Application of MICP in Water Stability and Hydraulic Erosion Control of Phosphogypsum Material in Slope

Abstract

Phosphogypsum is a kind of solid waste that occupies land resources and harms the environment. It can be used as a solidified material, but the utilization of phosphogypsum is limited by its impurities and weak strength performance. This study aimed to use microbial-induced carbonate precipitation (MICP) to improve the water stability, permeability, and hydraulic erosion resistance of phosphogypsum and evaluate its impact on the environment. In this paper, the phosphogypsum samples and artificial slopes were prepared and solidified by spraying various concentrations of bacteria solution and cementation solution to achieve microbial modification. The water stability and permeability test were used to calculate the mass of spalling under water shaking and the permeability coefficient. A rainfall scouring test was carried out to estimate the erosion resistance. The erosion degree was quantitatively calculated using 3D laser scanning technology. The results show that the microorganism treatment can improve water stability and reduce the permeability coefficient, while the differences between the content of CaCO₃ in the outermost layer and in the inner layer gradually increase with the increase in bacterial concentration, and the permeability coefficient was reduced uniformly. The sediment loss of the slope after MICP treatment was much less than that of the untreated slope, and the connection force between the particles was strengthened. By observing the morphology of the scoured samples, we found that the treated particles were aggregated and flocculated with more macropores, which led to the formation of erosion pits under scouring. The pH of the outflow of the modified slope was neutral, and the heavy metal elements were fixed by microbial action and carbonate, which is not harmful to the environment.

1. Introduction

Phosphogypsum is one of the largest solid wastes in the phosphate fertilizer industry. Each ton of P₂O₅ produces 4.5–5.5 t of by-product phosphogypsum [1,2]. According to statistics, worldwide PG production can reach 280 million tons per year. The annual output of phosphogypsum in China is 70 million tons, but the utilization rate is less than 15%, which causes serious land resource occupation and environmental pollution [3]. The acidity and the impurities in phosphogypsum will threaten the soil and groundwater environment, which incurs costs for its treatment and purification [4]. Phosphogypsum is a kind of gypsum material rich in phosphorus, and it contains soluble phosphorus, organic matter, and other impurities, which will affect the crystallization time of phosphogypsum crystals and reduce its mechanical properties. It is necessary to reduce the harmful influence by washing, flotation, calcination or adding other chemical stabilizers. A single treatment method cannot meet the conditions of use, and multiple methods are often required in combination, though this increases the treatment time and cost at the same time.

Some scholars have argued that phosphogypsum can be used in building materials, such as gypsum board, by adding an appropriate amount of modifier, and it is also commonly used as a cement retarder [5,6]. It is found that the use of pretreated phosphogypsum can replace natural gypsum as a cement retarder, which has no negative impact on cement strength [7]. Some studies recommend mixing phosphogypsum with other chemical additives for use as roadbed material, or as fillings to improve the mechanical strength and water stability of the composites [8,9,10,11]. By adding an artificial cementitious agent, including cement, lime and fly ash, to solidify phosphogypsum, the solubility of the pollutants is reduced, and it can be used as a backfill stabilizer and structure layer of pavement [12,13]. The increase in temperature can also remove the impurities and heavy metal ions on the surface so that it can be reused [14]. The main composition of phosphogypsum is CaSO₄·2H₂O, which can react with cement or lime to formulate pozzolanic products under an alkaline environment. The product of the chemical reaction between phosphogypsum and cement hydrate is ettringite, which provides the main skeleton with strength support, but the excessive production of by-products will expand the material and result in potential heaving failure [15,16]. At present, the treatment and application of solid waste is a research hotspot, and the modification and reuse of phosphogypsum are of great significance to the sustainable development of phosphogypsum resources.

With the rapid development of industrial society, lots of disturbed slopes with seriously damaged substrates have been formed, and they are easily degraded or even collapsed under gravity and rainfall [17]. Therefore, it is necessary to enhance the erosion resistance of the slope matrix and reduce soil loss. Adding chemical stabilizers, such as fly ash, cement, or polyacrylamide (PAM), to gypsum material can be used to improve soil properties [18,19]. However, the traditional curing methods will lead to different degrees of impact on the environment. The use of cement or lime requires a lot of energy consumption and discharges harmful gases [20,21]. Most of the chemical grouting materials, such as polyurethane and epoxy, are toxic and harmful to human health [22]. Microbial mineralization exists widely in nature. Compared with the traditional curing method, it has the advantages of having a controllable reaction, low energy consumption, and being environmentally friendly [23]. The usage of phosphogypsum is limited, and it is only used as an additive; therefore, its utilization rate still needs to be improved.

Microbial-induced carbonate precipitation (MICP) is a new method of material modification that has developed in recent years, and it can change the particle connection mode and pore structure characteristics of the material. The existing modification method of using an organic polymer chemical solution has the disadvantages of creating a toxic chemical solution and polluting groundwater environment. MICP uses the urease secreted by carbonate-mineralization microbes to decompose urea and generate carbonate ions, which combine with metal cations in the background, such as calcium ions and magnesium ions, to produce carbonate precipitation [24,25,26]. The precipitation can bond loose particles together to improve strength performance [27,28,29,30], and block pores in porous media to reduce the seepage in pores [31,32,33]. The phenomenon of microbial-induced carbonate precipitation is widespread in the natural environment, and there are many carbonate mineralization microbes distributed in the soil or seawater [34]. On the basis of this phenomenon, MICP is an environmentally friendly method.

Bacillus pasteurii with high urease yield is commonly used in MICP [35]. Many scholars use this strain to modify concrete or other loose materials, such as sand and soil, which can effectively reduce the water absorption and moisture permeability of porous medium and improve the strength and erosion resistance of materials [36,37]. The precipitation can be used as a filling material to repair cracks [38,39]. As early as 1992, Kantzas et al. used calcium carbonate induced by MICP to seal the pores between sand particles, which reduced the porosity of the sand column by 50% and the permeability coefficient by nearly 90% [40]. Previous studies have shown that MICP can significantly enhance the mechanical properties of various materials [41,42]. Yang Zuan et al. extracted mutagenic strains to prepare microbial mortar, and the highest strength reached 55 MPa, improving the splitting tensile strength and cyclic load resistance of the mortar [39]. Researchers found that the corrosion resistance of the materials was enhanced by MICP treatment. Ramakrishnan et al. used Bacillus pasteurii to form a calcium carbonate covering on the concrete surface, which reduced the permeability of the test block and improved the resistance to alkali, acid, freeze–thaw and drying shrinkage [43]. M. Maleki et al. studied the effectiveness of the MICP process to control wind erosion through a wind tunnel test, and the higher wind speed was associated with a more obvious control effect [44]. Volodymyr Ivanov et al. found that the sandy foreshore slope modified by MICP could resist 30 times the amount of tidal flow scouring [45]. As an emerging modification technology, MICP has proven its ability to improve the physical and chemical properties of materials. Numerous studies have shown that MICP is able to enhance the stability of materials and their resistance to hydraulic erosion [17,46].

In this paper, MICP was applied to the modification of phosphogypsum materials. We tested the water stability and permeability of phosphogypsum samples with microbial treatments, and solidified phosphogypsum to simulate an artificial slope using MICP. A laboratory rainfall scouring test was conducted to estimate erosion resistance performance and the environmental impact effect, the amount of sediment loss, the pH, the amount and type of harmful metal elements and the microstructure of the outflow, and the results were compared between the microbially treated and untreated slope. Moreover, 3D laser scanning technology was combined to calculate the slope erosion degree. This study will provide a preliminary experimental basis for applying phosphogypsum to slope protection.

2. Materials and Methods

2.1. Phosphogypsum Material

The phosphogypsum waste used in the experiment was taken from Kunming, Yunnan province in China. As shown in Figure 1, the initial water content of the material is high, and the grains are easily caked. Its principal physical parameter is shown in

Table 1. Principal physical parameter of phosphogypsum.

Liquid Limit (%)	Plastic Limit (%)	Plastic Index (%)	Maximum Dry Density (g/cm3)	Optimum Moisture Content (%)	Specific Gravity (g/cm3)	Particle Size Distribution
						d10	d30	d60
29.9	19.6	10.3	1.41	20.7	2.4	1.15	5.61	27.39

, d₅₀ = 18 μm, which is similar to fine-grained soil with good gradation.

Table 2. Chemical composition of phosphogypsum.

	CaO	SO3	SiO2	P2O5	Al2O3	Fe2O3	MgO	Na2O	K2O	F	Crystal Water	Organism
Wt (%)	30.12	40.33	9.71	0.70	0.20	0.09	0.01	0.22	0.24	0.20	18.10	0.08

presents its chemical composition. The phosphogypsum particles used in the experiment needed to pass through a 2 mm sieve.

2.2. Microbial Treatment Solution

The urease-positive microorganism used in this test was Sporosarcina pasteurii (ATCC 11859), and the size of bacteria is 0.5–3 μm [47]. This kind of bacterium has high urease activity and is ubiquitous in soil. It can produce large amounts of precipitates in a short period of time via the generation of adenosine triphosphate through the secretion of urease [45,48]. The bacteria were cultivated in a sterile aerobic NH₄-YE liquid medium consisting of 20 g/L yeast extract, 10 g/L ammonium sulfate, 0.13 M Tris (to keep the pH = 8–9). The organism grew for 24 h after inoculation was at a steady state, and the optical density value at 600 nm (OD₆₀₀) varied between 3.5 and 4.5. The urease activity of ureolytic bacteria culture ranges from 8 to 10 mM urea/min. The OD₆₀₀ of the bacterial solution used in the experiment was diluted to 0.1, 0.3 and 0.5, which is approximately 1.42 × 10⁷ cells/mL, 4.37 × 10⁷ cells/mL, and 7.1 × 10⁷ cells/mL [25], to prevent the high concentration of bacteria from blocking the pores. The final pH of the three bacteria solution was 7.8, 7.9 and 8.0, respectively.

The cementation solution is a mixture of urea and calcium chloride with equal concentration. The concentration of the solution is 0.5 M, meaning it has a moderate concentration. The cementation solution used in this experiment has the same concentration, and the pH of the solution is 9–10.

2.3. Water Stability Test

The sample size for the water stability test is 6.18 cm in diameter and 2 cm in height. The dried phosphogypsum was filled into the mold and compacted. The samples prepared were placed in a beaker and sprayed with 100 mL bacterial solution of different concentrations, namely, H₂O (E0), OD₆₀₀ = 0.1 (E2), OD₆₀₀ = 0.3 (E3), OD₆₀₀ = 0.5 (E4). After standing for 4 h, 100 mL of cementation solution was sprayed on the surface and kept for 12 h. After three cycles of treatment, the water stability test was carried out after drying at 45 °C. The prepared sample was placed in a beaker, and we then added water, the water level was added 2 cm below the top of the sample. We shook the beaker at a speed of 200 rpm, and measured the mass of the peeling samples and the content of calcium carbonate of the peelings during different time [49].

The same preparation method was used to prepare the samples for the penetration test. The sample size was 6.18 cm in diameter and 4 cm in height. The permeability coefficient of the phosphogypsum samples modified by different bacterial concentrations was measured. Three parallel tests were carried out for each type of sample to calculate the mean value (AVE), maximum value (MAX), minimum value (MIN) and the coefficient of variation (CV).

2.4. Environmental Impact Test

When phosphogypsum is used as a slope stabilization material, corrosivity caused by acid and its heavy metal elements may enter the environment through rain scouring and destroy the ecology. Therefore, the pH value and toxicity of the leaching solution are used as the environmental detection standards of phosphogypsum. The content of heavy metal elements in phosphogypsum samples was determined by inductively coupled plasma atomic spectrometry, and a pH meter was used to measure the pH value of the scouring fluid in the rainfall scouring test.

2.5. Artificial Slope Preparation

The artificial slope was simulated in a rectangular thickened plastic tray of 38 cm × 25 cm × 3.5 cm and solidified by the spaying method. The length of the slope was relatively long so it could simulate the runoff state of the surface. The sieved phosphogypsum was spread in the tray. In order to investigate the solidified effectiveness of the various concentrations of bacteria solution, the microbially modified slopes were treated by OD₆₀₀ = 0.1 (S2), OD₆₀₀ = 0.3 (S3) and OD₆₀₀ = 0.5 (S4), and the concentration of the cementation solution was 0.5 M. For MICP processes, 1 L of bacteria solution was uniformly sprayed on the surface and percolated for 8 h. After that, 1 L of cementation solution was sprayed on the surface in the same manner. The bacteria solution and cementation solution were fully mixed and reacted to form CaCO₃ for 12 h. However, the same amount of distilled water was sprayed on the phosphogypsum in S1 as the control group. After three cycles of treatment, the slope surface was air-dried for the laboratory rainfall scouring test.

2.6. Rainfall Scouring Test

The device designed for simulating slope erosion under rainfall is shown in Figure 2. The slope angle was set at 30°. The spraying device was set on the metal frame above the plastic tray. The nozzle of the spraying device was 50 cm away from the top of the slope, and could be evenly sprayed on the slope. There was another tray placed at the bottom to collect the slope sediment losses. A video was set in front of the device to continuously record the slope erosion pattern of the model slope. Although the scale of the experimental device was small and the slope border has a size effect, the difference between the microbially modified slope and the untreated slope can also be compared under the same conditions, so that the effectiveness of the modification can be confirmed.

In this experiment, the rainfall intensity was controlled at 300 mm/h, which was a rainstorm-level intensity [17]. The amount of the losses in the tray collected in initial 5 min, 10 min, 20 min, 30 min, 40 min, 50 min and 60 min was recorded as the erosion amount of the slope. Meanwhile, the pH of the outflow was also measured. The weight of the losses after drying was recorded as the erosion weight of the slope [49].

2.7. CaCO₃ Content and Micromorphology

After MICP treatment, CaCO₃ crystals were filled between the phosphogypsum particles. The CaCO₃ content of the consolidated samples was determined by immersing the dried sample in 1 M hydrochloric acid solution, and the reduction in mass is the weight of CaCO₃ [30,41]. The content of CaCO₃ is calculated by the difference between the initial mass (m₁) after curing and the residual mass (m₂) after soaking in acid [50]. The micromorphology of the scoured samples could be observed using a digital microscope, which only needed to be connected with the computer via the USB interface. The maximum magnification is ×1000.

$$\mathrm{c}\left({CaCO}_3\right)=\frac{\left(m_1-m_2\right)}{m_1}\times 100\%$$

(1)

2.8. Three-Dimensional Laser Scanning Technique

Recently, 3D laser scanning technology has been widely used to invert the surface morphology. In rock mechanics, 3D scanners are used to describe the rough profile of the structural plane and define the roughness coefficient (Rs) [51]. In this paper, a VIUscan 3D handy laser scanner was used to analyze the scoured surface of the phosphogypsum sample, and the scanning accuracy was 0.8 mm. The laser scanner could identify the point cloud of an object in a short time, and formed a digital model with a high resolution and high precision. The roughness of the target surface could be obtained by further processing and analyzing the point cloud. Rs is the ratio of the actual area of the sample surface to its projected area [52]. The higher value of Rs is associated with a rougher surface.

$$R_S=\frac{A_t}{A_n}$$

(2)

where A_t is the actual area of the surface and A_n is the projected area of the surface.

3. Results and Discussion

3.1. Heavy Metal Elements of Phosphogypsum Samples

Table 3. Content of heavy metal elements (mg/L) in different phosphogypsum samples.

Element	Untreated Phosphogypsum	Microbially Treated Phosphogypsum	Critical Value
Se	<0.01	<0.01	0.01
Cd	<0.003	<0.003	0.003
Pb	<0.05	<0.05	0.05
Ni	<0.01	<0.01	0.01
Hg	<0.01	<0.01	0.01
As	<0.1	<0.01	0.01
Cr	<0.01	<0.01	0.01
Be	0.045	<0.005	0.005
Ba	<0.004	<0.004	0.004
Ag	<0.1	<0.1	0.1
Cu	<0.01	<0.01	0.01
Zn	0.181	<0.006	0.006

shows the contents of some heavy metal elements in phosphogypsum samples before and after microbial modification. The contents of Be²⁺ and Zn²⁺ in the original phosphogypsum sample exceed the standard critical value, but they meet the emission standards after microbial treatment. In the process of microbial modification, the ion states of the heavy metals in phosphogypsum are extracted and fixed by combining them with carbonate through microbial complexation, which changes the chemical form of the heavy metal elements. The movement of harmful elements is reduced so as not to damage the ecological environment [53].

3.2. Water Stability and Permeability Characteristic of Phosphogypsum Sample

As shown in Figure 3, the residual mass of the samples decreased under continuous shaking in water. The cumulative spalling mass of the samples treated with water E1 was 37.24 g, which was greater than that of those treated with the microbial solution. The microbially treated samples had a similar change tendency: the sample E2 treated by OD₆₀₀ = 0.1 had the lowest peeling amount at 19.45 g, while the result of the OD₆₀₀ = 0.5 treated sample E4 was slightly larger at 24.78 g. It can be seen from the spalling results that the water stability of the samples treated by microorganisms is improved.

Figure 4 illustrates that the content of CaCO₃ at various stages of peeling has a decreasing trend, the outermost layer sample has the highest content of CaCO₃, and the internal transition gradually reduces. The linear curve was used to fit the downward trend. The outermost CaCO₃ content of the OD₆₀₀ = 0.5 treated sample is the highest. As the bacterial concentration decreased, CaCO₃ content slightly decreased. However, regarding the internal area of the sample, the content of CaCO₃ had the fastest decline in the E4 curve, and the gradient was the largest. This indicates that with the increase in bacterial concentration, the formation of CaCO₃ is more discrete, and the interior of the sample has a greater difference in CaCO₃ content.

The permeability characteristics of the samples modified with different bacterial concentrations decreased significantly, which is shown in

Table 4. Coefficient of permeability (cm/s) in different phosphogypsum samples.

Sample	K1	K2	K3	AVE	MAX	MIN	CV
OD600 = 0	1.49 × 103	1.15 × 103	7.92 × 104	1.14 × 103	1.49 × 103	8.02 × 104	0.61
OD600 = 0.1 + 0.5M	6.21 × 104	7.32 × 104	5.02 × 104	6.18 × 104	7.32 × 104	5.02 × 104	0.37
OD600 = 0.3 + 0.5M	8.79 × 104	6.66 × 104	1.06 × 103	8.68 × 104	1.06 × 103	6.66 × 104	0.45
OD600 = 0.5 + 0.5M	9.88 × 104	8.75 × 104	3.78 × 104	7.47 × 104	9.88 × 104	3.78 × 104	0.82

. The permeability coefficient of most samples decreases by one order of magnitude compared with that of the untreated sample. Compared with the untreated sample, the average value of the modified samples treated with OD₆₀₀ = 0.1, 0.3, and 0.5 decreased by 45.82%, 24.03%, and 34.59%, respectively. The minimum value decreased by 36.16%, 15.71%, and 51.58%, respectively. The OD₆₀₀ = 0.5 treated sample had the largest differentiation coefficient, but it has the lowest permeability coefficient among the samples. The calcium carbonate particles generated after microbial mineralization can seal up the pores in the particles and the flow channels of the internal fluid. However, with the increase in the bacteria concentration, the permeability coefficient decreased, but the heterogeneity of the sealing treatment increased.

3.3. Hydraulic Erosion Condition of Scoured Slope

Four modified slopes are shown in Figure 5. The loose phosphogypsum particles were solidified into a whole after treatment, and a thin layer of CaCO₃ was formed locally on S3 and S4. The rainfall scouring test lasted for one hour, and Figure 6 shows the morphology of the phosphogypsum slopes after scouring. The particles in S1 were connected by ordinary hydration, it can be seen that a hardened layer of a certain thickness was formed on the surface after hydration. In the early period of the test, only slight erosion occurred on the surface under the protection of the hardened layer. After 30 min, several cracks appeared on the hardened layer, and the non-hydrated loose particles were quickly washed away when the rainfall washed into the lower part of the hardened layer. More cracks were formed in the hardened layer under the scouring force of water and with the accumulation of shedding particles. The whole structure of the S1 slope was quickly destroyed. As for the microbially treated slopes S2, S3, and S4, the matrix was cemented tightly and uniformly without delamination, and while the structures of the slopes were not destroyed, the failure patterns were different. S2 was treated with OD₆₀₀ = 0.1, and the erosion type of slope is sheet erosion. The flakes were splashed, and the spalling area gradually expanded under rainfall scouring, meaning that the erosion surface was smooth. S3 was treated with OD₆₀₀= 0.3, and it showed obvious massive spalling. Erosion pits appeared on the surface and were arranged along the runoff direction with the increasing erosion time, which finally lead to rills. S4 was treated with OD₆₀₀= 0.5, and the solidification was inhomogeneous. The local surface had formed a compact overlay which was slightly damaged, while gullies appeared in the uncovered area, and the width of the gullies gradually widened with the continuous erosion, until the process finally became a form of gully erosion.

Comparing the damage patterns and erosion details of four phosphogypsum slopes after scouring, the different curing effects of microbial treatment and ordinary hydration could be analyzed. The hardened layer of S1 was gypsum dihydrate formed by the hydration of gypsum hemihydrate. The thickness was 6 mm with a low density. The formation of the hardened layer hindered the seepage of water, so the phosphogypsum particles at the lower part of the hardened layer could not completely hydrate and were still in a loose state. When the rainfall scoured into the interior of the slope, an extensive area of loose particles eroded under the scouring, and the whole structure of S1 was quickly destroyed. The phosphogypsum particles in S2, S3, and S4 were modified by the microbial solution and solidified into a whole. The matrix under the scouring layer was tightly cemented. The surface did not appear to have any apparent cracks and overall slip. It indicated that the connection between the particles had changed after the microbial treatment. Under the condition of multiple cycles of different concentrations of bacterial solution, the properties and morphology of each modified slope are different. It could be seen that the particles on S2 were closely connected with a uniform texture, and only weak flake peeling occurred under the scour of water flow. However, both S3 and S4 were scoured out with erosion pits or gullies. A thin layer of calcium carbonate was formed on S3 and S4 due to the incomplete infiltration of the microbial solution. The eroded surface is rough and uneven with bulky particles. Therefore, the erosion resistance of the phosphogypsum slope had been improved by MICP. When the concentration of the cement solution is 0.5 M, the slope modification effect with a bacterial solution concentration of 0.1 is better than that of 0.3 and 0.5.

The cumulative erosion weight and erosion velocity are shown in Figure 7 and Figure 8. For ordinary hydrated slope S1, the erosion velocity remained less than 1 g/min in the initial 30 min, but it increased significantly and peaked at 14.8 g/min at 60 min, and the final cumulative loss of S1 was 399.08 g. The increase in the erosion velocity was due to the loose particles in the lower layer being quickly washed away after the hardened layer was destroyed, and the slope suffered severe collapse and soil erosion. For the S2, S3, and S4 slopes, after microbial treatment, their cumulative erosion weight increased, but the erosion velocity showed a gradual decrease with time. Although the erosion velocity of S2 raised slightly over 20–30 min, the velocity was all less than 5 g/min during the test. The final erosion velocity of the treated slopes was close to 0. The cumulative loss of S2 was 39.34 g, which was the smallest of the three, while the loss of S3 was greater than that of S4. It can be seen from the morphology of the scoured slope that the erosion of S2 only occurred in the surface layer with thin sheet erosion and the smallest peel volume. However, the erosion of S3 produced different depths of erosion pits throughout the entire slope, which resulted in the largest cumulative erosion weight. The partial slope surface of S4 was covered by a mineralized CaCO₃ layer, which hindered scouring erosion, and the internal phosphogypsum was not corroded. The erosion was mainly reflected in the gully erosion, which was not covered by the CaCO₃ layer, so the loss of S4 was less than S3.

3.4. pH of Outflow Solution

The acidity of phosphogypsum can be adjusted by the alkaline modification solution used in MICP. The pH of the outflow solution was measured with time. The result is shown in Figure 9. The pH of S1 gradually decreased in the initial 30 min, but the decrease rate was increased after that. The amount of loss was small in the early period of scouring, and the scouring liquid was not mixed with phosphogypsum particles, so the pH of the outflow was neutral. However, after 30 min, the amount of loss sharply increased and mixed with water to form a soil mixture, meaning that the resulting pH of the outflow solution decreased and showed weak acidity. The pH of modified slopes S2, S3, and S4 was maintained at 7–8. The pH values of the bacteria solution and cementation solution were 8 and 10, respectively. The loss increased with time, but the pH did not show alkalinity. Previous studies had shown that the optimal pH environment for CaCO₃ formation was 6–9 [26,54]. In the process of urea hydrolysis that leads to the production of NH₃ and CO₃²⁻, NH₃ was hydrolyzed to produce OH⁻, providing favorable conditions for the nucleation and deposition of CaCO₃ [55]. It could be inferred that the alkalinity of the solution could neutralize the acidity of phosphogypsum, and the OH⁻ produced by NH₄⁺ hydrolysis was immobilized to form calcium carbonate. The pH of the effluent of treated phosphogypsum is neutral, which eliminates the environmental impact caused by the acidity of untreated phosphogypsum.

3.5. Three-Dimensional Laser Scanning Model and the Roughness of the Phosphogypsum Slope

The surface morphology of the four simulated phosphogypsum slopes after a rainfall erosion test can be scanned by a 3D laser scanner to obtain the point cloud and elevation information of the model. The simulated slope model is presented in Figure 10. As shown in the color bar, dark blue represents the top of the slope. As the elevation decreases, the blue will lighten and then turn dark red. The structure of the S1 was destroyed and the loose particles under the hardened layer were lost, but the hardened layer had not fallen off, so 3D scanning cannot obtain the information relating to the hollow bottom. With this condition, the bottom should be filled with a plane when processing the point cloud, which resulted in the simulated erosion degree, meaning that the erosion volume was smaller than the actual situation.

The elevation information of the simulated surface could be observed from the model analysis. The maximum elevation difference of the S1 model was 28 mm, and the red color shows that the erosion depth had reached the bottom of the tray and that the slope had been completely destroyed, which was consistent with the real condition. The models of the S2, S3, and S4 slopes did not show any red parts, which indicated that no deep erosion occurred in the model. The maximum height difference of S2 was 4 mm, and the color of the surface was almost blue and changed gradually, which shows that the change in the erosion elevation was slight. The maximum elevation difference of S3 was 7 mm. There were many light blue or white blocks on the surface, which were consistent with the erosion pits in the actual slope. The elevation difference of the S4 slope was approximately 6 mm, and the light blue areas were arranged in strips to form gullies.

Roughness measures can quantitatively evaluate the erosion degree of slope. The actual area and projected area of the surface can be obtained by importing the point cloud for analysis. The higher values of RS are associated with the rougher surface. The increase in the actual area indicates that the deeper the erosion depth, the greater the degree of erosion. Each artificial simulated slope is divided into several parts to calculate the roughness and maximum erosion depth, as shown in Figure 11.

It can be seen that the roughness of S1 was 1.87, and the maximum erosion depth was 28.54 mm, which was much greater than that of the microbially modified slope because some part of the slope was collapsed, so the different degree of each part of S1 is large. As for the microbially treated slopes, the erosion degree of S2 was the smallest and most uniform. The roughness of S3 was slightly larger than that of S4, but S4 has the largest erosion depth of the three, and the result was the same as the relationship of the actual measured erosion weight among four slopes.

3.6. Micromorphology of the Scoured Samples and CaCO₃ Crystals

Figure 12 shows the micromorphology of the scoured samples. The hydrated phosphogypsum crystals of S1 had a uniform size distribution and overlapped each other. The microbially treated slopes had different degrees of cementation and different pore structures. S2 appeared to have a large area of cementation, and the generated CaCO₃ filled the macropores while the particles were tightly connected. Lots of macropores in S3 cementation caused some of the particles to be connected through the bridging. However, the single particles were apparent in S4, and the cementation was limited to a few particles with high bonding force. The size of macropores between the particles was larger than otherwise.

The micromorphology shows that the connection mode between particles had changed after microorganism modification, and the bond force was stronger. Previous studies show that bacteria secrete acidic soluble organic matter during the growth process, which will affect the nucleation of crystals and the connection of particles [56,57]. The gypsum dihydrate particles formed by the hydration of gypsum hemihydrate are connected by intermolecular forces, which were easily destroyed under the continuous erosion of rainfall. The bacterial solution contains macromolecular organic matter, whose long chain could make the particles aggregate together. The functional groups in organic molecules such as -COOH and -OH will connect with phosphogypsum particles and calcium ions using hydrogen bonds or ionic bonds. This kind of connection needs a more destructive force to break.

Figure 13 shows the CaCO₃ content of the scoured sample. The mass loss of the S1 sample after soaking in hydrochloric acid might be due to the dissolution of some impurities in phosphogypsum. The mass losses of S2, S3, and S4 far exceeded that of S1. The CaCO₃ content of S2, S3, S4 was 26.25%, 29.34%, 33.81%, respectively. It can be seen from the results that the increase in bacterial concentration induces more CaCO₃. The conclusion here is the same as the conclusions of Okwadha et al. [58]. However, here, the distribution of CaCO₃ content in different parts became more uneven.

In order to explore the structure of CaCO₃ produced under different bacterial concentrations, the bacterial solutions of OD₆₀₀ = 0.1, OD₆₀₀ = 0.3, OD₆₀₀= 0.5, and 0.5 M cementation solution were mixed to induce CaCO₃ in a pure liquid environment, and the crystals were dried in the oven and magnified 1000 times for observation. As shown in Figure 14, although it is not clear whether a single particle in Figure 14 is a single CaCO₃ crystal at this magnification, it can be analyzed as the smallest CaCO₃ unit. Figure 14a shows that the CaCO₃ units formed by OD₆₀₀ = 0.1 were dispersed. A few particles were connected in a linear arrangement. CaCO₃ units in Figure 14b were densely clustered with a bridge structure. In Figure 14c, the size of the units in OD₆₀₀ = 0.5 was similar to that in OD₆₀₀ = 0.3. The particles were first connected by linearly linked and agglomerated flocculation, finally forming an overhead structure. There were many structural pores in the accumulation, which caused the fine particles in the pores to be easily washed away.

Al-thawadi et al. reported that, compared with low-concentration bacterial liquids, bacteria in high-concentration bacterial liquids are more likely to aggregate and flocculate, which induces the formation of larger-sized CaCO₃ clusters [59]. Gandhi et al. reported that the generation of a new crystal nucleus takes precedence over the growth of crystals under the same supersaturation in the solution [60]. Somani et al. had the same conclusion [61]. Snoeyink and Jenkins reported that the degree of supersaturation required for the growth of CaCO₃ crystals is greater than that of the formation of a new crystal nucleus [62]. Therefore, as the concentration of the bacterial solution increased, the increase in CaCO₃ content was caused by the rise in the number of crystals, and the unit size did not become larger, but the structure was formed as a flocculate structure and an overhead structure.

4. Discussion

Previous studies show that the effect of MICP mainly depends on the concentration of Ca²⁺, the concentration of CO₃²⁻, the pH, and the nucleation site [63]. Bacterial activity and urease activity can be affected by high concentrations or low concentrations of Ca²⁺ [37]. Soon et al. found that 0.5 M of cementation solution is suitable for mineralization [64]. It can be seen from Section 3.4 that the alkaline bacteria liquid and cementation solution neutralize the acidity of phosphogypsum, making the pH of the final mineralized environment range from 7 to 8, which does not inhibit the activity of the bacteria and the formation of CaCO₃. Therefore, the improvement of the hydraulic erosion resistance of phosphogypsum depends on the CaCO₃ content and the structure of the phosphogypsum connected to CaCO₃ crystals.

In our study, the water stability and hydraulic erosion resistance of microbially modified phosphogypsum materials are significantly improved compared with that of untreated materials. In the case of spray solidification, the bacterial solution is transported by gravity, and the bacteria will be adsorbed on the pore throat or the surface of the particles under the action of capillary suction and electrostatic force. As the concentration of bacteria increases, more soluble organic matter is carried in the solution, which makes the liquid viscous and increases the infiltration resistance of the liquid, thus reducing the permeability coefficient and improving the water stability. The bacterial density increase leads to the bacterial cells being prone to aggregation and flocculation, and the infiltration of the bacterial solution is limited by the size relationship of the bacteria and pores throats. During microbial treatments, the pores between the phosphogypsum particles are filled with CaCO₃ and the treatment solution, and the bacteria easily accumulate on the surface of the phosphogypsum slope. Compared with the slopes treated by OD₆₀₀ = 0.3 and OD₆₀₀ = 0.5, the migration and distribution of bacteria in the slope treated by OD₆₀₀ = 0.1 are relatively dispersed. CaCO₃ filled in the macropores and formed a uniform and dense solidified layer, decreasing the porosity of phosphogypsum and enhancing the erosion resistance of the slope. However, as for the phosphogypsum slope treated by higher bacterial concentrations, although the CaCO₃ content increases, the particles on the surface were agglomerated during cementation, which caused uneven solidification, and some macropores did not fill effectively, resulting in the infiltration of rainwater and more severe erosion during the rainfall scouring test.

5. Conclusions

In this paper, we used the method of cyclically spraying a bacteria solution with various concentrations (OD₆₀₀ = 0.1, OD₆₀₀ = 0.3, and OD₆₀₀ = 0.5) and a cementation solution (0.5 M) to modify the phosphogypsum material via MICP. We carried out a water shaking test and a permeability test to verify the water stability of the phosphogypsum material. The hydraulic erosion resistance of the modified phosphogypsum slope was studied through a rainfall scouring test, and the pH and heavy metal elements of the outflow were assessed. Three-dimensional laser scanning was used to analyze the micromorphology of the scoured slope. The following conclusions were obtained from the experimental results:

• The water stability and permeability of the sample treated by the microorganism are greatly improved. With the increase in the bacterial concentration, the difference between the content of CaCO₃ in the outermost layer and that in the inner layer gradually increases. The degree of permeability reduction is more uneven.
• Phosphogypsum can be used as solidified material to simulate artificial slopes. The structure of the ordinary hydrated phosphogypsum slope was completely destroyed in the rainfall scouring test. When the slope is treated with OD₆₀₀ = 0.1, the cementation of particles is more uniform, with fewer and smaller macropores, and the slope is eroded by sheet erosion. With the increase in the concentration of the bacterial solution, the bacterial will agglomerate and flocculate, resulting in uneven cementation and more macropores, which lead to erosion pits and gullies under rainfall scouring. The OD₆₀₀ = 0.1 treated slope had the best erosion resistance and the minimum erosion loss.
• Phosphogypsum slope treated by microbial treatment can greatly improve the water erosion resistance and reduce slope erosion. The erosion degree of the microbially treated slopes is much less than that of the untreated slope. When the content of calcium carbonate is high and unevenly distributed, it can lead to deeper erosion depth.
• The structure of calcium carbonate induced by different concentrations of treatment solution is different. Under the same concentration of cementation solution, the increase in the bacterial concentration is associated with higher CaCO₃ content, and the CaCO₃ flocculate to form an overhead structure. The micropores in the structure easily cause the loss of fine particles.
• The pH of the modified phosphogypsum outflow solution is neutral. The heavy metal elements meet the emission standards through microbial action and fixation after microbial treatment. This article proves the effectiveness of microbially modified phosphogypsum and provides a preliminary test basis for the utilization of phosphogypsum material in the reinforcement of slopes.