Laboratory Study on the Effectiveness of Limestone and Cementitious Industrial Products for Acid Mine Drainage Remediation

: Acid mine tailings may affect several environmental matrices. Here, we aimed to stabilize acid-generated mine tailings using several alkaline and cementitious amendments, which were tested in columns for 361 days. The alkaline amendments consisted of 10 and 20 wt.% limestone, while the cementitious amendments consisted of different binders at a total dosage of 5 wt.% binder. The different formulations for the cementitious amendments were: 50% Kruger ﬂy ash and 50% class F ﬂy ash; 20% ordinary Portland cement, 40% Kruger ﬂy ash, and 40% class F ﬂy ash; 80% ordinary Portland cement and 20% Kruger ﬂy ash; and 20% ordinary Portland cement, 40% Kruger ﬂy ash, and 40% ﬂy ash. Kinetic testing on the amendment formulations showed that the pH values increased from <2.5 to circumneutral values (~7.5). The mobility of various chemical species was greatly reduced. Cumulative Fe released from the unamended tailings was ~342.5 mg/kg, and was <22 mg/kg for the amended tailings. The main mechanisms responsible for metal(loid) immobilization were the precipitation of secondary phases, such as Fe-oxyhydroxides, physical trapping, and tailing impermeabilization.


Introduction
Mining activities generate large quantities of solid wastes such as waste rocks and tailings. Tailings are generated after ore treatment and processing, and are characterized by fine grain-size distributions and high specific surface areas [1][2][3]. The majority of tailings are stored at the surface in a tailing storage facility (TSF). Mine tailings often contain non-negligible amounts of iron sulfide minerals such as pyrite and pyrrhotite [4][5][6]. The oxidation of sulfide minerals occurs when they are exposed to water and oxygen, both in the presence or absence of microorganisms [5]. Sulfide oxidation can involve various and complex processes, including chemical, biological, and electrochemical reactions [7]. Sulfide oxidation increases the acidity of leachate waters, which can lead to the formation of acid mine drainage (AMD). The overall reaction of pyrite oxidation result produces four moles of H + (Equation (1)). FeS 2 + 3.75O 2 + 3.5H 2 O → Fe(OH) 3 + 2 SO 4 2− + 4H + (1) be used as an option for the mitigation of the studied mine site. Moreover, this technique can be easily used for other problematic mine tailings in other countries. Table 1. Some case studies of alkaline and cementitious amendments used to stabilize contaminated materials.

Materials and Study Site
The tailings used in this study were collected from the Joutel mine site, which is an inactive mine site located in Quebec, Canada. During operation, the Joutel mine extracted gold via floatation and cyanidation [57], with the waste tailings from this process being deposited in a 120 ha TSF. Operations at Joutel ended in 1994, which has allowed for significant oxidation and acid generation to occur in some localized areas of the TSF [58]. The tailings used in this study were collected from an acidic oxidized zone in the TSF. The samples were homogenized in situ and shipped to the laboratory.
The materials used as amendments in this study were limestone (LS); fly ash (FA) that was collected from the Boralex thermal energy station in Sennterre, Quebec; ordinary Portland cement (OPC); Kruger fly ash (CVK) that was collected from the biomass cogeneration power plant at the Kruger paper mill in Brompton, Quebec; and class F fly ash (FAF), which was provided by LafargeHolcim Canada. Limestone was formed mainly by calcite and dolomite [15]; fly ash was mainly composed of plagioclases and carbonates [15]; FAF contained mainly quartz [59], plagioclase, and mullite; and CVK was composed of lime, gehlenite, quartz, and calcite.

Physical, Chemical, and Mineralogical Material Characterizations
The grain-size distribution (GSD) of all samples was analyzed using a laser analyzer (Malvern Mastersizer, Malvern Panalytical, Rouyn-Noranda, Quebec, Canada). The chemical composition of the solid samples was evaluated using inductively coupled plasma atomic emission spectroscopy (ICP-AES) (Perkin Elmer Optima 3100 RL, USA) following complete digestion by HNO 3 /Br 2 /HF/HCl (HF and HCL to dissolve silicates; HNO 3 and Br 2 to oxidize sulfur). The ICP-AES was calibrated using an international standard (SY-4). The detection limits of ICP-AES are variable and depend on the analyzed element (60 ppm for Al and Ca, 15 ppm for Mg, 5 ppm for Mn, 1 ppm for Na and K, 10 ppm for Fe, 5 ppm for Li, Pb, As and Cr, 10 ppm for Cu and 55 ppm for Zn). Total carbon and sulfur were analyzed in pulverized samples to <74 µm via infrared spectroscopy following combustion in an induction furnace (ELTRA CS-2000; detection limit = 0.09%). The mineralogical composition of tailings was investigated thoroughly using mineralliberation analysis (MLA). MLA is technique based on scanning electron microscopy with an energy dispersive spectrometer (SEM-EDS) that uses special software to automate mineral abundance quantification [60]. This technique also enables quantitative analyses of mineral liberation, mineral association, and elemental deportment. In this study, MLA was used to quantify modal mineralogy; calcite and pyrite liberation; and Ca, Mg, Fe, and S deportments.

Toxicity Characteristic Leaching Procedure and Kinetic Column Tests
Toxicity characteristic leaching procedure (TCLP) was used as preliminary test to evaluate metal(loid) mobility within the different amendment formulations ( Table 2). TCLP was primarily designed to simulate the release of heavy metal(loid)s and organics from municipal wastes [61], but is frequently used on mine tailings. In brief, around 20 g of each sample was added to 400 mL of a leaching solution composed of acetic acid [62]. Each mixture was then placed in a high-density polyethylene (HDPE) bottle, and the bottles were tumbled for 18 h at a 30 ± 2 rpm. Samples were filtered through a 0.45 µm nylon filter, and the leachates were analyzed for their pH, electrical conductivity (EC), and chemical composition (by ICP-AES). Kinetic column tests (KCTs) are widely used to predict the environmental behavior of mine tailings, as well as to test different reclamation scenarios for acid-generating tailings [63]. Column tests are recognized as reproducible tests [64], and under the right conditions, they may be able to provide insight into the environmental behavior of wastes under field conditions. The column tests were monitored for more than 365 days. The long-term geochemical monitoring of columns presents several advantages, such as: (i) considering the possible dynamic changes, (ii) using a large quantity of sample (around 10 kg in this study), and (iii) allowing possible precipitation reactions. Seven amendment formulations were selected based on laboratory batch tests and tested in KCTs as indicated in Table 2. Each sample was placed in an acrylic column with a diameter of 14 cm and a height of 100 cm. Each column's inner wall was covered with grease to avoid the creation of preferential flow paths during the flushing of the samples [65]. Column C1 contained a reference sample of pure mine tailings. Columns C2 and C3 contained alkaline amendments based on the LS addition.
The leachates from each column were analyzed for pH and EC using individual probes, while the acidity and alkalinity were analyzed using an automated titrator (Metrohm Binkmann, 716 DMS Trinitro, Rouyn-Noranda, Canada). The chemical composition of the leachates was determined by ICP-AES for samples filtered to 0.45 µm and acidified to 2% v/v HNO 3 . The chemical quality of the leachates from the column tests was integrated in Visual Minteq (Version 3.1) to calculate the saturation indexes of the ferrihydrite and goethite [66]. Saturation indexes were calculated using the standard database of Visual Minteq.

Physical and Chemical Characteristics
The GSDs and chemical characteristics of the studied samples are presented in Table 3. Grain-size distributions were relatively heterogeneous. The D90 values, which correspond to the diameter at which 90% of particles pass on the cumulative grain-size distribution curve, were about 30.4, 4500, 46.7, 1500, 104.2, and 89.1 µm for the mine tailings, LS, OPC, FA, CVK, and FAF, respectively. The mine tailings had a similar GSD to that of the OPC. Table 3. Grain-size distribution and chemical composition of the studied samples (MT = mine tailing, LS = limestone, OPC = ordinary Portland cement, FA = fly ash, CVK = Kruger fly ash, FAF = fly ash class F, and DL = detection limit).  Table 3.

Mineralogical Composition of the Mine Tailings
The mineralogical composition of the highly oxidized mine tailings was investigated using mineral-liberation analysis. The tailings' modal mineralogy and mineral-liberation degrees are presented in Figure 1. The tailings sample contained approximately 31 wt.% of secondary iron oxides, 20 wt.% of quartz, 16 wt.% of albite, 11 wt.% of chlorite, and 9 wt.% of gypsum. Secondary iron oxides were present as a cementitious material that coated various minerals, especially pyrite and silicates (Figure 2A-D). Other minerals were detected in varying concentrations, such as pyrite, chalcopyrite, and calcite. The pyrite content was about 0.44 wt.%, despite a total S content of about 4.36 wt.%. This suggests that a major part of pyrite was oxidized in situ and that the resulting sulfate precipitated as gypsum. As a result of sulfide oxidation, carbonates dissolved to buffer the produced acidity. Thus, residual calcite was present in trace amounts of about 0.19 wt.%. Comparing the sulfide content with the carbonate content allows prediction of the acid-generating behavior for the mine tailings, since it is recognized that the balance between acid-forming and acid-consuming minerals can inform on the environmental behavior of mine tailings (acid-generating or not) [13,67,68]. However, the reactivity of these minerals is influenced by several parameters, such as their available reactive surface area [69][70][71][72][73]. Therefore, the degree of liberation for calcite and pyrite was analyzed by MLA, as shown in Figure 1B. The degrees of liberation for calcite and pyrite were only slightly different. Free (liberation degree = 100%) calcite and pyrite represented around 47.6 and 55.5% of the total calcite and pyrite, respectively, which meant that the completely available reactive surface of pyrite was slightly higher than that of calcite. The locked part of calcite and pyrite was about 23 and 20%, respectively. Around 7.5% of calcite and pyrite were mid-liberated (between 20 and 70%).  . Iron occurred mostly in iron oxides, followed by chlorite, actinolite, and pyrite. Around 84% of iron was present in iron oxides, 11% came from chlorite, and small amounts of iron were associated with actinolite and pyrite (<2%). Magnesium mainly occurred in chlorite and actinolite; 74% of magnesium came from chlorite, and 26% came from actinolite. Sulfur was mainly associated with gypsum and pyrite; 87.5% of S came from gypsum, and 12% of S came from pyrite. Small amounts of S were also associated with chalcopyrite (<0.2%).  Table 4 presents results of TCLP tests that were carried out to evaluate metal(loid) mobility following different amendment treatments. The performance of each amendment formulation was slightly different. Both alkaline (C2 and C3) and cementitious formulations (C3, C4, C5, C6, and C7) increased the pH of the leachates compared to the unamended tailings, indicating the dissolution of alkaline phases that buffered the produced acid. The pH of the leachates from the unamended tailings was about 2.5, while pH values of that from amended tailings ranged from 4.92 for C6 (80%OPC-20%CVK) to 6.46 for C2 (20 wt.% LS). The analyzed EC was similar for all samples, ranging between 7.81 mS/cm for C1 and 9.37 mS/cm for C2. The application of the different amendment formulations reduced Al, Fe, Ni, and Zn releases compared to the unamended tailings. Aluminum release from the unamended tailings was about 0.4 mg/L, but did not exceed 0.05 mg/L in any of the amended tailings. Iron releases were reduced from 0.5 mg/L from the unamended tailings to a maximum of 0.08 mg/L from C6. Nickel releases were around 0.01 mg/L for the unamended tailings and less than the detection limit of the analytical method for all amended tailings. Zinc concentrations were about 0.6, 0.34, 0.33, 0.30, 0.60, 0.31, and 0.30 for C1, C2, C3, C4, C5, C6, and C7, respectively. Finally, S releases were between a minimum value of 567 mg/L from C2 to a maximum value of 766 mg/L from the unamended tailings.

Kinetic Column Tests
The water quality of leachates from the KCTs are presented in Figures 4-6. The evolution of pH in the leachates from the KCTs ( Figure 4A) varied widely among the tested samples. The mine tailings showed an acidic behavior characterized by pH values ranging between 2.1 and 2.8. Thus, there was a relatively small variation in pH values from the beginning to the end of the tests. However, the amended formulations raised pH values to circumneutral values. Average pH values were about 7.7, 7.2, 7.7, 7.4, 7.7, and 7.5 for C2, C3, C4, C5, C6, and C7, respectively, and these values were stable throughout the duration of the KCTs. Average EC values ( Figure 4B) were about 4.2, 4.24, 4.93, 5.7, 6.8, 5.3, and 5.2 mS/cm for C1, C2, C3, C4, C5, C6, and C7, respectively. At the beginning of the leaching tests, EC values were notably higher than at the end of the tests. This is likely due to the redissolution and release of secondary mineral phases that are naturally present in preoxidized tailings. Acidity and alkalinity values were in stark contrast between the amended and unamended materials ( Figure 4C,D). Acidity in leachates from unamended tailings started at around 1400 mg CaCO 3 /L, then stabilized at around 300 mg CaCO 3 /L. However, average acidity values in the amended tailings were 23, 60, 33, 45, 31, and 42 mg CaCO 3 /L for C2, C3, C4, C5, C6, and C7, respectively. The results for acidity were mirrored by the results for alkalinity ( Figure 4D). Leachates from the unamended tailings (C1) showed negligible alkalinity values. The amended tailings showed leachates with non-negligible values.
At the end of the KCTs, water-infiltration speed was measured for the different column tests. The main objective was to compare the hydraulic conductivity reduction of the different formulations. The results from the water-infiltration tests are presented in Figure 7. Samples C1, C2, and C3 drained water more quickly than the other samples. At 4000 min, about 68 mL was drained from C1, 69 mL was drained from C2, and 60 mL was drained from C3. In contrast, 30 mL was drained from C4, 49 mL was drained from C5, 48 mL was drained from C6, and 38 mL was drained from C7. In general, the cementitious amendments decreased water-infiltration rates through the tailings the most. This is likely explained by tailing cementation and cohesion processes that reduced the hydraulic conductivity of the amended materials [76].

Discussion
The unamended tailings (C1) showed an acidic behavior; pH values were lower than 3 from the very first stages of leaching ( Figure 4). Consequently, the acidity of the leachates from the unamended tailings reached high levels compared to that of the amended tailings. This could be explained by the mineralogical composition of the mine tailings. Their carbonate content was lower than 0.2 wt.%, and the calcite liberation degree was around 50%. This meant that the reactive surface of calcite was reduced by half, as suggested by several authors who demonstrated that a mineral's reactivity is influenced by its liberation degree [69][70][71][72]77]. Consequently, the tailings' neutralization potential was likely negligible, assuming that the NP was influenced mainly by its carbonate content [5,11]. Mineralogical analyses of the mine tailings revealed that the acidity could be from two sources: (i) pyrite oxidation and (ii) latent acidity due to the release of exchangeable Fe [65,78,79]. Exchangeable Fe explains the high concentrations of Fe released within the unamended tailings during the first stages of the leaching tests. Moreover, considering environmental regulations in Canada, such as Directive 019 in Quebec [80], the mine tailings used in this study are considered as problematic and require a reclamation scenario to prevent environment contamination. Instantaneous Fe concentrations in leachates from the unamended tailings were around 78 mg/L, which exceeded the regulatory limit of 3 mg/L set by Directive 019 (Quebec, Canada). The same tendency was observed for Cu, Pb, and Zn, which also exceeded regulatory limits. Average instantaneous concentrations of Cu, Pb, and Zn were 20.3, 1, and 11.35 mg/L, respectively; while their average concentrations should not exceed 0.3, 0.2, and 0.5, respectively.
The geochemical behavior of limestone-and cement-based amendments (Table 2) enabled the neutralization of acidity in the Joutel mine tailings. Following the amendment, pH values were maintained at circumneutral values (around 7.5), and the leachates of the amended tailings were less concentrated in several significant, toxic metals. The two categories of amendments used in this study were: (i) alkaline amendments (10% and 20% limestone) and (ii) cementitious amendments (based on OPC, FAF, CVK, and FA). Acid buffering by all amendments was performed through carbonate dissolution [48,[81][82][83]. Depending on the pH of the system, two reactions of calcite (carbonate) dissolution could occur (Equations (3) and (4)): For the cement-based amendments, cement hydration and the precipitation of hydrated phases could produce additional alkalinity [23,43]. For example, the dissolution of two moles of tricalcium silicate (C 3 S) produced by cement hydration produces 8 moles of OHfollowing the reaction (Equation (5)): The mobilities of the chemical species, including Al, Cr, Cu, Fe, Pb, S, and Zn, were greatly reduced through the use of both the alkaline and cementitious amendments. Reduction factors, calculated using Equation (6), are presented for Al, Cr, Cu, Fe, Pb, S, and Zn in Figure 8. The reduction factor for each element was calculated based on its average concentration in leachates from the unamended and amended tailings [15]: where RF i is the reduction factor for element i; C a is the cumulative concentration (mg/kg) for the element i from amended column (C2 to C7); and C 0 is the cumulative concentration (mg/kg) for the element i in the leachates from oxidized tailings (control column; C1). The RF Al was around 100% for all amendments tested in this study. The RF Cr was about 100%, 81%, 100%, 100%, 100%, and 85 % for C2, C3, C4, C5, C6, and C7, respectively. RF Cu , RF Fe , RF Pb , and RF Zn values were >95% for all amendment formulations tested in this study. Sulfur was immobilized less compared to other elements. Indeed, S immobilization was only about 50%, 37%, 4%, 97%, 10%, and 13% for C2, C3, C4, C5, C6, and C7, respectively. Metals immobilization within cementitious and alkaline amendments is ensured by several chemical and physical mechanisms [15,27,48,84]: i.
Alkaline amendments: adding alkaline materials, such as limestone, enhances the neutralization potential of the mixture. This shifts the balance between the acid-generating potential and neutralization potential in a given tailing [20,85]. Acid buffering is reflected by an increase in the alkalinity and pH of the leachates (Figure 4). Changing the geochemical conditions of the system influences the elemental deportment. This means that a number of secondary phases may precipitate [13], the most notable of which are Fe-oxyhydroxides such as ferrihydrite and goethite, and sulfates such as jarosite [68,86]. Although several mechanisms are responsible for metal(loid) attenuation, the most discussed mechanisms are Fe-oxyhydroxide precipitation, and sorption mechanisms [87][88][89]. Moreover, saturation indexes for ferrihydrite and goethite calculated using Visual Minteq (V3.1) showed that these two minerals will likely precipitate (saturation index >0) for the amended tailings ( Figure 9A,B). However, for the unamended tailings, these two mineral species will remain in solution (saturation index <0). Dissolution for Fe-oxyhydroxides within the unamended tailings and their precipitation within the amended tailings explains the high concentrations of Fe analyzed within the leachates from the Joutel tailings ( Figure 6A). ii. Cementitious amendments: the same mechanisms discussed within alkaline amendments also occur within cementitious amendments, as they also neutralize acid and increase pH (Figure 4). However, the use of cementitious additives also improves the mechanical properties of the mixtures [23,47,50]. The use of cement and industrial byproducts increases the cohesion and the long-term impermeabilization of the mixture [15]. Consequently, contaminant mobility and diffusion are significantly limited by (i) physical trapping, and (ii) reducing the reactive surface area of the mixture. Tailing solidification reduces the available surface area of sulfide minerals, and thus their reactivity [35,36]. Performances of the tested amendments were significantly different considering results of the TCLP tests and KCTs. Column tests are more realistic than TCLP tests due to several factors: (i) sample size-within column tests, higher sample masses (>10 kg) are used compared to those used in TCLP tests (20 g); (ii) liquid/solid ratios (LSR)-the LSR in TCLP tests is about 20, while in column tests it is around 0.2; thus, dilution of elements may occur within the TCLP tests; (iii) test duration-KCTs and TCLP tests are performed on different time scales (18 h for TCLP tests and 361 days for columns); and (iv) leaching solutions-the solution used for the KCTs was deionized water, and that used for the TCLP tests was acetic acid. Therefore, the reactivity of the mineralogical phases will be different. For these reasons, KCTs may represent more accurate tests for evaluating the performance of amendments to stabilize acid-generating tailings. Moreover, Ghosh, Mukiibi [90] showed that TCLP tests may underestimate leaching of some chemical species such as arsenic.
Considering metal(loid) immobilization, both the alkaline and cementitious amendments showed relatively similar effectiveness. However, several factors make the formulations based on industrial byproducts better. The formulations based on limestone require a quarry for carbonate extraction. This represents an additional environmental impact to resolve another environmental problem (AMD). Consequently, the formulations based on industrial byproducts (e.g., fly ashes) could be more interesting. Substitution of byproducts for OPC could also reduce costs related to cement production and valorize some industrial wastes. Teixeira, Mateus [91] showed that the substitution of cement with fly ashes is a viable alternative because it allows the reduction of the environmental footprint of cement production. The use of industrial byproducts in the S/S technology is an ecofriendly option, but it must not affect the contaminants stabilization. This study proved that acid-generating and contaminated tailings can be stabilized using limestone and cementitious additives. Similarly, the use of industrial by-products in the stabilization of mine tailings needs more considerations during the rehabilitation of mine sites. Moreover, a technical and economic study is required to select the best scenario among those tested in this study. Additionally, geochemical modeling of the leachates from the different columns can be useful to predict the long-term efficiency (>10 y) of the tested formulations. Similarly, field tests are required to test the in situ effectiveness of the formulations used in this study. However,   [15] have already demonstrated that alkaline and cementitious amendments are efficient to neutralize acid-generating tailings and prevent contaminants release by AMD.

Conclusions
The reclamation and management of tailing storage facilities is a crucial step in the life cycle of a mine. The objective of TSF reclamation is to reduce or eliminate the risk of environment contamination due to acid and/or metal(loid) release. The mine tailings characterized in this study were acid-generating. Their neutralization potential was negligible due to carbonate depletion (calcite content <0.2 wt.%). Consequently, kinetic tests showed that the mine tailings generated AMD with no lag time. The mine tailings presented the risk of releasing Al, Cr, Cu, Fe, Pb, and Zn into the environment. Iron concentrations were significantly higher (78 mg/L) than the limit set by current regulations (3 mg/L). To stabilize these acid-generating tailings, amendments based on limestone, ordinary Portland cement, and fly ash were tested. All amendments showed promising results with respect to immobilizing potentially harmful chemical species. Indeed, pH values of the amended tailings immediately increased to circumneutral values (around 7.5). Moreover, the reduction factors of Al, Cr, Cu, Fe, Pb, and Zn releases were all higher than 80%. Application of this technique in the field at full scale will require a thorough economic evaluation of individual amendment formulations.