Chemical composition of coal fly ashes A and B used in AMD treatment experiments.
A treatment process for Acid mine drainage (AMD) using coal fly ash (CFA) was developed. AMD was treated with CFA as the alkaline agent at different CFA: AMD ratios and pH, electrical conductivity (EC) evolution monitored over time. In a separate experiment two AMD sources with differing chemistry were treated with the same CFA to evaluate the impact of AMD chemistry on the treatment process and product water quality. Various CFA: AMD ratios were stirred in a beaker for a pre-set time and the process water chemistry determined. pH was observed to increase in a stepwise manner with buffer zones observed at 4-4.5, 4.5-7 and 6-8. AMD with low concentration of Al3+, Fe2+, Fe3+ and Mn2+ didn’t exhibit these buffer zones. Two competing processes were observed to control the evolving pH of process water: dissolution of basic oxides (CaO, MgO) from CFA led to pH increase and hydrolysis of AMD species such as Al3+, Fe2+, Fe3+ and Mn2+ led to pH decrease. These processes initiated mechanisms such as precipitation, adsorption and ion exchange that led to decrease in inorganic contaminants as the treatment progressed. Inorganic contaminants removal was directly related to amount of CFA in reaction media. Precipitation of insoluble hydroxides and Al, Fe-oxyhydroxysulphates contributed to removal of major and minor chemical species. Precipitation of gypsum contributed to removal of sulphate. Na, K and Mg remained largely in solution after initial decrease. Significant leaching of B, Sr, Ba, and Mo from CFA was observed and was directly linked to amount of CFA in the reaction media. This will be a shortcoming of the treatment process since other processes may be required to polish up the product water. The treatment of AMD with CFA was observed to depend on CFA, AMD chemistry, treatment time and might therefore be site specific.
- acid mine drainage
- coal fly ash
- inorganic contaminants
- coal fly ash: acid mine drainage ratios
The mining industry in South Africa has a huge potential to impact negatively on the environment. Negative impacts include generation of reactive tailings and acid mine drainage (AMD). In South Africa, the Witwatersrand basin alone decants approximately 10–60 mL/day into nearby rivers . AMD is highly acidic (pH 2–4), sulfate-rich and frequently carries a high concentration of inorganic contaminants. AMD is extremely acidic (as low as pH 2.0) and enriched with iron, manganese, aluminum, sulfate, and metal species such as lead, mercury, cadmium, zinc [2, 3, 4, 5, 6]. When sulfide minerals such as pyrite (FeS2), its dimorph marcasite and pyrrhotite (Fe1−x S) are exposed to oxygen and water, they undergo a bacterially catalyzed oxidation reaction which lead to generation of acidity and increased Fe and sulfate concentrations in recipient water bodies (Eq. (1)). The overall reaction is often written as :
South Africa environmental regulations require these AMD streams to be treated to acceptable levels before discharging into surface water bodies. A range of active and passive remediation technologies have been adopted by various mining companies to reduce impact of AMD on ground and surface water resources. These technologies include active neutralization by lime, limestone, biological sulfate removal [8, 9, 10], eutectic freeze crystallization  and the alkaline barium calcium desalination process (ABC) . Passive AMD remediation technologies include alkalinity generating artificially constructed wetlands . Other innovative AMD management and remediation technologies have been developed which involve treatment and recovery of beneficial products such as gypsum and sulfur . Recently, innovative and sustainable management technologies for AMD have been developed which involve treatment and recovery of drinking water that is supplied to communities near the mines while the gypsum recovered is used in construction of gypsum boards and houses. This is to reduce the cost of treatment and make the process sustainable . Bhattacharyya and Gupta  observe that these AMD management technologies are still expensive and mining companies are still searching for cheaper treatment and management technologies.
South Africa uses more than 100 MT of low grade bituminous coal annually to produce electricity in coal powered utilities. These coal power utilities in turn produce 28 MT of coal fly ash annually that is disposed of on land as ash dumps or slurried to ash dams due to lack of alternative application of this coal combustion by-product . Fly ash is a powdery substance which consists of fine spherical particles which are either solid or hollow. It is a ferro-aluminosilicate material with particles ranging in size from 20 to 80 μm, its surface is enriched with trace elements of Si, Al, Fe, Ca, K and Na [17, 18]. Several authors have reported that South African coal fly ash consists mainly of aluminosilicate glass matrix in addition to crystalline mullite and quartz. This glass phase was observed to be associated with elements such as As, Na, Mg, K, Sr, B and Mo as soluble salts on the surface of the spheres [3, 19, 20]. The South African coal is sub-bituminous and generates coal fly ash that has low Fe content and the aqueous extracts of this coal fly ash are strongly alkaline (pH 12–12.5) due to the free lime content [16, 21, 22, 23, 24]. The highly soluble CaO occurring as sub-micron fragments on the fly ash particles accounts for the alkaline properties of the coal fly ash .
Traditionally, AMD neutralization and remediation has been carried out using liming agents such as limestone, lime and sodium hydroxide [9, 10, 13, 25, 26]. An important disadvantage of these liming agents is cost and generation of large volumes of sludge that requires disposal. Due to those limitations, mining companies are always on the lookout for cheaper and effective liming agents.
Several research reports have highlighted the application of fly ash for the control of acid generation from sulphidic rich wastes [27, 28] and for amendment of acidic soils . The studies involved blending the mine spoils, tailings and acidic soils with varying amounts of fly ash. These authors observed that the coal fly ash controlled the acid generation from the wastes by release of alkalinity over time to the system which neutralized the acidity produced. Several authors have reported on the successful application of coal fly ash for inorganic contaminants removal from AMD and acidic leachates [8, 19, 30, 31]. They observed that contaminants removal was directly proportional to CaO content of the fly ash. The authors concluded that coal fly ash due to its free lime content can be used as a neutralization and inorganic contaminants removal for AMD remediation.
Most of the power utilities generating coal fly ash in South Africa are located near coal mines supplying them with coal . These coal mines are also the sources of AMD that requires remediation. This chapter reports on work done to explore the possibility of utilizing coal fly ash for treatment and remediation of AMD. The fact that most of the coal-powered utilities are near the coal mines that supply them with coal and produce AMD, makes the proposed treatment process economically viable. A fundamental understanding of the solution chemistry and product water quality after neutralization of AMD with coal fly ash is a prerequisite for this treatment process.
2. Collection of samples, physicochemical analysis and experimental methods
2.1. Collection and preparation of samples
Coal fly ash (CFA) samples were collected from two coal-powered utilities in Mpumalanga, South Africa and stored in tightly lockable plastic containers. The AMD samples used were collected from two collieries and a government AMD treatment plant that remediates AMD seepage from an old abandoned mine in Witbank, South Africa. The AMD samples from the government AMD treatment plant were scooped from the seepage point while samples for the two collieries consisted of AMD pumped from underground old mine workings to a collection dam and underground mine voids, respectively.
Raw AMD samples were filtered by using 0.45 μm pore cellulose nitrate membrane filters and diluted with MilliQ (ultrapure) water to EC < 1.5 mS/cm and 3 drops of dilute HNO3 added and kept refrigerated at 4°C until analysis for cationic species. Samples for anion analysis were diluted with MilliQ water and refrigerated at 4°C until analysis.
2.2. Coal fly ash/acid mine drainage treatment experiments
The AMD/coal fly ash treatment experiments were designed to develop neutralization patterns that would indicate buffering properties of the AMD and also show the contact time required for the breakthrough to circum-neutral-alkaline pH in the product. Department of Water Affairs and Forestry  South Africa requires pH for treated water for release into surface water bodies to be in the range of 6–9. The batch treatment experiments were conducted by stirring a mixture of coal fly ash and AMD which was pre-determined to give a specific coal fly ash/AMD ratio (FA:AMD). An overhead stirring system was used to agitate the mixture for all experiments. The progress of the reaction was monitored by measuring the pH and EC with a Hanna HI 991301 portable pH/EC/TDS/Temperature metre.
A second set of treatment experiments were repeated at selected FA:AMD ratios to explore the product water chemistry and inorganic contaminants removal efficiency of the coal fly ash with increasing pH of the process water. The solid residues were separated by filtration and the process water samples were prepared for cationic and anionic species analysis. Cationic species were analyzed using ICP-MS (ELAN 6000) and the accuracy of the analysis monitored by the use of NIST water standards. Fe2+/Fe3+ analysis was done by the colorimetric method using 2, 2-bipyridal as the complexing reagent. Sulfate analysis was done turbidimetrically by a portable data logging spectrophotometer (Hach DR/2010) and ion chromatography (Dionex DX-120).
Chemical characteristics of the coal fly ash samples were ascertained by X-ray fluorescence spectroscopy (XRF) by fusing with lithium metaborate. The solid residues collected after reacting coal fly ash with AMD were analysed by Phillips PANalytical X-ray diffractometer (XRD) using Cu Kα radiation generated at 20 mA and 40 KV. Specimens were step scanned as random powder mounts from 5 to 85° 2θ integrated at 0.02° 2θ per second. Powder samples of the solid residues were also observed under a scanning electron microscope (SEM-EDX) equipped with an energy dispersive X-ray analysis system (Hitachi X-650 microanalyzer).
3. Results and discussion
3.1. Chemical composition of coal fly ash
The chemical composition of the coal fly ashes used in the AMD treatment experiments are presented in Table 1 .
|Coal fly ash A||Coal fly ash B|
|Species||% (w/w)||Species||ppm||Species||% (w/w)||Species||ppm|
|SiO2||53.4 ± 2.4||Cu||47.3 ± 6.6||SiO2||53.8 ± 0.29||Cu||57.9 ± 9.9|
|TiO2||1.34 ± 0.05||Mo||5.22 ± 0.14||TiO2||1.44 ± 0.11||Mo||6.56 ± 0.15|
|Al2O3||23.4 ± 1.1||Ni||93.4 ± 6.5||Al2O3||26.2 ± 2.52||Ni||58.2 ± 1.2|
|Fe2O3||4.72 ± 0.96||Pb||56.4 ± 13.6||Fe2O3||3.40 ± 0.24||Pb||29.1 ± 7.18|
|MnO||0.06 ± 0.002||Sr||1463.9 ± 111.8||MnO||0.05 ± 0.02||Sr||2056.0 ± 205|
|MgO||2.69 ± 0.05||Zn||57.3 ± 4.71||MgO||2.48 ± 0.58||Zn||25.4 ± 1.35|
|CaO||8.43 ± 0.57||Zr||488.1 ± 125.7||CaO||8.50 ± 1.75||Zr||536.1 ± 131.3|
|Na2O||0.35 ± 0.25||Co||18.3 ± 13.08||Na2O||0.49 ± 0.05||Co||10.4 ± 3.3|
|K2O||0.49 ± 0.03||Cr||179.2 ± 1.14||K2O||0.86 ± 0.07||Cr||122.7 ± 27.8|
|P2O5||0.35 ± 0.22||V||147.4 ± 38.9||P2O5||0.60 ± 0.22||V||145.8 ± 32.8|
|Cr2O3||0.03 ± 0.009||Ba||928.0 ± 91.9||Cr2O3||0.03 ± 0.006||Ba||1559.2 ± 346.7|
|NiO||0.011 ± 0.001||NiO||0.009 ± 0.002|
|V2O5||0.019 ± 0.002||V2O5||0.02 ± 0.003|
|ZrO2||0.052 ± 0.012||ZrO2||0.055± 0.003|
|LOI||2.36 ± 0.19||LOI||1.33 ± 0.36|
The three major oxides identified in the two coal fly ashes were Al2O3, Fe2O3 and SiO2. Coal fly ash B had higher Al2O3 content while coal fly ash A showed higher Fe2O3 content. These coal fly ashes can be classified as class F according to the American Society for Testing and Materials  since (SiO2 + Al2O3 + Fe2O3 ≥ 70%). The CaO content shows slight variation with coal fly ash B showing a slightly higher value. The total CaO content detected by XRF includes total CaO including that locked up in the aluminosilicate matrix. XRF does not distinguish the free lime from that trapped within the glass matrix. The free CaO content of CFA is important because of its availability for rapid dissolution, which has implications on its neutralization potential during AMD treatment. Substantial MgO was observed in both coal fly ashes and could contribute significantly in the neutralization of AMD. Both coal fly ashes showed high concentrations of Sr, Ba, Cr, Zr and Ni ( Table 1 ). Trace elements of Mo are also present. These species are likely to be released into solution during the AMD treatment process.
3.2. Physicochemical properties of acid mine drainage samples
The physicochemical properties of the AMD samples used in the treatment experiments are presented in Table 2 .
|pH||2.39 ± 0.05||2.46 ± 0.03||2.91 ± 0.02|
|EC (mS/cm)||10.83 ± 0.13||10.78 ± 0.15||10.02 ± 0.06|
|Acidity (mg/l CaCO3)||6950 ± 70.7||7000 ± 70.7||500 ± 0.0|
|TDS (mg/L)||16765 ± 50.5||19410 ± 76.8||8975 ± 60.5|
|B||1.37 ± 0.163||1.51 ± 0.08||2.29 ± 0.221|
|Na||358.7 ± 2.95||399.9 ± 21.3||4137.9 ± 233.0|
|Mg||2661.7 ± 35.0||2844.2 ± 148.1||388.7 ± 19.7|
|Al||1068.1 ± 11.28||1140.1 ± 61.58||60.0 ± 2.9|
|Si||82.01 ± 1.24||87.8 ± 5.87||69.7 ± 3.5|
|K||23.03 ± 2.86||19.3 ± 4.21||52.6 ± 3.6|
|Ca||653.3 ± 10.6||1012.3 ± 75.9||842.1 ± 117.4|
|Mn||226.3 ± 4.7||242.3 ± 12.9||31.6 ± 1.50|
|Fe||5599.9 ± 80.9||6115.9 ± 327.5||250.8 ± 11.2|
|Fe2+||3725.1 ± 30.5||2886.3 ± 20.7||153.1 ± 9.5|
|Fe3+||1451.9 ± 45.2||3344.6 ± 50.5||126.1 ± 6.5|
|Ni||6.95 ± 0.02||7.96 ± 1.16||2.35 ± 0.13|
|Cu||0.355 ± 0.007||0.345 ± 0.018||0.116 ± 0.012|
|Co||4.3 ± 0.11||4.57 ± 0.32||1.15 ± 0.063|
|Zn||48.99 ± 30.63||17.7 ± 0.65||9.52 ± 0.49|
|Sr||7.69 ± 0.226||8.39 ± 0.45||1.05 ± 0.058|
|Mo||0.04 ± 0.002||0.044 ± 0.003||0.036 ± 0.004|
|Ba||0.209 ± 0.002||0.189 ± 0.01||0.148 ± 0.018|
|SO4 2−||11888.1 ± 20.6||14949.7 ± 28.3||6155 ± 54.3|
|Cl−||729.3 ± 15.3||265.9 ± 10.6||720 ± 11.5|
|NO3 −||163.2 ± 21.6||41.6 ± 5.7||BDL|
The AMD samples are highly acidic with the pH ranging from 2.39 for Navigation samples to 2.92 for Brugspruit samples ( Table 2 ). All the samples exhibit high electrical conductivity (10.02–11.36 mS/cm), typical of most AMD impacted mine waters. This is attributed to high sulfate content which contributes to the high conductivity values of the samples. The sulfate correlated positively with the EC measurements for most of the acidic coal mine waters investigated in this study. The sulfate recorded in these samples ranged from 6155 to 14950 mg/L making this anion dominant in the AMD samples while major cationic species included Na, Ca, Mg, Al, Mn and Fe. It is worth noting that Navigation and Bank AMDs had significantly higher total Fe iron content than Brugspruit AMD.
3.3. Treatment of Fe-rich acid mine drainage samples using different coal fly ashes: the role of coal fly ash chemistry
The pH and electrical conductivity (EC) trends of the treatment reactions at a FA to AMD ratio of 1:3 are shown in Figure 1 .
The EC for treatment processes followed the same trend with continuous decrease as the agitation progressed. A zone was observed for the three treatment processes where the EC remained constant ( Figure 1 ). In terms of contact time, it coincided with the pH buffer zones.
Both coal fly ashes A and B showed potential capacity to neutralize AMD at CFA:AMD ratios of 1:3. The pH-neutralization trends for all reactions showed a strong pH buffering zone at pH 6.2–6.8 which was observed to last for different reactions times depending on the coal fly ash or AMD type used. This pH buffer zone is associated with the oxidation and hydrolysis of Fe2+ which releases H+ ions and delays the rise in pH  (Eq. (3)). Al3+ undergoes hydrolysis at pH ≈ 4.5 and will also contribute to pH buffer in this zone (Eq. (2)).
This pH buffering zone is a result of high concentrations of Fe2+ and Al3+ in the AMD (≈3000–4000 mg/L for Fe2+ and ≈1150 mg/L for Al3+) ( Table 2 ). It should be noted in Figure 1 that it took longer time for CFA-A to overstep the acidic buffer zone with Navigation AMD as compared to Bank AMD. This is attributed to the high concentrations of Fe2+ in Navigation AMD than Bank AMD ( Table 3 ) that could sustain the buffering capacity. It is observed that both solutions broke through to a pH of ≈9, with Navigation AMD taking 200 min and only 150 min for Bank AMD to break through. This indicates that the chemistry of the AMD is crucial in this treatment process. A small buffer zone was also observed at pH 4–4.5 for treatment reaction between CFA-B and Bank AMD. This buffer zone is attributed to the hydrolysis and precipitation of Fe3+ with subsequent consumption of H+ (Eq. (3)). It should be observed that Bank AMD had twice the concentration of Fe3+ as Navigation AMD. This aspect is also confirmed in the pH trend for the CFA-A and Bank AMD with a slight buffering at pH ≈5.5 ( Figure 1 ). Coal fly ash B appeared to have less free alkalinity than CFA-A. This is evident from the pH trends during the treatment process between both CFAs and Bank AMD. To break through to pH ≥ 8.7 ( Figure 1 ), contact time required was greater with CFA-B than CFA-A for the treatment of Bank AMD. The neutralization potential of CFA depends on free CaO available and in this case, CFA-A seems to have more free CaO than CFA-B. It should be noted that XRF gives the total CaO content of the coal fly ash and therefore results in Table 1 which do not reflect free CaO. This aspect confirms that the chemistry of the coal fly ash will also be an important factor in this treatment process.
3.4. Treatment of iron-rich and iron-poor acid mine drainage with coal fly ash: the role of CFA/AMD ratios and acid mine drainage chemistry
To evaluate the effect of AMD chemistry and CFA/AMD ratios on the treatment process and product water quality, CFA-B was reacted with Brugspruit and Navigation AMD. Brugspruit AMD had low total Fe, Al and Mn content while Navigation AMD had high concentration of total Fe, Al and Mn in addition to high SO4 2− content ( Table 2 ). Figure 2 shows the trends of pH and electrical conductivity (EC) for the treatment process between CFA-B and Brugspruit AMD at various CFA:AMD ratios.
The treatment of Brugspruit AMD with low ratios of CFA-B led to alkaline pH (pH ≈ 10) within 5 min of reaction time ( Figure 2 ). This indicates that even low amounts of FA could be used to achieve neutralization in some cases depending on the chemistry of AMD being treated. Treatment of Brugspruit AMD with CFA-B:AMD ratios of between 1:3.5 and 1:8 led to highly alkaline pH (pH > 12) in product water. The ratio of CFA used was directly proportional to the decrease of EC. The EC seemed to decrease within the initial 5 min of treatment and then stabilized at a minimum of 8.5–9 mS/cm. It should be observed all the CFA-B:AMD treatment ratios used led to a final pH ≥ 12 after 360 min of reaction. It should be noted that the buffer zones observed with CFA-A:Navigation AMD and CFA-B:Bank AMD treatment are not observed with Brugspruit AMD in all the CFA:AMD ratios employed ( Figures 1 and 2 ). This is a strong indication that AMD chemistry has a significant role to play in the final product water chemistry.
Figure 3 shows the trends of pH and electrical conductivity (EC) for the treatment process between CFA-B and Navigation AMD at various CFA:AMD ratios.
The pH and EC trends for the treatment process of Navigation AMD with CFA-B are distinctively different from those of the treatment process for Brugspruit AMD ( Figures 2 and 3 ). For the Navigation AMD treatment process, there is a stepwise increase in pH as the treatment progresses. Buffer zones are also observed at pH 3.5–4 and pH 5.5–6.5.
Despite the high CFA:AMD ratios employed in the treatment as compared to Brugspruit treatment, the breakthrough to pH > 10 only occurs after 210 min ( Figure 3 ) and for lower CFA:AMD ratio of 1:3, a breakthrough was not achieved after 360 min. It should be noted that the stepwise increase in pH with time is lacking for the Brugspruit AMD. Stepwise and gradual decrease in EC is noted for the Navigation AMD treatment process ( Figure 3 ). The initial decrease in EC for the Brugspruit treatment is not sustained and stabilizes at 8.67–9.56 mS/cm. Uhlmann et al.  observed that the buffering observed at pH 3.5–4 is due to hydrolysis of Fe3+ while oxidation and hydrolysis of Fe2+ contributes greatly to buffering at pH 5.5–6.5  (Eqs. (2) and (3)). At pH ≥ 4.5, Al3+ undergoes hydrolysis forming insoluble hydroxides and will buffer pH in the region of 3.5–4 (Eq. (2)) [34, 35]. It should be noted that Brugspruit AMD had very low concentration of Fe3+/Fe2+ and Al3+ as compared to Navigation AMD ( Table 2 ). This could explain the lack of stepwise decrease in pH and buffering during the treatment process. There are two opposing processes that finally control the final pH of the product water in this treatment process. The dissolution and hydrolysis of soluble alkaline oxides such as CaO and MgO ( Table 1 ) from coal fly ash will contribute to increase in pH (Eqs. (4) and (5)). This is also confirmed by the fact that the time taken for breakthrough of the buffer zone 5.5–6.5 reduced with increase of CFA in the treatment mixture. CFA:AMD ratio of 1:1.5 took the shortest time to breakthrough this buffer zone ( Figure 3 ).
Hydrolysis of AMD constituents such as Fe3+, Al3+ and Fe2+ releases protons (acidity increase) (Eqs. (2) and (3)) and offsets the pH increase attributed to the dissolution of the alkaline oxide from coal fly ash. The net result is a pH buffer zone, this buffer zone is only overstepped when the components leading to release of acidity are completely hydrolyzed. These hydrolysable components will finally determine if the final pH of the product water will be acidic or alkaline at any given treatment time. It will also be observed that the final pH of the product water also depended on the CFA:AMD ratio applied. Higher CFA:AMD led to higher pH of product water regardless of treatment time. It can therefore be concluded that factors that will control the product water chemistry and quality would be, CFA:AMD ratio, the treatment time and the chemistry of the AMD.
3.5. Inorganic contaminants removal: product water quality compared to South African water quality guidelines
Acid mine waters are highly reactive solutions that can dissolve most primary minerals when reacted with an alkaline solid waste material such as coal fly ash with subsequent formation of a variety of secondary minerals that can potentially adsorb trace chemical species. Dissolution of coal fly ash by AMD triggers several processes such as hydrolysis of soluble alkaline oxides leading to increase in pH. Several authors observe that pH is the most important parameter in coal fly ash solutions and accounts for significant inorganic toxic elements removal through precipitation, ion exchange and adsorption processes [13, 30].
Tables 3 and 4 present results of inorganic contaminants removal as a function of final solution pH and CFA:AMD ratios for Brugspruit and Navigation AMD treatment with CFA-B. The results are discussed with respect to pH of precipitation of the various metal species as determined from thermodynamic calculations and experimental observations from titration of solutions containing the stated species as reported by Britton .
3.6. Brugspruit AMD treatment with CFA-B: inorganic contaminants removal
Brugspruit AMD had different chemistry compared to that of Navigation AMD, it had low concentration of the major hydrolysable chemical species Al3+, Fe3+, Fe2+ and Mn2+. Major inorganic contaminants Fe, Mn, Al and Mg were reduced significantly in all the ratios investigated ( Table 3 ). Fe and Al were reduced by >95% at pH ≥ 9.16 attained in the final process waters for CFA:AMD ratios, 1:3, 1:5, 1:8, 1:2 and 1:3. It should be noted that final pH also depended on the treatment time employed. Mn removal was ≥90% for all CFA:AMD ratios evaluated except for 1:3 ratio at 80%. Jenke et al.  observed that at the pH of minimum solubility of the hydroxides of Fe3+ (pH 3.0), of Fe2+ (pH 6.0–8.0), of Mn2+ (pH 8.41–9.0) and of Zn2+ (pH 6.0–6.5), a significant proportion of the initial concentration should be precipitated out of solution. Gitari et al.  observed that at pH ≥ 4.5, most of the mineral phases bearing these species were at saturation or oversaturation (SI ≥ 1). At pH ≥ 12, a slight increase in Al, Fe and Mn was observed. At pH 12.0–12.5 attained for ratios 1:8, 1:2 and 1:3, the formation of hydroxy complexes was attributed to the decreased removal. Drever  observes that Al3+ hydroxide exhibits minimum solubility at pH 6–6.5, therefore at pH > 6.5, the Al-hydroxy complexes become more important .
Among the minor and trace species, Ni, Cu, Pb, Mg, Cr, Ni,Co, Cu, As and Cd were all observed to decrease as the pH increased. The Cu, Cr, Zn and Pb removal efficiency was observed to decrease at CFA:AMD ratio of 1:8 which attained a pH of 12.04. The decrease in removal of these species could be attributed to soluble hydroxo species formed at pH > 12.0. Another reason for the increase could be reduced adsorption due to reduced formation of amorphous Fe(OH)3, MnOOH and Al(OH)3 due to the low concentration of Fe3+, Mn and Al3+ in Brugspruit AMD. Mg2+ removal efficiency approached 100% at pH 9.16 with a decrease observed as the pH increased to 12.62. Decreased removal could be due to the formation of soluble hydroxo species at the high pH. Na and K were removed by 50% as the pH increased. Gitari et al.  observed that as pH increases and in presence of SO4 2-, several Na-K bearing mineral phases such as jarosite are precipitated and would account for this decrease. Sulfate did not register significant removal and in some CFA:AMD ratios (1:2) was observed to increase, this increase is attributed to release from the coal fly ash.
3.7. Navigation AMD treatment with CFA-B: inorganic contaminants removal
The pH increase with increase of CFA:AMD ratios was observed to be gradual for Navigation AMD treatment as compared to that of Brugspruit AMD ( Tables 3 and 4 ). Several inorganic contaminants removal trends are observed as the pH increases. At final pH 6.33 achieved for the CFA:AMD ratio of 1:3, a sharp decrease in concentration is observed for Fe, Al and Mn. These are the major inorganic contaminants that characterize Navigation AMD. With subsequent CFA:AMD ratios, the removal efficiency of ≥90% are observed. Gitari et al.  observed that as pH increases, mineral phases bearing these chemical species are precipitated and account for their increased removal in solution. Minor and trace inorganic species such as Si, Cr, Ni, Co, Cu, Zn, As, Se, Cd and Pb were all observed to decrease as the pH increased with subsequent CFA:AMD ratios. Their removal could be attributed to several mechanisms such as precipitation of mineral phases and adsorption to high surface area Fe and Al oxyhydroxy precipitates. Gitari et al.  observed that co-precipitation of Fe, Al-oxyhydroxides, oxyhydroxy sulfates and adsorption could be responsible for attenuation of Cu2+, Zn2+ species.
Britton  observes that the pH of minimum solubility of the hydroxides/oxyhydroxides of Fe3+, Fe2+, Al3+, Zn2+, Cu2+ and Ni2+ are 3.0, 6.0–8.0, 4–4.5, 6.0–6.5 and 6.66, respectively. The pH attained for CFA:AMD ratio of 1:3–1:2 is within this range. This would explain the high removal of these species within this pH range. However, Cu and Pb registered an increase in concentration in final process waters for CFA:AMD 1:2 and 1:2.5 ratios whose final pH ranged from 8.7 to 9.5 ( Table 4 ). This would be attributed to the formation of soluble hydroxo complexes for both species at this pH range.
Mg, Na, K and Ca removal increased with increasing pH of the final process water. These chemical species are highly soluble in the pH range generated in process waters and likely removed through incorporation in precipitating mineral phases. Gitari et al.  observed that mineral phases such as alunite, basaluminite, jurbanite, brucite and jarosite are likely to precipitate out in CFA:AMD reactions leading to removal of these species. Sulfate was likely removed through formation of gypsum, incorporation into Al, Fe-oxyhydroxysulfates. A direct relationship between percent sulfate removal and FA content in the reaction mixture suggests that dissolution of CaO and subsequent formation of gypsum accounts for the sulfate removal in both AMD treatments. XRD analysis identified gypsum in the solid residues of the CFA:AMD reactions [3, 39].
3.8. Leaching of chemical species from coal fly ash in the treatment process
Coal fly ash is a reactive material and on contact with aqueous solutions such as AMD undergoes dissolution with subsequent release of chemical species. An observation of results in Tables 3 and 4 show that interaction of coal fly ash and AMD leads to the release of B, Sr, Mo and Ba. The fact that these chemical species increased with increase in the ratio of CFA in the treatment mixture attests to the fact they were being released from the coal fly ash. Mattigod et al.  and Eary et al.  reported that the glassy phase in coal fly ash is enriched with As, Na, Mg, K, Sr, B, Mo and Ba as soluble salts which are released into solution when the CFA contacts aqueous solutions.
3.9. Comparison to South Africa water quality guidelines
The product water for various CFA:AMD ratios was compared to the Department of Water Affairs and Forestry  water quality limits for irrigation and domestic use. Treatments of Navigation AMD with CFA-B resulted in much cleaner water with Cu, K, Mo, Na and Zn being within the domestic water use limits for 1:3 CFA:AMD ratio and Co, Cu, K, Na and Zn being within the domestic water limits for 1:2.5 CFA:AMD ratio. Treatment of Brugspruit AMD with CFA-B produced less clean water although breakthrough to alkaline pH was established within less than an hour. Only Cu and K were within the domestic limits for 1:3 FA:AMD ratio. This comparison includes only the FA:AMD ratios that resulted in process water in the pH range 6–9. The alkali and alkaline earth metal species (Na, B, Mg, Ca, Sr and Ba) remained largely in the process water in concentrations beyond the guideline limits. B, Mg, Sr, Mo and Ba were largely released from dissolution of coal fly ash and additional treatment options such as reverse osmosis will be required to further clean the water to required standards.
The dissolution and hydrolysis of basic oxides such as MgO and CaO from coal fly ash on contacting AMD led to an increase in pH. On the other hand, hydrolysis of AMD chemical species such as Al3+, Fe3+, Fe2+ and Mn2+ led to release of acidity that counteracted the pH increase leading to stepwise increase in pH of process water and buffer zones as the treatment processes progressed. This was clearly observed for Navigation AMD as opposed to Brugspruit AMD that had low hydrolysable species. These two processes initiated several processes such as precipitation and adsorption that led to attenuation of the major and minor chemical species leading to much cleaner product water. Efficiency of the inorganic contaminants removal was directly linked to amount of FA in the reaction mixture and to the final pH in product water. Three processes were identified that would finally control the quality of the product water, these were (1) CFA:AMD ratios, (2) treatment time and (3) chemistry of the AMD. Most of the major contaminants were reduced to below South African water quality guidelines; however, the coal fly ash was observed to release other contaminants in solution and this could be a drawback for the proposed treatment process. Other factors that would improve the sustainability of the treatment process are beneficial application of the treatment residues as backfill material. Economic viability of this process would also depend on closeness of the AMD generating mines and the coal power utilities generating the coal fly ash.
The authors would like to acknowledge and thank Water Research Commission (WRC), National Research Foundation (NRF) and Coaltech 2020 for funding this research work.