**1. Introduction**

The demand for renewable heat and energy production using the CO2 neutral energy source wood has been growing enormously in Switzerland over the last decades—leading to strongly increasing amounts of wood ashes. In Switzerland, an annual load of 60,000 t wood ash arises from automatic firings through the energetic use of natural wood (e.g., forest) and from the thermal utilization of waste wood (e.g., coated, painted wood) [1]. A quarter thereof represents wood ash from waste wood enriched in heavy metals and Cr(VI). Depending on the incinerator plant and furnace, wood ash can be divided into up to three different fractions: grate ash, cyclone ash and filter ash [2]. The coarse-grained grate ash arises directly from the grate and is equivalent to bottom ash residing from municipal solid waste incineration (MSWI). This is the biggest fraction with roughly 60–90 wt % of the thermal residue [2]. The cyclone ash and the filter ash arise at the flue gas cleaning system and are often collected together and referred to as wood fly ash. Compared to grate ash, wood fly ash is enriched in volatile elements (e.g., Cl, heavy metals) since their low boiling point makes them evaporate during combustion and later precipitate at the flue gas cleaning system [3]. The chemical composition of wood ash is mainly dependent on wood quality and incineration conditions [3,4]. Factors affecting wood quality are wood type, compartment, growing environment and possible treatments

**Citation:** Wolffers, M.; Weibel, G.; Eggenberger, U. Waste Wood Fly Ash Treatment in Switzerland: Effects of Co-Processing with Fly Ash from Municipal Solid Waste on Cr(VI) Reduction and Heavy Metal Recovery. *Processes* **2021**, *9*, 146. https://doi.org/10.3390/pr9010146

Received: 30 November 2020 Accepted: 11 January 2021 Published: 13 January 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

(e.g., impregnation) prior to combustion [3,5]. While wood ash from natural wood may be enriched in organic pollutants (e.g., PAH, PCDD-/F [6]) and can contain naturally incorporated heavy metals in elevated concentrations, the contaminated waste wood ash is mostly characterized by highly elevated heavy metal concentrations (e.g., Zn, Pb, Cu, Cr) remaining from paints, coatings or impregnation [7]. Due to the oxidative conditions during combustion, Cr(III) compounds are mainly oxidized to very toxic and highly mobile Cr(VI) [8]. Thermal residues from waste wood which have been impregnated with Cr-bearing compounds are often severely enriched in Cr(VI). In water extraction tests, studies report Cr(VI) concentrations in waste wood ash that exceed the threshold value for landfilling [8,9].

Because of the possibly high contaminant load, wood ash is considered as waste and must be dumped on landfills, although opportunities for recycling are being sought (e.g., in concrete production [10]). The less polluted grate ash and fly ash from natural wood can be deposited without further treatment on landfill type D and E, according to the Swiss Waste Ordinance [11]. Waste wood fly ash (WWFA), however, must be treated before deposition due to the elevated concentrations in Cr(VI) and possibly environmentally harmful heavy metals. As there is currently not enough capacity available in Cr(VI) reducing facilities in Switzerland to treat the entire quantity of WWFA before landfilling, waste wood fly ashes can be deposited temporarily without a prior treatment on landfill type D or E (depending on their total organic carbon (TOC) content (<2 or <5 wt %, respectively) until 2023. From 2023 on, WWFA must be treated in order to reduce Cr(VI) and recover the heavy metals. Acid fly ash leaching with the FLUWA process [12,13] represents a promising method for treating WWFA prior to deposition. The FLUWA process represents the state-of-the-art process in Switzerland for recovering heavy metals (mainly Zn, Cd, Pb, Cu) from the similarly generated MSWI fly ash (MSWIFA). The ash is thereby leached with acid scrub water, acid from the flue gas cleaning system (~5% HCl). Filtration of the ash slurry yields a heavy metal enriched leachate that is then precipitated to a hydroxide sludge for subsequent heavy metal recovery and a filter cake depleted in heavy metals that is deposited on landfills. Because WWFA can yield heavy metal concentrations in the same range as MSWIFA, but occur in smaller quantities, a cotreatment of both ashes could be expedient. Beginning 2021, all Swiss MSWIFA must be treated before deposition [11] and depending on the heavy metal recovery guideline in revision, the use of an oxidant (e.g., hydrogen peroxide (H2O2)) will be necessary during the FLUWA process. Oxidizing conditions during metal extraction are a prerequisite to suppress the reductive precipitation of the redox-sensitive elements Cu, Pb, and to a minor extent, Cd. The co-processing of WWFA in the FLUWA process is already carried out at this study's investigation site Energiezentrale Bern, where it is economically favorable to co-process the arising WWFA40, a WWFA with 40% waste wood content. However, the heavy metal extraction efficiency of cotreating WWFA40 and the completeness of Cr(VI) reduction during the FLUWA process have not been investigated in detail.

The aim of the study was, therefore, to investigate the heavy metal recovery, as well as the Cr(VI) reduction efficiency of the FLUWA process when WWFA40 is co-processed. The actual state was investigated (reducing conditions) as well as the future state (oxidizing conditions) when oxidizing conditions become state-of-the-art for the FLUWA process. The industrial process was first simulated in laboratory-scale experiments in order to evaluate the leaching behavior of the different ash types (MSWIFA, WWFA40 and MFA (a mix of the two ashes to simulate the co-processing)) and to quantify the leaching efficiency in terms of heavy metal recovery and amount of neutralizing chemicals. The same experimental setups were later implemented at an industrial scale. Of special interest of the experiments was the heavy metal recovery efficiency, successful outcome of Cr(VI) reduction under reducing and oxidizing processing conditions and consumption of neutralizing chemicals (HCl, H2O2). The ashes used in the experiment were characterized with respect to their chemical and mineralogical composition, and their acid-neutralizing capacity (ANC) was determined. For comparison, the sample WWFA100 with 100% waste wood content was analyzed. To assess both the hazard potential of WWFA and the completeness of Cr(VI) reduction, the water-soluble and total content of Cr(VI) were determined for WWFA and filter cakes.

### **2. Materials and Methods**

### *2.1. Sample Origin, Sampling and Sample Preparation*

MSWIFA and WWFA40 samples were collected in 2017 at the waste and wood power plant Energiezentrale Bern. An annual amount of 135,000 t municipal solid waste and 65,000 t of wood (25% water content, fluidized-bed combustion) was incinerated in separate incinerators in 2017. MSW and wood are combusted separately but treated together prior to landfilling with the FLUWA process in two consecutive extraction reactors (1 m<sup>3</sup> each). The co-processing of WWFA40 in the FLUWA is economically favorable since excess acidity of their scrub water is consumed by the alkalinity of the WWFA40, and the use of lime milk is minimized. At present, the FLUWA process at the incineration plant is performed under reducing conditions (without the addition of H2O2) with MSWIFA and WWFA40 proportions in the ratio as they are produced. Adjustments of the ash ratio are made in the current process such that a favorable extraction pH of 3.8 is achieved. To perform industrial experiments at oxidizing conditions, a pumping system was installed for continuous dosing of H2O2.

In order to understand the geochemical differences between MSWIFA and WWFA40, three representative composite samples of each ash type were investigated in terms of chemical and mineralogical composition. The sampling duration varied between one and three weeks. Samples were taken twice a day and mixed into composite samples. Additionally, three samples of WWFA40 (weekly composite samples) and their corresponding filter cakes were made available for Cr(VI) analyses. For comparison, the sample WWFA100 (100% waste wood, monthly composite sample) from a Swiss biomass power plant was investigated.

Approximately 10 kg of ash was collected in each sampling campaign. The ashes were homogenized and split into 1 kg working batches and dried at 105 ◦C until constant weight for chemical analysis and at 40 ◦C for mineralogical analysis.

## *2.2. Chemical Analysis*

The elemental composition of the ashes was obtained through energy-dispersive X-ray fluorescence (ED-XRF) analysis performed on pressed powder pellets (4.0 g ash, 0.9 g wax as a binder) using a Spectron Xepos (SPECTRO, Kleve, Germany) spectrometer with matrix adjusted calibration. For quality control, the samples were analyzed in duplicates. The accuracy of the method was previously verified by the authors [14] through multiple determinations of similar ash samples and the analysis of the standard reference material BCR 176R [15]. The ED-XRF measurements showed good reproducibility within <2% for the elements Cu, Zn, Cd, Sb, Pb, Br, Sn, Ba, within 5% for Al, Si, S, Cl, Ca, Ti, Mn, Fe, Cr, Sr and within <10% for K, Na, Mg. Extract solutions were analyzed by inductively coupled plasma optical emission spectroscopy (ICP-OES) on a Thermo Scientific iCAP 700 Series (Waltham, MA, USA) after dilution with HNO3 1% and calibrated with the multielement standards CertiPUR IV and X (Merck, Kenilworth, NJ, USA). The analytical error accounts for ±5% for all elements except Na, K, Ca, Sb and S that showed ±10% error based on multiple measurements of certified standard solutions.

### *2.3. Mineralogical Analysis*

The mineralogical composition was obtained throughout X-ray powder diffraction (XRD) analysis using a Panalytical X'Pert Pro diffractometer (CuKα-radiation) (Malvern Panalytical, Almelo, Netherlands). 4 g of material was mixed with 1 g of internal standard (corundum) and ground dry for 6 min at 55 Hz in an XRD McCrone mill from Retsch (Haan, Germany). Measurement was performed on disoriented samples from 5 to 75◦ 2Theta at 40 kV acceleration voltage and at an electron generating current of 40 mA. An

automatic divergence slit was used. For quantification, TOPAS-Academic software (V6, Coelho Software, Brisbane, Australia) was used (Rietveld refinement). The extended uncertainty is attributed to 50% for concentrations <1%, 20% for concentrations <5% and 10% to concentrations >5%. The structural data (\*.cif files) of the inorganic crystal structure database (ICSD) was used, and phase concentrations and amorphous part were calculated based on the internal standard.

### *2.4. Acid-Neutralizing Capacity*

Acid-neutralizing capacity (ANC) titration was performed using a 785 DPM Titrino device from Metrohm (Herisau, Switzerland). with 2 g of ash 20 mL ultrapure water. Every 10 min, 1 mL of HCl 1 M was added under continuous stirring until the ash slurry reached a pH of 2.

### *2.5. Water-Extractable Cr(VI)*

The content of water-soluble Cr(VI) was determined in WWFA samples in order to make statements about the hazard potential of WWFA. Furthermore, the content of watersoluble Cr(VI) in the filter cakes was determined to examine the completeness of the Cr(VI) reduction during the FLUWA process. Cr(VI) analyses were performed on the eluates obtained with the standard eluate test F-22 [16]. The ash and filter cake samples were thereby eluted with ultrapure water for 24 h at a liquid to solid (L/S) ratio of 10 (4 g ash, 40 mL ultrapure water). For quality control, as well as to investigate transformations in the redox state during the eluate test, Cr(III) and Cr(VI) spikes were added, and the eluate test was performed two times per sample, each spiked and unspiked. Selected samples were further eluted as duplicates. For the spiking, the concentration of 50 mg/kg Cr(III) in the form of CrNO3·9H2O and 10 mg/L CrO4 <sup>2</sup><sup>−</sup> for the Cr(VI) spike were added for the eluate test. The comparison of the Cr(VI) concentrations of the two eluates per sample allowed calculating the recovery of the spike. The water-extractable Cr(VI) content in the eluates was determined spectrophotometrically with a Merck Spectroquant Pharo 100 after complexation with diphenylcarbazide (DPC) at an absorption maximum of 540 nm. For complexation, a Spectroquant chromate test set (Merck, No 1.14758, Darmstadt, Germany) was used. No determination of Cr(VI) in the FLUWA leachate was performed as the Cr concentration in the leachates were below the detection limit in all samples.

### *2.6. Total Cr(VI)*

For a more comprehensive assessment of the hazard potential of Cr(VI), the total content of Cr(VI) was determined for two WWFA samples (WWFA40\_2 and WWFA100). Total Cr(VI) extraction on WWFA was performed on duplicates by hot alkaline extraction (method 3060A [17]). For quality control, the standard reference material NIST 2701 [18] was analyzed for total Cr(VI) content, and the determined value was within the given uncertainty. The accuracy of the method on similar materials was previously verified [19] by the authors, and reproducibility within 10% was attributed to the method based on spike recoveries and multiple measurements of the standard reference material [20]. Analysis of Cr(VI) concentration in the extract solutions was performed with ICP-OES after the use of CHROMAFIX PS-H+ cation exchange cartridges (Macherey–Nagel, Düren, Germany) in order to retain Cr(III). The effectiveness of the cartridges on similar extract solutions was tested in previous studies [19]. In order to investigate Cr redox transitions during extraction, a parallel extraction with Cr(VI) spiking was performed for each sample by adding a concentration of 100 mg/kg PbCrO4.

### *2.7. Laboratory Experiments*

In order to investigate the effects of co-processing WWFA40 in the FLUWA process, the process was simulated on a laboratory-scale at conditions feasible on an industrial-scale. At a laboratory-scale, mass balances can be quantified precisely, and process conditions (pH, Eh) can more easily be controlled. Three different ash types were used for the experiment: MSWIFA (sample MSWIFA\_2), WWFA40 (sample WWFA40\_2) and MFA (the ash mix of WWFA40\_2 and MSWIFA\_2, mixed at a ratio of 1:2, to simulate a cotreatment). Two different experimental setups were performed at a laboratory-scale (Table 1): without the use of H2O2 (30%) and with a concentration of 40 L H2O2/t ash, which represents standard plant conditions. All experiments were performed two times in order to assure reproducibility.


**Table 1.** Experimental setups for laboratory- and industrial-scale experiments.

150 mL of artificial acid scrub water (HCl 5% with 25 g/L NaSO4) was heated in a beaker glass to 40 ◦C before adding the ash (50 g, liquid to solid ratio L/S = 3). Due to the exothermic reaction between the scrub water and the ash, the temperature rose immediately to 60 ◦C. Under continuous stirring, the ash slurry reacted for 60 min at 55–60 ◦C. The pH and Eh values (Ag/AgCl) were recorded temperature-compensated. The pH was controlled by adding HCl (32%). Before filtration, the pH was adjusted to a value of 3.8 with NaOH (65%). The slurry was filtered with a vacuum filter device. 100 mL of deionized water was used to wash the filter cake. For the experiments performed with H2O2, a concentration of 40 L/t H2O2 was added by 10 consecutive portions (at a 3 min interval) to the ash slurry in the first half of the experiment. The leachates were diluted with HNO3 1% and further analyzed by ICP-OES. The filter cake was weighed and dried at 105 ◦C until a constant mass was reached. The heavy metal recovery was determined from mass balance calculations.

### *2.8. Industrial-Scale Experiments*

Both the current state (reducing conditions) as well as the future state (oxidizing conditions) were investigated at an industrial scale (Table 1). Each ash type (MSWIFA, WWFA40 and MFA (ratio WWFA40 to MSWIFA of 1:2.5)) was treated separately without the addition of H2O2. Additionally, experiments with 40 and 60 L/t H2O2 (30%) (Merck, Darmstadt, Germany) were performed for MSWIFA and MFA. At laboratory-scale experiments, it was shown that WWFA40 required higher H2O2 dosages for oxidizing conditions to persist. Therefore, industrial-scale experiments were also conducted with 60 L/t. The efficiency of the cotreatment of WWFA40 and MSWIFA was compared to the scenarios where only WWFA40 or MSWIFA was treated solely. It was found that the treatment of wood ash solely is difficult at an industrial scale due to difficulties with filtration. Therefore, the

industrial-experiments with WWFA40 solely were not carried out with H2O2 and the focus was put on the cotreatment with MSWIFA.

The different experimental setups were performed at Energiezentrale Bern on different days for time periods between 4 and 24 h. A calculated amount of H2O2 was added continuously to the first extraction reactor. During each experiment, multiple samples (in 30- or 60-min intervals, depending on the experiment duration) were taken from the input ash and from the filter cake and combined into composite samples. The pH and Eh were monitored. An average L/S-ratio of 15 was calculated on annual mass flux balances since flow measurement of the leachate is not implemented. The ash and filter cake samples were dried at 105 ◦C and analyzed for their elemental composition with ED-XRF in order to calculate the recovery.

### **3. Results**

### *3.1. Chemical Composition*

The main constituents of MSWIFA are Ca, Cl, Si, S, Na (>75,000 mg/kg, Table 2). Zn, K, Al (>35,000 mg/kg) and Fe (>10,000 mg/kg) are subsidiary constituents. WWFA differ in chemical composition from MSWIFA, and there are also chemical differences between WWFA40 and WWFA100. WWFA generally shows higher Ca and Si concentrations (>100,000 mg/kg) than MSWIFA, followed by K, Al, Cl, S, Fe (>20,000 mg/kg). Notable is the higher concentration in S and heavy metals (mainly Pb, Zn, Cu) in WWFA100 compared to WWFA40, the latter showing higher concentrations in matrix elements (Ca, Si, K, Mg) in contrast. The concentrations of the main- and subsidiary constituents in the three MSWIFA samples vary within 10–20%, but Cu shows variations by more than 30%. The variations in chemical composition are smaller for WWFA40 than for MSWIFA. The element concentrations vary mostly within 10% between the three WWFA40 samples (except for about 20% for Cr, Mn and Pb). MSWIFA shows strongly elevated concentrations in the potentially harmful heavy metals Pb, Cu, Sb, Cd, some exceeding the threshold for landfilling by multiple times (Pb) or orders of magnitude (Sb, Cd). WWFA shows considerably lower Cd and Sb concentrations (lower by almost two and three orders of magnitude, respectively) but shows strongly elevated Pb concentrations. In two samples of WWFA40, Pb concentration is only 10% lower than in MSWIFA. In WWFA100, Pb concentrations exceed those in MSWIFA by more than double. Zn and Cu concentrations in WWFA40 are about one-third of the concentration in MSWIFA, and in WWFA100, about half. In contrast, WWFA shows higher concentrations in Ba, Cr, Fe, Mn and Ti concentrations than MSWIFA. WWFA100 further shows a very high TOC content of 74,400 mg/kg—which exceeds the threshold value for landfilling.

### *3.2. Mineralogical Composition*

The difference in chemical composition between MSWIFA and WWFA is also represented by a different mineralogical composition (Table 3). As a result of the high Cl content, Cl salts such as halite, sylvite and the Zn bearing K2ZnCl4 are important phases in MSWIFA. Anhydrite represents another main phase, together with several silicates (e.g., gehlenite) and calcite. The main mineralogical differences between WWFA40 and WWFA100 are in calcite and anhydrite content. WWFA40 show very high calcite concentrations (17–28 wt %) and high concentration in quartz (7–10 wt %). Sylvite, periclase and Ca-, Al-, Na- silicates form minor components in WWFA40. WWFA100 shows calcite, gehlenite, anhydrite and magnesite as main phases (>7 wt %), followed by minor concentrations in quartz and Ca-, Al-, Na- silicates. The presence of amorphous phases is clearly visible in all spectra by a bump in the background between 20 and 40◦ 2Theta and was calculated to make 35–50% of the total content.


**Table 2.** Elemental composition of the analyzed municipal solid waste fly ash (MSWIFA) and waste wood fly ash types WWFA40 and WWFA100, determined by energy-dispersive X-ray fluorescence (ED-XRF).

**Table 3.** Mineralogical composition of the analyzed MSWIFA, WWFA40 and WWFA100, determined by XRD.


### *3.3. Acid-Neutralizing Capacity*

Acid-neutralizing capacity (ANC) is similar for ashes of the same type but shows differences between the ash types (Figure 1). The amount mol H+ needed to reach an optimal extraction pH of 3 varies within 10% and 5%, respectively, for MSWIFA and WWFA40. On average, MSWIFA consumed 4.7 mol H+, WWFA100 6.2 mol H+, and WWFA40 with 8.9 mol, almost double the amount of H<sup>+</sup> to reach a pH of 3. The titration curve of WWFA40 is characterized by a small plateau at pH 12.7 and a big plateau at pH 7. Although WWFA100 shows a similarly high initial pH as WWFA40, the ANC titration curve resembles more that of MSWIFA than that of WWFA40, as the plateau starting at pH 7 is less pronounced. The pH starts dropping rapidly from an initial value of 12.7 to pH 7, where calcite buffering starts. For MSWIFA, the initial pH of the titration curve is lower (pH 11.5) and drops rapidly towards pH 5—where an almost linear decrease in pH initiates.

**Figure 1.** Titration curves of acid-neutralizing capacity (ANC) for MSWIFA, WWFA40 and WWFA100. The samples used for the leaching experiments are indicated with \*.

### *3.4. Water-Extractable and Total Cr(VI)*

Multiple determination of the water-soluble Cr(VI) concentration of the samples revealed reproducibility within 10%. The Cr(III) spike was fully retained during all eluate tests, which proves that no oxidation of Cr(III) occurred. The Cr(VI) spike was fully retained during the majority of the eluate tests. In the experiments where the Cr(VI) spike was not fully retained, redox transformations leading to the reduction of Cr(VI) occurred (indicated with \* in Table 4).

**Table 4.** Water-extractable Cr(VI) and total Cr(VI) concentrations of WWFA and the filter cakes of WWFA40. Cr(VI) concentrations indicated with \* are from extractions with poor Cr(VI) spike recoveries (<2%), implying erroneously low values. Where not indicated, the Cr(VI) spike was fully retained.


All eluates of the 6 analyzed WWFA40 samples showed water-extractable Cr(VI) concentrations that exceed the threshold limit for landfilling (0.5 mg/kg) by more than two orders of magnitude. The water-extractable Cr(VI) content made up for 10–20% of the Cr concentration in WWFA40. For WWFA100, the determined water-extractable Cr(VI) concentration was as low as 1 mg/kg. Since the Cr(VI) spike recovery was only 2% for WWFA100, this implies that a major part of the sample's native Cr(VI) also has been reduced. All six analyzed filter cakes from WWFA40, including all filter cakes from the experiments performed in this study, showed Cr(VI) concentrations below the given threshold for landfilling, independent of the applied scale (laboratory or industrial) and the redox conditions.

For the hot alkaline extraction of sample WWFA40\_2, none of the Cr(VI) spike was recovered, in neither of the duplicates. The total Cr(VI) concentration was expected to be similar or higher to the measured water-extractable Cr(VI) concentration. Instead, a concentration of 1 mg/kg was measured. For WWFA100, Cr(VI) spike recovery was 80% in the hot alkaline extraction and the measured total Cr(VI) concentration 87 mg/kg. The results of the double determination agreed within 3%. It is assumed that matrix interferences occurred during the hot alkaline extraction of the sample WWFA40\_2 (and to a minor extent in sample WWFA100), leading to a strong diminution in Cr(VI) concentration. This might have been favored by the strongly reducing conditions during the hot alkaline extraction with WWFA40\_2.

### *3.5. Laboratory-Scale Leaching Experiments*

The heavy metal recovery achieved for the two different experimental setups is shown in Figure 2. The reproducibility of the experiments performed in duplicates was very good (within 5–10%). Only for Cu, the reproducibility was within 20% since the solubility of Cu is strongly pH-dependent, and a small increase in the filtrate pH value can enhance precipitation of Cu hydroxides [21]. Given the attributed uncertainty of 10%, Zn recovery can be considered equal for MSWIFA and MFA, whereas it was lower by 30% for WWFA40. This lower yield is associated with a high leachate pH of 5.4, which assumedly led to Zn precipitation. For Cd, recovery is 40% lower for MFA compared to MSWIFA, whereas no Cd was recovered from WWFA40. No Pb and Cu were mobilized for any of the ash types without the use of H2O2.

**Figure 2.** Heavy metal recovery in % for the target heavy metals Zn, Cd, Pb, Cu for the 3 different ash types in laboratory experiments. The pH of the leachate is indicated in the bar for Zn recovery. (**a**) recovery without the use of H2O2 (**b**) recovery for experiments with 40 L/t H2O2.

When a quantity of 40 L/t H2O2 was used, Zn recovery did not change. The recovery for Cd could almost be doubled from MSWIFA, and it was achieved to recover Pb (53%) and Cu (38%). This significantly higher recovery for Pb and Cu (and to a minor extent Cd) when using H2O2 was observed in previous studies [14]. The recovery for MFA is equal (within the uncertainty) to that of MSWIFA for Zn and Cd. For Pb, recovery is lower by 25%, and almost no Cu was mobilized during the experiment. For WWFA40, the same recovery for Zn was achieved as for the other ash types, but any of the other heavy metals

could be recovered. In the experiment with H2O2, the amount of HCl 32% needed to keep extraction pH at a level of 2.5 was twice as high for MFA compared to the experiments with MSWIFA (17 vs. 9 mL, respectively), and the amount needed for WWFA40 was 33 mL.

The ashes showed a strong redox buffer, visible by the subsequent drop in redox potential after each H2O2 dosage (Figure 3). For MSWIFA, the redox potential dropped to strongly negative values shortly after the H2O2 dosage, whereas for WWFA40, Eh was still positive before the next H2O2 dosage. Thus, H2O2 consumption seemed to be slower. However, the amount of H2O2 added was not enough to maintain oxidative conditions over the entire extraction time for MFA and WWFA40. Only for MSWIFA, it was possible to maintain a stable positive redox potential over the entire experiment with 40 L/t H2O2.

**Figure 3.** Evolution of redox potential (Eh) during laboratory experiment with H2O2. **Left**: Eh before and after each dose of H2O2 (indicated with \*). **Right**: evolution of Eh after the last H2O2 dosage.

### *3.6. Industrial-Scale Leaching Experiments*

Taking into account the attributed uncertainty of 10%, the recovery for Zn was the same for the three ash types independent of the amount of H2O2 added. The recovery for Cd reflected the trends observed from the laboratory experiments: a lower Cd recovery by one-third for MFA and a negligible Cd recovery for WWFA40. As already observed in the laboratory experiments, the recovery for Pb and Cu was negligible without H2O2 (Figure 4a). With 40 L/t H2O2, 55% of Pb and 16% of Cu could be mobilized from MSWIFA, but only 12% Pb and 3% Cu from MFA (Figure 4b). The recovery of both Pb and Cu was thus significantly lower for MFA compared to MSWIFA.

**Figure 4.** Recovery for the industrial-scale experiments (**a**) without and (**b**) with H2O2 (40 and 60 L/t H2O2). The pH of the leachate is indicated in the bar for Zn recovery. The industrial-scale experiments with H2O2 were not performed for WWFA40.

With the higher dosage of 60 L/t H2O2, a higher Cd fluctuation in recovery could be observed, but considering the attributed uncertainty of 10%, the recoveries were comparable for the two ash types. For Pb recovery, a strong increase by a factor of three could be observed for MFA, whereas Pb recovery did not increase with the higher dosage for MSWIFA. A clear increase in Cu mobilization could be observed for both ash types, which resulted in recoveries 2.5 and 6 times higher than with 40 L, respectively. Table 5 lists the heavy metal concentrations of the ashes investigated in the industrial-scale leaching experiments. Considerable differences in heavy metal concentrations could be observed, which must be taken into account when comparing the recoveries.

experiments. The concentrations for the experiments with 40 and 60 L/t H2O2 are indicated with \* and \*\*, respectively. **Experiment without H2O2 Experiments with H2O2 mg/kg mg/kg**

**Table 5.** Heavy metal concentrations of the different ashes investigated in industrial-scale leaching

