3.1. HDS Elemental Content and Solubility
Manganese was the most abundant of all metals considered by the RSA guidelines (on both a mass and molar basis) (
Table 4). This was expected based on the natural abundance of elements in geology and soil in this province. After Fe, Mn is the most abundant transition metal, and AMD dissolves and mobilises it as it percolates through geological formations and soil. The Limestone/Lime Site HDS was particularly rich in total Mn, but with low Mn solubility compared to Limestone–Sites 1 and 2 HDS. The Mn content varied across the HDS products considered and seemed to be an AMD treatment signature. An HDS process without lime treatment generates HDS with appreciable soluble Mn (TCLP extractable Mn). Examples of this are HDS
CaCO3 and HDS
CaCO31 sludges with soluble Mn highlighted in grey. The mechanism for immobilisation is rapid oxidation of Mn
2+ to Mn
4+ and subsequent precipitation of sparingly soluble Mn(IV) oxides under high pH (9–9.5) conditions [
40]. The kinetics of oxidation is appreciably slower at the pH or alkalinity levels created by limestone, and results in incomplete oxidation of Fe
2+ and Mn
2+. Due to the specific interaction of Mn
2+ and Fe
2+ with carbonates, there is also the risk that these metals will temporarily precipitate with carbonates, most likely as surface precipitates on limestone particles, which will further decrease their propensity to be oxidized. X-ray diffraction (XRD) analysis (data not shown) for HDS
CaCO3, detected ankerite, Ca(Mg,Fe,Mn)(CO
3)
2, confirming the presence of ferrous iron and Mn(II), as well as incomplete oxidation of the Fe and Mn in the HDS products.
3.2. Assessment of HDS Using the RSA Guidelines
The assessment with RSA was achieved by comparing HDS data (on the right) to the thresholds (on the left) in
Table 5. For the sludge HDS
CaCO3, from Limestone–Site 1, the only element not classified as Type 3 or 4, was Mn (
Table 6). The TCLP extractable Mn of 259 mg L
−1 (highlighted in grey) in
Table 5 exceeded LCT3 (200 mg L
−1) resulting in a Type 0 (
Table 6) being allocated to this sludge, requiring treatment and reassessment according to RSA regulations before disposal in a lined facility. As discussed earlier, the high Mn solubility was attributed to the fact that the Limestone–Site 1 process only uses limestone. The higher total Mn for HDS
CaCO3 sludge can be attributed to various factors, for example, (1) the Mn concentration of the AMD may have been higher as a result of a general increasing trend or seasonal changes in mine water quality; (2) the limestone source used in the HDS process at the time of sampling. The Ni content of HDS
CaCO3 was more in-line with the HDS
Ca(OH)21 and can either be related to temporal changes in AMD or quality of limestone used [
13]. A more detailed discussion of this is beyond the scope of this study. A consistent and characteristic signature of the HDS
CaCO3 was Ba, as HDS
Ca(OH)21 exhibited low levels of this element. This could also have been a limestone signature.
The only total analyses given for Limestone–Site 2 were for Cr, Mn, Zn, and the TCLP analyses reported were also limited for both HDS
CaCO31 and HDS
CaCO32 sludges. The Limestone–Site 2 HDS classification was therefore largely incomplete. However, some general trends were evident. Similar to HDS
CaCO3, soluble Mn (
Table 5) was the element condemning the HDS
CaCO31 in a lined pond to Type 0 (
Table 6). However, the main difference between the two sludges was the higher soluble Ni in HDS
CaCO31 compared to that of HDS
CaCO3. Maree et al. [
9] reported TCLP extractable Mn (211 mg L
−1) in HDS
CaCO31 sludge that exceeded LCT3 (200 mg L
−1) and Ni exceeded its LCT2 threshold (7 mg L
−1).
The sludge, HDS
Ca(OH)21 from Limestone/Lime process had higher Co, Mn, Pb, Ni, and Zn content than HDS
CaCO3 sludge and all these metals exceeded the TCT0 thresholds. For this sludge, HDS
Ca(OH)21, it seemed that Pb persists and is not effectively removed by the neutralization process. For both HDS
Ca(OH)22 and HDS
Ca(OH)23 sludges, Pb content exceeded the TCT0 threshold of 20 mg kg
−1 (
Table 5). The RSA guidelines have set low minimum TCLP values (LCT0), especially for As, Cd, Pb, Hg, and Se. Due to detection limit difficulties, this has resulted in inconclusive results and technically also an incomplete classification (
Table 6) of the Limestone–Site 1, Limestone/Lime Site HDS for some of more important elements from an environmental point of view. A TCLP extract (or any other extract) from soil or solid waste at fairly low solution to solid ratios (20:1) creates a substantially more saline and a complex matrix. As a result, method detection limits (MDLs) are always higher (often an order of a magnitude) for extracts than for drinking water. The MDL is the lowest concentration of an element in a specific extractant/matrix where its signal is statistically separable from background “noise”. As, Cd, Pb, Hg, and Se are especially prone to matrix and spectral interferences resulting in, for example, false positive interferences. In order to measure LCT0 concentrations repeatedly with high confidence, the TCLP MDLs for these elements should be below LCT0 concentrations.
Lead in general shows higher MDLs in TCLP and other extracts. Kavouras et al. [
41] reported 0.3 mg L
−1 (Pb determined by atomic absorption spectroscopy using a graphite furnace), the Laboratory analysis of the Limestone/Lime Site, reported 0.1 mg L
−1, while Lin and Chang [
42] reported 0.016 mg L
−1 for Pb in TCLP. Apart from the latter article, all the other detection limits in TCLP were an order of a magnitude higher than the LCT0 for Pb. It is believed that a more extensive investigation on TCLP MDLs for commercial laboratories for these elements will confirm the trend that LCT0 levels for some or all these elements are below typical TCLP MDLs and are therefore of no meaning and should be revised.
3.3. Assessment of HDS Using Australian (New South Wales) Guidelines
The Australian guidelines define total F, Ag, Be levels where the RSA guidelines do not. As a result, analysis of these elements is not required for hazardous waste characterisation in RSA and a judgement on these elements with the Australian system could not be made. On the 1st screening stage (
Table 7) the Australian regulations indicate that Ni (108 mg kg
−1 highlighted in grey) for HDS
CaCO3, from Limestone–Site 1, exceeded SCC1 (40 mg kg
−1) allocating a Restricted Solid Waste status to the material, meaning it can pollute the environment. The material was then assessed against both SCC and TCLP, the 2nd Screening stage (
Table 8). The elements considered for this sludge were found to be below SCC1 thresholds, but when compared against TCLP thresholds, only Ni (2.9 mg L
−1 highlighted in grey) exceeded TCLP1 (2.0 mg L
−1) confirming the classification of the material as a Restricted Solid Waste. This categorization is equivalent to Types 1 to 2 in the RSA system. Only Ni was highlighted as an element of concern because the Australian guidelines do not consider Mn.
The sludge, HDS
CaCO31, from Limestone–Site 2 was allocated a hazardous status (equivalent to Type 0 of the RSA system) since its Ni (16.5 mg L
−1 highlighted in grey) exceeded TCLP2 (8.0 mg L
−1) and HDS
CaCO32 from the same site was categorized as a General Solid Waste, since all its TCs and LCs were below SCC1 and TCLP1 thresholds (
Table 7 and
Table 8). The other sludges, HDS
Ca(OH)21, from Limestone/Lime Site had TCs for Pb (143 mg kg
−1 highlighted in grey) and Ni (104.9 mg kg
−1 highlighted in grey) exceeding their thresholds (only in the 1st screening stage—
Table 7), HDS
Ca(OH)22 also had its TC for Pb (163 mg kg
−1 highlighted in grey) exceeding its threshold (only in the 1
st screening stage—
Table 7), categorising both materials as Restricted Solid Wastes, but HDS
Ca(OH)23 from the same site was categorized as a General Solid Waste (allowing exploration for use by either construction industry or agriculture).
3.4. Assessment of HDS Based on Canadian, US, and Chinese Guidelines
Guidelines for these countries rely only on TCLP data (
Table 9). To achieve the assessment, the TCLP extracted data presented on the right of
Table 9 were compared to the thresholds presented on the left portion of the same Table. When comparing HDS
CaCO3 sludge to the USEPA, Canadian, and Chinese guidelines, all of the LC values considered were below threshold levels, categorizing this material as non-hazardous waste. This suggested that no restrictions are needed for the disposal of this sludge and the potential for its use in agriculture or construction industry can be explored.
HDSCaCO31 from Limestone–Site 2 had an LC for Ni (16.5 mg L−1 highlighted in grey) exceeding Canada’s (Alberta) and China’s TCLP threshold (of 5 mg L−1) thereby classifying the material as hazardous waste. The other three systems classified this sludge as non-hazardous. The source of the acid waters treated in the Limestone–Site 2 has not been confirmed as being solely of coal mining origin, and it should be noted that a major metalliferous processing organisation is situated upstream of the Limestone–Site 2 treatment plant. Analyses for HDSCaCO32 were below TCLP thresholds for all countries, assigning the material a non-hazardous status. When evaluating sludges from Limestone/Lime Site (HDSCa(OH)21, HDSCa(OH)22 HDSCa(OH)23) against the USEPA, Canadian, and Chinese guidelines, none of the LCs exceeded TCLP thresholds of any of the guidelines.
3.6. Should Mn Form Part of Hazardous Waste Classification?
The HDSCaCO31 product, from Limestone–Site 2 was flagged by the majority of the guidelines (Canada, China, and Australia) based on soluble Ni (USEPA does not consider Ni). HDSCaCO3, from limestone–Site 1, and HDSCaCO31 were both flagged on the basis of Mn, but only by the RSA guidelines. The RSA guidelines are very thorough in the number of elements they consider, and sensibly omit Fe and Al. However, it is the only system that considers Mn. Like Fe, Mn forms sparingly soluble oxides and this is most likely the reason why most countries do not consider it to be an element of major concern. The critical aspect at HDS plants is whether lime or limestone has been used in the neutralizing process. Lime accelerates oxidation kinetics because of the higher pH. If limestone is used, more time is needed at the lower pH for complete oxidation and formation of Mn (IV) oxides. Once formed, Mn (IV) oxides are exceedingly insoluble as demonstrated by the HDSCa(OH)21 HDS which had almost 7000 mg kg−1 of Mn, yet the solubility was below 0.04 mg L−1.
Mn is also a common soil constituent especially in a South African context. This is another environment where the low solubility of Mn from Mn(IV) oxides is demonstrated. The best example is the manganiferous soils derived from the Malmani dolomites in RSA which have been used for irrigation for 150 years or longer and have been critical in providing food for the large urban and peri-urban Gauteng population. These soils span important agricultural areas in Gauteng, parts of the Northwest and Mpumalanga Provinces of South Africa contain up to 13,000 mg kg
−1 of Mn (more than double the Mn total content of the HDS
Ca(OH)21) [
43]. This means that these soils would be Type 1 (high risk) wastes if they were to be classified using the RSA system based on TC.
Apart from their low solubility, Mn (IV) oxides also have various other benefits. Their metal scavenging abilities are well-known and have a particularly high affinity for B-type cations (soft metals), especially Pb [
44]. They also have the ability to oxidize organic pollutants in the soil and are more likely playing a critical role in protecting environmental quality rather than harming it. Furthermore, the oxidizing propensity it lends to environments is well-known in soil research and commonly observed in dolomite derived soils [
40]. Mn(IV) oxides will not only help buffer ferric oxide reduction and dissolution but also actively oxidize (or re-oxidize) ferrous iron and Mn
2+. If As occurs in the waste, the presence of Mn will result in its oxidation to the less soluble As(V) arsenate [
45].
Based on the arguments made on the low solubility of Mn(IV) oxides and of the potential environmental benefits, it seems prudent to omit total Mn content from the South African system, as has already been done with Fe.