**4. Discussion**

#### *4.1. Potential Release of Trace Elements Based on Different Extractions/Leaching Test*

The results after 48 h of pHstat leaching in this study corresponds to the standard pHstat test CEN/TS 14997 test [14] at pH 4. Therefore, comparing the release of trace elements by the CH3COOH extraction, ammonium-EDTA extraction and after 48 h of pHstat leaching provides information on the potential release of trace elements in BA samples based on standardized extraction/leaching tests. The release of some selected trace elements from CH3COOH 0.43 mol·L−<sup>1</sup> and ammonium-EDTA 0.05 mol·L−<sup>1</sup> extraction compared to the amount of these trace elements extracted with pHstat leaching test (determined with 48 h) is displayed in Figure 5.

**Figure 5.** Release of Cu-Co-Mo (in AS1) and Pb (in AS2) after acetic acid 0.43 mol·L−<sup>1</sup> and ammonium-EDTA 0.05 mol·L−<sup>1</sup> extraction compared to the amount of these trace elements released by the pHstat leaching test (after 48 h).

Most of trace elements showed the highest extractability with CH3COOH, while As and Mo were most effectively extracted with ammonium-EDTA. The high amount of trace elements (except As and Mo) that were extracted by CH3COOH can be explained by the low pH (pH = 3.26–3.44) of the CH3COOH extract. Acidification has a pronounced effect on the release of trace elements from the studied BAs. Removal or recovery of trace elements from ash by acid leaching has been studied to find out the most effective leaching agen<sup>t</sup> [24,40,41]. However, this recycling option suffers from some drawbacks such as the use of large amounts of acid due to the high ANC of ashes and the generation of wastewater. The more important extraction of As and Mo by ammonium-EDTA compared to CH3COOH extraction and pHstat leaching at pH 4 is related to the fact that they may occur as oxyanions which are known to be leached more under alkaline conditions than in acidic conditions. The pH of the CH3COOH extract is 3.26 (AS1) and 3.44 (AS2), whereas the pH of the solution of the pHstat leaching test was continuously kept at 4. Contrarily, the pH of the ammonium-EDTA extract is neutral to slightly alkaline with values of 7.59 (AS1) and 7.98 (AS2). Some trace elements such as Cd, Cu, Ni, Pb, and Zn can form complexes with dissolved organic carbon (DOC) which may enhance the leaching of these elements. Increased Cu leaching from MSWI bottom ash by DOC complexation was observed by [42,43]. Cu in AS1 exhibited a slightly higher mobility at alkaline pH values (9–11) compared to neutral pH, which can be related to the fact that Cu forms organometal complexes with dissolved organic matter in the leachates of sample AS1 at high pH [19]. Unfortunately, data of DOC in the leachates are not available; hence, the possible effects of complexation with DOC could not be evaluated in the present study.

The relatively high-potential mobility of Cu, Pb and Mo deduced from the ammonium-EDTA extraction, in combination with elevated total concentrations of Cu and Mo (in AS1) and Pb (in AS2) indicate a higher possibility of recovery of these metals from the BA. Up to 250 mg Cu/kg and 27 mg Mo/kg (in AS1) and 251 mg Pb/kg (in AS2) were released by the ammonium-EDTA extraction.

Removal of contaminants from ashes by washing with chelating agents (e.g., citrate buffer, EDTA or oxalate buffers) has been suggested since chelating agents can remove significant amounts of contaminants present in mobile forms in the outer layer of the ash particles [22]. Pre-washing with water, increasing the concentration of EDTA and increasing the extraction time even increase the extractability [41]. A high extractability of Cu (100%), Pb (94%) and Zn (40%) was observed from the fly ash with pre-water washing followed by a 24 h extraction with EDTA 0.1 mol·L−<sup>1</sup> (end pH 8.2 and L/S 5.0 <sup>L</sup>·kg−1) [41]. However, it is worth noting here that results for fly ash are not directly transferable to the BA samples in this work due to the differences in particle size, chemical composition, and mineralogy. Extraction efficiency obtained in the present study can probably be improved, by increasing the concentration of the EDTA solution, adapting the extraction time, L/S ratio, temperature, etc. It was not the purpose of the present study to investigate the optimal condition for maximal element recovery from the BA samples. However, the result show that the extractions used in the present study can be used for a relatively cheap and fast screening of the potential of element recovery from waste materials.

#### *4.2. Kinetics of Trace Element Release during Leaching Test at pH 4*

Understanding the kinetics of trace elements release is important for predicting the environmental risks associated with these elements over time. Kinetic leaching experiments performed on solid materials have shown that element leaching can be described by two steps, representing a fast release process followed by a slow process [44]. In this study, time-dependent leaching behavior of trace elements could be divided into three types (Figure 6), considering fast and slow release processes.

Type 1 includes elements which are released slowly, and steady state is not reached (e.g., Co and Ni in both samples, as well as Cu in AS2 and Zn in AS1). This type of element release is related to the desorption of elements that are strongly bound to solid phases or to the slow dissolution of solid phases. Similar leaching patterns of Co and Ni from both samples support the hypothesis that they might come from the same host phase or that they are retained by similar binding mechanisms. Compositional analysis by EPMA-EDS indicated that Fe, Co, and Ni co-existed in the analyzed spots in sample AS1 [19]. The slow release of Ni and Co is most likely due to the fact that Ni and Co are occluded in stable phases such as Fe-alloys and Fe-oxides and thus showed a slow mobilization under acidic environmental conditions. Similar time-dependent leaching patterns of Fe, Co, and Ni (Figure 3) from sample AS1 support this hypothesis. Although the Fe content of sample AS2 was much higher (24%) compared to that of sample AS1 (4%), it should be mentioned here that the total concentration of Co and Ni in AS2 is not that high (63 mg Co/kg and 233 mg Ni/kg), and Ni and Co did not show a release pattern similar to Fe in sample AS2. In sample AS1, Zn occurs in Si-rich phases which might be related to quartz, mullite or amorphous phases [19]. The release pattern of Zn during the pHstat leaching test in this study (Figure 3) is quite similar to the release of Zn from glass phases as observed by [45]. Moreover, quartz and mullite are known as stable phases under neutral and mildly acid conditions while Zn was observed to leach up to 9% in pHstat leaching test (Table 4). Therefore, Zn might be associated with amorphous or poorly crystalline phases, which are more easily dissolved than stable phases.

**Figure 6.** Release of Fe (AS1), P (AS1), S (AS2) and some selected trace elements illustrating three different types of time-dependent leaching behavior during pHstat leaching (at pH 4). Co and Ni concentrations were multiplied by a factor of 2 for a better visualization.

Release of elements according to "type 2" is characterized by an initial release at the beginning of the pHstat leaching test, followed by a decrease in dissolved concentrations over time. This type of release is related to precipitation reactions or re-adsorption onto solid phases (e.g., Cr and P in both samples). This is the case for example when elements forming oxyanions (e.g., chromate) are re-adsorbed on positively charged reactive surfaces at acidic pH. This phenomenon could not be assessed in the single extraction test (e.g., CH3COOH extraction) due to the short duration of the extraction and the fact that only the final leachate was analyzed.

The last type of leaching pattern ("type 3") is related to elements which are released rapidly, and steady state seems to be reached after more or less 6 h (e.g., Cu–Mo in AS1 and Zn in AS2). Mo was also observed in Fe-rich phases in sample AS1 by FEG-EPMA; however, its leachability was higher compared to other elements associated with Fe-rich phases. The release kinetics of Mo (AS1) during pHstat leaching is also faster than Co and Ni (Figure 3) suggesting that this element might not be incorporated into crystalline phases but is distributed on the surface of Fe-rich phases. Although Zn-bearing phases were not identified by XRD and FEG-EPMA in sample AS2 [19], the release behavior of Zn during pHstat leaching was similar to S (Figure 6), suggesting that Zn may also exist in S-rich phases or that Zn and S are bound to a solid phase in a similar way.

It can be deduced from the pHstat test that release mechanism of elements from BA are probably related to the surface processes on the solid material such as desorption and re-adsorption, or to solubility of mineral phases. When pH is lower, desorption happens since the presence of hydrogen ions displaces metals that bound on the solid surfaces [46]. Precipitation or re-adsorption onto solid phases is observed for some elements which form oxyanions such as Cr and P. It was concluded by [47] that solid solution formation is frequently believed to be a controlling mechanism for oxyanion leaching, for example for Cr (VI) in MSWI bottom ash.
