**3. Results**

In this study, although results and discussion mainly focus on trace elements (As, Cd, Co, Cr, Cu, Ni, Mo, Pb, and Zn), major elements (Al, Ca, Fe, K, Mg, P, S, and Si) are sometimes mentioned because of their relevance for interpretation of release mechanisms of elements of interest.

#### *3.1. Single Extractions*

The results of ammonium-EDTA and CH3COOH 0.43 mol·L−<sup>1</sup> extractions for the studied BA samples are displayed in Table 3. Extractability is expressed in percent of an element extracted with ammonium-EDTA and CH3COOH extraction relative to its total concentration in the (solid) sample. Ammonium-EDTA extraction was used to determine the potential mobility of trace elements as a consequence of complexation and used as an estimation of the "pool" of a specific element that can deliver elements from the solid phase to the solution [13]. Moreover, it can also give a rough indication on the bioavailability of some trace elements and it is sometimes used to assess the availability of trace elements to plants [29,30].

Among the examined trace elements, Cd (in AS1) and Pb (in AS2) showed the highest EDTA extractability (27 and 31% of the total content in the samples, respectively) while EDTA extractability of some other trace elements such as Co (3%), Ni (2–3%), Cr (0.2–0.6%) and Mo (10%) is rather similar despite of their difference in total concentrations in both samples (Table 3).


**Table 3.** Extractability (as% of total concentration) of elements from ammonium-EDTA and acetic acid extractions.

Although Cd reached the highest extractability in AS1, its concentration in the EDTA extract was below limit of quantification (LOQ) in AS2. Arsenic also had a low EDTA-extractable fraction (2%) in sample AS2, but in sample AS1, 14% of the total As concentration was extracted. Similar to As and Cd, Zn also displayed a higher extractability in sample AS1 (20%) compared to sample AS2 (6%). Results of As and Cd were in accordance with Ca and S, since both Ca and S reached a high extractability in AS1 (87–98%), and a slightly lower extractability in AS2 (36–45%). A study of ash from coal combustion, Nugteren (2008) [31] reported that As, Cd, and Mo are belonging to the group of elements which are associated with calcium oxides and sulfates. However, FEG-EPMA analyses of the bottom ash samples indicated that Mo (in AS1) was related to Fe-bearing phases, while As and Cd could not be observed during solid-phase characterization with FEG-EPMA [19]. The extractability of Cu (AS1 and AS2) and Pb (AS2) was quite high compared to other trace elements. The high EDTA extractability of Cu and Pb (respectively 22% (in AS1) and 31% (in AS2) of their total content in the samples) might be explained by the high complexation constants for these two elements with EDTA (log K = 17.8 and 18.3, respectively) [12].

The concentrations of elements extracted with CH3COOH varied between the two BA. Besides Cd (in AS1) and Pb (in AS2), Zn and Cu showed a high extractability (29–65% of their total concentration). In contrast, Cr, As and Mo display a low extractability ( ≤6%). It should be mentioned that Pb and Cu are characterized by a higher stability for mononuclear monoligand and biligand complex systems with CH3COOH compared to other metals [32,33]. However, the high extractability of Cd, Pb, Zn, and Cu can be a combination of both high stability of the acetate complexes and the drop of pH during extractions. In both samples, Co and Ni were released in similar amounts (15–17% for AS1 and 25–27% for AS2, respectively) possibly because they originate from the same host phases, namely Fe-alloys or Fe-oxides [19].

For both single extractions, Ca and S (in AS1) showed the highest extractability among the examined elements. Ca and S in AS1 were totally extracted during the CH3COOH 0.43 mol·L−<sup>1</sup> extraction. In general, major cations present in the solid samples may be one of the factors affecting trace element extraction efficiency due to their competition to form complex compounds with EDTA [34]. The dissolution of calcite can consume EDTA in calcareous soils, lowering the extraction efficiency for trace elements [35]. In the present study, important amount of Ca (87%) was extracted with ammonium-EDTA (Table 3), possibly affecting the extraction efficiency of the reagent.

#### *3.2. Acid Neutralization Capacity (ANC) and Trace Element Release at pH 4*

The ANCpH4, 96 h (i.e., the amount of acid added to maintain a pH of 4 until 96 h after the start of the pHstat titration) of sample AS2 (936 mmol·kg−1), was nearly double to the ANCpH4, 96h of sample AS1 (510 mmol·kg−1). The higher amount of calcite as determined by XRD in sample AS2 (3.2%) most likely explains the difference in ANC between both samples. Although the pH change during the extraction with CH3COOH 0.43 mol·L−<sup>1</sup> might provide an indication of the ANC of the two BA samples, the difference in the final pH of the CH3COOH extracts was not that high (final pH of 3.26 and 3.44 for AS1 and AS2 respectively). The initial pH of the CH3COOH solution was 3.02. The reason for this small difference in pH, despite the important difference in ANC, might be the short duration of the CH3COOH extraction test (16 h) in which the slow buffering reactions are not fully considered [36]. Hence, pHstat leaching tests, performed for a longer period (96 h in the present study) likely allow a better estimation of ANC from BA samples. It should also be mentioned that weathering of BA (natural or artificial) is responsible for increasing the buffering capacity of the BA [15]. Several studies have shown that leaching for several metals appears to be less important than from fresh BA after weathering [37,38]. However, a study about the carbonation (artificial weathering) of BA from municipal solid waste incinerator, Van Gerven et al. (2005) [39] reported an increase in the leaching of Cr and a constant leaching of Mo and Sb after carbonation of BA. Therefore, ANC of a BA is very important and should be better estimation to investigate the release of trace metal upon external addition of H+.

The evolution of ANC and the release of Ca with time in both samples during pHstat leaching are displayed in Figure 1.

**Figure 1.** Evolution of ANC and Ca during pHstat leaching test (pH 4).

We performed XRD phase analysis on residual BAs after pHstat leaching (at pH = 4) to assess leaching related to changes in major solid phases. XRD phase analysis on the residual BAs after the pHstat leaching (at pH = 4) showed that some peaks of calcite (CaCO3) decreased in intensity (Figure 2). This suggests that main mineral phases in the BAs were stable at pH 4, except small change was observed for calcite. Possibly, the dissolution of other mineral phases was too small to be detected by the XRD technique.

**Figure 2.** XRD patterns of original sample and sample after the pHstat test (pH = 4, sample AS2).

In the following section, leachability refers to the concentration in the final leachate (after 96 h) expressed in percent of an element leached relative to its total concentration in the (solid) sample (Table 4) except for Mo in sample AS2 since its concentration in the leachates decreased to values below the LOQ from 3 h onward. For the latter, the concentration in the leachate after 1 h was used.

In both BA samples, despite the high total concentration, Al Fe, and P exhibit very low leachability (<0.5%) compared to other major elements, such as Ca, K, Mg, Mn, and S (>2%). This suggests that no significant dissolution of Al-Fe-P containing minerals occurred during the pHstat leaching test. Release of some selected major and trace elements during pHstat leaching are displayed in Figures 3 and 4. Cadmium concentrations in the leachates from both samples were below LOQ likely due to the low total Cd-concentrations (≤1.5 mg·kg−1).

**Table 4.** Leachability (as% of total concentration) of elements in the pHstat test (after 96 h or, except for Mo in sample AS2, after 1 h).


**Table 4.** *Cont.*

**Figure 3.** Release of major elements and some selected trace elements from AS1 during pHstat leaching (at pH 4).

**Figure 4.** Release of major elements and some selected trace elements from AS2 as a function of time during pHstat leaching (at pH 4).

In the leachate of sample AS1, As and Pb were below LOQs while released Cr concentration was very low (0.04% or 0.3 mg·kg−<sup>1</sup> was released at pH 4). The leachability of Co, Cu, Mo, and Ni varied between 1–3%. The highest leachability was observed for Zn (9%).

In sample AS2, most of trace elements were only detected in the leachates after 1 h except Cr, Mo and Zn which were already released immediately from the start of the experiment. Arsenic concentrations in the leachate varied just around LOQ while Mo was released at the beginning of the pHstat experiment, but the concentration decreased below LOQ after 3 h of leaching. Nickel exhibits the highest leachability (8%), while other trace elements such as Co, Cu, Pb and Zn show a moderate leachability (1–6%). Chromium exhibited a low leachability (0.02%) despite its high total concentration (804 mg·kg−1). This indicates that most of the trace elements in the BAs do not occur in readily soluble forms, even if the external pH is lowered to a value of 4.
