Effect of Temperature on Internal Shear Strength Mechanism of Needle-Punched GCL

: Needle-punched geosynthetic clay liner (NPGCL) has been widely used in landﬁlls. The internal strength of the GCL changes with temperature variation, which affects its application in landﬁlls. A large-scale temperature-controlled direct shear apparatus was developed to study the internal shear strength characteristics of GCL affected by temperature. The internal strength of the GCL was dependent on the bentonite, the ﬁbers, and the interaction between the ﬁbers and the bentonite. The inﬂuence of temperature on the internal strength of the GCL was mainly reﬂected in the displacement at peak strength. However, the peak strength was basically unchanged. The strength of the bentonite and the ﬁbers-reinforced bentonite increased when the temperature increased. The tensile strength of needle-punched ﬁbers decreased with increasing temperature. The peak strength displacement of the ﬁbers-reinforced bentonite decreased with increasing temperature.


Introduction
Geosynthetic clay liner (GCL) has been widely used in landfill liner structures [1,2]. Slip failure along the interface of liner materials is the main form of slope instability [3]. The internal shear strength of GCL and the shear strength characteristics of the interface between GCL and other geosynthetic materials are key factors affecting the safety of GCL-containing liner landfills [4].
GCL internal strength was measured by direct shear, ring shear, and inclined plate [5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20]. The direct shear apparatus with displacement control and stress control has some advantages and disadvantages. Fox and Stark [21] conducted a detailed review of the shear test equipment used to measure the internal and interfacial shear strength of GCL and concluded that direct shear remained the preferred test method. Many scholars conducted direct shear tests on GCL using large-scale apparatus as shown in Table 1. The materials used in the tests include various types of GCL.

NP (W/NW)GCL-Smooth and Textured Geomembrane
Fox [8] 406 mm × 1270 mm 6.9-279 kPa Room temperature Hydration GCL U (W/W) SB (W/W) NP (W/NW) Gilbert et al. [5] 290 mm × 430 mm 3.5-345 kPa Room temperature Hydration Currently, research on the shear strength characteristics of GCL was conducted primarily through indoor shear tests to study the product type, hydration, temperature and hydration time, consolidation time, shear rate, shear displacement, the drainage condition of the normal stress, the process of hydration and shear, hydrated solution, and other specific factors of equipment (such as the specimen clamping system) that influence the shear strength characteristics of the GCL alone or with other geosynthetic materials [4,8,22,28,34,35]. However, a few studies of the strength formation mechanism of needle-punched GCL have been conducted. The needle-punched GCL consisted of powder bentonite between a woven and a nonwoven geotextile that were needle-punched together to provide internal reinforcement. Therefore, the shear strength of needle-punched GCL is dependent on the strength of the bentonite, the tensile strength of needle-punched fibers, and the interaction between fibers and bentonite. The shear strength and initial tangent modulus of unsaturated soil increases with increasing temperature [36]. Shao [37] studied the effect of temperature on the shear strength of clay soil, and found that the shear strength decreased and increased with increasing temperature, respectively, when the moisture content was greater and less than the critical value. Shibasaki and Yamasaki [38] cooled and heated calcium bentonite, and stick-slip characteristics were observed at high temperature, indicating that the shear characteristics changed with temperature. There was no consensus on the effect of temperature on the soil strength. The mechanical properties of polymer Sustainability 2021, 13, 4585 3 of 17 materials were reduced with increasing temperature [30,[39][40][41][42] and increasing temperature may reduce the internal strength of GCL. Studies on fiber-reinforced soil mostly focus on the influence mechanism of the fiber-reinforced soil strength, the influence of freeze-thaw cycles on fiber-reinforced soil, and the influence of the fiber content and other factors [43][44][45][46][47]. Few papers have studied the strength characteristics of fiber-reinforced soil at high temperatures. Increasing temperature may not necessarily lead to the decrease of the shear strength of the GCL.
The shear process of GCL under the maximum normal stress (1000 kPa) and maximum temperature (70 • C) in landfills can be simulated through the large scale temperaturecontrolled direct shear apparatus. According to the structure and manufacture process of the GCL, the strength of GCL was dependent on the three parts. The temperaturecontrolled direct shear tests of the GCL, the bentonite, the fibers and the fibers-reinforced bentonite were conducted to study how the mechanism of temperature affected the shear characteristics of the GCL.

A Large Scale Temperature-Controlled Direct Shear Apparatus
A temperature-controlled direct shear apparatus was modified from the large scale direct shear apparatus that was used to conduct a shear test of geosynthetics at room temperature by Lin [23][24][25] and Shi [48]. The large scale temperature-controlled direct shear apparatus was developed for the following experimental conditions:
Maintain a constant temperature up to 70 • C with an accuracy of ±0.5 • C through water cycling.
The normal stress was applied by hydraulic cylinder which was controlled by the computer. The maximum normal stress could be up to 2.8 MPa. The horizontal shear force was powered by an electric machine which could pull the lower shear box forward at a constant displacement rate. The maximum horizontal force was 150 kN. The shear displacement, the vertical displacement, and the horizontal force of the sample were measured by the displacement transducers and the tension transducer, and recorded by computer automatically. Free-rolling steel balls were placed between the upper shear box and the lower shear box to ensure the rolling resistance was as small as possible. The calibration test showed that the rolling resistance of the apparatus would not exceed 0.1 kN. The inner diameters of the shear boxes were 300 mm.
During the test, water was flowed from the constant-temperature water tank through the circulation pipe to the circulation tank where the sample was placed, as shown in Figure 1. The circulation tank was attached to the lower shear box. The water circulated from the constant-temperature tank to the circulation tank by using water pump, the sample was warmed up gradually, and reached the set temperature finally. Two water level sensors were used to control the water level in the circulation tank around the sample so that the sample was completely immersed. Before testing, the time that the sample took to reach the set temperatures was measured. The temperature of the water was 15.2-18.2 °C. Under the condition of maximum heat transfer between the water and the air, it took 200 min for the sample to reach the temperature of 70 ℃ as shown in Figure 2. The error range was between −0.5 °C and +0.5 °C.

Materials
The needle-punched GCLs were supplied by Shanghai Renzhong Industrial Co., Ltd. Approximately 5 kg/m 2 of sodium bentonite had been sandwiched between a layer of woven geotextile and a layer of non-woven geotextile. The liquid limit and the plastic limit of the bentonite were 236.1% and 34.9%, respectively. The masses of the non-woven geotextile and the woven geotextile were 139 g/m 2 and 121 g/m 2 , respectively. Other physical and mechanical parameters are shown in Table 2. Before testing, the time that the sample took to reach the set temperatures was measured. The temperature of the water was 15.2-18.2 • C. Under the condition of maximum heat transfer between the water and the air, it took 200 min for the sample to reach the temperature of 70 • C as shown in Figure 2. The error range was between −0.5 • C and +0.5 • C. Before testing, the time that the sample took to reach the set temperatures was measured. The temperature of the water was 15.2-18.2 °C. Under the condition of maximum heat transfer between the water and the air, it took 200 min for the sample to reach the temperature of 70 ℃ as shown in Figure 2. The error range was between −0.5 °C and +0.5 °C.

Materials
The needle-punched GCLs were supplied by Shanghai Renzhong Industrial Co., Ltd. Approximately 5 kg/m 2 of sodium bentonite had been sandwiched between a layer of woven geotextile and a layer of non-woven geotextile. The liquid limit and the plastic limit of the bentonite were 236.1% and 34.9%, respectively. The masses of the non-woven geotextile and the woven geotextile were 139 g/m 2 and 121 g/m 2 , respectively. Other physical and mechanical parameters are shown in Table 2.

Materials
The needle-punched GCLs were supplied by Shanghai Renzhong Industrial Co., Ltd. Approximately 5 kg/m 2 of sodium bentonite had been sandwiched between a layer of woven geotextile and a layer of non-woven geotextile. The liquid limit and the plastic limit of the bentonite were 236.1% and 34.9%, respectively. The masses of the non-woven geotextile and the woven geotextile were 139 g/m 2 and 121 g/m 2 , respectively. Other physical and mechanical parameters are shown in Table 2.

Specimen Preparation and Test Procedure
1.
The GCL sample was cut with a diameter of 300 mm; 2.
The non-woven geotextile and woven geotextile of the GCL were bonded to the circular steel plates with special glue called T360 supplied by Dongguan Xianglian Group that provided excellent bonding between polymer and metal. The circular steel plate was perforated so that the water could drain away from the sample. The sample was cured for 24 h until the strength of the special glue reached the maximal strength.
In the first stage of hydration, the GCL sample was put into a water container, and the sample was immersed in water. A pressure of 1 kPa was applied to the GCL sample for 24 h to reduce the uneven expansion. After the first stage of hydration, the GCL sample was placed into the large temperature-controlled direct shear box. The water level was controlled above the upper surface of GCL, so that the GCL was immersed in the water. The strength of bentonite decreased sharply after the first stage of hydration, and a large amount of bentonite would be extruded from the side of the GCL due to the application of the normal stress. Therefore, according to the Rules of Geotechnical Test (SL237-1999) [51], the fast consolidation method was adopted to perform graded consolidation compression in the second step of hydration. The total time of the second stage of hydration was 24 h.

4.
Through the temperature-controlled device, the set temperature circulating water was fed into the direct shear water tank. The shear step could not be conducted until the temperature of the sample reached the set temperature.

5.
The shear displacement rate was set as follows: dry state 0.3 mm/min, hydration state 0.1 mm/min [15,23,49,52,53]. For the dry samples at room temperature, steps 3 and 4 were ignored. The shear step could not be conducted until the vertical displacement stabilized.
Before the test, repeated tests were conducted, so that the reproducibility of the method could be verified. Figure 3 showed that the maximum error of shear stress was 10%, under the normal stress of 200 kPa at room temperature.

Specimen Preparation and Test Procedure
1. The GCL sample was cut with a diameter of 300 mm; 2. The non-woven geotextile and woven geotextile of the GCL were bonded to the circular steel plates with special glue called T360 supplied by Dongguan Xianglian Group that provided excellent bonding between polymer and metal. The circular steel plate was perforated so that the water could drain away from the sample. The sample was cured for 24 h until the strength of the special glue reached the maximal strength. 3. The GCL samples were hydrated following a two-stage hydration procedure [21,23,50]. In the first stage of hydration, the GCL sample was put into a water container, and the sample was immersed in water. A pressure of 1 kPa was applied to the GCL sample for 24 h to reduce the uneven expansion. After the first stage of hydration, the GCL sample was placed into the large temperature-controlled direct shear box. The water level was controlled above the upper surface of GCL, so that the GCL was immersed in the water. The strength of bentonite decreased sharply after the first stage of hydration, and a large amount of bentonite would be extruded from the side of the GCL due to the application of the normal stress. Therefore, according to the Rules of Geotechnical Test (SL237-1999) [51], the fast consolidation method was adopted to perform graded consolidation compression in the second step of hydration. The total time of the second stage of hydration was 24 h. 4. Through the temperature-controlled device, the set temperature circulating water was fed into the direct shear water tank. The shear step could not be conducted until the temperature of the sample reached the set temperature. 5. The shear displacement rate was set as follows: dry state 0.3 mm/min, hydration state 0.1 mm/min [15,23,49,52,53]. For the dry samples at room temperature, steps 3 and 4 were ignored. The shear step could not be conducted until the vertical displacement stabilized.
Before the test, repeated tests were conducted, so that the reproducibility of the method could be verified. Figure 3 showed that the maximum error of shear stress was 10%, under the normal stress of 200 kPa at room temperature.  Figure 4 shows that the shear stress of the GCL decreased sharply after hydration at the same room temperature, which was caused by the strength of the bentonite after hydration [4,26]. The peak shear strength of the GCL decreased between 51.5% and 64.8% after hydration.   Figure 4 shows that the shear stress of the GCL decreased sharply after hydration at the same room temperature, which was caused by the strength of the bentonite after hydration [4,26]. The peak shear strength of the GCL decreased between 51.5% and 64.8% after hydration. Under the same hydration condition, due to the increasing temperature, the peak shear strength of GCL changed slightly, and the peak strength displacement decreased with increasing temperature as shown in Figures 5 and 6. At room temperature (26.3-27.6 °C), the peak strength displacements of dry GCLs ranged from 10 to 12.5 mm. Under the condition of hydration, the peak strength displacement decreased non-linearly with increasing temperature. As the temperature increased from 30 °C to 70 °C, the peak strength displacement decreased nearly 15 mm. Under the same hydration condition, due to the increasing temperature, the peak shear strength of GCL changed slightly, and the peak strength displacement decreased with increasing temperature as shown in Figures 5 and 6. At room temperature (26.3-27.6 • C), the peak strength displacements of dry GCLs ranged from 10 to 12.5 mm. Under the condition of hydration, the peak strength displacement decreased non-linearly with increasing temperature. As the temperature increased from 30 • C to 70 • C, the peak strength displacement decreased nearly 15 mm.

Results of Tests on the GCL
During the test, an obvious stress plateau appeared before the shear stress reached the peak strength under dry condition at room temperature and hydration conditions from room temperature to 40 • C. With increasing temperature, the stress plateau gradually decreased when it reached 50 • C, and disappeared completely when the temperature reached 70 • C.
The initial moisture content of the bentonite in the GCL was relatively stable, and remained within the range of 12-15%. After the first step of hydration, the moisture content of the bentonite in the GCL was about 110%. The moisture content of the bentonite in the GCL after shearing is shown in Figure 7, and was mainly affected by normal stress and temperature. At the same temperature, the moisture content of the bentonite decreased as the normal stress increased. Under the same normal stress, the moisture content after the test decreased as the temperature increased.  During the test, an obvious stress plateau appeared before the shear stress reached the peak strength under dry condition at room temperature and hydration conditions from room temperature to 40 °C. With increasing temperature, the stress plateau gradually decreased when it reached 50 °C, and disappeared completely when the temperature reached 70 °C.
The initial moisture content of the bentonite in the GCL was relatively stable, and remained within the range of 12-15%. After the first step of hydration, the moisture content of the bentonite in the GCL was about 110%. The moisture content of the bentonite in the GCL after shearing is shown in Figure 7, and was mainly affected by normal stress and temperature. At the same temperature, the moisture content of the bentonite decreased as the normal stress increased. Under the same normal stress, the moisture content after the test decreased as the temperature increased.   During the test, an obvious stress plateau appeared before the shear stress reached the peak strength under dry condition at room temperature and hydration conditions from room temperature to 40 °C. With increasing temperature, the stress plateau gradually decreased when it reached 50 °C, and disappeared completely when the temperature reached 70 °C.
The initial moisture content of the bentonite in the GCL was relatively stable, and remained within the range of 12-15%. After the first step of hydration, the moisture content of the bentonite in the GCL was about 110%. The moisture content of the bentonite in the GCL after shearing is shown in Figure 7, and was mainly affected by normal stress and temperature. At the same temperature, the moisture content of the bentonite decreased as the normal stress increased. Under the same normal stress, the moisture content after the test decreased as the temperature increased.  The moisture content of the bentonite in the GCL was taken from three different locations: the center of sample, the edge of sample, and between the center and the edge of the sample. The moisture content at the edge of the sample was the largest, and the least was at the center. The moisture content measured in the test was close to that of other papers, near the fitting curve ω = 4.94 × σ n −0.282 [8].
During the internal shear test of the GCL, the GCL samples were loaded into the shear apparatus after the completion of the first stage of hydration. From the beginning of the second stage of hydration to the end of the shear test, the bentonite in the GCL was extruded from the edge of the sample. After the test, the extruded bentonite was collected and dried. With increasing temperature, the amount of extruded bentonite decreased under the same normal stress as shown in Figure 8. With the increase of normal stress, the amount of bentonite extrusion increased. and shear stage were collected by linear displacement transducers. The results are shown in Figure 8. The sum of the vertical displacements in the compression stage and shear stage showed the same pattern as the rule of bentonite extrusion, so that the vertical displacements were related to the extrusion of the bentonite. The vertical displacement of the GCL was caused not only by compression of the bentonite, but also by extrusion of the bentonite inside the GCL. Bareither [22] conducted direct shear tests of five different kinds of GCLs. The temperature increased from 20 ℃ to 80 ℃, and the internal shear strength of needle-punched GCL was significantly reduced, with a maximum reduction of 40%. In that test, the woven surface and non-woven surface of the GCL were constrained with a pyramid-tooth steel plate. The effectiveness of the pyramid-tooth gripping plates to transfer shear stress into the internal region of a needle-punched GCL was dependent on the peel strength and normal stress. The gripping effectiveness increased as the normal stress increased for a given GCL, or as the GCL peel strength decreased for a given normal stress.
To ensure that the GCL and perforated steel plate were constrained compactly, the woven and non-woven geotextile of GCL was bonded with the circular plate. During the test, none of the fibers were pulled out. The internal fibers were stretched, and the shear failure surface always occurred in the bentonite. It could avoid the damage of fibers pulled out from the woven surface during the shear process. After the first stage of hydration, the second stage of hydration was also called the compression stage. The vertical displacements of the specimens in the compression stage and shear stage were collected by linear displacement transducers. The results are shown in Figure 8. The sum of the vertical displacements in the compression stage and shear stage showed the same pattern as the rule of bentonite extrusion, so that the vertical displacements were related to the extrusion of the bentonite. The vertical displacement of the GCL was caused not only by compression of the bentonite, but also by extrusion of the bentonite inside the GCL.
Bareither [22] conducted direct shear tests of five different kinds of GCLs. The temperature increased from 20 • C to 80 • C, and the internal shear strength of needle-punched GCL was significantly reduced, with a maximum reduction of 40%. In that test, the woven surface and non-woven surface of the GCL were constrained with a pyramid-tooth steel plate. The effectiveness of the pyramid-tooth gripping plates to transfer shear stress into the internal region of a needle-punched GCL was dependent on the peel strength and normal stress. The gripping effectiveness increased as the normal stress increased for a given GCL, or as the GCL peel strength decreased for a given normal stress.
To ensure that the GCL and perforated steel plate were constrained compactly, the woven and non-woven geotextile of GCL was bonded with the circular plate. During the test, none of the fibers were pulled out. The internal fibers were stretched, and the shear failure surface always occurred in the bentonite. It could avoid the damage of fibers pulled out from the woven surface during the shear process.

The Mechanical Behavior of the Components of GCL Internal Strength under the Influence of Temperature
To study the mechanism of the internal strength characteristics of GCL under the influence of temperature. The influence of the bentonite, the needle-punched fibers, and the interaction between the bentonite and fibers on the internal strength characteristics of the GCL was studied.

Shear Strength of the Bentonite
Temperature-controlled direct shear test of bentonite was conducted with the same apparatus in GCL direct shear test. The moisture content of the bentonite removed from the hydrated GCL sample after the first stage hydration was about 110%. The bentonite was removed from the GCL, and bentonite samples were prepared with the moisture content of 110%. The sample was loaded into the temperature-controlled direct shear box, and the rapid consolidation method was adopted to conduct graded compression. The total compression time was 24 h, so that the second stage hydration of GCL was simulated. The shear displacement rates in the dry state and the hydration state were 0.3 mm/min and 0.1 mm/min respectively.
Comparing the results of the dry bentonite at room temperature (16.1-18.0 • C) with the hydrated bentonite at room temperature (14.9-15.6 • C) in Figure 9a, the peak strength decreased approximately 70.5-80.1%. The main reason that the strength of the GCL decreased sharply after hydration was that the strength of the bentonite decreased after hydration.

Shear Strength of the Bentonite
Temperature-controlled direct shear test of bentonite was conducted with the same apparatus in GCL direct shear test. The moisture content of the bentonite removed from the hydrated GCL sample after the first stage hydration was about 110%. The bentonite was removed from the GCL, and bentonite samples were prepared with the moisture content of 110%. The sample was loaded into the temperature-controlled direct shear box, and the rapid consolidation method was adopted to conduct graded compression. The total compression time was 24 h, so that the second stage hydration of GCL was simulated. The shear displacement rates in the dry state and the hydration state were 0.3 mm/min and 0.1 mm/min respectively.
Comparing the results of the dry bentonite at room temperature (16.1-18.0 °C) with the hydrated bentonite at room temperature (14.9-15.6 °C) in Figure 9a, the peak strength decreased approximately 70.5-80.1%. The main reason that the strength of the GCL decreased sharply after hydration was that the strength of the bentonite decreased after hydration.  Under the condition of hydration, the peak strength of the bentonite increased by 20.5-33% when the temperature increased from 15 °C to 70 °C as shown in Figure 9. The friction angle increased from 5.4° to 7.95°. The peak strength of the bentonite was similar to the results obtained from the routine direct shear test at room temperature [24]. With increasing temperature, the moisture content of the bentonite with the same initial moisture content decreased after shearing, while the density of the bentonite increased, as shown in Figure 9. The shear strength and the friction angle of the bentonite increased with increasing temperature. The moisture content was the key factor of the shear strength of bentonite [54]. After the shear step, the water around the sample was cooled and released. The moisture content and density of the bentonite was measured from the shear Under the condition of hydration, the peak strength of the bentonite increased by 20.5-33% when the temperature increased from 15 • C to 70 • C as shown in Figure 9. The friction angle increased from 5.4 • to 7.95 • . The peak strength of the bentonite was similar to the results obtained from the routine direct shear test at room temperature [24]. With increasing temperature, the moisture content of the bentonite with the same initial moisture content decreased after shearing, while the density of the bentonite increased, as shown in Figure 9. The shear strength and the friction angle of the bentonite increased with increasing temperature. The moisture content was the key factor of the shear strength of bentonite [54]. After the shear step, the water around the sample was cooled and released. The moisture content and density of the bentonite was measured from the shear layer. The moisture content of the sample would not change by the evaporation of water in the sample.
The stress plateau appeared in the results of direct shear test on GCL. The peak strength of bentonite accounted for 66.7-93% of the plateau-stress as shown in Table 3, which indicated that the stress plateau phenomenon in the GCL direct shear test was mainly caused by the shear stress of the bentonite in the GCL reaching the peak strength of the bentonite [24], while other internal forces in the GCL had not fully played their roles.

Tensile Strength of Needle-Punched Fibers
In order to simulate the effect of needle-punched fibers inside GCL on the internal strength of GCL, the tensile strength of needle-punched fibers under the influence of temperature was studied. The GCL was cut with the diameter of 300 mm and the bentonite was fully removed from the GCL. The non-woven and woven surfaces of the sample were bonded to the circular steel plate with special glue as same as the GCL sample. To ensure that the woven surface would not be contacted with the non-woven surface, steel balls were placed inside the sample between the woven geotextile and the non-woven geotextile. Normal stress would not be applied in the test. The shear displacement rate was 0.1 mm/min.
With the increasing temperature, the modulus of the internal fiber decreased, and the peak strength displacement of the fibers increased as shown in Figure 10. At room temperature, the tensile properties of the fibers changed slightly under dry and hydration conditions. The tensile strength of the fibers was mainly affected by temperature. The tensile strength of the needle-punched fibers played a key role at the shear displacement of 10 mm at room temperature (22 • C) and 20 mm at 70 • C. Normal stress of the tensile test was zero, and the tensile strength of the fibers only influenced the cohesion of the GCL. Under the same conditions, the sum of the strength of the bentonite and the tensile strength of fibers was less than the internal shear strength of the GCL. Therefore, the interaction between the fibers and bentonite could not be ignored.
conditions. The tensile strength of the fibers was mainly affected by temperature. The tensile strength of the needle-punched fibers played a key role at the shear displacement of 10 mm at room temperature (22 °C) and 20 mm at 70 °C. Normal stress of the tensile test was zero, and the tensile strength of the fibers only influenced the cohesion of the GCL. Under the same conditions, the sum of the strength of the bentonite and the tensile strength of fibers was less than the internal shear strength of the GCL. Therefore, the interaction between the fibers and bentonite could not be ignored.

Shear Strength of Fibers-Reinforced Bentonite
To study the influence of the interaction between needle-punched fibers and bentonite on the internal strength of GCL, the direct shear test of fibers-reinforced bentonite was conducted. The bentonite and fibers were removed from the GCL, and mixed according to the proportion of the GCL. Other steps in the direct shear test of bentonite were followed.
When an equal proportion of fibers in the GCL were added to the bentonite, the strength of the fibers-reinforced bentonite increased with increasing temperature as shown in Figure 11. However, the peak strength displacement decreased when the temperature increased, which was consistent with the pattern of the GCL. The peak strength displacement of bentonite was constant. Interaction between fibers and bentonite reached the peak strength earlier along with an increasing temperature. The important factor that

Shear Strength of Fibers-Reinforced Bentonite
To study the influence of the interaction between needle-punched fibers and bentonite on the internal strength of GCL, the direct shear test of fibers-reinforced bentonite was conducted. The bentonite and fibers were removed from the GCL, and mixed according to the proportion of the GCL. Other steps in the direct shear test of bentonite were followed.
When an equal proportion of fibers in the GCL were added to the bentonite, the strength of the fibers-reinforced bentonite increased with increasing temperature as shown in Figure 11. However, the peak strength displacement decreased when the temperature increased, which was consistent with the pattern of the GCL. The peak strength displacement of bentonite was constant. Interaction between fibers and bentonite reached the peak strength earlier along with an increasing temperature. The important factor that the peak strength displacement decreased in the GCL direct shear test was that the peak strength displacement of the fibers-reinforced bentonite decreased with increasing temperature. Comparing with the results of the bentonite, the shear strength of the fibers-reinforced bentonite was greater than that of bentonite under the same condition as shown in Figure  12. The friction angle increased 0.47 • under the condition of hydration at 70 • C. The interaction between the fibers and bentonite could not be ignored. the peak strength displacement decreased in the GCL direct shear test was that the peak strength displacement of the fibers-reinforced bentonite decreased with increasing temperature. Comparing with the results of the bentonite, the shear strength of the fibersreinforced bentonite was greater than that of bentonite under the same condition as shown in Figure 12. The friction angle increased 0.47° under the condition of hydration at 70 °C. The interaction between the fibers and bentonite could not be ignored. Figure 11. Shear stress-displacement relationships for the fiber-reinforced bentonite.   An obvious stress plateau phenomenon appeared in the shear test of GCL. With increasing temperature, the stress plateau gradually decreased, and disappeared completely when the temperature reached 70 °C. The stress plateau phenomenon was mainly caused by the shear stress of the bentonite reaching the peak strength of the bentonite [22]. The curves of GCL and fibers-reinforced bentonite were compared as shown in Figure 13. Before the shear stress reached the peak strength, there was a period of slow increase of shear stress due to restructuring of fibers and bentonite. The specimen with higher density would be denser at a fixed polymer content, and the grains and polymer additive inside An obvious stress plateau phenomenon appeared in the shear test of GCL. With increasing temperature, the stress plateau gradually decreased, and disappeared completely when the temperature reached 70 • C. The stress plateau phenomenon was mainly caused by the shear stress of the bentonite reaching the peak strength of the bentonite [22]. The curves of GCL and fibers-reinforced bentonite were compared as shown in Figure 13. Before the shear stress reached the peak strength, there was a period of slow increase of shear stress due to restructuring of fibers and bentonite. The specimen with higher density would be denser at a fixed polymer content, and the grains and polymer additive inside this specimen contacted well, thereby providing a path for the external load to go through, thus resulting in the decreasing of the peak strength displacement [55]. Slow increase of shear stress of fibers-reinforced bentonite was another reason for the stress plateau phenomenon in the result of GCL. The temperature increased to 70 • C, and the stress plateau disappeared with slow increase of shear stress. this specimen contacted well, thereby providing a path for the external load to go through, thus resulting in the decreasing of the peak strength displacement [55]. Slow increase of shear stress of fibers-reinforced bentonite was another reason for the stress plateau phenomenon in the result of GCL. The temperature increased to 70 °C, and the stress plateau disappeared with slow increase of shear stress. The results in Table 4 show that the internal strength of the GCL was dependent on three components. Under the influence of temperature, the different shear characteristics of the GCL were caused by different components. Slight variation of the cohesion and shear angle of the GCL did not mean that the components changed slightly. The frictional The results in Table 4 show that the internal strength of the GCL was dependent on three components. Under the influence of temperature, the different shear characteristics of the GCL were caused by different components. Slight variation of the cohesion and shear angle of the GCL did not mean that the components changed slightly. The frictional angle of bentonite increased 2.55 • from room temperature (15 • C) to 70 • C, and the peak strength displacement was constant. The variation trend of the peak strength displacement of the fibers was opposite to that of the fibers-reinforced bentonite. The cohesion and friction angle of fibers-reinforced bentonite increased observably with the increasing temperature. Lin [23][24][25] conducted a shear test of GCL-geomembrane and found that the peak strength increased along with the normal stress non-linearly, which was attributed to the variation of failure modes of the composite liner.

Conclusions
A large scale temperature-controlled direct shear apparatus was developed to study the characteristics of GCL internal shear strength under the influence of temperature. In the test, the internal strength of the GCL was dependent on three parts: the strength of the bentonite, the tensile strength of the needle-punched fibers, and the strength generated by the interaction between the fibers and the bentonite. Shear tests were conducted on the three components to study the mechanism of temperature affected the shear strength of the GCL.

1.
Internal shear strength of GCL decreased sharply after hydration, mainly caused by the strength of bentonite decreasing. The phenomenon of the stress plateau gradually disappeared as the temperature increased. When the temperature increased, the internal peak strength of the GCL changed slightly, and the peak strength displacement gradually decreased. The stress plateau phenomenon in the GCL was mainly due to the shear stress of the bentonite closed to its peak strength, and the slow increase of interaction between fibers and bentonite.

2.
The shear strength of the bentonite increased with increasing temperature. The tensile strength of the fibers was greatly affected by the temperature. The tensile strength of the fibers only influenced the cohesion of the GCL. 3.
The interaction between fibers and the bentonite was not negligible. The shear strength of the fibers-reinforced bentonite increased with increasing temperature, and the peak displacement decreased. Therefore, the interaction between the fibers and bentonite was an important factor leading to the decrease of the peak displacement in the GCL with increasing temperature.

4.
Increasing temperature may not necessarily lead to a decrease in the shear strength of the GCL, which is related to many factors, such as the constraint mode of the sample and the type of the sample. Further studies are needed.