Solidification performances of contaminants by red mud-based cementitious paste filling material and leaching behavior of contaminants in different pH and redox potential environments Open Access

Coal resources occupy a leading position in China's energy supply (Haibin & Zhenling 2010; Zhao et al. 2020) and play a crucial role in the development of economic growth and social progress (Sun et al. 2012, 2015). However, improper exploitation of coal resources has brought about serious environmental problems. After coal mining, multi-mined-out areas are formed, then threaten the safety of surface buildings engineering activities above (Kesimal et al. 2004; Hesketh et al. 2010; Zhang et al. 2020). Further, groundwater gradually accumulates and is easy to form acidic goaf water (Jiang et al. 2020), and consequently generates disastrous water resource and ecological environmental risks. Therefore, comprehensive treatment of coal mined-out areas is necessary. At present, the main approach for governing goaf is grouting filling, and the cement-based materials are main materials of grouting filling method. However, this method has a high cement consumption, without environmental and economic advantages.

Red mud is a solid waste residue generated from the digestion process of the bauxite ores with caustic soda for alumina production. The effective disposal of red mud remains a worldwide issue, most of which is stored by the stacking method, bringing environmental challenges such as soil salinization, groundwater pollution, dust pollution, etc. (Sutar 2014). The alternative utilization efforts have been mainly focused on three aspects: the first is the extraction and recovery of the valuable metals (Şayan & Bayramoğlu 2004; Kumar et al. 2006; Lu et al. 2012; Yang et al. 2016); the second is the production of the construction materials (Kalkan 2006; Feigl et al. 2017; Tang et al. 2018) and the third is the production of absorbents (Gupta et al. 2001; Tor et al. 2006, 2009; Huang et al. 2008; Garau et al. 2011). In general, the utilization and safe disposal of the bulk amounts of red mud are currently needed.

Therefore, using red mud, fly ash and other solid waste to prepare cementitious filling materials for goaf treatment instead of cement-based materials cannot only consume a large amount of red mud to reduce environmental pollution, but also contributes to energy conservation and pollution reduction, showing the remarkable social and economic benefit. However, the presence of high alkalinity and polluting elements restrict the recovery and utilization of red mud from the industrial solid waste (Zhang et al. 2016). Previous studies have reported that the content of the heavy metal elements such as Pb, As, Cr and Cu in red mud is high, and F⁻ is the main polluting anion in the red mud production process (Liu et al. 2014; Xue et al. 2016). To achieve optimal recovery and utilization of red mud based on safety, it is necessary to study the mechanism of the solidification performance of RMFM. Therefore, the micro-morphology of RMFM samples were observed under SEM in this paper, and then analyzed its stabilization/solidification mechanisms of materials to contaminants.

When using alkaline RMFM in practice, it is inevitable to contact with acid goaf water, and then neutralization happens, thus affecting the solidification performance of RMFM for contaminants and the composition of goaf water, which in turn cause secondary pollution to the surrounding environment. Therefore, the possible impact on the surrounding environment caused by the contact between RMFM and goaf water must be taken into account.

Plenty of studies have shown that goaf water is produced as a result of the oxidation of sulfur-rich coal or gangue/spoils (He et al. 2019; Zhu et al. 2020) such as pyrite (FeS₂) (Lazareva et al. 2019), and subsequent contraction with water for the formation of sulfuric acid (Mokgehle et al. 2019). Basically, goaf water is characterized by low pH value. When neutralization between acid goaf water and alkaline RMFM happens, the solidification performance of RMFM and the composition of goaf water will be affected, and the pH of solution exerts a significant effect on the reaction result and extent. In addition, after coal is mined out, the original reducing environment is gradually changed into an oxidizing environment due to aeration, and after the goaf grouting filling treatment, the goaf can become a closed reducing environment again, but the change of redox environment has varying degrees depending on the filling compaction rate. E_h (redox potential) represents the relative degree of environmental oxidizability and reducibility, and is an index of environmental redox state reflected by potential. As a basic factor of chemical and biochemical action in the environment, redox can affect the chemical behavior, migration ability and biological effectiveness of element by changing its distribution and existent morphology, thus has become a crucial factor that influences environmental behaviors.

Therefore, the leaching tests of RMFM under different pH and E_h conditions were designed to investigate the effects of environment on the solidifying effect of Cu, Pb, Cr, As and F⁻ by RMFM. To simulate the actual condition, the static immersion tests of RMFM in goaf water was also designed to supply the theoretical evidence for the feasibility of RMFM. Finally, primarily investigated the treatment effectiveness of RMFM on acid goaf water in this study, which provides new ideas for comprehensive treatment of mined-out areas.

The raw materials used to make RMFM include red mud, fly ash, cement, slag powder and desulfurized gypsum. These materials were supplied by Shanxi YunQuan Geotechnical Engineering Technology Co. Ltd, and their chemical compositions are summarized in Table 1. Red mud was the product of the Bayer process, and its pH value was measured to be 12.73 by Soil Quality-Determination of pH (NY/T 1377-2007) analysis procedures.

Goaf water around an abandoned coal mine in Yangquan City, Shanxi Province was observed to be yellowish brown and the water quality proved to be poor. As can be seen from Table 2, the goaf water sample is acidic, and the ion concentrations of F⁻, Pb and Cu all exceed their respective contamination limits in groundwater (GB/T14848-2017), with the exceed multiples of 2.6, 15.6, 4.07 times respectively.

RMFM was synthesized at mixing red mud, fly ash, cement, desulfurized gypsum and slag powder ratios of 55:30:7.5:5:2.5, and water-solid ratio of 1:1.2 (the proportion can effectively meet the performance and mechanical properties needed from a filling material). For subsequent static immersion test, a batch of cylindrical samples with a diameter of 22.3 mm and a height of 53.5 mm were prepared. The raw materials were mixed with water for 30 min, poured into molds, shaken for 5 min to eliminate the bubbles introduced during the mixing process, and cured for 28 d at room temperature (∼23 °C) and relative humidity of 40–50%. In addition, the powder samples with a particle size less than 45 μm were prepared by grinding the cured cylindrical samples for leaching tests.

The microstructures of the sample sections were examined with a scanning electron microscope. First, the SEM images of RMFM samples that shielded desulfurized gypsum (at red mud: fly ash: cement: slag ratio of 55:35:7.5:2.5, water-solid ratio of 1:1.2) maintained for 3 d, 7 d and 28 d were observed, and the formation mechanism of hydration products in an alkali activated system was analyzed. In addition, the sulfate activated principle was analyzed by comparing the micro-morphology of fly ash particles in samples with and without desulphurization gypsum. The reaction mechanisms and hydration products of the samples were investigated to gain insights into the stabilization/solidification mechanism by RMFM to heavy metals.

The actual water environment in goaf is complex. In this study, the effect of the water environments with different pH and E_h values have been studied on the solidification performance of RMFM.

1 M HCl and NaOH (analytical grade) as well as deionized water were used to configure five groups solutions with different pH values. The measured initial pH values were 2.33, 4.11, 6.86, 9.48 and 12.03, which were marked as T1, T2, T3, T4 and T5, respectively. The prepared powder samples were continuously oscillated and extracted in each solution for 8 h with a mass ratio of 1:20, repeated twice. After standing for 16 h, the leachate was collected and pressure filtered through a 0.45 μm filterpaper. Subsequently, the pH value and ion concentrations of Cu, Pb, Cr, As and F⁻ were determined.

Using deionized water, hydrogen peroxide and solutions with addition of dissolved organic matter (DOM), five groups of solutions with different E_h value were configured with the platinum and Ag-AgCl (reference) electrodes, which were marked as T6, T7, T8, T9 and T10, respectively. The initial E_h value of T6, T7, T8 (representing different concentrations of the solution with addition of dissolved organic matter) were −94, 4 and 89 mV respectively, representing the reducing environment, whereas the initial E_h values of the T9 (aerated deionized water) and T10 (made of hydrogen peroxide and deionized water) were 220 and 430 mV respectively, representing the oxidizing environment. The leachate was collected and filtered after leaching, followed by the determination of the pH and E_h values as well as the ion concentrations of the Cu, Pb, Cr, As and F⁻ after a period of time.

The prepared powder RMFM samples and goaf water were mixed in conical bottles with a mass ratio of 1:20 and oscillated continuously on a horizontal oscillator. After leaching, the leachates were filtered and collected, repeated twice, and then the pH value and ion concentrations of Cu, Pb and As were determined.

To simulate the use of RMFM under actual condition, the cylindrical samples were statically immersed in goaf water, and the soaking duration was 8 h, 24 h, 3 d, 7 d and 14 d respectively, repeated twice. After soaking, the solution was filtered and collected, and then the concentrations of Cu, Pb, Cr, As and F⁻ were determined.

The microstructures of the sample sections were examined by using a scanning electron microscope (SX-40, Hitachi, Japan). The pH and E_h values were determined by using a pH meter (pH-3C, Shanghai Precision and Scientific Instrument Ltd, China). The concentrations of Cu, Pb and Cr were determined by using an atomic absorption spectrophotometer (AAS, ZA3000, Hitachi, Japan), whereas the As concentration was determined by employing an atomic fluorescence spectrophotometer (AFS, F-7100, Hitachi, Japan), and the F⁻ and SO₄²⁻ contents were measured through an ion chromatograph (IC-2010, Tosoh, Japan).

Alkali activation is the generic term which is applied to the reaction of a solid aluminosilicate (termed the ‘precursor’, most common aluminosilicate supplementary cementitious materials are suitable for use, with blast furnace slag, coalfly ash, calcined clays and natural pozzolans having been demonstrated to give good results) under alkaline conditions (induced by the ‘alkali activator’), to produce a hardened binder which is based on a combination of hydrous alkali-aluminosilicate and/or alkali-alkali earth-aluminosilicate phases (Provis 2018). As a kind of alkali-activated cementitious material, the high content of alkali in red mud as well as the cement hydration product Ca(OH)₂ could active the glass phase aluminosilicate in fly ash and slag, then dissolution and polycondensation occurred, thus, forming a geopolymer material with high strength cementitious network structure.

Through the SEM analysis of RMFM samples that shielded desulfurized gypsum (shielding the effect of ettringite crystals on hydration products) maintained for 3 d, 7 d and 28 d, the SEM images were compared to reveal the changes in their microstructures. As can be seen in Figure 1, with time, an increasing extent of the amorphous reticular hydration products was observed. As shown in Figure 1(a), after 3 d, a large extent of hydration products were produced and wrapped on the surface of the unreacted raw material particles, leading to the loose surface texture. As observed from Figure 1(b), for the samples maintained for 7 d, the formation of a large extent of vitreous hydration products was observed, with smooth, flat and dense cross-section. However, owing to the small volume, numerous vitreous units and extensive crystal phase interfaces, a large extent of grain boundaries could be easily destroyed by the external forces, thus leading to poor overall strength of the samples. As shown in Figure 1(c), for the samples maintained for 28 d, the hydrated products developed into a large-volume vitreous body with extremely dense structure and almost complete hydration. With time, the bonding and overall strength of the sample gradually increased.

The SEM results show that during the initial stage of reaction, a large fraction of alkaline substances in red mud gathered around the fly ash and slag particles and broke into the vitreous body, thus, resulting in the rupture and recombination of the Si-O and Al-O bonds as well as the formation of a large number of [SiO₄]⁴⁻ and [AlO₄]⁴⁻. Subsequently, the Ca²⁺ interacted with these ions to form hydration products such as C-S-H (calcium silicate hydrate), C-A-S-H (calcium silicoaluminate hydrate) gels, etc. The network structure of these hydration products tightly encased the particles. On one hand, these hydration products overlapped with each other to form a network skeleton. On the other hand, these hydration products were filled and interspersed in the pores of the material, reducing the porosity and optimizing the pore structure, improving the density of the material.

Figure 2(a) and 2(b) show the comparison of RMFM sample microstructures without desulfurized gypsum and those containing desulfurized gypsum maintained for 7 d. The addition of desulfurized gypsum introduced a large amount of SO₄²⁻ into the system, SO₄²⁻ could combine with AlO₂⁻ to form Aft, and consumed a large amount of calcium and aluminum ions in the solution, which significantly improved the dissolution rate of fly ash particles. As shown in Figure 2(b), the surface of fly ash particles was rougher, accompanied by the formation of pits, which can effectively stimulate the activity of fly ash.

Sulfate activation is essentially the combined actions of alkali and sulfate activation. In general, the activity of fly ash and slag is not enough to just add sulfate. Only when a certain amount of sulfate is added in the alkaline environment, the activity of the raw material can be brought into full play and a higher gelling strength can be obtained.

Under the combined actions of alkali and sulfate activation, the surface particles of the fly ash and slag in the material were seriously eroded, thus, releasing a large amount of activated silicon and aluminum, and the combination with the Ca²⁺, OH⁻ and SO₄²⁻ ions in the external environment led to the formation of the hydration products such as C-S-H, C-A-H (calcium aluminate hydrated), AFt (ettringite), etc. These amorphous gels and needle-like AFt were noted to be interspersed in the system, thus improving the overall connectivity and reducing the porosity. Macroscopically, it represents the enhancement of the mechanical properties and performance.

As a kind of alkali-activated cementitious material, RMFM can effectively stabilize/solidify heavy metals due to its special hydration mechanism and hydration products. And its solidification mechanisms of heavy metals mainly include physical adsorption, encapsulation, and chemical reaction as well as ion exchange.

The physical functions include gelation and physical encapsulation. In the process of hydration reaction, the amorphous aluminosilicate gels in fly ash and slag were hydrated and dissolved by alkaline substances, and gel reaction occurred to form oligomers, which were gradually dehydrated and polymerized to form a network skeleton. During this process, red mud, fly ash and other particles in the reaction could be encapsulated by oligomers in polymers, keeping it unable to contact with leaching medium, thus, the heavy metals contained in the materials could also be encapsulated during the gelation process. The precipitates formed by the reaction of alkaline hydroxide and heavy metals could also be encapsulated in this process. Furthermore, due to the formation of fish-scale-like products on the surface of heavy metal solidified bodies, these products are staggered and dense, thus, the heavy metal ions sealed by oligomers could be firmly contained in the polymers, and, further, the leakage of pollutants was avoided, showing good solidification performance macroscopically.

The chemical functions include chemical absorption and the formation of chemical precipitations. In the process of solidifying metal cations, these ions did not change the structures of [SiO₄]⁴⁻ and [AlO₄]⁴⁻, rather they participated in the formation of cementitious materials through ion balance. Through chemical interactions, metal cations such as Na⁺ and K⁺ were bonded to Al-O and Si-O bonds or existed in the skeleton cavities to maintain charge balance. And researches also demonstrated that when heavy metals such as lead ions have a similar ionic radii to alkali metal ions, alkali metal ions from the network structure can be replaced by these heavy metals to maintain the charge balance of the aluminum-oxide tetrahedron, thereby immobilizing the heavy metals. In addition to being directly encapsulated in the gel process, some heavy metals could form precipitates and then be encapsulated as well. For example, lead ions can react with the dissolved silicon to form Pb₃SiO₅ (Zhang et al. 2008).

The C-S-H gel is the main hydration product of cementitious materials, which has a large specific surface area and ion exchange capacity, in which the Ca ions could be replaced by Pb and Cr in the interlayers to connect with other elements. Further, the structure of the AFt crystals formed after the addition of desulfurized gypsum reveals that the heavy metal ions could be stabilized in the gaps and channels inside the crystal columns by chemical replacement. Other harmful ions could be precipitated as hydroxides or carbonates.

The mechanisms of RMFM stabilization/solidification of heavy metals, in addition to the above-mentioned main functions of physical encapsulation, physical adsorption, ion exchange as well as chemical precipitation, also include surface complexation and passivation. These functions complement each other and jointly realize the stabilization and solidification of heavy metals (El-Eswed et al. 2017; Zhang et al. 2019).

In this group of leaching tests, the pH values of T1–T5 in leachate after 8 h of continuous oscillations are shown in Table 3. As can be seen, the pH values of each treatment were close and exhibited alkaline.

Figure 3 shows the concentrations of leached Cu, Pb, Cr and As of the treatments 1–5 under different pH. As can be seen directly from Figure 3, with the increase of pH, the leaching concentrations of Cu, Pb, Cr and As decreased, generally showing a linear trend. Therefore, it is suggested that the amounts of the heavy metals leached from RMFM had a specific negative correlation with the pH value, and the highest concentrations all appeared at T1. Among them, the ion concentration of Cu in T1 was 0.0359 mg/L, which was 40.1%, 56.8%, 57.7%, 57.7% higher than that in T2, T3, T4 and T5, respectively; the leaching concentration of Pb in T1 was 0.0851 mg/L, which was 11.8%, 22.6%, 24.4%, 28.2% higher than that in T2, T3, T4 and T5, respectively; the leaching concentration of Cr in T1 was 0.0886 mg/L, which was 6.8%, 37.9%, 68.8%, 71.4% higher than that in T2, T3, T4 and T5, respectively; and the leaching concentration of As in T1 was 0.0005 mg/L, which was 40% higher than the other treatments. It is suggested that the concentrations of leached Cu, Pb, Cr and As have similar rules as a function of pH. The stronger is the acidity of the extractant, the higher is the concentration of leached heavy metals; while in neutral and alkaline environments had significantly lower concentrations of heavy metals.

In acidic conditions, the acidic liquid entered the RMFM samples through the capillary pores during the initial stage of leaching and changed the original alkaline environment, and then dissolved partial geopolymer gels, resulting in worse overall adhesion, thus affecting the physical and chemical solidification effects on heavy metals. Furthermore, the alkaline environment could enhance the degree of alkali activation of the system, thus promoting the formation of more hydration products and raising the effect of solidification performance to some extent.

It can be noted that even in T1 (pH = 2.33), the leaching concentrations of heavy metals were still at low levels (<0.15 mg/L, all at trace levels), indicating that RMFM had effectively resisted acid corrosion and maintained good stability/solidification of heavy metals.

Figure 4 shows the concentrations of F⁻ in the leachates of the treatments 1–5. As the figure shows, with the increase of pH, the general tendency of F⁻ concentrations gradually decreased and then gradually rose whereas the lowest concentration turns out in T3, which appears neutral with a pH of 6.86.

Therefore, in alkaline conditions, the stronger is the alkalinity, the greater the mobility of F⁻; in acidic environments, similar to the leaching behaviors of heavy metals, the acidic substances could partially dissolve and destroy the geopolymer gels, thus, affecting the ability of the material to immobilize the pollutants. From this it's clear that the neutral environment is beneficial in controlling the mobility of fluorine.

In this set of leaching tests, T6–T8 were subjected to the reducing environments, whereas T9 and T10 were treated with the oxidizing environments. Before this set of leaching tests, the E_h value of RMFM with deionized water as extractant was detected. The results showed that the E_h value of the leachate was 70 mV, showing reducibility. The reducing sulfur in the materials was in a larger proportion, indicating that the RMFM have a certain potential for acid forming.

Table 4 shows the E_h values of the leachates after extraction. The E_h values of the five treatments were noted to increase after continuous oscillation for 8 h, and after leaching, the pH values of T6-T10 were 7.90, 8.03, 8.72, 4.33 and 0.08, respectively. Under reducing treatment, the leachates of T6, T7 and T8 were neutral, while in oxidizing environment, the leachates of T9 and T10 showed acidity, especially in T10 had strong acidity.

After leaching, the concentrations of SO₄²⁻ in the leachate were detected by ion chromatograph. The concentrations of SO₄²⁻ in the leachates of T6-T10 were 1.414, 1.601, 1.885, 2.157 and 3.72 g/L, respectively. It is noticeable that the stronger is the oxidizability of the extractant, the higher is the SO₄²⁻ concentration in the leachate, which indicates that the RMFM samples underwent a typical oxidation acid forming process in the strongly oxidizing environments, and then resulted in high E_h and low pH, along with a large extent of heavy metals and sulfates.

Figure 5 shows the leaching concentrations of Cu, Pb, Cr and As in T6–T10 under different redox environments. As shown in Figure 5(a) and 5(b), the leaching rules of Cu and Pb exhibited similarly, both presented a gradual increment in concentration on enhancing the extractant oxidizability. The highest leaching concentrations all appeared at T10, which showed strongest oxidizing, and the concentrations were much higher than the other treatments. Among them, the leaching concentration of Cu ions in T10 was 0.4053 mg/L, which was 96.8%, 93.6%, 91.8%, 78% higher than that in T9, T8, T7 and T6, respectively; the leaching concentration of Pb ions in T10 was 0.4123 mg/L, which was 97.8%, 95.7%, 88.9%, 80.6% higher than that in T9, T8, T7 and T6, respectively. In oxidizing environments, the stronger the oxidizability of the leachate, the more acidic substances were produced, the more dissolution of geopolymer gels, which led to more leached Cu and Pb ions. However, the pH values in T6–T8 under reducing treatment were close and exhibited neutral, thus showing lower leaching concentrations of Cu and Pb. Further, although the pH values of the three reducing treatments were close, the stronger the reducibility of the leachate, the significantly lower the concentrations of Cu and Pb, indicating that the reducing environments could inhibit the release of pollutants to a certain extent.

Figure 5(c) shows that the highest leaching concentrations of Cr also appeared at T10, which was 4.531 mg/L, and was 40.1%, 56.8%, 57.7%, 57.7% higher than that in T6, T9, T7 and T8, respectively. Different from Cu and Pb, the lowest concentration appeared at T8. In oxidizing environments, the acid forming process of RMFM was also the main reason for the high leaching concentration of Cr. As one of the transition elements, trivalent Cr and hexavalent Cr are two main states of the Cr under natural conditions; the mobility depends on its specific chemical forms and on its binding state. Under the condition of acidity and low redox potential, Cr(VI) can be easily reduced to Cr(III), and the Cr(III) in solutions is noted to be the most thermodynamically stable, which is poorly soluble in water and often exists in the form of insoluble hydroxide. By contrast, Cr(VI) has high solubility and strong mobility, and has been an important factor for groundwater and soil pollution (Fendorf 1995; Shanker et al. 2005). Due to the strong oxidation, Cr was almost in hexavalent state in T10, and the strong acidity further enhanced Cr concentrations. In the other treatments, Cr existed mainly in trivalent state, resulting in obviously lower concentrations.

As was almost in hexavalent state in T10, which could form stable precipitates with ferric iron in the leachate, thus, free state of As could not be detected.

Figure 6 shows the leaching characteristics of F⁻ under different redox environments. The leaching concentrations of F⁻ in T6–T8 were 0.93 and 17.74 mg/L, respectively, and no free fluorine was detected in T9 and T10. The stronger the oxidizability of the leachate, the higher the F⁻ concentration was. It is known that in the presence of Ca²⁺, the concentration of F⁻ is closely related to pH value. The leachate of T6-T8 were neutral, only in the neutral environment, the activity of Ca²⁺ is high, due to which F⁻ led to the most stable form of precipitation, thus resulting in low concentration of F⁻, while in T9 and T10, F⁻ easily migrated in the acid-oxidizing environments, showing significantly higher concentrations.

From the leaching test results, it is obvious that the RMFM samples underwent a typical oxidation acid production process in the oxidizing environments, for which the leachates resulted in low pH and high E_h. Except for As, which formed stable precipitates with Fe(III) under the strong acid-oxidizing conditions, the other polluting elements could be dissolved largely from RMFM. As a result, the oxidation environment was confirmed to enable an enhanced extent of pollutant release from RMFM.

By comparison, the concentrations of polluting ions in the reducing environment were significantly lower, exhibiting the potential to control the release of the pollutants. This may be related to the high content of the dissolved organic matter, anaerobic microorganisms and partial phosphate resulting due to the degradation of the organic matter in the anaerobic environment. On the one hand, the soluble organic matter in the reducing extractant could make the surrounding environment anoxic due to the self-oxidation under the action of the anaerobic microorganisms, thus, weakening the oxidation ability to produce acid, while the relatively high pH does not lead to the dissolution and destruction of the cementitious materials. On the other hand, RMFM contains a large amount of Al and Fe, which could be precipitated in the phosphate form with a very low solubility, and the formed colloids exhibit a specific adsorption or co-precipitation effect on the soluble ions. In addition, the organic matter in the reducing extractant could significantly affect the migration of the metal ions, such as Fe and Mn, by the formation of complexes (complexation or chelation) as well as the number and permeability of ions (through microfiltration membranes) in the solution. As a consequence, the leachates under reducing treatment had lower contaminants.

To simulate the use of RMFM under actual condition, the cylindrical samples were statically immersed in goaf water. The ion concentrations of Cu, Pb, Cr, As and F⁻ leached from the cylindrical RMFM samples within 14 d are shown in Table 5.

As shown in Figure 7, with the increase of soaking time, the ion concentration of each pollutant had no significant change at the beginning. After soaking for 72 h, the concentrations began to decrease, but the trends were not pronounced. Through the components detection of goaf water, it can be found that the goaf water itself contained a variety of pollutants. By contrast, the concentrations of pollutant elements released by RMFM samples were negligible. As the soaking time went on, partial geopolymer gels near the surfaces of the samples were gradually corroded and dissolved, resulting in the increase of the permeability, and the internal hydrates came into contact with goaf water, then a small amount of polluting elements were stabilized/solidified by physical and chemical reactions. Thus contaminated elements concentrations decreased slowly, and gradually came to stability over time.

Table 6 shows the comparison of the Cu, Pb and As concentrations of the unmodified red mud and RMFM powder samples in goaf water after continuously oscillated for 8 h.

As can be seen from Table 6, but rather than stabilize/solidify the heavy metals in the goaf water, the unmodified red mud released its own pollutants into the surroundings, resulting in secondary pollution. Meanwhile, due to the high alkalinity of red mud, the pH value of the goaf water after reaction exceeded its limit in groundwater (6.5–8.5), showing as alkaline.

By comparison, the powder RMFM samples had significant treatment effect on goaf water. After reaction for 8 h, the pH value of the leachate met the Class I of standard for groundwater quality (GB/T14848-2017), and heavy metals had also been effectively solidified. With the prolonging of reaction time, the concentrations of the heavy metals gradually decreased, and the removal efficiency gradually increased. After leaching, the removal rates of Cu, Pb, and As reached 99.34%, 96.25% and 96.15%, respectively.

The test results show that the adsorption capacity of unmodified red mud is devoid, so it can hardly be directly used to treat wastewater containing heavy metals. Thus it can be seen that red mud needs to be modified or blended with aluminosilicate powder with pozzolanic activity to produce geopolymer cementitious materials, only then could effectively treat coal mine wastewater. Compared with the unmodified red mud, the alkali-activated cementitious material RMFM could stabilize/solidify heavy metals effectively in goaf water due to its special hydration process and hydration products.

• (1)

  Based on the combined actions of alkali and sulfate activation, RMFM's special hydration process and hydration products make it stabilize/solidify heavy metal ions effectively. These hydration products such as C-S-H, C-A-H and AFt overlapped with each other, filled and interspersed in the pores of the material, reducing the porosity and improving the density of the material, showing effective solidification performance macroscopically. The solidification mechanisms of heavy metals by RMFM mainly include physical adsorption, encapsulation, and chemical reaction as well as ion exchange.

• (2)

  The leaching tests of RMFM at different pH show that the acidic substances could dissolve a part of the polymer gels and damage the density of the system, thus, affecting the solidification performance of RMFM, while the alkaline environment could enhance the degree of alkali activation of the system and produce an enhanced extent of hydration products, thus, improving the ability of the material to immobilize the pollutants to a certain extent. It can be noted that even in the most acidic treatment T1 (pH = 2.33), the leaching concentrations of the heavy metals were still at low levels (<0.15 mg/L, all at trace levels).

• (3)

  Through the leaching tests of RMFM under different redox conditions, it can be found that the concentrations of the polluting elements in the reducing environments were much lower than those in oxidizing environments. The RMFM samples underwent a typical oxidation acid forming process in the oxidizing environment. The stronger was the oxidizability, the higher was the acidity, which led to the more release of the pollutants from RMFM, while the reducing environments exhibit the potential to effectively control the release of the pollutants.

• (4)

  The static immersion test shows that RMFM could effectively stabilize its own pollutants and would not cause secondary pollution to the surroundings when in the possible exposure environment. In addition, the solidification test in goaf water shows that RMFM powder had significant removal effects on heavy metals. After continuous oscillation for 8 h, the removal rates of Cu, Pb, and As all reached more than 96.15%, providing new ideas for comprehensive treatment of mined-out area.

The authors thank the Fundamental Research Funds of Taiyuan University of Technology (No. RZ18200030) and Shanxi YunQuan Geotechnical Engineering Technology Co. Ltd for supporting this project.

All relevant data are included in the paper or its Supplementary Information.