Polymer has gained increasing attention for soil reinforcement (i.e., improving soil poor performances such as cohesion and ductility) in recent years [1
] because of its excellent characteristics in terms of efficient performance, better sustainability, providing a clean and healthy environment, and ease of operation [4
]. The mechanism of polymer reinforcement soil is reported whereby the polymer is strongly absorbed on the surface of soil grains via physio-chemical force, and subsequently enwraps grains and interlinks them together by their molecules uncoiling [7
], which in turn enhances soil stability. Many researchers have studied the behaviors of polymer treated soil in the laboratory and concluded that polymer treatment can effectively enhance soil mechanical properties in terms of compressive strength, shear strength, and tensile strength as well as ductility [9
], also improving soil hydraulic properties in terms of water hold ability, anti-erodibility, and permeability [13
]. Moreover, some polymers have been successfully used in different geotechnical engineering applications, such as: stabilization treatment of clay slope topsoil [15
], enhancing the strength of weak soil [6
], injection for pavement slab stabilization [16
], and pavement subgrade remediation [17
]. However, further examination shows that previous studies mostly investigate the effects of polymer on the mechanical and hydraulic performances of soil. Furthermore, there are few studies on the behavior of desiccation cracking of polymer treated soil [18
Cracking is a common natural phenomenon during soil drying that results from water loss by evaporation. In general, the formation and increase of cracks alter soil performance, including mechanical and hydraulic properties, which consequently leads to damages for buildings and geotechnical engineering [19
]. Indeed, a soil mass was split into detached clods by the cracks, and this in turn weakened the integrity and strength of the soil [21
]. On the other hand, hydraulic properties of soil are significantly influenced by these cracks, considering that the flow rate and velocity are controlled by the size and tortuosity of cracks, while the distribution and the connectivity of these cracks determine the flow pathway [23
]. Rayhani et al [24
] indicated that cracks could lead to an increase on the hydraulic conductivity of clay soil by 12–34 times. In field projects, the cracks observed on the surface of a clayey soil slope could provide a channel for rainwater entering into a deeper level of the slope, and consequently decrease the strength of the soil, which may cause damage to the slope [25
]. Also, the cracks observed in the landfill clay liners could significantly accelerate water infiltration into clay liners, and consequently weaken the efficiency of clay liners [27
]. Owing to these vital effects of desiccation cracks and the wide application of polymer in different geotechnical projects, focusing on the cracking behavior of polymer treated soils makes significant sense for the safe and effective running of related projects.
Considering the effects of soil desiccation cracks, the parameter acquisition and quantitative analysis of cracks are of great significance to the study of soil cracking [29
]. Zein el Adedine and Robinson [31
] used the number of intercepts between a transect and crack to estimate the length of cracks. Miller et al. [32
] introduced the crack intensity factor (CIF) as a descriptor of the extent of surface cracking, which was defined as the ratio of cracks area to the total surface area of a drying soil mass. However, the original crack pattern is often disturbed by human activities and equipment, which results in large measurement errors. With the continuous improvement of computer computing ability and the continuous development of image processing technology and software, it has gradually become a research hotspot to use computer to process and calculate fracture images and obtain relevant crack parameters [32
]. However, because the soil is a mixture with complex structure, its cracking is affected by many factors. After mixing with polymer, the structure of soil is changed, and the cracking of soil is also affected. The existing research is insufficient, so it is of great significance and value to study the cracking of polymer mixed soil by using fine image processing software.
Given these considerations, this study investigated the desiccation cracking behavior of two different commonly used polymers (polyurethane and polyacrylamide) admixed with clayey soils. The effects of polymer concentration were also considered. A series of desiccation tests were conducted, and the water evaporation, crack initiation, and propagation processes were recorded during drying. With the aid of an image-processing technique, the geometrical and morphological characteristics of a crack pattern were analyzed quantitatively. In addition, the microstructure changes were also studied using scanning electron microscopy.
2. Materials and Methods
A clayey soil derived from Nanjing area of eastern China, widely distributed in middle and lower reaches of Yangtze River is used in present study. Its physical parameters were measured according to the national criterion for geotechnical tests in China (i.e., GB/T 50123–1999), and the results are presented in Table 1
Polyurethane and polyacrylamide are used in this study because of their wide application as soil conditioners, enhancing aggregates stability, declining soil erosion, and improving soil mechanical properties [35
]. The polyurethane used herein is prepared by the polymerization of poly-oxypropylene diol (PPG, Jining Hongming Chemical Reagent Co., Ltd., Jining, China), poly-oxyethylene glycol (PEG, Shanghai Ika Biotechnology Co., Ltd., Shanghai, China), and toluene diisocyanate (TDI, Nantong Runfeng Petrochemical Co., Ltd., Nantong, China), and exhibited as a light-yellow transparent emulsion (Figure 1
a). The PU could dissolve in water, emulsify, and disperse, and finally form a stable system (Figure 1
a). It has a viscosity of 650−800 MPa∙s, specific gravity of 1.15, solid content higher than 88%, and pH of 7, and contains a variety of functional groups such as –OH, −NCO. Additionally, thermogravimetric analysis was carried out with a heating rate of 10 °C/min, a heating range of 0–800°C, and the results are shown in Figure 1
An anionic polyacrylamide provided by Henan Liansheng Environmental Protection Technology Co., Ltd., Gongyi, China is used. This PAM is polymerized by high purity acrylamide monomer, pure water, initiator, and corresponding auxiliaries in a suitable pH, temperature, and nitrogen environment. After that, the colloidal substance is granulated, dried, and crushed, and the solid PAM was obtained. This white power has a high-molecular-weight of 8–12 × 106
Da and 20% of hydrolysis, also its solid content is higher than 88%. Note that the PAM solution was obtained by dissolving in tap water rather than deionized water, to enhance the dissolution of PAM [38
]. The used PAM and its dissolved solution are shown in Figure 1
b. Thermogravimetric analysis for PAM was also carried out with a heating rate of 10 °C/min, a heating range of 0–800 °C, and the results are given in Figure 1
2.2. Sample Preparation and Test method
The craw clayey soil was air-dried, crushed, and sieved at 2 mm in the laboratory. For the reference sample, the initial saturated slurry was prepared by thoroughly mixing dry soil powder and water. The water content of slurry is 60%, which is much higher than its liquid limit. For admixed samples, the polymer was dissolved in water first, followed by adding into dry soil and thoroughly stirring. The desired mixture was then transferred into a plastic container with an inner dimension of 10 × 10 × 5 cm (length × width × height). Afterwards, the mixture was further homogenized with the aid of a mechanical device lasting for 3–5 min. The container was carefully vibrated to remove the entrapped air bubbles in the mixture and the surface of slurry was flattened. The final mixture was then conserved at room condition for 2 days for curing. The final thickness of the slurry was about 12 mm. The prepared sample was left to evaporate at constant room temperature about 20 ± 2 °C and the relative humidity of 35% ± 3%. During evaporation process, the variation of sample weight was recorded by an electronic balance with an accuracy of 0.01 g, and the initiation and propagation of cracks was photographed by a digital camera for further analyses.
The polymer concentration was set as 0.25%, 0.5%, 0.75%, and 1% for PU, and 0.025%, 0.05%, 0.075%, and 0.1% for PAM, according to their consistency and viscosity characteristics as well as reinforcement efficiency.
2.3. Image Processing and Quantitative Analysis
The technique of digital image processing is usually applied to obtain quantitative and accurate data for further analyses of cracking behavior [23
]. In this study, the digital image-processing technique developed by Nanjing University and introduced by Tang et al. [23
] and Liu et al. [40
] was used. The procedure is presented in Figure 2
and could be described as: transferring original color image to grayscale → image binarization → image de-noising → crack skeletonizing. Based on the difference of gray level in the images and cluster analysis method, the experimental soil cracking images are processed to form binary images of cracks, as shown in Figure 2
c. The black area represents the cracks, and the white area represents the soil clods. As shown in Figure 2
d, closed operations are used to repair individual fracture gaps and remove the fine spots. For the crack area, the nodes of the crack can be determined according to the number of surrounding pixels. By tracing and identifying the adjacent nodes, the median axis of the crack can be obtained, which is also regarded as the skeleton of the crack. The determination of crack intersection, length, and number are based on the skeleton network formed by the skeleton of the crack. Note that, only the central part of 9 × 9 cm is used (in Figure 2
a) in order to eliminate container boundary effects [41
The following quantitative parameters obtained from these processed images were determined and calculated [23
2.4. Microstructure Observation
The microstructure of the reference and admixed samples is investigated using SEM images for further understanding the soil evaporation process and cracking behavior. After the completion of evaporation, the selected sample was prepared as a cube with size of 0.5 cm × 0.5 cm × 2 cm, which was selected from the volume center of the sample. Subsequently, the cube was subject to dehydration and spraying gold treatments in turn. Then, the SEM observation was performed.
Desiccation cracking is a complicated natural phenomenon for soil, mainly induced by water loss, that is, evaporation is a precondition for desiccation cracking. In the natural state, soil is a kind of porous medium with water in its pores. For clayey soil, due to isomorphic replacement, secondary mineral dissociation or selective adsorption, the surface of clay particles generally has negative charges, which are compensated by cations adsorbed on the crystal layer surface. Under the hydration of cations, the surface of clay particles is surrounded by a layer of bound hydration membrane. Due to the existence of hydration membrane, the clay particles do not directly contact with each other, but have spacing, which provides a lot of space for the shrinkage of cohesive soil. In the process of evaporation, evaporation always starts from the surface of soil. The first loss is the free water between soil particles. With continuous evaporation, capillary water will be produced in the soil, which makes the water in the lower soil transfer to the upper layer to maintain evaporation, and the water content of the soil decreases continuously. The distance between soil particles also gradually decreases, which leads to the formation of tensile stress field between soil particles. Due to the heterogeneity of soil, the complexity of composition, structural differences, and environmental factors, the distribution of tensile stress field is uneven, which makes it easy to cause stress concentration in the weak position of soil particle connection, and then form cracks. The factors affected the soil evaporation process could be concluded as external factors in terms of air flow rate, relative humidity and temperature, and inner characteristics of soil in terms of soil composition, pore structure, size distribution of soil grains and soil layer thickness [49
In the current study, the external condition was consistent for all samples with different states. The SEM observations show that the presence of additives led to changes on pore structure for a given soil. In general, the PU and PAM dissolve in water could adsorb a large amount of water and swell, finally exhibit as hydrogel [52
]. Their presence at the inter-aggregates voids really reduced the pore space to a certain extent. The schematic diagram of soil evaporation is displayed in Figure 13
. As shown, the presence of polymer additives influenced the formation of an effective water migration channel, and thus slowed down the evaporation rate [41
]. As shown in the Figure 3
, Figure 4
and Figure 5
, the presence of PU and PAM significantly slowed down the evaporation rate and prolonged the duration of the constant evaporation stage. Moreover, in comparison with natural soil (Figure 13
a), the presence of PU and PAM hydrogels enhanced the water hold capacity of admixed samples, which might be another reason for the decline of evaporation rate of admixed samples. In contrast, for the admixed samples with a higher concentration, it was prone to form cracks at the upper layer because the blocking effect caused by additives prevented water migration from the bottom to upper, and the cracks subsequently became another evaporation surface to accelerate the water evaporation (Figure 13
b). This might be the key reason why the presence of PU and PAM accelerated these samples entering into the residual evaporation stage (sees Figure 3
and Figure 4
The factor that affected the development of tensile stress field between soil particles would affect the behavior of desiccation cracking [19
]. In general, the PU and PAM addition both enhanced the tensile strength between soil aggregates via its absorption and bonding effects, which in turns could delay and prevent the increase of surface crack. However, the bonding effect of polymer usually worked well when it was in “dry” state, that is, when the additives transformed from a “rubbery” state to a “glassy” state by water evaporation [54
]. In this study, the initial water content was 60%, which is far beyond the liquid limit of the experimental clayey soil. A large amount of water in the sample makes the polymer in a “rubbery” state for a long time, and cannot be transformed into “dry” state in a short time. Therefore, the polymer additives existed in the form of hydrogel at initial period would result in that the roles of polymer were not fully gave into play; this might be the main reason why the polymer had no significant effect on cracking behavior. Nevertheless, PAM reduced the cracking and crushing degree of soil to a certain extent. The addition of PAM reduces the number of cracks by 1%–4% and increases the average area of clods by 50 mm2
to 250 mm2
, which could be seen in Figure 8
. However, the changes of cracking behavior had no good regularity with PAM concentration, which could be explained by the fact that the distribution density and distribution location (on the inter-aggregates voids or on the surface of aggregates) was influenced by the higher consistency and viscosity of PAM, and this in turn affected the effectiveness of polymer addition. However, for PU admixed samples, it was observed that the crack intensity factor was slightly higher than that of natural soil, and it increased from 19.12% to 20.53% with different concentrations. It also increases the number of cracks by 1–2%, and the average area of clods by 10 mm2
to 75 mm2
, which can be seen in Figure 7
. This might be resulting from the fact that, in addition to the water absorbed by clay mineral, the water absorbed by PU additives also provided space for soil shrinkage during drying [43
]. This might be the main reason for the width of crack increase. Thus, necessary attention should be paid to the desiccation cracking of polymer treated soil when applied in the field.
A series of desiccation tests were conducted on polyurethane and polyacrylamide admixed clayey soils to investigate the effect of polymer addition on desiccation cracking behavior. The water evaporation, crack initiation, and propagation processes were recorded and analyzed. The microstructure changes were also studied using SEM images. According to the results obtained, the following conclusions could be drawn:
(1) The soil evaporation process could be affected by polymer. PU and PAM addition both decreased the evaporation rate of initial stage might by storing water in voids and filling inter-aggregates voids, about 12%, and the effects were related to polymer dosage. However, the polymer addition with higher dosage accelerated water evaporation and shortened the duration of the water loss process by forming cracks due to the blocking effects caused by additives.
(2) Final cracks morphology analyses of polymer admixed soils show that PAM addition slightly reduced the cracking and crushing degree and kept the soil relatively intact, also reducing the number of cracks by 1–4% and increasing the average area of clods by 50 mm2 to 250 mm2. However, PU addition slightly enhanced the cracking and crushing degree of soil, and increases the number of cracks by 1–2%, and the average area of clods by 10 mm2 to 75 mm2. In addition, PU and PAM addition both could increase the width and length of cracks.
(3) The polymer created a great bonding between soil aggregates and enhanced soil resistance to cracking. However, it was observed that polymer had no significant effect on cracking behavior. This might because the existence of the hydrogel form led to the bonding effect to be insufficient. Additionally, the water absorbed by polymers also provided space for soil shrinkage during drying, and thus slightly enhanced the crack intensity factor and increased the crack width.
In this study, only the evaporation and cracking processes of soil under the condition of single initial water content were considered. Meanwhile, the polymer concentration was relatively low. In the follow-up study, the case of multiple initial water content and high polymer concentration will be fully considered. In addition, the difference of probability density function of different variables will be also the focus of follow-up research.