**1. Introduction**

When society moves towards more sustainable material cycles, recovery of metals from industrial waste can be an opportunity to turn waste into a valuable resource. It is estimated that numerous metals will be consumed in less than 50 years (Zn) or 100 years (Co, Cu, Ni, Mo, and Pb) if current rates of extraction are maintained [1]. However, sustainable reserves of these metals will remain for the future if they are recovered and the remaining reserves are used more efficiently [2]. Waste containing significant amounts of trace elements offers a potential for the recovery of these elements, since trace elements in anthropogenic matrices are often more labile compared to natural matrices (e.g., ores, rocks) [3]. Generally, the term "major elements" is used for elements which have concentrations greater than 1% and term "trace elements" has been used in geochemistry for chemical elements that occur in the Earth's crust in amounts less than 0.1% (1000 mg·k−1) [4]. In this study, the term "trace elements" is used to indicate particularly the elements As, Cd, Cr, Co, Cu, Mo, Ni, Pb, and Zn while the term "major element" is used to indicate the following elements: Al, Ca, Fe, K, Mg, P, S, and Si.

Waste incineration has been a preferred alternative in solid waste managemen<sup>t</sup> since landfilling became more difficult to site due to high costs, diminishing land availability, and stricter regulations [5]. The bottom ashes (BAs) from waste incineration are often treated to recover the metals or they are reused in the cement and concrete industries to produce road construction materials in Europe and developed countries [6,7]. However, without any pretreatment, the residues which contain high amounts of leachable potentially toxic elements are also classified as hazardous wastes [8].

One of the key factors when evaluating the risks related to the managemen<sup>t</sup> (e.g., disposal/reuse) of solid waste regarding environmental health and safety is the release of pollutants to the receiving environment. Co-disposal of different kinds of waste without knowing their leaching characteristics may pose a threat to environment, for example, the disposal of ash in an acidic environment increased the leaching of heavy metals and contaminated the ground water at the disposal site [9]. For this reason, leaching/extraction tests are commonly applied because they provide information about the potential release of constituents from solid materials to the liquid phase [10]. There are several extraction/leaching tests for solid materials, each with different purposes and implications. Among these tests, an ammonium-EDTA (ethylenediaminetetraacetic acid) extraction is usually used to determine the potential mobility of trace elements in soils and sediments as a consequence of complexation. Different types of EDTA (EDTA free acid, sodium-EDTA, and ammonium-EDTA) are used as extraction solutions [11]. Acetic acid (CH3COOH) in different concentrations (0.43 mol·L−<sup>1</sup> or 0.11 mol·L−1) is often used to determine the fraction of elements in a soil or sediment that is susceptible to changes in acidity of the environment (e.g., through acid rain) [12]. Beside acetic acid, nitric acid is currently employed in pH-dependence or pHstat tests to analyze the leachability of an element under acidic conditions since acetic acid is a weak acid that may form complexes.

For the sake of harmonization, ammonium-EDTA 0.05 mol·L−<sup>1</sup> and CH3COOH 0.43 mol·L−<sup>1</sup> were selected as extraction solutions by the Standards Measurement and Testing (SM&T) Program to indicate respectively the "mobilizable/potentially available fractions" or the "acid extractable fractions" of elements in sediments and soils [13]. Since the release of trace elements from soils, sediments and waste materials is strongly influenced by pH, different types of tests are available to assess the release of elements due to changes in pH. While single extractions (e.g., the extraction with CH3COOH) allow estimating element release at a single pH value, which is determined by the reagen<sup>t</sup> and by the acid neutralizing capacity of the sample, pHstat leaching tests allow investigating element release at a pre-defined pH value. In the standard pHstat test [14], the release of trace elements is only assessed at one moment in time (after 48 h), which does not allow addressing the kinetics of the release of elements [14]. Hence, a detailed batch leaching test where the pH is kept at a preset value by using an automatic titration over time in conjunction with the analysis of the leachates at various times allows determining the release kinetics of elements due to changes in pH. This kind of test provides a more detailed assessment of time-dependent leaching behavior of trace elements.

Element leaching from BA has been the subject of many studies in different countries. An overview of some selected papers dealing with trace element leaching from BA which are relevant for the present study is provided in Table 1. It was found that leaching of trace elements from BA is pH-dependent and affected by accelerated ageing [15–17]. In developing countries, BA is sometimes treated differently from what is stipulated in legislation. For instance, some hazardous waste incinerators leave the BA covered at a dump site next to the plants (e.g., Ghana [18] and Vietnam). As for most developing countries, rapid industrial development has led to an increase in the generation of various types of industrial waste in Vietnam in recent years. However, data on the composition of BA from hazardous waste incinerators in Vietnam, as well as the potential release of trace elements from these materials are rather scarce, even though such studies provide essential information to select the most sustainable managemen<sup>t</sup> options for this kind of waste, and eventually also contribute to the protection of the environment.


**Table 1.** Total concentrations (dry weight basis) of major and trace elements, mineral composition, organic carbon content (OC), and pH, and of studied BA compared to other studies. Average ± standard deviation of 2 replicates.

> \* = BA sample 1; \*\* = BA sample 2.

In the present study, BA samples that were previously partly characterized using chemical, mineralogical, and physical approaches [19] were further studied through a complementary extraction/leaching approach with focus on the potential release of trace elements under influence of acidification and complexation. Single extraction with ammonium-EDTA (0.05 mol·L−1) and CH3COOH (0.43 mol·L−1), which are commonly applied to soils and sediments, were performed. The results of the single extractions were compared with the release of elements during pHstat titration leaching at pH 4. This pH was chosen since it represents the worst-case scenario when the samples would become exposed to acidification at the disposal site (e.g., acid rain conditions, mixing with other acid wastes, etc.). The comparability of different extractions and leaching tests is determined by the main characteristics of the tests (the chemical reagen<sup>t</sup> used, the duration of the tests, the liquid/solid (L/S) ratio). The tests used in this study are all conventional batch tests in which the leaching solution is not renewed, and the mixing is performed over a relatively short time period (hours to days) with the aim of reaching equilibrium conditions. The comparison of potential mobility of trace elements by different and extractions is not completely straightforward since operational conditions and reagents are different. The estimation of the potential trace elements' mobility is thus operationally defined by the extracting agents used [23].

The aim of the study is to investigate the usefulness of a different extraction/leaching approach focusing on the potential release of trace elements from BAs from hazardous waste incinerators under influence of acidification and complexation. The kinetic of release of the various elements under acidic conditions is also discussed and related to the solid-phase composition of the BA. It is not the purpose of the present paper to investigate the metal recycling from BA, since this would also require a detailed

economic and technological assessment. However, strategies for disposal or recycling of solid waste are also often based on the understanding of the leaching behavior of these materials [10]. In this study, the information deduced from the different methods was compared and evaluated. Results of the present study are helpful for the first steps in the evaluation the potential treatment strategies for these BAs. Results from extraction methods applied in the present study might be indicative for evaluation of options for metal recovery. Additionally, the investigation of release kinetics of trace elements under acidic conditions by a detailed batch leaching test is helpful to gain a better understanding of acid neutralization capacity as well as the solid-phase composition of the BA. Since this is just the first estimation for the usefulness of the CH3COOH and ammonium-EDTA extraction, we use standardized conditions, using more diluted extraction agents and high liquid/solid (L/S) ratios instead of severe conditions with more concentrated agents and low L/S ratios. Moreover, mild extraction conditions are used since the chemical consumption should be minimized during BA washing for multiple reasons. Several studies have dealt with bottom ash from municipal solid waste incineration; however, bottom ash from hazardous waste incineration has not been the subject of many studies before. This paper will improve our understanding of the release of trace elements under different external factors that BAs may encounter during their processing or management. Extraction and leaching methods applied in the present study are standardized tests that have been developed for soils and sediments. Developing novel extractions for waste materials is not the purpose of the present study. However, the approach used in this study is useful to evaluate the fraction of metals that can be leached out from BA, in view of metal recovery and treated BA disposal/use. This approach is necessary for the development of appropriate waste managemen<sup>t</sup> options, especially in countries facing inefficient waste and waste water treatment technologies, resulting in residual waste materials with considerable concentrations of valuable elements.

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

#### *2.1. Material Characterization*

The BA samples in the present study were collected from two hazardous waste incinerators in Ho Chi Minh City, South Vietnam. At the time of sampling, source material from the plant from which sample AS1 was collected included sludge from the waste water treatment of textile and printing ink production factories, and chemical containers. The general input for the incineration facility where AS2 was collected was sludge from waste water treatment of a textile dying factory, cloths containing chemical and lubricants, and out of date chemicals. After sampling, the BA materials were stored in plastic bags and brought to KU Leuven, Belgium for further treatment and laboratory analysis. More details about sampling and sample pretreatment and the determination of chemical and mineralogical composition of the samples can be found elsewhere [19]. In general, the composition of BA varies greatly and depends on the input material and the applied incineration technology. The BA samples in the present study were collected from two-stage incinerators having a quite low combustion temperature (550–650 ◦C) which can influence the complete burning of organic matter and the forming of metal oxides. Moreover, inputs of hazardous waste incinerators vary widely depending on the type of plants.

Some general characteristics (pH, total elemental concentrations, organic carbon content and mineralogical components) of these samples compared to other studies are summarized in Table 1. The fraction >2 mm determined by sieving was 12% for both samples. Mineralogical characterization was carried out by X-ray diffraction (XRD). According to XRD results, the studied BAs were mainly composed of oxides and Ca-, and Si-minerals. Magnetite (Fe3O4), quartz (SiO2), and calcite (CaCO3) were found in both samples. Corundum (Al2O3) and mullite (Al6Si2O13) were detected in sample AS1 and hematite (Fe2O3) and rutile (TiO2) were found in sample AS2 (Table 1). Total element concentrations in these samples were compared to the composition of other BA from different sources including industrial wastes, MSWI, coal, peat and forest residues, and hospital medical

waste incinerators. Both samples were characterized by a low concentration of Ca and P compared to the composition of other BA. Iron concentration in AS2 was much higher compared to its concentration in AS1 and other BA (Table 1). The total element concentrations of some trace elements in the two BA samples were compared to the Flemish limit values for recycling of granular material in construction applications to evaluate whether the studied BAs would be suitable for being used in construction applications. Results indicated that Cu (sample AS1 and AS2), Ni (sample AS1) and Zn (sample AS2) exceeded the Flemish limit values for recycling of granular material in construction applications [24]. Moreover, according to the Vietnamese National Technical Regulation on Hazardous Waste Thresholds, As and Pb (sample AS2) exceeded the limit values [25]. This might raise concerns regarding the potential hazard of using this BA as a construction material if no pretreatment is applied.

## *2.2. Extractions/Leaching Test*

Extractions and leaching test were performed at room temperature (20 ◦C). The ammonium-EDTA extraction followed the protocol of the Standards, Measurement and Testing (SM&T) Program [26]. Ammonium-EDTA 0.05 mol·L−<sup>1</sup> solution was added to the dry sample with a liquid/solid (L/S) ratio of 10 (L·kg−1). The suspension was shaken for 1 h in a reciprocal shaker, centrifuged (3500 rpm, 10 min), decanted and filtered (0.45 μm, Chromafil® PET-45/25, Macherey, Düren, Germany). The acid extractable fraction (generally considered to consist of exchangeable elements and elements bound to carbonates) was also based on the procedure described by the SM&T programme using an extraction with CH3COOH 0.43 mol·L−<sup>1</sup> [27]. The CH3COOH 0.43 mol·L−<sup>1</sup> solution was added to the dry sample to obtain a liquid/solid (L/S) ratio of 40 (L·kg−1). The suspension was shaken for 16 h in a reciprocal shaker, centrifuged (3500 rpm, 10 min), decanted off and filtered (0.45 μm, Chromafil® PET-45/25, Macherey, Düren, Germany). After extraction, the pH of the extracts was measured with a pH Hamilton single-pore electrode.

The pHstat leaching test was carried out employing an automatic multititration system (Titro-Wico Multititrator, Wittenfield and Cornelius, Bochum, Germany). 80 g of dried sample was put in an Erlenmeyer flask together with 800 mL of Milli–Q water (L/S ratio = 10 <sup>L</sup>·kg−1). Element release was measured at regular time intervals (0, 1, 3, 6, 12, 24, 48, 72 and 96 h). Bottom ash samples were first shaken for 30 min at the natural BA-pH (without acid addition) before the effective pHstat- experiment was started. A sample of 5 mL of the suspension was taken over a filter (0.45 μm, Chromafil® PET-45/25, Macherey-Nagel GmbH & Co. KG, Düren Germany) by means of a syringe attached to a flexible tube at regular time intervals. This pHstat test corresponds to the CEN/TS 14997 test [14], except that the leachate is not only sampled after 48 h, but at regular time intervals. Moreover, in the present study, the pHstat tests were only performed at pH 4. The objective of this test was to examine the kinetic release of trace elements under acidic conditions, while the pH of the suspension (bottom ash suspended in water) is kept at a constant value (pH = 4) by continuous titration with HNO3 (1 mol·L−1). Due to the variety of processes occurring at landfills or open dumps such as sulfide oxidation, microbial activity, acidic deposition and reaction with atmospheric CO2, the pH of ash in landfills may drop to a value as low as 3 to 5 [28]. Hence pH 4 was chosen to address a worst-case scenario of acidification.

#### *2.3. Analysis and Quality Control*

Single extractions with ammonium-EDTA 0.05 mol·L−<sup>1</sup> and CH3COOH 0.43 mol·L−<sup>1</sup> were done in duplicate and the results are presented as a mean value. Blank solutions were also inserted in each batch of extractions. A certified reference material (BCR 483) was also analyzed for quality control (Table 2).


**Table 2.** Comparison of the results of the ammonium-EDTA 0.05 mol·L−<sup>1</sup> extraction and CH3COOH 0.43 mol·L−<sup>1</sup> extraction of BCR 483 (average ± standard deviation of 2 replicates) and certified values.

The extracts/leachates were acidified immediately after the experiments, with concentrated HNO3 to bring the pH < 2. The EDTA extracts were kept at 4 ◦C and not acidified prior to analysis to prevent precipitation of EDTA salts at low pH. Elemental concentrations (Al, Ca, Fe, K, Mg, P, S, As, Cd, Co, Cr, Cu, Mn, Mo, Ni, Pb and Zn) were measured by ICP-OES (Varian 720-ES). A Varian 720-ES instrument supplied with double-pass glass cyclonic spray chamber, concentric glass nebulizer SeaSpray and "extended high solids" torch was used. Solutions were introduced into the spectrometer using the Varian SPS3 sample preparation system. Calibration solutions were prepared from certified multi-element ICP standard stock solutions and from Plasma HIQU (High Quality) single element solutions from CHEM-LAB (Belgium). Blanks were also included in the calibration. All solutions were prepared from 18 M Ω·cm<sup>−</sup><sup>3</sup> ultra-pure water supplied from Millipore system and stabilized with ultra-pure nitric acid (CHEM-LAB). Sensitivity, linear dynamic range, and freedom from spectral interferences were taken into consideration during wavelength selection for each element. Each measurement was carried out with three replicates.
