Dynamic Compression Properties and Deterioration of Red-Sandstone Subject to Cyclic Wet-Dry Treatment

State Key Laboratory for Geomechanics and Deep Underground Engineering, China University of Mining and Technology, Xuzhou 221116, Jiangsu, China Institute of Architectural Engineering Technology, Jiangsu Vocational Institute of Architectural Technology, Xuzhou, Jiangsu 221116, China Jiangsu Collaborative Innovation Center for Building Energy Saving and Construction Technology, Xuzhou, Jiangsu 221116, China


Introduction
Weathering is defined as the alteration and destruction process of physical and chemical effects on rock and soil materials at or near the earth's surface. is phenomenon is widespread in rock materials [1][2][3].Due to the deterioration induced by various weathering treatments, the assessment on engineering properties of rock mass is significantly important.As a typical type of the weathering process, cyclic wet and dry are prevalent in rock engineering, especially in fields of tunnelling, slopes, dams, and mining, where rock materials are usually subjected to periodical changes in the moist condition (namely, seasonal changes in groundwater level, water level changes in reservoirs, recurrent rainfall, and ebb and flow of tides) [4,5].Several test methods have been developed in the past decades to simulate the cyclic wet and dry actions [6][7][8][9][10][11]. Actually, an intact cycle of wet-dry consists of wet process (the transition from dry to wet) and dry process (the transition from wet to dry) [5].To accelerate the simulation process, the wet process is usually conducted by submerging rock materials in water for different time (e.g., 12 h, 24 h, 32 h, 48 h, and 3 d).e dry process is mainly controlled by the dry way and dry time.e air-dried and oven-dried are adopted widely.Although there is no universal standard for dry time t, the most common dry time are 12 h, 24 h, and 48 h.In addition, for the oven-dried type, the temperature is generally controlled below 105 °C to reduce the temperature effect on mechanical properties of rocks [7,8].
Based on the accelerated wet-dry method, many researchers have conducted the study on mechanical and physical properties of rock materials after cyclic wet-dry treatment [5,7,9,10,[12][13][14][15][16], such as density, porosity, water absorption, slake durability index, P-wave velocity, static uniaxial compression strength (UCS), tensile strength, and fracture toughness.Although it is impractical to test each property index of weathered rocks, it is universally accepted that cyclic wet-dry reduces both mechanical and physical properties of rock materials to varying degrees.As for physical properties, a substantial experiment concluded that, with the increasing wet-dry cycles, the P-wave velocity, density, and slake durability index show an increasing trend, while the porosity and water absorption increase due to the wet-dry induced by deterioration.In terms of mechanical properties, investigations were mainly concentrated on UCS, tensile strength, and fracture toughness.For example, the unconfined compression strength of sandstone specimens was measured by Hale and Shakoor [13] after 50 wet-dry cycles in total.ey pointed out that there is no obvious reduction in UCS.Özbek [15], Khanlari and Abdilor [16], and Zhao et al. [17] obtained similar results after different wet-dry cycles.However, a sharp decrease in uniaxial compression strength (reduced by 86.1%, after 8 cycles), tensile strength (reduced by 73.42%, after 10 cycles), and fracture toughness (reduced by 52.4%, after 7 cycles) was reported by Yao et al. [18], Liu et al. [19], and Hua et al. [20], respectively.e main causes for this discrepancy are the different mineralogical composition and microstructure (e.g., voids, cementations, texture, and fissures), which affect the sensitivity to cyclic wet-dry treatment [17].As the testing technique progresses, the wet-dry induced mechanism in the microscale has been visualized with the aid of microcomputed tomography (CT), scanning electron microscopy (SEM), and nuclear magnetic resonance (NMR).Zhang et al. [21], Zhou et al. [22], and Liu et al. [23] utilized the SEM technique to observe the microstructure of the rock specimen subjected to different wet-dry cycles and concluded that the combination of rock grains was weakened by the microcracks and argillization.Liu et al. [19] conducted a series of CT whole-section scanning on sandstone specimens under dry-wet cycles and obtained that with the increasing dry-wet cycles, the density decreases whereas the porosity increases.Water distribution in saturation and dry processes was concluded from NMR images by Zhou et al. [24].e test results explained the tensile strength differences under the same water content but different water treatments.However, the above research studies mainly focus on static mechanical and physical properties of rock materials.
ere is rare research on dynamic behaviors of weathered rock, especially the study on cyclic wet-dry cycles.Actually, rock materials may be broken dynamically to some extent in the engineering application, due to the impact loads induced by blasting, drilling, vibration, rock-bursts, or earthquake.erefore, it is meaningful to study the dynamic properties of rocks subjected to cyclic wet-dry.Recently, researchers have paid attention to this issue.Zhou et al. [5,22] analyzed the dynamic compression strength and dynamic tensile strength of sandstone after every 10 wet-dry cycles (50 cycles in total) by the modified split-Hopkinson pressure bar system and concluded that both dynamic compression and tensile strength decrease with the increase of wet-dry cycles.Considering combined effects of deterioration and loading rate in wet-dry cycles, a new decay model was established by Zhou et al. [22] to forecast the long-term dynamic tensile strength against many more cyclic wet-dry cycles.However, on account of varying rock types and different mechanical properties, more work should be done with cyclic wet-dry treatment and dynamic properties.
To investigate the cyclic wet-dry effect on static and dynamic compression properties of rocks, a series of experiments on red-sandstone were conducted after 0, 5, 10, 15, and 20 wet-dry cycles in this work.
e dynamic compression tests were conducted by using the split-Hopkinson pressure bar system, of which the impact velocity was also considered to assess the loading rate effect in this study.
e density, P-wave velocity, and the surface microscopic features were also analyzed to characterize the deterioration caused by cyclic wet-dry.Moreover, in order to reflect the loading rate and wet-dry cycles, a deterioration model was built to forecast the long-term dynamic compression strength of red-sandstone.

Specimen Preparations.
In this study, red-sandstone specimens were selected from Linyi, Shandong province.Results of X-ray diffraction indicated that this kind of redsandstone mainly consists of quartz (66.4%), feldspar (18.3%), calcite (8.6%), hematite (6.3%), and small amounts of chlorite, illite, and mica.Specimens were all extracted from the same red-sandstone block which has the good petrographic uniformity and geometrical integrity.Besides, ultrasonic detection was applied to minimize variations and guarantee good homogeneousness among specimens.According to the recommendation of International Society for Rock Mechanics (ISRM) [25,26], red-sandstone was cut into normal cylinder specimens, Φ50 × 25 mm for dynamic compression tests and Φ50 × 100 mm for static compression tests, respectively.All the specimens were carefully polished to satisfy the dimensional standard suggested by the ISRM.Table 1 shows the essential strength and physical properties of red-sandstone.

Cyclic Wet-Dry Setup.
In this study, a single wet-dry cycle was performed by submerging red-sandstone specimens into purified water for 24 hours.en, specimens were dried in an oven for 24 hours with a constant temperature of 60 °C and cooled to the room temperature.ese specimens were divided into five groups, subjected to 0, 5, 10, 15, and 20 wet-dry cycles, for static and dynamic compression tests.

Ultrasonic and SEM Tests.
As another physical property of rocks, the ultrasonic P-wave velocity is sensitive to inner microdeterioration caused by cyclic wet-dry.e P-wave velocity was measured by using a portable ultrasonic detector (RSM-SY6).It is developed by the Institute of Rock and Soil Mechanics, Chinese Academy of Science, with a primary frequency of 50 kHz and the sampling precision of 0.2 μs.Vaseline was applied to the top and bottom surfaces of specimens for a good coupling effect between the transmitter 2 Advances in Civil Engineering and the receiver.Besides, information of longitudinal waves was automatically recorded for the following analysis.

Mechanical Experiments.
Static compression tests were conducted by using the rock mechanics testing system of RMT-150B.e loading mode was controlled by the displacement, and the loading rate was set at 0.05 mm/min.Displacements and loads were recorded automatically and processed for calculating the uniaxial compression strength.As shown in Figure 1, the dynamic compression tests are performed by the SHPB test system of Φ50 mm. e SHPB system mainly consists of an incident bar, a striker bar, and a transmitted bar, with a length of 290 mm, 2400 mm, and 1400 mm, respectively.e specimen is placed between the incident and transmitted bars, which are composed of high strength 40Cr steel with an elastic modulus of 210 GPa and a density of 7800 kg/m 3 .When the strike bar is launched with the help of a nitrogen gasholder, a compression longitudinal incident wave (ε i ) is formed at free end of the incident bar and propagated along with the incident bar soon afterwards.When reaching the bar-specimen interface, some of the waves propagate through the specimen and transmit to the transmitted bar (ε t ), whilst other waves are reflected back to the incident bar (ε r ).All waves are recorded by two strain gauges which are fixed on the incident bars and transmitted bars.
e shaping technique was employed to guarantee the dynamic equilibrium according to methods suggested by the ISRM [26,27].A soft rubber sheet was set on the free end of the incident bar as a pulse-shaping device, which was 1.0 mm in thickness and 10 mm in diameter.Figure 2 illustrates the typical impact pulse wave in the SHPB compression test.It can be clearly seen that the sum of incident wave and reflected wave is nearly equal to the transmitted wave, indicating that the stress equilibrium has been achieved during the impact process, and the inertial effect can be ignored [28,29].According to the 1D wave theory and verification of dynamic stress equilibrium, the dynamic parameters, such as stress, strain, and strain rate, can be calculated through the three-wave analysis [28].

Index Property Variations.
e variation of P-wave velocity, density, and static uniaxial compression strength is listed in Table 2. e data clearly illustrate that, with the increase of wet-dry cycles, three index properties all decrease to varying degrees.After 20 wet-dry cycles, compared with untreated specimens, the density, P-wave velocity, and static UCS are reduced by 3.62%, 24.93%, and 18.28%, respectively.Besides, the most obvious decrease occurs in the P-wave velocity.Similar to many other studies [5,22], it is found that the variation of three index properties changes quickly in the first few cycles, and then gets slowly with the increasing wet-dry cycles.It can be inferred that the first few wet-dry cycles have a greater responsibility for the rock deterioration.
Figure 3 shows typical static stress-strain curves of redsandstone specimens subjected to different wet-dry cycles.It can be seen that curves under varying cycles are similar in shape, indicating an increasing trend with the increase of strain.When the strength reaches the maximum, there is a sharp drop, corresponding to the typical feature of brittle failure.Otherwise, in the initial stages, there is a compaction process, which is expressed with a concave curve.With the increasing wet-dry cycles, the compaction process is more obvious, resulting from the deterioration induced by wetdry.
e increasing critical strain also reflects this point.Above all, the rocks are more soft and ductile after the wetdry treatments. is is primarily caused by the expansion and growth of internal microcracks.

Microscopic Observations.
As discussed in Section 3.1, the cyclic wet-dry treatment has an adverse effect on the physical and mechanical properties of red-sandstone.To illustrate the deterioration mechanism induced by wet-dry cycles, the scanning electron microscope (SEM) technique was adopted to visualize the microstructure changes.Five cylinder samples with the dimension of 10 mm were prepared for the SEM test, before which samples were subjected to 0, 5, 10, 15, and 20 wet-dry cycles, respectively.Figure 4 shows typical SEM images of red-sandstone samples after different wet-dry cycles.It can be found that the microstructure of untreated specimen is intact, and loose granules [30][31][32] and new microcracks are hard to be observed (Figure 4(a)).For specimens subjected to 5 wet-dry cycles, a part of loose granules begin to separate from main grains, as shown in Figure 4(b).When 10 wet-dry cycles are reached, two new-formed microcracks are clearly seen around the grain (Figure 4(c)), demonstrating the partial failure of connection between two grains.With the increasing wet-dry cycles, not only the growth and expansion of microcracks but also the number and density of loose granules increase significantly, as shown in Figures 4(d) and 4(e).Additionally, microholes which are possibly caused by the dissolution of soft grains appear after 5 wet-dry cycles, as shown in Figures 4(c)-4(e).Due to the increasing microcracks and microholes, the grains, cements, or the connection between them are weakened or even destroyed, mainly resulting in the decay of physical and mechanical properties.e observation results are consistent with previous studies [5,22,23].Advances in Civil Engineering and equilibrium verification on red-sandstone specimens subjected to different wet-dry cycles, the dynamic compression parameters are obtained, including the loading rate and dynamic compression strength, as shown in Table 3. Figure 5 deeply presents the variation of dynamic compression strength with the loading rate subjected to 0, 5, 10, 15, and 20 wet-dry cycles, respectively.e static compression strength is also included for a better comparison.As can be seen from Figure 5, the dynamic compression strength is affected not only by the loading rate but also by the number of wet-dry cycles.For specimens subjected to same wet-dry cycles, assuming that the dynamic compression strength of red-sandstone has an increasing trend with the increase of the loading rate, a significant loading rate-independent effect was observed.It was also observed that the increasing rate of dynamic strength increased quickly in the first few cycles, then increased slowly in the following cycles.Test results are in accordance with the previous literature by Zhou et al. [5].For specimens under the similar loading rate, with the increase of wet-dry cycles, the dynamic compression strength decreases gradually.However, from the above data alone, it is impossible to conduct the quantitative research on strength reduction induced by wetdry cycles, because loading rates are not the same under equal conditions.In order to solve this problem, exponential equations are adopted to fit variation rules of loading rate and dynamic compression strength.Figure 5 illustrates fitting curves, and the equations are listed as follows:

Dynamical Compression Strength of Red-Sandstone after Different Wet-Dry Cycles. After a range of tests
where σ d is the dynamic compression strength; _ ε and n are the loading rate and number of wet-dry cycles, respectively.
From the fitting results above, a good correlation coefficient is found between fitting curves and scatter points, indicating that the exponential equation can effectively reflect the nonlinear relationship between the loading rate and dynamic compression strength of red-sandstone exposed to different wet-dry cycles.To analyze wet-dry effects on dynamic properties, dynamic compression strength of specimens under the same loading rate (100 Advances in Civil Engineering 5 250 s −1 , and 300 s −1 ) is obtained based on equation (1). Figure 6 further illustrates the relationship between dynamic compression strength and number of wet-dry cycles at di erent loading rates.It can be seen that, regardless of the loading rate, the dynamic compression strength decreases with the increasing wet-dry cycles, verifying the deterioration e ect on dynamic compression strength of redsandstone.In general, cyclic wet-dry treatment has an adverse e ect on both static and dynamic compression strengths of red-sandstone.

Decay Function Model of Dynamic Compression Strength
after Di erent Wet-Dry Cycles.Limited to many nonarti cial factors, such as laboratory conditions, research grants, and test period, laboratory tests cannot simulate site conditions 6 Advances in Civil Engineering completely, especially the weathering time which can last for several years.erefore, the long-term prediction which cannot be tested in laboratory condition for dynamic properties of rock is addressed.To conduct the long-term prediction for mechanical properties and provide a more reliable guidance for the design and construction in rock engineering, several decay models subjected to di erent weathering conditions were established [33][34][35][36].For instance, to investigate the decay e ect of rock against cyclic heating-cooling and freezing-thawing, a decay function model, consisting of decay constant (λ) and number of cycles (N), was proposed by Mutlutürk et al. [33], and the model was con rmed to satisfy the experimental data well.On the basis of this decay model, considering the strain rate, an improved deterioration model was proposed by Wang et al. [37,38] for the prediction of dynamic mechanical degradation of rocks exposed to longterm weathering.According to loading rate and number of wet-dry cycles, the decay function model of red-sandstone against wet-dry cycles is presented as follows: where σ dn is the dynamic compression strength against n wet-dry cycles; σ d0 is the initial dynamic compression strength free from wet-dry treatment; and λ(_ ε) is the decay constant in consideration of the loading rate.
For a more straightforward understanding on the deterioration induced by wet-dry, the half-life (n 1/2 ), revealing that the number of wet-dry is needed to reduce the dynamic compression strength to its half, is also adopted in this paper, as follows: Based on the above test data, the relationship between wet-dry cycles and normalized value of dynamic compression strength (i.e., σ dn /σ d0 ) is illustrated in Figure 7.
With the aid of the regression analysis method, the value of σ dn /σ d0 is the exponential tted, and the decay constant (λ) and half-life value (n 1/2 ) are obtained, as listed in Table 4.As the decay constant and half-life value are associated with the loading rate, the relationship between two parameters and loading rate is illustrated in Figures 8 and 9, and the scatters are also exponential tted, as expressed in equations ( 4) and (5).As can be seen from Figures 8 and 9, with the increase of the loading rate, λ decreases gradually whereas n 1/2 increases accordingly: N 1/2 172.76 − 106.96e −0.0011_ ε , R 2 0.998.
As illustrated above, the decay function model of redsandstone against di erent wet-dry cycles with the loading rate is achieved by substituting equation (4) into equation (2):

Conclusions
In this paper, a series of static and dynamic compression tests after 0, 5, 10, 15, 20, and 25 wet-dry cycles were carried out, and the deterioration mechanism was explained through SEM.Besides, a decay function model taking the loading rate into account was developed for the long-term prediction of dynamic compression strength.e conclusions are as follows: (1) Cyclic wet-dry treatment signi cantly a ects the mechanical and physical properties of red-sandstone.

Advances in Civil Engineering
With the increase of wet-dry cycles, the P-wave velocity, dried density, and static compression all decrease to varying degrees.(2) Dynamic compression strength of red-sandstone is in uenced by loading rate and number of wet-dry cycles.When the loading rate remains at a similar value, the dynamic compression strength decreases with the increase of wet-dry cycles.For specimens subjected to the same wet-dry cycles, the dynamic compression strength increases with the increasing loading rate.(3) e SEM technique is adopted to study the microdeterioration mechanism due to cyclic wetdry.
e results indicate that the deterioration mainly contributes to the weaken connection between grains or cements, and the deterioration degree is also strengthened with the increase of wetdry cycles.(4) Considering the loading rate, a decay function model is proposed for the long-term prediction on dynamic compression strength of red-sandstone subjected to cyclic wet-dry.Two parameters, namely, the decay constant (λ(_ ε)) and half-life value (n 1/2 ), are expressed against the loading rate.It is concluded that the higher the loading rate, the smaller the value of λ, and the more cycles needed to reduce the strength to its half.

Figure 2 :
Figure 2: Dynamic stress equilibrium check in a typical dynamic compression test.

Figure 4 :
Figure 4: SEM images of red-sandstone free from and after different numbers of wet-dry cycles: (a) untreated specimen; (b) specimen after 5 wet-dry cycles; (c) specimen after 10 wet-dry cycles; (d) specimen after 15 wet-dry cycles; (e) specimen after 20 wet-dry cycles.Note that the magnification ratio is 1200.

Figure 6 :
Figure 6: Dynamic compression strength against number of wetdry cycles at di erent loading rates.

Figure 7 :
Figure 7: Normalized UCS against number of wet-dry cycles at di erent loading rates.

Figure 8 :Figure 9 :
Figure 8: Variation of decay constant against loading rate.

Table 1 :
Variation of physical and strength properties of red-sandstone.

Table 2 :
Index properties variation of red-sandstone subjected to di erent wet-dry cycles.

Table 3 :
Dynamic compressive strength of red-sandstone with corresponding loading rate.

Table 4 :
Decay constant and half-life of dynamic UCS of redsandstone after cyclic W-D cycles.