588585
research-article2015
WMR0010.1177/0734242X15588585Waste Management & ResearchArends et al.
Original Article
Characterisation and materials flow management for waste electrical and electronic equipment plastics from German dismantling centres
Waste Management & Research 1–10 © The Author(s) 2015 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0734242X15588585 wmr.sagepub.com
Dagmar Arends1, Martin Schlummer1, Andreas Mäurer1, Jens Markowski2 and Udo Wagenknecht3
Abstract Waste electrical and electronic equipment is a complex waste stream and treatment options that work for one waste category or product may not be appropriate for others. A comprehensive case study has been performed for plastic-rich fractions that are treated in German dismantling centres. Plastics from TVs, monitors and printers and small household appliances have been characterised extensively. Based on the characterisation results, state-of-the-art treatment technologies have been combined to design an optimised recycling and upgrade process for each input fraction. High-impact polystyrene from TV casings that complies with the European directive on the restriction of hazardous substances (RoHS) was produced by applying continuous density separation with yields of about 60%. Valuable acrylonitrile butadiene styrene/polycarbonate can be extracted from monitor and printer casings by nearinfrared-based sorting. Polyolefins and/or a halogen-free fraction of mixed styrenics can be sorted out by density separation from monitors and printers and small household appliances. Emerging separation technologies are discussed to improve recycling results. Keywords Waste electrical and electronic equipment plastics, dismantling, characterisation, recycling, separation, density, near-infrared
Introduction Desirable properties as strength, flexibility, low costs and durability allow the use of polymers for a wide range of applications (Thompson et al., 2009a). One of the growing segments with an increasing demand for plastics is the sector for electrical and electronic equipment (EEE) (Wang and Xu, 2014). However, the steadily growing polymer production results in an increasing amount of plastic waste that has to be treated (Al Salem et al., 2009). According to official European statistics, 3.6m t of waste electrical and electronic equipment (WEEE) were collected in 29 European Union (EU) member states in 2010 (Eurostat, 2014). Based on an estimated overall plastic content of 22%, there were 792,000 t of plastics in European WEEE. Of the total WEEE plastics, 85% are contained in three product categories: information and communication technology (ICT), large household appliances and consumer electronics (Delgado et al., 2007). WEEE resembles a very complex waste stream (Dimitrakakis et al., 2009). Nevertheless, for 2012, the EU reported a recycling and reuse rate of 82% for WEEE in official EU statistics (Eurostat, 2014); however, this rate is mainly achieved by metal recycling. In contrast, recycling of WEEE plastics is rare, despite of the high resource potential mentioned above. In Germany, only about 10% of the WEEE plastics were recycled in 2013 (Consultic, 2014).
High-impact polystyrene (HIPS), acrylonitrile butadiene styrene (ABS), blends of ABS and polycarbonate (ABS-PC) and polypropylene (PP) are the dominating plastic types in WEEE (Wang and Xu, 2014) and the most promising target fractions or plastics recycling. Every plastic type is available in innumerable grades owing to their wide range of applications (Chancerel and Rotter, 2009; Thompson et al., 2009b). This material diversity hampers an effective plastics recycling. Furthermore, the waste stream contains different metals and noteworthy concentrations of hazardous substances, which hinder marketability of recycled plastics (Mark, 2006; Menad et al., 2013; Santos et al., 2011).
1Department
of Recycling Polymers, Fraunhofer Institute IVV, Freising, Germany 2Brandenburg University of Technology Cottbus-Senftenberg, Cottbus, Germany 3Processing Department, Leibniz Institute of Polymer Research Dresden, Dresden, Germany Corresponding author: Martin Schlummer, Department of Recycling Polymers, Fraunhofer Institute IVV, Giggenhauser Strasse 35, 85354 Freising, Germany. [email protected]
Downloaded from wmr.sagepub.com at University of Otago Library on July 4, 2015
2
Waste Management & Research
Regarding plastics recycling, several regulations have to be considered. According to the European Directive 2012/19/EU, WEEE products have to be collected in different categories. Furthermore, certain substances, mixtures and components have to be removed from the WEEE stream. The directive also dictates recovery as well as recycling rates for the different product categories (Directive 2012/19/EU, 2012). Since the waste composition strongly depends on its origin, it does make sense to keep waste streams from different sectors separated (Gent et al., 2011). Flame retardants, for example, are mainly used in ICT devices (Wang and Xu, 2014). The European Directive 2011/65/EC (RoHS 2) restricts the maximum allowable concentrations to 0.1% for lead, mercury, chromium (VI), polybrominated biphenyls (PBB) and polybrominated diphenyl ethers (PBDE) and to 0.01% for cadmium in new EEE products (Directive 2011/65/EU, 2011). Directive 2003/11/EC (2003) sets the same thresholds for pentabromodiphenyl ether (PentaBDE) and octabromodiphenyl ether (OctaBDE) for all products placed on the market. According to Wäger et al. (2012), contents of cadmium and PBDE are likely to exceed European thresholds for ABS, HIPS, PC/ABS and PP from WEEE. A considerable amount of WEEE containing bromine or heavy metals is exported to Asian and African countries instead of being treated properly. High bromine levels indicate that significant numbers of items do not fulfil the requirements for PBDE of the above mentioned directions (Sindiku et al., 2014). Legal requirements and the incompatibility of many plastic types necessitate an effective separation to produce high-quality recyclates (Dodbiba et al., 2005). Depending on the limitations of a specific sorting technology, certain input compositions are necessary with regard to material diversity, colour, additives and so on (Schlummer, 2014). Thus, many mechanical separation processes require an upstream concentration of the target polymer to guarantee an efficient recycling process (Gent et al., 2011). For most separation technologies, a pre-treatment-like crushing or pre-classification (e.g. by size or type) of the waste stream is necessary prior to sorting in order to achieve maximum yield and selectivity. Mark describes shredding and manual dismantling as the two possible treatments to produce plastic-rich streams from WEEE (Mark, 2006). The European Directive 2012/19/EU (2012) stipulates the separation of certain product parts. Therefore, in Germany, TVs, monitors, printers and some small household appliances (SHA) like vacuum cleaners and coffee machines are sorted out and/or dismantled manually to remove panel glass, cartridges and motors. The outcome of manual disassembly is a rather large size of plastic parts. Collecting these parts in separate batches results in a pre-concentration of plastic types and grades. Manual dismantling can be combined with hand-held spectroscopic devices like near-infrared (NIR) spectroscopy and/or X-ray transmission (XRT) spectroscopy. Besides manual separation, density separation, NIR spectroscopy and electrostatic separation can be considered as state-of-the-art recycling processes for WEEE (Krämer et al., 2009; Schlummer et al., 2010). NIR-based sorting is a widespread dry separation method for post-consumer plastics, especially for streams of packaging
waste (Hopewell et al., 2009). The waste stream is exposed to NIR illumination and a sensor or camera system detects the material-specific reflectance or transmission spectrum to compare it with a certain reference spectrum (Massen, 1998; van den Broek et al., 1998). Effective singularisation of particles and a certain range of particle sizes enhance efficient blowing out. Dark particles normally contain carbon black, which absorbs the NIR radiation to a large extent so that no reflected radiation can reach the sensors (Huth-Fehre et al., 1995). Depending on software and hardware specifications, a certain particle size range is preferable. High process yields with NIR sorting can only be achieved for lightly coloured inputs, with high contents of the target polymer(s) with a desirable particle size. For sorting black coloured plastic particles laser-based systems can be used (e.g. with laser induced breakdown spectroscopy – LIBS), however they are still under development and not ready for commercial use. Density separation is the most commonly applied recycling process for post-consumer plastics. It is mostly run as a wet process. This technology can be applied to separate polymer types, polymers from other material and different polymer grades of one and the same type (Mäurer and Schlummer, 2006). Pure homogeneous target fractions cannot be achieved for polymer mixtures with overlapping density ranges (Schlummer, 2014). Mäurer and Schlummer apply density separation to sort halogenated styrenics from halogen-free ones (Mäurer and Schlummer, 2006). In terms of throughput and/or purity, centrifugal systems are more effective than static ones. Nevertheless, static systems are commonly applied industrially (Gent et al., 2011). Waste streams with high plastic contents are obtained by the inevitable manual dismantling of certain waste electronic devices, such as TV sets, monitors or vacuum cleaners (Schlummer et al., 2007). The composition of such plastic fractions is limited to polymer types and grades that are used for these specific devices. In this article, a case study is described in which WEEE fractions from German dismantling centres were characterised, sorted and recompounded by a consortium of German dismantlers, waste recyclers and research centres. The approach included an intelligent materials flow management, taking advantage of the manual treatment. Properly input-tailored recycling processes were deduced from the input composition of different WEEE streams to produce high-quality and contaminant-free recyclates with high recovery rates. Depending on the input, combinations of handheld material identification, crushing, automated NIR and continuous density separation were tested. Analytical and mechanical product tests were performed to evaluate the products’ qualities and the efficiency of the developed process cascades.
Materials and methods Sampling The input material was provided by five German dismantling centres. Casings from monitors and printers (M&P), TVs and SHA and components of these WEEE fractions were examined.
Downloaded from wmr.sagepub.com at University of Otago Library on July 4, 2015
3
Arends et al. Table 1. Amounts used of each fraction. WEEE product
Description
Labelling
Amounts
TV casings
Unsorted Halogenated Halogen-free Die-cut fronts Die-cut backs Die-cut samples unsorted Unsorted Halogenated Halogen-free
TV TV-X TV-noX TVfronts TVbacks TVunsorted M&P M&P-X M&PnoX Mfeet Mbacks Mfronts SHA SHAcoffee SHAvacuum
550 kg and 550 kg 550 kg 550 kg 100 parts 100 parts 100 parts 500 kg and 550 kg 350 kg 550 kg
Monitor and printer casings
Small household appliances
Monitor feet Monitor backs Monitor fronts Unsorted casings Coffee machines Vacuum cleaner casings
73 parts and 30 kg scrap 251 parts and 30 kg scrap 260 parts 400 kg and 200 kg and 15 kg Sorted from 500 kg SHA Sorted from 500 kg SHA
M&P: monitor and printer; SHA: small household appliances.
The amounts that were used of each fraction are presented in Table 1.
Casings of monitors and printers Monitor casings were pre-sorted manually into their device components: monitor fronts (Mfronts), backs (Mbacks) and feet (Mfeet) for material characterisation. For material separation three dismantling centres provided monitor and printer (M&P) casings. Two dismantlers supplied 500 kg and 550 kg, respectively. One dismantler provided 350 kg of halogenated (M&P-X) and 550 kg of halogen-free (M&P-noX) casings, sorted out by sliding spark spectroscopy.
TV casings For material characterisation, die-cut samples from pre-sorted TV fronts (TVfronts), backs (TVbacks) and unsorted (TVunsorted) TV casings were provided by one dismantling centre. For material separation three dismantling centres provided TV casings. One dismantler pre-sorted 550 kg of halogenated (TV-X) and 550 kg of halogen-free (TV-noX) casings by sliding spark spectroscopy. The other two dismantlers provided 550 kg each of unsorted TV casings.
Small household appliances Two dismantlers supplied batches of 200 kg and 400 kg of mixed SHA casings. Another focus was put on: (a) non-dismantled coffee machines and coffee dispensers (SHAcoffee); and (b) dismantled casings of vacuum cleaners (SHAvacuum) that were separated from a batch of 500 kg of mixed SHA by hand sorting.
Material characterisation and analysis Input and output material streams were characterised by a number of laboratory methods. X-ray fluorescence (XRF) spectroscopy
with the Spectro X-Lab 2000 (Spectro Analytical Instruments GmbH & Co.KG, Kleve, Germany) was applied to determine element contents of bromine and chlorine, and the RoHS listed heavy metals lead, cadmium, chromium and mercury. Energy dispersive x-ray flourescence spectroscopy (ED-XRF) is recognised as an analytical approach with a precision of ±20% if applied in screening mode, which is considered to be sufficient for the assessment of whether a plastic part contains more than 0.1% of bromine, mercury, chromium and lead, or more than 0.01% of cadmium (Sindiku et al., 2014). For quality assurance, two polymeric-certified reference materials (polyethylene, EC 681, IRMM, Geel, Belgium, and polyvinyl chloride, PLPVC-3, Breitlaender, Wesel, Germany) covering a concentration range from 0.001 to 0.1% were analysed on each day of the analytical screening procedure. RoHS compliance was confirmed for a sample if the measured levels were safely below the threshold levels of 0.1 (PBB, PBDE, chromium, mercury, lead) or 0.01% (cadmium), respectively. PBDE and PBB levels were certified on the basis of bromine. Decabromodiphenyl ether (DecaBDE), OctaBDE and Decabromobiphenyl (DecaBB) contain 83%, 79% or 85% of bromine, respectively, indicating a level of 0.1% of each flame retardant at 0.083%, 0.079% or 0.085% of bromine. Considering an accuracy of ±20% of ED-XRF, concentrations were considered to be safely below the thresholds if they were below 80% of these levels. To determine the contents of bromine and chlorine at the dismantling centres, the sliding spark spectrometer SSS3 (Iosys, Ratingen, Germany) was used. To determine polymer types, an FTIR instrument (Spectrum One, ATR, golden gate, Perkin Elmer, Waltham, MA, USA) and the handheld NIR device Kusta 4004 (LLA Instruments GmbH, Berlin, Germany) were applied. Particle size classification with screens of 12 mm and 45 mm were performed for the output fractions of the crushing process to determine the particle size distribution. To characterise the big batches of M&P casings, TV casings and SHA static density fractionation was performed on a laboratory
Downloaded from wmr.sagepub.com at University of Otago Library on July 4, 2015
4
Waste Management & Research
Figure 1. Treatment scheme for the WEEE input fractions monitors and printers (M&P), SHA and TVs, comprising manual dismantling, hand-sorting, handheld spectroscopic devices, crushing, NIR or continuous density separation and combinations thereof. ABS: acrylonitrile butadiene styrene; Br: bromine; M&P: monitors and printers; PC/ABS: polycarbonate/acrylonitrile butadiene styrene; PS: polystyrene; SHA: small household appliances.
scale with samples from the coarse particle size reduction. Dipotassium phosphate was dissolved in water to create a density media of 1.0 g cm-3, 1.03 g cm-3, 1.06 g cm-3, 1.07 g cm-3, 1.08 g cm-3 and 1.09 g cm-3. In the first step, the lightest density fraction was separated from 10 kg of each batch, the material was dried and the mass balances were recorded. The bromine content of the light fraction was measured by XRF. The heavy fraction of this first step was then introduced to the density medium of 1.03 g cm-3, the material was dried, mass balances were recorded and the bromine content of the lighter fraction was measured. This procedure was repeated with media of higher densities until the bromine value exceeded 0.1% to determine which density cuts are expedient on a technical scale to produce legally marketable recyclates with high process yields.
Production of test specimen for mechanical product evaluation To determine tensile and impact strength of recycled styrenics and polyolefins, test specimen were produced. The recyclates were homogenised and cleaned in a single-screw extruder with a screen of 150 µm. The extrusion temperatures varied between 180 °C and 230 °C and pressures up to 200 bar were applied. The
extruded products were granulated and test specimen were produced by injection moulding at 240 °C for ABS, polystyrene (PS) and ABS-PC, and 190 °C for PP. The injection moulding dies had temperatures between 45 °C and 60 °C. Tensile and impact (Charpy method) tests were performed in accordance to standardised methods (DIN EN ISO 3167, DIN EN ISO 179).
Technical scale separation Manual dismantling and halogen detection. Dismantling of TVs, monitors, printers and vacuum cleaners was performed manually to remove the panel glass, cartridges and motors in German dismantling centres. Coffee dispensers and vacuum cleaners were separated by hand sorting from batches of SHA. One dismantler applied sliding spark spectroscopy to identify halogens in casings of TVs and M&P, and to separate halogenated parts from halogen-free ones. Material sorting. For material sorting, manual dismantling, hand-sorting, handheld spectroscopic devices, crushing, automated NIR or continuous density separation and combinations of these techniques were applied. Figure 1 depicts the treatment scheme for the different input streams. Two basic approaches were chosen for material sorting. M&P casings were sorted by NIR separation and the particle size of the
Downloaded from wmr.sagepub.com at University of Otago Library on July 4, 2015
5
Arends et al.
Results and discussion Material characterisation
Figure 2. Polymer type composition of monitor and TV components and mixed SHA.
ABS: acrylonitrile butadiene styrene; PC/ABS: polycarbonate/acrylonitrile butadiene styrene; PS: polystyrene.
sorted ABS and PS fractions was further reduced to separate halogenated parts by density separation. The NIR-based separation approach was also tested for SHA casings, but was not pursued owing to insufficient process efficiency. Because of the high content of dark parts, TV casings, mixed SHA fractions, SHAcoffee and SHAvacuum were directly introduced to the density separation process. Size reduction and separation. The separated casings were pressed and shipped to a business partner who crushed the material in a rotary shear and granulator (both Andritz-MEWA GmbH, Gechingen, Germany). Fine grains were separated by a zigzag air-classifier (AUT GmbH, Chemnitz, Germany). Iron was extracted by an overhead suspension magnet and a neodymium magnet (both Steinert Elektromagnetebau, Cologne, Germany), and non-ferrous metals were removed by an eddy current separator. Batches for the downstream continuous density separation were introduced to a hammer mill (Herbold, Meckesheim, Germany) for crushing to particle sizes below 8 mm. Automated NIR separation was tested on a technical scale with a UNISORT® NIR separation system (RTT-Steinert GmbH, Zittau, Germany). The first sorting step was performed with a multiplex sensor. The remaining fraction of parts that could not be identified was sorted once more by applying a camera system with an increased sensor resolution. Technical scale NIR separation was performed for M&P casings and for a batch of mixed SHA. On a technical scale, continuous density separation was performed with the Sorticanter® (Flottweg, Vilsbiburg, Germany). The density media were produced by dissolving dipotassium phosphate in water to produce a density media of 1.0 g cm-3, 1.06 g cm-3 and 1.08 g cm-3. Which density cuts were chosen for which fraction was decided according to the results of the material characterisation. An input of 100 kg was the minimum amount for the Sorticanter® to obtain separation results with repetitious accuracy. Continuous density separation was applied to sort halogen-free PS fractions from TV casings and halogen-free ABS and PS fractions of NIR pre-sorted M&P casings. From a batch of mixed SHA and from the sub-fractions SHAcoffee and SHAvacuum, polyolefins and halogen-free styrenics were separated.
Handheld spectroscopic polymer-type identification was used to get an idea of the polymer-type composition of monitor and TV components and mixed SHA. Figure 2 depicts the results. Figure 2 reveals a high ABS-PC and ABS content for all monitor components. PS was mainly used for Mfeet. In addition to the polymer type identification, Mfeet, Mfronts and Mbacks of 908 monitor casings were characterised by handheld sliding spark spectroscopy to identify the presence of halogens. A total of 29% of Mfronts contained chlorine and/or bromine and 28% of Mbacks, respectively. Less halogens were found in Mfeet, only 10% of the parts contained bromine and/or chlorine. The XRF results of the scrap material correlate with these results. For Mbacks, a total bromine content of 2.4% was measured and 0.15% for the feet. The results of the 100 unsorted TV samples shown in Figure 2 reveal that 86 specimens were composed of HIPS, nine of ABS and five of other polymer types. The XRF identified bromine contents above 0.1% in 36% of the HIPS samples. The chlorine content exceeded 0.5%, even though the bromine level was below 0.1% for another 8% of the HIPS samples. A cadmium content of 0.01% was exceeded for 19% of the samples, but only if either the bromine or chlorine content was too high as well. The bromine content of all ABS samples exceeds 0.1% and the legal threshold for cadmium is exceeded for most ABS samples. Manufacturers could be assigned to all TV parts analysed and ABS was almost exclusively used by a certain manufacturer. From these results, a halogen-free HIPS target fraction of 50% of the total TV fraction could be expected. Since no other halogenfree styrenics could be found, density separation appears to be a promising treatment to produce a pure halogen-free HIPS recyclate. The polymer type composition of TVfronts and TVbacks is similar. Comparing the XRF results of TVbacks and TVfronts reveals that 37% of the backs exceeded bromine levels of 0.1% and only 6% of the fronts exceeded this value. A higher process yield could be achieved if only TVfronts were used, but since TVbacks make up a bigger fraction of the WEEE stream, treating all TV components is recommended. Figure 2 depicts a rather high variety of materials in a mixed batch of SHA. Contents between 16% and 22% were identified for ABS, PS and polyolefins. Polyamide and polycarbonate could also be identified and a content of 33% of other plastics, metals and other material. The results of the static laboratory density fractionation are illustrated by Figure 3. For M&P casings, the bromine content stays below 0.1% for material densities of 1.08 g cm-3 and below, but the yield is only approximately 22% for both batches analysed. For the two batches of unsorted TV casings, legal thresholds for bromine can be kept when a density cut of 1.06 g cm-3 is applied. At this density cut, the yield is 69% and 77%. While the density fractions of both dismantlers show similar compositions for TV casings as well as for M&P casings, the two batches of SHA differ remarkably. For SHA 1 the bromine
Downloaded from wmr.sagepub.com at University of Otago Library on July 4, 2015
6
Waste Management & Research
Figure 3. Yield and bromine contents of density fractions. SHA: small household appliances; M&P: monitors and printers.
content stays below 0.1%, even in higher density fractions. Anyhow, thresholds for lead and cadmium are exceeded in this batch. The bromine content of batch SHA 2 exceeds 0.2% in the density fraction 1.07 g cm-3
research-article2015
WMR0010.1177/0734242X15588585Waste Management & ResearchArends et al.
Original Article
Characterisation and materials flow management for waste electrical and electronic equipment plastics from German dismantling centres
Waste Management & Research 1–10 © The Author(s) 2015 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0734242X15588585 wmr.sagepub.com
Dagmar Arends1, Martin Schlummer1, Andreas Mäurer1, Jens Markowski2 and Udo Wagenknecht3
Abstract Waste electrical and electronic equipment is a complex waste stream and treatment options that work for one waste category or product may not be appropriate for others. A comprehensive case study has been performed for plastic-rich fractions that are treated in German dismantling centres. Plastics from TVs, monitors and printers and small household appliances have been characterised extensively. Based on the characterisation results, state-of-the-art treatment technologies have been combined to design an optimised recycling and upgrade process for each input fraction. High-impact polystyrene from TV casings that complies with the European directive on the restriction of hazardous substances (RoHS) was produced by applying continuous density separation with yields of about 60%. Valuable acrylonitrile butadiene styrene/polycarbonate can be extracted from monitor and printer casings by nearinfrared-based sorting. Polyolefins and/or a halogen-free fraction of mixed styrenics can be sorted out by density separation from monitors and printers and small household appliances. Emerging separation technologies are discussed to improve recycling results. Keywords Waste electrical and electronic equipment plastics, dismantling, characterisation, recycling, separation, density, near-infrared
Introduction Desirable properties as strength, flexibility, low costs and durability allow the use of polymers for a wide range of applications (Thompson et al., 2009a). One of the growing segments with an increasing demand for plastics is the sector for electrical and electronic equipment (EEE) (Wang and Xu, 2014). However, the steadily growing polymer production results in an increasing amount of plastic waste that has to be treated (Al Salem et al., 2009). According to official European statistics, 3.6m t of waste electrical and electronic equipment (WEEE) were collected in 29 European Union (EU) member states in 2010 (Eurostat, 2014). Based on an estimated overall plastic content of 22%, there were 792,000 t of plastics in European WEEE. Of the total WEEE plastics, 85% are contained in three product categories: information and communication technology (ICT), large household appliances and consumer electronics (Delgado et al., 2007). WEEE resembles a very complex waste stream (Dimitrakakis et al., 2009). Nevertheless, for 2012, the EU reported a recycling and reuse rate of 82% for WEEE in official EU statistics (Eurostat, 2014); however, this rate is mainly achieved by metal recycling. In contrast, recycling of WEEE plastics is rare, despite of the high resource potential mentioned above. In Germany, only about 10% of the WEEE plastics were recycled in 2013 (Consultic, 2014).
High-impact polystyrene (HIPS), acrylonitrile butadiene styrene (ABS), blends of ABS and polycarbonate (ABS-PC) and polypropylene (PP) are the dominating plastic types in WEEE (Wang and Xu, 2014) and the most promising target fractions or plastics recycling. Every plastic type is available in innumerable grades owing to their wide range of applications (Chancerel and Rotter, 2009; Thompson et al., 2009b). This material diversity hampers an effective plastics recycling. Furthermore, the waste stream contains different metals and noteworthy concentrations of hazardous substances, which hinder marketability of recycled plastics (Mark, 2006; Menad et al., 2013; Santos et al., 2011).
1Department
of Recycling Polymers, Fraunhofer Institute IVV, Freising, Germany 2Brandenburg University of Technology Cottbus-Senftenberg, Cottbus, Germany 3Processing Department, Leibniz Institute of Polymer Research Dresden, Dresden, Germany Corresponding author: Martin Schlummer, Department of Recycling Polymers, Fraunhofer Institute IVV, Giggenhauser Strasse 35, 85354 Freising, Germany. [email protected]
Downloaded from wmr.sagepub.com at University of Otago Library on July 4, 2015
2
Waste Management & Research
Regarding plastics recycling, several regulations have to be considered. According to the European Directive 2012/19/EU, WEEE products have to be collected in different categories. Furthermore, certain substances, mixtures and components have to be removed from the WEEE stream. The directive also dictates recovery as well as recycling rates for the different product categories (Directive 2012/19/EU, 2012). Since the waste composition strongly depends on its origin, it does make sense to keep waste streams from different sectors separated (Gent et al., 2011). Flame retardants, for example, are mainly used in ICT devices (Wang and Xu, 2014). The European Directive 2011/65/EC (RoHS 2) restricts the maximum allowable concentrations to 0.1% for lead, mercury, chromium (VI), polybrominated biphenyls (PBB) and polybrominated diphenyl ethers (PBDE) and to 0.01% for cadmium in new EEE products (Directive 2011/65/EU, 2011). Directive 2003/11/EC (2003) sets the same thresholds for pentabromodiphenyl ether (PentaBDE) and octabromodiphenyl ether (OctaBDE) for all products placed on the market. According to Wäger et al. (2012), contents of cadmium and PBDE are likely to exceed European thresholds for ABS, HIPS, PC/ABS and PP from WEEE. A considerable amount of WEEE containing bromine or heavy metals is exported to Asian and African countries instead of being treated properly. High bromine levels indicate that significant numbers of items do not fulfil the requirements for PBDE of the above mentioned directions (Sindiku et al., 2014). Legal requirements and the incompatibility of many plastic types necessitate an effective separation to produce high-quality recyclates (Dodbiba et al., 2005). Depending on the limitations of a specific sorting technology, certain input compositions are necessary with regard to material diversity, colour, additives and so on (Schlummer, 2014). Thus, many mechanical separation processes require an upstream concentration of the target polymer to guarantee an efficient recycling process (Gent et al., 2011). For most separation technologies, a pre-treatment-like crushing or pre-classification (e.g. by size or type) of the waste stream is necessary prior to sorting in order to achieve maximum yield and selectivity. Mark describes shredding and manual dismantling as the two possible treatments to produce plastic-rich streams from WEEE (Mark, 2006). The European Directive 2012/19/EU (2012) stipulates the separation of certain product parts. Therefore, in Germany, TVs, monitors, printers and some small household appliances (SHA) like vacuum cleaners and coffee machines are sorted out and/or dismantled manually to remove panel glass, cartridges and motors. The outcome of manual disassembly is a rather large size of plastic parts. Collecting these parts in separate batches results in a pre-concentration of plastic types and grades. Manual dismantling can be combined with hand-held spectroscopic devices like near-infrared (NIR) spectroscopy and/or X-ray transmission (XRT) spectroscopy. Besides manual separation, density separation, NIR spectroscopy and electrostatic separation can be considered as state-of-the-art recycling processes for WEEE (Krämer et al., 2009; Schlummer et al., 2010). NIR-based sorting is a widespread dry separation method for post-consumer plastics, especially for streams of packaging
waste (Hopewell et al., 2009). The waste stream is exposed to NIR illumination and a sensor or camera system detects the material-specific reflectance or transmission spectrum to compare it with a certain reference spectrum (Massen, 1998; van den Broek et al., 1998). Effective singularisation of particles and a certain range of particle sizes enhance efficient blowing out. Dark particles normally contain carbon black, which absorbs the NIR radiation to a large extent so that no reflected radiation can reach the sensors (Huth-Fehre et al., 1995). Depending on software and hardware specifications, a certain particle size range is preferable. High process yields with NIR sorting can only be achieved for lightly coloured inputs, with high contents of the target polymer(s) with a desirable particle size. For sorting black coloured plastic particles laser-based systems can be used (e.g. with laser induced breakdown spectroscopy – LIBS), however they are still under development and not ready for commercial use. Density separation is the most commonly applied recycling process for post-consumer plastics. It is mostly run as a wet process. This technology can be applied to separate polymer types, polymers from other material and different polymer grades of one and the same type (Mäurer and Schlummer, 2006). Pure homogeneous target fractions cannot be achieved for polymer mixtures with overlapping density ranges (Schlummer, 2014). Mäurer and Schlummer apply density separation to sort halogenated styrenics from halogen-free ones (Mäurer and Schlummer, 2006). In terms of throughput and/or purity, centrifugal systems are more effective than static ones. Nevertheless, static systems are commonly applied industrially (Gent et al., 2011). Waste streams with high plastic contents are obtained by the inevitable manual dismantling of certain waste electronic devices, such as TV sets, monitors or vacuum cleaners (Schlummer et al., 2007). The composition of such plastic fractions is limited to polymer types and grades that are used for these specific devices. In this article, a case study is described in which WEEE fractions from German dismantling centres were characterised, sorted and recompounded by a consortium of German dismantlers, waste recyclers and research centres. The approach included an intelligent materials flow management, taking advantage of the manual treatment. Properly input-tailored recycling processes were deduced from the input composition of different WEEE streams to produce high-quality and contaminant-free recyclates with high recovery rates. Depending on the input, combinations of handheld material identification, crushing, automated NIR and continuous density separation were tested. Analytical and mechanical product tests were performed to evaluate the products’ qualities and the efficiency of the developed process cascades.
Materials and methods Sampling The input material was provided by five German dismantling centres. Casings from monitors and printers (M&P), TVs and SHA and components of these WEEE fractions were examined.
Downloaded from wmr.sagepub.com at University of Otago Library on July 4, 2015
3
Arends et al. Table 1. Amounts used of each fraction. WEEE product
Description
Labelling
Amounts
TV casings
Unsorted Halogenated Halogen-free Die-cut fronts Die-cut backs Die-cut samples unsorted Unsorted Halogenated Halogen-free
TV TV-X TV-noX TVfronts TVbacks TVunsorted M&P M&P-X M&PnoX Mfeet Mbacks Mfronts SHA SHAcoffee SHAvacuum
550 kg and 550 kg 550 kg 550 kg 100 parts 100 parts 100 parts 500 kg and 550 kg 350 kg 550 kg
Monitor and printer casings
Small household appliances
Monitor feet Monitor backs Monitor fronts Unsorted casings Coffee machines Vacuum cleaner casings
73 parts and 30 kg scrap 251 parts and 30 kg scrap 260 parts 400 kg and 200 kg and 15 kg Sorted from 500 kg SHA Sorted from 500 kg SHA
M&P: monitor and printer; SHA: small household appliances.
The amounts that were used of each fraction are presented in Table 1.
Casings of monitors and printers Monitor casings were pre-sorted manually into their device components: monitor fronts (Mfronts), backs (Mbacks) and feet (Mfeet) for material characterisation. For material separation three dismantling centres provided monitor and printer (M&P) casings. Two dismantlers supplied 500 kg and 550 kg, respectively. One dismantler provided 350 kg of halogenated (M&P-X) and 550 kg of halogen-free (M&P-noX) casings, sorted out by sliding spark spectroscopy.
TV casings For material characterisation, die-cut samples from pre-sorted TV fronts (TVfronts), backs (TVbacks) and unsorted (TVunsorted) TV casings were provided by one dismantling centre. For material separation three dismantling centres provided TV casings. One dismantler pre-sorted 550 kg of halogenated (TV-X) and 550 kg of halogen-free (TV-noX) casings by sliding spark spectroscopy. The other two dismantlers provided 550 kg each of unsorted TV casings.
Small household appliances Two dismantlers supplied batches of 200 kg and 400 kg of mixed SHA casings. Another focus was put on: (a) non-dismantled coffee machines and coffee dispensers (SHAcoffee); and (b) dismantled casings of vacuum cleaners (SHAvacuum) that were separated from a batch of 500 kg of mixed SHA by hand sorting.
Material characterisation and analysis Input and output material streams were characterised by a number of laboratory methods. X-ray fluorescence (XRF) spectroscopy
with the Spectro X-Lab 2000 (Spectro Analytical Instruments GmbH & Co.KG, Kleve, Germany) was applied to determine element contents of bromine and chlorine, and the RoHS listed heavy metals lead, cadmium, chromium and mercury. Energy dispersive x-ray flourescence spectroscopy (ED-XRF) is recognised as an analytical approach with a precision of ±20% if applied in screening mode, which is considered to be sufficient for the assessment of whether a plastic part contains more than 0.1% of bromine, mercury, chromium and lead, or more than 0.01% of cadmium (Sindiku et al., 2014). For quality assurance, two polymeric-certified reference materials (polyethylene, EC 681, IRMM, Geel, Belgium, and polyvinyl chloride, PLPVC-3, Breitlaender, Wesel, Germany) covering a concentration range from 0.001 to 0.1% were analysed on each day of the analytical screening procedure. RoHS compliance was confirmed for a sample if the measured levels were safely below the threshold levels of 0.1 (PBB, PBDE, chromium, mercury, lead) or 0.01% (cadmium), respectively. PBDE and PBB levels were certified on the basis of bromine. Decabromodiphenyl ether (DecaBDE), OctaBDE and Decabromobiphenyl (DecaBB) contain 83%, 79% or 85% of bromine, respectively, indicating a level of 0.1% of each flame retardant at 0.083%, 0.079% or 0.085% of bromine. Considering an accuracy of ±20% of ED-XRF, concentrations were considered to be safely below the thresholds if they were below 80% of these levels. To determine the contents of bromine and chlorine at the dismantling centres, the sliding spark spectrometer SSS3 (Iosys, Ratingen, Germany) was used. To determine polymer types, an FTIR instrument (Spectrum One, ATR, golden gate, Perkin Elmer, Waltham, MA, USA) and the handheld NIR device Kusta 4004 (LLA Instruments GmbH, Berlin, Germany) were applied. Particle size classification with screens of 12 mm and 45 mm were performed for the output fractions of the crushing process to determine the particle size distribution. To characterise the big batches of M&P casings, TV casings and SHA static density fractionation was performed on a laboratory
Downloaded from wmr.sagepub.com at University of Otago Library on July 4, 2015
4
Waste Management & Research
Figure 1. Treatment scheme for the WEEE input fractions monitors and printers (M&P), SHA and TVs, comprising manual dismantling, hand-sorting, handheld spectroscopic devices, crushing, NIR or continuous density separation and combinations thereof. ABS: acrylonitrile butadiene styrene; Br: bromine; M&P: monitors and printers; PC/ABS: polycarbonate/acrylonitrile butadiene styrene; PS: polystyrene; SHA: small household appliances.
scale with samples from the coarse particle size reduction. Dipotassium phosphate was dissolved in water to create a density media of 1.0 g cm-3, 1.03 g cm-3, 1.06 g cm-3, 1.07 g cm-3, 1.08 g cm-3 and 1.09 g cm-3. In the first step, the lightest density fraction was separated from 10 kg of each batch, the material was dried and the mass balances were recorded. The bromine content of the light fraction was measured by XRF. The heavy fraction of this first step was then introduced to the density medium of 1.03 g cm-3, the material was dried, mass balances were recorded and the bromine content of the lighter fraction was measured. This procedure was repeated with media of higher densities until the bromine value exceeded 0.1% to determine which density cuts are expedient on a technical scale to produce legally marketable recyclates with high process yields.
Production of test specimen for mechanical product evaluation To determine tensile and impact strength of recycled styrenics and polyolefins, test specimen were produced. The recyclates were homogenised and cleaned in a single-screw extruder with a screen of 150 µm. The extrusion temperatures varied between 180 °C and 230 °C and pressures up to 200 bar were applied. The
extruded products were granulated and test specimen were produced by injection moulding at 240 °C for ABS, polystyrene (PS) and ABS-PC, and 190 °C for PP. The injection moulding dies had temperatures between 45 °C and 60 °C. Tensile and impact (Charpy method) tests were performed in accordance to standardised methods (DIN EN ISO 3167, DIN EN ISO 179).
Technical scale separation Manual dismantling and halogen detection. Dismantling of TVs, monitors, printers and vacuum cleaners was performed manually to remove the panel glass, cartridges and motors in German dismantling centres. Coffee dispensers and vacuum cleaners were separated by hand sorting from batches of SHA. One dismantler applied sliding spark spectroscopy to identify halogens in casings of TVs and M&P, and to separate halogenated parts from halogen-free ones. Material sorting. For material sorting, manual dismantling, hand-sorting, handheld spectroscopic devices, crushing, automated NIR or continuous density separation and combinations of these techniques were applied. Figure 1 depicts the treatment scheme for the different input streams. Two basic approaches were chosen for material sorting. M&P casings were sorted by NIR separation and the particle size of the
Downloaded from wmr.sagepub.com at University of Otago Library on July 4, 2015
5
Arends et al.
Results and discussion Material characterisation
Figure 2. Polymer type composition of monitor and TV components and mixed SHA.
ABS: acrylonitrile butadiene styrene; PC/ABS: polycarbonate/acrylonitrile butadiene styrene; PS: polystyrene.
sorted ABS and PS fractions was further reduced to separate halogenated parts by density separation. The NIR-based separation approach was also tested for SHA casings, but was not pursued owing to insufficient process efficiency. Because of the high content of dark parts, TV casings, mixed SHA fractions, SHAcoffee and SHAvacuum were directly introduced to the density separation process. Size reduction and separation. The separated casings were pressed and shipped to a business partner who crushed the material in a rotary shear and granulator (both Andritz-MEWA GmbH, Gechingen, Germany). Fine grains were separated by a zigzag air-classifier (AUT GmbH, Chemnitz, Germany). Iron was extracted by an overhead suspension magnet and a neodymium magnet (both Steinert Elektromagnetebau, Cologne, Germany), and non-ferrous metals were removed by an eddy current separator. Batches for the downstream continuous density separation were introduced to a hammer mill (Herbold, Meckesheim, Germany) for crushing to particle sizes below 8 mm. Automated NIR separation was tested on a technical scale with a UNISORT® NIR separation system (RTT-Steinert GmbH, Zittau, Germany). The first sorting step was performed with a multiplex sensor. The remaining fraction of parts that could not be identified was sorted once more by applying a camera system with an increased sensor resolution. Technical scale NIR separation was performed for M&P casings and for a batch of mixed SHA. On a technical scale, continuous density separation was performed with the Sorticanter® (Flottweg, Vilsbiburg, Germany). The density media were produced by dissolving dipotassium phosphate in water to produce a density media of 1.0 g cm-3, 1.06 g cm-3 and 1.08 g cm-3. Which density cuts were chosen for which fraction was decided according to the results of the material characterisation. An input of 100 kg was the minimum amount for the Sorticanter® to obtain separation results with repetitious accuracy. Continuous density separation was applied to sort halogen-free PS fractions from TV casings and halogen-free ABS and PS fractions of NIR pre-sorted M&P casings. From a batch of mixed SHA and from the sub-fractions SHAcoffee and SHAvacuum, polyolefins and halogen-free styrenics were separated.
Handheld spectroscopic polymer-type identification was used to get an idea of the polymer-type composition of monitor and TV components and mixed SHA. Figure 2 depicts the results. Figure 2 reveals a high ABS-PC and ABS content for all monitor components. PS was mainly used for Mfeet. In addition to the polymer type identification, Mfeet, Mfronts and Mbacks of 908 monitor casings were characterised by handheld sliding spark spectroscopy to identify the presence of halogens. A total of 29% of Mfronts contained chlorine and/or bromine and 28% of Mbacks, respectively. Less halogens were found in Mfeet, only 10% of the parts contained bromine and/or chlorine. The XRF results of the scrap material correlate with these results. For Mbacks, a total bromine content of 2.4% was measured and 0.15% for the feet. The results of the 100 unsorted TV samples shown in Figure 2 reveal that 86 specimens were composed of HIPS, nine of ABS and five of other polymer types. The XRF identified bromine contents above 0.1% in 36% of the HIPS samples. The chlorine content exceeded 0.5%, even though the bromine level was below 0.1% for another 8% of the HIPS samples. A cadmium content of 0.01% was exceeded for 19% of the samples, but only if either the bromine or chlorine content was too high as well. The bromine content of all ABS samples exceeds 0.1% and the legal threshold for cadmium is exceeded for most ABS samples. Manufacturers could be assigned to all TV parts analysed and ABS was almost exclusively used by a certain manufacturer. From these results, a halogen-free HIPS target fraction of 50% of the total TV fraction could be expected. Since no other halogenfree styrenics could be found, density separation appears to be a promising treatment to produce a pure halogen-free HIPS recyclate. The polymer type composition of TVfronts and TVbacks is similar. Comparing the XRF results of TVbacks and TVfronts reveals that 37% of the backs exceeded bromine levels of 0.1% and only 6% of the fronts exceeded this value. A higher process yield could be achieved if only TVfronts were used, but since TVbacks make up a bigger fraction of the WEEE stream, treating all TV components is recommended. Figure 2 depicts a rather high variety of materials in a mixed batch of SHA. Contents between 16% and 22% were identified for ABS, PS and polyolefins. Polyamide and polycarbonate could also be identified and a content of 33% of other plastics, metals and other material. The results of the static laboratory density fractionation are illustrated by Figure 3. For M&P casings, the bromine content stays below 0.1% for material densities of 1.08 g cm-3 and below, but the yield is only approximately 22% for both batches analysed. For the two batches of unsorted TV casings, legal thresholds for bromine can be kept when a density cut of 1.06 g cm-3 is applied. At this density cut, the yield is 69% and 77%. While the density fractions of both dismantlers show similar compositions for TV casings as well as for M&P casings, the two batches of SHA differ remarkably. For SHA 1 the bromine
Downloaded from wmr.sagepub.com at University of Otago Library on July 4, 2015
6
Waste Management & Research
Figure 3. Yield and bromine contents of density fractions. SHA: small household appliances; M&P: monitors and printers.
content stays below 0.1%, even in higher density fractions. Anyhow, thresholds for lead and cadmium are exceeded in this batch. The bromine content of batch SHA 2 exceeds 0.2% in the density fraction 1.07 g cm-3
