Magnesium silicate hydrate (M-S-H) cementitious material is regarded as a potential alternative to conventional Portland cement due to its excellent sulfate erosion resistance, heavy metal ion solidification performance, and low pH value during hydration. However, the industrialization process of M-S-H cementitious materials is currently restricted due to the high cost and unstable supply of traditional silicon sources that rely on industrial by-products (such as silica fume, fly ash, and slag). Natural layered silicates such as montmorillonite and kaolin have a similar tetrahedral-octahedral stacking unit structure to M-S-H, and can exhibit a good pozzolanic activity after mechanical or thermal activation. Exploring the use of calcined montmorillonite and kaolin as alternative silicon sources to partially or completely replace silica fume broadens the source of raw materials for M-S-H cementitious materials and reduces production costs. This work was to investigate the effects of partial or complete replacement of silica fume by calcined montmorillonite (MT) and calcined kaolin (KL) on the structure and strength of M-S-H cementitious materials, clarifying the action mechanism of aluminum in the two clay minerals on the M-S-H hydration process and layered structure.

Light-burned magnesia (LM) was used as a magnesium source and silica fume (SF) as a reference silicon source. Montmorillonite and kaolin were calcined in a muffle furnace at 750 ℃ for 1 h at a heating rate of 10 ℃/min to obtain calcined montmorillonite and calcined kaolin. The reference sample was prepared via mixing MgO and silica fume at a mass ratio of 4 : 6. On this basis, different mass fractions (i.e., 10%, 30%, 50%, 70%, and 100%) of calcined montmorillonite and kaolin were used to replace silica fume, and the samples were named as MT-10~MT-100 and KL-10~KL-100, respectively according to the replacement ratio. The mortar strength specimens were prepared with the samples above, superplasticizer, tap water, and standard sand, and the compressive strength was tested according to the standard GB/T 17671—1999 “Methods of testing cement- Determination of strength”. The paste was prepared at a water-to-solid ratio of 0.5 and 0.7% superplasticizer based on the total mass of solids, cured in sealed glass tubes with liquid paraffin, and terminated hydration with isopropanol at a certain age, then dried in a vacuum oven at 40 ℃ for 48 h. The mineral phase, thermal behavior, functional groups, microstructure, and element distribution of the samples were characterized by X-ray diffraction (XRD), thermogravimetric analysis (TGA), Fourier transform infrared spectroscopy (FTIR), scanning electron microscopy (SEM), and solid-state ²⁹Si and ²⁷Al magic-angle spinning nuclear magnetic resonance (MAS NMR).

Calcined kaolin exhibits a great early strength enhancement effect, and its compressive strengths at 3 d and 7 d have a positive correlation with the dosage (i.e., correlation coefficients are 0.996 and 0.905, respectively). At the dosage of calcined kaolin of 70%, the strength at 3 d reaches 39.37 MPa and the strength at 7 d reaches 49.58 MPa, which are 464.0% and 400.3% higher than those of the reference sample, respectively. In contrast, the strength enhancement effect of calcined montmorillonite is weak. The strengths at 3 d and 7 d increase slowly with the increase of montmorillonite dosage. At the montmorillonite dosage of 70%, the strength at 3 d is only 8.94 MPa and the strength at 7 d is 14.28 MPa, which are 36.5% and 53.7% higher than those of the reference sample, respectively. The strength increases slowly in the montmorillonite dosage range of 10%–50%.

The microscopic analysis shows that calcination at 750 ℃ causes more thorough damage to the crystal structure of kaolin than that of montmorillonite. The XRD diffraction peaks of kaolin disappear completely after calcination, and the layered structure collapses completely, resulting in a higher pozzolanic activity. After calcination, the interlayer spacing of montmorillonite decreases, and some characteristic diffraction peaks are still retained, indicating that its crystal order is not completely destroyed and the pozzolanic activity is lower than that of calcined kaolin. The FTIR and NMR analysis further confirms that the structural hydroxyl groups of calcined kaolin are completely removed, and a large number of active Al and Si sites are exposed. Although montmorillonite also undergoes dehydroxylation and structural damage after calcination, the degree of disorder is lower than that of kaolin.

Aluminum in both calcined montmorillonite and kaolin can enter the M-S-H phase in the early stage of hydration, and form magnesium aluminosilicate hydrate (M-S-A-H) gel via substituting Si⁴⁺ in the silicate tetrahedron or Mg²⁺ in the magnesium-oxygen octahedron, which significantly improves the polymerization degree of the product. The formed M-S-A-H is external hydration products, mainly formed through element diffusion between particles. The analysis by the XRD, FTIR, and DTG shows that Mg-Al layered double hydroxide (Mg-Al LDH) is generated within the first 28 d of hydration in both raw material systems. The enhancement of cementitious material strength by calcined kaolin and montmorillonite is related to the formation of M-S-A-H. The strength improvement effect of calcined kaolin is higher than that of calcined montmorillonite due to the more thorough destruction of the crystal structure, higher degree of network disorder, more active Al and Si sites, and higher content of Al(Ⅴ).

Calcined kaolin and montmorillonite could be used as alternative silicon sources for M-S-H cementitious materials. Calcined kaolin had a great early strength enhancement effect, and its compressive strength had a positive correlation with the dosage. At the dosage of 70%, the compressive strength at 7 d was increased by more than 400%, compared with the sample prepared with pure SF. The strength enhancement effect of calcined montmorillonite could be weak, and the strength was only increased by 53.7% at the same dosage and age. The difference in strength enhancement effect between calcined kaolin and montmorillonite could be due to the fact that the calcination at 750 ℃ caused more thorough damage to the crystal structure of kaolin than that of montmorillonite, resulting in a higher pozzolanic activity. Aluminum in both calcined montmorillonite and kaolin could enter the M-S-H phase in the early stage to form M-S-A-H gel, significantly improving the polymerization degree of the product. The formed M-S-A-H were external hydration products, and Mg-Al LDH was generated within the first 28 d of hydration.

The treatment of dye wastewater remains a formidable challenge within the realm of industrial wastewater management. The existing methods for treating dye wastewater include adsorption, electrochemical oxidation, photocatalysis, biodegradation, and membrane separation. Among these, membrane separation is particularly advantageous due to its energy efficiency, high efficacy, and ease of process control. Separation membranes can be categorized into two primary types, i.e., organic and inorganic membranes. Inorganic membranes have attracted significant attention due to their high mechanical strength, controllable size distribution, resistance to high temperature and pressure, and chemical stablility. Despite these benefits, the advancement of conventional inorganic ceramic membranes is hindered by their complex preparation processes and high energy demands. This limitation underscores a urgent need for the development of simple and cost-effective inorganic membranes. A promising innovation in this field is a geopolymer-zeolite composite membrane, which is synthesized through in-situ hydrothermal conversion following the formation of reactive silica-aluminum materials under alkali or acid excitation. Coal gangue (CG), as a solid waste from coal mining primarily composed of SiO₂ and Al₂O₃, can be used as a raw material for the production of geopolymer-zeolite composite membrane. In this paper, A cost-effective geopolymer-zeolite composite membrane was prepared by a polymerization-hydrothermal method with CG as the main raw material for the separation of RhB in water. The objective was to propose a novel utilization strategy for CG and to lay a foundation for the development of membrane separation technology.

Coal gangue (CG) (Zhunger, Inner Mongolia, China) was calcinated at 800 ℃ for 2 h, resulting in the formation of calcined coal gangue (CCG). The CCG was then uniformly mixed with other additives in a mass ratio of CCG:NaOH:Na₂SiO₃:H₂O (100:10:40:45). This mixture was poured into polytetrafluoroethylene moulds with the diameters of 50 mm and thicknesses of 5 mm. The samples were cured at 70 ℃ for 24 h to produce coal gangue-geopolymer membrane (CCG-GM). Afterwards, the CCG-GM was hydrothermally treated in a 100 mL 1 mol/L NaOH solution at 140 ℃ for 12 h, resulting in a coal gangue-geopolymer-zeolite composite membrane (CCG-GZCM).

The prepared CCG-GZCM was incorporated into a custom-built terminal filtration system, and its performance was evaluated via measuring the pure water flux, membrane flux, and RhB removal rate under varying conditions (i.e., system pressure, initial RhB solution concentrations, and pH values of the RhB solution). In addition, the circular utilization of CCG-GZCM was also examined to assess its reusability. The selectivity of CCG-GZCM for the removal contaminants (i.e., RhB, methylene blue (MLB), methyl violet (MV), methyl blue (MB), methyl orange (MO), and eosin Y (EY)) was determined.

The XRD patterns reveals that CG primarily consists of quartz and kaolinite. Upon calcination at 800 ℃ for 2 h, kaolinite in CG transforms into metakaolinite, thus froming CCG. For alkali activation, a broad diffraction peak of CCG initially appears at 15°–25°, and then shifts to 20°–40°, indicating the conversion of CCG into a geopolymer, thus forming CCG-GM. Subsequently, CCG-GZCM, with a main phase composition of NaP1 zeolite, is obtained via in-situ hydrothermal transformation of CCG-GM. The FTIR spectra of CCG-GZCM confirm the presence of NaP1 zeolite absorption bands. The XPS spectra further identify Na, Si, O, and Al as the predominant elements on the surfaces of both CCG-GM and CCG-GZCM. Note that CCG-GZCM exhibits a larger specific surface area and pore volume, compared to CCG-GM, albeit with a smaller average pore size. The TG-DSC spectra indicate that the weight loss of CCG-GZCM primarily occurs at 30–253 ℃, indicating its good thermal stability. In addition, the compressive strength of CCG-GM also increases from 18.84 MPa to 47.01 MPa (CCG-GZCM) after hydrothermal transformation.

CCG-GZCM exhibits varying removal efficiencies for different dye types. At an initial concentration of 10 mg/L and a pH value of 7, the pure water flux and membrane flux to RhB solution of CCG-GM and CCG-GZCM increase linearly with the increase of system pressure, while the RhB removal rate decreaseslinearly. At a system pressure of –0.08 MPa and a pH value of 7, the RhB removal rate by CCG-GZCM gradually decreases with the increase of initial concentration within 60 min. Conversely, at the same system pressure and an initial RhB concentration of 10 mg/L, the RhB removal rate by CCG-GZCM gradually increases with increasing pH value of the RhB solution within 60 min, and the maximum is 99.11%. However, when the system pressure maintains at –0.08 MPa, at an initial RhB concentration of 10 mg/L and an initial pH value of 11, the RhB removal rate by CCG-GZCM decreases with an increasing number of cycles, reducing to 94.25% after 6 cycles. In addition, CCG-GZCM demonstrates different removal rates within 30 min for different dye types. It achieves a high removal rate of exceeding 99% for cationic dyes including RhB, MLB, and MV. In contrast, the removal rates for anionic dyes, including MB, MO, and EY, are 70.00%, 25.82%, and 42.69%, respectively.

The pore structure of geopolymer membrane was modified and converted it into CCG-GZCM with larger specific surface area, higher pore volume and smaller average pore size by a hydrothermal method. The CCG-GZCM was characterized by an amorphous geopolymer matrix with NaP1 zeolite and a minor presence of NaA zeolite on the surface. At a system pressure of –0.08 MPa, an initial RhB concentration of 10 mg/L, and a pH value of 11, the CCG-GZCM achieved a remarkable RhB removal rate of 99.11% for 60 min. In addition, the CCG-GZCM also demonstrated a high circular utilization efficiency for RhB removal, maintaining a removal rate of exceeding 99% for cationic dyes for 30 min. The CCG-GZCM could be used as a promising approach for the effective utilization of coal gangue, having a significant potential for application in water pollutant treatment.

The use of steel slag as a cementitious material is the most likely field to achieve large-scale engineering utilization. However, the inherent defects of low activity and poor grindability caused by the high iron oxide content of steel slag limit its dosage in cement. One important direction to improve the activity and grindability of steel slag is to obtain a product similar to granulated blast furnace slag (GBFS), mainly composed of glass, by melting and reducing iron oxides in steel slag. This GBFS-like residue exhibits lower early activity due to its lower alkalinity coefficient (K=n(CaO)+ n(MgO)/n(SiO₂)+ n(Al₂O₃)). In order to improve the early activity of the residue, this paper focuses on the effect of CaO content on the formation of clinker minerals. Coal gangue is used as a reducing agent to reduce iron oxides in steel slag for making full use of industrial solid waste.

Based on the principle of complete reduction of iron oxides in steel slag and coal gangue, the appropriate ratio of coal gangue and steel slag is calculated according to the chemical composition of steel slag and the fixed carbon content of coal gangue. CaO was obtained by calcining and analytical reagent CaCO₃ at 1500 ℃. Different amounts of resultant CaO (10, 20, 30, 40 g) were incorporated into 100 g mixture of steel slag and coal gangue, respectively. The reference sample without CaO and the above four samples are named as C0, C10, C20, C30 and C40, and the K of samples are 1.05, 1.33, 1.61, 1.89, 2.17, respectively. After thoroughly mixing each sample, 200 g mixture containing CaO was placed into a corundum crucible and calcined at 1500 ℃ for 30 min. After calcination, the crucible was take out from the high temperature furnace for water quenching. Iron alloy particles are obtained by crushing, peeling, and magnetic separation from water quenched slag. The remaining water quenched residue (WQR) is used for mineral phase analysis, cement mortar strength, soundness, and other tests.

Iron alloy particles in C0, C10, C20, C30 sample can be easily peeled off. But it is difficult for C40 sample and some small metal particles can be observed at the crucible bottom. This is because the increase of K leads to an increase in the viscosity of the melt, making it difficult for iron particles to sedimentation and aggregation.

The analysis of the chemical composition of the WQR after stripping iron particles shows that the reduction rate δ and recovery rate η of the reference sample C0 are about 92%. For C10 (K=1.33), the η and δ of Fe reach maximum values. As the CaO content continues to increase, the η and δ gradually decrease. But, these two parameters of C20 still exceed that of C0. The f-CaO content and soundness of the WQR meet the requirements of relevant national standards, although these two values will slightly increase with the increase of dosage of CaO.

The results of XRD Rietveld refinement for quantification indicate that C0 and C10 samples are mainly consist of glass and contain a small amounts of Gehlenite, spinel et al. With the increase of CaO content, the diffraction peak of the crystalline phase gradually increases, while the amorphous envelope peak gradually weakens. For C20, the main crystalline phases are C₂S and bredigite, and diffraction peaks of periclase can be observed. In the C30 sample, the main mineral phases are C₂S, C₃A, iron, and periclase. Meanwhile, C₃S diffraction peak is relatively low. The C₃S diffraction peak in the C40 sample is significantly enhanced, with a content of 50%. FTIR shows that as the CaO content increases from 20 g to 40 g, the absorption band of β-C₂S decreases, while the absorption band of M3 C₃S gradually increases.

BSE image show that the C10 sample is mainly composed of glass matrix, with a small amount of C₂S grain, serrated-like fine stuff formed at the edge of C₂S grains, snowflake like or dendritic substance surrounding the C₂S, and iron particles. EDS results show that the order of Al content in several phases is C₂S < fine stuff < dendritic substance < matrix. The order of Ca and Si content is C₂S >fine stuff > dendritic substance > matrix.

In the C30 sample, sharp edged plate-like particles with a Ca/Si of about 3 formed, indicating that C₂S has begun to transform into C₃S. The formation of C₃S and C₂S not only significantly reduces the Si content in the matrix, but also leads to an increase in the relative content of Al in the intermediate matrix. When the dosage reaches 40%, C₃S has already formed in large quantities. The Ca/Al ratio of the intermediate phase is about 1.53, indicating the formation of C₃A.

The main conclusions of this paper are summarized as following. When K value of WQR is less than 1.6, increasing the CaO content appropriately can improve the reduction rate and recovery rate of iron. The C10 sample is mainly composed of glass. When K is 1.3, a small amount of β-C₂S is formed in the WQR, which gradually grows by absorbing CaO and SiO₂ from the glass matrix. With the increase of K, Periclase begins to form in the zone where silicate minerals are more abundant. When the CaO dosage reaches 40 g, the amount of C₃S continues to increase to 50%, and C₃A form simultaneously. Although high alkalinity leads to the formation of C₃S and C₂S minerals in the system, it also results in the formation of periclase. To avoid the potential long-term unsoundness, it is more appropriate to control K value at around 1.3.