The present article examines flows and stocks of Stockholm Convention regulated pollutants, commercial penta- and octabrominated diphenyl ether (cPentaBDE, cOctaBDE), on a city level. The goals are to (1) identify sources, pathways, and sinks of these compounds in the city of Vienna, (2) determine the fractions that reach final sinks, and (3) develop recommendations for waste management to ensure their minimum recycling and maximum transfer to appropriate final sinks. By means of substance flow analysis (SFA) and scenario analysis, it was found that the key flows of cPentaBDE stem from construction materials. Therefore, end-of-life (EOL) plastic materials used for construction must be separated and properly treated, for example, in a state-of-the-art municipal solid waste (MSW) incinerator. In the case of cOctaBDE, the main flows are waste electrical and electronic equipment (WEEE) and, possibly, vehicles. Most EOL vehicles are exported from Vienna and pose a continental, rather than a local, problem. According to the modeling, approximately 73% of cOctaBDE reached the final sink MSW incinerator, and 17% returned back to consumption by recycling. Secondary plastics, made from WEEE, may thus contain significant amounts of cOctaBDE; however, uncertainties are high. According to uncertainty analysis, the major cause is the lack of reliable values regarding cOctaBDE concentrations in European WEEE categories 3 and 4, including cathode ray tube monitors for computers and televisions. We recommend establishing a new, goal-oriented data set by additional analyses of waste constituents and plastic recycling samples, as well as establishing reliable mass balances of polybrominated diphenyl ethers’ flows and stocks by means of SFA.
Polybrominated diphenyl ethers (PBDEs) were routinely applied as flame retardants in polymer materials in a wide spectrum of applications for several decades. The group of PBDEs consists of 209 congeners that differ by a configuration of bromine atoms on the two benzene rings. Commercial mixture of pentabromodiphenyl ether (cPentaBDE) refers to a complex mixture, in which the main components are denominated as BDE-47, BDE-99, and BDE-100/85. These congeners have the highest concentration by weight with respect to the other components of the mixture. They are classified as persistent organic pollutants (POPs), posing a risk for human health and the environment, and have been listed under the Stockholm Convention as “new POPs” (UNEP 2009a). The main components in commercial octabromodiphenyl ether (cOctaBDE) mixtures, listed as new POPs (UNEP 2009b), are BDE-153, BDE-154, BDE-170, BDE-171, and BDE-183; other components in this mixture are BDE-196, BDE-197, and BDE-203.1 Decabromodiphenyl ether, which is not regulated by the Stockholm Convention, is not part of this analysis.
On the European level, cPentaBDE and cOctaBDE are administered by two key directives (EC 2003a, 2003b). The European framework established removal of pollutants as a basic treatment rule and restricted recycling of waste electrical and electronic equipment (WEEE) containing cPentaBDE and cOctaBDE above a threshold value of 0.1 weight percent (1 gram per kilogram) (0.1 wt% [1 g/kg]),2 also referred to as a maximum concentration value (MCV). Ten WEEE categories were created for monitoring purposes, with category 3 (information technology [IT] and telecommunication equipment) and category 4 (consumer equipment) being those relevant with respect to POP-PBDEs (Wäger et al. 2011; UNEP 2012a).
From European waste hierarchy and life cycle assessment perspectives, recycling of plastic materials is a preferred option; however, contaminants such as POP-PBDEs must be strictly excluded from recycling streams. Various recycling technologies for PBDE-containing plastics have been continuously developed during the past decade, employing either mechanical or chemical processes. The technologies3 that allow for production of recycled plastics from mixed plastic streams, with POP-PBDEs’ levels below MCV, are in development or pilot scale. No information was available (as of July 2012) on any full-scale operation, approximately ten years after industry announced this approach as an aim (UNEP 2012b).
Application of cPentaBDE, in terms of significance and prominence, was as a fire retardant in polyurethane (PUR) foam, used in vehicles and building insulation, and cOctaBDE was predominantly applied to acrylonitrile butadiene styrene (ABS) polymer casings in electrical and electronic equipment (EEE). POP-PBDEs are expected in high concentrations in cathode ray tube (CRT) computer monitors and televisions. Other less-known uses of POP-PBDEs include polyvinyl chloride (PVC) and polyethylene (PE) polymers for construction and, possibly, high-impact polystyrene (HIPS), polybutylene terephthalate (PBT), and polyamide (PA) polymers for vehicles (table 1). Because of the relatively long service times of these products, large amounts of POP-PBDEs are still present in consumption stock of cities.
Table 1. Former uses of cPentaBDE and cOctaBDE in different sectors
Compound and sector
Mass flow estimates
Mass flow estimates vary in different sources, especially in the case of cOctaBDE, and are therefore only indicative. The table is based on UNEP (2012a), Morf et al. (2003), Leisewitz and Schwarz (2000), and Lassen and Løkke (1999). cPentaBDE = commercial pentabrominated diphenyl ether mixture; cOctaBDE = commercial octabrominated diphenyl ether mixture; PUR = polyurethane; PVC = polyvinyl chloride; HIPS = high-impact polystyrene; PBT = polybutylene terephthalate; PA = polyamide; PE = polyethylene; ABS = acrylonitrile butadiene styrene; WEEE = waste electrical and electronic equipment; CRT-PCs and TVs = cathode ray tube computer monitors and televisions; EU = European Union.
Upholstery of seats, ceiling, headrest, textile back-coating
PUR foam in vehicles and construction accounts for 90% to 95% of total cPentaBDE use.
Textiles, printed circuit boards, cable sheets, conveyor belts, and so on
Other applications less than 5% of total use
HIPS, PBT, PA
Dashboard and steering wheel
Estimates range from minor to major use
WEEE categories 3 and 4, with focus on CRT-PCs and TVs
Major use, estimated up to 95% of total cOctaBDE use in the EU
Various problems are associated with cPentaBDE and cOctaBDE. First, despite the regulatory framework, cOctaBDE concentrations exceeding MCV were observed in polymers in European WEEE streams (e.g., Wäger et al. 2011; Schlummer et al. 2007; Morf et al. 2005; see also Pizzol et al. 2012). Second, the presence of POP-PBDEs was reported in the environment, including indoor dust, soils, sediments, and sewage sludge (e.g., UBA 2008). Third, information about the fate of POP-PBDEs in the city of Vienna, as in every other city, is quite limited.
The main objective of the management of POP-PBDEs is to dispose of them in environmentally sound final sinks, such as complete mineralization by incineration or storage in an underground site. A “final sink” can be defined in two ways: either by transformation into other substances so that the original does not exist anymore, as in incineration, or in a biochemical process. It also applies to the storage of a substance in water, air, or soil with a residence time of >10,000 years (Döberl and Brunner 2004). PBDEs are organic semivolatile compounds, which can be effectively transformed into harmless substances by advanced thermal treatment and air pollution control. Possible formation of emissions, such as dioxins and furans, can be controlled by state-of-the-art technologies, which are used in today's municipal solid waste (MSW) incinerator (Morf et al. 2003; UNEP 2012b). For POP-PBDE-containing polymers, waste to energy is thus a preferred treatment option, as long as no automated full-scale operation capable of their removal is in place, and state-of-the-art incineration is an appropriate final sink for POP-PBDEs.
This article presents a case study of POP-PBDEs in the city of Vienna. The city level for the analysis was chosen for four reasons. First, cities are major sources of PBDE emissions. PBDE concentrations on sites close to emission sources, such as road soils in urban areas, are much higher than at remote sites (e.g., Luo et al. 2009; Gevao et al. 2011). Second, by choosing a city, flows of contaminants to the so-called hinterland4 are exposed and thus become more apparent (Brunner and Rechberger 2004). Third, the knowledge of contaminants’ fate from the source to the final sink is crucial for the development of waste management strategies at the urban level; the municipality has the means to manage POP-PBDEs by designing waste management systems. Last, municipalities can often supply information about flows and stocks of goods and substances within its regime.
This analysis takes into consideration previously published data; therefore, no laboratory analyses were conducted for this particular study. The goals of the present article are to (1) identify sources, pathways, and sinks, (2) determine the fractions of cPentaBDE and cOctaBDE that reach final sinks, and (3) develop recommendations for goal-oriented waste management in order to ensure minimum recycling and maximum transfer of POP-PBDEs to final sinks.
In order to achieve the desired objectives, substance flow analysis (SFA) methodology and scenario analyses have been applied. Flows and stocks of cPentaBDE and cOctaBDE were modeled with the SFA software, STAN (STAN 2012), taking uncertainties into account. The purpose of scenario analysis was to address the problems of missing data and substantial uncertainties.
Substance Flow Analysis
SFA, as defined in Brunner and Rechberger (2004), is based on the law of conservation of matter. The mass balance principle allows for linking inputs, outputs, and stocks. Hence, SFA facilitates the verification of data and enables the estimation of missing data by making use of the redundancy in a material flow system. SFA has been widely applied to trace pollutants, such as PBDEs, heavy metals, and nutrients, through urban regions (e.g., Morf et al. 2003; Tasaki et al. 2004; Baccini and Brunner 2012; see also Hansen and Lassen 2002). Results from SFA modeling can be compared with experimental results, such as concentrations of substances in specific flows of goods. For instance, Morf and colleagues (2005) showed, in a case study regarding PBDE flows in a recycling plant, that SFA-based results match field results quite well. For the SFA of cPentaBDE and cOctaBDE in this study, the administrative borders of the municipality of Vienna and the year 2010 were taken as system boundaries in space and time.
STAN (STAN 2012) software allows for modeling on two layers: The first layer handles economic goods with a positive or negative market value, such as plastic materials where POP-PBDEs are present or expected to be present (Table 1), and the second layer handles analyzed substances in a chemical sense (chemical elements or compounds), in this case cPentaBDE and cOctaBDE. Further, the STAN model allows for structuring the SFA system according to key processes and stocks on several levels of processes and subordinate systems. Three processes have been defined on the first level: Consumption; Waste Management; and Environment. Each of these processes is modeled as a subsystem that holds several other processes (figure 1).
The process (or subsystem) Consumption serves to quantify stocks and flows of POP-PBDEs in products that are in use (construction, vehicles, and electrical and electronic equipment [EEE]), and it includes consumer emissions. The processes Construction and Vehicles have no imports, because PBDE-flame retardants were only added to construction and vehicles in the past. Today in Vienna, both stocks are decreasing and will finally, after decades, reach a low value close to zero. The PBDE will either be transferred to the process of Waste Management, which is, by far, the most prominent flow, or to the subsystem of Environment. The two import flows of figure 1 (virtual import 1 and 2) are virtual flows that have a purely formal purpose because the STAN software requires an import for each process within a given system or an internal flow from another process in the system. A special case is the process Use of EEE, because flows of POP-PBDEs can return to this process by reused or recycled goods. This recycling flow is of a particular concern because it prolongs the lifetime of POP-PBDEs, and it exposes consumers to these POPs by recycling goods long after they have been abandoned. The virtual processes 1 and 2 present in subsystem Consumption are no physical processes in the sense of a process chain, but they are included in the model as an aid to sum up all POP-PBDEs present in various types of WEEE (CRT-PCs [computer monitors], CRT-TVs [televisions], WEEE-3r, and WEEE-4r) and, similarly, in various types of construction materials (PUR, PVC, and PE).
Imports into the subsystem Environment consist of the consumer emissions. Emitted PBDEs are transported to the atmosphere and are subsequently transferred, by dry and wet deposition, to soils, municipal waste water, and hydrosphere. Consumer emissions are determined by multiplying existing POP-PBDE stocks with experimentally determined annual emission factors, which are available in the literature, and range from 0.054% for cOctaBDE to 0.39% for cPentaBDE (Morf et al. 2003; UNEP 2006). Water bodies in Vienna account for approximately 5% of the surface area. Because of this relatively small area, and because of low emission factors, depositions to the hydrosphere within the system boundaries have not been taken into account. Therefore, the subsystem Environment is composed of two processes: atmosphere and soil. Sedimentation and resuspension flows are very close to zero and are therefore not included in the system. This may seem to be in contrast with measurements in sediments of the river Danube, which showed 30 times higher concentrations for the sum of 18 selected PBDE congeners downstream of Vienna versus concentrations upstream (UBA 2004). The reason for this increase is likely attributable to a former industrial source located at the confluence of the rivers Schwechat and Danube outside of the city limits and downstream of Vienna (UBA 2004).
A large share of POP-PBDE-containing materials have already entered waste management; therefore, the consideration of waste management options plays a crucial role in the overall assessment (UNEP 2012b). The key process and an appropriate final sink in the subsystem Waste Management is state-of-the-art incineration, where POP-PBDEs are effectively destroyed (Vehlow and Mark 1997). There are three imports to the subsystem Waste Management, the first being construction wastes, which are allocated between incineration and landfill, with the help of a virtual process. In this case, the virtual process gives a possibility to set the incineration/landfill ratio. The second are WEEE plastics, which are essentially divided between reuse, export for treatment abroad, incineration, and recycling; their end-of-life (EOL) pathway starting from collection centers in Vienna is described in detail in the Supporting Information available on the Journal's Web site. The third import is deposition by rainfall on the city. POP-PBDEs are brought this way to the municipal waste water treatment plant, where the majority accumulate in sewage sludge as a result of their low solubility. Following the setup of the system, a subsequent inventory analysis of the products in use is conducted.
Data uncertainties are often a key factor in SFA and therefore require special attention. The software, STAN, allows for the consideration of uncertainties and accepts data input in the form of value ± uncertainty value. It is assumed that uncertain data are normally distributed, given by their mean value, μ, and standard deviation, σ. This approximation, which is sometimes not entirely appropriate, offers the possibility to use methods such as error propagation and data reconciliation (Cencic and Rechberger 2008). In reality, however, the higher the uncertainties are, the less symmetric the error intervals become (Hedbrant and Sörme 2000). Our approach to cope with high uncertainties and/or missing information is scenario analysis.
Three cases have been identified where data were either missing or limited to such an extent that using it may have resulted in a wrong interpretation. The first case is cOctaBDE concentrations in goods in WEEE categories 3 and 4, as split between (a) CRT PC monitors and TVs, where high concentrations are expected, and (b) other products excluding screen devices of WEEE categories 3 and 4, where lower concentrations are expected. The second case is cOctaBDE occurrence in vehicles, and the third case is waste flows of construction polymers. The impact of various scenarios on the system as a whole was evaluated. This was done by choosing a “most realistic” scenario for each of cases 1 to 3 and varying one parameter at the time.
Scenario Case 1: cOctaBDE Concentrations in Polymers of WEEE Categories 3 and 4
WEEE category 3 refers to IT and telecommunications equipment, including computers and monitors (CRT, liquid crystal display [LCD], and plasma), printer units, copying machines, telephones, calculators, and other products. WEEE category 4 refers to consumer equipment, including TVs (CRT, LCD, and plasma), radio sets, video, hi-fi recorders, and other products for recording or reproducing sound or images (for a complete list, see Annex 1B of EC 2003a). As previously mentioned, CRT PC monitors, also abbreviated as CRT-PCs in this article, and CRT-TVs produced before 2005 are target products to be addressed by the inventory and are expected to contain more than 50% of the total cOctaBDE present in the EEE (UNEP 2012a). Despite the importance of this product group, there is a limited knowledge of cOctaBDE concentration in these products; currently, no comprehensive data set is available. The United Nations Environment Programme (UNEP) inventory guidance recommends using a data set by Wäger and colleagues (2010), consisting of relatively few samples (n = 7 for PC monitors and n = 5 for TVs) with very high variance (0.14 to 10.6 grams per kilogram [g/kg] for CRT-PCs and 0.05 to 3.54 g/kg for TVs), where mean and median values are significantly different. The details will be discussed further in the article. The scenarios considered for case 1a, CRT-PCs and TVs, are (1) mean cOctaBDE concentration (taken for most realistic) and (2) median concentration, from Wäger and colleagues (2010). Additionally, the extreme end values (minimum and maximum) will be also tested for a complete overview. For an in-depth inventory, a possible presence of cOctaBDE in other products in WEEE categories 3 to 4 is examined, now excluding screen devices. In this “rest” of the two categories, denominated for this purpose as WEEE-3r and WEEE-4r, the concern is the same as in the case of CRTs: The information about cOctaBDE concentration in WEEE-3r and -4r is limited to very few samples (n ≤ 3) from Wäger and colleagues (2010). The scenarios considered for case 1b, WEEE-3r and -4r, are equally (1) mean cOctaBDE concentration (taken for most realistic) and (2) median concentration, from Wäger and colleagues (2010). The impact of minimum and maximum values is also tested.
Scenario Case 2: The Occurrence of cOctaBDE in Vehicles
Regarding the use of POP-PBDEs in the transport sector, the literature focuses primarily on cPentaBDE formerly used in PUR foam in upholstery of seats, headrests, ceilings, and applications in textile back-coating. The calculation of the cPentaBDE is straightforward, based on the number of vehicles and literature (minimum, mean, and maximum) values of cPentaBDE concentration per vehicle (UNEP 2012a). When considering the cOctaBDE mixture, the extent to which it was used in vehicles is not clear. cOctaBDE was not used in PUR polymers, because it is not compatible with flexible polyurethane foam (FPF) products. Luedeka (2011) believes that reported detection of OctaBDE in FPF products results from FPF absorption properties, possibly related to contact and transfer from plastic scrap having PBDE content, or from misidentification. Nonetheless, cOctaBDE was used in other polymer parts, such as HIPS, PBT, and PA, in dashboards and steering wheels (UNEP 2012b). Three scenarios are considered: (1) no cOctaBDE was used in cars in Europe, as in UNEP (2012a), (2) a worst-case scenario based on Morf and colleagues (2003), who determined that approximately 20% (143 tonnes [t])5 of the total cOctaBDE stock in Switzerland in 1998 was in vehicles. This estimate was based on a Danish study of Lassen and Løkke (1999), which determined the content of PBDEs in the plastic fraction in vehicles without distinction between PentaBDE and OctaBDE and, consequently, distributed cPentaBDE and cOctaBDE according to the world production data. This approach, acknowledged by the POPs Review Committee (UNEP 2008), will be used for calculation of the upper value of the interval for the case of Vienna, and (3) an average between (1) and (2) is taken for the most realistic case.
Scenario Case 3: Waste Flows of Construction Polymers
Finally, there is a lack of information regarding the fate of polymers used for insulation purposes in the construction sector after they have been treated by collection and sorting plants. PUR insulation foam is difficult to recycle, and because of its high calorific content, it is ideally—and likely—incinerated. But, it may as well be that some polymer fractions are landfilled together with other construction materials. In Switzerland, a country with a long and successful incineration history, the incineration/landfilling ratio was estimated at approximately 80:20. This can be used as a first, and realistic, scenario for Vienna. The relation of how much PBDE-treated polymer ends up in the incinerator and how much ends up in a construction landfill is a straight (linear) line; therefore, it makes sense to choose a second scenario far away from the first one, at an inverse ratio 20:80. It has to be noted that the second scenario is an extreme case, and not realistic for Vienna.
Results and Discussion
Waste Management as the Key Process within and outside of Vienna
According to the STAN model, the largest substance flows occur from subsystem Consumption to subsystem Waste Management, consisting mainly of construction wastes in the case of cPentaBDE and WEEE and vehicles in the case of cOctaBDE (figure 2). Vehicles are not treated within Vienna Waste Management, but are partly exported and partly treated in Austrian car shredders. Consumer emissions to the environment are very small and will no longer play a significant role, neither for cPentaBDE (<0.01 tonnes per year [t/a]) nor cOctaBDE (<0.02 t/a). In addition, they will continue decreasing with time as the consumption stock is depleted.
Figure 2 shows the amounts of cPentaBDE and cOctaBDE in the consumer stock, estimated at 76 ± 22 t for cPentaBDE and 22 ± 42 t for cOctaBDE. Both stocks are declining, with approximately the same velocities: dStock is –3.2 ± 0.4 t/a for cPentaBDE and −3.2 ± 5.4 t/a for cOctaBDE. Assuming a static and linear future development, it can be calculated that the two reservoirs of cPentaBDE and cOctaBDE will be depleted in 24 and 7 years, respectively, with high uncertainties for cOctaBDE.
Final Sinks for cPentaBDE and cOctaBDE
In the case of cPentaBDE, the largest flows were found in construction polymers (mostly PUR insulation foam): 2.6 ± 0.4 t/a. The quantity of construction polymers that ended up in landfills together with construction materials is not known (scenario case 3); therefore, the fractions that are incinerated and, respectively, landfilled cannot be discerned. Thus, it is basically unknown which amount of cPentaBDE has reached the appropriate final sink MSW incinerator. Because landfilling preserves the potentially hazardous cPentaBDE, and should be avoided, we recommend separating polymers—especially PUR foam insulation and PVC duroplastic sheeting—from construction wastes and to incinerate this plastic waste fraction. We also recommend separation and incineration of PE polymer waste fraction—especially PE roof sheeting—which may account for cOctaBDE flows into landfills.
The cOctaBDE system is more complex. The main flows are WEEE (1.3 ±2.9 t/a) and, possibly, vehicles (1.8 ± 0.9 t/a). Most Austrian vehicles are exported abroad for second-hand use or dismantling; therefore, the flow has no environmental impact on the city of Vienna. Rather, it points to a supranational challenge, which is not part of this case study. What was described as scenario case 2, the occurrence of cOctaBDE in vehicles, has thus no impact on the system Vienna. According to STAN modeling, 73% of cOctaBDE entering waste management ends up in MSW incineration, with a high uncertainty of 1.2 ± 5.4 t/a. The focus in the case of cOctaBDE is on the flows of WEEE (figure 3). Through reuse and recycling, some cOctaBDE returns into Consumption. The reuse flows are insignificant (sum <0.01 t/a), compared to the recycling flow. Based on input values, the amount of cOctaBDE returning into Consumption by recycled plastics amounts to 0.3 t/a (or 17% of cOctaBDE entering waste management). The rest corresponds, in equal parts, to cOctaBDE exports (5%) and cOctaBDE, which ends up in landfills (5%). Recycling flow uncertainty calculated by STAN is very high (± 6.1 t/a); thus, it is necessary to take a detailed look at the scenario case 1.
Scenario analysis has shown that varying the cOctaBDE input concentrations in polymers of CRT-PCs, TVs, and WEEE-3r and WEEE-4r has a large effect on the results (table 2). The biggest impacts on cOctaBDE flows have variations in the concentration of CRT PC monitors and WEEE-3r. Nevertheless, all four flows affect the results if input concentrations are varied.
Table 2. Results of the scenario analysis, where different cOctaBDE input concentrations in polymer fractions of CRT PCs and TVs (case 1a) of WEEE-3r and -4r (case 1b) were considered
Impact on the system:
Average flow of polymer (t/a),
Scenario 1a and 1b
average cOctaBDE treated fraction (%)
in the fraction (g/kg)
flow estimate (t/a)
Values from measurements in single categories 3 “Information and Communications Technology equipment without screens” and 4 “consumer equipment without screens” and mixed categories 3 and 4 were used. Mean values were chosen for “the most realistic case” and based on single categories only, as done in UNEP (2012a). The median, minimum, and maximum values were chosen from a data set considering also mixed category 3 and 4 measurements. All data are from Wäger et al. (2010). cOctaBDE = commercial octabrominated diphenyl ether mixture; CRT-PCs = cathode ray tube computer monitors; CRT-TVs = cathode ray tube televisions; WEEE = waste electrical and electronic equipment; WEEE-3r = WEEE category 3 excluding screen devices (“rest”); WEEE-4r = WEEE category 4 excluding screen devices (“rest”); t/a = tonnes per year; g/kg = grams per kilograms; Min. = minimum; Max. = maximum.
Min. c = 0.14
0.19 ± 6.36
Max. c = 10.6
0.68 ± 6.63
Mean c = 2.54
0.29 ± 6.06
Median c = 0.66
0.20 ± 6.29
Min. c = 0.05
0.23 ± 6.17
Max. c = 3.54
0.52 ± 6.24
Mean c = 0.87
0.29 ± 6.06
Median c = 0.66
0.28 ± 6.08
Min. c = 0.05
0.24 ± 6.04
Max. c = 1.56
0.96 ± 6.23
Mean c = 0.18
0.29 ± 6.06
Median c = 0.38
0.39 ± 6.07
Min. c = 0.15
0.29 ± 6.06
Max. c = 1.56
0.46 ± 6.10
Mean c = 0.15
0.29 ± 6.06
Median c = 0.38
0.32 ± 6.06
Need for Better Data about cOctaBDE Concentrations
Being that cOctaBDE flows, including WEEE recycling flow, could only be estimated with a great uncertainty, the origins of the uncertainties related to the process “Use of EEE” have been investigated. Part of the uncertainty is caused by the fact that there is no information about the amount of old appliances stockpiled in households; per capita statistics from Switzerland had to be used for the purpose of determining stocks in CRT-PCs and CRT-TVs in Vienna. The high dStock uncertainty is influencing the recycling flow estimate (figure 3). The flows of substances are calculated as flow of goods, multiplied by polymer fractions, and further multiplied by substance concentrations in the polymer. Thus, the uncertainties originate from three parameters, and they are additive. First, the uncertainties in the flow of goods are low, because detailed data about the flows of goods in WEEE-3 and WEEE-4 are available (see the Supporting Information on the Web). Second, the polymer fraction of a product group is, depending on the product, sometimes in a relatively narrow range (21% to 26% for products in the WEEE-4 category), and in other cases, the range is much wider (13% to 38% for CRT PC monitors, similarly for TVs). Because these ranges depend on product design, they cannot be influenced. Third, the main origin of the uncertainty is the wide cOctaBDE concentration range in European WEEE polymers. At the current state of knowledge, the difference between minimum and maximum value is sometimes as high as a factor of 100 (0.14 to 10.6 g/kg in CRT monitors, or 0.05 to 3.54 g/kg in CRT-TVs). In order to narrow down this uncertainty, it is necessary to provide more and improved data, resulting in better mean values and confidence levels (e.g., 1σ, 2σ, and so on). This allows for choosing concentration intervals corresponding to a confidence level satisfactorily for SFA requirements. To support the hypothesis that the current knowledge is insufficient, cOctaBDE concentration data sets available at the present time were reviewed. They include European WEEE streams (Wäger et al. 2010), European housing and mixed WEEE shredder residues (Schlummer et al. 2007), and CRT-PCs and TVs imported to Nigeria (Sindiku et al. 2012). All available cOctaBDE data from these studies are summarized in table 3. The analysis of cOctaBDE flows in a Swiss recycling plant (Morf et al. 2005) will be looked at separately.
Table 3. Overview of measured cOctaBDE concentrations in polymers of CRT-PCs, CRT-TVs, and WEEE categories 3 and 4 without screens (Wäger et al. 2010), in housing shredder residues from CRT-glass recycling (Schlummer et al. 2007), and CRT PCs and TVs polymers (single housing samples) of European origin imported to Nigeria (Sindiku et al. 2012)
Wäger and colleagues (2010) analyzed polymers from European WEEE recycling plants for Eurpean Union (EU)-regulated chemicals, including cOctaBDE. Mixed plastic samples were taken from maximally 20 t of input WEEE material stemming from pretreatment of single WEEE categories 3 and 4 (without screen devices), mixed WEEE category 3 and 4 (also without screen devices), and from specific single products, such as CRT-PCs and TVs, which were expected to contain particularly high levels of contaminating substances. In CRT PC monitors (population, n = 5), an extremely large value, relative to the rest of the data, was found (10.6 g/kg). In some cases, outliers may result from measurement instruments or methodological problems. Because the sampling procedure was conducted according to existing standards, the outlier represents likely a true extreme value of distribution in this case and indicates a high variance of cOctaBDE concentration in polymers of CRT monitors. The presence of an outlier in a sample of n = 5 requires additional samples to be taken to gain more information about the variance of cOctaBDE concentrations in CRT-PCs, and the same is the case for CRT-TVs. For products in single categories WEEE 3 and 4, n = 2 and n = 1 sampling campaigns, respectively, have no statistical value and are therefore not conclusive. This is especially true because by comparing the two single categories, and a mixed WEEE 3 and 4 category results, an outlier of 1.56 g/kg makes the decision of which value to choose as the most representative difficult. The data discussed here are recommended by UNEP as a part of its inventory guidance intended for country reports (national implementation plans) under the Stockholm Convention on POPs, simply because no more data sets exist.
Schlummer and colleagues (2007) examined cOctaBDE levels in plastic fractions of European housing shredder residues (HSRs) and mixed WEEE shredder residues (WSRs). HSRs were defined as polymer fractions obtained by shredding polymeric housing materials only, isolated from WEEE in the course of the recycling of CRT-glass. WSR refers to polymer-rich fractions generated during metal recovery processes in state-of-the-art WEEE processing plants. OctaBDE was identified in 5 of 7 HSR samples, ranging from 2.9 to 13.8 g/kg, and in 7 of 8 WSR samples, ranging from 0.8 to 4.4 g/kg. In this data set, the data were close to normal distribution, with mean values close to median values.
Sindiku and colleagues (2012) analyzed cOctaBDE concentrations in polymers of CRT-PCs (22 samples) and TVs (36 samples) that were exported from Europe to Nigeria. Because of the fact that single housing polymers were analyzed, not mixed plastics as in the cases of Wäger and colleagues (2011) and Schlummer and colleagues (2007), the measured concentrations were expected to contain either a high level of cOctaBDE or no cOctaBDE at all. Only if a specific housing was produced from recycled plastics, which had been contaminated with cOctaBDE upon recycling, a concentration below 1 g/kg might be detected. In the CRT TV samples, the presence of cOctaBDE above the detection limit was reported in 4 of n = 36 samples, all being CRT-TVs from the 1980s. These samples contained cOctaBDE in mostly very high concentrations: 6.6, 59.3, 64.1, and 290 g/kg (mean concentration, 11.7). In CRT PC samples, the presence of cOctaBDE above the limit of detection was not reported in any of n = 22 samples.
SFA methodology and scenario modeling are, in general, appropriate tools for estimating the amount of cOctaBDE in recycling streams. This case study shows that the current data on cOctaBDE concentrations are insufficient for assessing cOctaBDE recycling flow of WEEE polymers within reasonable uncertainty levels. We recommend a larger-scale cOctaBDE concentration-measuring campaign in CRT monitors (PCs and TVs), as well as WEEE of categories 3 and 4. This is needed for better understanding of flows and stocks and hence for avoiding risks arising from cOctaBDE for human health and the environment.
Need for Monitoring of cOctaBDE Flows in Recycling Plants
Our results indicate a possible return of cOctaBDE to the consumption stock by recycled WEEE in the city of Vienna. In Switzerland, a robust experimental analysis at a recycling plant of small WEEE polymers, representing the relevant appliances in WEEE, followed a nation-wide SFA analysis for PentaBDE, OctaBDE, and other contaminants (Morf et al. 2003, 2005). The average cOctaBDE concentration in this WEEE stream amounted to 0.53 ± 0.03 g/kg and agreed fairly well with the SFA-calculated concentration of 0.39 g/kg. Further, cOctaBDE concentrations corresponding to PC-screen housings and TV housing rear covers at the output of the plant were determined at 11 g/kg of cOctaBDE in both cases. In a Swiss experience almost a decade ago, polymers from PC-screen and TV housings posed a problem equivalent to exceeding MCV by a factor of ten and thus should not have been recycled.
In Austria, such analysis has not yet been published. The Austrian law obliges the relevant institution on the federal state level to undertake control of plants that treat hazardous waste at a minimum of every five years (AWG 2002). Information about substance flows is also needed for plastic recycling plants in order to follow POP-PBDEs from sources to final sinks and to ensure that the goals of “clean cycles” and “appropriate final sinks” can be reached by waste management.
SFA modeled by the software, STAN, allows for following the flows and stocks of POP-PBDE on a city level. The precondition is a minimum data set about flows and stocks of goods containing POP-PBDEs and about concentrations of these chemicals in the corresponding goods. In the city of Vienna, in cases of missing information, available data from other regions were applied. These data show wide ranges, so that SFA modeling resulted in calculated flows with, in part, high uncertainties. For future waste management decisions, it is recommended to collect more and better data. This is all the more required because waste management is the key process for a city that has no more inputs of POP-PBDEs as a result of successful regulations, but still has a large stock because of the legacies of the past. In Vienna, by far, the largest flows of POP-PBDEs are contained in various wastes, such as WEEE, construction wastes, and EOL vehicles. MSW incineration with advanced air pollution control represents a safe transformation process for disposal and can be called an appropriate final sink. In Vienna, 73% of cOctaBDE entering waste management ends up in incineration. Landfilling, which is an important process for cPentaBDE-containing construction wastes, releases only minor amounts of POP-PBDEs, but stores them for very long periods. A considerable fraction of POP-PBDEs containing waste is recycled, with little information about the fate during recycling. In order to prevent the cycling of hazardous POP-PBDEs back to the consumer, more information about the recycling processes is required. Equivalent to incineration processes, data about the transfer coefficients (partitioning) of POP-PBDEs in existing recycling plants has to be supplied for comprehensive understanding of the fate during waste management and beyond. It is recommended to establish a comprehensive knowledge base about flows and stocks of plastic materials containing POP-PBDEs, with particular emphasis on EOL products and waste recycling plants.
The authors appreciate comments related to POP-PBDE inventories from Dr. Roland Weber (POPs Environmental Consulting) and from Dipl.-Ing. Franz Neubacher (UV&P Umweltmanagement) regarding the fate of EOL vehicles in Austria. The authors also thank David Laner (TU Vienna) for his comments, Gerda Fischer (Statistics Austria) for explanations related to vehicle statistics, and Andreas Schuh (EAK-Austria) for guidance through the flows of WEEE in Vienna. This work was supported by the Austrian Science Fund (FWF; I 549-N21).
For comprehensive tables on all congeners, contained in both commercial mixtures, and for individual mass shares of each congener in them, refer to UNEP (2012a).
One gram (g) = 10−3 kilograms (kg, SI) ≈ 0.035 ounces (oz). One kilogram (kg, SI) ≈ 2.204 pounds (lb).
See density separation, two-stage density separation, X-ray transmission, laser sorting, temperature/emissivity separation, and CreaSolv® Process (e.g., Schlummer et al. 2006; Allen et al. 2003; UNEP 2012b).
The term hinterland indicates all entities outside of a given system and connected with it through flows of matter and energy (Baccini and Brunner 2012).
One tonne (t) = 103 kilograms (kg, SI) ≈ 1.102 short tons.
Dana Vyzinkarova was an M.Sc. student at the Vienna University of Technology, Vienna, Austria, at the time the article was written. She is currently a Ph.D. student at the Department of Environmental Engineering, Technical University of Denmark, Kgs. Lyngby, Denmark.
Paul H. Brunner is professor and head of the Institute for Water Quality, Resources, and Waste Management at the Vienna University of Technology.