Using Cement and Calcium Lignosulfonate to Improve the Mechanical Properties and Microstructure of Loess in a Seasonal Freezing Zone

Abstract

The cement composite calcium lignosulfonate is used to enhance the mechanical properties and the freeze–thaw resistance of loess. Based on an unconfined compressive test under different freeze–thaw cycles, the influence of cement dosage, curing age, and freeze–thaw cycles on compressive strength are discussed. The results indicate that the strength of loess can increase by up to 13 times, and the loss of strength is reduced from 72% to 28% under the reinforcement of cement dosage and curing age. The loss of strength is mainly concentrated in the initial 5 freeze–thaw cycles, and the structure gradually stabilizes after 10 freeze–thaw cycles. In addition, according to the X-ray diffraction test, it is found that the stabilized loess exhibits a comparatively more stable mineral composition. The scanning electron microscope results reveal that hydration products enveloped the soil particles, forming a mesh structure that strengthens the connection between the soil particles. The freeze–thaw damage makes the small and medium pores turn into large pores in loess, while the stabilized loess changes micro and small pores into small and medium pores, with no large pores found. It is feasible to improve loess with the cement composite calcium lignosulfonate, which can provide references for the reinforcement treatment of loess.

1. Introduction

Loess is widespread in the provinces of Gansu, Shanxi, and Qinghai and in other north-western regions of China. Engineering safety is thus affected by some of its characteristics, particularly loose porous structure, poor water stability, strong collapsibility, and weak adhesion [1]. At the same time, the properties of loess are further deteriorated under repeated freeze–thaw cycles, which can trigger a series of threatening incidents such as ground subsidence, fissures, and landslides in a seasonal frozen soil region [2]. Therefore, in order to improve the engineering performance of loess, it is essential to select appropriate curing materials for incorporation into the loess.

Freezing initiates the transformation of mobile water within a soil structure into ice, leading to soil expansion. Conversely, during the thawing process, the water transitions from solid ice to a liquid state, softening the soil [3]. The process induces a deterioration in strength, cohesion, internal friction angle, and microstructure characteristics, primarily occurring within the first five cycles; then, the structure stabilizes and tends to obtain a new balance [4,5].

Traditional reinforcing materials include lime, fly ash, and cement. Lime or fly ash can enhance the mechanical strength of loess [6,7,8]. However, lime is prone to easy cracking and displays unstable performance; fly ash shows poor improvement in the water stability of loess and requires the alkaline environment created by cement to maximize its effectiveness. Cement is widely used in comparison to lime and fly ash, owing to its enhanced mechanical properties, reduced environmental impact, and superior technical efficiency [9,10,11]. Guthrie et al. [12] found that soil treated by a sufficient amount of cement effectively resisted frost heave effectively through a series of strength tests and frost heave tests. Zhang et al. [13] conducted freeze–thaw durability tests and unconfined compressive strength tests on fly ash or cement stabilized loess. It was found that the cement stabilized loess maintained nearly zero mass loss during the freeze–thaw cycles, demonstrating superior performance in terms of strength enhancement and freeze–thaw resistance compared to the fly ash stabilized loess. Cheng et al. [14] analyzed the reinforcement mechanism of cement-improved loess using an indoor compressive strength test and proposed an optimal cement content based on a field test. Yuan et al. [15] found that the addition of cement can alter the physical properties, enhance both compressive and shear strength, and change the size and shape of pores by generating C-S-H gels and ettringite.

Various new materials, including ionic curing agents [16], polymer materials [17], and nanomaterials [18], have also yielded improved results in the reinforcement of loess. However there are a series of challenges such as increased costs, uncertainties in field performance, and limitations in application. Calcium lignosulfonate (CL) has strong dispersibility, cohesion, and chelation properties [19]. With advancements in research, scholars have discovered that CL yields promising results in soil reinforcement treatments. Li et al. [20] compared the solidification effect of CL, sodium hydroxide, and sodium silicate stabilizers on loess. Their results indicated that the addition of CL led to reductions in mineral crystal spacing and particle double layer thickness, consequently increasing compactness and enhancing compressive strength, thereby optimizing the improvement effect. Ji et al. [21] carried out mechanical property tests and microstructure analyses on clay modified by CL. The results showed that as the CL content increased, the unconfined compressive strength initially increased and then decreased, reaching peak values at a CL content of 0.5%. Yang et al. [22] also found advantages, such as improving the soil compressive strength and controlling the mass loss during freeze–thaw cycles in lignin-stabilized soils. CL can promote the hydration reaction of cement, which contributes to the overall durability and performance of the structure. Zou et al. [23] investigated the impact of CL on cement strength through a uniaxial compression test. The results revealed that the addition of CL increased the maximum destructive stress of the specimen by 54.9%, and the failure mode shifted from being predominantly tensile- to shear-dominated. The presence of CL can facilitate the growth of ettringite crystals by providing more Ca²⁺, contributing to the strength of the cement [24]. CL has the advantages of being non-toxic, having a rich reserve, having a low cost, having strong reproducibility, and having stable chemical properties [25].

In summary, the modified effect of cement on loess is superior to other traditional materials in terms of strength and frost resistance, and CL can enhance the performance of cement. At present, there is a lack of relevant research on the use of cement composite CL to improve the mechanical properties of loess under freeze–thaw conditions. In view of this, this study employs this method to stabilize the loess surrounding Tunnel No.1 in Longnan City, Gansu Province, a seasonal freeze area. Combined with laboratory mechanical testing tests and microscopic tests, the enhancement mechanism was analyzed to provide a theoretical foundation for engineering applications.

2. Materials and Methods

2.1. Materials

The test soil was taken from the Guangou No. 1 Tunnel Wudu District, Longnan City, Gansu Province, and was pale yellow in color with a small amount of snail crushed shells. According to the test specification [26], the basic physical properties of loess were tested and are shown in

Table 1. Physical properties of loess.

Natural Moisture Content /%	Natural Dry Density /g·cm3	Liquid Limit WL /%	Plastic Limit WP /%	Liquidity Index IL	Plasticity Index IP	Optimum Moisture Content /%	Maximum Dry Density /g·cm3	Particle Size Distribution/%
								<5 μm	5 μm–75 μm	75 μm–2 mm
9.89	1.45	28.93	13.63	−0.24	15.3	14.63	1.84	15.05	81.43	3.52

.

The coefficient of collapsibility for the loess under various vertical pressures is shown in Figure 1. With the gradual increase in pressure, the coefficient of collapsibility shows a trend of initially rising and then falling, and it reaches the maximum coefficient of 0.081 at 300 kPa. Notably, the initial collapse pressure is 70.1 kPa, indicating a strong collapsibility of the loess.

The test cement employed P.O 42.5 ordinary Portland cement.

Table 2. Index of cement.

Initial Setting Time/min	Final Setting Time/min	Compressive Strength/MPa		Break Off Strength/MPa
197	285	3 d	28 d	3 d	28d
		27.1	42.5	5.7	6.5

displays the physical and mechanical indexes.

Figure 2 shows calcium lignosulfonate (CL), chemical formula C₂₀H₂₄CaO₁₀S₂, with a purity of more than 98% and a small amount of impurities. It is a brownish-yellow powdery solid that dissolves readily in water. This anionic surfactant contains hydrophobic groups and functions effectively as a stabilizer [27].

2.2. Specimen Preparation

The collected loess was naturally air-dried, crushed, and subsequently oven-dried at 105 °C. It was then crushed once more and passed through a 2 mm sieve. The addition of cement changed the maximum dry density (MDD) and optimum moisture content (OMC). The MDD and OMC for each improved loess were determined by heavy compaction testing. The specimen was prepared according to the optimum water content and a compaction degree of 95%. The cement contents were 4%, 6%, 8%, and 10%. According to the study by Ji et al. [21], the dosage of CL was determined to be 0.5%. This dosage represents an engineering recommendation to ensure certain performance and economic efficiency. The curing ages were 7 d, 14 d, 21 d, and 28 d.

Firstly, mixing of the loess and cement was ensured to achieve uniformity. The pre-measured CL and distilled water were combined in a beaker and stirred continuously to form an admixture solution. This solution was poured into a spray pot and uniformly applied onto the soil sample. Subsequently, the samples were placed in the mixer and agitated for 3–5 min. The specimens were compacted into a cylindrical shape with a 39.1 mm diameter and an 80 mm height using the bumping method. The samples were placed in a constant temperature and humidity curing box for the curing process. After reaching the curing age, they were submerged in water for 48 h. This procedure was performed to simulate extreme freeze–thaw conditions, and then, the specimens were employed for the unconfined compressive strength test and freeze–thaw cycle test.

2.3. Test Method

2.3.1. Compact Test

Table 3. Parameters of the compactor.

Hammer Bottom Diameter	Hammer Quality /kg	Fall Height /mm	Layer	Number of Hits per Layer	Compaction Cylinder Size		Sample Size
					Diameter /mm	Height /mm	Height /mm	Volume /cm3
51	4.5	457	3	94	152	116	102	2103.9

and Figure 3 show the heavy-duty electric compactor and the instrument parameters. Samples with different water contents were compacted for both the loess and improved loess. Subsequently, the water content and dry density of each sample were remeasured to determine the MDD and OMC of the samples.

2.3.2. Freeze–Thaw Cycle Test

Resistance to freeze–thaw cycles serves as an important indicator for assessing the effectiveness of soil improvement. According to the “Microstructure response to shear strength deterioration in loess after freeze-thaw cycles” [5], the number of freeze–thaw cycles was set to 0, 1, 5, 10, and 15 cycles. Figure 4 shows the meteorological data of Longnan city in recent years. The lowest air temperature recorded was approximately −20 °C. Given that the surface temperature typically exceeds the air temperature, the freezing temperature used in the experiment was set at −18 °C. The samples were wrapped in plastic film to prevent water loss. The specimen was then frozen at −18 °C for 12 h and subsequently thawed in water at around 20 °C for another 12 h, thus completing one freeze–thaw cycle.

2.3.3. Unconfined Compressive Strength Test

The test design is shown in

Table 4. Test design.

Specimen	Cement Content (%)	CL Dosage	Water Content (%)	Freeze–Thaw Cycles	Curing Age (d)
S0	0	0	14.6	0, 1, 5, 10, 15	7, 14, 21, 28
S4	4	0.5	16.5	0, 1, 5, 10, 15	7, 14, 21, 28
S6	6	0.5	17.6	0, 1, 5, 10, 15	7, 14, 21, 28
S8	8	0.5	18.4	0, 1, 5, 10, 15	7, 14, 21, 28
S10	10	0.5	19.1	0, 1, 5, 10, 15	7, 14, 21, 28

. The test was carried out using a self-designed fully automated triaxial apparatus, as shown in Figure 5. According to the test specification [26], a confined pressure of 0 kPa and a shear rate of 0.8 mm/min were set, and the test was ended when the axial strain reached 8%. Due to a strong water sensitivity and the weak reinforcement effects in loess, tests were promptly conducted immediately after specimen preparation without curing. The value is regarded as constant to facilitate comparative analyses.

2.3.4. XRD Test

The specimen was analyzed with a Rigaku Ultima IV instrument, as shown in Figure 6. After 28 days of curing, 50 g of samples S0 and S10 were collected and crushed into powder for the XRD test. This step aimed to qualitatively analyze the changes in the internal composition of the soil samples.

2.3.5. SEM Test

Figure 7 shows the ZEISS-GeminiSEM500 (Carl Zeiss AG, Oberkochen, Germany) scanning electron microscope apparatus used for the test. Samples S0 and S10 were analyzed at 28d by scanning electron microscopy after undergoing 0 and 15 freeze–thaw cycles. The specimens were dried using a vacuum dryer; cut to the dimensions of 1 cm in length, 1 cm in width, and 0.5 cm in height; and then sprayed with gold for the SEM test.

3. Results and Discussion

3.1. MDD and OMC

It can be seen from Figure 8 that with the increase in cement dosing, the MDD decreased from 1.85 g/cm³ to 1.76 g/cm³, while the OMC increased from 14.6% to 19.1%. Because the cement needs to absorb more water, and additional water is held in the flocculent structure between the cement and loess [28], the effect of cement dosing on the MDD and the OPC cannot be ignored when preparing the samples.

3.2. Unconfined Compressive Strength Test

3.2.1. Curing Age

The unconfined compressive strength (UCS) of the stabilized loess at various dosages and curing ages is depicted in Figure 9. The UCS of specimen S0 is merely 148.5 kPa. The strength significantly increases with the addition of cement and CL, accompanied by an extended curing age.

The strength of the S4 specimen increases to 353 kPa, 455 kPa, 726.5 kPa, and 870.7 kPa at 7 d, 14 d, 21 d, and 28 d, respectively, representing increments of 138%, 206%, 389%, and 486%, respectively. The S6 specimen exhibits increments of 239%, 347%, 548%, and 714% at different curing ages. The S8 specimen demonstrates remarkable increments of 516%, 638%, 911%, and 1129% in strength. Similarly, the increments in strength of the S10 specimen are 817%, 943%, 1169%, and 1303%, respectively. It can be seen that the UCSs of the specimens are relatively close at 7 days and 14 days. However, at 21 days, there is a significant increase in UCS, particularly notable in the S8 and S10 specimens. After curing for 28 days, the rate of increase in strength slows and the UCS continues to rise until stabilization.

This is because during the initial phase of cement generation, the colloidal particles are primarily encapsulated on the surface of the soil particles. At this stage, the generation of a hydration reaction product is less and has not been fully stable. After 21 days of curing, these hydration products precipitate and accumulate in the voids between the particles with the increase in curing time, leading to the formation of new and more stable hydrated crystals [29]. Crystals depend on the variety of gravitational attraction to form a stable structure, thereby enhancing the macroscopic strength.

3.2.2. Cement Dosage

The relationship between UCS and cement dosage is plotted in Figure 10. After curing for 28 d, the UCS values for the S4 and S6 specimens were 870.7 kPa and 1209.5 kPa, respectively. When the cement dosage increased to 8% and 10%, the strength could reach 915.5 kPa and 1362.2 kPa after 7 days of curing, surpassing that of the S4 and S6 specimens at 28 days of curing. This indicates that the cement content contributes more to UCS than the curing age. This can be attributed to the limited formation of stable hydration products in specimens with low cement contents, failing to adequately bond with loess. Consequently, this leads to the creation of localized weak areas within the sample, thereby impacting the overall compressive strength. Increasing the cement content accelerates the reaction with the mineral components of loess, facilitating the formation of stable substances that effectively fill the voids within the soil matrix. Hence, it is advisable to select a cement content of over 8% and a curing period exceeding 21 days to ensure engineering quality.

3.2.3. Freeze–Thaw Cycles

Samples S0, S4, and S10 were analyzed to investigate the damaging effects of the freeze–thaw cycles. As illustrated in Figure 11, the strength of the loess specimen decreased from 148.5 kPa to 41.9 kPa after 15 freeze–thaw cycles, with the final strength loss reaching approximately 72%. This signifies poor frost resistance and instability, failing to meet the necessary engineering requirements.

Even after enduring 15 freeze–thaw cycles, the stabilized loess specimens maintain high strength. After curing for 7 days, specimens S4 and S10 retained strengths of 165.7 kPa and 720.6 kPa after 15 cycles. By 28 days of curing, the UCS reached 452.6 kPa and 1491.7 kPa, respectively, surpassing those of the loess.

Figure 12 indicates the strength loss of the soil at each freeze–thaw stage. For the S4 specimen with low cement dosages, as shown in Figure 12a, the greatest strength decline, approximately 19%, occurred following the first freeze–thaw cycle. The frost damage was lightly suppressed by prolonging the curing age. This is primarily attributed to the limited number of hydration products and hardened crystals formed within the specimen, which still exhibit loess-like properties. Strength loss was concentrated in the first five freeze–thaw cycles, with a notable decrease in the rate of strength degradation. Similarly, Li et al. [30] found drastic damage to the UCS in the initial three freeze–thaw cycles. This rate gradually slowed down after 5–10 cycles. After 10 cycles, the average rate of strength loss per freeze–thaw cycle reduced to below 2%. The final residual strength remained within the range of 47% to 54%.

The augmentation of cement content significantly enhances the resistance to freeze–thaw cycles. Ma et al. [31] also obtained the same conclusion. As illustrated in Figure 12b, the rate of strength loss for sample S10 notably diminished compared to S4 sample. The loss in strength after the first freeze–thaw cycle is within 10%, and the cumulative loss in strength after five cycles amounts to about 20%. With the prolongation of the curing age, the rate of strength degradation begins to decrease gradually, improving the freeze–thaw resistance of the samples. This preserves the residual strength of the soil matrix within the range of 55% to 72%.

3.3. Stress–Strain Curves

The freeze–thaw cycles change the structure of the soil and its stress–strain curve. The stress–strain curve typically begins with a smooth “transition zone” during the initial loading phase, representing the compaction process, followed by the linear elastic stage, and culminates in the failure stage after reaching the peak stress. In Figure 13a, for specimens S4-7d, a considerable amount of internal porosity persisted, indicating that they undergo a long compacting process, which ends at A ($\mathsf{\epsilon }=1.8\%$). The freeze–thaw cycles prolong the process, which enters into the linear elastic phase when at B ($\mathsf{\epsilon }=2.7\%$). And, the stiffness of the soil body increases with the increase in curing age and cement dosing. The stress–strain curve will go through a short compacting process, for example, as shown in Figure 13b, the process for sample S4-28d before freezing and thawing ends at A ($\mathsf{\epsilon }=0.9\%$), and at B ($\mathsf{\epsilon }=1.7\%$). In particular, the transition zone of sample S10-28d ends before the axial strain $\mathsf{\epsilon }=0.8\%$. The shortening of the compacting process indicates that the soil body becomes consolidated and increasingly resistant to the effects of freezing and thawing.

3.4. Modulus of Deformation

The modulus of deformation is a crucial parameter for assessing the deformability of geotechnical materials, with E₅₀ representing the slope of a straight line at half of the peak stress [32]. The calculation formula is as follows:

$$E_{50}=\frac{{\sigma }_{50\%}}{{\epsilon }_f}$$

(1)

where σ_50% is the 50% peak stress and ε_f is the axial strain corresponding to the 50% peak stress.

Figure 14 illustrates the correlation between the elastic modulus and the number of freeze–thaw cycles. The elastic modulus of the loess specimen is less than 1 MPa. As the cement content increases and the curing duration extends, the elastic modulus increases continuously. For example, the modulus of elasticity of the S4-7d specimen is 5.9 MPa, and that of the S10-28d specimen is 53.2 MPa, which is about nine times higher. The elastic modulus experiences a significant drop after the initial freeze–thaw cycle, followed by a gradual decline with an increase in the number of freeze–thaw cycles. Similar results were obtained by Xie et al. [33]. The elastic modulus of the S10-28d specimen exhibited a slight increment; this phenomenon may be due to its more stable structure, which is less susceptible to the effects of freezing and thawing. These variations are within an acceptable margin of error.

3.5. Damage Pattern Analysis

Figure 15 summarizes the typical damage forms of specimens in the test. For the S0 specimen, it shows twisted damage, which is due to the small friction of the specimen, and the particles are easily misaligned with each other, so it shows as twisted and accompanied due to bulging damage in the form of a bulging belly in the absence of lateral constraints. For the soil specimen S4-7d without freezing and thawing, the cohesive force is improved compared to the loess specimen, and the rigidity is increased, but the internal composition still has some medium and large pores. In the process of vertical compression, the specimen is slightly compacted, followed by swelling and bulging damage. After 15 freeze–thaw cycles, the structure of specimen S4-7d becomes looser, with serious structural damage, showing fragmentation damage. After increasing the cement dosage to 10% and prolonging the curing age to 28d, the specimen has the strongest rigidity and the most compact structure, In the process of compression, shear damage occurs on the weakest surface of the specimen, such as X-type conical damage or 72° diagonal damage. For specimen S10-28d after 15 freeze–thaw cycles, the structure has a little damage, and it is manifested as cleavage damage with large fragmentation.

3.6. X-ray Diffraction Test

X-ray diffraction can detect complex transformations within the microstructure of crystals, with each crystal corresponding to a unique diffraction pattern, thus allowing for a material analysis in terms of its physical phases. According to the results from the XRD diffractograms of loess and modified loess presented in Figure 16, there is a notable distinction between the loess and the stabilized loess. This suggests that the addition of cement and CL interacts with the loess, resulting in the formation of new substances and subsequent structural alterations.

The mineral composition of the loess specimen is mainly dominated by quartz [SiO₂], plagioclase feldspar (loess exists in the form of sodium feldspar [Na(AlSi₂O₈)] and potassium feldspar [K(AlSi₃)O₈]), calcite [CaCO₃], and mica [KAl₂(AlSi₃O₁₀)(OH)₂]. Concurrently, it contains a minor fraction of soluble salts, and these soluble salts disintegrate upon contact with water, thereby exhibiting strong water sensitivity.

After the improvement in the cement–CL composite, not only did the diffraction peaks corresponding to the existing phases increase, but new diffraction peaks also emerged. The emergence of these new peaks can be attributed to two factors. The first is the generation of gel groups. As shown in Figure 17, the pH value of the solution with a 10% cement dosage was 12.3; the pH value of the solution with a 0.5% CL dosage was about 4.4, and that of the mixed solution is 11.8. The robust alkaline environment is conducive to the formation of products such as C-S-H and C-A-H, as shown in Equations (1) and (2). The subsequent generation of hydration products, like calcium aluminates, exerts an adsorptive effect on the surrounding soil particles, which enhances the cohesion of the soil. As a result, the enhancement of the 2(2θ = 27.91°, d = 3.193), 3(2θ = 29.40°, d = 3.034), and 4(2θ = 8.85°, d = 9.982) diffraction peaks are observed. Another reason is the formation of ettringite, known as AFt and represented by the chemical formula 3CaO·Al₂O₃·3CaSO₄·32H₂O, as depicted in Equation (4). It can effectively fill the fine voids and enhance densification. Concurrently, its crystal structure, characterized by needle-like clusters, offers a stronger bonding and friction effect compared to the inter-granular interactions of loess. Such effects contribute to the improvement in soil strength, which is evidenced by the additional diffraction peak (2θ = 16.68°, d = 5.310).

$$C{a}^{2+}+2O{H}^{-}+Si{O}_2\to CaO\cdot Si{O}_2\cdot H_{2}O\left(\mathrm{C}\text{-}\mathrm{S}\text{-}\mathrm{H}\right)$$

(2)

$$C{a}^{2+}+2O{H}^{-}+A{l}_2{O}_3\to CaO\cdot A{l}_2{O}_3\cdot H_{2}O\left(\mathrm{C}\text{-}\mathrm{A}\text{-}\mathrm{H}\right)$$

(3)

$$A{l}_2{O}_3+3S{O}_4^{2-}+6C{a}^{2+}+6O{H}^{-}\to AFt$$

(4)

$$Ca Si{O}_5\left(\mathrm{C}\text{-}\mathrm{S}\text{-}\mathrm{H}\right)+CaS{O}_4\cdot C_4{H}_8{O}_{6}S(\mathrm{CL})+C{a}_{3}A{l}_2{O}_6\left(\mathrm{C}\text{-}\mathrm{A}\text{-}\mathrm{H}\right)\to AFt$$

(5)

Figure 18 illustrates a three-view representation of the nascent crystal structure generated by the VESTA software.

Ettringite possesses a crystal structure that comprises a central column flanked by grooves. Chalcocite is composed of aluminum ions bound to three hydroxyl groups to produce [Al(OH)₆]³⁻ aluminum octahedra, which are connected to the calcium ions (see in Figure 18a). The central column is composed of each aluminum octahedron and the three calcium polyhedra. A considerable quantity of chemical bonds, including hydrogen bonds (e.g., those between aluminum hydroxyls and water molecules and sulfate ions at the polyhedra and H of water molecules and O of sulfate ions), are present. Additionally, van der Waals forces, ionic–covalent bonds, and ligand–covalent bonds contribute significantly to the stability of the calcite crystals [34]. The conditioning temperature was suitable for this experiment, resulting in the formation of elongated needle and rod crystals exclusively.

Calcite is a constituent of the tripartite crystal system, which is illustrated in Figure 18b through its rhombic crystal structure encircled by C-O bonds on the outside. The O site activity of Ca²⁺, which is situated in the interior of the cube, is greater than that of the Ca²⁺ site [35]. Due to the presence of hydrogen bonding conditions, the water molecules in the hydration products are more susceptible to adsorption via the clustering effect.

Calcium feldspar is a component of the triclinic crystal system, as illustrated in Figure 18c. This system consists of silica–oxygen tetrahedra or aluminum–oxygen tetrahedra connected by co-topping to form a three-dimensional shelf structure. Ca²⁺, whose charge number is greater than that of Na⁺, exerts a competitive inhibitory effect on Na⁺. As a result, Ca²⁺ consistently supplants Na⁺, thereby contributing to the formation of calcium feldspar crystals that are both more stable and pure. From the perspective of the crystalline phase, the solidified loess particles appear more solid in comparison to the soluble salt constituent of the loess.

3.7. SEM Test

The microstructure of soil comprises its particles, size, and arrangement. Lei [36] categorized the pores into four distinct groups based on the diameter of the pore. Pores measuring less than 2 μm are classified as micro pores; pores ranging in diameter from 2 to 8 μm are regarded as small pores; pores measuring between 8 and 32 μm are categorized as medium pores; and pores exceeding 32 μm are classified as large pores. As illustrated in Figure 19a,b, the particles of unfrozen, thawed loess are angular in shape and organized in a mosaic pattern. Point–surface contact and surface–face contact predominate, and the overall structural arrangement remains relatively compact, characterized by robust mechanical occlusion. Nevertheless, subsequent to 15 freeze–thaw cycles, the soil became more loosened as a result of water expansion into ice crystals. This led to a weakening of the particle bonds, the appearance of fissures accompanied by particle rolling and shedding phenomena, and a gradual transition from face-to-face interaction to point-to-point contact. The incorporation of cement–CL significantly enhances the microstructure of the loess. The stabilized loess exhibits a more compact structure in comparison to the loess specimens, as illustrated in Figure 19c,d. Additionally, the original shelf pores and mosaic pores are extensively filled with micro and small pores. The joint structure remains intact even after 15 cycles of freezing and thawing, and only minor cracks and overhead pores have developed in comparison to the condition of the joint prior to freezing and thawing.

An enhanced analysis through the incorporation of electron microscope images with greater magnification is illustrated in Figure 20a. The specimen is covered with C-S-H and C-A-H hydration productions, which exhibits favorable bonding characteristics. These could envelop and adsorb the adjacent soil particles while perpetually extending in all directions to establish an interlaced mesh structure. In contrast to the laminar mechanically arranged structure of the loess specimen, the interlocking of particle angles enhances the ability of the particles to occlude one another. Concurrently, the contact area of the particulates is expanded, resulting in a significant enhancement in the soil’s friction strength. Additionally, the enhancement prevents freeze–thaw damage to the soil structure. As shown in Figure 20b, the hydration process of cement generated ettringite crystals resembling needles and rods in the overhead pore space. These crystals were stacked atop one another to fill the pore space of the particles and, in conjunction with the surrounding particles, formed a skeleton that offered a specific level of support. Therefore, it retained its high strength even after cycles of freezing and thawing.

In this paper, three microscopic parameters were selected for analysis: pore area, fractal dimension, and roundness. The grayscale processing of the SEM images at 500× was performed using IPP6.0 software. This resolution allows for greater observation of the cross-sectional features and enhances the precision. The resultant grayscale processed image is shown in Figure 21.

(1)

Pore area

The pore distribution of the loess, as depicted in Figure 22a, is dominated by small and medium-sized pores prior to freezing and thawing, accounting for approximately 66.5% of the total pores, with macro pores comprising 15.7%. This pore structure reflects the characteristic propensity of loess for strong collapsibility. After 15 freeze–thaw cycles, it can be seen that the total pore area of loess increases from 1145.51 μm² to 1634.5 μm². Notably, the proportion of large pores predominates, comprising up to 30.5%, indicating that freeze–thaw cycles progressively enlarge the original medium and small pores into larger pores, weakening the strength. In contrast, the stabilized loess exhibits a predominance of micro and small pores, constituting about 87.5%, with the rest being medium pores. Through freeze–thaw cycles, micro and small pores are transformed into small and medium pores, with no large pores.

(2)

Fractal dimension and roundness

Fractal dimension reflects the complexity of the soil pores [31]. The equivalent perimeter P and the equivalent area A of the pore are obtained by the software to determine the number of fractal dimensions. A straight line is then fitted in double logarithmic coordinates, and the fractal dimension value of the pore can be derived from the slope and intercept of this double logarithmic straight line. The formula is given below:

$$\lg \mathrm{P}=\frac{\mathrm{D}}{2}\lg \mathrm{A}+\mathrm{C}$$

(6)

where P is the equivalent perimeter of the pore, A is the equivalent area of the pore, C is a constant, and D is the fractal dimension of the pore.

As shown in Figure 23, the fractal dimension of S0 specimen is 1.372, and the fractal dimension after freezing and thawing is 1.480, with a large change amplitude, indicating that the freezing and thawing effect led to the development of irregularity in the pores of loess, and there are many fissure damages in the inner part of the loess, which have altered the original structure and increased the fractal dimension. While the pore fractal dimension numbers of S10 before and after a freeze–thaw cycle are 1.410 and 1.446, respectively, the variation is only 0.036. This suggests that the pore structure is regularized, inhibiting freeze and thaw deformation by the addition of cement–CL.

(3)

Roundness

Roundness is a parameter that describes the shape of a pore and represents how closely the pore shape resembles a perfect circle. The roundness are plotted in

Table 5. Roundness of the specimen.

Parameter	S0 (Pre)	S0 (Post)	S10 (Pre)	S10 (Post)
Roundness	1.83	2.36	1.58	1.76

.

The roundness of loess before and after experiencing freezing and thawing was 1.83 and 2.36, respectively, an increment of 0.53, while the roundness of modified loess before and after freezing and thawing is 1.58 and 1.76, respectively, an increment of only 0.18. The growth of ice crystals during freezing and thawing causes the extrusion of soil particles, leading to a dislodgement of the particles, a change in the particle arrangement, an irregular enlarging of the pore space, and increased roundness. The structure is strengthened by the addition of cement–CL, mitigating freeze–thaw damage to some extent. Consequently, the microscopic parameters increase lightly.

3.8. Discussion

Given that cement serves as a crucial medium for reinforcing soils, this study solely contemplates the influence of cement dosage. This method greatly improves the UCS of the soil and the resistance to freeze–thaw cycles. However, the quantity of CL employed is predicated on the optimal doses established in the literature [21]. Subsequent investigations could delve into the impact of CL dosage on the mechanisms underpinning soil reinforcement. Moreover, coupling with dry and wet cycles and dynamic loads should also be further analyzed.

4. Conclusions

In this paper, cement–CL was utilized to improve the physical characteristics of loess, which effectively enhances the mechanical properties and resistance to frost, leading to the following conclusions:

(1)

The addition of cement–CL can significantly improve the mechanical properties of loess, and the UCS of the soil increases with the augmentation of cement dosage and the prolongation of the curing age. It can reach up to 13 times the strength of loess at the promotion of dosing and aging. The positive effect of cement dosage on the improvement in compressive strength is greater than the curing age. When the cement dosage increases to 8% and the maintenance age reaches more than 21 days, the UCS of the soil increases significantly.

(2)

The freeze–thaw damage effect results in larger particle pores and a gradual shift in the particle contact pattern from face-to-face contact to point-to-face contact. The first freeze–thaw cycle causes the greatest structural damage, and the structure tends to be stable after 10 freeze–thaw cycles. The strength loss of loess specimens is as high as 72%, which can be reduced to 28% after stabilization. The addition of an appropriate amount of cement–CL significantly enhances the resistance to freeze–thaw of the soil.

(3)

The improvement mechanisms associated with cement–CL include particle filling, the formation of a mesh structure through hydration, the generation of stable crystals via ion-exchange reactions, and the facilitation of calcium ion by CL. Analyses of the physical phase composition, intergranular contacts, pore content, and pore arrangement corroborate the feasibility of the improvement strategy. This scheme is predicated on loess as the primary matrix, with cement hydration providing stabilizing substances and lignocellulose furnishing a calcium source to facilitate the hydration reaction process.