The mechanism of synergistic hydration performance improvement of the ternary composite cementitious material composed of hemihydrate phosphogypsum, blast furnace slag and quicklime

To realize the resource utilization of solid waste hemihydrate phosphogypsum (HPG) and significantly enhance its mechanical strength and water resistance as a construction material, this study employs the synergistic activation of HPG with blast furnace slag (BFS) and quicklime. A high-strength and highly water-resistant HPG-based ternary composite cementitious material (HPG-CCM) was prepared. Systematic investigations on its leaching toxicity, hydration characteristics, and microstructural evolution were conducted through ICP-OES, XRD, SEM-EDS, and TG-DTG analyses. The results demonstrate that when the HPG-to-BFS mass ratio was 6:4 with 3% externally added quicklime, the 28-day compressive strength and softening coefficient of HPG-CCM reached 52.8 MPa and 0.97, respectively. Under this optimal formulation, the relative proportions of C-S-H gel and ettringite in the reaction system were effectively modulated, resulting in significantly enhanced structural compactness, suppressed volumetric expansion, and improved water resistance. Furthermore, when the external quicklime content exceeded 3%, the leaching concentrations of phosphorus (P) and fluorine (F) remained below the thresholds stipulated by the Chinese standard GB 8978-1996, confirming the material’s exceptional safety. Mechanistic analysis reveals that the hydration process of HPG-CCM comprises three distinct phases: Early Hydration Stage: Quicklime neutralized the inhibitory effects of soluble P and F within HPG on its hydration activity, facilitating the formation of interlocking CaSO₄·2H₂O crystals that established the primary strength framework. Densification Stage: Active silicates and alumina from BFS reacted with quicklime and CaSO₄·2H₂O to generate C-S-H gel and ettringite, which filled interstitial pores within the CaSO₄·2H₂O matrix, yielding a dense microstructure. Encapsulation Stage: Continuous hydration of BFS produced C-S-H gel that progressively enveloped CaSO₄·2H₂O crystals, thereby augmenting both mechanical robustness and water resistance.

1 Introduction

With the increasing adoption of the new hemihydrate-based wet-process phosphoric acid (WPA) technology, a distinct type of solid waste—hemihydrate phosphogypsum (HPG)—is generated as a direct by-product of this emerging route. In this process, HPG (CaSO₄·0.5H₂O) is precipitated within the CaSO₄–H₃PO₄–H₂SO₄–H₂O equilibrium system by controlling the sulfuric acid dosage, maintaining the reaction temperature at 90 °C–110 °C (Jia et al., 2021), and adjusting the phosphoric acid concentration to approximately 45% (see Equation 1). These specific operating conditions fundamentally differentiate the hemihydrate WPA route from the conventional dihydrate process and result in HPG exhibiting unique phase composition and impurity characteristics.

${\text{Ca}}_5{\left({\text{PO}}_4\right)}_3\mathrm{F}+10{\mathrm{H}}_2{\text{SO}}_4+5{\mathrm{H}}_2\mathrm{O}\to 6{\mathrm{H}}_3{\text{PO}}_4+10{\text{CaSO}}_4·0.5{\mathrm{H}}_2\mathrm{O}+2\text{HF}\uparrow (1)$

With the increasing adoption of the hemihydrate process in WPA production, the resulting by-product—HPG—has attracted considerable attention due to its inherent potential cementitious activity (Jiang et al., 2018). HPG can hydrate and harden at ambient temperature, offering the possibility of partially replacing conventional cementitious materials such as Portland cement (Wang et al., 2025a). However, its practical application is hindered by several critical drawbacks. First, the surface and crystal lattice of HPG are enriched with water-soluble P₂O₅ and F⁻ impurities, which can be readily released during prolonged stockpiling or under rainfall, posing potential threats to both surface and groundwater (Jiang et al., 2024). Second, its hydration products are predominantly dihydrate gypsum, which exhibits a loose microstructure and high solubility in water, resulting in poor water resistance and insufficient mechanical performance for engineering applications—its softening coefficient typically ranges from only 0.3 to 0.5 (Fornés et al., 2021; Tao et al., 2024; Wang et al., 2025b). Given the broad demand for gypsum-based binders in the construction industry, the inadequate water resistance of HPG significantly compromises its long-term stability in humid or water-exposed environments, thereby restricting its large-scale implementation in building and infrastructure projects. Consequently, enhancing the water resistance of HPG is not only the key technical bottleneck in its utilization as a construction material, but also a critical step toward advancing the resource utilization of phosphogypsum and mitigating its associated environmental risks.

To enhance the water resistance of gypsum-based binders, recent studies have explored the incorporation of hydraulic binders (e.g., blast furnace slag and fly ash) in combination with alkaline activators to promote the formation of C-S-H gels and minor amounts of ettringite crystals, thereby improving material densification and long-term water stability (Zhang et al., 2022; Chen P. et al., 2022). Blast furnace slag (BFS), a latent hydraulic material rich in amorphous SiO₂ and Al₂O₃, can exhibit markedly increased reactivity under alkaline conditions, yielding a variety of cementitious hydrates (Wang et al., 2022; Zheng et al., 2023). Previous investigations have reported promising results by coupling slag with carbide slag (Zhou et al., 2024) or quicklime (Park et al., 2016) to co-utilize with PG. However, the fundamental mechanism in these systems is that slag-generated C-S-H phases encapsulate gypsum particles, while PG itself does not actively participate in the cementitious reaction. As a result, the gypsum content in such composites remains constrained, limiting the potential for high-volume resource utilization of PG.

In contrast, HPG possesses inherent cementitious activity, and activating its latent reactivity could overcome the limitation of its treatment as an “inert aggregate.” However, most existing studies on HPG have primarily focused on improving its mechanical strength, with limited attention given to enhancing its water resistance, understanding its environmental behavior, or elucidating the underlying synergistic reaction mechanisms. Moreover, previous reports indicate that improper use of alkaline activators can lead to excessive formation of ettringite phases, resulting in deleterious volume expansion and cracking (Lin et al., 2024). Consequently, the key research challenge lies in developing a composite system that can both stimulate the reactivity of HPG and regulate the structure of its hydration products, thereby improving overall water resistance and ensuring environmental stability. Although extensively studied in supersulfated cement systems, the synergistic activation behavior of blast furnace slag and quicklime in hemihydrate phosphogypsum remains insufficiently understood.

Given the above, this study proposes a composite material in which quicklime and BFS are synergistically combined with HPG. Quicklime serves to create an alkaline environment that neutralizes soluble P and F impurities while simultaneously promoting the activation of BFS. This approach enables regulation of the relative formation of C-S-H gels and ettringite phases within the reaction system, thereby enhancing microstructural densification, suppressing volumetric expansion, and improving both mechanical strength and water resistance. The investigation begins with an evaluation of macroscopic properties—including uniaxial compressive strength (UCS), softening coefficient, expansion rate, and ionic leaching—followed by the characterization of HPG’s reactive evolution using electrical conductivity and hydration heat measurements. Furthermore, a combination of X-ray diffraction (XRD), thermogravimetric and differential thermogravimetric (TG-DTG), Fourier transform infrared spectroscopy (FTIR), pore size distribution analysis, and scanning electron microscopy-energy dispersive X-ray spectroscopy (SEM-EDS) is employed to elucidate the synergistic regulation mechanisms by which quicklime and BFS influence the hydration products and microstructure of HPG. The findings aim to provide a novel pathway for the high-volume incorporation of industrial solid wastes into high-water resistance cementitious materials.

2 Materials and methods

2.1 Raw materials

The raw materials employed in this study comprised HPG, BFS, quicklime, and water.

The HPG sample was sourced from the production line of Guizhou Chanhen Chemical Co., Ltd. Its particle size ranged from 0.40 to 210.50 μm, with a median particle size (d₅₀) of 53.69 μm and a specific surface area of 154.27 m²/kg. The primary chemical constituents were SO₃, CaO, and SiO₂ (see Table 1), accompanied by minor amounts of environmentally sensitive impurities such as F and P₂O₅. The crystalline water content was determined to be 5.46%. XRD analysis indicated that the dominant phase was CaSO₄·0.5H₂O, while SEM images revealed polycrystalline particles with smooth surfaces and dense structures (see Figure 1).

Table 1

Table 1. Chemical components of raw materials.

Figure 1

Figure 1. Microstructural and phase composition analysis of HPG: (a) SEM image; (b) XRD pattern.

The BFS was obtained from a steel plant in Yunnan Province and was typical water-quenched slag. Its major chemical components were CaO, SiO₂, and Al₂O₃ (see Table 1), meeting the activity requirements for S95-grade slag powder (Xia et al., 2025). The specific surface area was 481.30 m²/kg. As shown in Figure 2, the particle size ranged from 0.51 to 100.80 μm, with a d₅₀ of 17.55 μm and a relatively narrow particle size distribution, meeting the fineness requirements for cementitious activity.

Figure 2

Figure 2. Particle size distribution of raw materials.

Quicklime was procured from Guizhou Chuandong Chemical Co., Ltd., and was of industrial grade. The primary active component was CaO, with an available CaO content of 95% (see Table 1). It had a specific surface area of 544.30 m²/kg, a particle size range of 0.21–89.16 μm, and a d₅₀ of 8.33 μm, indicating high reactivity and suitable fineness for cementitious applications.

The mixing water used in this study was local tap water, employed for slurry preparation.

2.2 Experimental design

The samples were prepared according to the mix proportions listed in Table 2, with HPG, BFS, and quicklime as the primary constituents. Predetermined masses of HPG, BFS, and quicklime were blended in the proportions shown in Table 2 to produce the HPG-based composite cementitious materials (HPG-CCM). The HPG mass fractions were set at 100%, 80%, 70%, and 60%, corresponding to BFS contents of 0%, 20%, 30%, and 40%, respectively. Quicklime was used as an additional component, with contents of 0%, 1%, 3%, 5%, and 7%, calculated by mass percentage of the total binder. The resulting mixtures were designated N1-N13, with N1 serving as the control group to evaluate the effects of BFS and quicklime on the water resistance of HPG-CCM. The blended powders were mixed with water at a mass concentration of 50% and stirred for 1 min to form a homogeneous slurry, which was then cast into three-gang steel molds measuring 40 mm × 40 mm × 160 mm. Following casting, the specimens were cured in a standard curing chamber (20 °C ± 2 °C, relative humidity≥95%) for 24 h, demolded, and then immersed in local tap water at ambient temperature (20 °C ± 2 °C) for curing until the designated ages of 1, 3, 7, 14, and 28 days. At each curing age, uniaxial compressive strength, softening coefficient, and expansion rate tests were performed.

Table 2

Table 2. Material proportion.

Based on the results of these tests, the optimal mix proportions of the three materials were selected for further experiments, including toxicity leaching tests and microstructural characterization. The overall experimental design is illustrated in Figure 3.

Figure 3

Figure 3. The specific experimental process.

2.3 Test methods

2.3.1 Setting time test

At a mass concentration of 50%, the raw materials were precisely weighed according to the proportions outlined in Table 2, thoroughly mixed with water, and subsequently poured into a circular mold. In accordance with the Chinese standard GB/T 17669.4-1999, the initial and final setting times of the samples were determined using a Vicat apparatus.

2.3.2 UCS test

Specimens measuring 40 mm × 40 mm × 160 mm were cured in a standard curing chamber for durations of 1, 3, 7, 14, and 28 days. At each curing age, three specimens were tested to ensure the representativeness of the results. Compressive strength tests were performed using a cement concrete compression testing machine (YAW-300D, China) at a loading rate of 0.6 kN/s. The average value of the three measurements was taken as the compressive strength.

2.3.3 Softening coefficient test

Specimens cured for up to 28 days were removed and immersed in water for 24 h to ensure full saturation before testing the saturated compressive strength, following the standard testing procedures outlined in JCT 698-2010. Another set of specimens was dried in an oven (DHG-9023A, China) at 40 °C–45 °C for 24 h to achieve complete dryness. After cooling to room temperature, their dry compressive strength was measured. The softening coefficient was calculated according to Equation 2. All tests were performed in triplicate to ensure data reliability.

$K_{\mathrm{s}}=\frac{{\sigma }_{csat}}{{\sigma }_{cdry}}(2)$

Where K_s is the softening coefficient, σ_csat is the saturated compressive strength, and σ_cdry is the dry compressive strength.

2.3.4 Expansion rate test

The thoroughly mixed slurry was cast into three-gang molds measuring 40 mm × 40 mm × 160 mm. After curing at a constant temperature of 20 °C ± 2 °C for 1 day, the specimens were demolded and their initial lengths (L₀) were measured using a vernier caliper with a precision of 0.01 mm. The specimens were then immersed in water and cured until 28 days. After wiping the surface moisture, the length at 28 days (L_T) was measured again. The expansion rate of the specimens was calculated according to Equation 3.

$\epsilon =\frac{L_T-L_0}{L_0}(3)$

Where L₀ is the initial length and L_T is the specimen length after 28 days of curing.

2.3.5 Toxic leaching test

The concentrations of leachable toxic ions from samples at different hydration ages were determined following the HJ 557–2010 standard. Samples were crushed into fragments smaller than 3 mm in diameter and soaked in deionized water at a solid-to-liquid ratio of 1:10. The mixtures were placed on a horizontal shaker (HZQ-3, China) and agitated at 110 rpm for 8 h, followed by a static period of 16 h. The supernatant was then filtered through a 0.45 μm microporous membrane to obtain the leachate, which was subsequently analyzed using inductively coupled plasma optical emission spectrometry (ICP-OES).

2.3.6 XRD

XRD patterns were obtained using a Bruker D8 Advance diffractometer (Germany) with a scanning range of 5°–60° and a scan rate of 10°/min. All samples were ground into fine powders to ensure homogeneity. Approximately 50 mg of each sample was used. During the measurements, the temperature of both the samples and the instrument was maintained at 25 °C ± 2 °C to minimize peak shifts caused by temperature fluctuations.

2.3.7 TG-DTC

Thermal performance analysis was conducted using a simultaneous thermal analyzer (STA 449F3, Netzsch, Germany) to assess the samples at different hydration ages. The measurements were carried out under a nitrogen atmosphere, with a heating rate of 10 °C/min from room temperature to 1000 °C. During the experiment, the flow rate of the atmosphere was maintained at 30 mL/min, and the sample mass was kept below 10 mg. Prior to each test, the equipment was calibrated to ensure the accuracy of the thermogravimetric results, with the measurement error controlled within ±2%.

2.3.8 FTIR

FTIR analysis was performed using a Nexus spectrometer (Thermo Fisher Scientific, United States) to investigate the bonding characteristics of the hydration products. Thin pellets were prepared by pressing a mixture of sample powder and KBr at a ratio of 1:100. Spectra were recorded over the range of 4000 to 400 cm⁻¹. Each sample was measured in triplicate, and the average spectrum was used to enhance the reliability of the results. The spectral resolution error was controlled within ±1 cm⁻¹.

2.3.9 Conductivity

The dissolution of cementitious material particles was evaluated by monitoring the rate of increase in electrical conductivity, while a subsequent decrease in conductivity served as an initial indication of hydration product formation (Wang et al., 2024). Under a solid-to-liquid ratio of 1:10, both the control group (N1) and experimental samples were directly mixed into beakers for measurement. Electrical conductivity values were automatically recorded using a conductivity meter (YSI 3200, United States). All experiments were conducted in a temperature-controlled environment, with temperature variations maintained within ±1 °C.

2.3.10 Calorimetry

The hydration heat of the samples was measured using an isothermal microcalorimeter (TAM Air C80, SETARAM, France). Data were recorded at 1 min intervals to determine the hydration rate, which was further used to calculate the relationship between heat evolution and hydration time. Tests were conducted under controlled conditions at a temperature of 20 °C ± 2 °C, relative humidity of 60% ± 5%, and a water-to-cement ratio of 0.5. The measurement error was maintained within ±5%.

2.3.11 Pore size measurements

The pore structure of the hardened specimens was analyzed using a Tristar II 3030 nitrogen adsorption analyzer (Micromeritics, United States). Prior to each measurement, samples were degassed under vacuum at 60 °C for 6 h. During testing, the temperature and humidity of the environment were carefully controlled and maintained within a ±3% margin of error to ensure data reliability.

2.3.12 SEM-EDS

Prior to SEM analysis, the dried samples were sputter-coated with a thin layer of gold to enhance electrical conductivity. Microstructural characterization was performed using a JSM-6510 scanning electron microscope operated at an accelerating voltage of 20 kV and a resolution of 3 nm, enabling detailed observation of morphological features and microstructural details. The SEM system was equipped with EDS for elemental analysis, allowing simultaneous evaluation of compositional changes alongside morphological variations.

3 Results and discussion

3.1 Effects of BFS and quicklime on the water resistance of HPG

3.1.1 Setting time analysis

Figure 4 presents the results of the setting time tests for HPG-CCM mixtures with different proportions. It can be observed that the initial setting time for the sample without quicklime and BFS was 239 min, while the final setting time was 321 min. When the HPG:BFS ratio was 8:2, the setting times gradually decreased with increasing quicklime content. The initial setting time reduced from 239 min (without quicklime) to 141 min, and the final setting time decreased from 321 min to 181 min (see Figure 4a). This reduction is attributed to the increase in alkalinity with higher quicklime content, which accelerated the neutralization of water-soluble phosphorus (P) and fluorine (F) impurities adsorbed on HPG, thereby promoting the hydration reaction of HPG and shortening the setting times. Under the condition of 3% quicklime content, when the HPG:BFS ratio was reduced from 8:2 to 6:4, the initial setting time of HPG-CCM decreased from 187 min to 133 min, and the final setting time decreased from 242 min to 149 min (see Figure 4c). This indicates that with higher HPG content, more quicklime is required to neutralize the P and F impurities. Insufficient quicklime content may lead to excessively long setting times.

Figure 4

Figure 4. Setting time of HPG-CCM at different quicklime contents. (a) HPG:BFS(8:2), (b) HPG:BFS(7:3), (c) HPG:BFS(6:4).

3.1.2 Mechanical performance analysis

Figure 5 presents the compressive strength results of HPG-CCM at different curing ages. As shown in Figure 5a, at a hydration age of 1 day, with a quicklime content of 1% and an HPG-to-BFS ratio of 6:4, the compressive strength of HPG-CCM reached 2.37 MPa. Increasing the quicklime content initially enhanced the mechanical strength, followed by a decline at higher dosages. This behavior is attributed to the presence of soluble P- and F-containing impurities in HPG, which suppress the hydration of HPG and result in the low strength observed in the control group. After quicklime addition, these impurities react with CaO to form insoluble P- and F-bearing precipitates, thereby reducing their inhibitory effect and promoting the hydration and hardening of HPG.

Figure 5

Figure 5. Compressive strength of HPG-CCM at different curing ages: (a) 1 d, (b) 3 d, (c) 7 d, (d) 14 d, (e) 28 d, and (f) strength development across curing ages at the optimal mix ratio.

It should be noted that both Ca₃(PO₄)₂ and CaHPO₄ are possible precipitation products under alkaline conditions. Specifically, within the pH range of approximately 9.8–14, the dominant species are CaHPO₄ and Ca₃(PO₄)₂, whereas at pH values above 14, Ca₃(PO₄)₂ predominates. Therefore, it can be inferred that the system mainly contains Ca₃(PO₄)₂, CaHPO₄, and CaF₂ (see Equations 4–7) (Xu et al., 2023; Wu et al., 2022). However, the relative proportions of these two phases remain undetermined and require further detailed structural analysis. In this study, the identification of the phosphate precipitates was primarily inferred from thermodynamic considerations and the alkaline pH environment generated by CaO addition. Direct structural characterization of the precipitates has not yet been conducted, and thus the specific phase composition cannot be conclusively determined at this stage. This point will be further clarified and supplemented in future work. Excessive CaO addition led to the presence of unreacted quicklime during curing, which may induce excessive ettringite formation and result in specimen expansion and a subsequent decrease in compressive strength.

$\text{CaO}+{\mathrm{H}}_2\mathrm{O}\to \text{Ca}{\left(\text{OH}\right)}_2(4)$

$3\text{Ca}{\left(\text{OH}\right)}_2+2{\mathrm{H}}_3{\text{PO}}_4\to {\text{Ca}}_3{\left({\text{PO}}_4\right)}_2\downarrow +6{\mathrm{H}}_2\mathrm{O}(5)$

$\text{Ca}{\left(\text{OH}\right)}_2+{\mathrm{H}}_3{\text{PO}}_4\to {\text{CaHPO}}_4\downarrow +2{\mathrm{H}}_2\mathrm{O}(6)$

$2{\mathrm{F}}^{-}+{\text{Ca}}^{2+}\to {\text{CaF}}_2\downarrow (7)$

Furthermore, as the proportion of HPG relative to BFS increases, the mechanical strength of HPG-CCM exhibits a similar trend of initially increasing and then decreasing with different quicklime content. This observation indicates that during the early hydration stage, the mechanical strength of HPG-CCM primarily derives from the hydration and hardening of HPG; thus, higher HPG content corresponds to greater strength, while BFS acts mainly as an inert aggregate at this stage.

When the hydration age is extended to 3 days, the results shown in Figure 5b reveal that within the HPG-to-BFS ratio range of 6:4 to 8:2, the mechanical strength of HPG-CCM again initially increases and then decreases with increasing quicklime content, reaching a maximum compressive strength at 3% quicklime addition. This phenomenon suggests that with prolonged hydration, samples with initially lower HPG content gradually catch up in strength to those with higher HPG content after 3 days. This is attributed to the alkali activation of BFS during this period, which significantly enhances the hydration process.

When the hydration age was further extended to 7 days, the results in Figure 5c indicate that at an HPG-to-BFS ratio of 6:4, the increase in mechanical strength was significantly greater than that observed for the 8:2 ratio, with the peak of the strength surface shifting towards the left. At a quicklime content of 3%, the compressive strength increased to 35.55 MPa. This suggests that a higher proportion of BFS facilitates the formation of more strength-bearing hydration products, markedly enhancing the mid-term mechanical performance of HPG-CCM.

Figures 5d,e display the mechanical strength of HPG-CCM after 14 and 28 days of hydration, respectively, within an HPG-to-BFS ratio range of 6:4 to 8:2 and quicklime content between 1% and 5%. The peak strength regions on the response surfaces are located on the left side. Compared with the early hydration results in Figure 5a, the strength peak shifts from the right side at early stages to the left side at later stages. This indicates that the early strength developed primarily through HPG hydration is progressively supplemented by strength contributions from the hydration of BFS and quicklime during later stages. Under the experimental conditions, the optimal HPG-to-BFS mass ratio was determined to be 6:4. Subsequent investigations focused primarily on samples N10∼N13.

At an HPG-to-BFS ratio of 6:4, the compressive strength development of HPG-CCM across curing ages is presented in Figure 5f. The results demonstrate that increasing quicklime content initially enhances compressive strength, which subsequently declines at higher dosages. A quicklime content of 3% yielded a peak 28-day compressive strength of 52.8 MPa. This behavior can be attributed to the effective neutralization of deleterious impurities such as P and F by an appropriate amount of quicklime, mitigating their inhibitory effects on the hydration reaction and promoting the hardening process of HPG-CCM. Moreover, with prolonged curing, quicklime further activates the latent pozzolanic reactivity of BFS under alkaline conditions, generating additional hydration products and thereby continuously improving the mechanical properties of the sample (Park et al., 2016; Antonio et al., 2010). Therefore, an HPG-to-BFS mass ratio of 6:4 combined with 3% quicklime addition provides an optimal formulation for maximizing the mechanical performance of HPG-CCM.

3.1.3 Softening coefficient analysis

Figure 6 shows the variation of the softening coefficient of HPG-CCM under different mix proportions. Without modification by quicklime and BFS, the softening coefficient of HPG was only 0.37, indicating low strength and poor water resistance. This behavior is primarily attributed to the slow hydration rate of HPG, where most particles fail to fully react with water to form an interconnected network of CaSO₄·2H₂O crystals. Instead, the structure is maintained predominantly by weak physical forces such as van der Waals interactions, resulting in significantly compromised mechanical properties and water resistance of the hydrated matrix (Gao et al., 2022). When the HPG-to-BFS mass ratio was 6:4 with 3% quicklime addition, the softening coefficient of HPG-CCM reached a maximum value of 0.97, representing a 162.16% improvement compared to the unmodified sample, demonstrating excellent water resistance.

Figure 6

Figure 6. Softening coefficient of HPG-CCM at different quicklime contents.

As quicklime content increased, the softening coefficient of HPG-CCM exhibited an initial rise followed by a decline. Under the synergistic effect of an alkaline environment and SO₄^2- ions, BFS was effectively activated, producing hydration products such as C-S-H gel and ettringite (Zhang et al., 2017; Yoon et al., 2023). These products fill the pores within the HPG-CCM matrix, enhancing its compactness and thereby significantly improving the softening coefficient. However, when quicklime content exceeded 5%, the softening coefficient decreased markedly. Consistent with the observations in Section 3.1.2, this decline is attributed to excessive quicklime, which induces overproduction of ettringite within the system. This leads to expansion, increased internal porosity and cracking, weakening the structural integrity of the hardened body and subsequently reducing both saturated compressive strength and water resistance.

3.1.4 Volume stability analysis

Figure 7 shows the variation in expansion rate of HPG-CCM under different quicklime contents. The results demonstrate that the expansion rate initially increases and then decreases with increasing quicklime dosage. When quicklime content ranges from 1% to 3%, HPG-CCM maintains high compressive strength while keeping the expansion rate below 1%, indicating an optimal balance between mechanical performance and volumetric stability within this range. However, when quicklime content is increased to 5%, and the HPG-to-BFS mass ratios are 6:4, 7:3, and 8:2, the expansion rates rise sharply to 3.38%, 5.63%, and 2.35%, respectively, with the 7:3 ratio exhibiting the highest expansion. This behavior is attributed to the excessive quicklime continuously reacting with CaSO₄·2H₂O in the alkaline environment, further activating the aluminate phases in the glassy matrix of BFS and generating a large quantity of needle-like ettringite crystals (Lin et al., 2024). The volumetric expansion of ettringite induces internal stress accumulation within the hardened matrix (Yang et al., 2022; Qi et al., 2023), ultimately causing significant expansion and cracking. Therefore, excessive quicklime addition compromises the volumetric stability of the HPG-CCM hydration matrix, rendering it unsuitable for practical engineering applications.

Figure 7

Figure 7. Volume expansion rates of HPG-CCM at different quicklime dosages.

3.2 Environmental impact assessment of enhanced water resistance in HPG-CCM

Figure 8 presents the leaching concentrations of harmful P and F ions from HPG-CCM samples at 28 days of curing, compared against the regulatory limits set by GB 8978-1996 (P ≤ 0.5 mg/L, F ≤ 10 mg/L). The control sample (N1) exhibited extremely high leachate concentrations of P and F, measuring 239 mg/L and 110.02 mg/L, respectively, exceeding the permissible limits by approximately 478 and 11 times (see Figures 8a,b). These findings indicate that untreated HPG releases substantial quantities of hazardous ions, posing a significant threat to soil and groundwater systems.

Figure 8

Figure 8. Leaching concentrations of P and F ions in HPG-CCM at different curing time. (a) P, (b) F.

Upon the addition of 1% quicklime (sample N10), the leaching concentrations of P and F decreased markedly to 1 mg/L and 10 mg/L, respectively. While the fluoride concentration reached the regulatory threshold, the phosphorus concentration remained slightly above the allowable limit, suggesting that this low quicklime dosage offers partial but insufficient stabilization. Increasing the quicklime content to 3%, 5%, and 7% further reduced the P concentrations to 0.05 mg/L, 0.02 mg/L, and 0.01 mg/L, respectively—well below the regulatory limit of 0.5 mg/L—demonstrating excellent phosphorus immobilization. Correspondingly, fluoride concentrations decreased to 0.78 mg/L, 0.64 mg/L, and 0.51 mg/L, all significantly lower than the 10 mg/L limit, indicating effective fluoride mitigation.

3.3 Mechanism analysis of water resistance improvement

3.3.1 Hydration product evolution

3.3.1.1 Mineralogical composition analysis

Figure 9 presents the XRD patterns of HPG-CCM with different mix proportions after 28 days of curing. The primary diffraction peaks in sample N1 correspond to CaSO₄·0.5H₂O, CaSO₄·2H₂O, and SiO₂. The presence of CaSO₄·0.5H₂O indicates incomplete hydration of HPG under these conditions. Following the incorporation of quicklime and BFS, sample N10 exhibits diffraction peaks for ettringite, CaSO₄·2H₂O, and CaSO₄·0.5H₂O, suggesting that at a quicklime content of 1%, some CaSO₄·0.5H₂O remains unconverted to CaSO₄·2H₂O. The incomplete hydration can be primarily attributed to two factors: insufficient alkalinity to fully neutralize the soluble P and F ions in HPG, which inhibits hydration, and the encapsulation of unreacted HPG particles by formed C-S-H gel, which impedes further hydration.

Figure 9

Figure 9. XRD patterns of HPG-CCM with different mix proportions after 28 days of curing.

When the quicklime content is increased to 3%, the CaSO₄·0.5H₂O diffraction peak disappears in sample N11, indicating that HPG has largely fully hydrated to CaSO₄·2H₂O. Under sufficiently alkaline conditions, Ca²⁺ and OH⁻ ions facilitate the migration of impurity P and F ions from the surface or lattice of HPG particles, forming insoluble precipitates with Ca²⁺, which continuously promotes HPG hydration. Comparing samples N10 and N12 reveals similar ettringite peak intensities.

3.3.1.2 TG-DTG analysis

The TG-DTG curves of HPG-CCM after 28 days of curing are presented in Figure 10. As shown, the unmodified HPG-CCM (without quicklime and BFS) exhibits a distinct weight loss peak around 148.5 °C, corresponding to the dehydration of CaSO₄·2H₂O formed during HPG hydration. In contrast, HPG-CCM samples containing quicklime and BFS display a new weight loss peak near 118.63 °C. This characteristic peak is primarily attributed to the synergistic effect of the supplementary materials, which not only accelerate the hydration of HPG but also promote the formation of new mineral phases, specifically ettringite, whose dehydration temperature is centered around 118 °C. This conclusion is further supported by the characteristic ettringite diffraction peak observed at approximately 9.1° in the XRD patterns (Qu et al., 2017; Wei et al., 2022).

Figure 10

Figure 10. TG-DTG curves of HPG-CCM with different mix proportions after 28 days of curing. (a) DTG, (b) Weight.

The mass loss associated with hydration products for HPG-CCM at different mix proportions is summarized in Table 3. With increasing quicklime content, the weight loss attributable to CaSO₄·2H₂O shows a continuous upward trend, indicating that quicklime significantly enhances the conversion of HPG to CaSO₄·2H₂O. However, the weight loss related to ettringite exhibits a non-linear pattern, increasing initially and then decreasing. At low quicklime content (sample N10, 1% quicklime), the alkalinity of the system is insufficient, limiting both the hydrolysis rate of BFS and the hydration degree of HPG, thereby suppressing ettringite formation. Conversely, at excessive quicklime content (sample N13, 7% quicklime), although the hydration rate of HPG-CCM is markedly increased, the rapid formation of ettringite tends to coat BFS particles, hindering further hydrolysis reactions and resulting in a decline in ettringite formation at later stages. Overall, sample N11 exhibits the highest ettringite-associated weight loss, indicating that an optimal proportion of quicklime and BFS exerts a synergistic effect that maximizes the formation of both key hydration products.

Table 3

Table 3. Mass loss of hydration products in samples with different mix proportions.

3.3.1.3 FTIR analysis

Figure 11 presents the FTIR spectra of hydration products in HPG-CCM with different mix proportions. The FTIR analysis was employed to identify the functional groups and mineral phase transformations occurring during hydration. The absorption bands at 3,553 cm⁻¹ and 3,408 cm⁻¹ correspond to the stretching vibrations of O-H groups. In the quicklime-free N1 sample, these peaks are relatively weak. However, their intensities increase markedly upon quicklime incorporation, indicating that the alkaline environment enhances the reaction between HPG and water, leading to the formation of CaSO₄·2H₂O. This observation is consistent with the XRD results presented in Section 3.3.1.1.

Figure 11

Figure 11. FTIR spectra of HPG-CCM with different mix proportions after 28 days of curing.

The bands at 1683 cm⁻¹ and 1619 cm⁻¹ are attributed to the bending vibrations of H-O-H in gypsum (Nishiyama and Oinuma, 1977). In addition, the absorption peaks at 1150 cm⁻¹ and 1116 cm⁻¹ are associated with the V₃ vibrations of SO₄^2-, consistent with the sulfate-related signals originating from HPG (Xu et al., 2016). The peaks at 672 cm⁻¹ and 598 cm⁻¹ correspond to the symmetric and asymmetric vibrations of Al-OH groups in ettringite, further confirming the presence of ettringite (Trezza and Lavat, 2001; Alvarez-Ayuso and Nugteren, 2005; Qian et al., 2008). Meanwhile, the peak at 468 cm⁻¹ is assigned to the Si-O vibrations in SiO₄ tetrahedra, corroborating the formation of C-S-H gel and compensating for the limited sensitivity of XRD in detecting the fibrous C-S-H phase (Re et al., 2023).

When the quicklime dosage reaches 7%, the characteristic peaks of C-S-H gel progressively weaken and eventually disappear, suggesting that excessive quicklime may hinder the hydration of BFS, thereby suppressing C-S-H formation. Overall, the FTIR findings align well with the XRD results, confirming that the primary hydration products of HPG-CCM are CaSO₄·2H₂O, ettringite, and C-S-H gel.

3.3.2 Microstructural evolution

3.3.2.1 Microstructure analysis

Figure 12 presents the microstructural morphology of HPG-CCM after 28 days of hydration. In the control group (Figure 12a), the hydration products primarily consist of irregular block-like CaSO₄·0.5H₂O and a small amount of plate-like CaSO₄·2H₂O. The relatively low content of hydrated CaSO₄·2H₂O and the absence of effective interlocking between crystals indicate incomplete hydration, which accounts for the low compressive strength observed at 28 days.

Figure 12

Figure 12. SEM images of HPG-CCM after 28 days of curing and the EDS mapping analysis results of sample N11. (a) quicklime:0%, (b) quicklime:1%, (c) quicklime:3%, (d) quicklime:5%, (e) quicklime:7%, (f) Element scatter plot.

When 1–3 wt% quicklime was incorporated (Figures 12b,c), the HPG-CCM samples exhibited a comparatively dense microstructure, characterized by the coexistence of C-S-H gel and CaSO₄·2H₂O. The abundant C-S-H gel encapsulated needle-like ettringite crystals and densely filled the pores, resulting in the near absence of visible ettringite in samples N10 and N11. This microstructural densification effectively hindered water ingress into the CaSO₄·2H₂O crystals, thereby enhancing both compressive strength and water resistance. However, when the quicklime content exceeded 5 wt% (Figures 12d,e), the hydration products clearly contained blocky CaSO₄·2H₂O, needle-like ettringite, and only small amounts of C-S-H gel. Excessive quicklime led to an overly alkaline environment, which inhibited the formation of C-S-H gel, resulting in a less compact product structure. Furthermore, the high alkalinity promoted ettringite formation, inducing volumetric expansion and severe cracking in the hydration products. These microstructural defects caused a marked reduction in both compressive strength and water resistance, consistent with the macroscopic observations described in Section 3.1.

For the sample with an HPG-to-BFS mass ratio of 6:4 and 3 wt% quicklime, the EDS elemental mapping after 28 days of hydration is shown in Figure 12f. The analysis revealed that needle-like regions exhibited an Al:Ca molar ratio of 0.3–0.4, matching the stoichiometric composition of ettringite (Zhou et al., 2024), confirming that CaSO₄·2H₂O, quicklime, and BFS reacted chemically to form ettringite. These elongated ettringite crystals interlocked within the CaSO₄·2H₂O framework, promoting microstructural densification and contributing positively to the mechanical performance of HPG-CCM. In addition, flocculent gel phases displayed Si:Ca molar ratios of 0.4–0.6, consistent with the typical stoichiometry of C-S-H gel (Xia et al., 2025). This indicates that, under alkaline conditions, SiO₂ from BFS participated in the hydration reaction to generate C-S-H gel. These gels uniformly coated the CaSO₄·2H₂O crystalline framework, further enhancing skeletal compactness and significantly improving both the mechanical strength and water resistance of HPG-CCM.

3.3.2.2 Pore structure analysis

Figure 13 shows the pore size distribution characteristics of HPG-CCM after 28 days of curing with different mix proportions. As shown in Figure 13a, when the quicklime content is 1%, the pore size distribution of sample N10 is primarily concentrated between 3–4 nm, with a predominant presence of aerogel pores (less than 4.5 nm). This is mainly attributed to the insufficient quicklime content, which results in the hardened matrix being primarily composed of interlocked CaSO₄·2H₂O crystals, with only a small amount of C-S-H gel and ettringite crystals filling the pores, leading to a concentration of smaller pore sizes.

Figure 13

Figure 13. Pore size distribution of HPG-CCM with different mix proportions after 28 days of curing: (a) pore size distribution; (b) cumulative pore volume.

When the quicklime content is increased to 3% (sample N11), the pore size distribution shifts to the 2.5–4 nm range, indicating a further filling of the pores by C-S-H gel and ettringite crystals generated during the hydration process. This shift promotes an increase in gel pore volume, thereby improving the densification and compressive strength of the HPG-CCM matrix. However, when the quicklime content is further increased to 5% and 7% (samples N12 and N13), the peak of the pore size distribution shifts significantly toward larger pore sizes, with a concentration around 25 nm. This change can be attributed to the high alkalinity induced by the excessive quicklime content and the combined effect of CaSO₄·2H₂O, which significantly accelerates the BFS hydration reaction, resulting in an overproduction of ettringite crystals. Given that ettringite crystals are significantly larger than C-S-H gel, their excess generation in the microstructure leads to internal expansion (Yuichiro et al., 2021; Zhe et al., 2025), forming a larger number of harmful macropores and causing a decline in mechanical strength and softening coefficient. As seen in Figure 13b, the cumulative pore volume of sample N10 is the highest, reaching 0.044 cm³/g. With increasing quicklime content, the cumulative pore volume of the HPG-CCM hydration products follows a “decrease first, then increase” trend, which is highly consistent with the changes in mechanical strength shown in Figure 5.

3.3.3 Hydration reaction process analysis

3.3.3.1 Conductivity analysis

The evolution of electrical conductivity in HPG-CCM pastes with different mix proportions is presented in Figure 14. For the unmodified N1 sample (i.e., without quicklime and BFS), the conductivity exhibited a rapid initial increase followed by a gradual decline. This behavior can be attributed to the rapid dissolution of Ca²⁺, SO₄^2-, and minor amounts of soluble P and F impurities from HPG during the early stage, which sharply increased the ionic concentration in the pore solution, thereby causing a pronounced rise in conductivity. Subsequently, Ca²⁺ ions in solution reacted with PO₄^3- and F⁻ ions to form insoluble precipitates, reducing the concentration of free ions and leading to a gradual stabilization of conductivity. When 1 wt% quicklime and 40 wt% BFS were incorporated (sample N10), the conductivity also rose sharply at the early stage but quickly leveled off within a short period. This can be explained by the rapid dissolution of quicklime during hydration, which released large amounts of Ca²⁺ and OH⁻ ions, causing an initial surge in conductivity (Yang et al., 2022). As PO₄^3- and F⁻ anions reacted with Ca²⁺ to form low-solubility precipitates, the ionic concentration in the solution reached a new equilibrium, resulting in stabilized conductivity.

Figure 14

Figure 14. Electrical conductivity of HPG-CCM pastes at different curing times.

It is noteworthy that when the quicklime dosage exceeded 3wt%, the conductivity of the HPG paste increased significantly more than that of the unmodified sample. This indicates that higher quicklime contents markedly promote the dissolution of HPG, accelerating ion release and thereby effectively enhancing its hydration reactivity. The process subsequently led to the formation of hydration products such as CaSO₄·2H₂O, ettringite, and C-S-H gel, followed by a dissolution–crystallization equilibrium stage, during which the conductivity tended to stabilize in the later period.

3.3.3.2 Hydration heat analysis

The hydration heat curves of HPG-CCM with different mix proportions are presented in Figure 15. As shown in Figure 15a, three main exothermic peaks appear during the hydration process at approximately 0–0.25 h, 1.5–7.5 h, and 7.5–70 h. In the initial stage (0–0.25 h), the peak heat release significantly increases with rising quicklime content, primarily attributed to the exothermic reaction of quicklime hydration forming Ca(OH)₂.

Figure 15

Figure 15. Hydration heat analysis of HPG-CCM: (a) heat flow during 0–70 h hydration; (b) heat flow during 1.5–7.7 h hydration; (c) heat flow during 7.5–70 h hydration; (d) cumulative hydration heat.

Figure 15b reveals that the unmodified HPG sample exhibits no distinct exothermic peaks, indicating a relatively low hydration rate. In contrast, quicklime-modified HPG samples show pronounced exothermic peaks within the 1.5–7.5 h interval. The heat release during this stage mainly arises from the reaction of quicklime with soluble P and F impurities in the solution, generating insoluble calcium salts such as Ca₃(PO₄)₂ and CaF₂. These precipitates do not form a dense barrier layer on the HPG crystal surfaces, thereby enabling rapid hydration of HPG into CaSO₄·2H₂O, a strongly exothermic process (Qian et al., 2008; Re et al., 2023). This result demonstrates that an appropriate amount of quicklime can substantially enhance the early hydration performance of HPG.

Figure 15c shows that in the later stage (7.5–70 h), the third exothermic peak shifts toward shorter curing times with increasing quicklime content, reflecting an accelerated hydration reaction. When the quicklime dosage is 1% (sample N10), only a weak exothermic peak appears at 37.5 h, suggesting incomplete neutralization of P and F impurities and a hindered hydration rate. Increasing the dosage to 3% (sample N11) advances the exothermic peak to 30 h with a markedly enhanced peak value. This is because at this dosage, impurities are fully neutralized, and the quicklime synergistically activates the hydration of BFS alongside the formation of CaSO₄·2H₂O, leading to abundant C-S-H gel and ettringite crystals that effectively fill pores and improve structural density and strength. Further increasing the dosage to 5% and 7% causes excess Ca(OH)₂ to react with CaSO₄·2H₂O earlier, triggering BFS hydration sooner and shifting the exothermic peak to approximately 12 h (Xue et al., 2018; Chen X. et al., 2022).

The cumulative hydration heat release is showed in Figure 15d. Over 0–72 h, sample N10’s cumulative heat release reaches only 52.36 J/g, close to that of unmodified HPG. Sample N11’s cumulative heat significantly increases to 194.29 J/g, a 271.05% improvement over the unmodified sample. When the dosage rises to 5% and 7%, cumulative heat decreases to 144.00 J/g and 163.10 J/g, respectively. Notably, within the first 20 h, samples with 5% and 7% quicklime show higher cumulative heat than the 3% sample, but after 20 h, the 3% sample’s heat release rapidly increases to the highest value. These results indicate that optimizing the quicklime and BFS proportions is critical for balancing early activation and long-term hydration potential, thereby enhancing the water resistance of HPG-CCM.

3.4 Hydration mechanism

Based on the previous research findings, the hydration process of the HPG-quicklime-BFS ternary cementitious system can be divided into three main stages: the early hydration stage, the hydration densification stage, and the hydration encapsulation stage (Figure 16). In the early hydration stage, quicklime reacts rapidly with water to form Ca(OH)₂, releasing a significant amount of OH⁻ ions, which greatly increases the alkalinity of the HPG-CCM system. In this highly alkaline environment, soluble impurities such as P and F in the HPG dissolve quickly and react with OH⁻ and Ca²⁺ ions to form insoluble Ca₃(PO₄)₂ and CaF₂ precipitates (Xu et al., 2023; Wu et al., 2022). Meanwhile, during the hydration of HPG, Ca²⁺ and SO₄^2- ions are released, and when their concentration reaches the solubility product of CaSO₄·2H₂O, CaSO₄·2H₂O nuclei are formed and continue to grow under hydration (Equations 8, 9). These interlocking CaSO₄·2H₂O crystals gradually form the early strength framework. In the hydration densification stage, the high alkalinity provided by quicklime promotes the dissolution of the glass phase in the blast furnace slag, releasing a large amount of SiO₄^2- and Al(OH)₄^- ions. SiO₄^2- reacts with Ca²⁺ and OH⁻ to form C-S-H gel (Equation 10). At the same time, Al(OH)₄^- reacts with Ca²⁺, SO₄^2-, and OH⁻ to form ettringite crystals (Equation 11). These C-S-H gels and ettringite crystals fill the pores in the CaSO₄·2H₂O framework, significantly densifying the system structure. In the hydration encapsulation stage, as the hydration reaction continues, the content of C-S-H gel increases and gradually encapsulates the CaSO₄·2H₂O crystals. This process not only further enhances the mechanical properties of the HPG-CCM system but also effectively improves its water resistance.

${\text{CaSO}}_4·0.5{\mathrm{H}}_2\mathrm{O}\to {\text{Ca}}^{2+}+\mathrm{S}{{\mathrm{O}}_4}^{2-}+0.5{\mathrm{H}}_2\mathrm{O}(8)$

${\text{Ca}}^{2+}+\mathrm{S}{{\mathrm{O}}_4}^{2-}+2{\mathrm{H}}_2\mathrm{O}\to {\text{CaSO}}_4·2{\mathrm{H}}_2\mathrm{O}(9)$

${\text{SiO}}_2+x\text{Ca}{\left(\text{OH}\right)}_2+y{\mathrm{H}}_2\mathrm{O}\to x\text{CaO}·{\text{SiO}}_2·y{\mathrm{H}}_2\mathrm{O}(10)$

$\begin{array}{ll}\hfill & \hfill {\text{Al}}_2{\mathrm{O}}_3+3{\text{CaSO}}_4·2{\mathrm{H}}_2\mathrm{O}+3\text{Ca}{\left(\text{OH}\right)}_2+23{\mathrm{H}}_2\mathrm{O}\\ \hfill & \hfill \to 3\text{CaO}·{\text{Al}}_2{\mathrm{O}}_3·3{\text{CaSO}}_4·32{\mathrm{H}}_2\mathrm{O}\end{array}(11)$

Figure 16

Figure 16. The three hydration stages of the HPG-quicklime-BFS ternary cementitious system.

4 Conclusion

This study systematically investigated the synergistic effects of quicklime and BFS on the water resistance and hydration characteristics of the HPG-CCM. The main conclusions are summarized as follows:

1. Without quicklime modification, the soluble P₂O₅ and F impurities in HPG significantly hinder its hydration process, resulting in negligible early strength development. After 28 days of curing, the compressive strength of the unmodified HPG-CCM was only 3.03 MPa with a softening coefficient of 0.37, indicating poor water resistance and limiting its application in moisture-exposed industrial environments.

2. The synergistic activation of HPG by quicklime and BFS, with an optimized HPG to BFS mass ratio of 6:4% and 3% quicklime addition, yielded remarkable performance improvements. The HPG-CCM exhibited an early compressive strength of 8.96 MPa at 1 day and reached 52.8 MPa at 28 days, demonstrating excellent mechanical strength and water resistance.

3. Quicklime rapidly neutralizes soluble P and F impurities in HPG while providing abundant Ca²⁺ ions, thereby accelerating the early hydration of HPG. When the quicklime content exceeded 3%, partial hydration was achieved within 1 day, leading to the development of considerable cementitious strength. Moreover, the contents of P and F in the resulting HPG-CCM complied with the relevant Chinese regulatory limits, confirming its environmental safety.

4. The hydration of HPG-CCM under the combined influence of quicklime and BFS occurs in three stages. In the early stage, quicklime converts P and F impurities into inert forms and releases Ca²⁺ ions, promoting the transformation of HPG into CaSO₄·2H₂O. In the densification stage, BFS reacts with quicklime and CaSO₄·2H₂O to form C-S-H gel and ettringite, filling pores and enhancing compactness. Finally, in the encapsulation stage, C-S-H gel coats the CaSO₄·2H₂O crystals, further improving the sample’s water resistance and mechanical properties.

Data availability statement

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

Author contributions

JL: Formal Analysis, Writing – original draft, Investigation, Supervision. YL: Formal Analysis, Investigation, Supervision, Data curation, Writing – review and editing. LY: Writing – review and editing, Software, Methodology, Investigation, Funding acquisition, Resources, Project administration. JC: Resources, Project administration, Supervision, Writing – review and editing.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This work was financially supported by the Guizhou Provincial Science and Technology Key Project (Grant No. LH (2025) 010).

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

^{Abbreviations:}WPA, wet-process phosphoric acid; PG, phosphogypsum; HPG, hemihydrate phosphogypsum; HPG-CCM, hemihydrate phosphogypsum-based composite cementitious material; BFS, blast furnace slag; UCS, uniaxial compressive strength; SEM-EDS, scanning electron microscopy-energy dispersive X-ray spectroscopy; XRD, X-ray diffraction; TG, thermogravimetric, DTG, differential thermogravimetric; FTIR, Fourier transform infrared spectroscopy.

References

alvarez-Ayuso, E., and Nugteren, H. W. (2005). Synthesis of ettringite: a way to deal with the acid wastewaters of aluminium anodising industry. Water Res. 39 (1), 65–72. doi:10.1016/j.watres.2004.07.029

PubMed Abstract | CrossRef Full Text | Google Scholar

Antonio, A., Neto, M., Cincotto, M., and Repette, W. (2010). Mechanical properties, drying and autogenous shrinkage of blast furnace slag activated with hydrated lime and gypsum. Cem. Concr. Comp. 32, 312–318. doi:10.1016/j.cemconcomp.2010.01.004

CrossRef Full Text | Google Scholar

Chen, P., Ma, B., Tan, H., Wu, L., Zheng, Z., He, X., et al. (2022a). Improving the mechanical property and water resistance of β-hemihydrate phosphogypsum by incorporating ground blast-furnace slag and steel slag. Constr. Build. Mat. 344, 128265. doi:10.1016/j.conbuildmat.2022.128265

CrossRef Full Text | Google Scholar

Chen, X., Wang, Q., Wu, Q., Xie, X., Tang, S., Yang, G., et al. (2022b). Hydration reaction and microstructural characteristics of hemihydrate phosphogypsum with variable pH. Constr. Build. Mat. 316, 125891. doi:10.1016/j.conbuildmat.2021.125891

CrossRef Full Text | Google Scholar

Fornés, V., Vaičiukynienė, D., Nizevičienė, D., and Doroševas, V. (2021). The improvement of the water-resistance of the phosphogypsum by adding waste metallurgical sludge. J. Build. Eng. 43, 102861. doi:10.1016/j.jobe.2021.102861

CrossRef Full Text | Google Scholar

Gao, D., Zhang, D., Peng, Y., Diao, Hu., and Wang, W. (2022). Effects of different carboxyl additives on the growth habits of hemihydrate gypsum crystals. Constr. Build. Mat. 316, 126037. doi:10.1016/j.conbuildmat.2021.126037

CrossRef Full Text | Google Scholar

Jia, R., Wang, Q., and Luo, T. (2021). Reuse of phosphogypsum as hemihydrate gypsum: the negative effect and content control of H₃PO₄. Resour, Conserv Recy 174, 105830. doi:10.1016/j.resconrec.2021.105830

CrossRef Full Text | Google Scholar

Jiang, G., Wu, A., Wang, Y., and Lan, W. (2018). Low cost and high efficiency utilization of hemihydrate phosphogypsum: used as binder to prepare filling material. Constr. Build. Mat. 167, 263–270. doi:10.1016/j.conbuildmat.2018.02.022

CrossRef Full Text | Google Scholar

Jiang, G., Zhao, L., Li, H., Wu, S., Cheng, H., Sun, W., et al. (2024). Immobilization of phosphorus and fluorine from whole phosphogypsum-based cemented backfill material with self-cementing properties. Constr. Build. Mat. 438, 137072. doi:10.1016/j.conbuildmat.2024.137072

CrossRef Full Text | Google Scholar

Lin, R., Huang, G., Ma, F., Pan, T., Wang, X., Han, Y., et al. (2024). Investigation of phosphogypsum-based cementitious materials: the effect of lime modification. Dev. Built Environ. 18, 100477. doi:10.1016/j.dibe.2024.100477

CrossRef Full Text | Google Scholar

Nishiyama, T., and Oinuma, K. (1977). Infrared spectra in the region of OH stretching vibration of clay minerals. J. Min. Pet. Sci. 13, 129–139. doi:10.2465/gkk1952.13

CrossRef Full Text | Google Scholar

Park, H., Jeong, Y., Jun, Y., Jeong, J., and Oh, J. (2016). Strength enhancement and pore-size refinement in clinker-free CaO-activated GGBFS systems through substitution with gypsum. Cem. Concr. Comp. 68, 57–65. doi:10.1016/j.cemconcomp.2016.02.008

CrossRef Full Text | Google Scholar

Qi, H., Yan, X., Ma, B., Tan, H., He, X., Su, Y., et al. (2023). Improving the volume stability of the β-hemihydrate phosphogypsum-cement system by incorporating GGBS. Constr. Build. Mat. 408, 133807. doi:10.1016/j.conbuildmat.2023.133807

CrossRef Full Text | Google Scholar

Qian, G. R., Shi, J., Cao, Y. L., Xu, Y. F., and Chui, P. C. (2008). Properties of MSW fly ash-calcium sulfoaluminate cement matrix and stabilization/solidification on heavy metals. J. Hazard Mat. 152 (1), 196–203. doi:10.1016/j.jhazmat.2007.06.118

PubMed Abstract | CrossRef Full Text | Google Scholar

Qu, E., Thomas, A., and Westphal, T. (2017). Phase assemblage in ettringite-forming cement pastes: a X-ray diffraction and thermal analysis characterization. J. Build. Eng. 12, 37–50. doi:10.1016/j.jobe.2017.05.005

CrossRef Full Text | Google Scholar

Ren, Z., Wang, L., Wang, H., Liu, S., and Ren, J. (2023). Stabilization and solidification mechanism of Pb in phosphogypsum slag-based cementitious materials. Constr. Build. Mat. 368, 130427. doi:10.1016/j.conbuildmat.2023.130427

CrossRef Full Text | Google Scholar

Tao, L., Deng, F., Xu, K., Dong, X., Xu, D., Liao, Y., et al. (2024). Effect of calcium sulfoaluminate cement on the hydration process and water resistance of phosphogypsum slag cement. J. Mat. Res. 33, 5375–5383. doi:10.1016/j.jmrt.2024.10.166

CrossRef Full Text | Google Scholar

Trezza, M. A., and Lavat, A. E. (2001). Analysis of the system 3CaO⋅Al₂O₃–CaSO₄⋅2H₂O–CaCO₃–H₂O by FT-IR spectroscopy. Cem. Concr. Res. 31 (6), 869–872. doi:10.1016/s0008-8846(01)00502-6

CrossRef Full Text | Google Scholar

Wang, J., Tan, H., He, X., Zhang, J., Jian, S., Du, C., et al. (2022). Influence of wet grinded slag on the hydration of phosphogypsum-slag based cement and its application in backfill tailings. Constr. Build. Mat. 360, 129509. doi:10.1016/j.conbuildmat.2022.129509

CrossRef Full Text | Google Scholar

Wang, Y., Huang, X., Zhang, S., Ma, W., and Li, J. (2024). Utilization of ultrafine solid waste in the sustainable cementitious material for enhanced performance. Constr. Build. Mat. 417, 135239. doi:10.1016/j.conbuildmat.2024.135239

CrossRef Full Text | Google Scholar

Wang, Z., Wang, Y., Wu, A., Zhang, L., Ruan, Z., Li, J., et al. (2025a). Mitigating natural aging in hemihydrate phospho-calcium sulfate with sodium tripolyphosphate: sustaining long-term cementitious performance. Constr. Build. Mat. 487, 142156. doi:10.1016/j.conbuildmat.2025.142156

CrossRef Full Text | Google Scholar

Wang, Z., Wang, Y., Dino, G., Zhang, L., Ruan, Z., Zhang, M., et al. (2025b). Degradation of hemihydrate phosphogypsum-based backfill in underground mining: mechanical and microstructural insights on the effects of pH and temperature of mine water, process. Saf. Environ. Prot. 193, 272–285. doi:10.1016/j.psep.2024.11.003

CrossRef Full Text | Google Scholar

Wei, H., Jing, L., Qun, L., Qiang, H., Hui, Z., Yu, C., et al. (2022). Calcium transfer process of cement paste for ettringite formation under different sulfate concentrations. Constr. Build. Mat. 348, 128706. doi:10.1016/j.conbuildmat.2022.128706

CrossRef Full Text | Google Scholar

Wu, F., Chen, B., Qu, G., Liu, S., Zhao, C., Ren, Y., et al. (2022). Harmless treatment technology of phosphogypsum: directional stabilization of toxic and harmful substances. J. Environ. Manage. 311, 114827. doi:10.1016/j.jenvman.2022.114827

PubMed Abstract | CrossRef Full Text | Google Scholar

Xia, Y., Tang, L., He, Z., Tang, J., Yu, J., Cao, D., et al. (2025). Study on mechanical properties, hydration mechanism and carbon emission of phosphogypsum-based sustainable cementitious materials. Sustain. Chem. Pharm. 45, 102016. doi:10.1016/j.scp.2025.102016

CrossRef Full Text | Google Scholar

Xu, H., Ming, J., Xiao, Z., and Ci, T. (2016). Mechanical properties and hydration mechanisms of high-strength fluorogypsum-blast furnace slag-based hydraulic cementitious binder. Constr. Build. Mat. 127, 137–143. doi:10.1016/j.conbuildmat.2016.09.152

CrossRef Full Text | Google Scholar

Xu, J., Xu, F., Jiang, Y., Jiao, Y., Sun, T., Yang, F., et al. (2023). Mechanical properties and soluble phosphorus solidification mechanism of a novel high amount phosphogypsum-based mortar. Constr. Build. Mat. 394, 132176. doi:10.1016/j.conbuildmat.2023.132176

CrossRef Full Text | Google Scholar

Xue, C., Jian, G., Chuan, L., and Ya, Z. (2018). Effect of neutralization on the setting and hardening characters of hemihydrate phosphogypsum plaster. Constr. Build. Mat. 190, 53–64. doi:10.1016/j.conbuildmat.2018.09.095

CrossRef Full Text | Google Scholar

Yang, J., Zeng, J., He, X., Zhang, Y., Su, Y., and Tan, H. (2022). Sustainable clinker-free solid waste binder produced from wet-ground granulated blast-furnace slag, phosphogypsum and carbide slag. Constr. Build. Mat. 330, 127218. doi:10.1016/j.conbuildmat.2022.127218

CrossRef Full Text | Google Scholar

Yoon, S., Jeon, C., and Choi, S. (2023). Durability of blast-furnace slag concrete mixed with expansive and swelling admixtures. Case Stud. Constr. Mat. 18, e02179. doi:10.1016/j.cscm.2023.e02179

CrossRef Full Text | Google Scholar

Yuichiro, K., Haruka, T., and Sadayuki, W. (2021). The long-term suppression effects of fly ash and slag on the expansion of heat-cured mortar due to delayed ettringite formation. Constr. Build. Mat. 310, 125235. doi:10.1016/j.conbuildmat.2021.125235

CrossRef Full Text | Google Scholar

Zhang, B., Shen, W., Wang, J., Zhang, W., Zhu, W., and Liu, Y. (2017). Slight-expansive road base course binder: properties, hydration and performance. Constr. Build. Mat. 150, 626–633. doi:10.1016/j.conbuildmat.2017.06.002

CrossRef Full Text | Google Scholar

Zhang, L. J., Mo, K. H., Yap, S., Gencel, O., and Ling, T. (2022). Mechanical strength, water resistance and drying shrinkage of lightweight hemihydrate phosphogypsum-cement composite with ground granulated blast furnace slag and recycled waste glass. Constr. Build. Mat. 345, 12832. doi:10.1016/j.conbuildmat.2022.128232

CrossRef Full Text | Google Scholar

Zhe, Z., Yuchen, H., Lian, X., and Guo, G. (2025). The influence of portlandite, calcite, quartz and ettringite inclusions on the multiscale mechanical behaviors of C-S-H matrix. Cem. Concr. Res. 189, 107781. doi:10.1016/j.cemconres.2024.107781

CrossRef Full Text | Google Scholar

Zheng, P., Li, W., Ma, Q., and Xi, L. (2023). Mechanical properties of phosphogypsum-soil stabilized by lime activated ground granulated blast-furnace slag. Constr. Build. Mat. 402, 132994. doi:10.1016/j.conbuildmat.2023.132994

CrossRef Full Text | Google Scholar

Zhou, Z., Li, H., Liu, N., Peng, L., and Jiang, Z. (2024). Development and property optimization of a sustainable phosphogypsum-based cementitious system with ground-granulated blast furnace slag and carbide slag. Constr. Build. Mat. 449, 138498. doi:10.1016/j.conbuildmat.2024.138498

CrossRef Full Text | Google Scholar