Comparative study of acid mine drainage using different neutralization methods

Two-stage acid mine drainage (AMD) neutralization residue yield is 2.57 t/t acid equivalent using limestone and lime, 15.74% less than the 3.05 t/t acid equivalent yield from single-stage lime neutralization residue. XRD results shows that the main residue component generated by two-stage neutralization is CaSO4•0.5H2O, with a molecular weight 15.70% lower than the CaSO4•2H2O generated in single-stage neutralization. The differing amount of gypsum crystallization water from the different neutralization processes is the main cause of the different residue yields. While the actual removal rate and pH of Fe3þ are consistent with the theory, those for Zn2þ, Cu2þ and Al3þ are inconsistent with the theoretical values, the error rate increasing with increasing pH. Co-precipitation with and adsorption by Fe(OH)3, mainly generated during neutralization, are the main reasons for the difference. The cost of two-stage neutralization is 16.60% less than that of single-stage neutralization when the initial reaction pH is 3.4, but the unit cost begins to increase if the initial reaction pH is raised above 3.4. When the initial pH is 4.0, the cost of two-stage neutralization exceeds that of single-stage neutralization by 4.07%.


INTRODUCTION
Acid mine drainage (AMD) is a major challenge to the mining industry due to its environmental consequences. The main cause of AMD is oxidation of sulfide mineral ores, initially exposed by mining activities; for example, through open-pit mining areas, waste rock disposal, and so on. The components of AMD differ (Torres & Auleda 2013) depending on the mine's geological conditions and the metal minerals present. Pyrite is one of the main sources of AMD due to its relative ease of oxidization (Kefeni et al. 2017). AMD is produced by both active and closed mines ( Johnson & Hallberg 2005;Le Pape et al. 2017), and AMD yield usually increases sharply during mine development.
Treatment processes differ according to AMD chemical composition and purpose. Neutralization is a traditional and widely used method in treating AMD (Kumar et al. 2008;Romero et al. 2011;Zheng et al. 2011;Alakangas et al. 2013;Akinwekomi et al. 2017Akinwekomi et al. , 2020Masindi et al. 2019). In this process, calcium hydroxide, limestone, sodium hydroxide or alkaline tailings are commonly used as neutralization reagents. The metal and sulfate ions are removed from the AMD as metal hydroxide and gypsum residues, respectively (Tolonen et al. 2014). The water is then recycled or discharged, and the solid residue discharged to tailings. Sulfide precipitation is usually used to recover metal ions such as Cu 2þ , Zn 2þ , etc. from AMD and Na 2 S, NaHS, H 2 S, and so on, are the reagents normally used (Chen et al. 2014;Luo et al. 2017). Ion exchange (Luo et al. 2017) and adsorption (Mohan & Chander 2006;Groudev et al. 2008;Freitas et al. 2011;Zhang 2011) have also been studied widely in treating low concentration AMD. Sulfate-reducing bacteria (SRB) (Foucher et al. 2001;Sun et al. 2020) use sulfate as a terminal electron acceptor, SO 4 2À being converted into H 2 S, and then H 2 S-selective recovery of some metals, such as copper and zinc, as pure sulfides. SRBs are seldom used because of their low efficiency and high cost.
No matter which AMD treatment process is used, a neutralization process must be used to raise the water's pH above 7.0, using alkaline reagents, before discharge. The price of limestone in China is just one-third that of lime so, to reduce AMD treatment costs, two-stage neutralization by limestone and lime instead of single-stage lime neutralization is valuable. However, neutralization of AMD by these two methods is seldom compared. In this work, residue composition and yield, removal rate, reagent consumption and cost by these neutralization methods are compared.

Materials
AMD came from one of the Zijin Mining Group mines in Chinaits average content is shown in Table 1.
The effective calcium oxide and calcium carbonate contents of lime are 75.67% and 11.90%, respectively, and the calcium carbonate content is 96.23% in limestone -Table 2. The average particle sizes were 150 and 74 μm. The XRD traces for lime and limestone are shown in Figure 1. Calcium oxide is the main component of lime - Figure 1(a)with small amounts of calcium hydroxide, calcium carbonate and quartz. Calcium carbonate is the main component of limestone (Figure 1(b)).  The tests were carried out in a 400 m 3 /h process chain. Figure 2 comprises the flowcharts for (a) the single-stage neutralization process using lime and (b) two-stage neutralization using lime and limestone.
Single-stage (lime) neutralization: 10% (w/w) lime was added in the first and second reaction tanks, and the pH was controlled at 4.5 and 7.0 in them, respectively. The slurry from the second neutralization tank was thickened and the underflow pumped to the plate-frame pressure filter. Following pressure filtration, the residue was discharged to the tailings pile and the overflow discharged or reused after the pH reached 7.0.
Two-stage neutralization: AMD was pumped continuously into the first neutralization tank and neutralized using 10% (w/w) limestone. The slurry from here flowed automatically to the second neutralization tank, where it was reacted with 10% lime. The slurry from the second neutralization tank flowed to the thickener and the underflow was pumped to the plate-frame pressure filter. The pressure-filter residue was discharged to the tailings pile, and the overflow discharged or reused after the pH reached 7.0.

Analytical methods
Liquid and slurry samples were collected on-line continuously. Wet residue was dried at 50°C for 8 hours before chemical analysis. Liquid and dry residue elements were analyzed by atomic absorption spectrometry (AAS, iCE3400, Thermo Fisher, USA). The phase composition of the neutralization   residue was analyzed by X-ray diffraction (XRD) (X'Pert Powder-DY3624, PAN Analytical, Netherlands), scanning electron microscopy (Quanta 650, FEI, USA) and energy spectrum analyzer (EDAX APOLLO Xjime USA).

RESULTS AND DISCUSSION
Theoretical precipitation pH Two-stage neutralization theoretical precipitation pH: the sulfuric acid in the AMD was neutralized with limestone, and a hydrolytic precipitation process of the Cu 2þ , Fe 3þ , Al 3þ and Zn 2þ present began as the pH increased gradually. In the second neutralization tank, the slurry pH was further increased to 7.0 after continuous lime neutralization. The limestone/H 2 SO 4 neutralization reaction is shown in Formula (1), and the solubility product constants arising for each ion during hydrolysis by formulae (2) to (5) (Ma & Chen 2005). Single-stage neutralization theoretical precipitation pH: Ca(OH) 2 slurry was formed when calcium oxide was digested in water -Formula (6). The sulfuric acid concentration in the AMD decreased and the slurry pH increased gradually as the sulfuric acid was neutralized -Formula (7). Hydrolytic precipitation of the Cu 2þ , Fe 3þ , Al 3þ and Zn 2þ began at the same time (formulae (2) to (5)). (1) Table 3 represents the initial and complete theoretical precipitation pH results for Fe 3þ , Al 3þ , Cu 2þ , Zn 2þ on the basis of the solubility product constants determined from formulae (2) to (5), the initial AMD ion concentrations (Table 1) and assuming the completed precipitation is 1.0 Â 10 À5 mol/L.

Theoretical reagent consumption by equivalent acid
The concentration of the various ions in the AMD varied during the trialssee Figure 3 and it was difficult to calculate limestone and/or lime consumption accurately by AMD volume. To compare reagent consumption, the limestone or lime consumption by Fe 3þ , Zn 2þ , Al 3þ , Cu 2þ during neutralization hydrolysis was converted into the equivalent consumption of sulfuric acid, using reaction formulae (1) to (7). The results are shown in Table 4. Effect of pH on removal rate Figure 4 shows the removal rates of H 2 SO 4 , Fe 3þ , Al 3þ , Zn 2þ and Cu 2þ increasing as the initial pH of two-stage neutralization increased, while their concentrations decrease. Figure 5 shows that the theoretical and actual removal rates of Fe 3þ were very similar, but the actual Zn 2þ , Cu 2þ and Al 3þ removal rates were much higher than predicted by theory. The difference arises because Fe 3þ was removed first and precipitated out as Fe(OH) 3 , as the initial precipitation pH of Fe 3þ is much lower than those of Al 3þ , Cu 2þ and Zn 2þ . Because of the large specific surface, strong adsorption and negative charge of Fe(OH) 3 in the acidic system, the Zn 2þ , Cu 2þ and Al 3þ were co-absorbed and co-precipitated into the residue by the Fe(OH) 3 generated before reaching    their theoretical initial precipitation pHs. As a result, the actual removal rate exceeds the theoretical rate, and the initial precipitation pHs of Zn 2þ , Cu 2þ and Al 3þ were lower than the theoretical values.

Effect of pH on reagent consumption
Normally, limestone consumption increases as the pH increases; however, reagent consumption is higher when the pH is 3.2 than when it is 3.4 or 3.6 ( Figure 6(a)) because of the various ionic   concentrations (see Figure 3). Converting reagent consumption by AMD volume to the acid consumption equivalent shows that limestone consumption increased while lime consumption decreased with increasing pH. The optimal limestone and lime consumptions were determined as 0.79 and 0.33 kg/kg acid equivalent respectively at pH 3.4.

Neutralization residue yield and reaction mechanism
The neutralization residue yield is 2.57 t/t acid equivalent in the two-stage process using limestone and lime, which is 15.74% lower than the 3.05 t/t equivalent from single-stage neutralization using lime (Figure 7). The free water content of the two-stage neutralization residue varied from 42.25% to 42.63% with an average of 42.51%, some 5.84% higher than the 36.67% average free water content of the single-stage neutralization residue (range 35.85% to 37.01%). The residue yield in the limestone and lime, two-stage neutralization is lower than that in the single-stage (lime) neutralization reported by Li (2007), or Ding & Ding (2004), but the mechanism causing the difference in residue yield was not further studied.  (8) and (9). The residue crystals are fine, which is consistent with the crystal properties of calcium sulfate di-hydrate (Xiang 1998). Figure 8(b) and 8(d) show that that the main neutralization residue component from two-stage neutralizationlimestone and limeis CaSO 4 •0.5H 2 O. The residue's crystal morphology is flaky, irregular, and relatively loose, consisting mostly of secondary particles composed of single, fine grains. This is consistent with the crystal structure of β-CaSO 4 •0.5H 2 Othat is, calcium sulfate hemihydrate (Li 2007). The molecular weight of CaSO 4 •0.5H 2 O is 15.70% less than that of CaSO 4 •2H 2 O, consistent with the test results shown in Figure 7(a), which shows that the residue yield from two-stage neutralization is 15.74% lower than that from single-stage neutralization.
It is clear from Figure 5(a) that more than 97% of the sulfuric acid was removed by calcium carbonate in the first reaction during two-stage neutralization. The reactionsee Formula (10)produced a large amount of CO 2 gas as bubbles, which then cover the CaSO 4 crystal surface and prevent contact between it and H 2 O. In two-stage neutralization, the acidity of the first reaction system exceeds that in single-stage neutralization while the temperature is lower in the first stage. The CaSO 4 •0.5H 2 O reaction mechanism is shown in Formula (11).
Effect of neutralization methods on cost Figure 9 shows that the reagent cost of two-stage neutralization falls as the pH increases from 3.0 to 3.4. At pH 3.4, the reagent cost for two-stage neutralization is 0.227 yuan (RMB)/kg equivalent acid (0.035 USD), its lowest level. The reagent cost of single-stage neutralization at pH 3.4 is 16.60% more than that of two-stage neutralization at the same pH. However, two-stage neutralization reagent costs increase as the pH is increased from 3.4 to 4.0. At pH 4.0, two-stage neutralization reagent cost exceeds that of single-stage neutralization, at the same pH, by 4.07%, so the initial reaction pH is the key to achieving cost savings using two-stage neutralization rather than the singlestage process.

CONCLUSIONS
(1) During AMD neutralization using limestone and lime, the theoretical and actual removal rates of Fe 3þ were basically similar, but the actual removal rates of Zn 2þ , Cu 2þ and Al 3þ were much higher than theoretical predictions. This arose because Fe 3þ was removed first and precipitated as Fe(OH) 3 , and, due to its large specific surface, strong adsorption and negative charge in the acidic system, the Zn 2þ , Cu 2þ and Al 3þ were co-absorbed and co-precipitated before reaching their theoretical initial precipitation pHs.
(2) The XRD results show that the main components in two-stage and single-stage neutralization residues are CaSO 4 •0.5H 2 O and CaSO 4 •2H 2 O, respectively. The molecular weight of CaSO 4 •0.5H 2 O is 15.70% lower than that of CaSO 4 •2H 2 O. The different crystalline water content in the gypsum produced by the different neutralization methods caused the residue yield from two-stage neutralization to be 15.74% less than that from single-stage neutralization. (3) The limestone and lime consumptions were 0.79 and 0.33 kg/kg acid equivalent respectively.
When the initial reaction pH is 3.4, the reagent cost is 16.60% less in two-stage than singlestage neutralization, but the saving rate begins to decrease when the initial reaction pH exceeds 3.4. The cost of two-stage neutralization is 4.07% higher than single-stage neutralization when the initial reaction pH is 4.0.