**4. Discussion**

### *4.1. Chemical and Mineralogical Differences and its Effects on Acid-Neutralizing Capacity*

The chemical and mineralogical characterization showed considerable differences between the ash types. As the difference in matrix elements, WWFA showed significantly higher Ca and Si contents and significantly lower concentrations in Cl than MSWIFA. Within WWFA, chemical differences were also observed, depending on the waste wood content. The S content, for example, was as high as in MSWIFA for WWFA100, but less than half the concentration in WWFA40. The fluctuations in the elemental composition of MSWIFA are attributed to the compositional differences in waste input. Remarkable is the constant chemical and mineralogical composition between the different WWFA40 samples, as waste wood is also a very heterogeneous feedstock [22]. This could be an effect of the low waste wood content in WWFA40. The significantly higher Cl concentration in MSWIFA is associated with the combustion of plastics (PVC). It is known that elevated Cl concentrations in the flue gas favor the evaporation and transport of heavy metals (e.g., Cd, Cu, Pb, Zn, Sb [23–25]), and elevated heavy metal concentrations in MSWIFA have been observed by many authors (e.g., [15,26]). The high concentrations of these hazardous metals in MSWIFA are associated with the combustion of, e.g., batteries, paints, alloys, plastics [27]. The high Sb concentrations in MSWIFA are problematic due to their high toxicity (especially of the trivalent species [28]). Antimony is not soluble at the low pH conditions prevailing in the FLUWA process and accumulates in the filter cake. To fully assess the hazard potential of Sb and to evaluate Sb mobilization during the FLUWA process, additional studies are in progress. The high Pb and Cr(VI) concentrations in WWFA are problematic and justify the need for treatment prior to deposition. WWFA100, the WWFA with higher waste wood content, showed significantly higher heavy metal concentrations, which reflects findings made in previous studies on WWFA (e.g., [22]). The elevated Pb and Zn concentration in WWFA is probably due to pigments (e.g., PbCO3, ZnO) from paints and coatings, whereas the high Cu content in WWFA could arise from the combustion of pickled wood products [29].

During combustion, alkali- and alkaline earth metals in the wood transform to oxides and are subject to successive hydrogenation and carbonation during cooling [30], which

could explain the high calcite content in WWFA40. The high Si content stands in relation to the usage of quartz sand as bed material during fluidized bed combustion [10]. In the presence of SO2 and O2, CaO often forms sulfate compounds (e.g., anhydrite) [31], as present in WWFA100. WWFA100 further showed higher TOC content (associated with incomplete combustion [32]) and different matrix composition (e.g., less calcite, more sulfates) than WWFA40, the latter affecting ANC. The high ANC of WWFA40 can be explained by the very high calcite content and dissolution of Ca-silicates (e.g., gehlenite, belite). The high initial pH of all the three ash types suggests the presence of non- or microcrystalline CaO or Ca(OH)2 since no CaO or lime was identified in all ashes. WWFA100 and MSWIFA both showed poorly acid buffering sulfates and chlorides as main constituents, which explains the lower amount of H+ needed to reach the acid conditions required for the FLUWA process.

Thus, although WWFA showed different chemical and mineralogical characteristics than MSWIFA, there are also different geochemical properties within WWFA, depending on their waste wood content and certainly also depending on differences in the waste wood composition.

### *4.2. Water-Extractable and Total Cr(VI) in WWFA and Filter Cakes*

The measured average water-extractable Cr(VI) concentration of 93 mg/kg in WWFA40 justifies the strong need for Cr(VI) reduction prior to landfilling. The treatment of WWFA40 with the FLUWA process successfully reduced water-soluble Cr(VI). All analyzed filter cakes showed Cr(VI) concentrations below the threshold limit for landfilling, even when the experiments were performed under oxidizing conditions with H2O2. It was shown that acidic conditions are sufficient to dominate the reduction of water-extractable Cr(VI), as also observed in other studies [33]. It was also reported that H2O2 could act as a reductant in acidic solutions [34]. It is assumed that the water-soluble Cr(VI) is being reduced to Cr(III) during the FLUWA process, followed by precipitation as Cr(III) phase, probably as hydroxide. This is supported by the absence of dissolved Cr in the filtrate. Unfortunately, no Cr phases could be identified with XRD as the concentrations are very low and the precipitated phases possibly amorphous. It was further shown that the high Cr(VI) concentrations in the eluates were not caused by the oxidation of Cr(III) during the eluate test. However, Cr(VI) reduction occurred during the eluate test for WWFA100 due to its highly reductive character—leading to erroneously low water-extractable Cr(VI) concentrations. Similar matrix interferences occurred during the hot alkaline extraction for WWFA40. These observations show that spiking during Cr(VI) extraction tests with reactive material such as WWFA is indispensable. Poor Cr(VI) spike recovery is not indicative of method failure, but rather an indication for the potential of the sample to reduce the spiked Cr(VI) and not sustain its native Cr(VI) [35]. They further report that the presence of high TOC contents, as well as considerable S2<sup>−</sup> or Fe2+ concentrations in the sample, are most likely the reason for low Cr(VI) spike recovery or reduction of native Cr(VI). Although oxidative conditions prevail during combustion, it cannot be excluded that locally reducing conditions occur, where Fe2+ and S2<sup>−</sup> persist. While no mineral phases containing Fe2+ and S2<sup>−</sup> were observed, their presence in minor concentrations cannot be excluded. For further interpretation, a more detailed investigation on possible reductants other than Corg in WWFA must be performed, with special focus on the content of Fe2+ and S2−.

### *4.3. Leaching Experiments: Heavy Metal Recovery and Consumption of Neutralizing Chemicals*

The laboratory-scale experiments were able to predict well the recovery trends of the industrial-scale. Differences in the recoveries between laboratory- and industrial-scale are primarily attributed to differences in pH and to element contents of the ash, as well as to a larger L/S in the industrial scale, which will increase the recovery. For WWFA40, a higher H2O2 dosage was needed to achieve oxidizing conditions during extraction. This could be caused due to the high content in organic matter or the presence of metals in their metallic form—leading to rapid consumption or even catalytic destruction of the added H2O2. The high ANC of WWFA40 led to higher acid consumption in the FLUWA compared to MSWIFA (3× higher). Thus, WWFA40 can represent a heavy metal-rich replacement for the often-used pH neutralizing agent lime milk. The high acid consumption is even more pronounced when H2O2 is used, since the oxidation of, e.g., metallic compounds consumes H+.

For the element Zn, recovery was equally high for the different ash types independent of the H2O2 dosage, as Zn mobility is independent of the redox conditions during extraction but controlled by pH and binding form [36]. The achieved Zn recovery was lower by 15% in the laboratory-scale experiments compared to the industrial-scale experiment. It is assumed that this is due to the higher L/S used in the industrial-scale experiment (L/S of 15 compared to L/S of 3). The low Zn recovery of WWFA40 (laboratory-scale) is associated with enhanced precipitation of Zn due to the high leachate pH. Remarkable is the fact that the Zn recovery seems not strongly affected by the Zn concentration in the ash. As observed in other studies [36], the Zn yield stagnates at about 70% (in this study at about 65% in the industrial-scale-experiments). It is assumed that the majority of the Zn in the ashes is readily available for dissolution (e.g., as Cl- or S-salts). The remaining 30–35% of the Zn seems, however, to be present in the insoluble form under these conditions (e.g., as glassy particles, as Ca replacement in gehlenite or associated with iron [36]). The recovery of Cd showed the same trends for laboratory- and industrial-scale, but with lower recoveries by 15–50% in the laboratory-scale. The higher recoveries at the industrial-scale are again attributed to the higher L/S. The lower Cd yields in the laboratory-scale experiment without H2O2 are associated with the higher leachate pH. Since Cd concentration is about one order of magnitude lower than the other elements, it is subject to larger fluctuations as inhomogeneities in Cd concentration in the sample are more pronounced. The low Cd recovery from WWFA40 is attributed to the very low Cd concentration in the ash. An increase in Cd recovery can be observed for each ash type when the experiments are performed under oxidizing conditions, as observed in other studies [36]. As WWFA40 showed a higher redox buffer than MSWIFA, it is assumed that the lower Cd recoveries in the experiments without H2O2 are a result of reductive precipitation of Cd since Cd recoveries are comparable to those of MSWIFA when using H2O2. The mobility of Pb and Cu is highly dependent on the redox conditions, as well as on the pH (especially for Cu). The recoveries for Pb and Cu for 40 L H2O2 are higher on the laboratory-scale than on the industrial-scale. Besides differences in elemental concentrations in the ashes, the pH and redox conditions are more easily controllable on the laboratory-scale than on the industrial-scale. Additionally, the industrial-scale experiment runs over longer timespans and is subject to fluctuations of ash input and neutralizing chemicals. The recoveries for Pb and Cu were significantly lower for MFA compared to MSWIFA, which is attributed to the higher redox buffer of WWFA40. However, further data are required for quantifying the negative effects.

The FLUWA process nevertheless represents a valuable option for treating WWFA as the heavy metal concentrations in WWFA are in the same range as for MSWIFA, and WWFA shows comparable heavy metal recoveries for Zn and Cd. The negative effects of the higher consumption of H2O2 affecting the Pb and Cu recovery may be diminished by a cotreatment of lower WWFA ratios.

### **5. Conclusions**

Heat and energy production in Switzerland using waste wood incineration is growing, and new treatment pathways must be implemented to recover heavy metals from the ashes and to reduce Cr(VI) content. Acid leaching, already established for MSWIFA, was found to be a valuable option for the treatment of WWFA. Laboratory-scale experiments were found to be suitable when evaluating the co-processing of MSWIFA and WWFA before implementing at the industrial scale.

Comparison of the chemical and mineralogical composition of WWFA with MSWIFA showed that WWFA could contain heavy metals (especially Pb) in elevated concentrations

similar to that of MSWIFA. The investigated WWFA samples showed Cr(VI) concentrations more than two orders of magnitude above the threshold value for landfilling. It was found that the concentrations in heavy metals, Cr(VI) and matrix minerals differed within the two WWFA types, depending on waste wood content. The elevated heavy metal and Cr(VI) concentrations in WWFA justify the need for treatment prior to deposition. The treatment with the FLUWA process allowed to successfully reduce the Cr(VI) in the filter cake until below the threshold value for landfilling, even when the process was performed under oxidizing conditions. The co-processing of WWFA required higher acid dosages due to its high ANC, but the Zn and Cd recovery were not negatively affected by the co-processing. Nevertheless, the co-processing of WWFA had a particularly negative effect on the recovery of the redox-sensitive elements Pb and Cu, as WWFA showed a strong redox buffer and thus a higher consumption of the oxidant H2O2. Therefore, higher dosages of H2O2 are needed to maintain oxidizing conditions during the process required for Pb and Cu mobilization. The use of a stronger oxidizing agent (e.g., permanganate) could be expedient and should be further tested with regard to successful Cr(VI) reduction. Alternatively, smaller percentages of WWFA could be co-processed in existing FLUWA plants in order to diminish the negative effects due to the higher demand for neutralization chemicals.

Within the next years, the implementation of co-processing the two ash types could contribute significantly to the growing demand for treatment capacities in Switzerland.

**Author Contributions:** Conceptualization, U.E., G.W. and M.W.; methodology, M.W., U.E. and G.W.; formal analysis, M.W.; investigation, M.W.; resources, M.W.; data curation, M.W.; writing original draft preparation, M.W.; writing—review and editing, U.E. and G.W.; visualization, M.W.; supervision, G.W. and U.E.; project administration, M.W.; funding acquisition, U.E. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was partially funded by the Swiss Federal Office for the Environment (FEN) and Office for Water and Waste (AWA, Canton of Berne).

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Acknowledgments:** Special thanks to Thomas Bücherer, Thomas Andres and Roland Furrer from Energiezentrale Bern for enabling the project, providing sample material and process details, as well as for their technical support during sampling and industrial-scale experiments. Many thanks to Stefan Schlumberger (ZAR) and Kaarina Schenk (FEN) for project feedback and discussion. Analytical support and/or assistance with industrial-scale experiments by Anna Zappatini, Wolfgang Zucha, Christine Lemp (University of Bern) and Stephan Fromm (ZAR) is highly acknowledged.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**

