Characterizing Architectural Options for Electronic Waste Recycling Systems Susan A. Fredholm, Jeremy R. Gregory and Randolph E. Kirchain

Abstract—Several electronic waste recycling systems now exist worldwide in many different forms. In order to determine the optimal system structure for a location, a methodology for comparison must be established. This paper outlines a framework developed to compare the environmental and economic performance of recycling systems through analysis of both context and system architecture options. An overview of the available architectural options for designing an e-waste system is presented, and then the comparison framework is applied to an analysis of systems operating in Switzerland, Sweden, the Netherlands, and the US States of California, Maine and Maryland. The analysis presented here focuses on the quantity of e-waste collected by each system. The results of applying this framework to a collection analysis show that while all systems examined appear to still be increasing their mass collected per capita each year, the mass collected per electronic item at end of life may have reached a plateau in the older systems. Furthermore, as a predictor of mass of e-waste collected per capita, the number of collection points available may be significant; however it is not readily apparent whether the number of collection points per capita or per unit area is more significant, nor why some systems appear to be outliers of the otherwise apparent trends. Index Terms—electronics, e-waste, recycling, WEEE

E

I. INTRODUCTION

LECTRONIC waste (e-waste) recycling offers several environmental and health benefits when processing is done responsibly. The controversy around e-waste recycling systems rarely debates such benefits; rather, it focuses on the characteristics of an effective system and whether benefits outweigh costs. Historically, the costs of recycling most electronics have outweighed the value of the scrap materials recovered. Thus, in the absence of legislation, e-waste recycling systems have been limited to private recycling of high-value waste with only limited consumer participation. E-waste recycling systems now exist in many locations worldwide and the amount of related legislation continues to

This work was sponsored by the National Science Foundation Graduate Research Fellowship Program, the MIT-Portugal Program and HP. S. Fredholm, J. Gregory, and R. Kirchain are all members of the Materials Systems Laboratory at the Massachusetts Institute of Technology, Cambridge MA 02139 USA. (email: fredholm; jgregory; kirchain @mit.edu) S. Fredholm is a graduate student in the Technology & Policy Program of MIT’s Engineering Systems Division. J. Gregory is a researcher with the MIT Energy Initiative. R. Kirchain is an assistant professor in the department of Materials Science and Engineering and the Engineering Systems Division of MIT.

increase. The US does not yet have national e-waste legislation, but a sparse patchwork of legislation exists at the state level. This amount of state level activity appears to be rapidly increasing: 79 pieces of e-waste legislation were introduced in 33 states in 2007, compared with 54 bills introduced in 27 states in 2006 [1]. Numerous approaches have been proposed including landfill bans, extended producer responsibility (EPR) and consumer advance recovery fee (ARF) funded recycling systems. With the enactment of the WEEE (Waste Electric and Electronic Equipment) directive in the European Union in 2003, all EU member states are now required to provide an ewaste system. The WEEE directive further mandates that electronics manufacturers, or importers of electronics into the State, finance the transportation and processing of the WEEE. As a result of this directive, there are now many national ewaste systems in existence; however, as will be discussed in this paper, they take many different forms. Given this variety of implementation options, policymakers are left with the question of how to best setup a system for their jurisdiction. They want to know both: What system architecture will drive the most recovery? and What will be the most economically efficient? To answer these questions policymakers need to know how to evaluate the performance of existing systems, and furthermore, how to use this information to design new systems. Our research objective is therefore to develop a methodology for comparing the performance of different recycling system architectures. This paper describes a framework that may be used for such comparisons. This framework builds upon the framework for evaluating the economic performance of recycling systems presented in [2]. In the remainder of the paper we characterize possible architectural options for e-waste recycling systems and then use the framework to examine how the choice of various options correlates with performance in existing systems. The correlation analysis presented will focus on quantity of e-waste collected as a measure of performance. II. CHARACTERISTICS OF RECYCLING SYSTEMS The primary goals of any e-waste recycling system are to collect e-waste so as to divert it from landfill or inappropriate disposal, and process it such that its component materials are recycled. The performance of a recycling system is therefore characterized in terms of both environmental efficiency (e.g. the amount or percentage of waste recovered or reused) and

economic efficiency (e.g. the costs of the recycling system). Ultimately, both environmental and economic metrics are a function of both a given recycling system architecture and the context in which the system exists. Contextual factors include the amount of waste generated, population density, labor rates, trade restrictions and other regulations. As no single architecture can provide the optimal performance in all contexts, the system architecture for a given location should be chosen with respect to the contextual factors of that location. There is a wide range of activities available to system architects to satisfy each of these functions, resulting in a diverse set of potential (and, in fact, extant) system architectures. The primary options available to system architects can be categorized into product scope, collection methods, management structure, and financial structure. Each of these categories will be further defined in the following subsections. In addition to these categories, e-waste recycling systems may differ in their choice of transportation logistics and physical processing methods chosen for reuse, recovery, and recycling of the material. The implications of these transportation and processing decisions are beyond the scope of this work.

jurisdictions require that all electronics retailers collect all ewaste, whereas others only mandate collection of waste from current customers. In Switzerland, all electronics retailers are required to take-back, free-of-charge to the customer, all household waste electronic goods brought to them. In Portugal and the Netherlands, retailers are only required to accept waste items from customers who are either buying a new similar item from that store, or can prove that the waste item was originally purchased in that store.

A. Product Scope The scope of products included in current e-waste systems varies significantly. The WEEE directive of the European Union, defines ‘EEE’ as “equipment which is dependent on electric currents or electromagnetic fields in order to work properly”[3], but is colloquially remembered as “anything with a cord.” Thus, each EU member country must handle all types of WEEE, but may choose to separate certain types of WEEE into different systems. For example, in the Netherlands, ICT-Milieu handles the Category 3, IT and Telecommunications Equipment, while NVMP is responsible for all other categories of Dutch WEEE. In other locations around the world, the scope of e-waste products handled is much smaller. The state of Maine, in the United States, began a system in January 2006 which only collects TVs, computer monitors, and laptop computers.

D. Financial Structure Recycling systems may be financed directly by the government, by the consumers of electronic products, either at the time of product purchase or product disposal, or by the manufacturers of the products. End-of-life (EOL) consumer fees are used in many areas of the United States. Alternatively, locations including California and Switzerland use Advance Recovery Fees (ARFs) to collect money from consumers at the time of the new product’s purchase. Extended Producer Responsibility (EPR) based systems hold manufacturers responsible for financing the e-waste recycling system. In Maine, the manufacturer of each collected e-waste product is documented and manufacturers are then billed for processing their proportionate share of the collected waste. Sweden and Germany instead charge each manufacturer a percentage of the most recent e-waste recycling costs based upon the current sales market share of that manufacturer. Other systems, such as NVMP in the Netherlands, simply tax each electronic item brought into the country for sale [4]. Most existing European collection systems allow multiple producers to share responsibility for their waste, which is referred to as collective producer responsibility (CPR). Thus, rather than holding each producer responsible for only those goods which that producer manufactured, producers may band together in order to process their collective, unsorted goods. Collective producer responsibility organizations can achieve greater financial economies of scale than manufacturers operating individually. Conversely, Maine, in the US, applies the principles of individual producer responsibility, tallying the brand of each waste product collected and then charging the manufacturer for its unit share of the current waste. Any brand documentation or sorting of collected products adds

B. Collection Methods Collection methods offered for household waste are often different from those available to businesses. A system architect may include e-waste collection as a part of regular curbside pickup within a municipality, require consumers to bring their e-waste to designated drop-off collection points, use retail stores for new electronics as collection points, have products shipped back to their original manufacturer, or any combination of these methods. Curbside pick-up of household e-waste is rare, but several existing e-waste systems require each municipality to provide a local collection point. Most often, this results in adding e-waste to the scope of products already collected at existing recycling or transfer stations. Where retailers are used as collection centers, some

C. Management Recycling systems can be run by producers, recyclers, a governmental entity, or third party organizations. System management responsibilities typically include establishment and collection of recycling fees, hiring of transportation logistics firms and processors, certification of processors, and advertising to increase public awareness of the system. Systems often differ with respect to the number of options they provide to those held financially responsible. For example, Sweden requires all logistics and processors be hired through El-Kretsen, whereas Germany has over 20 system managers each choosing their own logistics and processing providers.

cost to a collection system, but incentivizes design for recycling by holding producers individually accountable for the end-of-life treatment of their own products. III. A FRAMEWORK FOR COMPARING RECYCLING SYSTEMS We propose organizing the characteristics of recycling systems into the framework outlined in Figure 1. This framework groups system characteristics and performance metrics into one of 3 categories: System Architecture, Context, and Performance. System Architecture describes the design characteristics of the e-waste system. Context describes characteristics of the geo-economic landscape in which the system operates. Contextual factors are constraints on the ewaste system design, and unlike system architecture, cannot be directly modified. Finally, performance metrics are used to evaluate the system. A successful e-waste recycling system will achieve environmental goals with economic efficiency. Therefore, our performance metrics include both an analysis of system costs, and the quantity of goods processed.

proper handling. Unlike Switzerland, Sweden and the Netherlands are a part of the European Union and therefore must use the electronics producers, rather than the consumers, as the financiers of their systems. The Netherlands’ ICTMilieu is similar to Switzerland’s SWICO in that they both only handle a fraction of electronic waste product types, whereas El-Kretsen in Sweden handles all types of WEEE products. El-Kretsen is also known for annually collecting the greatest mass of WEEE per capita in the world [5]. California, Maine and Maryland are the first US States to establish statewide e-waste recycling systems and make available their performance data. The scope of waste collected in each of these states is considerably smaller than that collected by any of the European countries studied. We collected data describing the context, system architecture, and performance of each system from a variety of sources. Contextual data was found in [6]-[10]. System architecture and performance data for the European systems was collected from [4], and [11]-[15]. For the US systems, this data was obtained from [2], and [16]-[18]. Selections of this collected data are shown in Figure 2. Using this table, we examine the differences in system performance relative to system architecture and contextual factors for insight into the influence of system architecture options. In this paper we specifically focus our analysis on the environmental metric of quantity of e-waste collected by each system as an illustration of the utility of the framework. This method of comparison provides insight into the relative performance of systems and allows system architects to incorporate new information into the design of their systems. V. CASE STUDY ANALYSIS

Figure 1. An outline of the framework proposed for comparing recycling systems. The characteristics of a recycling system are organized into 3 categories: system architecture, context, and performance.

IV. CASE STUDY BACKGROUND A case study comparing electronics recycling systems in three European countries and three American states was carried out to test the framework and compare the context, architecture, and performance of systems on separate continents. The European countries in the case study are Switzerland, Sweden, and the Netherlands and the American states are California, Maine, and Maryland. We chose this set of jurisdictions in order to demonstrate the variety of architectural options in use and because data on their performance was readily available. Switzerland’s SWICO is the oldest national e-waste system in the world, having begun operation in 1994. The SWICO Recycling Guarantee is supported by the voluntary participation of electronics manufacturers. Member companies agree to add a standard ARF, visible to the consumer, to each of their products sold in Switzerland, and return all e-waste collected to SWICO for

Before comparing performance metrics of each system, it is important to note some of the differences between each system’s architecture and context as shown in Figure 2. Of the systems being compared, California’s system covers by far the largest population, yet has a lower than average population density. The area of both California and Sweden is also an order of magnitude larger than that of the other systems. Differences in average wages between jurisdictions can also influence a system’s economic performance. At over 25 US Dollars per hour, Switzerland pays workers a much larger average wage than the other jurisdictions studied. With respect to system architecture, the scope of products collected by the American systems is much smaller than that of the European systems. California and Maine both limit their collection to display devices (TVs, monitors and laptop computers), while Maryland adds desktop computers to this list. The European systems studied here additionally include all other telecommunications equipment (computer accessories, telephones, fax machines, etc) in their scope. With respect to collection methods, the group of systems evaluated here represents three different options for using electronic retail stores as collection points. Switzerland requires that electronic retailers take-back, free of charge to

Figure 2. A comparison of e-waste recycling systems in 2006 using our proposed framework. California and Maryland were assumed to have the same percentage of Category 3 WEEE in their total WEEE collected as Maine.

Figure 3. Non-retail collection points provided per capita and per unit area. The wide, dark bars correspond to non-retail collection points per thousand people shown on the left axis. The narrow, light bars correspond to nonretail collection points per thousand square kilometers shown on the right axis.

Figure 4. Kg of Category 3 Waste collected per capita

the individual, all e-waste brought to them, whereas the Netherlands only requires this for customers purchasing new

items, and the other systems do not systematically use the retail stores at all. Sweden’s El-Kretsen, which does not use retail stores as collection points, provides the largest number of non-retail collection points. The number of collection points is an important architectural choice as it determines the level of convenience provided to individuals eligible for using the e-waste system. Research has shown that the more convenient recycling is, the more likely people are to recycle [19]. Figure 3 shows the correlation between the architectural choice of number of collection points with both population and geographic area. While Sweden provides the greatest absolute number of collection points, Maine provides the largest number of collection points per person, and Switzerland provides the greatest number per unit area. Having examined contextual and architectural differences between the systems, we can begin to examine system performance. A simple, and commonly used, metric of environmental performance is mass collected per capita. Figure 4 shows this metric for each of the six systems in the case study. In order to compensate for differences in the scope of products collected by each system, only the mass of products which belong to the EU-defined Category 3 WEEE, IT and Telecommunications Equipment, was plotted. The Netherlands’ ICT-Milieu only collects Category 3 equipment and thus, this category was chosen as the greatest common scope of each system. This category includes notebook and desktop computers, monitors, printers, other computer accessories, telephones and fax machines. Figure 4 demonstrates that all existing systems, even Switzerland’s SWICO which has been in operation since 1994, are continuing to increase the average amount of ewaste they are collecting per person. To determine if this result is an artifact of increasing efficiency of the system or due to increasing consumption and disposal of electronic

Figure 5. Kg of Category 3 WEEE collected per PCs in Use 3 and 5 years earlier. The number of PCs currently in use is an indicator of the quantity of electronics which will be retired in the future. In A) the mass of category 3 WEEE collected each year is divided by the quantity of PCs estimated to have been in use 3 years earlier. In B) a 5 year time delay is applied. Both A) and B) show the same general trends within each system.

materials, we compare the mass collected per amount of electronics usage in each jurisdiction in Figure 5. “PCs in use” refers to the total number of personal computers, new and old, that have not yet reached their end-of-life, in a given country during a particular year. The rate of growth in computer usage per capita over time is similar in all jurisdictions [20]. A lag between growth in usage of electronic equipment and retirement of that equipment is expected. Therefore, Figure 5 presents the mass of Category 3 WEEE collected by each system investigated normalized by the quantity of PCs in use in each jurisdiction both 3 years and 5 years earlier. Notably, both figures show comparable trends, indicating that WEEE mass, at least within this context, is growing roughly in step with the rate of consumption of new electrical and electronic equipment. Unlike Figure 4, Figure 5 shows that collection in the oldest system, Switzerland, has reached a plateau relative to the amount of WEEE arising. The next oldest system, present in the Netherlands, continues to increase only slightly. Sweden’s collection rate however, appears to still be growing rapidly. With only two years of collection data currently available, little can be observed with respect to the American systems other than the fact that they appear to be increasing collection relative to computer use as well. Figure 6 shows additional metrics for collection rates as

Figure 6. Kg of Category 3 WEEE per capita collected from non-retail collection points vs. the number of non-retail collection points. In (A), the number of collection points is normalized to population, and in (B) points are normalized to area. Multiple points attributed to the same system refer to different years of operation.

related to the number of collection sites available. In both cases the number of sites refers to the number of non-retail sites available. In some areas, including Switzerland and the Netherlands, retail stores are used as additional collection points; however, the quantity of retail sites being used as additional collection points is unknown. In the Netherlands, only 5% of the e-waste treated is collected at non-retail collection points [4]. In Switzerland, the non-retail collection points have contributed up to 42% percent of the total e-waste collected each year [12]. Retail sites are not a significant source of collection for any of the other systems presented in Figure 6 and therefore the number sites counted below are assumed to collect 100% of the waste for the remaining systems. Figure 6(a) normalizes the number of non-retail collection sites available to population, whereas Figure 6(b) normalizes sites to geographic area. Both compare these normalized values with the mass of waste collected per capita. In both cases there appears to be an upward trend with significant outliers. In Figure 6(a), Maine, with a large number of points per capita, yet relatively small mass collected, appears to be an outlier of the general trend. In 7(b), Sweden’s large mass collected relative to its available points per unit area appears to be an outlier. Further work is needed to determine the significance of the apparent trends and reasons why some systems do not follow the trends.

VI. CONCLUSION

[9]

The system architecture and contextual factors vary significantly between locations of e-waste systems. As a result, there are more differences between most existing systems than there are similarities. This paper has presented a uniform framework for comparing the system architecture, context, and performance of different e-waste recycling systems. The results of applying this framework to a collection analysis show that while all systems examined appear to still be increasing their mass collected per capita each year, the mass collected per electronic item at end of life may have reached a plateau in the older systems. Furthermore, as a predictor of mass of e-waste collected per capita, the number of collection points available may be significant; however it is not readily apparent whether the number of collection points per capita or per unit area is more significant, nor is it readily apparent why some systems appear to be outliers of the otherwise apparent trends. Additional data on these and other systems, specifically including the number of retail collection points in each system, is necessary to better understand the perceived trends and outliers. The large variation in existing system characteristics limits opportunities to isolate case study variables as necessary to gain insight into their effects on system performance. In order to best understand how choices in system architecture affect the performance of the system, the knowledge gained through case study data collection must be abstracted to a model in which individual options can be exercised independently. The case studies completed here can be used to inform such a model. Research in this area is presented in [21].

[10]

REFERENCES [1] [2]

[3]

[4]

[5]

[6] [7] [8]

J. Gast. Making a Mark, in E-Scrap News. 2008. p. 24-28. J. R. Gregory and R. E. Kirchain. "A Comparison of North American Electronics Recycling Systems." in Proceedings of the 2007 IEEE International Symposium on Electronics & the Environment, 2007, pp. 227-232. Available: http://ieeexplore.ieee.org/xpls/abs_all.jsp?isnumber =4222839&arnumber=4222888&count=55&index=48 The European Parliament and the Council of the European Union, "Directive 2002/96/EC of the European Parliament and of the Council of 27 January 2003 on waste electrical and electronic equipment (WEEE)," 2003. Available: http://www.epeat.net/Docs/EU%20WEEE%20Directive.pdf National Electronics Product Stewardship Initiative (NEPSI), "Financing And Infrastructure Model Characteristics," Center for Clean Products and Clean Technologies, University of Tennessee, Knoxville 2002. Available: http://www.cleanproduction.org/library/EPR_dvd/Matrix_NEPSI.pdf DTI Global Watch Mission, "WEEE recovery: the European story," UK Department of Trade and Industry June 2006. Available: http://www.europeanleadfree.net/SITE/UPLOAD/Document/WEEE_DT I_Mission.pdf Central Intelligence Agency, "The 2008 World Factbook," 2008. Available: https://www.cia.gov/library/publications/the-world-factbook/ US Census Bureau, "Area and Population," in State and Metropolitan Area Data Book: 2006: US Census Bureau, 2006, p. 3. Available: http://www.census.gov/prod/2006pubs/smadb/smadb-06tablea.pdf Eurostat, "Population change: absolute numbers and crude rates," in Population and social conditions: European Commission, 2007. Available: http://epp.eurostat.ec.europa.eu/

[11] [12]

[13] [14]

[15]

[16]

[17]

[18]

[19]

[20] [21]

LABORSTA, "37 - Recycling Wages," in Yearly Statistics: 5B Wages in Manufacturing, B. o. Statistics, Ed.: International Labour Office, 2005. US Department of Labor, "Electrical and Electronic Equipment Assemblers Mean Hourly Wage," in November 2004 State Occupational Employment and Wage Estimates, B. o. L. Statistics, Ed.: US Census Bureau, 2004. Available: http://www.bls.gov/oes/2004/november/oessrcst.htm El-Kretsen, "Annual Report of 2006," El-Kretsen, Stockholm, Sweden, 2006. Available: http://www.el-kretsen.se/upload/English/Documents/ Newsletters_reports/Annual_Report%202006.pdf SWICO Recycling Guarantee, "Activity Report 2006," Swiss Association for Information, Communication and Organizational Technology, Zurich, 2006. Available: http://www.swico.ch/filesforweb07/swico_jahresbericht_2006.pdf ICT Milieu, “Facts and Figures,” 2007. Available http://www.ictoffice.nl/index.shtml?ch=MIL&id=4534 Ökopol GmbH - Institute for Environmental Strategies, "The producer responsibility principle of the WEEE Directive," Ökopol GmbH, The International Institute for Industrial Environmental Economics, and Risk & Policy Analysts, 19 August 2007. Available: http://ec.europa.eu/environment/waste/weee/studies_weee_en.htm J. Huisman et al, "2008 Review of Directive 2002/96 on Waste Electrical and Electronic Equipment (WEEE)," United Nations University, Bonn, Germany 05 August 2007 2007. Available: http://ec.europa.eu/environment/waste/weee/studies_weee_en.htm J. R. Gregory and R. E. Kirchain. "A Framework for Evaluating the Economic Performance of Recycling Systems: A Case Study of North American Electronics Recycling Systems." Environmental Science and Technology., submitted for publication. Northeast Recycling Council (NERC), "Setting Up & Operating Electronics Recycling/Reuse Programs: A Manual for Municipalities & Counties," April 2002. Available: http://www.nerc.org/documents/survey/index.html Technology Administration and Office of Technology Policy, "Recycling Technology Products: An Overview of E-Waste Policy Issues," D. o. Commerce, Ed., July 2006. Available: http://www.technology.gov/reports.htm H. Nixon, O. A. Ogunseitan, J.-D. Saphores, and A. A. Shapiro, "Electronic Waste Recycling Preferences in California: The Role of Environmental Attitudes and Behaviors," in Proceedings of the 2007 IEEE International Symposium on Electronics & the Environment, 2007, pp. 251-256. Available: http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=4222892 Euromonitor International. “Personal computers (PCs) in use: Euromonitor International from International Telecommunications Union/national statistics.” Euromonitor International, 2008. J. Dahmus. S. Fredholm. E. Olivetti. J. Gregory. R. Kirchain. “Modeling the Economic and Environmental Performance of Recycling Systems.” in Proceedings of the 2008 IEEE International Symposium on Electronics & the Environment, 2008, to be published