Waste Management xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Waste Management journal homepage: www.elsevier.com/locate/wasman
Quantification of the resource recovery potential of municipal solid waste incineration bottom ashes Elisa Allegrini a,⇑, Alberto Maresca a, Mikael Emil Olsson a, Maria Sommer Holtze b, Alessio Boldrin a, Thomas Fruergaard Astrup a a b
Technical University of Denmark, Department of Environmental Engineering, Building 115, 2800 Lyngby, Denmark Afatek Ltd., Selinevej 18, 2300 Copenhagen S, Denmark
a r t i c l e
i n f o
Article history: Received 28 February 2014 Accepted 8 May 2014 Available online xxxx Keywords: MSWI Bottom ashes Non-ferrous metals Recovery efficiency REE Critical elements
a b s t r a c t Municipal solid waste incineration (MSWI) plays an important role in many European waste management systems. However, increasing focus on resource criticality has raised concern regarding the possible loss of critical resources through MSWI. The primary form of solid output from waste incinerators is bottom ashes (BAs), which also have important resource potential. Based on a full-scale Danish recovery facility, detailed material and substance flow analyses (MFA and SFA) were carried out, in order to characterise the resource recovery potential of Danish BA: (i) based on historical and experimental data, all individual flows (representing different grain size fractions) within the recovery facility were quantified, (ii) the resource potential of ferrous (Fe) and non-ferrous (NFe) metals as well as rare earth elements (REE) was determined, (iii) recovery efficiencies were quantified for scrap metal and (iv) resource potential variability and recovery efficiencies were quantified based on a range of ashes from different incinerators. Recovery efficiencies for Fe and NFe reached 85% and 61%, respectively, with the resource potential of metals in BA before recovery being 7.2%ww for Fe and 2.2%ww for NFe. Considerable nonrecovered resource potential was found in fine fraction (below 2 mm), where approximately 12% of the total NFe potential in the BA were left. REEs were detected in the ashes, but the levels were two or three orders of magnitude lower than typical ore concentrations. The lack of REE enrichment in BAs indicated that the post-incineration recovery of these resources may not be a likely option with current technology. Based on these results, it is recommended to focus on limiting REE-containing products in waste for incineration and improving pre-incineration sorting initiatives for these elements. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Waste incineration has a prominent role in several countries in Northern Europe; for example, 88% of Danish domestic waste was incinerated in 2009 (DEPA, 2011). While modern waste incineration represents a robust technology for volume and mass reduction of the waste, combined with efficient energy recovery, increasing concern regarding the potential loss of valuable resources in waste
Abbreviations: Al, Aluminium; BA, Bottom ash; ECS, Eddy current separator; Fe, Ferrous metals; HNFe, Heavy non-ferrous metals; ISS, Inductive sorting system; MFA, Material flow analysis; MSWI, Municipal solid waste incineration; NFe, Nonferrous metals; PGM, Platinum group metals; REE, Rare earth elements; SFA, Substance flow analysis; SS, Stainless steel; STAN, Software for substance flow analysis; WEEE, Waste electrical and electronic equipment; WW, Wet weight; XSS, X-ray sorting system. ⇑ Corresponding author. Tel.: +45 45251572; fax: +45 45932850. E-mail address: [email protected] (E. Allegrini).
exists in society (European Commission 2008 and European Commission 2010; DEPA, 2013). The main material flow from waste incineration is bottom ash (BA), typically representing 15– 30% of the input waste mass (Hjelmar et al. 2010). BA contains two types of resources – metals and aggregates – the former of which may be present either as scrap metal (i.e. metals in metallic form, present as individual pieces or pieces which may subsequently be cleaned from aggregates) or as metals bound in mineral form within the aggregate matrix. While larger pieces of scrap metals, mainly magnetic metals, are removed from BA in most countries (e.g. Crillesen and Skaarup, 2006), the recovery of rarer or more valuable metals, or the utilisation of the aggregates, may not be common practice. This potentially represents a loss of resources and subsequently lowers the environmental performance of waste incineration. Detailed data documenting the presence of valuable resources in incineration BA are needed, in order to provide more accurate assessments of resource recovery from
http://dx.doi.org/10.1016/j.wasman.2014.05.003 0956-053X/Ó 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Allegrini, E., et al. Quantification of the resource recovery potential of municipal solid waste incineration bottom ashes. Waste Management (2014), http://dx.doi.org/10.1016/j.wasman.2014.05.003
2
E. Allegrini et al. / Waste Management xxx (2014) xxx–xxx
the waste incineration process and to determine possibilities for optimisation. Scrap metals are typically distinguished between ferrous (Fe) and non-ferrous (NFe) metals, based on their magnetic properties. Fe scrap is separated by magnets and consists mainly of steel and iron pieces, while NFe scrap is separated by eddy-current separators (ECSs) or inductive sorting systems (ISSs) and includes aluminium (Al), stainless steel (SS) and other conductive metals with higher atomic mass than Al (e.g. Cu, Zn and precious metals often termed ‘heavy non-ferrous metals’, HNFe). Economic interests and changes in raw materials availability have promoted interest over recent decades in the recovery of scrap metals; however, significant environmental savings have also supported this development (e.g. Damgaard et al., 2009). Based on the existing literature, typical ranges for Fe and NFe scrap content in BA can be found at 7–15% and 1–2%, respectively (Grosso et al., 2011). Current recovery efficiencies for Fe scrap are typically above 80%, while NFe recovery efficiencies vary considerably and are implemented to a much lesser extent (Heinrichs et al., 2012). Compared with Fe recovery, NFe recovery is a more sensitive activity and demands a far more complex recovery system than its Fe counterpart. For NFe recovery, ashes have to be divided into more homogeneous sub-streams (i.e. with narrow grain size ranges) before applying the ECS (e.g. European Commission, 2006). Consequently, the configuration of the recovery facility itself is critical for the efficiency of the overall process. Heinrichs et al. (2012) reported sorting efficiencies ranging from 29% to 75% for NFe metals at German state-of-the-art facilities. In the Netherlands, the development of new, advanced sorting systems allows the recovery of metals in the very fine fraction (down to 1 mm, De Vries et al., 2012), while in Amsterdam, for example, the use of wet sorting has increased recovery efficiencies by up to 73% for NFe metals (Rem et al., 2004; Muchova and Rem, 2006). Probably the highest recovery efficiency employed today is represented by the Swiss approach (above 90% for NFe), where metal recovery is performed on BA discharged from the incinerator furnace in dry form, i.e. avoiding wet quenching of the material (Büchi et al., 2013). However, dry BA is landfilled after metal recovery as opposed to wet quenched BA which can be used as construction material after metal recovery and weathering. This suggests that not only the recovery techniques themselves but also the ash generation processes prior to resource recovery may affect recovery efficiencies. The composition of waste input into the incineration process is also important for the presence and quality of resources in BA (e.g. Biganzoli et al., 2012; Hu et al., 2011; Astrup et al., 2011). Source segregation schemes (i.e. level of service, sorting guides for households, collection frequency, etc.) may potentially affect the composition of waste routed to incineration (e.g. presence of electronic equipment, type of packaging materials, etc.), thereby also affecting grain sizes and the abundance of metal resources present in the ashes. Based on material flow analysis (MFA), a recent study indicated that the contents of precious elements in the waste input into a Swiss incinerator (with dry BA discharge) were too low to consider any form of direct recovery action on waste prior to incineration (Morf et al., 2013). While recyclable resources, including metals, preferably should be sorted out from waste prior to incineration, valuable resources may still end up in the waste routed to incineration, due to misplacement or inefficient sorting schemes. For these resources, post-incineration recovery may therefore be the only realistic option available. While metal recovery from BA generally has the highest priority due to prices and environmental considerations, the remaining mineral fraction also has resource value, for example as a substitute for natural aggregates in construction works (e.g. Astrup, 2007) or as aggregate in concrete production (e.g. Van Der
Wegen et al., 2012). While the utilisation of mineral fraction is very high in some European countries (e.g. Denmark, Germany and the Netherlands, with utilisation rates up to 98% – as reported by Crillesen and Skaarup, 2006), other countries do not obtain the full benefits of this resource, which is instead landfilled. Improved sorting of metals from remaining mineral fraction generally improves the resource quality of aggregates for utilisation; for example, the presence of metallic Al may generate H2, thus causing loss of strength in any concrete containing affected aggregates (e.g. Pecqueur et al., 2001; Sorlini et al., 2011). At the same time, metals remaining in minerals upon landfill disposal or utilisation, for example in road construction, clearly represent a loss of resources. The overall aim of this paper is to provide a detailed account of metal resources in waste incineration BA, based on a full-scale facility in Denmark. The intention is thereby to offer a more consistent platform for future environmental assessments of incineration technologies and associated metal recovery. The specific objectives are: (i) a detailed quantification of resource flows within a BA recovery system, (ii) the quantification of recovery efficiencies and associated variability between ashes at the same recovery facility, (iii) the systematic determination of the resource potential (scrap metals, precious metals and aggregates) of ashes from a range of incinerators and (iv) an evaluation of important optimisation potential for improved recovery.
2. Material and methods 2.1. MSWI BA recovery facility The study was based on a Danish municipal solid waste incineration (MSWI) BA recovery facility located in the Copenhagen area. The system was managed by AFATEK Ltd., treating BA from six incinerators with various capacities and catchment areas, representing approximately 30% of the annual Danish production of BA. The system consisted of four subsequent sections (see Fig. 1. and Fig. A.1 in the Supplementary material), namely (i) Fe recovery, (ii) Fe upgrading, (iii) NFe recovery and (iv) NFe upgrading. In the Fe recovery section, raw (wet-quenched) BA delivered to the site from incinerators was screened in a 50 mm trommel and the two resulting streams passed separately through magnetic separators. Coarse, non-magnetic material was crushed and recirculated on top of the system, residue which could not be crushed was handsorted to separate coarse metal items for recycling and unburned materials were returned to the incineration plant. Non-magnetic material below 50 mm was stored in outdoor heaps for an average period of four months, during which time carbonation and weathering occurred. This storage period – known as ‘ageing’ – is necessary to obtain a material that complies with the Danish statutory order for use of residual materials in construction works (N. 1662:2010). Ageing happened before NFe recovery for logistic reasons (limited capacity of the NFe recovery section) and because of the resulting lower moisture content of the BA, which facilitated further processing. Magnetic material was upgraded into a second section, employed to attain a grade of steel suitable for secondary steel production. The Fe upgrading section consisted of a rotating drum where friction between the drum and the scrap metals detached impurities sticking to the metal which were then separated from the clean scrap through a 10 mm trommel screen. The scrap flow was then refined further through two magnetic separators in series. The rejected non-magnetic material from the two magnets was a mixture of heavy metallic items enriched in copper coils, which were then hand-sorted for copper recovery. After the ageing period, non-magnetic material below 50 mm was fed into the NFe recovery section. The material was first screened and then divided into four grain size fractions whose size
Please cite this article in press as: Allegrini, E., et al. Quantification of the resource recovery potential of municipal solid waste incineration bottom ashes. Waste Management (2014), http://dx.doi.org/10.1016/j.wasman.2014.05.003
E. Allegrini et al. / Waste Management xxx (2014) xxx–xxx
3
Fig. 1. MFA results: values are in Mg of material (e.g. MSWI BA, scrap metals) on a wet basis, and the uncertainty values represent the relative standard uncertainty. The Fe recovery, Fe upgrading and NFe recovery processes are defined by subsystems where all machinery included in each section is considered. See Fig. A.3 in Appendix A to see the subsystem schemes.
range was of approximately 0–2 mm, 2–8 mm, 8–16 mm and 16– 50 mm, based on actual load conditions. The fine fraction was not treated for metal recovery, while the remaining three sizes passed through three ECSs, and the 16–50 mm was fed additionally into an ISS. Non-conductive material was collected together with the fine material and residue from other sections, and then it was utilised as aggregate, substituting for virgin gravel in road sub-bases. Conductive material from ECS output was transported to an NFe upgrading section. This section was an industrial facility managed by the company Scanmetals Ltd., consisting of a complex sequence of screens – ECSs, ISSs, X-ray sorting systems (XSS) and sorting tables – aiming at increasing the scrap metal grade, separating light NFe metal (i.e. Al) from HNFe and separating clean scrap in three grain sizes of appropriate quality for secondary aluminium and copper production. 2.2. Sampling campaign Based on historical data on ash quantities and characteristics, taken from AFATEK Ltd., a sampling and measurement campaign was designed to fill as best as possible a number of data gaps regarding material flows within the recovery facility itself (i.e. the three first sections, excluding the NFe upgrading system). Nine measurement campaigns were performed over one year, each covering the
treatment of approximately 100 Mg of raw BA from various incinerators. During the campaigns, all relevant input and output flows were weighed and recorded by means of front loaders. Coarse materials output from the Fe recovery section were hand-sorted during each campaign to quantify the amount of unburned materials and metals as well as their type. Furthermore, heavy residue from the Fe upgrading system was hand-sorted to estimate the amount of copper coil. Numerous samples were collected during the measurement campaigns as well as during normal system operations from three main streams: (i) non-magnetic BA below 50 mm output from the Fe recovery section (41 samples), (ii) treated BA output from the NFe recovery section (14 samples) and (iii) fine BA below 2 mm bypassing the ECSs (23 samples). The sampling procedure followed the principles for representative sampling as stated in Danish regulation N. 1662:2010. A detailed overview of the sampling procedure and the number of samples collected is provided in Supplementary material (see Section A.1 in Appendix A). 2.3. Characterisation of BA samples BA samples obtained from the sampling campaign were characterised for the following parameters: (i) moisture content, following the EN 1097-5:2008 standard and (ii) metal scrap content, following a procedure developed at the Swiss Institute for Environ-
Please cite this article in press as: Allegrini, E., et al. Quantification of the resource recovery potential of municipal solid waste incineration bottom ashes. Waste Management (2014), http://dx.doi.org/10.1016/j.wasman.2014.05.003
4
E. Allegrini et al. / Waste Management xxx (2014) xxx–xxx
mental and Process Engineering, UMTEC (2012). The latter procedure was used to quantify the content of magnetic and non-magnetic scrap in BA by means of crushing, screening and handsorting material samples with minimum grain size of 0.5 mm. The same analyses were carried out on subsamples obtained by sieving the first two types of samples (non-magnetic BA below 50 mm and treated BA) in the three grain sizes 2–8 mm, 8– 16 mm and 16–50 mm, in order to simulate the three individual fractions treated in the NFe recovery system. Additional details regarding experimental procedures are provided in Supplementary material (see Section A.2.). 2.4. Analytical techniques for analysing precious and critical elements To assess the resource potential of precious and critical elements, a smaller number of samples (four samples of the treated BA and four of BA below 2 mm) were selected for further analysis of total content of a wide range of inorganic elements. The analysed elements included precious metals and elements listed by the European commission as critical (European Commission, 2010), with the exception of Y, which was used as an internal standard during the analyses, and Os and Pm, which were included in the internal standard solution used. For these analyses a 5 kg subsample from each sample was obtained by applying mass reduc-
tion techniques, as reported by Petersen et al. (2004). The samples were dried in an oven at 105 °C until a constant weight was achieved (about 24 h), and then the grain size was reduced in a jaw crusher, and after reducing the sample mass by means of a riffle-splitter, the material was pulverised by means of vibratory disc mill (Wolframe Carbide discs). Coarse metallic items which could not be size-reduced were removed from the sample material. A 0.2 g powder sample was digested by microwaveassisted digestion (Multiwave Anton Paar 3000), and then analysed by ICP-MS (7700x, Agilent Technologies). Every sample material was digested and analysed in duplicate. A single digestion method was applied to address a wide range of elements since the achieved detection limits for all elements of interest fitted the scope of the study. Samples were digested with HNO3, HCl, HF and H3BO3 following the standard EN 13656:2003 procedure, and the results were validated by means of three certified reference materials: fly ash (BCR-176R); sediments (GBW-07318); soil (NCS DC 78302). Additional information on analytical procedure is reported in A.2.2 in Appendix A. 2.5. MFA and SFA methodology Material flow analysis (MFA) and substance flow analysis (SFA) were carried out based on data obtained through the analysis of
Fig. 2. SFA results for NFe scrap: values are in Mg of material on a wet basis, and the uncertainty values represent the relative standard uncertainty. The Fe recovery, Fe upgrading and NFe recovery processes are defined by subsystems where all machinery included in each section is considered. See Fig. A.3 in Appendix A to see the subsystem schemes.
Please cite this article in press as: Allegrini, E., et al. Quantification of the resource recovery potential of municipal solid waste incineration bottom ashes. Waste Management (2014), http://dx.doi.org/10.1016/j.wasman.2014.05.003
E. Allegrini et al. / Waste Management xxx (2014) xxx–xxx
5
Fig. 3. SFA results for the Fe scrap: values are in Mg of material on a wet basis, and the uncertainty values represent the relative standard uncertainty. The Fe recovery, Fe upgrading and NFe recovery processes are defined by subsystems where all machinery included in each section is considered. See Fig. A.3 in Appendix A to see the subsystem schemes.
Table 1 MSWI BA metal potential (%ww: wet weight basis) and recovery efficiency. Uncertainties are shown as relative standard uncertainty. Scrap metal potential in raw MSWI BA Fe 7.2%ww ± 18% NFe 2.2%ww ± 2% Al 1.4%ww ± 2%
Cu
Scrap metal recovery efficiency: Danish state-of-the-art Fe 85% ± 28% NFe 61% ± 2.8% Al 62% ± 2.4% Cu
SFA was 5000 Mg of raw (wet) MSWI BA (referred to hereafter as BA batch), a choice based on the fact that most of the historical data were reported according to this unit, as required by Danish regulation N. 1662:2010.
0.24%ww ± 6%
3. Results and discussion 61% ± 11%
historical data and the measurement campaigns, as well as the analysis of samples obtained from the facility. The data were aggregated, elaborated and entered into the MFA software package STAN (Cencic and Rechberger, 2008). The analysis was carried out at two different layers: in the first layer, data about the wet mass of material circulating in the system were entered (MFA), while in the second one the concentration of a defined substance (Fe or NFe) was applied to each material flow (SFA). The calculation algorithm applied to the MFA and SFA was IAL-IMPL2013 (www.industrialgorithms.com), implemented in STAN. Additional information on input data and assumptions in the MFA and SFA is reported in Section A.3 of Appendix A. The functional unit used for the MFA and
3.1. Fe and NFe recovery potential and efficiencies Figs. 1–3 report the results of the MFA/SFA according to the main processes (see scheme in Fig. A.1), while additionally Table 1 summarises the main conclusions which can be drawn based on the analysis of the MFA/SFA figures. In Supplementary material, the SFA for Al and Cu scrap also is reported in Figs. A.4 and A.5. Potential Fe and NFe content in the raw BA was estimated to be 7.2%ww and 2.2%ww, respectively, while estimated recovery efficiencies were 85% for Fe and 61% for NFe. More specifically, raw BA consisted of 1.4%ww and 0.24%ww Al and Cu scrap, respectively, and the recoveries of these two types of metals were estimated to be 62% and 61%, respectively. Recovery efficiencies were estimated by considering the total amount of scrap sorted from the BA in various system steps and the total amount of the
Please cite this article in press as: Allegrini, E., et al. Quantification of the resource recovery potential of municipal solid waste incineration bottom ashes. Waste Management (2014), http://dx.doi.org/10.1016/j.wasman.2014.05.003
6
E. Allegrini et al. / Waste Management xxx (2014) xxx–xxx
Table 2 Content of scrap metals in MSWI BA. Results of metal scrap content analyses carried out on MSWI BA samples collected before and after the NFe recovery unit, and on individual grain size fractions. The results are reported as a percentage of wet weight (%ww), and thus moisture content values are also reported. Uncertainty is shown as relative standard uncertainty. Moisture content
NFe%ww
Fe%ww
MSWI BA before NFe recovery Totalb 12% ± 22% 2–8 mm 9.6% ± 19%a 8–16 mm 5.5% ± 13%a 16–50 mm 4.9% ± 0.63%a
2.1% ± 2.9% 3.1% ± 2.6% 3.5% ± 6.5% 3.4% ± 67%
0.83% ± 4.9% 0.63% ± 50% 0.56% ± 18% 0.99% ± 24%
MSWI BA after NFe recovery 12% ± 29% Totalb 2–8 mm 9.6% ± 19% 8–16 mm 5.5% ± 13% 16–50 mm 4.9% ± 0.63% 17% ± 13% Fine fractionc
0.82% ± 4.9% 0.80% ± 27% 0.87% ± 5.2% 0.64% ± 40% 0.75% ± 8%
0.51% ± 9.8% 0.49% ± 3.5% 0.75% ± 50% 1.0% ± 53% 0.39% ± 17%
a The moisture content of the BA fractions before NFe recovery was assumed to be identical to the one measured in the BA fractions after NFe recovery. b The metal content of the total sample was calculated including the value measured in the fine fraction and assuming that the fine fraction represents 37% of the MSWI BA before NFe recovery, and 40% of the MSWI BA after NFe recovery (assumption based on data on grain size distribution reported in Section A.2.3 in Appendix A). c The metal content in the fine fraction was calculated assuming that no scrap metals are present in the fraction below 0.5 mm and assigning the measured amount of scrap metals in the fraction 0.5–2 mm to the whole initial weight of the fine fraction.
same scrap metal estimated in the initial raw BA. Thus, untreated BA (i.e. fine ash) was included in determining recovery efficiency as a loss. The calculated efficiencies covered only the recovery of Fe and NFe within the BA recovery facility, so losses during the incineration process and during subsequent recycling processes (e.g. aluminium secondary smelter) were excluded from the system boundary of this study. Fe and NFe metal flows reported in Figs. 2 and 3 included impurities and forms of scrap metal with no value from a recycling perspective. For example, Al scrap flows following the ‘‘Al fractioning’’ section included the flow of metallic aluminium as well as aluminium oxide (a non-recyclable form of aluminium). According to personal communication with the secondary aluminium sector, Al losses after recycling Al incineration scrap can be estimated at around 33–35%, 27–30% and 18–20%, respectively, for the sizes below 5 mm, between 5 and 10 mm and above 10 mm. These losses are mainly due to Al oxidation. Table 2reports input data for the SFA resulting from scrap metal content analyses on BA before and after NFe recovery and on individual grain size fractions. The results are reported on a wet basis, and moisture content values are shown to allow recalculations and comparison with other studies. Data concerning individual grain size fractions were based on two measurements only, indicating that the values are less robust than those for the total and fine fractions, which were based on a large number of measurements (see Appendix A for details). Nevertheless, the results provide an indication of the distribution of NFe metals within BA and of the recovery efficiencies achieved for each BA grain size: recovery efficiency for each ECS – treating individual BA fractions (2–8, 8–16 and 16– 50 mm) – was above 70% (from 74% to 81%), increasing in line with the BA grain size.
The sorting efficiencies for Al and Cu were estimated on the basis of the respective SFAs (see Figs. A.4 and A.5), based on the results of the analyses carried out on NFe material sorted from the BA samples during the metal scrap content analyses (see Table 3). The light parts of the NFe, mainly Al, represented from 65% to 68% of NFe, which is similar to previous results reported in the literature (e.g. Mitterbauer et al., 2009). The HNFe fraction was between 25% and 32%, increasing in line with decreasing grain size, as also reported by other authors (e.g. Berkhout et al., 2011). In addition, the red part of the HNFe, i.e. copper and copper alloys, increased from the coarse to the fine fraction of the MSWI BA. This concentration of HNFe metals in fine MSWI BA is currently driving the increasing interest in recovering metals from these material sizes, especially below 2 mm. In fact, while the Al fraction is of less interest because of high oxidation levels (e.g. Biganzoli et al., 2013a), the heavy portion potentially has great economic value because of the high concentration of valuable metals such as Cu and precious metals (Biganzoli et al., 2013b; Muchova et al., 2009), which are not affected significantly by oxidation. The moisture content of the total MSWI BA before and after NFe recovery was approximately 12%, and the fine fraction contributed the most, with a moisture content of 17%. The high moisture content of fine fraction is one of the greatest challenges for metal recovery from this fraction (e.g. De Vries et al., 2012). The fine fraction represented approximately 37% of the MSWI BA before NFe recovery, thus, when considering the content of NFe in this fraction (0.76%ww), approximately 12% of NFe is lost in the fine fraction, if no recovery is carried out. 3.2. Recovery potential for critical elements The results of the chemical analyses performed on MSWI BA after NFe recovery (treated BA), and on its individual grain size fractions, are reported in Table 4. Out of a wide set of analysed elements, Table 4 reports a selection of chemical elements on the basis of their criticality from a resource and environmental perspective. As result of the digestion procedure used in this study the total concentration of rare earth elements (REEs) in the treated BA was approximately 110 mg kg 1, with the most abundant elements being La, Ce and Nd with concentrations over 20 mg kg 1. REE concentrations were found to be within ranges quoted in the literature (see Table 4). REE levels were of the same order of magnitude as their abundance in the Earth’s crust, but generally they were two or three orders of magnitude lower than typical levels in REE concentrated ores. Landfill mining as a source of REE is gaining increasing interest; however, waste residues of interest are industrial waste such as metallurgical slag, red mud, and mine tailings (Binnemans et al., 2013), which can reach higher REE concentrations (e.g. up to 5% of rare earth oxides) and be available in large volumes. Based on the results obtained with the applied digestion method, REE concentrations appeared too low in the BA samples to consider recovery; furthermore, applying hydrometallurgical processes to residues with large amounts of impure metals, such as in MSWI BA, is not suitable (Wang et al., 2011). Most of the REEs were found to be enriched in the coarser MSWI BA material; in particular, Sc, Ce, Sm, Dy, Ho, Tm, Yb and Lu
Table 3 Composition of NFe metals in individual MSWI BA samples based on sorting analyses. Values as a% of the NFe fraction. Uncertainty is shown as relative standard uncertainty. Light fraction (Al)
8–50 mm 2–8 mm 0.5–2 mm
68% ± 3.6% 65% ± 2.7% 65% ± 3.8%
HNFe
25% ± 9.8% 29% ± 4.9% 32% ± 7.8%
HNFe constituents Red metals (Cu)
Yellow metals (brass and others)
White metals (SS and others)
5.1% ± 14% 11% ± 9.1% 20% ± 8.5%
12% ± 14% 13% ± 8.8% 8.8% ± 14%
7.9% ± 15% 4.7% ± 15% 3.2% ± 27%
Please cite this article in press as: Allegrini, E., et al. Quantification of the resource recovery potential of municipal solid waste incineration bottom ashes. Waste Management (2014), http://dx.doi.org/10.1016/j.wasman.2014.05.003
7
E. Allegrini et al. / Waste Management xxx (2014) xxx–xxx Table 4 Total content of critical and precious metals and metals of typical environmental concern in treated MSWI BA and in individual BA fractions. Values reported in mg kg relative standard deviation, existing literature ranges and typical ore concentrations. Fractions of treated BA Treated BA
Literature dataa
Ore concentration
1
with
16–50 mm
8–16 mm
2–8 mm
0–2 mm
min–max
min–max
Ref
5.6 ± 7% 22 ± 7% 51 ± 7% 5.5 ± 10% 21 ± 26% 3.1 ± 5% 0.74 ± 6% 3.05 ± 8% 0.46 ± 4% 2.6 ± 2% 0.51 ± 6% 2.5 ± 15% 0.22 ± 8% 1.5 ± 7% 0.21 ± 7%
4.1 ± 5% 17 ± 11% 46 ± 3% 4.1 ± 7% 16 ± 8% 2.4 ± 12% 0.64 ± 3% 2.4 ± 10% 0.42 ± 5% 2.0 ± 7% 0.38 ± 13% 2.69 ± 23% 0.17 ± 13% 10% 0.17 ± 13%
4.0 ± 3% 21 ± 20% 43 ± 2% 5.05 ± 34% 21 ± 22% 2.2 ± 6% 0.64 ± 8% 2.6 ± 21% 0.48 ± 34% 2.0 ± 13% 0.34 ± 7% 2.3 ± 13% 0.15 ± 9% 1.0 ± 10% 0.14 ± 5%
4.0 ± 4% 24 ± 8% 42 ± 5% 6.7 ± 11% 27 ± 7% 2.4 ± 4% 0.66 ± 5% 2.9 ± 8% 0.43 ± 8% 2.35 ± 10% 0.37 ± 5% 1.7 ± 29% 0.15 ± 6% 1.25 ± 16% 0.15 ± 6%
1.3–22 (6) 2–30 (5) 11–51 (4) 1.1–10 (4) 4.0–37 (4) 0.93–5 (4) 0.25–2.6 (4) 0.88–5 (4) 0.18–3 (4) 0.54–3 (3) 0.11–0.45 (2) 0.31–2 (3) 0.01–0.18 (2) 0.31–5 (4) 0.02–0.23 (2)
20–130,000 2600–180,000 18,000–300,000 1600–33,000 6000–98,000 690–16,000 220–1100 650–21,000 43–5300 430–47,000 16–11,000 13–29,000 3–4900 3–34,000
Contents lists available at ScienceDirect
Waste Management journal homepage: www.elsevier.com/locate/wasman
Quantification of the resource recovery potential of municipal solid waste incineration bottom ashes Elisa Allegrini a,⇑, Alberto Maresca a, Mikael Emil Olsson a, Maria Sommer Holtze b, Alessio Boldrin a, Thomas Fruergaard Astrup a a b
Technical University of Denmark, Department of Environmental Engineering, Building 115, 2800 Lyngby, Denmark Afatek Ltd., Selinevej 18, 2300 Copenhagen S, Denmark
a r t i c l e
i n f o
Article history: Received 28 February 2014 Accepted 8 May 2014 Available online xxxx Keywords: MSWI Bottom ashes Non-ferrous metals Recovery efficiency REE Critical elements
a b s t r a c t Municipal solid waste incineration (MSWI) plays an important role in many European waste management systems. However, increasing focus on resource criticality has raised concern regarding the possible loss of critical resources through MSWI. The primary form of solid output from waste incinerators is bottom ashes (BAs), which also have important resource potential. Based on a full-scale Danish recovery facility, detailed material and substance flow analyses (MFA and SFA) were carried out, in order to characterise the resource recovery potential of Danish BA: (i) based on historical and experimental data, all individual flows (representing different grain size fractions) within the recovery facility were quantified, (ii) the resource potential of ferrous (Fe) and non-ferrous (NFe) metals as well as rare earth elements (REE) was determined, (iii) recovery efficiencies were quantified for scrap metal and (iv) resource potential variability and recovery efficiencies were quantified based on a range of ashes from different incinerators. Recovery efficiencies for Fe and NFe reached 85% and 61%, respectively, with the resource potential of metals in BA before recovery being 7.2%ww for Fe and 2.2%ww for NFe. Considerable nonrecovered resource potential was found in fine fraction (below 2 mm), where approximately 12% of the total NFe potential in the BA were left. REEs were detected in the ashes, but the levels were two or three orders of magnitude lower than typical ore concentrations. The lack of REE enrichment in BAs indicated that the post-incineration recovery of these resources may not be a likely option with current technology. Based on these results, it is recommended to focus on limiting REE-containing products in waste for incineration and improving pre-incineration sorting initiatives for these elements. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Waste incineration has a prominent role in several countries in Northern Europe; for example, 88% of Danish domestic waste was incinerated in 2009 (DEPA, 2011). While modern waste incineration represents a robust technology for volume and mass reduction of the waste, combined with efficient energy recovery, increasing concern regarding the potential loss of valuable resources in waste
Abbreviations: Al, Aluminium; BA, Bottom ash; ECS, Eddy current separator; Fe, Ferrous metals; HNFe, Heavy non-ferrous metals; ISS, Inductive sorting system; MFA, Material flow analysis; MSWI, Municipal solid waste incineration; NFe, Nonferrous metals; PGM, Platinum group metals; REE, Rare earth elements; SFA, Substance flow analysis; SS, Stainless steel; STAN, Software for substance flow analysis; WEEE, Waste electrical and electronic equipment; WW, Wet weight; XSS, X-ray sorting system. ⇑ Corresponding author. Tel.: +45 45251572; fax: +45 45932850. E-mail address: [email protected] (E. Allegrini).
exists in society (European Commission 2008 and European Commission 2010; DEPA, 2013). The main material flow from waste incineration is bottom ash (BA), typically representing 15– 30% of the input waste mass (Hjelmar et al. 2010). BA contains two types of resources – metals and aggregates – the former of which may be present either as scrap metal (i.e. metals in metallic form, present as individual pieces or pieces which may subsequently be cleaned from aggregates) or as metals bound in mineral form within the aggregate matrix. While larger pieces of scrap metals, mainly magnetic metals, are removed from BA in most countries (e.g. Crillesen and Skaarup, 2006), the recovery of rarer or more valuable metals, or the utilisation of the aggregates, may not be common practice. This potentially represents a loss of resources and subsequently lowers the environmental performance of waste incineration. Detailed data documenting the presence of valuable resources in incineration BA are needed, in order to provide more accurate assessments of resource recovery from
http://dx.doi.org/10.1016/j.wasman.2014.05.003 0956-053X/Ó 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Allegrini, E., et al. Quantification of the resource recovery potential of municipal solid waste incineration bottom ashes. Waste Management (2014), http://dx.doi.org/10.1016/j.wasman.2014.05.003
2
E. Allegrini et al. / Waste Management xxx (2014) xxx–xxx
the waste incineration process and to determine possibilities for optimisation. Scrap metals are typically distinguished between ferrous (Fe) and non-ferrous (NFe) metals, based on their magnetic properties. Fe scrap is separated by magnets and consists mainly of steel and iron pieces, while NFe scrap is separated by eddy-current separators (ECSs) or inductive sorting systems (ISSs) and includes aluminium (Al), stainless steel (SS) and other conductive metals with higher atomic mass than Al (e.g. Cu, Zn and precious metals often termed ‘heavy non-ferrous metals’, HNFe). Economic interests and changes in raw materials availability have promoted interest over recent decades in the recovery of scrap metals; however, significant environmental savings have also supported this development (e.g. Damgaard et al., 2009). Based on the existing literature, typical ranges for Fe and NFe scrap content in BA can be found at 7–15% and 1–2%, respectively (Grosso et al., 2011). Current recovery efficiencies for Fe scrap are typically above 80%, while NFe recovery efficiencies vary considerably and are implemented to a much lesser extent (Heinrichs et al., 2012). Compared with Fe recovery, NFe recovery is a more sensitive activity and demands a far more complex recovery system than its Fe counterpart. For NFe recovery, ashes have to be divided into more homogeneous sub-streams (i.e. with narrow grain size ranges) before applying the ECS (e.g. European Commission, 2006). Consequently, the configuration of the recovery facility itself is critical for the efficiency of the overall process. Heinrichs et al. (2012) reported sorting efficiencies ranging from 29% to 75% for NFe metals at German state-of-the-art facilities. In the Netherlands, the development of new, advanced sorting systems allows the recovery of metals in the very fine fraction (down to 1 mm, De Vries et al., 2012), while in Amsterdam, for example, the use of wet sorting has increased recovery efficiencies by up to 73% for NFe metals (Rem et al., 2004; Muchova and Rem, 2006). Probably the highest recovery efficiency employed today is represented by the Swiss approach (above 90% for NFe), where metal recovery is performed on BA discharged from the incinerator furnace in dry form, i.e. avoiding wet quenching of the material (Büchi et al., 2013). However, dry BA is landfilled after metal recovery as opposed to wet quenched BA which can be used as construction material after metal recovery and weathering. This suggests that not only the recovery techniques themselves but also the ash generation processes prior to resource recovery may affect recovery efficiencies. The composition of waste input into the incineration process is also important for the presence and quality of resources in BA (e.g. Biganzoli et al., 2012; Hu et al., 2011; Astrup et al., 2011). Source segregation schemes (i.e. level of service, sorting guides for households, collection frequency, etc.) may potentially affect the composition of waste routed to incineration (e.g. presence of electronic equipment, type of packaging materials, etc.), thereby also affecting grain sizes and the abundance of metal resources present in the ashes. Based on material flow analysis (MFA), a recent study indicated that the contents of precious elements in the waste input into a Swiss incinerator (with dry BA discharge) were too low to consider any form of direct recovery action on waste prior to incineration (Morf et al., 2013). While recyclable resources, including metals, preferably should be sorted out from waste prior to incineration, valuable resources may still end up in the waste routed to incineration, due to misplacement or inefficient sorting schemes. For these resources, post-incineration recovery may therefore be the only realistic option available. While metal recovery from BA generally has the highest priority due to prices and environmental considerations, the remaining mineral fraction also has resource value, for example as a substitute for natural aggregates in construction works (e.g. Astrup, 2007) or as aggregate in concrete production (e.g. Van Der
Wegen et al., 2012). While the utilisation of mineral fraction is very high in some European countries (e.g. Denmark, Germany and the Netherlands, with utilisation rates up to 98% – as reported by Crillesen and Skaarup, 2006), other countries do not obtain the full benefits of this resource, which is instead landfilled. Improved sorting of metals from remaining mineral fraction generally improves the resource quality of aggregates for utilisation; for example, the presence of metallic Al may generate H2, thus causing loss of strength in any concrete containing affected aggregates (e.g. Pecqueur et al., 2001; Sorlini et al., 2011). At the same time, metals remaining in minerals upon landfill disposal or utilisation, for example in road construction, clearly represent a loss of resources. The overall aim of this paper is to provide a detailed account of metal resources in waste incineration BA, based on a full-scale facility in Denmark. The intention is thereby to offer a more consistent platform for future environmental assessments of incineration technologies and associated metal recovery. The specific objectives are: (i) a detailed quantification of resource flows within a BA recovery system, (ii) the quantification of recovery efficiencies and associated variability between ashes at the same recovery facility, (iii) the systematic determination of the resource potential (scrap metals, precious metals and aggregates) of ashes from a range of incinerators and (iv) an evaluation of important optimisation potential for improved recovery.
2. Material and methods 2.1. MSWI BA recovery facility The study was based on a Danish municipal solid waste incineration (MSWI) BA recovery facility located in the Copenhagen area. The system was managed by AFATEK Ltd., treating BA from six incinerators with various capacities and catchment areas, representing approximately 30% of the annual Danish production of BA. The system consisted of four subsequent sections (see Fig. 1. and Fig. A.1 in the Supplementary material), namely (i) Fe recovery, (ii) Fe upgrading, (iii) NFe recovery and (iv) NFe upgrading. In the Fe recovery section, raw (wet-quenched) BA delivered to the site from incinerators was screened in a 50 mm trommel and the two resulting streams passed separately through magnetic separators. Coarse, non-magnetic material was crushed and recirculated on top of the system, residue which could not be crushed was handsorted to separate coarse metal items for recycling and unburned materials were returned to the incineration plant. Non-magnetic material below 50 mm was stored in outdoor heaps for an average period of four months, during which time carbonation and weathering occurred. This storage period – known as ‘ageing’ – is necessary to obtain a material that complies with the Danish statutory order for use of residual materials in construction works (N. 1662:2010). Ageing happened before NFe recovery for logistic reasons (limited capacity of the NFe recovery section) and because of the resulting lower moisture content of the BA, which facilitated further processing. Magnetic material was upgraded into a second section, employed to attain a grade of steel suitable for secondary steel production. The Fe upgrading section consisted of a rotating drum where friction between the drum and the scrap metals detached impurities sticking to the metal which were then separated from the clean scrap through a 10 mm trommel screen. The scrap flow was then refined further through two magnetic separators in series. The rejected non-magnetic material from the two magnets was a mixture of heavy metallic items enriched in copper coils, which were then hand-sorted for copper recovery. After the ageing period, non-magnetic material below 50 mm was fed into the NFe recovery section. The material was first screened and then divided into four grain size fractions whose size
Please cite this article in press as: Allegrini, E., et al. Quantification of the resource recovery potential of municipal solid waste incineration bottom ashes. Waste Management (2014), http://dx.doi.org/10.1016/j.wasman.2014.05.003
E. Allegrini et al. / Waste Management xxx (2014) xxx–xxx
3
Fig. 1. MFA results: values are in Mg of material (e.g. MSWI BA, scrap metals) on a wet basis, and the uncertainty values represent the relative standard uncertainty. The Fe recovery, Fe upgrading and NFe recovery processes are defined by subsystems where all machinery included in each section is considered. See Fig. A.3 in Appendix A to see the subsystem schemes.
range was of approximately 0–2 mm, 2–8 mm, 8–16 mm and 16– 50 mm, based on actual load conditions. The fine fraction was not treated for metal recovery, while the remaining three sizes passed through three ECSs, and the 16–50 mm was fed additionally into an ISS. Non-conductive material was collected together with the fine material and residue from other sections, and then it was utilised as aggregate, substituting for virgin gravel in road sub-bases. Conductive material from ECS output was transported to an NFe upgrading section. This section was an industrial facility managed by the company Scanmetals Ltd., consisting of a complex sequence of screens – ECSs, ISSs, X-ray sorting systems (XSS) and sorting tables – aiming at increasing the scrap metal grade, separating light NFe metal (i.e. Al) from HNFe and separating clean scrap in three grain sizes of appropriate quality for secondary aluminium and copper production. 2.2. Sampling campaign Based on historical data on ash quantities and characteristics, taken from AFATEK Ltd., a sampling and measurement campaign was designed to fill as best as possible a number of data gaps regarding material flows within the recovery facility itself (i.e. the three first sections, excluding the NFe upgrading system). Nine measurement campaigns were performed over one year, each covering the
treatment of approximately 100 Mg of raw BA from various incinerators. During the campaigns, all relevant input and output flows were weighed and recorded by means of front loaders. Coarse materials output from the Fe recovery section were hand-sorted during each campaign to quantify the amount of unburned materials and metals as well as their type. Furthermore, heavy residue from the Fe upgrading system was hand-sorted to estimate the amount of copper coil. Numerous samples were collected during the measurement campaigns as well as during normal system operations from three main streams: (i) non-magnetic BA below 50 mm output from the Fe recovery section (41 samples), (ii) treated BA output from the NFe recovery section (14 samples) and (iii) fine BA below 2 mm bypassing the ECSs (23 samples). The sampling procedure followed the principles for representative sampling as stated in Danish regulation N. 1662:2010. A detailed overview of the sampling procedure and the number of samples collected is provided in Supplementary material (see Section A.1 in Appendix A). 2.3. Characterisation of BA samples BA samples obtained from the sampling campaign were characterised for the following parameters: (i) moisture content, following the EN 1097-5:2008 standard and (ii) metal scrap content, following a procedure developed at the Swiss Institute for Environ-
Please cite this article in press as: Allegrini, E., et al. Quantification of the resource recovery potential of municipal solid waste incineration bottom ashes. Waste Management (2014), http://dx.doi.org/10.1016/j.wasman.2014.05.003
4
E. Allegrini et al. / Waste Management xxx (2014) xxx–xxx
mental and Process Engineering, UMTEC (2012). The latter procedure was used to quantify the content of magnetic and non-magnetic scrap in BA by means of crushing, screening and handsorting material samples with minimum grain size of 0.5 mm. The same analyses were carried out on subsamples obtained by sieving the first two types of samples (non-magnetic BA below 50 mm and treated BA) in the three grain sizes 2–8 mm, 8– 16 mm and 16–50 mm, in order to simulate the three individual fractions treated in the NFe recovery system. Additional details regarding experimental procedures are provided in Supplementary material (see Section A.2.). 2.4. Analytical techniques for analysing precious and critical elements To assess the resource potential of precious and critical elements, a smaller number of samples (four samples of the treated BA and four of BA below 2 mm) were selected for further analysis of total content of a wide range of inorganic elements. The analysed elements included precious metals and elements listed by the European commission as critical (European Commission, 2010), with the exception of Y, which was used as an internal standard during the analyses, and Os and Pm, which were included in the internal standard solution used. For these analyses a 5 kg subsample from each sample was obtained by applying mass reduc-
tion techniques, as reported by Petersen et al. (2004). The samples were dried in an oven at 105 °C until a constant weight was achieved (about 24 h), and then the grain size was reduced in a jaw crusher, and after reducing the sample mass by means of a riffle-splitter, the material was pulverised by means of vibratory disc mill (Wolframe Carbide discs). Coarse metallic items which could not be size-reduced were removed from the sample material. A 0.2 g powder sample was digested by microwaveassisted digestion (Multiwave Anton Paar 3000), and then analysed by ICP-MS (7700x, Agilent Technologies). Every sample material was digested and analysed in duplicate. A single digestion method was applied to address a wide range of elements since the achieved detection limits for all elements of interest fitted the scope of the study. Samples were digested with HNO3, HCl, HF and H3BO3 following the standard EN 13656:2003 procedure, and the results were validated by means of three certified reference materials: fly ash (BCR-176R); sediments (GBW-07318); soil (NCS DC 78302). Additional information on analytical procedure is reported in A.2.2 in Appendix A. 2.5. MFA and SFA methodology Material flow analysis (MFA) and substance flow analysis (SFA) were carried out based on data obtained through the analysis of
Fig. 2. SFA results for NFe scrap: values are in Mg of material on a wet basis, and the uncertainty values represent the relative standard uncertainty. The Fe recovery, Fe upgrading and NFe recovery processes are defined by subsystems where all machinery included in each section is considered. See Fig. A.3 in Appendix A to see the subsystem schemes.
Please cite this article in press as: Allegrini, E., et al. Quantification of the resource recovery potential of municipal solid waste incineration bottom ashes. Waste Management (2014), http://dx.doi.org/10.1016/j.wasman.2014.05.003
E. Allegrini et al. / Waste Management xxx (2014) xxx–xxx
5
Fig. 3. SFA results for the Fe scrap: values are in Mg of material on a wet basis, and the uncertainty values represent the relative standard uncertainty. The Fe recovery, Fe upgrading and NFe recovery processes are defined by subsystems where all machinery included in each section is considered. See Fig. A.3 in Appendix A to see the subsystem schemes.
Table 1 MSWI BA metal potential (%ww: wet weight basis) and recovery efficiency. Uncertainties are shown as relative standard uncertainty. Scrap metal potential in raw MSWI BA Fe 7.2%ww ± 18% NFe 2.2%ww ± 2% Al 1.4%ww ± 2%
Cu
Scrap metal recovery efficiency: Danish state-of-the-art Fe 85% ± 28% NFe 61% ± 2.8% Al 62% ± 2.4% Cu
SFA was 5000 Mg of raw (wet) MSWI BA (referred to hereafter as BA batch), a choice based on the fact that most of the historical data were reported according to this unit, as required by Danish regulation N. 1662:2010.
0.24%ww ± 6%
3. Results and discussion 61% ± 11%
historical data and the measurement campaigns, as well as the analysis of samples obtained from the facility. The data were aggregated, elaborated and entered into the MFA software package STAN (Cencic and Rechberger, 2008). The analysis was carried out at two different layers: in the first layer, data about the wet mass of material circulating in the system were entered (MFA), while in the second one the concentration of a defined substance (Fe or NFe) was applied to each material flow (SFA). The calculation algorithm applied to the MFA and SFA was IAL-IMPL2013 (www.industrialgorithms.com), implemented in STAN. Additional information on input data and assumptions in the MFA and SFA is reported in Section A.3 of Appendix A. The functional unit used for the MFA and
3.1. Fe and NFe recovery potential and efficiencies Figs. 1–3 report the results of the MFA/SFA according to the main processes (see scheme in Fig. A.1), while additionally Table 1 summarises the main conclusions which can be drawn based on the analysis of the MFA/SFA figures. In Supplementary material, the SFA for Al and Cu scrap also is reported in Figs. A.4 and A.5. Potential Fe and NFe content in the raw BA was estimated to be 7.2%ww and 2.2%ww, respectively, while estimated recovery efficiencies were 85% for Fe and 61% for NFe. More specifically, raw BA consisted of 1.4%ww and 0.24%ww Al and Cu scrap, respectively, and the recoveries of these two types of metals were estimated to be 62% and 61%, respectively. Recovery efficiencies were estimated by considering the total amount of scrap sorted from the BA in various system steps and the total amount of the
Please cite this article in press as: Allegrini, E., et al. Quantification of the resource recovery potential of municipal solid waste incineration bottom ashes. Waste Management (2014), http://dx.doi.org/10.1016/j.wasman.2014.05.003
6
E. Allegrini et al. / Waste Management xxx (2014) xxx–xxx
Table 2 Content of scrap metals in MSWI BA. Results of metal scrap content analyses carried out on MSWI BA samples collected before and after the NFe recovery unit, and on individual grain size fractions. The results are reported as a percentage of wet weight (%ww), and thus moisture content values are also reported. Uncertainty is shown as relative standard uncertainty. Moisture content
NFe%ww
Fe%ww
MSWI BA before NFe recovery Totalb 12% ± 22% 2–8 mm 9.6% ± 19%a 8–16 mm 5.5% ± 13%a 16–50 mm 4.9% ± 0.63%a
2.1% ± 2.9% 3.1% ± 2.6% 3.5% ± 6.5% 3.4% ± 67%
0.83% ± 4.9% 0.63% ± 50% 0.56% ± 18% 0.99% ± 24%
MSWI BA after NFe recovery 12% ± 29% Totalb 2–8 mm 9.6% ± 19% 8–16 mm 5.5% ± 13% 16–50 mm 4.9% ± 0.63% 17% ± 13% Fine fractionc
0.82% ± 4.9% 0.80% ± 27% 0.87% ± 5.2% 0.64% ± 40% 0.75% ± 8%
0.51% ± 9.8% 0.49% ± 3.5% 0.75% ± 50% 1.0% ± 53% 0.39% ± 17%
a The moisture content of the BA fractions before NFe recovery was assumed to be identical to the one measured in the BA fractions after NFe recovery. b The metal content of the total sample was calculated including the value measured in the fine fraction and assuming that the fine fraction represents 37% of the MSWI BA before NFe recovery, and 40% of the MSWI BA after NFe recovery (assumption based on data on grain size distribution reported in Section A.2.3 in Appendix A). c The metal content in the fine fraction was calculated assuming that no scrap metals are present in the fraction below 0.5 mm and assigning the measured amount of scrap metals in the fraction 0.5–2 mm to the whole initial weight of the fine fraction.
same scrap metal estimated in the initial raw BA. Thus, untreated BA (i.e. fine ash) was included in determining recovery efficiency as a loss. The calculated efficiencies covered only the recovery of Fe and NFe within the BA recovery facility, so losses during the incineration process and during subsequent recycling processes (e.g. aluminium secondary smelter) were excluded from the system boundary of this study. Fe and NFe metal flows reported in Figs. 2 and 3 included impurities and forms of scrap metal with no value from a recycling perspective. For example, Al scrap flows following the ‘‘Al fractioning’’ section included the flow of metallic aluminium as well as aluminium oxide (a non-recyclable form of aluminium). According to personal communication with the secondary aluminium sector, Al losses after recycling Al incineration scrap can be estimated at around 33–35%, 27–30% and 18–20%, respectively, for the sizes below 5 mm, between 5 and 10 mm and above 10 mm. These losses are mainly due to Al oxidation. Table 2reports input data for the SFA resulting from scrap metal content analyses on BA before and after NFe recovery and on individual grain size fractions. The results are reported on a wet basis, and moisture content values are shown to allow recalculations and comparison with other studies. Data concerning individual grain size fractions were based on two measurements only, indicating that the values are less robust than those for the total and fine fractions, which were based on a large number of measurements (see Appendix A for details). Nevertheless, the results provide an indication of the distribution of NFe metals within BA and of the recovery efficiencies achieved for each BA grain size: recovery efficiency for each ECS – treating individual BA fractions (2–8, 8–16 and 16– 50 mm) – was above 70% (from 74% to 81%), increasing in line with the BA grain size.
The sorting efficiencies for Al and Cu were estimated on the basis of the respective SFAs (see Figs. A.4 and A.5), based on the results of the analyses carried out on NFe material sorted from the BA samples during the metal scrap content analyses (see Table 3). The light parts of the NFe, mainly Al, represented from 65% to 68% of NFe, which is similar to previous results reported in the literature (e.g. Mitterbauer et al., 2009). The HNFe fraction was between 25% and 32%, increasing in line with decreasing grain size, as also reported by other authors (e.g. Berkhout et al., 2011). In addition, the red part of the HNFe, i.e. copper and copper alloys, increased from the coarse to the fine fraction of the MSWI BA. This concentration of HNFe metals in fine MSWI BA is currently driving the increasing interest in recovering metals from these material sizes, especially below 2 mm. In fact, while the Al fraction is of less interest because of high oxidation levels (e.g. Biganzoli et al., 2013a), the heavy portion potentially has great economic value because of the high concentration of valuable metals such as Cu and precious metals (Biganzoli et al., 2013b; Muchova et al., 2009), which are not affected significantly by oxidation. The moisture content of the total MSWI BA before and after NFe recovery was approximately 12%, and the fine fraction contributed the most, with a moisture content of 17%. The high moisture content of fine fraction is one of the greatest challenges for metal recovery from this fraction (e.g. De Vries et al., 2012). The fine fraction represented approximately 37% of the MSWI BA before NFe recovery, thus, when considering the content of NFe in this fraction (0.76%ww), approximately 12% of NFe is lost in the fine fraction, if no recovery is carried out. 3.2. Recovery potential for critical elements The results of the chemical analyses performed on MSWI BA after NFe recovery (treated BA), and on its individual grain size fractions, are reported in Table 4. Out of a wide set of analysed elements, Table 4 reports a selection of chemical elements on the basis of their criticality from a resource and environmental perspective. As result of the digestion procedure used in this study the total concentration of rare earth elements (REEs) in the treated BA was approximately 110 mg kg 1, with the most abundant elements being La, Ce and Nd with concentrations over 20 mg kg 1. REE concentrations were found to be within ranges quoted in the literature (see Table 4). REE levels were of the same order of magnitude as their abundance in the Earth’s crust, but generally they were two or three orders of magnitude lower than typical levels in REE concentrated ores. Landfill mining as a source of REE is gaining increasing interest; however, waste residues of interest are industrial waste such as metallurgical slag, red mud, and mine tailings (Binnemans et al., 2013), which can reach higher REE concentrations (e.g. up to 5% of rare earth oxides) and be available in large volumes. Based on the results obtained with the applied digestion method, REE concentrations appeared too low in the BA samples to consider recovery; furthermore, applying hydrometallurgical processes to residues with large amounts of impure metals, such as in MSWI BA, is not suitable (Wang et al., 2011). Most of the REEs were found to be enriched in the coarser MSWI BA material; in particular, Sc, Ce, Sm, Dy, Ho, Tm, Yb and Lu
Table 3 Composition of NFe metals in individual MSWI BA samples based on sorting analyses. Values as a% of the NFe fraction. Uncertainty is shown as relative standard uncertainty. Light fraction (Al)
8–50 mm 2–8 mm 0.5–2 mm
68% ± 3.6% 65% ± 2.7% 65% ± 3.8%
HNFe
25% ± 9.8% 29% ± 4.9% 32% ± 7.8%
HNFe constituents Red metals (Cu)
Yellow metals (brass and others)
White metals (SS and others)
5.1% ± 14% 11% ± 9.1% 20% ± 8.5%
12% ± 14% 13% ± 8.8% 8.8% ± 14%
7.9% ± 15% 4.7% ± 15% 3.2% ± 27%
Please cite this article in press as: Allegrini, E., et al. Quantification of the resource recovery potential of municipal solid waste incineration bottom ashes. Waste Management (2014), http://dx.doi.org/10.1016/j.wasman.2014.05.003
7
E. Allegrini et al. / Waste Management xxx (2014) xxx–xxx Table 4 Total content of critical and precious metals and metals of typical environmental concern in treated MSWI BA and in individual BA fractions. Values reported in mg kg relative standard deviation, existing literature ranges and typical ore concentrations. Fractions of treated BA Treated BA
Literature dataa
Ore concentration
1
with
16–50 mm
8–16 mm
2–8 mm
0–2 mm
min–max
min–max
Ref
5.6 ± 7% 22 ± 7% 51 ± 7% 5.5 ± 10% 21 ± 26% 3.1 ± 5% 0.74 ± 6% 3.05 ± 8% 0.46 ± 4% 2.6 ± 2% 0.51 ± 6% 2.5 ± 15% 0.22 ± 8% 1.5 ± 7% 0.21 ± 7%
4.1 ± 5% 17 ± 11% 46 ± 3% 4.1 ± 7% 16 ± 8% 2.4 ± 12% 0.64 ± 3% 2.4 ± 10% 0.42 ± 5% 2.0 ± 7% 0.38 ± 13% 2.69 ± 23% 0.17 ± 13% 10% 0.17 ± 13%
4.0 ± 3% 21 ± 20% 43 ± 2% 5.05 ± 34% 21 ± 22% 2.2 ± 6% 0.64 ± 8% 2.6 ± 21% 0.48 ± 34% 2.0 ± 13% 0.34 ± 7% 2.3 ± 13% 0.15 ± 9% 1.0 ± 10% 0.14 ± 5%
4.0 ± 4% 24 ± 8% 42 ± 5% 6.7 ± 11% 27 ± 7% 2.4 ± 4% 0.66 ± 5% 2.9 ± 8% 0.43 ± 8% 2.35 ± 10% 0.37 ± 5% 1.7 ± 29% 0.15 ± 6% 1.25 ± 16% 0.15 ± 6%
1.3–22 (6) 2–30 (5) 11–51 (4) 1.1–10 (4) 4.0–37 (4) 0.93–5 (4) 0.25–2.6 (4) 0.88–5 (4) 0.18–3 (4) 0.54–3 (3) 0.11–0.45 (2) 0.31–2 (3) 0.01–0.18 (2) 0.31–5 (4) 0.02–0.23 (2)
20–130,000 2600–180,000 18,000–300,000 1600–33,000 6000–98,000 690–16,000 220–1100 650–21,000 43–5300 430–47,000 16–11,000 13–29,000 3–4900 3–34,000
