rock burst. More detailed information concerning the IRR characteristics of rock-like materials under deformation is summarized in Table 1. It is widely shared that compression and shear loads lead to the increase of IRR temperature, while the tensile stress results in a temperature reduction.

Table 1 Summary of experimental studies on the IRR features of rock-like materials
Full size table

However, the above-mentioned experimental studies have been limited to dry state. As known to all, water is one of the most crucial and relevant environmental factors controlling rock properties. For in-situ conditions, rock masses are commonly exposed to wet condition in essence, due to the rain, seepage, underground water, and even damp air (Zhou et al. 2018a; Ma et al. 2019a; Cai et al. 2020). Thus, it is highly worthwhile to investigate the effects of water saturation on the IRR behavior of rocks. Moreover, previous research documented that water weakens the rock strength and stiffness (Hawkins and McConnell 1992; Yang et al. 2013; Jiang et al. 2014; Zhou et al. 2016, 2018a, 2019; Zhao et al. 2017; Tang et al. 2018; Wasantha et al. 2018; Bian et al. 2019; Cai et al. 2019, 2020; Weng et al. 2019; Yao et al. 2019; Jia et al. 2019; Liu et al. 2019a; Dang et al. 2019). The loss extent of strength and stiffness is strongly dependent on rock types (Wong et al. 2016; Cai et al. 2019). Whether water shows a pronounced and distinctive impact on the IRR behavior for different rock types is still not quite clear. The IRR characteristics sensitivity to water for different rock types should be revealed in a more detailed manner.

The primary objective of this study is to explore the effects of water saturation on IRR characteristics of rocks as well as the water sensitivity of IRR behavior for different rock types. Three typical rocks were adopted for tests. Uniaxial compression tests were conducted on rock samples under oven-dried and water-saturated conditions to determine the mechanical properties. Meanwhile, the IRR temperature evolution on rock surface was detected and recorded by an infrared thermographic camera. Some important IRR parameters, such as the heating rate in elastic deformation phase, the temperature increment at peak stress, and the maximum temperature on fracture surface before failure, were analyzed.

2 Test Materials and Methods

2.1 Material Characterization and Sample Preparation

Three types of rock commonly encountered in rock engineering were selected for tests, i.e., sandstone, granite, and marble. Figure 2 presents optical polarized micrographs of the three rocks. Their mineral compositions and corresponding grain sizes are listed in Table 2. Table 3 illustrates some essential physical parameters. More detailed descriptions about the three rocks can be referred to our previous publications (Cai et al. 2019).

Fig. 2
figure 2

Optical plane-polarized (left) and cross-polarized (right) micrographs of a sandstone, b granite, and c marble (Q quartz, Kfs kalium feldspar, Pl plagioclase, R rock debris, Ms Muscovite, Bt biotite, Dol Dolomite)

Full size image
Table 2 Percentage and grain size of mineralogical compositions for three rock types (Cai et al. 2019)
Full size table
Table 3 Some essential physical parameters of the tested rocks
Full size table

For each rock type, all samples were obtained from one rock block without visible geological weakness to minimize the scatters of the physical properties across the sample sets. Samples were first sliced into 50 × 50 × 100 mm3 rectangular prisms, and then, six surfaces of each sample were grinded to be smooth and parallel. The ultrasonic wave velocity was measured to collect samples with similar physical properties (Song et al. 2018). Ten samples for each rock type (a total of 30 samples) were used for tests. After that, all samples were placed in a vacuum oven at 105 °C for more than 48 h for the purpose of complete water evaporation.

To determine the immersion time for water saturation, a preliminary immersion test was performed. Three dry rock samples (one for each rock type) were first submerged in purified water for free soaking, and then, they were taken out and weighed every hour until their masses remain unchanged. The continual variations of water content against immersion time for the three rock samples are shown in Fig. 3. It can be clearly observed that the water content of samples remains unchanged in 24 h. Then, the samples are regarded as reaching the completely saturated state. Hereby, half of rock samples were immersed in water for more than 24 h for full saturation in this study.

Fig. 3
figure 3

Variation of water content versus immersion time

Full size image

2.2 Experimental Set-Up

Uniaxial compression tests were conducted using an electric servo-hydraulic material testing machine (INSTRON 1346) installed at the Advanced Research Center at Central South University (see Fig. 4). The displacement rate was controlled at 0.24 mm/min for all tests, i.e., the sample was tested with a constant strain rate of 4 × 10–5/s.

Fig. 4
figure 4

Photographic view of experimental setup

Full size image

An uncooled infrared thermographic camera (SC7000 FLIR) was applied to detect the IRR temperature evolution during the loading process. This camera can detect the infrared radiation for the wavelengths ranging from 3.7 to 4.8 μm with the temperature sensitivity of 0.01 °C. It was operated to capture the IRR images with 112 × 194 pixel resolution for a 62 × 108 mm2 actual area at a 100 Hz frame rate. In tests, the IR camera was placed at a secure distance (1.0 m) from the sample to avoid the damage and disturbance caused by the flying rock fragments. Specific measures were adopted to minimize the environmental impact on the IRR temperature detection as follows: (1) the samples were placed in the laboratory at least 24 h before tests to have the same temperature with the environment; (2) the sample and infrared thermographic camera were placed in a cuboid aluminum isolation box to prevent disturbances from outside, such as light, air motion, and radiation from other objectives; (3) black paint was sprayed on the inner surfaces in the aluminum box to avoid the light reflection by aluminum plates. Additionally, prior studies suggested that the region selected for temperature analysis should be near the sample center which is more sensitive to temperature variation (Zhao and Jiang 2010; Wang et al. 2016). In this paper, we computed the variations of average infrared radiation temperature (△T) in a zone of interest (ZOI) with 84 × 100 pixel resolution in the sample center, as shown in Fig. 5, to represent the global temperature change on the sample surface.

Fig. 5
figure 5

Schematic view for infrared thermographic technology (Ge germanium, ZOI zone of interest)

Full size image

The rock sample also emits acoustic waves due to deformation or damage in tests. The acoustic waves can be detected by the acoustic emission (AE) sensors and then converted into electrical signals, so that they can be recorded. In recent years, the AE signals are widely used to assess the damage condition inside rock (Zhou et al. 2018b; Ma et al. 2019b; Song et al. 2020; Zhang et al. 2020) and locate any defects (Zhou et al. 2020). In the present study, the AE technique was applied to monitor and record the AE counts to facilitate the analysis of crack initiation and evolution. Two AE sensors (Nano 30) and pre-amplifiers (PAC 2/4/6) were used. A threshold value of 45 dB was set to eliminate the potential noises. During whole testing process, the room temperature was kept at 25 °C and the humidity is controlled at 60% by air conditioner to minimize the data discreteness.

2.3 Principle of Thermal Infrared Radiation Detection

An object emits electromagnetic radiation with a large range of wavelengths, in which only the IRR band has the thermal effect. In the infrared thermographic camera, a germanium lens with the specific refraction coefficient can separate the IRR wave from all categories of electromagnetic radiation (see Fig. 5), which indicates that only the IRR waves can be captured by the camera. Subsequently, the radiant energy of IRR waves is received by the focal plane array detector element and then transformed into temperature information according to the Stefan–Boltzmann law (Boltzmann 1884):

$$W_{e} = \varepsilon \alpha T^{4} ,$$
(1)

where \(W_{e}\) is the radiant energy intensity; \(\varepsilon\) is the emissivity of the object, for a blackbody, \(\varepsilon = 1\), for a graybody, \(0 < \varepsilon < 1\); \(\alpha \;{ = }\;5.6696 \times 10^{{{ - }8}} W/(m^{2} \cdot K^{4} )\) is the Stefan-Boltzmann constant; \(T\) is the absolute temperature of the object (K). The output IRR image shows the surface energy intensity map.

Rocks are typically porous-elastic–plastic. Under external stresses, the deformation and fracture of rock will alter its internal energy and trigger energy transition into IRR energy. Hence, the variation of IRR temperature can, to a certain extent, characterize the stress state of rock. The temperature evolution of rock during loading process is dominated by coupled thermo-mechanical effects (Minh Phono Luong 1990; Wu et al. 2006b; Luong 2007), including pore gas escaping, thermal-elastic effect, fracture effect, frictional heat effect, and heat conduction effect.

3 Test Results and Discussion

In this section, we analyze and discuss the experimental results of different rock types in oven-dried and water-saturated states under the following five sub-sections.

3.1 Water-Induced Changes in Mechanical Properties

Table 4 displays the average values of the mechanical properties, including uniaxial compressive strength (UCS), elastic modulus, and critical strain, of the three rock types. All the three mechanical properties are deteriorated by water. The loss extent of these parameters was determined accordingly and is shown in Fig. 6. It can be observed from Fig. 6 that the water-induced reductions in the UCS for the three rocks are markedly distinct. Among the three rock types, the sandstone has undergone the largest decline in UCS, up to 29.8%. The UCS of the granite slightly dwindles by 3.58% after water saturation. However, an approximately negligible UCS loss for the marble is observed, indicating that the UCS of the marble is not apparently affected by water. In addition, the elastic modulus and critical strain follow a similar weakening pattern as for UCS. Under water-saturated condition, the elastic modulus and critical strain of the sandstone display the most drastic drop; however, for the granite, these parameters only undergo a relatively gentle dip while the marble exhibits the least loss in terms of the two properties. The water-weakening mechanisms are not discussed herein, since we have elucidated the reasons in the previous publication (Cai et al. 2019).

Table 4 Average test results of rocks under dry and saturated conditions
Full size table
Fig. 6
figure 6

Loss percentage of mechanical properties of three rock types due to water saturation (UCS uniaxial compressive strength, EM elastic modulus, CS critical strain)

Full size image

3.2 Relationship Between the Variation in Surface Temperature, Uniaxial Stress, and AE Counts

Figure 7 demonstrates the histories of axial stress, AE counts, and the variations of AIRT (△T) versus time for each rock type under dry and saturated conditions. Evidently, the evolution manner of AE counts is similar for all samples regardless of moisture condition and rock types. According to the slope of stress and the occurrence of ring counts, loading history can be divided into four phases. In each phase, the evolution characteristics of the AIRT variation are as follows:

  1. 1.

    Crack closure phase (phase I): where initial defects (such as pores and micro-cracks) are closed (Peng et al. 2019), and quite limited AE counts occur. At the start of loading, the temperature variation of dry rocks may be negative, accompanying drastic fluctuation. This behavior is particularly obvious for the dry sandstone. This is likely due to the shrink of volume and the pore breaking, gases (e.g., CO2, O2, and SO2) primordially enclosed in pore escape and take the heat away. Theoretically, the more porosity or absorbed gases, the more the IRR temperature reduction as well as the more violent fluctuation (Wu et al. 2006b). However, for saturated samples, water occupies the majority of pore spaces instead of gases, the heat loss induced by the escape of pore gases is limited. Hence, the AIRT variation of the three rocks rises at the loading initiation.

  2. 2.

    Elastic deformation phase (phase II): where rock samples deform elastically and micro-cracks initiate steadily. The axial stress on the samples increases fast. The AIRT of rocks continues to rise due to the thermal-elastic effect. This will be analyzed in the next section.

  3. 3.

    Unstable crack propagation phase (phase III): where the increasing rate of the axial stress slightly dwindles. The AE counts exponentially boom as the stress rises, which infers that the storage of elastic energy is strong enough to trigger the eruption of an abundance of micro-cracks (Martin and Chandler 1994; Song et al. 2019; Zhang et al. 2020). In this phase, the variation of AIRT is dominated by the frictional heat effect and fracture effect. As stated by Liu et al. (2006), the generation of cracks (fracture effect) consumes the energy, which will further reduce the AIRT. Contrarily, the friction between cracks, joints, and grains generates much more heat, which, in turn, monotonically increases the AIRT. As shown in Fig. 7, the AIRT of all samples increases in this phase. This is attributed to the produced frictional heat energy compensating the dissipative energy used for crack initiation and propagation. Another reason is due to the fact that the crack initiates from the interior of rock rather than the surface, which postpones the decrease of the detected AIRT on rock surface.

  4. 4.

    Post-peak phase (phase IV): where the axial stress suddenly drops to zero accompanied by the majority of AE counts occurrence. In this phase, the AIRT variation of dry and saturated samples is markedly distinct, which will be further discussed in Sect. 3.5.

Fig. 7
figure 7

Histories of axial stress, AE counts, and temperature variation versus time for different rock types under dry and saturated conditions

Full size image

3.3 Thermal–Mechanical Response Before Peak Stress

Figure 8 plots the thermo-mechanical response of different rocks before peak stress. Generally, the variation of AIRT (△T) rises with the increase of uniaxial stress. During phase I, the variation of temperature shows a rapid rise. Afterwards, the increasing rate of AIRT variation seems to slightly decrease during phase II. This is likely due to the fact that the plastic deformation caused by the closure of initial defects will produce more heat. The stress at the turning point of the increasing rate of △T can be determined as the crack closure threshold.

Fig. 8
figure 8

Thermo-mechanical response of different rocks: a sandstone, b granite, and c marble

Full size image

Figure 8 also illustrates that the variation in AIRT rises linearly with the uniaxial stress (\(\sigma\)) during the elastic deformation phase (phase II). This indicates that the thermal-elastic effect dominates the temperature increment. According to thermal-elastic theory (Wu et al. 2006b; Luong 2007), the variation of the IRR temperature of an isotropic-elastic solid is linearly related to the variation of the sum of principal stresses. Rock material is regarded as an elastic solid in this phase, so the relationship between \(\sigma\) versus △T could be expressed by the following function:

$$\Delta T = a\sigma + b,$$
(2)

where a and b are fitting constants. The fitting curves plotted in Fig. 8 is highly consistent with the testing data set, which indicates that △T increases versus stress with a constant rate. The slope a can be defined as the heating rate.

Figure 9 illustrates the average heating rate in the elastic deformation phase. Under dry state, the sandstone and granite samples hold a higher heating rate compared to the marble. This could be explained from different mineralogical compositions. As listed in Table 2, the sandstone and granite contain some relatively soft minerals, such as mica, smectite, and chlorite, in the former or muscovite and chlorite in the latter. The soft and ductile minerals hold a larger deformation compared to the hard grains under the same stress to generate more heat energy (Salami et al. 2017). However, the marble is mainly composed of the dolomite with high resistance. The stronger dolomite makes the IRR temperature of marble rises in a slighter pattern as the stress increases.

Fig. 9
figure 9

Heating rate of dry and saturated rocks during elastic deformation phase

Full size image

From Fig. 9, we can also clearly observe that the heating rates of the three rocks remarkably increase after water saturation. This indicates that the presence of water facilitates the release of thermal energy during elastic deformation. The potential mechanisms can be interpreted from the following viewpoints:

  1. 1.

    The softening of minerals: the deformation of ductile materials produces more heat compared with hard ones exposed to the same stress state (Vergani et al. 2014; Salami et al. 2017). Because of the water presence, physical weathering occurs for some minerals, especially swelling clay minerals (e.g., smectite) (Hawkins and McConnell 1992; Wasantha et al. 2018; Cai et al. 2019). This renders rock samples more ductile. This mechanism can be characterized by the water-induced reduction in elastic modulus (see Fig. 6). The elastic modulus of the sandstone holds the largest decline due to the highest content of smectite (Cai et al. 2019), while the drop for granite and marble is relatively gentle. Thus, this effect is not pronounced in the granite and marble.

  2. 2.

    Chemical reactions: some of the constituent minerals can chemically react with water under wet environment. According to the mineral composition, two reactions, i.e., the hydrolysis of quartz (Eq. 3) (Atkinson and Meredith 1981; Eppes and Keanini 2017; Zhou et al. 2018c) and the dissolution of calcite (Eq. 4) (Ciantia et al. 2015; Liu et al. 2019b), are dominant as the following formulae:

$$( - {\text{Si}} - {\text{O}} - {\text{Si}} - ) + ({\text{H}} - {\text{O}} - {\text{H}}) \to ( - {\text{Si}} - {\text{OH}} - {\text{HO}} - {\text{Si}} - ),$$
(3)
$${\text{CaCO}}_{3} + {\text{CO}}_{2} + {\text{H}}_{2} {\text{O}} \to 2{\text{HCO}}_{3}^{ - } + {\text{Ca}}^{2 + } .$$
(4)

Both of two reactions are exothermic due to the heat content of the products which is lower than that of the reactants (Bondi and Simkin 1956; Smith et al. 2004). Based on the second law of thermodynamics (Elliott and Lira 2007), the decrease in heat content of products must be compensated by releasing heat energy. In theory, the more quartz and calcite contained in rock, the more heat generated from chemical reactions. Figure 10 plots the relationship between the water-induced increase in heating rate and the content of quartz or calcite mineral. The increasing extent of heating rate linearly rises with the increase of calcite content (see Fig. 10a). However, the correlation between the increasing extent of heating rate and the calcite content is ambiguous, as shown in Fig. 10b. The heating rate increment of the granite containing considerable quartz content is exceptionally lower than that of the marble without quartz. The possible reason is that the calcite dissolution produces much more heat than quartz hydrolysis, although the content of calcite is tiny. The other reason is that the quartz of hydrolysis is limited. The quartz is soluble when the water can flow into the new-formed fracture tip zone (Atkinson and Meredith 1981). Cai et al. (2019) elucidated that the granite is tested under un-drained condition because of the extremely low porosity while the sandstone under drained one. In other words, the water is unable to migrate freely in the granite sample, leading to a limited solution rate of quartz.

  1. 3.

    The increase of shear crack: Hawkins and McConnell (1992) compared the microscopic failure patterns of a large amount of rocks under dry and saturated conditions. They stated that the water presence triggers a transform from axial cleavage (tensile) in dry state into shear or a shear/tensile mixed mode due to the reduction in frictional coefficient at grain boundaries. On the microscopic scale, several researchers identified the crack type by means of acoustic emission method (Cai et al. 2019; Liu et al. 2019b; Yao et al. 2019). They discovered that the tensile crack is dominant under dry condition, while more shear cracks are observed in the water-saturated rock samples. The increasing shear cracks result in more severe crack sliding to produce much more thermal energy.

Fig. 10
figure 10

Water-induced increase in heating rate versus the contents of a calcite and b quartz minerals

Full size image

To sum up, among the three rocks, the water-saturated sandstone exhibits the most drastic increase in terms of heating rate due to the coupled effects of mineral softening, chemical reactions, and increasing shear crack. The dissolution of calcite and the increase of shear crack lead to a moderate increase in the heating rate for the marble. The least increase in heating rate for the granite is attributed to the water-induced shear crack increase.

3.4 Water Saturation Effects on the AIRT Increment at Peak Stress

Figure 11 displays the AIRT increment at peak stress (△Tf) for dry and saturated rocks. Under dry condition, the granite samples hold the largest value of △Tf among the three rocks, followed by the marble and then the sandstone. It is clear that the △Tf of the three rocks dramatically rises after water saturation. This further proves that the presence of water promotes the release of heat energy of rock under compression. The increasing extent of the marble is the most remarkable, up to 114.0%, and those of the sandstone and the granite are 73.3% and 57.9%, respectively.

Fig. 11
figure 11

AIRT increment at peak stress of dry and saturated rocks

Full size image

From Fig. 12, it is evident that the AIRT increment at peak stress not only depends on the moisture condition but also on the rock strength. The temperature increment at peak stress exhibits a linearly positive correlation with rock strength, i.e., it increases with the rise of rock strength under both dry and saturated conditions. The primary reason might be that the brittle rock with higher strength stores abundant strain energy before failure. More energy is potentially dissipated in a thermal manner. Consequently, the increase of rock strength rises the AIRT at failure.

Fig. 12
figure 12

AIRT increment at peat stress versus rock strength under dry and saturated conditions (the gray arrow line denotes saturated data)

Full size image

3.5 Temperature Variation on Fractures

Figure 13 shows the spatial–temporal evolution of the IRR temperature of dry and saturated rocks under uniaxial compression. When the load level is low, the IRR temperature variation on sample surface is quite limited and evenly distributed. Afterwards, the overall temperature rises with the increase of external load. Interestingly, the temperature increment at bottom of sample is apparently higher than that at top. This is probably due to the frictional effect between the sample ends and the loading platen. Before peak stress is reached, no apparent temperature anomalies can be observed on the sample surfaces. This lends support to our hypothesis that crack initiates inside the rock rather than on the surface.

Fig. 13
figure 13figure 13

Evolution of temperature variation and IRR images: a sandstone, b granite, and c marble (figures in the left column represent the overall history of temperature variation, and those in the right column show the temperature variation in post peak)

Full size image

After peak stress, the AIRT variation of rocks violently fluctuates with the decrease of uniaxial stress. The temperature variation depends on the competition between the heat produced by the frictional effect and the heat dissipation caused by the stress decline and the fracture effect. Soon after, surface fragments spall from the rock body. The new-formed fractures with high temperature emerge in the view of the IR thermographic camera, resulting in the abrupt change of temperature. This means that the temperature inside rock is greater than that on rock surface due to the severe frictional effect. At this time, the mean value of the temperature variation in the ZOI is not appropriate for characterizing the temperature increment inside the rock, because the fracture surface area for each sample appearing in the ZOI is different. Hence, we measured the maximum temperature increment (△Tmax) in the final frame of IRR image to characterize the temperature variation inside the rock.

Figure 14 illustrates the maximum temperature increment on fracture surface of dry and saturated rocks. Apparently, the presence of water increases the △Tmax on fracture surface. For the final circumstance, the uniaxial stress drops zero. The thermo-elastic effect is absent due to the stress relaxation. Therefore, the temperature increment of fracture surface is controlled by the fracture effect and the frictional heat effect. On one hand, the water decreases the number of crack creation, this will be reflected by the cumulative ring counts (see Table 4). Moreover, as reported by Zang et al. (1996) and Li et al. (2019), the trans-granular fracture frequently occurs in dry rocks, while the inter-granular fracture commonly happens in saturated rocks probably as a consequence of the reduced grain boundary friction. They render the less energy dissipation in the fracture of saturated rocks. On the other hand, more heat generated from the crack sliding due to the increasing shear cracks during the failure of saturated samples. These two factors dominantly lead to the increase in △Tmax of rock fracture after water saturation.

Fig. 14
figure 14

The max temperature increment on fracture surface of dry and saturated rocks

Full size image

4 Summary and Conclusions

To study the water saturation effects on the thermal IRR characteristics of rocks and reveal the water sensitivity of different rock types, three representative rocks (sandstone, granite, and marble) were selected for tests. Uniaxial compression tests were conducted on oven-dried and water-saturated rock samples. Water-induced variations of their mechanical parameters and AE features were captured. An infrared thermographic camera was applied to monitor and record the evolution of IRR temperature during the whole loading process. The relation between the AIRT variation and the applied stress was obtained. The changes in IRR temperature features due to water saturation such as the heating rate in elastic deformation phase, the temperature increment at peak stress, and the maximum temperature in the fracture surface were analyzed. The main conclusions of this research are summarized as follows:

  • 1. The presence of water weakens the mechanical properties (including UCS, elastic modulus, and critical strain) of rocks in different extents. Among the three tested rocks, the mechanical parameters for the sandstone are the most susceptible to water, and then followed by the granite and the marble.

  • 2. For a given rock type, the surface AIRT of saturated samples is significantly higher than that of dry ones under the same stresses. This indicates that water promotes the release of heat energy within rocks under uniaxial deformation, which further leads to the increase in the heating rate in elastic deformation phase, the AIRT increment at peak stress, as well as the temperature on the new-formed fracture surface.

  • 3. During elastic deformation phase, the water-induced increase in heating rate primarily depends on the mineralogy and secondly on the porosity. For water-saturated condition, the joint effects of mineral softening, chemical reactions, and increasing shear cracks are responsible for the most considerable increase of the heating rate for the sandstone. A moderate increase in heating rate for the marble is mainly induced by calcite dissolution and the increase of shear crack. For the granite, the mineral softening and calcite dissolution effects do not work anymore due to the absence of swelling clay minerals and calcite. The quartz hydrolysis is also limited by the low porosity. Consequently, only the effect of increasing shear crack slightly enhances the heating rate.

  • 4. Besides the moisture state, the UCS of rock also affects the AIRT increment at peak stress. Under both dry and saturated conditions, the AIRT increment at peak stress shows a linearly positive correlation versus UCS.

  • 5. The present research suggests that water plays an important role in controlling the thermal IRR features of rock materials. However, the tested rocks are limited to hard brittle ones. Similar tests will be further conducted on other rock types, such as coal, limestone, mudstone, and shale to testify whether these findings are more generally applicable.