Global and Local Shear Behavior of the Frozen Soil–Concrete Interface: Effects of Temperature, Water Content, Normal Stress, and Shear Rate

Abstract

The interface between soil and concrete in cold climates has a significant effect on the structural integrity of embedded structures, including piles, liners, and others. In this study, a novel temperature control system was employed to conduct direct shear tests on this interface. The test conditions included normal stress (25 to 100 kPa), temperature (ranging from 20 to −6 °C), water content (from 10 to 19%), and shear rates (0.1 to 1.2 mm/min). Simultaneously, the deformation process of the interface was continuously photographed using a modified visual shear box, and the non-uniform deformation mechanism of the interface was analyzed by combining digital image correlation (DIC) technology with the photographic data. The findings revealed that the shear stress–shear displacement curves did not exhibit a discernible peak strength at elevated temperatures, indicating deformation behavior characterized by strain hardening. In the frozen state, however, the deformation softened, and the interfacial ice bonding strength exhibited a positive correlation with decreasing temperature. When the initial water content was 16% and the normal stress was 100 kPa, the peak shear strength increased significantly from 99.9 kPa to 182.9 kPa as the test temperature dropped from 20 °C to −6 °C. Both shear rate and temperature were found to have a marked effect on the peak shear strength, with interface cohesion being the principal factor contributing to this phenomenon. At a shear rate of 0.1 mm/min, the curve showed hardening characteristics, but at other shear rates, the curves exhibited strain-softening behavior, with the softening becoming more pronounced as shear rates increased and temperatures decreased. Due to the refreezing of interfacial ice, the residual shear strength increased in proportion to the reduction in shear rate. On a mesoscopic level, it was evident that the displacement of soil particles near the interface exhibited more pronounced changes. At lower shear rates, the phenomenon of interfacial refreezing became apparent, as evidenced by the periodic changes in interfacial granular displacement at the interface.

1. Introduction

The structure–soil interface contact problem is widespread in various geotechnical engineering projects. In cold climates, due to the thermodynamic changes of water and ice at different temperatures, the specific engineering properties of frozen soil, as well as its interaction with structural elements, undergo significant alterations upon freezing, causing channels, embankments, pipelines, and pile foundations to face serious frost damage, which results in substantial differences in their service life [1,2,3].

In cold climates, when the temperature falls below freezing, the shear resistance at the interface between the structure and the soil changes due to alterations in ice adhesion. The maximum shear resistance is referred to as freezing strength [4]. The dynamic changes in freezing strength under different conditions, along with the interaction between frost-heave force and bearing capacity, can lead to varying degrees of structure stability. For example, the lining of a water conveyance channel undergoes cyclic deformation due to frost heave and freezing forces, eventually causing cracks or fractures, which lead to water leakage. Over time, this results in severe damage to the lining structure [5]. A plethora of research studies have been conducted to examine the resilience of the interface between building materials and frozen soil, as well as the factors that contribute to its strength. Penner et al. [6] conducted a field-based analysis focusing on the freezing resistance between clay and steel pipe interfaces. They provided insights into the effects of temperature on freezing strength and derived an empirical formula for calculating this strength. In contrast, Parameswaran et al. [7,8] explored the freezing resistance at the interface of structural piles with sand or ice through pull-out tests. Their findings indicated a positive correlation between freezing strength and the rate of loading. Furthermore, Ladanyi et al. [9,10,11] proposed that freezing strength is dependent on various factors, including the soil’s physical attributes, temperature, the nature of the interface, the rate of loading, and the type of loading. Two common test methods are employed to evaluate the macroscopic interaction between soil and building materials: the pull-out test [12,13,14] (which assesses the shear behavior between a curved structure and the soil interface) and the direct shear test [15,16,17,18] (which evaluates the shear behavior between a plane structure and the soil interface). Desai et al. [19] employed a direct shear apparatus to investigate various issues related to the interaction between soil and structural materials. Liu et al. [20,21] conducted a series of experiments on frozen soil–concrete interfaces. They classified the shear stress–displacement curve into five distinct stages ranging from elastic deformation to stable residual strength. Lu et al. [22] proposed that the freezing strength of the soil–concrete interface is dependent on a multitude of factors, including normal stress, soil moisture content, concrete surface roughness, interface temperature, and soil physical properties, based on their investigation of the interface freezing-strength mechanism. Sun et al. [23,24] explored how the freezing strength of the interface is affected under varying conditions, discovering that the stress-deformation characteristics of this interface are influenced by factors such as water content, temperature, and normal stress. Zhao et al. [25] conducted the direct shear tests on the artificial frozen silt sand–structure interface, and their findings indicate that the freezing temperature significantly influences both the peak and ultimate interfacial strengths. In a study by Pan et al. [26], direct and cyclic shear tests were performed on cemented sand–structure interfaces under various freeze–thaw cycles. Their results pertained to the shear stress at this interface and also explored the formation and variation mechanisms of shear stress through microscopic analysis. Fang et al. [27,28] performed experiments on the frozen soil–concrete interface and conducted post-shear microstructural analysis. Their objective was to elucidate how concrete surface roughness affects the shear strength between concrete and frozen soil, with a focus on uncovering the causes and nature of stick–slip failure. Wang et al. [29,30] examined soil–concrete interfacial interaction through a combination of freeze–thaw cycles and direct shear tests, exploring the correlation between the stress–displacement curve, shear strength, and interface parameters as the number of freeze–thaw cycles increased.

The macroscopic interaction of the soil–structure interface is influenced by the microscopic deformation characteristics of the shear band at the contact surface. Numerous methods exist to study the microscopic deformation characteristics of the structure–soil interface, including photoelastic technology, X-ray imaging, and CT scanning technology based on the slice method, among others [31,32,33,34,35,36]. The aforementioned methods have high operational requirements and provide relatively limited microscopic research parameters. Particle image velocimetry (PIV) [37] and digital image correlation (DIC) [38] were initially used in fluid mechanic research. Later, White et al. [37] introduced this technology into the study of pile–soil interaction and developed the corresponding Go-PIV analysis software. Research on the microscopic response of the interface has not only yielded significant findings regarding particle movement and evolution mechanisms but also enabled the continuous quantification of the particle-structure parameters, interface shear-band thickness, etc., [39,40,41,42]. The application of PIV technology by Abdi [43] allowed for the examination of soil deformation and particle displacement at the reinforcement–soil interface and near the pile tip. In addition, the influence of factors such as soil particle size, overburden pressure, and bolt height and position on microscopic deformation was thoroughly analyzed. Zhang et al. [44] employed PIV technology to investigate the frost-heave deformation of fine-grained fillers with varying moisture contents under different temperature gradients. Yao [45] conducted a model test on the interaction between flat-ended piles and sand, using PIV to obtain a displacement vector field and contour map of the sand surrounding the pile boots. Sun [46] employed digital image acquisition and processing technology (DIC) to capture the strain field evolution of the BFRP reinforcement during pull-out from a concrete surface, subsequently calculating the bond stress at the interface. Wang [47] utilized the DIC system to examine the effect of graded loading on slope displacement change under various working conditions.

At present, few micro-scale experimental studies have addressed the interface between soil and structure in the cold environments. Pan et al. [48] utilized PIV technology to investigate the deformation behavior of the interface within a direct shear experiment involving silty clay–concrete specimens. Their research contributed to a computational approach for estimating the shear band thickness at the soil–concrete interfaces. However, the analysis of interface deformation and strain localization phenomena remains incomplete. Research on the performance of the frozen soil–concrete interface has received considerable attention. Prior research has indicated that the initial water content, temperature, and surface roughness exert a considerable influence on the interface strength. However, the majority of these studies concentrate on the analysis of macroscopic mechanical properties, with relatively little research conducted on the microscopic properties, the formation of shear bands, and the localization of strains. Accordingly, this paper examines the shear mechanical properties of the frozen soil–concrete interface through temperate-controlled direct shear tests and compares and discusses the shear stress and shear-displacement-curve characteristics, cohesion, and friction angle under different temperatures, normal stresses, water contents, and shear rates. Concurrently, the digital image correlation (DIC) method is employed to conduct a qualitative and quantitative analysis of the interfacial mesoscopic deformation mechanism. The findings can provide scientific parameters for the stability analysis of structures in frozen regions.

2. Materials and Methods

2.1. Sample Preparation

The soil used in this test was collected from Lanzhou City, Gansu Province. Based on the soil classification standard [49], it is classified as silt. The particle size-distribution curve of the soil is illustrated in Figure 1. The basic physical properties of the soil are listed in

Table 1. Basic physical and mechanical properties.

Soil Classification	Maximum Dry Density (g/cm3)	Optimum Water Content (%)	Liquid Limit (%)	Plastic Limit (%)
Silt	1.91	13.3%	28.2	14.3

.

The soil sample preparation process, based on standards [50] (GB/T 50123-1999), involves the following steps: ① The original soil is naturally air-dried, thoroughly mixed, and then sieved through a 2 mm mesh to remove larger particles, and then the initial water content is determined by the drying method; ② To ensure uniform distribution of water content in the soil samples, test soils with different water contents are prepared according to the experimental design. The samples are then placed in sealed bags and left to equilibrate for 24 h.

The concrete sample preparation process includes the following: ① Apply an even layer of petroleum jelly inside a square stainless-steel ring with a side length of 100 mm and a height of 25 mm. ② Mix concrete mortar made of PO 32.5 common Portland cement mixed with natural river sand (gravel with larger particles are removed) in a ratio of cement, sand, and water of 1:3:0.65. ③ Pour the concrete mortar evenly into the steel ring, smooth it so that the thickness of the concrete sample is equal to the height of the steel ring, and cure for 28 days according to standard procedures. ④ Take the concrete sample out of the steel ring, and select the one with uniform surface and good texture as the concrete roughness test sample.

Roughness is a crucial parameter that influences the geotechnical behavior of the interface. After the concrete sample is completed, the concrete surface profile is measured using a CMOS micro laser displacement sensor (with a laser diameter is approximately 50 μm and a repeatability accuracy of around 10 μm) to obtain the parameters of the concrete surface contour. Figure 2a shows a typical contour line of the sample surface. Using the evaluation method of R. Tse and D.M. Cruden [51], the joint roughness coefficient (JRC) can be calculated [52]. The calculation equation is as follows:

$$JRC=32.2+32.47\log Z_2$$

(1)

$$Z_2={\displaystyle \frac{1}{L}}\sqrt{{{\int }_{X=0}^{X=L}\left({\displaystyle \frac{\mathrm{dy}}{\mathrm{dx}}}\right)}^2}={\left[{\displaystyle \frac{1}{M(\Delta X{)}^2}}{\sum }_{i=1}^M{\left(y_{i+1}-y_i\right)}^2\right]}^{{\displaystyle \raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$2$}\right.}}$$

(2)

where L is the projection length of the joint profile line in the x direction, $\mathsf{\Delta }x$ is the sampling interval along the x direction, $\mathsf{\Delta }y=y_{\mathrm{i}+1}-y_{\mathrm{i}}$ is the height difference between two adjacent sampling points on the profile line in the y direction, and M is the total number of sampling intervals.

As shown in Figure 2b, when compared to the Barton standard joint profiles [53], most of the prepared samples fall within the “smooth” range of the profile (JRC = 0~2). Therefore, their surface roughness can be considered similar.

Finally, the concrete sample was placed in the mold, and the soil sample was compacted using the layered compaction method to create a concrete–soil interface sample with a height of 50 mm and a width of 100 mm (Figure 3). Both the concrete and soil samples were 25 mm in height, with the interface aligned with the upper and lower shear-box interfaces. The sample was then wrapped with cling film and placed in a quick freezer for 24 h. After demolding, it was wrapped with cling film and stored in a constant temperature chamber at the target temperature for 24 h in preparation for testing.

2.2. Test Apparatus

To achieve precision control of the test temperature, a dual temperature control system was employed, including a cold room and a constant temperature chamber. The direct shear apparatus was installed in the constant temperature chamber, as shown in Figure 4. The cold room mainly includes temperature control and ventilation systems, and its exterior is constructed from insulation panels with excellent thermal insulation performance. The dimensions are 5 m × 2.8 m × 2.8 m, and the temperature control accuracy is ±0.5 °C. The constant temperature chamber uses a 1~3.5 Hp DC variable frequency condensing unit with an overall size of 2.0 m × 1.2 m × 1.8 m, providing temperature control precision up to ±0.2 °C.

The direct shear test instrument uses a liquid crystal micro-controlled shear instrument (Figure 5), including a vertical load unit, a horizontal load unit, a data acquisition system, a shear box, a displacement meter, etc. The shear box size is 100 mm × 100 mm × 50 mm, and the upper and lower boxes’ heights are both 25 mm. In order to conduct mesoscopic mechanism research, a 100 mm × 20 mm opening is made on the side wall of the shear box for shear-band image photography, as shown in the red area in Figure 5. At the same time, an acrylic sheet is placed at the opening to prevent the soil from being squeezed out during the test.

The mesoscopic deformation characteristics of the interface were investigated using DIC, or digital speckle correlation. DIC is a technique that uses the initial image within a series during the deformation of the specimen as a reference for the derivation of deformation data of the relevant region through correlation calculations [37]. Its basic principle involves meshing the relevant region in the deformed image, matching the seed points, and treating each sub-region as a rigid motion. For each sub-region, correlation calculations based on a correlation function are performed using a predefined search method. The region in the deformed image with the highest mutual correlation coefficient within the sub-region is identified; this represents the post-deformation position of the sub-region, from which the displacement of the sub-region is derived. Comprehensive deformation information for the entire field can be obtained by processing all subareas. With advantages such as full-field measurement, robust anti-interference capabilities, and high measurement accuracy, DIC has been widely applied in geotechnical engineering [54,55,56].

2.3. Test Method

The design of the direct shear test scheme was based on existing test studies and geotechnical test specifications, aiming to investigate the shear characteristics of the interface between construction materials and soil under different conditions [20,21]. The test process is shown in Figure 6. The test conditions include normal stress (σ_n = 25 kPa, 50 kPa, 75 kPa, 100 kPa), soil water content (the optimal water content of the test soil was found to be 13.3%, so the test set the water content at w = 10%, 13%, 16%, 19%), test temperature (T = 20 °C, −2 °C, −4 °C, −6 °C), and shear rate (v = 0.1 mm/min, 0.4 mm/min, 0.8 mm/min, 1.2 mm/min).

DIC mesoscopic images and direct shear data were collected simultaneously, with different acquisition intervals for varying shear rates. Taking a shear rate of 0.8 mm/min as an example, the acquisition interval is 6 s, and the image size obtained by the camera is 5568 pixels × 3712 pixels, with a resolution of 300 dpi. Higher resolution results in more pixels and speckle patterns within the image, leading to more precise analysis outcomes [57]. Since the soil particles are very small and have highly similar appearances, they do not meet the requirements for DIC processing. To address this, after sample preparation, the side of the sample was evenly and alternately sprayed with black paint and white paint to prepare the speckle pattern on the side of the sample.

3. Results

3.1. Shear Stress-Shear Displacement Curve

Figure 7 shows the interface shear stress (τ) versus shear displacement (u) curves under different normal stresses (T = 20 °C and −6 °C, w = 13% and 19%, v = 0.8 mm/min). At T = 20 °C, the τ-u curves at different water contents exhibit strain-hardening behavior without a distinct peak shear strength. The τ value increases with u and eventually stabilizes. As shown in Figure 7c, at T = −6 °C and w = 13%, the τ-u curve presents a clear peak strength and strain-softening behavior. After τ reaches its peak, it decreases significantly with increasing u. This phenomenon occurs because the deformation of the interface surpasses the peak strength displacement. Subsequently, the cemented ice at the interface is subjected to gradual shearing, which is a characteristic of brittle materials. This leads to a reduction in shear stress. Subsequently, as u increases, the interface enters the softening stage, and τ remains basically stable [17,18,20]. In Figure 7d, where w = 19%, the τ-u curve shows an overall softening phenomenon. With high water content, the amount of interfacial cemented ice increases markedly. After τ reaches its peak, the curve drops more sharply. As w increases, the τ-u curve transitions from plastic failure to brittle failure [17,20,58]. From Figure 7c,d, under strain-softening conditions, the shear displacement at peak shear strength is approximately 3.2 mm for all normal stress and water content levels.

Figure 8 shows the τ-u curve of the interface at different temperatures and water contents (σ_n = 50 kPa, v = 0.8 mm/min). As can be seen from Figure 8a, when w = 10%, as the temperature decreases from positive to negative temperature, the interface τ-u curve gradually transitions from hardening to softening, and the change is obvious because the interface cemented ice is lessened when the water content is low. As w gradually increases (Figure 8b–d), the effect of temperature on the interface τ-u curve gradually increases. At high water content, as T decreases, the peak strength gradually increases, and the strain-softening behavior of the interface also gradually increases. When w = 10%, the peak shear strength increases by 29.8 kPa from 20 °C to −6 °C, an increase of 59.6%. When w = 19%, the peak shear strength increases by 70 kPa from 20 °C to −6 °C, an increase of 110%, and the increase is significantly increased. Nevertheless, the impact of T on the residual strength of the interface is not readily apparent, as the residual strength is primarily derived from the friction between soil particles and the concrete surface. Additionally, the contribution of soil particle cohesion and ice cementation is minimal [26,27,29]. It can be seen from Figure 8 that the u corresponding to the peak shear strength is about 3.2 mm. Under negative temperature, the interface shear stress enters a stable stage (residual stage) at a shear displacement of about 6 mm. When w = 10%, the residual stress increases by 10.6 kPa when T decreases from −2 °C to −6 °C, an increase of 21%. When w = 19%, it increases by 12.8 kPa, an increase of 23%. It is evident that the influence of temperature on residual stress is not particularly significant.

Figure 9 shows the interface τ-u curve at different shear rates and temperatures (σ_n = 75 kPa, w = 13%). As observed from Figure 9a, the τ-u curve at a temperature of 20 °C shows hardening characteristics at various shear rates. Although the curve shape remains relatively consistent as the shear rate v increases, the peak stress gradually rises. In Figure 9b, the τ-u curve shows hardening characteristics when v = 0.1 mm/min, but shows strain-softening characteristics at other shear rates. This softening phenomenon becomes more pronounced as v increases. At lower strain rates, the soil skeleton and cemented ice have creep deformation, and brittle failure does not occur easily, but at higher shear rates, the time for particle rearrangement on the interface is shorter, and the shear strength also increases accordingly [59,60,61]. Comparing Figure 9b–d, obviously, the greater the influence of the shear rate on the interface τ-u curve and shear strength, the lower the temperature. Due to the refreezing of interfacial cemented ice, the residual shear strength increases with the decrease in the shear rate. Under negative temperature conditions, it is clear that the shear stress of v = 0.1 mm/min after shear stabilization (residual stage) is greater than that of v = 0.4 mm/min. This is because when v = 0.1 mm/min, the complete shear process of the sample takes 2 h, and the shear process is completely carried out in a constant temperature room. Part of the cemented ice that is constantly broken at the interface will refreeze during the shear process. This phenomenon will be more obvious at the lower temperature, resulting in the shear stress of v = 0.1 mm/min being greater than that of v = 0.4 mm/min after the shear stabilizes.

3.2. Curve Peak Shear Strength

Figure 10 shows the peak shear strength (v = 0.8 mm/min) at virous σ_n, T, and w. It can be observed that the peak interface shear strength rises in conjunction with an increase in normal stress for a range of soils with varying water contents. As the normal stress rises, the contact areas between the soil and concrete also rise during the shear process, increasing the friction of the interface. The particles inside the interface soil also rearrange with the upsurge of normal stress, which increases the interface shear resistance [62,63]. As T declines, the peak shear strength of the interface demonstrates an upward trajectory. The presence of a considerable amount of free water within the soil structure allows for the gradual transformation of the majority of this water into cemented ice as the temperature decreases. This process contributes to an enhancement in the peak shear strength of the interface [28,29]. The overall change amplitude is the weakest when w = 10% (Figure 10a). Given the low free water content of the interface and the concomitant reduction in interface cementation, it is evident that the interfacial peak shear strength will be adversely affected. As the water content increases, the amplitude of change in the interfacial peak shear strength also becomes more pronounced. When w is 10%, 13%, 16%, and 19%, the peak shear strength at T = −6 °C and σ_n = 100 kPa increases by 258%, 237%, 246%, and 325%, respectively, compared with the peak shear strength at T = 20 °C and σ_n = 25 kPa.

Figure 11 shows the peak shear strength (w = 13%) under different σ_n, T, and v. The peak shear strength gradually increases as the shear rate increases. When the shear rate is low, there is a certain interface creep deformation between concrete and soil. This deformation is not sufficient to cause brittle failure, resulting in a low peak shear strength. Conversely, at higher shear rates, the time available for rearrangement of soil particles at the interface is reduced, which results in increased shear strength.

An increase in the shear rate speeds up the interaction time between the soil and concrete, resulting in a slight increase in shear strength at 20 °C. However, at lower temperatures, the effect of the shear rate (v) on peak shear strength becomes more pronounced. As ice is a time-dependent material, the interfacial cemented ice content is greater at the lower temperature, and it is more likely that the interfacial cemented ice will fail brittlely at the faster shear rate, thereby increasing the shear strength. When the normal stress is σ_n = 25 kPa and σ_n = 100 kPa, the peak shear strength at T = −6 °C and v = 0.1 mm/min is 248% and 101% higher than the peak shear strength at T = 20 °C and v = 1.2 mm/min, respectively.

3.3. Shear Strength Parameters

The interface strength parameters can be obtained by fitting the Mohr–Coulomb strength theory (Equation (3)) [17] as follows:

$$\mathsf{\tau }=\mathsf{\sigma }\tan \mathsf{\phi }+\mathrm{c}$$

(3)

where $\mathsf{\tau }$ is the interfacial peak shear strength, $\mathsf{\sigma }$ is the normal stress, $\mathsf{\phi }$ is the interfacial internal friction angle, and c is the interfacial cohesion.

By fitting the data, the peak shear strength parameter fitting correlation R² under different conditions is greater than 0.974, and the residual shear strength parameter fitting correlation R² is greater than 0.943, indicating the rationality of data fitting. Figure 12 illustrates the variation in interfacial peak cohesion and friction angle under diverse water contents and temperature conditions. As displayed in Figure 12a, as the temperature decreases, the interfacial peak cohesion exhibits a gradual upward trend, particularly pronounced at higher water content levels. As depicted in Figure 12b, upon a reduction in temperature, the interfacial peak friction angle exhibits a slight upward trend. However, the impact of varying water contents is not notably significant, and a distinct variation pattern remains elusive. According to existing studies, as the interface temperature decreases, free water turns into ice, and the interfacial cohesion provided by ice cementation is significantly enhanced, but the effect on the interfacial friction angle under freezing conditions is small [22,64]. As illustrated in Figure 12c, at 20 °C, an increase in water content is accompanied by an initial rise in interfacial peak cohesion, followed by a subsequent decline. This is because at high water contents, the interfacial free water will play a lubricating role, resulting in a decrease in cohesion [65]. However, under negative temperature conditions, due to the contribution of interfacial cementation, the interfacial peak cohesion increases with the growth of water content. As evidenced in Figure 12d, the interfacial friction angle exhibits a slight decrease with a rise in water content at positive temperatures, but the change pattern is not obvious at negative temperatures. The friction angle under negative temperature conditions is slightly greater than that under normal temperature conditions.

As illustrated in Figure 13a,b, with a reduction in temperature, the cohesion at the concrete–soil interface results in an overall increase in the shear rate. The interface friction angle shows no obvious change as the temperature decreases. As shown in Figure 13c,d, as the shear rate increases, the cohesion increases at all temperatures, and as the temperature drops off, the magnitude of the increase becomes progressively larger due to the rate-dependent properties of ice. The friction angle shows little influence from the shear rate. At positive temperatures, it is slightly decreased with the rising shear rate. That is because the faster shear rate makes less time for internal particle rearrangement, resulting in a decrease in the friction angle. While, at negative temperatures, it is slightly increased with the rising shear rate, that is due to the fact that ice cementation is a low- or no-friction material [66].

The residual cohesion and friction angle under other conditions are shown in

Table 2. Residual shear strength index of the frozen soil–concrete interface under different water contents and temperatures.

Water Content (%)	Temperature (°C)	Cohesion of Residual Strength (kPa)	Internal Friction Angle of Residual Strength (°)
10%	−2	11.05	41.5
	−4	9.1	44.2
	−6	25.45	38.1
13%	−2	22.95	36.4
	−4	26.35	40.4
	−6	40.65	26.2
16%	−2	28.9	35.1
	−4	32.25	35.9
	−6	42.3	31.1
19%	−2	16.6	37.3
	−4	18.6	40.4
	−6	17.3	45.1

and

Table 3. Residual shear strength index of the frozen soil–concrete interface under different shear rates and temperatures.

Shear Rates (mm/min)	Temperature (°C)	Cohesion of Residual Strength (kPa)	Internal Friction Angle of Residual Strength (°)
0.4	−2	24.15	31.7
	−4	13.15	40.28
	−6	17.45	41.65
0.8	−2	22.95	36.4
	−4	26.35	40.4
	−6	40.65	26.2
1.2	−2	28.5	38.4
	−4	33.3	38.9
	−6	43.55	35.7

. A noteworthy observation is the substantial increase in residual cohesion, whereas no discernible pattern emerges in the residual friction angle. In the negative-temperature environment, the cemented ice undergoes brittle failure, so the residual cohesion in the residual stage is mainly the residual friction between the soil particles and the interface and the refreezing of the interface. However, the effect of refreezing on the residual strength, cohesion, and friction angle of the interface is very small and can be ignored.

3.4. Mesoscopic Deformation Characteristics of Shear Bands

In this test, scattered spots were sprayed on the sample surface using paint, which can better identify the analysis area in the DIC analysis. The distance between the sample and the camera was about 25 cm, and a cold LED light was used to supplement the light on the sample surface, keeping the light intensity of the light source stable during all of the tests. In the process of sample shearing, a Nikon digital camera was used to take pictures of the deformation of the soil surface in the process of shearing, and the image size of the captured area was 5568 pixels × 3712 pixels, with an image resolution of 300 dpi. In order to avoid the influence of the boundary effect on the mechanical properties of the interface, a certain gap was left outside the DIC analysis area. The shear rate of the specimen was 0.8 mm/min, and the images were collected every 6 s until the end of the test. The DIC analysis was performed using the analysis software developed by Xintuo 3D Technology Co. Ltd. (Shenzhen, China) (www.xtop3d.cn), which has the same general principles as other DIC software or PIV software but has been optimized in terms of visualization and parameter settings, and its capability of fixing a group of images at a fixed shooting position is automatically calibrated in the software. Only the analysis area and the tracking area range need to be set during the analysis process.

To investigate the mesoscopic movement patterns of soil particles adjacent to the interface within the analysis zone during the shearing process, a selection of 15 tracking points (labelled 1 to 15) and five tracking sections (A–A’, B–B’, C–C’, D–D’, and E–E’) have been made, as illustrated in Figure 14. This approach aims to examine the soil particle kinematics during shearing and the formation of shear bands.

3.4.1. Non-Uniform Deformation of the Interface

Figure 15 shows the variation of the horizontal displacement X (δ_h) and vertical displacement Y (δ_v) of the tracking points during the interface shearing process (at v = 0.8 mm/min, T = −4 °C; σ_n = 25 kPa; w = 16%). As can be seen from Figure 15a–c, δ_h increases rapidly in the initial stage and reaches a peak value at a shear displacement of about 3.2 mm. Then, δ_h changes suddenly and enters the shear residual stage, and the point displacement near the interface changes more suddenly. At this time, the interfacial cemented ice undergoes brittle failure, and the shear displacement corresponds to the peak shear-strength displacement of the macro test data. In addition, the closer to the interface, the larger δ_h is, which is due to the interlocking of the interface between concrete and soil during the shearing process and the formation of the interface shear band formed by the soil particles frozen by cemented ice. The δ_h of points 1, 6, and 11 in the figure is greater than that of other tracking points, and the δ_h of points 5, 10, and 15 is the lowest, indicating that the soil is gradually compacted in the shear direction.

As shown in Figure 15d–f, with the elevation of shear displacement, the δ_v change trends of the points (AA’, BB’, CC’, DD’, EE’) on the same section are similar, and all change in the negative direction of the Y axis, and the soil as a whole reaches shear dilation during shearing. The vertical displacement of each tracking point reaches a peak at about u = 4 mm, the δ_v change slows down at 3.2 mm, and the soil’s anti-deformation ability weakens and gradually reaches a stable state. During the shearing process, the δ_v of the tracking points on the left (such as 1, 6, and 11) is greater than that of the tracking points on the right (such as 5, 10, and 15), and the left interface soil is continuously compacted. During the shearing process, the normal stress is transmitted from top to bottom, and the δ_v of the tracking points close to the interface (such as 11–15) is lower than that of other positions (such as 1–10), indicating that a shear band is formed in the soil at the interface.

3.4.2. Effect of Temperature

Figure 16 shows the variation of horizontal displacement X (δ_h) and vertical displacement Y (δ_v) during the interface shearing process (at 20 °C and −6 °C under the conditions of v = 0.8 mm/min, σ_n = 100 kPa, and w = 13%). It can be observed from Figure 16a,b that at 20 °C, with the increase in shear displacement, the δ_h gradually increases, and the final value of δ_h at all monitoring points is between 0.77 mm and 1.03 mm. At −6 °C, the variation of δ_h at each monitoring point significantly differs from that in the melting state. There is a significant mutation of δ_h near u = 3 mm because cemented ice forms inside the soil at the interface during freezing, and the soil is frozen as a whole. The frozen soil exhibits a high degree of stiffness. Following the attainment of the peak value, the interface undergoes a brittle failure, resulting in the soil rebounding (as shown by the dotted line in Figure 16b). This phenomenon gives rise to a mutation in the displacement of the interface soil. Subsequently, δ_h gradually increases, and the final value is between 0.09 mm and 0.28 mm.

As shown in Figure 16c,d, the interface is dilatant at 20 °C, and the interface strain localization is significant. The tracking points (11~15) close to the interface enter the stable state earlier than the tracking points far from the interface. For example, points 11~15 enter the stable state at about u = 4 mm, and tracking points 1~5 enter the stable state at about u = 8 mm. At −6 °C, the δ_v of the tracking points can be clearly divided into two parts: in the first part, after the shear deformation begins, δ_v gradually becomes dilatant, and a mutation occurs near u = 3 mm (corresponding to the mutation point of horizontal displacement, and as shown by the dotted line in Figure 16d). δ_v decreases, and some points experience shear contraction. In the second part, after the δ_v mutation, δ_v gradually increases again (dilatant) as u increases, finally reaching a stable value at about u = 7 mm. As the contact surface transitions into the residual shear stage, where the shear stress undergoes a steady variation, the alterations in δ_v tend to gradually diminish [67,68,69]. The final δ_v of the tracking points close to the interface is close to 0, and the δ_v of the tracking points far from the interface gradually increases. The soil of the interface has uneven deformation, the displacement on the left side of the interface is more than that on the right side, and the soil is gradually compacted in the shear direction.

3.4.3. Effect of Water Content

Figure 17 shows the variation of horizontal displacement X (δ_h) and vertical displacement Y (δ_v) during shearing under different water contents (at T = −4 °C, σ_n = 50 kPa). It is noticeable from Figure 17a,b that the growth in water content has a significant effect on the change before the δ_h mutation. When the water content is 10%, the δ_h mutation value is about 0.19 mm; when the water content is 19%, the δ_h mutation value is about 0.23 mm, and the change rate of δ_h is faster at 19%.

Figure 17c,d is the δ_v variation curve. The δ_v has the maximum value of 0.6 mm at w = 10%, while the maximum value at 19% is 0.35 mm. It is evident from the figure that the increase in water content makes the shear dilation phenomenon less obvious. This is because with the rise of water content, the free water content inside the soil increases; it condenses into ice at −4 °C, the cemented ice content increases, and the deformation capacity inside the soil is weakened. The cemented ice breaks, and there are gaps between the interface soil particles, ice, and water, which makes the soil show shear dilation characteristics.

3.4.4. Effect of Shear Rate

Figure 18 contains curves showing the changes in horizontal displacement X (δ_h) and vertical displacement Y (δ_v) of the tracking point under different shear rates (at T = −4 °C, σ_n = 50 kPa, w = 13%). From the changes in δ_h in Figure 18a–c, it can be observed that as the shear rate grows, the degree of dispersion, as indicated by the value of δ_h, gradually decreases. Furthermore, the dispersion of δ_h at each monitoring point in the interface tends to aggregate. This phenomenon is caused by the combined effect of normal stress and a slower shear rate. The slower shear rate and the applied normal stress enhance the interaction between soil particles inside the soil, resulting in the dispersion of each δ_h tracking point. The displacement fluctuation phenomenon at a rate of 0.1 mm/min is no longer obvious at a 0.4 mm/min rate. After the interfacial cemented ice of the frozen soil is destroyed, the slow shear rate and the effect of normal stress cause the interface to repeatedly refreeze, resulting in continuous displacement fluctuations (as shown by the dotted line in Figure 18a).

From the changes in δ_v in Figure 18d–f, it is apparent that the increase in the shear rate gradually makes the interface shear dilation phenomenon weaken. For example, the maximum δ_v is 0.8 mm at a rate of 0.1 mm/min, and the maximum δ_v is 0.2 mm at a rate of 1.2 mm/min. It can be observed that an elevated shear rate impedes the rolling of soil particles under identical temperatures and normal stress conditions, and the shear phenomenon is less pronounced. At a higher shear rate, the time required for the rearrangement of soil particles is shorter, resulting in a minor alteration in the overall δ_v of the soil. When v = 0.1 mm/min, similar to δ_h, δ_v also shows periodic fluctuations. It is also due to the slow shear rate and the effect of normal stress that the interface repeatedly refreezes (as shown by the dotted line in Figure 18d).

4. Conclusions

The frozen soil–concrete interface is widely present in the engineering of pile foundations, roadbeds, linings, and panels in cold regions, and the deformation behavior and strength of the interface is one of the key factors affecting the structural load transfer. Under the changing conditions of temperature, moisture content, loading rate, etc., the shear behavior and deformation mechanism of the interface become complex and variable, which troubles the engineering design. Therefore, this paper employs a dual temperature-controllable apparatus to conduct direct shear tests on the interface between frozen soil and concrete under diverse hydro-thermal conditions with varying shear rates. It elucidates the deformation law of interface shear strength under different conditions and utilizes the digital image correlation (DIC) method to investigate the microscopic deformation mechanism of the interface. The principal findings are as follows:

• It is evident that there are discernible distinctions in the characteristics of the shear stress–shear displacement curve at positive temperatures (20 °C) and negative temperatures (−2 °C, −4 °C, −6 °C). The absence of an evident stress peak at 20 °C suggests that the material displays stress-strain-hardening characteristics. At negative temperatures, the adhesive ice enhances interfacial cohesion, thereby causing the shear stress–shear displacement curve to exhibit strain softening. This phenomenon becomes increasingly pronounced as the temperature decreases.
• In the frozen state, there is a notable variation in the interface shear strength due to the distance ice-water phase transition after the temperature drops below freezing. At a temperature of −6 °C, the water content rises from 10% to 19% at 100 kPa, accompanied by a corresponding rise in the peak shear strength, which increases from 139.2 kPa to 194.3 kPa. A decrease in temperature from 20 °C to −6 °C was observed at 16% water content and 100 kPa, accompanied by an increase in shear strength from 99.9 kPa to 182.9 kPa. It was observed that the residual shear strength was significantly influenced by the normal stress. When the water content was 19% and the temperature was −6 °C, the normal stress increased from 25 kPa to 100 kPa, and the residual shear strength increased from 44.5 kPa to 107.6 kPa.
• The rate-dependent and temperature-dependent properties of ice result in a significant influence of the shear rate and temperature on the peak shear strength, with the primary factor being the effect on the interface cohesion. The curve demonstrates hardening characteristics when v = 0.1 mm/min; however, it exhibits strain-softening characteristics at other shear rates. As v increases and T decreases, the softening phenomenon becomes more pronounced. In a negative temperature environment, a minor proportion of the cemented ice that is continually disrupted at the interface will refreeze during the shearing process. The lower the temperature, the more pronounced this phenomenon will be, resulting in greater shear stress at a rate of 0.1 mm/min than at a rate of 0.4 mm/min after the shear has stabilized. At a higher shear rate, the time available for particles at the interface to rearrange is shorter, allowing for the gradual increase in interface shear strength to be explained.
• The influence of the stress and deformation transfer process on the interfacial soil is evidenced by the significant strain-localization characteristics observed in shear, as evidenced by the DIC analyses conducted at a temperature of −4 °C, under a normal stress of 25 kPa and a water content of 16%. The closer the point of observation is to the interface, the greater the X-axis displacement. The peak value is reached at a shear displacement of 3.2 mm, at which point the interfacial cemented ice undergoes brittle failure. This is accompanied by a sudden change in the X-axis displacement, which then enters the shear residual stage. This indicates that the DIC analysis corresponds to the macro test data. The soil as a whole undergoes shear compaction during shearing. Prior to the shear displacement of 4 mm, the soil experiences gradual shear compaction as the shear progresses, reaching a peak at 4 mm. At this point, the soil’s ability to resist deformation is diminished, and it reaches a stable state.