Tracking the Flow of Resources in Electronic Waste - ACS Publications

Article pubs.acs.org/est

Tracking the Flow of Resources in Electronic Waste - The Case of Endof-Life Computer Hard Disk Drives Komal Habib,* Keshav Parajuly, and Henrik Wenzel Department of Chemical Engineering, Biotechnology and Environmental Technology, University of Southern Denmark, Campusvej 55, DK-5230, Odense M, Denmark S Supporting Information *

ABSTRACT: Recovery of resources, in particular, metals, from waste flows is widely seen as a prioritized option to reduce their potential supply constraints in the future. The current waste electrical and electronic equipment (WEEE) treatment system is more focused on bulk metals, where the recycling rate of specialty metals, such as rare earths, is negligible compared to their increasing use in modern products, such as electronics. This study investigates the challenges in recovering these resources in the existing WEEE treatment system. It is illustrated by following the material flows of resources in a conventional WEEE treatment plant in Denmark. Computer hard disk drives (HDDs) containing neodymium−iron−boron (NdFeB) magnets were selected as the case product for this experiment. The resulting output fractions were tracked until their final treatment in order to estimate the recovery potential of rare earth elements (REEs) and other resources contained in HDDs. The results further show that out of the 244 kg of HDDs treated, 212 kg comprising mainly of aluminum and steel can be finally recovered from the metallurgic process. The results further demonstrate the complete loss of REEs in the existing shredding-based WEEE treatment processes. Dismantling and separate processing of NdFeB magnets from their end-use products can be a more preferred option over shredding. However, it remains a technological and logistic challenge for the existing system.

■

INTRODUCTION Mineral resources, in particular, metals, have played a crucial role in the development of modern society. Metals are chosen due to their specific properties, imperative for the functioning of a product. For a big part of human history, only a few metals were commonly used, such as iron, copper, lead, and tin. With the advancement of technology, especially during the last century, more and more metals were brought into use. Today, almost all of the stable elements of the periodic table are used in the modern products.1,2 The unprecedented growth in metals production during the last century has raised concerns regarding long-term availability of metals to meet the future generation’s demand. In this context, the concept of urban mining−recovery of resources from the waste streams is widely seen as an apt solution to deal with the depletion issue of resources.3 However, enhancing the recovery of resources from the waste flows such as waste electrical and electronic equipment (WEEE) is not straightforward due to the immense complexity of modern products, in both design and material aspects.4 Though the overall recycling rate of WEEE has improved as a result of improved collection and processing, it mainly accounts for the base metals, such as steel, aluminum, and copper. It does not necessarily claim the proportional improvement in the recovery of other valuable resources. The complexities in design © 2015 American Chemical Society

features and material composition, including hazardous substances, lack of technically and economically efficient processing techniques, and uncertainties in the market of recycled resources, are contributing to the loss of resources. It has resulted in the down cycling of specialty metals,5 which are used in small amounts in the final products to achieve specific functionalities, such as rare earth elements (REEs).6,7 REEs have been classified as critical resources by a number of governing bodies and research institutes due to their high supply risk and increasing importance in modern applications.8−13 The current recycling rate of REEs from end-of-life (EoL) products is reported to be less than 1%.14,15 This is mainly because of the long lifetimes of their major end-use products, such as wind turbines and passenger vehicles. Another reason for the current negligible recycling rate of REEs is their small amount in the existing waste flows.16 During recent years, the focus on enhancing the recovery of REEs from different waste streams has increased exponentially.17 Several studies have focused on estimating the recovery potential of REEs from neodymium−iron−boron (NdFeB) magnets Received: Revised: Accepted: Published: 12441

May 6, 2015 September 8, 2015 September 9, 2015 September 9, 2015 DOI: 10.1021/acs.est.5b02264 Environ. Sci. Technol. 2015, 49, 12441−12449

Figure 1. Composition of different output fractions resulting from treating the HDDs in a WEEE treatment facility. The pie charts represent the component composition (%) and the bar charts present the material composition of various output fractions (kg). The figure presents the treatment plant layout accompanied by the flow of HDDs throughout the plant. The width of arrows with respect to the output fractions is representative of the share of a particular fraction in the total weight of the output fractions. The dashed arrows represent the subsequent processes for the different output fractions resulting from the WEEE treatment plant.

Environmental Science & Technology Article

12442

DOI: 10.1021/acs.est.5b02264 Environ. Sci. Technol. 2015, 49, 12441−12449

Article

Environmental Science & Technology

Material Flow Analysis. Material flow analysis (MFA)29 is used as a methodological approach to track the flow of REEs present in the NdFeB magnets, along with other resources contained in the computer HDDs. MFA is a widely used tool to balance the incoming and outgoing flows of resources within defined spatial and temporal boundaries. A conventional WEEE treatment facility located in Denmark was chosen to track the end-of-life flow of resources contained in the electronic waste. Figure 1 shows the process flow layout of the WEEE treatment plant, where all the WEEE processing steps can be seen in order. The WEEE processed at this facility is first unloaded in a fully covered area, from where it is passed on to an input conveyer belt. The WEEE reaches the first manual presorting section, where hazardous and/or components containing valuable materials, such as screens, batteries, and printed circuit boards (PCBs), are disassembled from the products. The remaining flow then enters the rotating chain shredder via a conveyer belt. This shredder has a size-adjustable exit hole, which is used to optimize the size of the material coming out of the shredder. As visible from Figure 1, the shredder is equipped with air suction extractors that extract the dust and light material that enters the three-step air-filtering system. The output material from the shredder passes through the first overbelt magnet, where the ferrous fraction is sorted from the rest. The remaining fraction after this magnetic separator enters the size-sorting section, where the size-sorting equipment separates the incoming material into two streams, larger and smaller than 10 cm2. The first magnetic separator and the size-sorting equipment are attached to the same air filtration system as the chain shredder. The material smaller in size than 10 cm2 passes through two magnetic separators. The first one is an overbelt magnet, whereas the second is a drum magnet. At this point, most of the ferrous material is sorted from the rest. The remaining material passes through an eddy-current separator to separate aluminum from other materials. The material larger in size than 10 cm2 enters the fourth magnetic separator, which is an overbelt magnet separator. Here again, the ferrous materials are separated from the rest, which then passes through the second eddy-current separator to sort aluminum from the residual fraction. In this experiment, the second eddy current was not used because of the high aluminum content of the residual fraction coming from the fourth magnetic separators. For this experiment, a total of 244 kg of HDDs, representing 700 2.5” HDDs and 350 3.5” HDDs, was processed in the WEEE treatment facility. The weight of HDDs was selected considering the 1 ton h−1 operating capacity of the WEEE treatment plant, where this experiment took 15 min. Moreover, 244 kg of HDDs was considered an appropriate sample size, due to the logistics, time, and economic constraints regarding fine sorting of different output material fractions of the WEEE treatment plant. The experiment was performed April 3, 2014. The PCBs were taken out of the HDDs through manual sorting before the HDDs were fed into the shredder due to their high economic value and thus are not represented in the abovementioned total weight of HDDs. The input to the shredder and the output material fractions were weighted within the WEEE treatment plant, using a weighting scale with ±1 kg precision. The resulting 10 various output fractions were collected in large bags and transported back to the sorting lab situated at the University of Southern Denmark (SDU). It is worthwhile to mention here that fraction 3 shown in Figure 1 is not an outcome of the size-sorting equipment but originates as

contained in various end-use products.16,18−20 Nevertheless, development of efficient REEs recovery plants at commercial scale is yet to be realized. This is mainly because of the low concentrations of REEs in the products, which are further diluted into the outgoing recyclates from WEEE treatment facilities. Two recent studies have highlighted this issue.21,22 REEs contained in the NdFeB magnets stick to the ferrous surfaces due to their magnetism. Therefore, they become part of the outgoing ferrous fraction. Bandara et al. (2015)23 have shown the existence of traces of neodymium in the slag from steel mills, where the input material for these mills has been the output ferrous fraction of an electronic waste shredder. This study tracks the EoL flow of resources (metals, alloys, plastics, and others) in electronic waste, demonstrated with the help of REE-based permanent magnets, i.e., NdFeB magnets contained in computer hard disk drives (HDDs). The motivation behind it is to ascertain the final fate of resources in the existing WEEE handling and treatment system. The main reasons for selecting the computer HDDs as a case study are (a) HDDs are often reported to be the largest end-user of NdFeB magnets (∼30%),24,25 (b) almost all the HDDs found in the existing WEEE streams contain the NdFeB magnets, and (c) HDDs represent the maximum amount of REEs present in the existing waste flows.16 According to Sprecher et al. (2014)19 HDDs offer the most feasible option for the large-scale recovery of neodymium. In general, the HDDs contain two different kinds of NdFeB magnets: a high performance sintered magnet found in the voice coil actuator assembly of the HDDs and a low-quality NdFeB epoxy-bonded magnet found in the spindle motor.26,27 Due to the low content of REEs in the bonded magnets, only the sintered magnets are considered in this study. To the best of the authors’ knowledge, this is the first study presenting a detailed flow of resources contained in the HDDs, such as REEs, aluminum, steel, copper, and mix plastics, in a WEEE treatment facility. The study shows the empirical results regarding the composition of various output fractions of the WEEE treatment facility. Another innovative element of the study is that it further tracks the resulting output fractions of the WEEE treatment facility until the smelting plants in order to estimate the recovery potential of REEs as well as the other materials present in HDDs. This contribution presents the detailed flow of resources from treatment of EoL HDDs to the final recovery of resources, whereas the figures related to the experiments as well as the detailed material composition of the output fraction can be found in the Supporting Information (SI).

■

METHODS Dismantling of HDDs and Material Composition Analysis of Components. To define the general material composition of computer HDDs, a sample of 20 HDDs, representing different manufacturers and age groups (see Tables S1 and S2 in the SI), were collected from the WEEE stream and analyzed. This sample included 10 HDDs from desktop computers (sized 3.5”) and another 10 from laptops (sized 2.5”), randomly collected from the local WEEE stream. These HDDs were manually disassembled with the help of hand tools, and the individual components were weighed. The elemental composition analysis of the dismantled components was carried out with the help of an X-ray fluorescence (XRF) spectroscopy system.28 The detailed composition of NdFeB magnets was derived from the work of Habib et al. (2014).16 12443


Article

Environmental Science & Technology Table 1. Material Composition of 3.5” and 2.5” HDDs per Componenta component cover base casting platters magnets magnet carrying plates voice coil actuators spindle motor PCBs others total weight

HDD

weight (g)

share of total weight (%)

3.5” 2.5” 3.5” 2.5” 3.5” 2.5” 3.5” 2.5” 3.5”

108.14 9.04 228 29.44 42.53 8.16 13.35 2.75 51.13

19.3 9.4 40.7 30.7 7.59 8.5 2.4 2.9 9.12

2.5” 3.5” 2.5” 3.5” 2.5” 3.5” 2.5” 3.5” 2.5” 3.5” 2.5”

8.87 19.13 7.07 47.18 11.15 32.69 11.83 18.02 7.52 560.16 95.829

9.25 3.41 7.4 8.4 11.64 5.8 12.34 3.22 7.84 100 100

material composition mostly steel comprising >80% Fe and 15−18% Cr; the rest is Mn, V, Co, and Ni Al alloy with a mixture of 90−97% Al, and small amounts of Ni, Fe, Mn, Cr, Co, and V Al alloy with a mixture of 93−96% Al, and small amounts of other alloying elements, such as Fe, Cu, Zn, Ni, Cr, Mn, and V pure Al core coated with thin layers containing Fe, Cr, Co, Ni, Zn, and V mixture of glass and ceramics NdFeB magnets with a composition of 4.6% Pr, 30.8% Nd, 1.6% Dy; the rest is mainly Fe NdFeB magnets with a composition of 4% Pr, 30.4% Nd, 2.4% Dy; the rest is mainly Fe steel coated with Ni

actuator coil consists of Cu, the axis mainly comprises steel, and the arms are made of Al alloy mainly consists of Al with small amount of Fe, Cu, and Mn − 25−30% are screws made of different steel alloys, and thr rest is a mix plastic fraction, containing paper, plastic, and foam

a

Detailed data regarding the source and age of HDDs and the weight of different components found in the computer HDDs can be found in the Supporting Information, Tables S1 and S2.

tablet was estimated using a wavelength dispersive X-ray fluorescence spectrometer (WD-XRF),30 with a detection limit of 0.001−0.01% (w/w) and ±15% precision.

a result of designed leak to avoid any hindrance to the normal flow of materials through different sections in the plant. The first two fractions resulting from the filters were not sorted any further because they mainly comprised dust apart from small plastic and rubber pieces. The remaining eight output fractions were further sorted manually to visualize the material composition of each fraction (see sections S2 and S3 of the SI for figures of output fractions). As shown in Figure 1, magnetic dust containing REEs was present in five different fractions. This dust was collected and analyzed further to reveal the chemical composition (see the next section for more details). The fine sorting of eight different fractions led to creating a mass balance for all the components found in HDDs throughout the process flow, representing different material compositions. After performing this initial mass balance, the subsequent processing, i.e., the final treatment of these output fractions, was also taken into account. The final fate of these fractions was traced until the smelting step, based on the data obtained from WEEE recycling industry, in order to estimate the recovery potential of different metals and alloys contained in the HDDs. On the basis of the expert advice obtained from the WEEE recycling industry regarding the material composition of the output fractions, a 5% material processing loss was assumed for the smelting process in order to calculate the maximum recoverable amount of metals and alloys, e.g., aluminum and steel (see Figure 1). Chemical Characterization of Magnetic Dust. The magnetic dust was manually isolated from the various shredded components with the help of a brush. This dust, found in four ferrous fractions resulting from the magnetic separators plus fraction 3, was collected. The screws and other metal components were taken out from the collected dust. Five representative samples were collected from each fraction. Each homogenized sample was pressed into a tablet (size, 32 mm; pressure, 30 tons for 10 s). The elemental composition of the

■

RESULTS The aim of this research was to explore the recovery challenges of critical resources, such as REEs, from the urban mines by highlighting the drawbacks in the existing WEEE treatment system. This was accomplished by tracking the flow of computer HDDs containing REEs in the form of NdFeB magnets in a conventional WEEE treatment facility. The results are divided into four main sections: material composition of HDDs; overall mass balance of HDDs, with a focus on the material sorting efficiency of the plant; final treatment of the output fractions; and, finally, the flow of REEs in the WEEE treatment plant. Material Composition of 3.5” and 2.5” HDDs. Table 1 presents the detailed composition of 3.5” and 2.5” HDDs, where it becomes evident that aluminum is the dominant constituent of both types of HDDs. Nearly 50% of 3.5” HDDs weight and 40% of 2.5” HDDs weight are represented by aluminum alloy, out of which almost 80% is consumed by a single component, i.e., the base casting. Another component made of aluminum alloy is the top cover of HDDs, though the majority of the 3.5” HDDs showed steel as a basic material for their cover (see Table 1). Furthermore, the platters (data storage discs) were also found to be made of aluminum coated with magnetic layers. The majority of the platters found in 2.5” HDDs are manufactured of glass and ceramics. The spindle motor and the voice coil actuators mainly consist of aluminum, along with a small amount of copper coils and steel (see Table 1 for more details). Overall Material Flow Analysis. Figure 1 displays on the overall mass balance of HDDs throughout the treatment plant, as well as the performance of the treatment facility with respect 12444


Article


Figure 2. Mass flow of different materials contained in the computer HDDs along with their estimated recovery amounts.

stream. The same is the case with voice coil actuators and the spindle motors, which were not fully liberated from the adjoining components and thus appeared in the ferrous fractions (see section S4 in SI for the detailed material composition of nonliberated components and miscellaneous categories mentioned in Figure 1). Final Treatment of Output Fractions. The 10 different output fractions are sent to the final material and/or energy recovery facilities depending on their material composition. These facilities may be located within or outside the country; for example, the energy recovery facilities are mostly situated inside Denmark. However, there are no metal smelters present in Denmark to recover the pure metals or the alloys from the different material fractions resulting from WEEE treatment facilities. Figure 1 reveals the fate of different output fractions with the help of dashed arrows and boxes. Figure 2 shows the simplified flow of different materials present in HDDs from the WEEE treatment plant to the final treatment of resulting output fractions. Figures 1 and 2 show that fractions 1 and 2 are directly sent to the local incinerator because they mainly consist of dust and some mix plastics. The remaining fractions are sent either directly to smelters, such as fractions 3 and 8, or are first sent to the local scrap dealers for fine sorting of different materials and then sent to the smelters sited outside Denmark. As shown in Figure 1, the output fractions 3 and 8 were mixed together because almost 86% and 99% of their weight was dominated by aluminum, respectively. Due to the high material purity of these fractions, it is not required to send them to further shredding and sorting processes, and thus, they were directly sent to the aluminum smelters outside Denmark. The total amount of material entering the smelters was 97.8 kg, where 96.5 kg is aluminum and the remaining impurities consisted of steel, copper, mix plastics, and the magnetic dust (see Figures 1and 2 for details). Rigamonti et al.31 have documented the smelter efficiency for aluminum as 83.5% considering the aluminum contained in the municipal solid waste (MSW). However, on the basis of the high material quality of output fractions and the information collected from

to the separation of various components and materials into the different output fractions. The first thing to be noticed is the difference of nearly 7.3 kg in the output amount compared to the input amount. A big share of this mass difference is the materials that did not come out of the shredder (stayed at the bottom of shredder). This missing weight of HDDs in the shredder is expected to come out of the shredder with the next batch of WEEE processed in the plant. The remaining loss may have happened due to some components being stuck at various points in the plant. The largest output fraction was fraction 8, equivalent to 90.8 kg, resulting from the Eddy-current separator. This fraction mostly consisted of aluminum, which is no surprise, as aluminum dominates the total mass of HDDs (see Table 1). The second largest fraction was fraction 10, originating from the magnetic separator. This fraction weighed almost 56.3 kg and comprised mainly steel, followed by aluminum. The major source of steel in this fraction is the top cover of HDDs and the plates used as a base for magnets. Fraction 4 was the third biggest output fraction that resulted from the first magnetic separator. The total weight of this fraction was 34.4 kg. As visible from Figure 1, this fraction mainly consisted of steel coming from the top and bottom plates for magnets and the top cover of HDDs. Nearly 25% of this fraction consisted of the partially broken HDDs and the residue consisting of spindle motors, screws, spacer rings, and broken glass and ceramics from the platters. In general, it can be seen that none of the output fractions consist of a single material. For example, the purpose of installing the magnetic separators in the WEEE treatment facility is to isolate the ferrous metals from the rest. However, all four fractions originating from the magnetic separators comprised various materials apart from iron and steel, such as platters. The majority of the platters are made of aluminum coated with thin magnetic layers containing chromium, cobalt, iron, nickel, and zinc to enhance the magnetic storage capacity of these platters. Due to these thin layers, the platters are attracted by the magnetic separators and end up in the ferrous 12445


Article

Environmental Science & Technology the experts in the WEEE treatment and final processing (smelters) industry, we have assumed 5% processing loss for the aluminum and steel smelters. Thus, the total amount of aluminum that can be recovered from fractions 3 and 8 is equivalent to 91.67 kg. Fraction 10the second largest outputmainly consisted of aluminum (47%) and steel (52%). This fraction is first sent to the local scrap dealer, where 0.26 kg of pure copper from the voice coil actuators and the spindle motors is separated and sent to the copper smelter. The processing loss for the copper smelter is 1%.32 The remaining amount is then sent to a local shredding (jaw crushing) plant to crush the material into small pieces, hence making it convenient to liberate materials from each other, such as steel and aluminum. These fine-sorted fractions are then sent to the aluminum and steel smelters located outside Denmark. The finally recovered amount of aluminum and steel is estimated to be 24.95 and 28.04 kg respectively, assuming 5% processing loss in the aluminum and steel smelters. As shown in Figure 2, fraction 4 is a ferrous stream resulting from the first magnetic separator. The total weight of this fraction was 34.4 kg, out of which nearly 31 kg was steel. The remaining consisted of aluminum, copper, plastics, and magnetic dust. This fraction is also sent to the local shredding and sorting plant and eventually reaches the final smelters. At this point, there lies an opportunity to separate 0.64 kg of copper and 2 kg of aluminum from the steel. The rest is sold to the steel smelters located outside Denmark, where 29 kg of steel can be finally recovered. Fractions 5−7 and 9 consist of mixed materials, mainly the nonliberated and the miscellaneous components. These fractions are mixed together and then sold to a local scrap dealer, who then sends the material to a local shredding and sorting plant. This process allows the partially broken HDDs to be shredded into smaller pieces, leading to enhanced liberation of various components from each other. At this step, 1.9 kg of copper can be separated as a pure copper fraction. The fine sorted fractions are then sent to aluminum and steel smelters situated outside Denmark. The total amount of steel and aluminum entering these smelters is 17.4 and 17.7 kg, respectively, where the finally recovered amount of steel and aluminum is estimated to be 16.53 and 16.81 kg, respectively. Almost 10 kg of residue material from these fractions is sent to the local incinerator. Finally, the results presented in Figure 2 make it clear that 212 kg of metals and alloys is recovered from an input stream of 244 kg of HDDs. Out of these 212 kg of refined metals and alloys, 135.4 kg is aluminum, 73.6 kg is steel, and 2.87 kg is copper. This implies that nearly 32 kg is lost throughout the process chain, from EoL HDDs to the refined materials. Out of this, nearly 7.3 kg did not appear in the output fractions of WEEE treatment facility, and the rest is lost during the subsequent processes, such as additional shredding and sorting followed by smelting. Tracking the Flow of REEs. Figure 1 shows a detailed process flow of REEs contained in the NdFeB magnets present in computer HDDs across the WEEE treatment facility. The NdFeB magnets found in HDDs are brittle in nature and convert into powder when they pass through the shredder. This powder retains its magnetism and is prone to stick to the ferrous surfaces. Results presented in Figure 1 reveal that all four output fractions originating from the magnetic separators and fraction 3 do contain a considerable amount of this powder

sticking to the surfaces of the shredded ferrous stream. This dustlike powder forms lumps due to the attraction of dust particles caused by the magnetic nature of particles forming this dust (see Figure S20 in SI). The total amount of collected dust from the four output fractions was only 2.7 kg. Comparing this amount to the average weight of NdFeB magnets in the input HDDs, which is equivalent to 6.6 kg, highlights that almost two-thirds of the input magnet fraction was lost during the processing of HDDs. This missing amount of dust containing REEs has high potential to stick to the internal walls of different processing equipment (e.g., chain shredder, pipes, sides of the conveyer belts, and the collection containers), as they are mainly made of ferrous metals. Apart from this, nearly 0.28 kg of intact NdFeB magnets was found in the partially shredded HDDs in the nonliberated component of different output fractions (see Figure S5 and Table S3 in the SI). This corresponds to 0.09 kg of neodymium and 0.006 kg of dysprosium. A detailed elemental analysis of the REEs containing dust collected from five different fractions (see Figure 1) is presented in Table S4 of the SI. The results impart that the amount of REEs present in this dust fraction was negligible, where neodymium and dysprosium made up only 0.9 and 0.1% of the total dust weight, respectively. This translates into 0.02 kg of neodymium and 0.003 kg of dysprosium in the dust fraction. The estimated amount of neodymium and dysprosium in the NdFeB magnets contained in the input stream of HDDs was equivalent to 1.98 kg of neodymium and 0.13 kg of dysprosium. Comparing this to the amount of neodymium and dysprosium in the output dust fractions reveals that almost 99% of the input REEs is lost during processing of HDDs in the WEEE treatment facility. This entire loss of REEs can be attributed to two main reasons: (1) The NdFeB magnets convert into powder in the shredding process and this powder retains the magnetism of the magnet. Due to this magnetic nature, the powder sticks to the ferrous surfaces of different equipment of the treatment plant. Already at this point almost 90% of the REEs are lost, and the remaining amount ends up in the ferrous fractions. (2) During the shredding process, apart from the NdFeB magnets, other different, easily breakable components, mainly plastic and ceramics, also convert into powder, depending on the retention time in the shredder. This resulting powder further dilutes the REEs-containing dust and disperses over various output fractions, and thus, REEs are completely lost during the process. The shredding-based treatment of HDDs is not efficient to recover the small quantity of REEs from the magnet. A better approach for this could be to separate the magnets from HDDs before they enter the shredder. Currently, manual disassembly of HDDs is performed to take out the PCBs due to their high economic value. The NdFeB magnets can also be recovered manually from the HDDs at the same time. However, it this not practiced due to the difficulty in removing these magnets manually because of their strong magnetism. The manual separation may not be always economically feasible considering the current low concentration of REEs and high labor costs in Denmark.16

■

DISCUSSION The overall material flow analysis of computer HDDs highlights a number of obstacles in the efficient recovery of different resources. One of them is the design of products that hinders 12446


Article


composition vary depending on the size and manufacturer of the end-use products. This does not allow the existing WEEE processing technology to effectively concentrate the recyclates suitable for REEs recovery. Furthermore, these magnets are not easy to disassemble due to the design features and connections. Their strong magnetism is another obstacle in separating these magnets from their end-use products and transporting them for further processing. These challenges lead the recyclers to treat the EoL products containing NdFeB magnets in a traditional WEEE treatment process, i.e., shredding. The shredding process, however, results in the complete loss of NdFeB magnets and the REEs contained in them. In order to tackle these issues related to REEs recovery, it is imperative to separate magnets form their EoL products. This will allow concentrating the target elements, such as REEs, present in the EoL product, leading to their efficient recovery. However, as shown by Habib et al.,16 the small amount of NdFeB magnets contained in the computer HDDs is not economically attractive for the WEEE treatment plants in Denmark. The additional labor and time required to separate magnets from the EoL products is seen as a key hindrance with respect to the current volatile market price of REEs. Nevertheless, the number of EoL products containing NdFeB magnets is likely to increase in the future.16,34 Moreover, future developments regarding commercial-scale recycling technologies can make it possible to separate the NdFeB magnets and to make the final recovery of REEs from these magnets an economically efficient process. To enhance the recovery of critical resources from WEEE, we suggest the improvement on both ends of the value chain, beginning with the design of products and ending with the handling and treatment of EoL products. Designing the products in a way to enhance the liberation of different components and materials contained in the products is the key for maximizing the resource efficiency. Simultaneously, the improvement of existing WEEE handling and treatment system is necessary to ensure the efficient recovery of critical resources. A solution here could be installing robots for automatic disassembly of products, followed by separation of components containing the critical resources. This could be a technically feasible solution in the case of HDDs due to the almost identical design of HDDs, where it becomes easier for the robots to cut and open the specific part of HDDs containing the NdFeB magnets. These parts can be then heat-treated35 to demagnetize the magnets for further sorting and processing. Treating the computer HDDs with hydrogen to release the REEs mixture from the product is another promising solution.36 Finally, the authors are aware of the data uncertainties associated with the material composition of the input stream to the WEEE treatment facility. The exact composition of 244 kg of computer HDDs was not analyzed, because disassembly and sorting of various components and materials would have been a prerequisite for such an analysis. This could have manipulated the actual performance of the WEEE treatment plant. For this reason, the composition of the incoming flow was generalized on the basis of detailed composition analysis of 20 HDDs. As the computer HDDs often follow the same design configuration and material composition, the sample size of 20 HDDs from different manufacturers and age groups was considered a sufficient sample size for this kind of study. Furthermore, the model uncertainties related to the processing of HDDs, such as the difference of 7.3 kg between the input and output of the

the effective recovery of various materials contained in the products. This issue can be better explained with the results presented in this study. The comprehensive data regarding the composition of eight output fractions resulting from the WEEE treatment plant shows that all of these fractions contain unintended materials. For example, the fractions resulting from the magnetic separators consist of aluminum and copper apart from their purposive output product, i.e., ferrous materials. The complex design of modern products hinders the complete liberation of components and, thus, the materials from each other. In the case of HDDs the screws and other fixtures are the main obstacles in separation of components like the top cover and the base casting from each other. This results in aluminum contamination in the ferrous fraction and vice versa. The WEEE treatment technologies are not yet mature enough to recover the elements from the complex mixture of different materials in the products that are not designed for the EoL material recovery. The current WEEE treatment system is focused on materialcentric recycling33 that aims at recycling of bulk materials (metals and alloys), such as aluminum, copper, iron, and steel, and the precious metals, such as gold, silver, and platinum group metals. Critical raw materials, such as REEs, are not on the priority list of materials recycling, due to both their current volatile market price and the small and dispersed amounts in existing waste flows. The complexity of modern products in terms of material composition and design features cannot be addressed by the existing manner of WEEE recycling. This holds more true during the initial processing of WEEE, where different types of products are shredded together to generate material streams that follow the material-centric recycling chain. The product-centric approach is seen as a potential solution to maximize the material recovery from WEEE an attempt to close the material cycle. The increasing focus on design for recycling, design for EoL, design for metallurgy, design for sustainability, and similar approaches seems to be promising in order to ensure increased resource efficiency in future. Regarding REEs recovery, the semiquantitative data collected from the WEEE treatment companies in Denmark show that in 2014 nearly 60 Mg of HDDs were received at the WEEE treatment plants in Denmark. Out of these, almost 85% were 3.5” and the remaining were 2.5” HDDs. Taking an average weight of 13.35 and 2.75 g of NdFeB magnets found in the 3.5” and 2.5” HDDs, respectively, this amount translates into a total of 1.6 Mg of NdFeB magnets present in the computer HDDs reaching the WEEE treatment plants in Denmark in 2014. However, Habib et al.16 showed that the current maximum theoretical recovery potential of NdFeB magnets contained in the computer HDDs was estimated to be 4.5 Mg. This implies that in reality only 35% of the EoL computer HDDs succeeds in reaching WEEE treatment facilities in Denmark. The remaining amount of the EoL computers does not enter the local WEEE treatment chain and is transported to WEEE treatment facilities located outside Denmark. This means that almost two-thirds of the recovery potential of REEs is lost at the beginning of the whole recycling process chain. The next challenge, as highlighted by Habib et al.,16 is the very small amount of REEs present in the different end-use products. The weight and composition of NdFeB magnets vary between different product types within the same product category, such as between computer HDDs and the cell phones within the IT and telecommunication category. Even within the same product type, the NdFeB magnets weight and 12447


Article


(11) Bauer, D.; Diamond, D.; Li, J.; Sandalow, D.; Telleen, P.; Wanner, B. Critical Materials Strategy; US Department of Energy: Washington, DC, 2010. (12) National Research Council. Minerals, Critical Minerals, and the US Economy; National Academies Press: Washington, DC, 2008. (13) Habib, K.; Wenzel, H. Reviewing resource criticality assessment from a dynamic and technology specific perspective − using the case of direct-drive wind turbines. J. Cleaner Prod. 2015, DOI: 10.1016/ j.jclepro.2015.07.064. (14) Graedel, T.; Allwood, J.; Birat, J.-P.; Buchert, M.; Hagelüken, C.; Reck, B. K.; Sibley, S. F.; Sonnemann, G. Recycling Rates of Metals: A Status Report; United Nations Environment Programme: Paris, France, 2011; isbn 9280731610. (15) Reck, B. K.; Graedel, T. Challenges in metal recycling. Science 2012, 337 (6095), 690−695. (16) Habib, K.; Schibye, P. K.; Vestbø, A. P.; Dall, O.; Wenzel, H. Material Flow Analysis of NdFeB magnets for Denmark: A comprehensive waste flow sampling and analysis approach. Environ. Sci. Technol. 2014, 48 (20), 12229−12237. (17) Tsamis, A.; Coyne, M. Recovery of Rare Earths from Electronic Wastes: An Opportunity for High-Tech SMEs; European Union: Brussels, 2015. (18) Rademaker, J. H.; Kleijn, R.; Yang, Y. Recycling as a Strategy against Rare Earth Element Criticality: A Systemic Evaluation of the Potential Yield of NdFeB Magnet Recycling. Environ. Sci. Technol. 2013, 47 (18), 10129−10136. (19) Sprecher, B.; Kleijn, R.; Kramer, G. J. Recycling Potential of Neodymium: The Case of Computer Hard Disk Drives. Environ. Sci. Technol. 2014, 48 (16), 9506−9513. (20) Yano, J.; Muroi, T.; Sakai, S.-i. Rare earth element recovery potentials from end-of-life hybrid electric vehicle components in 2010−2030. J. Mater. Cycles Waste Manage. 2015, 1−10. (21) Widmer, R.; Du, X.; Haag, O.; Restrepo, E.; Wäger, P. A. Scarce Metals in Conventional Passenger Vehicles and End-of-Life Vehicle Shredder Output. Environ. Sci. Technol. 2015, 49, 4591−4599. (22) Bandara, H. M. D.; Darcy, J. W.; Apelian, D.; Emmert, M. H. Value Analysis of Neodymium Content in Shredder Feed: Toward Enabling the Feasibility of Rare Earth Magnet Recycling. Environ. Sci. Technol. 2014, 48 (12), 6553−6560. (23) Bandara, H. M. D.; Mantell, M. A.; Darcy, J. W.; Emmert, M. H. Closing the Lifecycle of Rare Earth Magnets: Discovery of Neodymium in Slag from Steel Mills. Energy Technology 2015, 3 (2), 118−120. (24) Du, X. Y.; Graedel, T. E. Global Rare Earth In-Use Stocks in NdFeB Permanent Magnets. J. Ind. Ecol. 2011, 15 (6), 836−843. (25) Kara, H.; Chapman, A.; Crichton, T.; Willis, P.; Morley, W. Lanthanide Resources and Alternatives; report DFT-01 205 issue2.doc for Department for Transport and Department for Business, Innovation and Skills; Oakdene Hollins Research and Consulting: Aylesbury, UK, 2010. (26) Ueberschaar, M.; Rotter, V. S. Enabling the recycling of rare earth elements through product design and trend analyses of hard disk drives. J. Mater. Cycles Waste Manage. 2015, 17, 266−281. (27) Hatch, G. Seagate, Rare Earths and the Wrong End of the Stick. http://www.techmetalsresearch.com/2011/07/seagate-rare-earthsand-the-wrong-end-of-the-stick/; 2011. (28) Thermo Scientific, Niton XL3t XRF analyzer. (29) Brunner, P. H.; Rechberger, H. Practical handbook of material flow analysis. Int. J. Life Cycle Assess. 2004, 9 (5), 337−338. (30) Philips X-ray fluorescence spectrometer PW2400. (31) Rigamonti, L.; Grosso, M.; Sunseri, M. C. Influence of assumptions about selection and recycling efficiencies on the LCA of integrated waste management systems. Int. J. Life Cycle Assess. 2009, 14 (5), 411−419. (32) United Nations University. 2008 Review of Directive 2002/96 on Waste Electrical and Electronic Equipment (WEEE): Annex to the Final Report; United Nations University: Bonn, Germany, 2007. (33) Reuter, M. A.; Kojo, I. V. Challenges of metals recycling. Materia 2012, 2, 50−56.

treatment facility, are not assessed extensively in this work. More empirical data is required in order to address these issues, which is an ambition for future research.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.est.5b02264. A description of the detailed material composition of the output fractions, the chemical characterization results for the magnetic dust, and pictures related to the experiment (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]; phone: +45 24409707. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We would like to thank the projects INNOSORT (http:// innosort.teknologisk.dk/) and TOPWASTE (www.topwaste. dk) for providing us the opportunity to carry out this research and facilitating our access to the necessary data. The support provided by Ciprian Cimpan and Peter Klausen Schibye regarding the experiment and sorting is highly acknowledged. Similarly, the authors are thankful to Tom Ellegaard for facilitating the experiment. The authors pay their deepest gratitude to Lorie Hamelin and the two anonymous reviewers for providing valuable comments.

■

REFERENCES

(1) Greenfield, A.; Graedel, T. E. The omnivorous diet of modern technology. Resources, Conservation and Recycling 2013, 74 (0), 1−7. (2) Graedel, T. E.; Harper, E. M.; Nassar, N. T.; Nuss, P.; Reck, B. K. Criticality of metals and metalloids. Proc. Natl. Acad. Sci. U. S. A. 2015, 112 (14), 4257−4262. (3) Graedel, T.; Gunn, G.; Espinoza, L. T. 1. Metal resources, use and criticality. In Critical Metals Handbook, 1st ed.; John Wiley & Sons, Ltd.: New York, 2014. (4) Reuter, M. A.; Hudson, C.; van Schaik, A.; Heiskanen, K.; Meskers, C.; Hageluken, C. Metal Recyling: Opportunities, Limits, Infrastructure; United Nations Environment Programme: Paris, France, 2013; pp 1−320. (5) Graedel, T.; Allwood, J.; Birat, J.-P.; Buchert, M.; Hagelüken, C.; Reck, B. K.; Sibley, S. F.; Sonnemann, G. Recycling Rates of Metals: A Status Report. United Nations Environment Programme: Paris, France, 2011. (6) Nelen, D.; Manshoven, S.; Peeters, J. R.; Vanegas, P.; D'Haese, N.; Vrancken, K. A multidimensional indicator set to assess the benefits of WEEE material recycling. J. Cleaner Prod. 2014, 83, 305− 316. (7) Wang, F.; Huisman, J.; Meskers, C. E. M.; Schluep, M.; Stevels, A.; Hagelüken, C. The Best-of-2-Worlds philosophy: Developing local dismantling and global infrastructure network for sustainable e-waste treatment in emerging economies. Waste Manage. 2012, 32 (11), 2134−2146. (8) European Commission. Report on Critical Raw Materials for the EU: Report of the Ad Hoc Working Group on Defining Critical Raw Materials; European Commission: Brussels, Belgium, 2014. (9) European Commission. Critical Raw Materials for the EU; European Commission: Brussels, Belgium, 2010; pp 1−85. (10) Bauer, D. D, D.; Li, J.; McKittrick, M.; Sandalow, D.; Talleen, P. Critical Materials Strategy; US Department of Energy: Washington, DC, 2011. 12448


Article

Environmental Science & Technology (34) Habib, K.; Wenzel, H. Exploring Rare Earths supply constraints for the emerging clean energy technologies and the role of recycling. J. Cleaner Prod. 2014, 84, 348−359. (35) Ralf Holzhauer, L. B.; Spiecker, T. Recycling of rare earth magnets, theoretical and experimental results from sorting experiments. 13th International Electronics Recycling Congress IERC 2014, Vienna, Austria, 22−24th January,2014. (36) Harris, I. R.; Williams, A.; Walton, A.; Speight, J. Magnet recycling. US Patent US 20120137829 A1, 2014.

12449
