Investigation of shear strength of sand–bentonite mixtures with boron additives at high temperature for energy geo-structures

The soils surrounding energy geo-structures are exposed to high temperatures and temperature cycles. Changes in the engineering properties of soils should be investigated under thermal effects and soils that are highly durable against temperature changes are needed for thermo-active geo-structures. Generally, bentonite or sand–bentonite mixtures (SBMs) are preferred as natural barrier soil materials. Hence, the engineering properties of these natural soil materials against high temperatures should be improved. Boron, which has high thermal resistivity, reduces the heat expansion of materials, when added to soils may increase the durability of buffer materials at high temperatures. In the present study, the effects of tincal and ulexite additives were observed on the shear strength behavior of SBMs at 80 °C and room temperature. The general results showed that with the contribution of boron, the shear strength of the SBMs increased with increasing temperature. The effect was more pronounced for 20% SBMs at high temperature. Tincal and ulexite can be used to increase the shear strength of SBMs at high temperatures.


Introduction
Fossil fuels are expensive and pollute the environment. In addition, with increasing need for energy parallel to increase in population of the world, fossil fuel-based energy resources, such as coal, oil, and gas, are constantly consumed. Running out of energy resources will eventually become a major problem for human societies (UNFCC 2015). Renewable energy sources, on the other hand, are the only solution for use as an alternative to fossil fuels (Tiwari and Mishra 2011). For that reason, the types and numbers of thermo-active geo-structures have both increased in recent 1 3 256 Page 2 of 17 years. Heat storage systems, buried high-voltage cables, oil and gas pipelines, energy piles, CO 2 sequestration plants, geothermal energy plants, nuclear waste disposal repositories, ground improvement using heating and freezing techniques, and sealed underground buildings (subway and tunnels for fire prevention) can be counted among the types of thermo-active structures. Energy geo-structure projects are related to heat transfer that takes place through soil mass. Hence, there is a need to understand soil behavior with thermal phenomena in terms of mechanical and environmental variables. The number of studies on the thermo-mechanical behavior of soils has increased over the past two decades.
The conceptual design of nuclear waste repositories comprises multi-barrier systems. The systems considered for high-level radioactive wastes (HLWs) have three basic materials: host rock, backfill, and bentonite. A multi-barrier system should provide sufficient conditions for the isolation and controlled release of radionuclides (Wang 2010). In HLW projects, it is thought that the canister can release heat for thousands of years and the buffer performance surrounding the canister is therefore important. According to many studies, the buffer should have low hydraulic conductivity, good chemical buffering, and self-healing capacity (Imbert and Villar 2006;Ouhadi et al. 2010). Especially with the effects of earthquakes, rock-shear movements may occur, and another function of the buffer is to protect the canister from rock movements (Sellin and Leupin 2014).
In nuclear waste storage and isolation, it is crucial that buffer materials be mechanically stable and chemically resistant, as well as having low hydraulic permeability (IAEA 2001). Since bentonite and sand-bentonite mixtures (SBMs) meet the above-mentioned requirements, these are used as buffer and filling materials to isolate underground nuclear waste repositories. The reasons for using SBMs to construct an impermeable and resistant buffer material are that the sand component will form a load-carrying skeleton and the bentonite component will provide impermeability by filling the gaps in the sand (Mollins et al. 1996).
Under high temperatures, the strength loss of soils around energy structures can cause damage. The long-term engineering properties of soils at impermeable barriers, such as nuclear power plants, energy piles, and heat storage facilities, should not change, especially in terms of hydraulic conductivity, compressibility, and strength at high temperatures. Former studies showed that temperature changes affect the permeability, compressibility, and shear strength parameters of soils (Pusch et al. 1990;Abuel-Naga et al. 2006). Zheng et al. (2015) investigated the effect of high temperatures (100 °C and 200 °C) on the bentonite and host rock in a radioactive waste repository and reported that illitisation of bentonite and clay soils increased as the temperature increased. Previous studies showed that temperature increase has detrimental effects on the volume deformation of clayey soils.
Inter-particle forces and the viscous resistance of adsorbed water play important roles in the changes of fabric behavior of clayey soils at varying temperatures (Abuel-Naga et al. 2007). The volumetric changes in muddy clay were investigated under heat up to 100 °C by Chen et al. (2016). It was determined that soils tend to decrease their volume at high temperatures. The reasons for this phenomenon are the steam outflow of water, the steam outlet, the displacement of the grains, the fracture of the grains, and the degradation of organic matter.
Temperature increase also has effects on the shear strength of soils. The magnitude of this effect depends on density, water content, soil type, and mineralogical and chemical composition (Mitchell 1969). Drained shear tests were carried out to investigate the effect of temperature on the shear strength of calcareous sand and quartz sand (He et al. 2021). Temperature increase caused densification in both of these sands and the particle breakage of the calcareous sand increased with increasing temperature. In addition, increasing temperature reduced the peak internal friction angle. However, an improvement was observed in the strength of the quartz sand with increasing temperature. It was reported that the different behaviors of the two sands were due to the fact that the particle breakage in the calcareous sand under the effect of temperature was dominated by densification (He et al. 2021). The change in residual deviatoric stress of soft Bangkok clay at high temperatures (70 °C, 90 °C) was observed by Abuel-Naga et al. (2006). In that study, higher peak shear strength values were obtained at high temperatures. The residual deviatoric stress values of samples at different temperatures had similar values at large strains. In addition, soil samples that were subjected to shearing under high temperature had lower volumetric strains. The changes of shear strength parameters in previously heated samples were investigated and the internal friction angle of the kaolin was found to be significantly increased due to the increasing temperature. As a result of heating bentonite up to 100 °C, there was a sharp increase in the internal friction angle and a decrease in cohesion value. Thus, the highly cohesive bentonite became a soil with low cohesion and had high internal friction angle which are specific to granular soils with increasing temperature (Wang et al. 1990). De Bruyn and Thimus (1996) reported that as temperature increased up to 80 °C, the mechanical strength of Boom clay decreased. The strength of the buffer (sand-bentonite mixtures) also decreased when the temperature was increased up to 100 °C (Lingnau et al. 1996). However, this decrement was not much (less than 10%). In addition, the buffer maintained normally consolidated behavior when the temperature was increased up to 100 °C (Lingnau et al. 1996). The structural strength of clayey soil reduced when the temperature was increased to 50 °C and thermal softening behavior was observed in the soil after heating (Gu et al. 2014). In addition, there was an improvement in the strength of the soil after a heating-cooling cycle (20-30-40-50-40-30-20 °C). For that reason, it was concluded that some thermo-mechanical behaviors of the soil were irreversible. Hong et al. (2013) reported that the effects of high temperature on the shear strength of clay were highly dependent on the volume change due to heating. It was also noted that thermal expansion decreased soil strength, while thermal contraction increased the shear strength. In a study on solid waste, strength parameters at increasing temperatures were investigated by triaxial tests. While the variation of the internal friction angle of the solid waste with temperature was negligible, the cohesion decreased linearly with increasing temperatures (20-60 °C) (Shi et al. 2021). The direct shear behavior of granite fractures under normal stress between 0.2 and 10 MPa was investigated at different temperatures from 20 to 800 °C. The peak shear strength decreased linearly with increasing temperature. In addition, as the normal stress increased, the thermal effect became less pronounced (Tang and Zhang 2020).
To increase the thermal durability of SBMs, boron minerals can be used. The most important features that play a role in the selection of boron are heat resistance and low thermal expansion at high temperatures. Boron is one of the most widely used raw materials in industry and it can also absorb neutrons. Boron oxide (B 2 O 3 ) and boric acid (H 2 BO 3 ) are the simplest forms of boron compounds. Colemanite bonds with calcium and ulexite and tincal bond with calcium-sodium and sodium, respectively (Sallı-Bideci 2016). Boron increases the strength and resistance of glasses against thermal shocks (Sugozu et al. 2016). It also significantly reduces the expansion of glass with temperature and protects glass from acid and scratching (Yigitbasoglu 2004). Boron compounds and boron fibers are used for high strength and flexibility in plastics or metals. For example, the addition of boron increases the hardness and strength of steel (Boren 2022). Boron compounds are particularly effective in increasing thermal stability as well as preventing radiation transmission (Guzel et al. 2016). It is a neutron-absorbing material; for that reason, it is preferred in nuclear power plants. Former studies reported that boron is held strongly by the aluminum or silicon tetrahedron portion in the clay structure. Adsorption of boron by clay reached a maximum at pH 9-9.97 (Privett 1987). The highest boron adsorption is done by illite, and the least is done by kaolinite between illite and kaolinite. The surface area of montmorillonite is higher than that of illite and it is accepted that montmorillonite can adsorb much more boron than illite (Keren and Mezuman 1981).
Sand-bentonite mixtures are used especially in impermeable barrier applications. Hence, the engineering properties of buffer soil materials should be improved against thermal effects. In the present study, it was aimed to develop boron-SBMs in soils surrounding energy geo-structures that are durable in terms of shear strength at high temperatures. In this study, mixtures formed by adding 10% and 20% boron (tincal and ulexite) to 10% and 20% SBMs were used. After the addition of boron to the SBMs, shear strength parameters at room temperature (RT) were determined and tests were conducted at 80 °C to observe the shear strength behavior of the SBMs under various temperatures.

Materials
In the present study, the boron minerals tincal and ulexite, which are commonly used in industry, were used. The bentonite was Na-bentonite and supplied by a local company. Tincal and ulexite were obtained from Eti Mining Directorate of Turkey. The grain size distributions of the sand, bentonite, and additives are given in Fig. 1. The sand was well-graded according to the Unified Soil Classification System (ASTM D2487 2017). The physico-chemical properties of the used materials are given in Table 1.
The structure of tincal was observed using scanning electron microscopy (SEM). SEM analyses were performed using a COXEM EM-30 Plus device. Samples were dried in a freeze-dryer for 24 h to prevent shrinkage. After the drying process, SEM analyses of the samples were performed. SEM photographs of the boron minerals are given in Fig. 3. A different structure in the tincal mineral resembling cubic particles was clearly observed at high temperature (Fig. 3b). No significant change was observed in the structure of the ulexite mineral with increasing temperature (Fig. 3c, d).
The materials used this study are indicated in Fig. 4. The Na-bentonite and sand were dried for 24 h in an oven. The sand was used by sieving through No. 6 (3.35 mm) mesh after drying (Fig. 4a). Bentonite was used by sieving through No. 200 (0.075 mm) mesh (Fig. 4b). Tincal was used as supplied without drying (Fig. 4c). Similarly, ulexite was used without drying (Fig. 4d). Furthermore, the tincal was crushed with a jaw crusher and sieved through No. 40 (0.425 mm) for the experiments. Since ulexite was supplied in ground form, it was used without any crushing or sieving processes. The natural water contents of the borons were determined at the beginning of each test and these values were considered during sample preparation.

Mixture preparation
The total dry weights of the samples included 10% and 20% boron and the rest of the dry weight of the mixture was weighed as bentonite and sand. Details of the mixtures are presented in Table 2. Samples are referred in the text with abbreviations where "B" stands for bentonite, "S" for sand, "T" for tincal, and "U" for ulexite. For example, sample 8B-72S-20U contains 8% bentonite, 72% sand, and 20% ulexite.

Methods
The direct shear tests were performed according to ASTM D3080 standard (2018). The samples were prepared for the direct shear tests with w opt + 2% water content and dry unit weight value (γ d ) corresponding to Standard Proctor Test results (ASTM D698 2012). In the direct shear tests, the dry materials were mixed in a container. Water was then added to the mixture (w opt + 2%) and mixed in homogeneously with the help of a shovel. The samples were placed into a mold (6 × 6 cm) in three layers by compacting each layer at its γ d value. The shear box was filled with water and the samples were kept under submerged condition for 24 h. During this period, a pressure of 25 kPa was applied to prevent swelling. The samples were consolidated under 49, 98, and 196 kPa normal stresses and then sheared.
The direct shear tests were performed at RT and 80 °C. In the shear box, the high temperature was supplied using a heat rod (Fig. 5a). The temperature of the water in the cell was continuously controlled and maintained at 80 °C by means of a thermostat (Fig. 5b, c).
When the temperature of the water decreased, the thermostat activated the heat rod, and when the temperature was 80 °C, it disabled the heat rod. Thus, the temperature remained constant throughout the tests. Two K-type thermocouples were used to measure and record the temperature of the water and soil together with a digital thermometer (Fig. 6). The schematic representation of the test set-up is given in Fig. 7. The tests performed for this study are shown in Fig. 8.

Results and discussion
Direct shear tests were conducted on 10% and 20% SBMs with 10% and 20% tincal and ulexite as boron minerals. The direct shear tests were performed both at RT and high temperature of 80 °C. The effects of the boron minerals on the shear strength behavior of the SBMs at high temperature were examined.

High-temperature effect on shear strength parameters of sand-bentonite mixtures
The direct shear tests were conducted on additive-free SBMs at both RT and high temperature. The shear stress-strain graphs of 10% and 20% SBMs at RT and 80 °C are given in Fig. 9. When the bentonite content was increased from 10 to 20%, the maximum shear stress value decreased, while the  Table 3. Although the ϕ′ value of the 10B-90S mixture showed an insignificant change with increasing temperature, the ϕ′ of the 20B-80S mixture increased from 7.1° to 13.5°. In addition, the cohesion (c′) of the 20B-80S mixture decreased from 38.5 to 20.0 kPa when the temperature was increased. Similarly, Wang et al. (1990) reported that the cohesion of air-dried bentonite was ten times higher compared to bentonite heated to 100 °C. It was noted that as a result of heating, bentonite, which has high cohesion, behaved like a granular soil with low cohesion and high internal friction angle (Wang et al. 1990). In another study on the sand-bentonite mixtures showed that buffer strength decreased as a result of triaxial tests up to 100 °C (Lingnau et al. 1994). Similarly, Magsoodi et al. (2020) reported that sand did not affected from temperatures up to 60 °C; however, the cohesion of kaolin increased. The results of the present study are in good agreement with the existing literature so that as temperature increases strength of the sand-bentonite mixtures decreases for 10B-90S; however, increases for 20B-80S which highlights that bentonite content plays an important role on this behavior.
The SEM analyses of additive-free 10B-90S and 20B-80S samples were performed at RT and 80 °C (Fig. 10). When Fig. 10a, b is examined, a change was not observed in the structure of the 10B-90S mixture with increasing temperature. When the temperature was increased, however, it is clearly seen that the pores decreased and a more contracted structure was observed in the 20B-80S mixture (Fig. 10c, d).
Previous studies showed that increase in temperature transform smectite into more constant silicate phases by cementation and illitisation (Wersin et al. 2006). As a result, the water-retaining ability of the smectite reduces. phenomenon can be explained by the viscous shear resistance of water and fabric changes due to an increase in temperature. As temperature increases, the viscosity of water, that of around the bentonite particles, decreases.

High-temperature effect on shear strength behavior of sand-bentonite mixtures with tincal and ulexite additives
Direct shear tests were performed on the 10B-90S and 20B-80S mixtures at RT and 80 °C. For 10B-90S samples, the ϕ′ value of the 9B-81S-10 T mixture did not change significantly with increasing temperature, while as the tincal percentage was increased (8B-72S-20 T), ϕ′ increased from 33.0° to 37.7°. The shear stress-strain graph for 20B-80S samples with tincal at both temperatures is given in Fig. 11. It is clearly seen that the 20B-80S samples completed the elastic deformation stage at a lower shear strain when tincal was added (dashed line in Fig. 11). The maximum shear stress and ϕ′ values of 20B-80S samples increased with tincal addition at both RT and high temperature compared to the additive-free 20B-80S mixture. At RT, the effect of tincal was more significant.
The c′ value decreased with increasing temperature for 10B-90S and 20B-80S samples with tincal. For example, the c′ of the 18B-72S-10 T mixture decreased to one-third when temperature increased. The maximum shear stress value (τ max ) for 9B-81S-10 T decreased from 160.1 to 148.0 kPa with increasing temperature; however, when the tincal percentage was increased to 20% (8B-72S-20 T), τ max increased from 150.0 to 160.3 kPa. It can be said that 20% tincal is more effective than 10% tincal for 10B-90S mixtures at 80 °C in terms of shear strength. It can be concluded that tincal addition has positive effects on the shear strength of SBMs at high temperatures (Alpaydin and Yukselen-Aksoy 2021). Ulexite addition increased the ϕ′ and cohesion values of 10B-90S samples at 80 °C (Alpaydin and Yukselen-Aksoy 2019). Only the c′ values of the 10B-90S mixtures with ulexite did not change significantly with increasing temperature. A significant increase in ϕ′ values was observed; for example, the ϕ′ of 8B-72S-20U increased from 22.4° (RT) to 32.6° (80 °C). When the temperature increased to 80 °C, the angles of internal friction of 20B-80S mixtures nearly doubled; however, there was no significant change in the cohesion values in the presence of ulexite. For mixtures with 20% bentonite with 10% and 20% ulexite, the ϕ′ values of 9.5° and 9.9° at RT increased to 15.9° and 20.0° at 80 °C, respectively.
The shear stress-strain graph of 20B-80S samples with ulexite at RT and 80 °C is shown in Fig. 12. According to Fig. 12, the peak and residual shear stress values of the 18B-72S-10U and 16B-64S-20U mixtures increased at 80 °C. The ulexite-containing SBMs completed the elastic deformation stage at higher τ max values than the other mixtures at 80 °C. The addition of ulexite increased the shear strength at RT, as well. However, as can be clearly seen from Fig. 12, the ulexite additive behaved much more effectively in terms of maximum shear strength at 80 °C when compared to values at RT. In addition, as the ulexite content increased, the shear strength increased. The shear failure envelopes of all samples at RT and 80 °C are given in Fig. 13.
The change of maximum shear stress values (τ max ) of 10% and 20% SBMs with ulexite with increasing temperature is given in Fig. 14. A significant increase in τ max was observed with increasing temperature for the 9B-81S-10U and 8B-72S-20U mixtures. The τ max of 9B-81S-10U increased from 152.5 to 161.1 kPa, which was higher than the τ max of additive-free 10B-90S at both temperatures. Overall, the results show that the τ max values of mixtures with ulexite at 80 °C were higher than the values obtained at RT.   Table 4.
According to the results of the present study, boron additives reduced cohesion values at RT because of the non-cohesive structure of boron. Maghsoodi et al. (2019) reported that as temperature increases, the cohesion of clay increases. However, the results of the present study have shown that as the temperature increased, the cohesion values of additive-free 10B-90S and 20B-80S mixtures, borons, and boron-containing mixtures all decreased.
According to the results of the present study, the shear strength of both SBMs generally increased with tincal and ulexite addition at high temperature. The boron is adsorbed by the aluminum or silicon tetrahedron parts in the bentonite structure (Keren and Mezuman 1981). For that reason, the boron was held by bentonite particles. Eventually, the shear strength of sand-bentonite increased as a result of both boron adsorption and the material replacement, because when the boron was added to the sample, both bentonite and sand contents were decreased. The bentonite material was replaced with boron mineral, which has low thermal expansion (boron: 0.0000083 cm/cm/°C (0 °C)). The thermal expansion coefficient of expansive soils is approximately 10 -4 /°C −1 (Romero et al. 2005). Such low thermal expansion increases soil strength (Hong et al. 2013). As a result of this, shear strength increased with the contribution of boron. The representation of the material replacement is given in Fig. 15.

Influence of boron percentage on the shear strength behavior of sand-bentonite mixtures
The effect of boron percentage on shear strength (τ max ) and shear strength parameters (ϕ′, c′) was analyzed. When the tincal content was increased from 10 to 20%, the ϕ′ value slightly decreased at RT, whereas it slightly increased at 80 °C for 10B-90S mixtures. However, the increase in tincal percentage from 10 to 20% caused a slight decrease in ϕ′ values for 20B-80S mixtures at 80 °C. The increase in tincal percentage had no significant effect on c′ values for almost all mixtures at RT and 80 °C.
When the tincal percentage was increased from 10 to 20% at RT, τ max decreased for 10B-90S and 20B-80S samples. The change in τ max according to tincal percentage in SBMs at 80 °C is shown in Fig. 16. When tincal content increased from 10 to 20%, τ max increased for 10B-90S mixtures, while it decreased for 20B-80S mixtures at 80 °C.
The ϕ′ values of the 10B-90S mixtures decreased when the ulexite content was increased at both temperatures. When the ulexite content was increased from 10 to 20%, there was no significant change in the ϕ′ values of the 20B-80S mixtures at RT. In addition, as the  ulexite content was increased from 10 to 20%, the ϕ′ value increased, too, for 20B-80S mixtures at 80 °C. When the ulexite percentage was increased from 10 to 20%, the c′ value decreased for 10% and 20% SBMs at both temperatures in general.
The maximum shear stress values of the 10B-90S mixtures decreased as the ulexite content increased from 10 to 20% at both temperatures (Fig. 17). Additionally, as ulexite content increased for 20B-80S mixtures at 80 °C, τ max increased.

Influence of boron mineral type on the shear strength behavior of sand-bentonite mixtures
The effect of boron mineral type (i.e., tincal or ulexite) on maximum shear stress (τ max ) and shear strength parameters (ϕ′, c′) was investigated.
The highest c′ values were obtained using ulexite for 10% and 20% SBMs at both temperatures. The c′ of 10B-90S with 20% ulexite was 29.4 kPa, while in the presence of 20% tincal, it was 6.9 kPa at 80 °C. When the results of tests performed at RT were analyzed, the boron mineral that caused the lowest τ max was 10% ulexite. The highest τ max values of the 10B-90S and 20B-80S mixtures were, respectively, obtained as 160.1 kPa and 95.4 kPa using tincal at RT. The tincal content was more effective than ulexite in increasing τ max at RT. When the   Fig. 18. With the contribution of 20% boron, the τ max of the 20B-80S mixtures reached higher values with the contribution of ulexite. When the SEM analyses of the 18B-72S-10T mixture, one of the tincal-containing SBMs, were examined, needlelike structures were seen to form with increasing temperature (Fig. 19b). When SEM analysis of the 8B-72S-20U sample, one of the ulexite-containing SBMs, was conducted at both temperatures, the structure, which had seemed smooth and void-free at RT, could not preserve its structure at high temperature (Fig. 19c, d).

Conclusions
To develop soil mixtures that do not lose their strength at high temperatures, boron may be used because of its superior performance in thermally resistant materials. In this study, the shear strength behavior of 10% and 20% SBMs was investigated with the contribution of boron (i.e., tincal and ulexite) at RT and 80 °C. The following results were obtained: • Although the 10B-90S mixture showed an insignificant change with increasing temperature, the ϕ′ value of the 20B-80S mixture dramatically increased. The c′ value of all SBMs decreased with increasing temperature. In addition, the τ max of 10B-90S decreased slightly and τ max values did not show a significant change for 20B-80S when the temperature was increased to 80 °C. • By adding boron to 10% SBMs at RT, ϕ′ values were generally decreased. In 20% SBMs at both temperatures, ϕ′ values increased with the contribution of both boron additives. The c′ values decreased with tincal at RT and high temperature. • The τ max values decreased for 10B-90S mixtures with boron, although they increased for 20B-80S mixtures at RT in general. At 80 °C, when tincal and ulexite minerals were added to the mixtures, τ max increased. The general results show that with the contribution of boron, the shear strength of the SBMs increased with increasing temperature. It can be concluded that the addition of boron increased the shear strength of SBMs at high temperature. In the field applications, boron can be used to increase the shear strength of SBMs; however, the effect of tincal may be more pronounced when compared to ulexite.