Study on the Improvement Performance of Different Clay Components with Desulfurization Gypsum-Containing Cementitious Material

Abstract

The use of cementitious materials to improve clay is a common technique in engineering. However, the effectiveness of these materials, particularly desulfurized gypsum, on clays with different mineral compositions remains unclear, resulting in a lack of theoretical basis for their application in engineering. This study investigated the synergistic effects of clinker–metakaolin–desulfurized gypsum on clays with various mineral compositions through a series of macroscopic and microscopic laboratory tests. The results revealed that the stress–strain relationships of all clay samples exhibited softening characteristics. The softening was most pronounced in kaolinite samples, followed by illite and bentonite samples. For single-phase clays, the unconfined compressive strength followed the order of kaolinite > illite > bentonite. For multiphase clays, the order was illite + kaolinite > bentonite + illite + kaolinite > bentonite + kaolinite > bentonite + illite. The strength enhancement in the improved soils was primarily due to kaolinite and illite. As the content of desulfurized gypsum increased, the ettringite crystals in the improved soils transformed from cluster-like to framework-like structures. When the gypsum content exceeded 10%, the macroscopic performance of the improved soils decreased. These findings provide valuable insights for related engineering applications.

1. Introduction

China’s coastal and riverside regions contain deep deposits of soft clay. Due to the high compressibility, high void ratio, and high water content of soft clay, extensive use of cementitious materials is required for improvement during construction. Researchers have systematically studied the modification of soft clay with inorganic cementitious materials and have made significant progress in this field [1,2,3]. Wang et al. [4], Zhang et al. [5], and Sun et al. [6] investigated the role of metakaolin in cement-modified soil. They found that metakaolin contains a large amount of highly active aluminum, which can increase the abundance of hydration products. The active oxides react with Ca²⁺ in an alkaline environment, producing pozzolanic reactions [7,8]. This reaction consumes Ca²⁺ and generates cementitious substances such as C-S-H and C-A-H, thereby improving the mechanical properties of the modified soil [9,10]. With further research into metakaolin cementitious materials [11,12], binary cementitious materials of cement and metakaolin have gradually been developed.

Desulfurization gypsum is a waste byproduct generated in industrial processes, primarily composed of dihydrate gypsum (CaSO₄•2H₂O). In several developed countries, the utilization rate of desulfurization gypsum is notably high. According to the “2020 National Report on Solid Waste Pollution Control in Major Cities” [13], China produces approximately 1.3 × 10⁸ tons of desulfurization gypsum annually. As the second largest solid waste in the country, the comprehensive utilization rate of desulfurization gypsum remains low. The extensive stockpiling and landfilling of this material lead to significant environmental pollution. However, it has been demonstrated that desulfurization gypsum can effectively improve the properties of clay, offering promising solidification effects [14,15]. Therefore, researching the application of desulfurization gypsum aligns with China’s environmental protection policies. Based on the binary cementitious materials of cement and metakaolin, desulfurization gypsum was introduced to form a ternary cementitious material system of cement–metakaolin–desulfurization gypsum.

Analysis of existing research indicates that studies have mainly focused on the effects of cementitious material ratios on the physical and mechanical properties of improved clay [16,17,18], while the impact of clay mineral composition has been overlooked. Clay is primarily composed of minerals such as montmorillonite, illite, and kaolinite.

Table 1. Composition of clinker, metakaolin, and desulfurized gypsum.

Region	Clay Mineral Composition
East China Sea coast [22]	Includes 65.1% illite, 26.6% chlorite, 8.9% kaolinite, and 4.3% montmorillonite
Changdu, Tibet [23]	Illite predominates, followed by mixed-layer illite–smectite, with kaolinite and chlorite present in the lowest amounts
Lianyungang coast [24]	Includes 44% mixed-layer illite–smectite, 29% illite, 13% kaolinite, and 14% chlorite

summarizes the clay mineral compositions from different regions, highlighting significant variations in mineral composition. These variations result in the same cementitious material producing markedly different improvement effects on clays from different regions [19,20,21]. Therefore, studying the improvement performance of cementitious materials for clays with different compositions is crucial for the development of construction projects in regions with soft clay in China.

It is noteworthy that to avoid the influence of natural gypsum in cement on the research results, clinker was used as a substitute for cement in this study. Artificial clays with different compositions were prepared. A series of macro and micro laboratory tests were conducted to explore the performance of these clays when improved with the ternary cementitious material composed of clinker, metakaolin, and desulfurization gypsum. This study aimed to reveal the mechanisms by which the ternary cementitious material containing desulfurization gypsum enhances the mechanical properties and microstructural characteristics of clays with different compositions.

2. Materials and Method

2.1. Experimental Materials

2.1.1. Cementitious Materials

The ternary cementitious materials used in the experiments consist of clinker (C), metakaolin (MK), and desulfurization gypsum (G). X-ray fluorescence (XRF) spectroscopy was performed using a Japanese ZSX Primus III+ instrument to determine the primary chemical components. The loss-on-ignition test was conducted at a controlled temperature of 950 °C. The test results are presented in

Table 2. Chemical composition and content of clinker, metakaolin, and desulfurization gypsum (%).

Dioxide	CaO	SiO2	Al2O3	Fe2O3	SO3	TiO2	Loss on Ignition
clinker	37.9%	34.7%	11.5%	5.1%	0%	0.5%	10.23
metakaolin	0.2%	51.3%	45.9%	0.7%	0%	0.9%	0.91
desulfurized gypsum	37.1%	2.2%	0.8%	0.3%	48.5%	0.1%	10.87

. Clinker is sourced from a cement processing plant in Chongqing, China. The metakaolin is mined from a location in Shijiazhuang, Hebei Province. The desulfurization gypsum is obtained from the flue gas desulfurization ash of a power plant in Fujian Province. Based on existing research findings [25,26], the amount of clinker was set at 14% of the total mass of clay, and the amount of metakaolin was set at 25% of the total mass of clinker. The amount of desulfurization gypsum, considered as a variable, was determined according to the amount of clinker. In this study, the G/C ratios were set at 0%, 5%, 10%, 15%, and 20%. The water content is determined as 1.5 times the liquid limit.

2.1.2. Artificial Clay

In this study, bentonite, illite, and kaolinite, which are widely distributed, were used to prepare clays with different mineral compositions. In this study, a laser particle size analyzer, model Malvern Mastersizer 2000, was used to analyze the particle size distribution of the soil materials. Figure 1 presents the resulting particle gradation curves. The basic properties of the different minerals were tested according to the “Standard for Geotechnical Testing Methods (GB/T 50123-2019) [27]” and are presented in

Table 3. Basic properties of bentonite, illite, and kaolinite.

Clay Mineral	Bentonite	Illite	Kaolinite
Liquid limit (%)	84.0	33.3	30.0
Plastic limit (%)	48.0	17.0	18.7
Plasticity index (%)	36.0	16.3	11.3
Main mineral components	Montmorillonite	Illite	Kaolinite

.

2.2. Experimental Method

2.2.1. Unconfined Compressive Strength Test

The unconfined compressive strength test was conducted using a universal testing machine to study the improvement performance of the ternary cementitious material on clays with different compositions. The test group configurations are shown in

Table 4. Test conditions.

Test Condition	Clay Minerals	Clinker	Metakaolin	Curing Time	Water Content	Desulphurization Gypsum/Clinker (%)
	Bentonite/Illite/Kaolin	(%)	(%)	(d)	(%)	0	5	10	15	20
B1I0K0	1:0:0	14	3.5	7, 28, 56	126.0	0	0.7	1.4	2.1	2.8
B0I1K0	0:1:0				75.0
B0I0K1	0:0:1				67.5
B1I1K0	1:1:0				98.1
B1I0K1	1:0:1				94.4
B0I1K1	0:1:1				71.3
B1I1K1	1:1:1				87.9

. During sample preparation, the prepared clay and cementitious materials were thoroughly mixed and added to the saturator in three stages. The saturator was then placed in a curing room. After 10 h, the samples were demolded and wrapped in plastic wrap. The samples were cured in a constant temperature and humidity environment (20 ± 3 °C and 95%) for 7, 28, and 56 days, respectively. The samples were cylindrical, with a diameter of 35 mm and a height of 80 mm. Tests were conducted when the samples reached the target curing age.

During the test, the vertical displacement rate was controlled at 1 mm/min, and the test was stopped when the axial strain reached 10%. Each test group included three parallel samples, and the analyzed indicators were the average values.

2.2.2. Microstructure Tests

XRD and SEM tests were conducted to elucidate the mechanism by which the ternary cementitious material improves the performance of clays with different compositions from a microstructural perspective. After completing the unconfined compressive strength test, the crushed samples were dried and sealed for use in the microstructure tests.

X-ray diffraction (XRD) analysis was conducted using a Smartlab 9 kW high-resolution X-ray diffractometer. The scanning angle was set from 5° to 85°, with a scanning speed of 1°/min. During sample preparation, the powdered samples were first placed onto the slide. The samples were then compacted, and excess material outside the groove was removed. Finally, the cover slip was pressed tightly against the slide, and the samples were positioned on the test stage to initiate scanning. SEM tests were conducted using a Hitachi Regulus 8100 electron microscope with magnifications of 500× and 5000×. The 500× magnification was used to analyze changes in pore size and number, while the 5000× magnification was used to analyze the hydration products of the improved soil.

3. Compressive Strength Characteristics

3.1. Stress–Strain Relationship

The stress–strain relationship aims to reflect the variation in the characteristics of soil stress with strain under uniaxial compression. Figure 2 presents the stress–strain curves of clay samples with different compositions after being improved by ternary cementitious materials at a curing age of 56 days. It can be observed that clay samples with different mineral compositions exhibit strain-softening characteristics. The stress–strain curves show a reverse U-shaped profile, which can be divided into three stages: Elastic deformation stage: the axial stress increases linearly with axial strain. Plastic deformation stage: the axial stress increases nonlinearly with axial strain, and the growth rate gradually slows down. Yielding and failure stage: after reaching peak strength, the axial stress rapidly decreases with increasing axial strain, exhibiting strain-softening characteristics. The addition of desulfurization gypsum significantly alters the stress–strain characteristics of the samples, gradually reducing the plastic deformation stage of the curve. This indicates that adding desulfurization gypsum can significantly improve the compressive and deformation resistance of the soil while, to some extent, weakening the plasticity of soft soil. This phenomenon exemplifies the significant enhancement of clay properties through the use of ternary cementitious materials.

Due to the inherent strength differences in minerals, the strain-softening characteristics vary among the three single-phase mineral composition clays (B1I0K0, B0I1K0, B0I0K1). Kaolinite exhibits the most pronounced softening, followed by illite, while montmorillonite shows the least noticeable softening. As the amount of desulfurization gypsum increases, the strain-softening characteristics of different clay samples become more pronounced, as indicated by changes in peak strength. The peak strengths of kaolinite, illite, and the illite–kaolinite mixture samples reached their maximum at a G/C ratio of 10%, measuring 1.93 MPa, 1.40 MPa, and 1.63 MPa, respectively. For the bentonite sample, the peak strength peaked at a G/C ratio of 20%, achieving 0.22 MPa. The peak strengths of the bentonite–illite, bentonite–kaolinite, and bentonite–illite–kaolinite samples reached their maximum at a G/C ratio of 15, recording values of 0.70 MPa, 0.98 MPa, and 1.33 MPa, respectively.

3.2. Relationship between Peak Strength and G/C Ratio

Figure 3 presents the relationship curves between the peak strength of clays with different compositions and the G/C ratio. At a curing age of 7 days, the peak strength gradually increases with an increase in the G/C ratio. When the G/C ratio increases from 0% to 20%, the ranges of peak strength variations for montmorillonite, illite, kaolinite, montmorillonite + illite, montmorillonite + kaolinite, illite + kaolinite, and montmorillonite + illite + kaolinite clay samples are 0.066~0.124 MPa, 0.245~0.309 MPa, 0.478~0.682 MPa, 0.134~0.245 MPa, 0.221~0.373 MPa, 0.327~0.417 MPa, and 0.221~0.384 MPa, respectively. From the perspective of single-phase composition clays, when the G/C ratio is 0%, the peak strength of kaolinite reaches 0.478 MPa, increasing cumulatively by 0.204 MPa when the G/C ratio reaches 20%, significantly higher than the other samples. This indicates that ternary cementitious materials have the most optimal improvement effect on kaolinite, followed by illite, and least on montmorillonite. This improvement effect similarly influences the peak strength of multiphase composition clays.

At curing ages of 28 days and 56 days, the relationship between peak strength changes in different composition clay samples and desulfurization gypsum (G) content is not linear, showing the presence of a threshold G/C ratio. The G/C thresholds are 10% for illite, kaolinite, and illite + kaolinite samples and 15% for montmorillonite + illite, montmorillonite + kaolinite, and montmorillonite + illite + kaolinite samples. In contrast, the peak strength of montmorillonite samples continues to show an increasing trend when the G/C ratio reaches 20%, although the rate of increase significantly decreases.

Based on the above analysis,

Table 5. Optimal desulfurization gypsum content for clay samples with different compositions.

Curing Time	B1I0K0	B0I1K0	B0I0K1	B1I1K0	B1I0K1	B0I1K1	B1I1K1
7d	20%	20%	20%	20%	20%	20%	20%
28d	20%	10%	10%	15%	15%	10%	15%
56d	20%	10%	10%	15%	15%	10%	15%

summarizes the optimal desulfurization gypsum (G) content for clay samples with different compositions.

3.3. Relationship between Peak Strength and Curing Time

The reaction of cementitious materials in improving soil is generally divided into three stages: the hydration reaction stage and the secondary hydration reaction stage, corresponding to curing ages of 0–7 days and 7–28 days, respectively. The third stage, corresponding to a curing age of 28–56 days, is typically considered where the strength of ordinary cemented soil no longer increases. The cementitious material used in this study is a ternary material, and its reaction characteristics are not yet fully understood. Therefore, this study explores the relationship between peak strength and curing age for clays with different compositions.

Figure 4 illustrates the relationship between the rate of increase in peak strength and curing age for single-phase clay samples. Taking the G/C ratio as 20% as an example, the rates of the increase in peak strength for the three reaction stages are as follows: for montmorillonite samples, they are 0.01770 MPa/day, 0.00450 MPa/day, and 0.00126 MPa/day; for illite samples, they are 0.04410 MPa/day, 0.0300 MPa/day, and 0.01850 MPa/day; and for kaolinite samples, they are 0.09740 MPa/day, 0.03710 MPa/day, and 0.01820 MPa/day. It is evident that with increasing curing age, the rates of the increase in peak strength for single-phase clay samples show a decreasing trend. Compared to montmorillonite and kaolinite samples, illite samples exhibit the smallest rate of decline in strength increase.

Figure 5 depicts the relationship between the rate of increase in peak strength and curing age for multiphase component clays. Taking the G/C ratio as 20% as an example, the rates of increase in peak strength for montmorillonite + illite samples, montmorillonite + kaolinite samples, illite + kaolinite samples, and montmorillonite + illite + kaolinite samples in the first reaction stage (0–7 days) are 0.03500 MPa/day, 0.05329 MPa/day, 0.05957 MPa/day, and 0.05486 MPa/day, respectively. In the second reaction stage (7–28 days), the rates of increase in peak strength are 0.01700 MPa/day, 0.02264 MPa/day, 0.02286 MPa/day, and 0.02407 MPa/day, respectively. In the third reaction stage (28–56 days), the rates of increase in peak strength are 0.00653 MPa/day, 0.00318 MPa/day, 0.01750 MPa/day, and 0.00885 MPa/day, respectively. It is evident that the rates of increase in peak strength for multiphase component clay samples are between those of the included single-phase clay samples and closer to the high-performance single-phase clay mineral samples. This can be attributed to the particle size distribution of different component clays.

Based on the analysis above, it is evident that ternary cementitious materials exhibit superior improvement effects on kaolinite and illite, resulting in a more pronounced increase in peak strength. Additionally, unlike conventional cement-stabilized soil, the peak strength of the modified clay in this study continues to increase after 28 days of curing and demonstrates further growth potential even after 56 days. This phenomenon is attributed to the ability of desulfurization gypsum within the ternary cementitious materials to accelerate specific ion exchange reactions. This process generates a network-like gel that effectively fills particle voids and bonds the particles together, thereby enhancing the compressive strength of the modified soil.

Furthermore, a quantitative relationship formula is established between the peak strength of modified soil samples and clay components:

$$UCS=B{x}_1+I{x}_2+K{x}_3$$

(1)

In the formula, $UCS$ represents the peak strength of the modified soil; $B$, $I$, and $K$ denote the contribution values of montmorillonite, illite, and kaolinite to the peak strength, respectively. $x_1$, $x_2$, and $x_3$ represent the proportions of montmorillonite, illite, and kaolinite in the clay, respectively. Under different G/C ratios and curing periods, the fitting relationships of peak strength for different component-modified soil samples are summarized in

Table 6. Fitting relationships of peak strength for different component clays.

Curing Time	G/C (%)	$\mathit{B}$	$\mathit{I}$	$\mathit{K}$	Correlation Coefficient (R2)
7 d	0	0.0372	0.2492	0.5016	0.99
	5	0.0515	0.2771	0.5739	0.99
	10	0.1073	0.2905	0.6517	0.99
	15	0.1377	0.2625	0.5193	0.99
	20	0.1174	0.2410	0.4546	0.99
28 d	0	0.0490	0.5002	0.8118	0.98
	5	0.0783	0.6323	1.0999	0.97
	10	0.2071	0.7507	1.2007	0.99
	15	0.1377	0.2625	0.5193	0.99
	20	0.2167	0.6643	1.1295	0.98
56 d	0	0.0377	0.8484	1.2637	0.98
	5	0.0167	1.1301	1.6196	0.96
	10	0.0415	1.3478	1.8531	0.98
	15	0.1887	1.2704	1.8170	0.99
	20	0.1651	1.1766	1.6301	0.99

.

4. Microstructural Characteristics Analysis

Scanning electron microscopy (SEM) and X-Ray diffraction (XRD) analyses were conducted to investigate the pore structures and hydration products of bentonite, illite, and kaolinite samples modified with ternary cementitious materials after a curing period of 56 days.

4.1. Void Analysis

The internal pore structures of the three single-component clay samples, observed at 500× magnification, are shown in Figure 6.

Figure 6 illustrates the internal pores of the bentonite sample. It can be seen that the soil contains numerous large pore structures. As the desulfurized gypsum content increases, both the number and size of pores within the bentonite sample gradually decrease. However, when the G/C ratio reaches 20%, a small number of pores still remain.

Figure 7 shows the internal pore structures of the illite samples. Significant variations in pore structure are observed with different desulfurized gypsum contents. Specifically, when the desulfurized gypsum content is 0%, there are some large pore structures within the sample. As the gypsum content increases, both the size and number of pores decrease. At 10% gypsum content, it becomes challenging to directly observe large pores or micro-cracks, and the overall microstructure appears relatively smooth. However, with further increases in gypsum content, small cracks begin to appear, and the particle morphology becomes rougher.

Figure 8 presents the internal pore structures of the kaolinite samples. Similar to the illite samples, as the desulfurized gypsum content increases from 0% to 10%, the pore size and number in the kaolinite samples decrease, achieving a dense structure. However, with further increases in gypsum content, the original structural morphology is disrupted, resulting in micro-cracks and roughened particle surfaces.

From the analysis, it can be concluded that at low desulfurized gypsum content (0–10%), the pore size and number in the samples follow a decreasing order of bentonite, illite, and kaolinite. When the gypsum content reaches 15% or higher, illite and kaolinite samples exhibit swelling and cracking, while the pores in bentonite samples remain incompletely filled. This discrepancy in pore filling is a significant factor contributing to the differences in mechanical properties among the three types of samples.

4.2. Analysis of Hydration Products

Figure 9 and Figure 10 show the SEM and XRD results of bentonite samples at 5000× magnification, respectively. It can be observed that bentonite samples exhibit loosely layered structures and abundant hydration product crystals under different desulfurized gypsum contents. The diffraction peaks are found within the ranges of 5–10°, 15–20°, and 60–65°. With increasing gypsum content, needle-shaped ettringite crystals begin to appear within the bentonite samples, although the overall structure remains loose. This loose structure is the primary reason for the lower mechanical performance of the modified bentonite samples.

Figure 11 and Figure 12 present the SEM and XRD results, respectively, of illite samples modified with ternary cementitious materials at 5000× magnification. The illite samples exhibit numerous flaky structures and hydrated product crystals at different desulfurization gypsum contents. The incorporation of desulfurization gypsum leads to the formation of network structures of calcium silicate hydrate (C-S-H) and calcium aluminosilicate hydrate (C-A-S-H), along with a small number of needle-like and clustered ettringite crystals. These formations enhance the macroscopic performance of the modified soft soil. As the desulfurization gypsum content increases, the ettringite crystals gradually transition from a clustered to a framework structure. Consequently, when the desulfurization gypsum content exceeds 10%, the mechanical properties of the modified illite samples slightly decline.

Figure 13 and Figure 14 show the SEM and XRD results, respectively, of kaolinite samples modified with desulfurization gypsum at 5000× magnification. After the addition of desulfurization gypsum, the kaolinite samples exhibit irregular flaky structures and numerous hydration products. As the desulfurization gypsum content increases, the calcium aluminosilicate hydrate (C-A-S-H) in the samples begins to interweave and forms a framework structure on the kaolinite particles. When the desulfurization gypsum content ranges from 0% to 10%, the diffuse peak in the diffraction angle range of 5° to 10° gradually disappears.

5. Conclusions

Through a series of macro and micro indoor tests, the modification performance and microstructural changes in different clay components using a ternary cementitious material composed of clinker, metakaolin, and desulfurization gypsum were studied. The mechanisms of clay modification by these materials were investigated. The main conclusions are as follows:

• The stress–strain relationships of various clays modified by the ternary cementitious material exhibit softening characteristics. The softening phenomenon is most pronounced in kaolinite samples, followed by illite and bentonite samples. Desulfurization gypsum in the ternary material accelerates specific ion exchange reactions, forming network gels that effectively fill particle pores and bond particles together, thereby enhancing the compressive strength of the modified soil.
• Increasing the desulfurization gypsum content in the ternary cementitious material improves the rigidity of different clay samples while reducing their plasticity. There is an optimal gypsum content for each clay component, which varies with curing age.
• The unconfined compressive strength (UCS) of clays modified by the ternary cementitious material varies among different components. For single-component clays, the UCS follows the order of kaolinite > illite > bentonite. For multi-component clays, the UCS follows the order of illite + kaolinite > bentonite + illite + kaolinite > bentonite + kaolinite > bentonite + illite. Kaolinite and illite are more effective in enhancing peak strength. A quantitative relationship between peak strength and clay components was established.
• The incorporation of desulfurization gypsum in the modified soil leads to the formation of network structures of calcium silicate hydrate (C-S-H) and calcium aluminosilicate hydrate (C-A-S-H), along with a small number of needle-like and clustered ettringite crystals. As the gypsum content increases, the ettringite crystals transition from a clustered to a framework structure. This explains the decrease in the mechanical properties of illite and kaolinite samples when the gypsum content exceeds 10%.