Feasibility of Using Radio Frequency Identification to Facilitate Individual Producer Responsibility for Waste Electrical and Electronic Equipment
Summary
Regulatory measures that hold producers accountable for their products at end of life are increasingly common. Some of these measures aim at generating incentives for producers to design products that will be easier and cheaper to recover at the postconsumer stage. However, the allocation of recovery costs to individual producers, which can facilitate realization of the goals of these policies, is hindered by the practical barrier of identification and/or sorting of the products in the waste stream. Technologies such as radio frequency identification (RFID) can be used for brand or model recognition in order to overcome this obstacle. This article assesses the read rate of RFID technology (i.e., the number of successful retrievals of RFID tag data [“reads”] in a given sample of tagged products) and the potential role of RFID tags in the management of waste electrical and electronic equipment (WEEE) at current levels of technical development.
We present the results of RFID trials conducted at a civic amenity site in the city of Limerick, Ireland. The experiment was performed for fixed distances up to 2 meters on different material substrates. In the case of white goods (i.e., large household appliances), a 100% read rate was achieved using an RFID handheld reader. High read rates were also achieved for mixed WEEE. For a handheld scan of a steel cage containing mixed WEEE, read rates varied from 50% to 73% depending on the ultrahigh frequency (UHF) metal mount tag employed and the relative positioning of the tags within the cage.
These results confirm that from a technical standpoint, RFID can achieve much greater brand or model identification than has been considered feasible up to now, and thus has a role to play in creating a system that allocates recovery costs to individual producers.
Introduction
In recent decades, mass consumption and a tendency for electronic products to have shorter lifetimes have led to a massive increase in waste electrical and electronic equipment (WEEE) (Gupta 1995). A measure that could help decrease the environmental impacts of WEEE would be to extend producer responsibility throughout the entire product life cycle (Fishbein 1998). In 2001 the Organization for Economic Cooperation and Development (OECD) defined extended producer responsibility (EPR) by stressing that producer responsibility should be extended to the postconsumer stage of a product's life cycle. In addition to the basic definition, the OECD enlarged EPR to “providing incentives to producers to incorporate environmental considerations in the design of their products” (OECD 2001, 9).
However, in order for there to be incentives for improved design, producers should be held individually responsible for financing the recovery of their own products, a concept known as individual producer responsibility (IPR) (Mayers et al. 2011). If the producer pays only for waste recovery of the products that it has produced, there is indeed an incentive to design products that are easier and cheaper to recover (Dempsey et al. 2010). Without IPR, the link between design effort and reduced recovery costs would be lost and others would benefit from one's efforts since costs would be based on averages across products.
Governments around the world have embraced the concept of EPR in environmental policy to pursue life cycle thinking and preventative legislation (Atasu and Van Wassenhove 2012; Tojo 2004). Regulations such as the European Union's (EU's) WEEE Directive (Directive 2002/96/EC) introduced the notion of producer responsibility in an effort to create an incentive for producers to design products that are easier to recover at their end of life (EOL). In practice, however, such measures have had limited success. One obstacle is the difficulty of identifying brands in order to allocate costs to producers when used products enter the waste stream. This has led producer responsibility organizations (PROs) operating in EU member states to charge producers for recovery costs based on their market shares (i.e., the firm's share of sales of new products within the relevant product category) instead of their return shares in the waste stream (i.e., the firm's share of products recovered for recycling). Equally important, PROs charge average recovery costs per tonne across brands, hardly an incentive to design products that would be easier to recover (Dempsey et al. 2010).
Given recent improvements in radio frequency identification (RFID) technology, the aim of this research was to assess the read rate (i.e., the number of successful RFID tag reads in a given sample of tagged products) that can be achieved in WEEE management. RFID directly provides an alternative solution for brand identification (e.g., compared to physical sorting) and can therefore facilitate cost allocation based on actual product recovery costs or characteristics. RFID can indirectly enable recovery of desired substances (e.g., gold on circuit boards) by recognition of products known to contain those substances. It can also allow identification of products that contain hazardous materials that must be properly disposed of. In addition, it can provide information on the product's quality and material components, which may help determine the optimal mode of recovery. Of course, this could require proper sorting after identification.
The technical feasibility of RFID has been challenged (CECED 2003; Thomas 2009), since according to the literature the reader needs to be closer than 1 meter (m) to the tag on the product and the amount of metal surroundings need to be limited (Thomas 2003).1 The read rate of RFID reported in the current article was not measured in a lab, but in a realistic EOL waste environment. The experiment was carried out for fixed distances up to 2 m on different material substrates in an EOL environment for mixed WEEE and white goods. Our study concentrated on business-to-consumer (B2C) WEEE, because this is the only category that contributes to regulatory collection targets. Very little business-to-business (B2B) WEEE is reported to the EU, as per existing regulations (Eurostat 2009).
The article is structured as follows. In the next section we review the literature on the history of EPR/IPR and the challenges faced, the role of RFID in the management of WEEE, and how RFID can facilitate EPR/IPR. A description of the existing recovery system for B2C WEEE in Ireland is provided. An assessment of the potential read rate achievable for RFID tags within a mixed WEEE and white goods environment is conducted using a field experiment. We present the results of preliminary lab tests conducted for tag selection, including the rationale for selecting certain tags for the field trials. After reporting the results of an RFID trial at a collection site in Limerick, Ireland, we present a summary and our conclusions.
Literature Review
Extended Producer Responsibility and Individual Producer Responsibility
Since the term EPR was first used in a report for the Swedish Ministry of the Environmental and Natural Resources in 1990 (Lindhqvist and Lidgren 1990), environmental regulations across the globe dealing with EOL products have been based on the EPR concept (Lifset 1993; Lindhqvist 2000). A 1991 packaging ordinance in Germany incorporated the concept of EPR by shifting the cost of collecting, sorting, and recycling used packaging from municipalities to producers, holding the latter responsible for managing packaging waste (Fishbein 1998). The EU Packaging Directive (Directive 94/62/EC) followed Germany's packaging ordinance and promoted the concept of EPR in packaging (without actually requiring it) to the member states.
Environmental regulations on WEEE in many countries have embraced the notion of EPR. Specifically, in the EU, Directive 2002/96/EC imposes the concept of EPR for WEEE on the member states. However, the implementation of IPR through the WEEE directive has encountered two major challenges. First, the directive allows mixed WEEE in the same waste flow, making segregation by product category and by producer very difficult. Second, despite the fact that the WEEE directive introduces producer responsibility in an effort to create incentives for producers to design products that will be easier to recover, there are hurdles to implementation. For example, in most EU member states the PROs charge producers based on their present market share for the collective financing of the recycling systems. Consequently the benefit of any individual investment in designing easy-to-recover products will be shared by all producers, killing all the good intentions of the directive (Dempsey et al. 2010).
IPR can be implemented by holding each producer responsible for financing the recovery of its own products either through an individual recycling system or by joining a PRO that organizes and finances collection and recycling (Mayers et al. 2011). In the latter instance, this can be facilitated if there is brand identification. Although there is some skepticism as to whether IPR is feasible in practice (Sander et al. 2007), according to Dempsey and colleagues (2010), examples of systems that integrate elements of IPR abound, including the Specified Home Appliances Recycling Law (SHARL) and the PC Recycling System in Japan, the WEEE recovery systems in Maine and Washington in the United States, and operations in the Information and Communication Technology Milieu system in the Netherlands up to 2003. An alternative system to allocate costs to producers that does not require brand recognition or sorting is suggested by Mayers and colleagues (2012). All these systems aim to reduce the flow of waste to landfills and incentivize producers to design products that can be more efficiently recovered (Mayers et al. 2005). Argentina has proposed a WEEE law that relies on IPR (Arditi 2011), according to which producers can join collective systems or develop individual recovery systems, together with a differentiation in the fee per category of WEEE.
Radio Frequency Identification Technology
Clearly, brand identification in the collected WEEE could facilitate recovery cost allocation to the producers, and it is here that RFID, with its multiple implementations, has a role to play. Wal-Mart mandated that its top 100 suppliers put RFID tags on pallets and cases shipped to its stores as of 2003. The U.S. Department of Defense, Target, Albertson's, and Best Buy also use RFID technology (Delen et al. 2007). The Massachusetts Institute of Technology (MIT) Auto-ID Labs focus on developing low-cost RFID technology and global standards to track objects automatically and ubiquitously.
The potential of RFID technology for facilitating WEEE management has already been explored to some extent. Binder and colleagues (2008) explored the potential of RFID for improving the waste and resource management system in Switzerland and found that implementing RFID in WEEE management could increase the recycling rate. Saar and colleagues (2004) investigated the benefits from identification of product codes for recycling cell phones. Identification of the brand and model can be linked to a database that provides information on how to dismantle the product and the location of hazardous materials. Abdoli (2009), Kahhat and colleagues (2008), Kulkarni and colleagues (2005), and Thomas (2009) described the advantages of using RFID in recycling and refurbishing enterprises to increase the recycling rate and select the most appropriate EOL option. The use of RFID has been found to increase collection rates if coupons are provided to consumers who drop their used products into a recycling bin where the RFID reader reads the product's characteristics (Thomas 2007). Delen and colleagues (2007) found that RFID technology could improve supply chain visibility by benefiting inventory management and asset utilization since the inventory information (e.g., on the shelf, at the point of sale) could be distributed in real time. Zhou (2009) established the benefits of RFID technology for supply chain processes, including more efficient control through increased information accuracy, better knowledge of customer behavior, improved tracking of quality problems, and tighter management of perishable items and returns.
Beyond its application to EOL, RFID offers considerable benefits throughout the life cycle of electrical and electronic equipment (EEE). In the context of the adoption of an RFID waste management infrastructure, Thomas (2009) estimated labor cost savings for computer recyclers of between US $0.07 and US $0.14 per recycled computer.2 According to Thomas (2009), the benefit of RFID technology is less about saving money for recyclers, and more about reducing environmental impacts through increasing reuse and recycling. The cost of RFID tags has also been reduced significantly in the last decade. For example, in 2002 the cost per tag was $0.3 to $0.5 (Saar and Thomas 2002); in 2009 it was $0.1 (Thomas 2009). As Thomas (2007) pointed out, an RFID system may also be cost effective for removing hazardous and recovering valuable components from the waste stream.
Dempsey and colleagues (2010) and Thomas (2003) reviewed the different technologies used for the identification of EOL products, comparing optical barcodes with chipless tags and RFID tags. Barcoding technology is associated with the following advantages: (1) it is the most widely used tagging technology, so there are high volumes in the market; (2) the cost is lower; (3) the tag is easy to incorporate; and (4) it does not have to contact the reader and can be located a long way from it. The drawbacks of this technology are (1) barcodes can become obscured, dirty, or damaged, making them difficult or impossible to read; (2) they require line of sight, which makes identification of mixed loads of WEEE difficult; (3) they are read only; and (4) no data modification is possible. Many of the benefits of barcoding technology compensate for the shortcomings of RFID technology and vice versa, while chipless tag3 technology lies somewhere in between.
RFID trials have also been conducted in other research fields. Bjorninen and colleagues (2011) developed and tested flexible passive ultrahigh frequency (UHF) tags for water bottle item-level identification. Their experiment, which included multiple bottle configurations, found no “dead” zones, despite the fact that nearby tags affected the tag backscatter radiation pattern. Medeiros and colleagues (2011) developed and tested a passive UHF RFID tag for suitcase tracking and identification at airports: a lab conveyer belt test was conducted and 100% read rates were achieved when suitcases passed in front of two RFID antennas. Polycarpou and colleagues (2011) investigated a UHF system within a health care environment. The trials examined patient identification, real-time location service of medical assets, and drug inventory control. De Blasi and colleagues (2010) conducted a performance evaluation of UHF tags in the pharmaceutical supply chain. These studies demonstrated that the use of far-field4 UHF tags combined with a suitable reading system configuration was able to guarantee performance, except in the presence of considerable quantities of metal.
Challenges of Radio Frequency Identification Technology
The application of RFID technology to WEEE management presents technological challenges (Dempsey et al. 2010; Thomas 2009), the main ones being hard-to-define read conditions and the large amount of metal nearby. Where read conditions are not clearly defined, the position of the reader cannot be guaranteed and thus the identification of items in mixed WEEE is not clear. The metal in the surroundings significantly reduces readability, especially of UHF tags (Derbek et al. 2007; Dobkin and Weigand 2005; Singh et al. 2009). Another limitation is the RFID read range (Thomas 2003). Extremes of temperature and humidity can also impede tag detection and undermine the reliability of the system (Ahson and Ilyas 2008). In sum, the literature indicates that physical limitations require that the reader be closer than 1 m to read a RFID tag on a product and that the amount of metal in proximity be limited (Thomas 2003).
Another technical challenge that hinders the use of RFID in WEEE management is that tags need to be read regardless of their orientation (Thomas 2007). Delen and colleagues (2007) conducted a study using actual RFID data that revealed a common problem: a valid tag passed within the prescribed range of an RFID reader but the reader failed to read it. The aforementioned limitations were observed in pilot studies involving reverse logistics for WEEE. For example, the “Multi Life Cycle Centre (MLCC) for electronic and electric equipment” project conducted in 2007 explored the use of RFID as a mechanism to optimize the flow of EEE product-related information once it had entered the waste stream (Knoth 2007)5. Initially, problems arose due to antennas detuning from their operating resonant frequency, most likely as a result of radio frequency variations, signal strength losses due to metal proximity, harmonic effects, and signal reflection. These had a detrimental influence on the tag/reader communication distance. Second, problems occurred in the labeling of containers. Passive RFID tags could not be read when mounted on metal. Similar problems occurred in the tagging of WEEE due to the large percentage of metal components in the products. With an overall read rate of 30%, the study concluded that the introduction of RFID was not feasible for B2C WEEE management. Further studies investigating the tagging of EEE durables have also reported reflection, absorption, and detuning effects caused by metal (Derbek et al. 2007; Singh et al. 2009).
Fortunately there has been significant research and development in the field of RFID tags and many of the problems experienced in the past have been resolved. One improvement is the emergence of metal-mount tags (Odin Technologies 2008; RFIDTags.com. 2010). Initial developments altered the tag design to incorporate a spacer to shield the tag antenna from the metal, in the process creating bigger tags. New techniques focus on a specialized antenna design that utilizes the metal interference and signal reflection. The metallic surface is used as the ground plane of the antenna or as an energy-improving reflector for a longer read range than similar-size tags attached to nonmetal objects (Rao et al. 2008).
A range of information technology (IT) asset management tags specifically designed for use with EEE has recently become available. These reportedly provide a long read range for their size and a suitable form factor for application to EEE products (Odin Technologies 2008). However, performance measurements that are included in data sheets (e.g., read distance, orientation sensitivity, and transfer rate, gathered within anechoic chambers) bear little relation to real-world operating conditions, characterized by multiple tags within a densely populated waste environment. Anechoic chambers are designed to stop reflections of either sound or electromagnetic waves and prevent exterior sources of interference.
A study conducted by the European Committee of Domestic Equipment Manufacturers (CECED) in 2003 concluded that no tagging system is available now or in the foreseeable future to meet the current operational requirements for disposal and logistics of WEEE (CECED 2003). However, long-term benefits were considered probable in the future (20 years plus), assuming a substantial initial investment. Of the RFID solutions, low frequency systems were identified as the best suited for WEEE given their superior read performance within the presence of metals when compared to the other RFID frequency classes. In the current article we demonstrate that UHF systems today provide an effective read range and material independence suitable for deployment in mixed WEEE environments.
The contribution of this article is to show that a read rate higher than 30% can now be achieved using RFID in a harsh EOL waste environment as opposed to lab conditions. It also questions some of the technical challenges of RFID presented in the literature, specifically that the reader needs to be closer than 1 m and that the amount of metal surroundings needs to be limited. Moreover, the article assesses these technical challenges for different tag types. First, the read rate and orientation sensitivity of five UHF metal-mount tags at fixed distances up to 2 m on different material substrates are analyzed in preliminary lab experiments. Second, the read rates of the best performing metal-mount tags from these experiments are investigated in an EOL white goods environment and in a densely populated mixed WEEE environment. The experimental work completed for the case study investigates the capability of RFID tags to technically facilitate EOL product identification within the context of a system that would properly allocate recovery costs to individual producers, thereby approaching IPR and providing incentives for producers to design products that are less costly to recover. This has only been speculated upon in the literature (Dempsey et al. 2010; Thomas 2009).
The Existing Take-Back System for Business-to-Consumer Waste Electrical and Electronic Equipment in Ireland
Under B2C compliance systems in Ireland, a consumer has numerous options regarding the disposal of used products. These include (1) returning the WEEE directly to a civic amenity (CA) site; (2) returning the product to a retailer when purchasing a new, equivalent product; or (3) returning products via events such as curbside collection and on selected days each year (“open days”) when old equipment is collected at designated areas (although the WEEE collected this way ultimately ends up in a CA site).
The 45 CA sites with WEEE collection facilities in Ireland collected more than 15,000 tonnes in 2010. Limerick (city and county) has three CA sites, which collected more than 10% of the total WEEE, approximately 1,500 tonnes in 2010. Some 933 tonnes were collected at the Mungret facility, making it one of the ten biggest facilities. At the Mungret site, operated by Indaver, B2C WEEE is divided into five categories: cathode ray tubes (CRTs), IT equipment,6 large domestic appliances (LDAs), cold LDAs,7 and mixed WEEE. Both LDAs and cold LDAs can be placed under the umbrella of white goods. No other categories of WEEE are recovered. LDAs represent the highest tonnage of WEEE recovered (76%), and IT equipment the lowest (1.6%). CRTs account for a significant tonnage (21%), but are considered “historic WEEE,”8 and therefore fall outside the scope of this case study. Currently, in Ireland's waste management system, the WEEE cost allocation is based on EPR, where producers pay an average recovery cost based on market share, with no brand differentiation of the recovered WEEE.
Where products are returned to a retailer or collected via curbside collection/open days, brand identification by means of RFID is relatively easy to accomplish. Reflection, absorption, and detuning effects do not pose a significant problem for these take-back cases since product handling occurs at an item level (assuming EOL identification through an RFID handheld reader). However, in most cases individual items that are returned directly to the CA site or are collected during curbside collection and open days are not scanned, but rather are stored in open-top steel cages, with contents placed in the cage as they are returned by consumers (as mixed WEEE). It is assumed that all items are RFID tagged and readable at an item level. Given the random nature of the placement in storage cages, some unreadable tags are inevitable. As a result, EOL identification is more difficult at the CA site, hence brand identification in this take-back case is the focus of the experiments reported here.
Radio Frequency Identification Experimental Trials at the Limerick Civic Amenity Site
Preliminary Lab Experiments
- Tag: tag size, tag type, read distance, tag frequency, orientation sensitivity, tag-to-tag interference
- Reader: reader type (handheld), reader orientation, read time, reader power
- Environment: cage type, number of contents, content positioning, material dependency, temperature
Reader and Tag Type Selection
Taking typical product characteristics and the operating environment at CA sites into account, the tag size, tag type, tag frequency, and reader type could all be predetermined from analysis of the RFID manufacturer datasheets prior to actual testing. UHF RFID technology was chosen over the other frequency types, as this provides the most suitable read range (3–5 m) for the operating environment. All the tags selected were passive UHF EPC Gen2 RFID metal-mount tags, which are specifically designed to withstand a variety of environmental operating conditions, such as those associated with WEEE. It was decided to use a leading industry standard handheld reader (Motorola MC9090) for the tests, as handheld readers permit an increased level of flexibility compared with their fixed counterparts. Metal-mount tag types were perceived as the best choice for tagging appliances due the heterogeneous mix of materials that constitute white goods and mixed WEEE. The experiments were conducted by the University of Limerick.
The tags investigated in the lab are listed in table 1. While SARC-3 and SL tag types operate within both U.S. and EU frequency ranges, IT asset management and the Emerson and Cuming (E&C) tags operate only within the defined EU frequencies.
| Tag type | Manufacturer | Dimensions (mm) | Operational band (MHz) | Cost (>10,000 piece pack) |
|---|---|---|---|---|
| SARC-3 | Alien Technology | 100 mm × 14 mm × 14 mm | 840–960 | $1.00/piece |
| E&C | Emerson & Cuming | 66 mm × 25 mm × 5 mm | 865–868 | $3.00/piece |
| SL High Temperature G (SL) | LPR Global | 40 mm × 32 mm × 8 mm | 860–960 | $4.50/piece |
| SL Mini Metal A (SL mini) | LPR Global | 24 mm × 19 mm × 4.6 mm | 860–960 | $2.00/piece |
| IT Asset Management | Emerson & Cuming | 36 mm × 16 mm × 3 mm | 865–868 | $3.50/piece |
Notes
- One millimeter (mm) = 10-3 meters (m, SI) ≈ 0.039 inches; MHz = megahertz.
As seen in table 1, SARC-3 tags are the largest tags, but also the cheapest, costing approximately $1 each for purchases of 10,000 units or more. E&C tags are more expensive, but also the least rigid of the five tag types. The flexible nature of E&C tags makes them more suitable for appliances with curved surfaces. SL tags have a rigid exoskeleton that can withstand severe outdoor conditions, wide temperature variations, and greater shocks than the other tag types listed, but they are also the most expensive, with a cost of $4.50 per unit. The IT asset management tags are physically the smallest tags and are specifically designed for use in asset management of IT equipment during the life of the product.
The use of such tags in these tests may be open to criticism, since the tags were new and therefore had not undergone functional degradation during the appliance's use—it cannot be guaranteed that tags more than 5 years old will exhibit the same read rates as new tags. This is one limitation of the experiment, since in order for the supply chain to benefit fully from RFID use, the RFID tag must be on the product when it is sold, remain on it to gather information (RFID tags have limited write capability) throughout its life, and be read at EOL (Saar and Thomas 2002). Furthermore, variance and inconsistency between the same types of tags were not considered, as each tag was tested prior to the field experiments to ensure readability.
Preliminary Lab Experiments
As explained in the previous section, the tag size, tag type, tag frequency, and reader type were duly selected. Tags were placed on each piece of WEEE. For the lab tests, it was possible to study the readability of the tags taking into account the tag type and the maximum read distance for two different materials. Reader power was maintained at 2 watts (W) effective radiated power throughout the experiments (the maximum power permitted in Europe). The material substrates examined were plastic and steel. The handheld reader was fixed 1 m above ground level, directly facing the tagged substrate for these tests.
The results of the lab experiments are shown in table 2. SARC-3 tags exhibited the largest read distance, providing read ranges in excess of 3 m when applied to both plastic and steel substrates. SL tags provided the second longest read distance (2 m) when applied to the same materials. They exhibited better orientation sensitivity (+90°, −90°) when compared with SARC-3 tags (+45°, −60°). The orientation sensitivity and maximum read distance were determined from the experiments for each specific combination of tag type and material. The IT asset management and SL minitags had a limited read range on both metal and plastic, and therefore were not used for the field trials.
| Tag type | Material | Maximum read distance (m) | Orientation sensitivity (°) |
|---|---|---|---|
| SARC-3 | Steel | 3 | +45°, −60° |
| Plastic | 3.5 | +45°, −60° | |
| E&C | Steel | 2 | +45°,−45° |
| Plastic | 2 | +45°,−45° | |
| SL | Steel | 2 | +90°,−90° |
| Plastic | 2 | +90°,−90° | |
| SL mini | Steel | 1 | +45°,−45° |
| Plastic | .5 | +45°,−45° | |
| IT asset management | Steel | 1 | +45°,−50° |
| Plastic | .5 | +45°,−50° |
Notes
- m = meters; ° = degrees.
Case Study: Radio Frequency Identification Experimental Trials at the Limerick Civic Amenity Site
The field trials were conducted at the Mungret CA site on the outskirts of Limerick, where mixed WEEE and white goods entering the site are separated before their respective storage and loads are weighed using a weighbridge. White goods are housed in 20- to 40-foot containers, while mixed WEEE is stored in open-top steel cages before transport to the recycling facility. No records of individual appliances at an item or batch level entering the system are maintained.
Field Experiments
Field trials were conducted by the University of Limerick, with the support of the Limerick County Council, Indaver, and the European Recycling Platform (ERP). The RFID tags were preprogrammed by the authors, giving each product a unique identification number; in the future, this operation will be conducted by the product manufacturer. When the worker holding the RFID reader scans around the cage, the serial number of each product that is read appears on the monitor of the RFID reader. In the future, reading of the serial number will allow the worker at the CA site to connect to the “EPC Discovery Services”9 and to source the manufacturer of the product and access relevant information specific to that product, such as the bill of materials.
The experiments concerning white goods focused on six common household appliances (washing machines, dryers, refrigerators, freezers, ovens, and dishwashers), while the trials for mixed WEEE dealt with the actual appliances contained in the WEEE cages during the trial days. These included household appliances (e.g., vacuum cleaners, kettles, toasters, coffee machines), sports and entertainment equipment (e.g., electric golf caddies, video cassette recorders, radios, CD players, speakers), and gardening and do-it-yourself equipment (e.g., nail guns, jigsaws). Although IT and lighting equipment normally have designated cages, the cage selected for the experiment contained eight keyboards, six printers, two game consoles, an external hard drive, and three types of lighting fixtures. A photograph of the mixed WEEE items in the cage is shown in figure 1.
Experimental Setup for White Goods
At the CA site, white goods are usually stacked side by side and sometimes stacked vertically in 20- to 40-foot containers. For the experiment on the white goods, the independent variables that influenced tag readability were tag type, product material, read distance, orientation sensitivity, and power attenuation. For this experiment the read distance was restricted by the storage technique employed in the 40-foot container and the power attenuation was maintained constant (2 W effective radiated power).10
The three best performing tags listed in table 2 were placed in various orientations on the front, side, and back panel of the six aforementioned appliances (washing machines, dryers, refrigerators, freezers, ovens, and dishwashers) in a 20-foot container. For the handheld sweep, the appliances were scanned in a consistent sweeping motion while walking up and down the container, with no restriction imposed on the orientation of the reader. Read distance was limited to approximately 2 m, with three scanning sweeps taken to ensure a reliable comparison.
Experimental Setup for Mixed Waste Electrical and Electronic Equipment
Mixed WEEE is stored in metal cages (weighing 40 kilograms [kg]) at the Mungret CA site. Unlike white goods, where product casings are largely made of aluminum or steel constituents, mixed WEEE casings are predominately plastic. Due to the heterogeneous nature of mixed WEEE, there is no structure to the stacking arrangement—the appliances are haphazardly placed in each cage. In the mixed WEEE, a steel collection cage as shown in figure 2 was used (standards for cages are currently in development). The number of products and the product materials are constantly changing due to the nature of mixed WEEE collection (on the trials dates, a WEEE cage containing 78 appliances was examined, with a mixed proportion of metal and plastic appliances). Another limitation of this experiment is that the product positioning is not taken into account due to the almost infinite number of relationships between the different products.
All mixed WEEE appliances (78 in total) were tagged with the three different tags and then put back into the cage. A scanning sweep was conducted (i.e., the cage was scanned in a consistent sweeping motion from the front, from the back, and from the top of the cage for a 60-second period). The items contained in the cage were then reshuffled (i.e., taken out and put back in a different order) and a further scanning sweep was administered to obtain an average readout for each tag type. This experiment was repeated once more to increase the accuracy of the average readability obtained.
Results
With a time limit of 60 seconds imposed for the scanning sweep, 100% readability was achieved for the white goods. The handheld reader checked an inventory of 25 white goods in a 20-foot container with 100% accuracy, regardless of the type of metal-mount tag used (SARC-3, SL, E&C). The accuracy is attributed to the bulk size and uniform stacking technique (i.e., side by side in a container) employed for this WEEE category. The experimental results for the mixed WEEE trials are given in table 3. It should be noted that only 55 of the 78 appliances were tagged for each of the scanning sweeps carried out in the case of the SL tags (due to their high cost).
| Scanning sweep no. | Tag type | No. of tagged appliances | No. of products identified (handheld configuration) | Read rate percentage |
|---|---|---|---|---|
| 1 | SARC-3 | 78/78 | 57 | 73% |
| 2 | SARC-3 | 78/78 | 52 | 67% |
| 3 | SARC-3 | 78/78 | 55 | 71% |
| 1 | E&C | 78/78 | 43 | 55% |
| 2 | E&C | 78/78 | 46 | 59% |
| 3 | E&C | 78/78 | 39 | 50% |
| 1 | SL | 55/78 | 52 | 67% |
| 2 | SL | 55/78 | 41 | 53% |
| 3 | SL | 55/78 | 45 | 58% |
Conclusion
This study assessed the read rate that can be achieved in WEEE management using RFID technology. The results of the lab experiments are shown in table 2. SARC-3 tags exhibited the longest read distance, providing read ranges in excess of 3 m when applied to both plastic and steel substrates. SL tags recorded the second longest read distance (2 m) when applied to the same materials, and exhibited improved orientation sensitivity (+90°, −90°) when compared with SARC-3 tags (+45°, −60°). The orientation sensitivity and the maximum read distance were determined from the experiments for each specific combination of tag type and material. The IT asset management and SL minitags had limited read ranges on both metal and plastic and were therefore not used for the field trials.
The results indicate that current RFID technology can support EEE product identification in the B2C WEEE management domain, given the high read rates achieved. From a technical standpoint, RFID can contribute to the identification of brands and therefore support a system that allocates recovery costs to individual producers, thereby facilitating IPR, even for mixed WEEE. In the white goods WEEE category, 100% readability was accomplished, which meets the requirements of the ideal IPR scenario with 100% brand recognition. Complete identification can be attributed to the bulk size, storage, and stacking techniques employed for EOL white goods. Currently, integration of RFID within the white goods sector appears to be at the concept level, with the development of the internet of things and smart appliances.11 Full-scale adoption and standardization within the white goods industry would facilitate the identification of brands for IPR following the incorporation of embedded UHF RFID tags.
In the case of mixed WEEE, 100% brand identification was not achieved in the field trials. Different tags on different products were read in each scanning sweep. Mixed WEEE was placed into steel WEEE cages in a random fashion, leading to a vast number of possible product placement combinations that could not be accounted for in the testing. When tags are densely surrounded by products in an almost enclosed environment, the antenna's behavior is altered, reducing the power linked to the integrated circuit on the tag, which consequently undermines the probability of a positive readout. Read rates achieved varied from 50% to 73%, depending on the UHF metal-mount tag employed and the relative positioning of the tags within the cage. Overall, higher read rates than in the past were obtained, suggesting that further investigation for the incorporation of RFID tags in EEE for EOL management is warranted.
These experiments have specific limitations. As indicated, the tags used were new and therefore had not undergone functional degradation during the use of the appliance. Further experimentation is required in order to see if the read rates are affected by the age of the tags. The experiment was not carried out in a lab environment, but in a real EOL waste environment. While this allowed us to establish the practical viability of the technology, it did not allow accurate control of various factors affecting the read rate. Future research should also be structured as a controlled experiment in order to study the impact on read rates of various critical technical specifications, such as multidirectional tag orientations, the distance from the reader, extremes of temperature and humidity, the number of products in the cage, and product positioning within the cage. The results also differentiated between the reading reliability of tag designs from different manufacturers. Further research is required to determine the most suitable tag for each appliance so that identification at EOL can be maximized while ensuring required functionality throughout earlier phases of the life cycle.
UHF RFID is poised for significant growth (Bridge 2007). Analysis conducted under the Bridge Project predicts the deployment of 170,000 passive RFID readers, with a total of 3 billion tags installed within the next 5 years. By 2022, an estimated 6 million readers are anticipated to be in use, with 86 billion tags purchased annually. Standardization is required to avoid creating perverse incentives for producers to use unreadable tags. Reading reliability must be prioritized as a short-term research goal, ensuring standardized performance between tag batches as well as reader technology variations. The economic feasibility of RFID offers another avenue for future work. IPR that incorporates RFID will require industry agreement on the cost, technology, and privacy and security concerns before the technology infrastructure can be rolled out.
Acknowledgments
This work was supported by the Environmental Protection Agency (EPA) of Ireland through the Science, Technology, Research and Innovation for the Environment (STRIVE) Programme, and also by ZeroWIN (www.zerowin.eu), a project supported by the European Commission's 7th Framework Programme (grant agreement no. 226752). The authors wish to thank the European Recycling Platform and Limerick County Council, as well as the employees of Indaver for their assistance during the trials.
Notes
Biographies
Maurice W. O'Connell is a postgraduate research student in the Electronic and Computer Engineering Department at the University of Limerick, Limerick, Ireland.
Stewart Hickey is a postdoctoral research fellow in the Electronic and Computer Engineering Department at the University of Limerick.
Maria Besiou was a postdoctoral research fellow at the Social Innovation Centre at INSEAD, Fontainebleau, France, at the time the article was written. She is currently assistant professor of logistics at the Kuehne Logistics University, Hamburg, Germany.
Colin Fitzpatrick is a lecturer in the Electronic and Computer Engineering Department at the University of Limerick.
Luk N. Van Wassenhove is a professor of technology and operations management at INSEAD.
