Recycle Of Weee

Resources, Conservation and Recycling 45 (2005) 60–69

The recovery of recyclable materials from Waste Electrical and Electronic Equipment (WEEE) by using vertical vibration separation Nusruth Mohabuth ∗ , Nicholas Miles School of Chemical, Mining and Environmental Engineering, University of Nottingham, Nottingham NG7 2RD, UK Received 10 December 2004; received in revised form 7 February 2005; accepted 14 February 2005 Available online 11 April 2005

Abstract In this article, the vertical vibration technique is described as it is used to separate a mixture of plastic and bronze in water. When a mixture of two equally sized granular materials is vertically vibrated, they often separate into two distinct layers. Plastic and bronze were used to mimic the situation of Waste Electrical and Electronic Equipment (WEEE) materials. At low frequency, a bronze rich layer is formed on top of a plastic layer, while at higher frequency the bronze remained sandwiched between two layers of plastic. A similar result was obtained when equivalent size shredded WEEE materials were vibrated. These results were compared with those in the plastic–bronze mixture. The WEEE mixture separates into a copper rich layer on top. The observations and possible mechanisms of this separation are discussed. © 2005 Elsevier B.V. All rights reserved. Keywords: Vibrated bed; WEEE; Separation process; Density segregation

1. Introduction Waste Electrical and Electronic Equipment (WEEE) consists of a wide range of electronic devices that can be classified into three groups, namely white goods, brown goods and Information Technology (IT) scraps (Sch¨afer et al., 2003). White commodities are mainly ∗

Corresponding author. E-mail addresses: [email protected] (N. Mohabuth), [email protected] (N. Miles).

0921-3449/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.resconrec.2005.02.001

N. Mohabuth, N. Miles / Resources, Conservation and Recycling 45 (2005) 60–69

61

Nomenclature a f g Γ ω

amplitude (m) frequency of vibration acceleration due to gravity (m/s2 ) ratio of vibrational to gravitational acceleration angular frequency (s−1 )

large household appliances, such as fridges and cookers with a high metal content and because of their size they have a relatively high collection quota of around 90%. In 1998, an arising of 392,000 tonnes was recorded in the UK (ICER, 2000). The second group is brown goods, which are household electrical entertainment appliances (CD Players, TVs, camcorders, radios, etc.) of which there were 80,000 tonnes produced in 1998 in the UK. Of these goods, only 50% were collected (ICER, 2000). Brown goods easily fit into the household bins and are thus more conveniently disposed of with other domesticwaste. As for the IT scraps, 357,000 tonnes were produced in 1998 and only 26% were collected (ICER, 2000). Due to rapid technology changes, computers are now featuring as a significant component in WEEE. It was estimated that in the United States, 20 million computers became waste in 1998 within an overall E-waste of 5–7 million tonnes (Puckett et al., 2002). WEEE makes up about 4% of the European municipal waste and is estimated to be increasing three times faster than municipal waste (ICER, 2000). The WEEE Directive 2002/96/EC will be enacted in September 2005. The focus of this directive is about prevention of WEEE and its reduction through increasing reuse, recycling and recovery of materials. The directive has set a target to recover 75% and to reuse and recycle 65% by an average weight per appliance by December 2006. Therefore, there is an urgent need to recover the maximum amount of recyclable material from WEEE, which typically consists of ferrous metals (47%), plastics (22%), glass (6%) and non-ferrous metals (4%) (ICER, 2000). To recover valuable materials from WEEE, the feed material initially needs to be liberated by mechanical processing so that the desirable fractions can be separated. Hammer mills and shredders are the most commonly used communication devices to reduce WEEE to finer fractions, thus, liberating the phases. Typical methods used to separate these liberated materials include manual sorting, magnetic separation, eddy current separation and air table sorting. Sch¨afer et al. (2003) reported that these techniques have shown limited efficiency due to enormous loss of materials (Sch¨afer et al., 2003). For example, an eddy current separator would separate non-ferrous metals. However, other metals can also be influenced by the magnetic field and affect the purity of the end product. Since there is strict specification for the reuse and recycling of the materials, efficient sorting is of great importance (Bledzki et al., 2002). WEEE materials, when reduced to finer grain sizes consist of non-uniform particles with varying properties such as size, density, shape and particle resilience. Differences in particle properties can be used to propagate a separation in vibrated system. In fact, wet and dry vibration techniques have been widely used for separation for many years by the mineral

62

N. Mohabuth, N. Miles / Resources, Conservation and Recycling 45 (2005) 60–69

industry. Here, techniques such as jigging have been used to process coal, iron ore and gold (Nesbitt et al., 2004). With the varying materials constituent and grain sizes of the WEEE, vertical vibration was considered as a potential separation technique. It is well known that when granular materials are subjected to vibration segregation occurs. This phenomenon has been studied by a number of researchers but the mechanisms have not yet been fully understood (Rosato et al., 1987; Jullien and Meakin, 1992; Knight et al., 1993; Duran, 2000; Duran et al., 1993,1994). The most common segregation is size segregation, which is known as the “Brazil Nut Effect” (BNE), where the larger size particles appear on top of smaller ones when subjected to vibration. Williams (1976) proposed four mechanisms for size segregation: (i) trajectory segregation, (ii) percolation of fine particles, (iii) rise of coarse particles on vibration and (iv) elutriation segregation. The Brazil Nut Effect has been studied by using a single large particle, most commonly known as the intruder particle, in a fine granular bed. It has been also observed that in some cases, a Reverse Brazil Nut Effect occurs. Shinbrot and Muzzio (1998) reported that at high amplitude in a vibrated bed a large heavy particle moves to the top of the bed but an equally large light particle sinks to the bottom of the bed. However, this phenomenon remains unexplained. When the bed vertical vibration is above a threshold amplitude (1.2 times the gravitational acceleration), the bed is fluidised and the granulated particles move into a convection flow. Convection was first reported by Faraday (1831) and was found to be an important driving mechanism in the segregation process (Shinbrot and Muzzio, 1998). It was also one of the major causes for density segregation (Burtally et al., 2003). This occurs due to differences in the momentum between different density particles (Akiyama et al., 2000). Burtally et al. (2002, 2003) have reported that a mixture of equal sized particles of bronze and glass spheres segregate into two clear-cut layers under vertical vibration in the presence of air due to differences in density. It was found that the denser materials, in this case bronze, appeared on top of a layer of glass spheres, the lower density material. However, sandwich separation and inverse segregation have also been observed depending on the frequency and acceleration of vibration. During vertical vibration, the bed is also influenced by the horizontal component of the fluid flow, which causes the bed to tilt (Faraday, 1831). This is known as Faraday tilting. The medium in which the granular materials are vibrated also influences the separation process. Even though most of the studies carried out so far were in air, the medium can be replaced by other fluids. Leaper et al. (2005) reported that replacing air by water at ambient temperature increased the viscosity of the medium by 50 times. As a result, it is expected to separate larger particles of about 1 mm diameter. These particles are seven times larger than that would be separated in air (Leaper et al., 2005). Using water as the separation medium also helps to eliminate the electrostatic effect, which is commonly observed in dry separation after prolonged vibration. This may affect the quality of separation and should be eliminated (Leaper et al., 2005). Electrical and electronic equipment are identified as a high copper containing mass stream (Sch¨afer et al., 2003). In this paper, bronze and plastic particles were used to mimic the most common components of WEEE. Bronze has similar properties as copper and plastics reflects other lighter materials that may be present in the WEEE. Recently, Burtally et al. (2003) separated a mixture of bronze and glass in the size range 90–125 ␮m diameter. However in this paper, coarser particle sizes, ranging from 250 to 400 ␮m, were used since

N. Mohabuth, N. Miles / Resources, Conservation and Recycling 45 (2005) 60–69

63

it was considered that WEEE materials would not be shredded to such fine size. To enhance the separation water was used as the medium for separation.

2. Materials and methods The present study was carried out using vertical vibration apparatus similar to the one in the Nottingham Granular Dynamics Group, Physics and Astronomy, University of Nottingham, UK (Burtally et al., 2003). The apparatus consisted of two loud speakers vertically mounted in a wooden framework, which were connected to a pair of accelerometers and an amplifier in order to control the vibration. The whole assembly was attached to a massive concrete block to prevent horizontal motion during vibration. The experiments were carried out in a rectangular vessel of thickness 10 mm, height 40 mm and width 40 mm. The vessel was aluminium framed and glass lined so that the behaviour of the materials could be easily viewed. The glass box was then fixed by screws to the bottom of the metal frame in between the two loud speakers. To investigate the separation process, three sets of experiments were conducted. The first two sets were carried out using combination of mixed plastic particles of mean density, ρp = 1600 kg/m3 , and bronze spheres of density, ρb = 8900 kg/m3 . The mean diameter of the mixture was 325 ␮m (size range, 250–400 ␮m). In the first set of experiments, a 75% plastics:25% bronze by volume mixture was used and in the second a 50:50% plastic–bronze mixture. To study the separation of copper in WEEE materials, a third set of experiments was conducted on a sample of shredded WEEE materials consisting mainly of circuit boards and electrical wiring. The WEEE scraps were obtained from Master Magnet Ltd. and were in a wide range of shapes and sizes, ranging from 100 to 150 mm. They were gradually reduced to smaller size fractions of 50, 25 and 10 mm. It should be noted that no glass was present in the shredded mixture. This material was further shredded, sieved to produce a fraction in the size range 250–400 ␮m, washed and dried before being placed into the cell. For each set of experiments, the bed depth was maintained at about 20 mm to allow for comparison. The experiments were conducted under atmospheric pressure with vibrational frequency ranging from 10 to 120 Hz and the dimensionless vibrational acceleration, Γ (Γ = aω2 /g, where a is the amplitude of the oscillation, g the acceleration due to gravity and ω = 2πf, is the angular frequency) in the range 2–10. The experiments were carried out in water at ambient temperature with the rectangular vessel initially being filled with the dry materials followed by water. The vessel was then shaken to mix the materials and to liberate air bubbles. The level of water was adjusted to the top and the cell attached on the loud speaker. The materials were expected to be well mixed before the start of the experiments but perfect mixing could not be guaranteed due to the sedimentation of the materials in water when shaken to achieve mixing. Interestingly enough, this is mimicking the jigging process used in the minerals industry where the high-density mineral, in this case bronze, appears at the bottom with the lower density one on top (plastic). The process was observed through the glass wall of the vessel and the time taken to achieve a visible separation was noted. At different frequencies, f, and Γ , the separation behaviour was observed and a graph of Γ against f was plotted to show the boundaries between different behaviours.

64

N. Mohabuth, N. Miles / Resources, Conservation and Recycling 45 (2005) 60–69

3. Observation of separation This section describes the results of the separation of the different mixtures when subjected to vibration and the effects of varying Γ and f. 3.1. Separation of plastic and bronze In most cases, the mixture separated into two layers of bronze and plastic. At some combinations of f and Γ , the separation was very good, with a pure bronze layer on top of a pure plastic layer, with very little contamination. This kind of separation is known as a “complete separation”. An example of such separation is shown in Fig. 1a–f for f = 40 Hz and Γ = 8.3. Once the cell was subjected to vibration, a bronze rich layer was formed in about 4 min. At the initial stage of separation clusters of bronze were formed and coalesced until they gathered into a tilted layer sandwiched between the plastic particles (Fig. 1b). The upper plastic layer moved in an anticlockwise convection current. This layer gradually became thinner as the plastic passed through the bronze layer until it eventually resulted in a bronze rich layer on top of a plastic layer (Fig. 1f). It should be noted that at the end of the experiment the bed with the bronze on top layer, remained tilted. Fig. 2 shows the behaviours of the 75% plastic:25% bronze mixture at different frequencies, f, and Γ values. Boundary lines are drawn to distinguish between the different regions. The line ‘α’ in Fig. 2, shows the boundary at which separation started to occur within 120 s of vibration. Below this line, in region A, there was no separation or separation is very slow.

Fig. 1. The behaviour of 25:75% mixture of bronze and plastic, mean size 325 ␮m at Γ = 8.3 and f = 40 Hz. The vibration was stopped and photographs were taken after (a) 0 s, (b) 30 s, (c) 90 s, (d) 150 s, (e) 210 s and (f) 250 s of vibration.

N. Mohabuth, N. Miles / Resources, Conservation and Recycling 45 (2005) 60–69

65

Fig. 2. Schematic diagram showing the behaviour of a 25:75% mixture of bronze and plastic, mean size 325 ␮m as a function of frequency and Γ . Region A: no separation, region B: area of complete separation, region C: tilted sandwich separation, region D: dome shape sandwich separation.

Above ‘α’, the sandwiched bronze rich layer was formed after a few seconds only. The higher the frequency and Γ values, the faster the formation of the bronze layer. However ‘complete separation’ with the bronze on top occurred only in region B. As for regions C and D, the bronze layer remained sandwiched between the two plastic layers. In region C, the sandwich bronze layer, remained tilted, but at very high frequency and Γ values, as in region D, Faraday tilting was no longer observed. Instead, the bed was dome shaped which may be due to the balance created by the convection current observed on both sides of the cell. When the composition of the bronze and plastic was changed to a 50:50% mixture the separation process occurred within 2.5 min. It proceeded in the same way as the 75% plastics:25% bronze mixture, with the formation of a tilted sandwich layer which then turned into a bronze rich layer on top. This mixture consisted of smaller amount of plastic particles than the previous one. Therefore, the separation was quicker since less plastic particles need to go through the bronze layer to achieve complete separation. Complete separation with a bronze rich layer on top occurred at a low frequency. 3.2. Separation of WEEE materials After considering the behaviour of bronze and plastic under vibration, it was important to study the separation of real materials. It was unclear how the varying composition of the WEEE material and the irregular nature of the shredded particles would influence the separation process. Experimental procedures were similar to those used for the bronze and plastic mixture. An example of the behaviour of the WEEE material under vibration in water at f = 32 Hz and Γ = 8.3 is shown in Fig. 3a–i.

66

N. Mohabuth, N. Miles / Resources, Conservation and Recycling 45 (2005) 60–69

Fig. 3. Behaviour of shredded WEEE materials, size range 250–400 ␮m, under Γ = 8.3 and f = 32 Hz. The pictures form a time sequence from (a) to (i) and correspond to 0, 10, 20, 40, 60, 80, 110, 130 and 190 s of vibration.

Separation occurred within 3 min of vibration and the results were found to be analogous to the bronze and mixed plastic system. The bed moved into a wave form pattern and after 20 s of vibration the metal fragments, which consisted mainly of copper, started to gather up (Fig. 3c). Eventually, a tilted bed was formed with a copper rich layer sandwiched in between two layers of non-metallic materials, though some copper were still in the lower layer (Fig. 3d). The bed moved in an anticlockwise convection current forcing the top materials into the lower layer. This caused the bed to take a pile shape until it equilibrated into a tilted bed (Fig. 3g). The convection current continued in the anticlockwise direction until all the upper materials were forced down through the copper rich layer (Fig. 3h). The direction of the convection current is as illustrated in Fig. 4b. This convection current eventually carried the remaining copper fragments into the top layer (Fig. 3i).

N. Mohabuth, N. Miles / Resources, Conservation and Recycling 45 (2005) 60–69

67

Fig. 4. Convection pattern.

4. Discussion The above observations suggest that bronze and plastic materials can be separated into two distinct layers using the vertical vibration technique. ‘Complete separation’ with a bronze rich layer on top occurred at low frequencies and low Γ . Two different mixture ratios were used with separation occurring for both mixtures. It was observed that even when the high-density materials were as low as 25% of the mixture, it still separated. This mixture was found to take about 2 min longer to separate than a 50:50% mixture. Burtally et al. (2003) separated a mixture of glass and bronze, where the shape of the particles were spherical. In this study, even though the mixed plastic particles were of irregular shape and the bronze uniform spheres good separation was still observed. However, the bed remained tilted throughout the separation process whereas in the study of Burtally et al. (2003), the bed was horizontal after vibration. This may be explained by considering that the plastic materials have a lower density than glass and may get more easily entrained in the convection current. This separation was carried out in water, which is more viscous and has an important effect on the bed shape. Convection may also be influenced by the relative density of the materials. Above the line ‘α’ in Fig. 2, the particles start to fluidise and a convection current was set up. Fig. 4a and b shows the pattern of this convection current. The top and bottom layer moved in an opposite direction to the sandwiched layer. Convection is considered to be one of the main mechanisms contributing to separation. The convection current was set up within each individual layer rather than in the bed as a whole. It was observed that the convection entrained the heavier materials and they eventually gathered up into a layer. The convection current was observed to be dominant in the sandwich and the top layers. After complete separation, the convection current remained in the same direction as shown in Fig. 4b and the two layers did not remix. Since the particles are of different densities, the separation process may be explained by the effect of the momentum and drag forces on the particles during vibration. Williams (1976) stated that the bigger the particle the further it travels through a fluid. Thus, the distance travelled by the particles does not depend on the materials but on their size. Similarly, the momentum of a particle is proportional to the mass of the particle (Akiyama et al., 2000). Therefore, the higher density particles are thrown higher than the lower ones during vibration and are further carried by the convection current. As a result, the lower density

68

N. Mohabuth, N. Miles / Resources, Conservation and Recycling 45 (2005) 60–69

particles get more easily carried into the convection current, and thus, causing separation between the two different density materials. The higher density size particles have greater momentum when it is vibrated at low frequency and high amplitude for a given Γ . Thus, ‘complete separations’ with the bronze on top were observed only at low frequencies as shown in Fig. 2. When comparing the result of WEEE materials with the plastics and bronze mixture, it was observed that the behaviours are analogous. The separation occurred with the heavier layer on top of the lighter one. The top layer was a copper rich layer, but it also contained a small fraction of other metals that were present in the WEEE materials. The lower layer consisted mainly of plastic and other lighter materials of non-uniform shape. At the end of the separation, the bed exhibited the Faraday tilt similar to the one in the previous experiments with pure materials and the convection current followed the same pattern as the one discussed above. However, before stabilizing into the tilted shape the bed followed a wave form, and then produced a peak in the middle of the bed. This may be due to the fact that the copper gathered up in equal proportions on both sides of the bed thus forming two individual convection currents in the copper rich layer. During the experiments, some very fine dust-like particles remained suspended in the water during vibration. These materials did not take part in the separation process. 5. Conclusion Using the vertical vibration technique to separate materials, it was observed that for both cases, the bronze–plastic combination and the WEEE material, the metal components can be separated. When particle sizes are larger, water can be used as a medium to enhance separation. Convection and differences in momentum between the different density particles were the major driving force for separation. The mixture of 25% bronze showed that even low concentration of bronze separated successfully under vibration. Thus, this could form the basis of a practical method for separating copper from WEEE. Further studies will be carried out to give more detailed quantitative analysis for the separation. Work is currently underway to develop a laboratory-based continuous system. Acknowledgements The authors are thankful to Helena Webster and Philip Hall for their help and support. N. Mohabuth was supported by the Developing Solutions Research Scholarship, University of Nottingham. References Akiyama T, Yamamuro K, Okutsu S. A solid–solid extraction. Powder Technol 2000;110:190–5. Bledzki AK, Sperber VE, Wolff S. Methods of pretreatment. In: La Mantia F, editor. Handbook of plastic recycling. Shorpshire, UK: Rapra Technology Limited; 2002. Burtally N, King PJ, Swift MR. Spontaneous air-driven separation in vertically vibrated fine granular mixtures. Science 2002;295:1877–9.

N. Mohabuth, N. Miles / Resources, Conservation and Recycling 45 (2005) 60–69

69

Burtally N, King PJ, Swift MR, Leaper M. Dynamical behaviour of fine granular glass/bronze mixtures under vertical vibration. Granular Matter 2003;5:57–66. Duran J. Sands, powders, and grains: an introduction to the physics of granular materials. New York: SpringerVerlag Inc; 2000. Duran J, Mazozi T, Cl´ement E, Rajchenbach J. Size segregation in a two-dimensional sandpile: convection and arching effects. Phys Rev E 1994;50:5138–43. Duran J, Rajchenbach J, Cl´ement E. Arching effect model for particle size segregation. Phys Rev Lett 1993;70:2431–6. Faraday M. On a peculiar class of acoustical figures. Phil Trans R Soc Lond 1831;50:299–340. ICER (2000) UK Status Report on Waste From Electrical And Electronic Equipment. Industry Council for Electronic Equipment Recycling. London. Jullien R, Meakin P. Three-dimensional model for particle-size segregation by shaking. Phys Rev Lett 1992;69:640–3. Knight JB, Jaeger HM, Nagel RS. Vibration induced size separation in granular media: the convection connection. Phys Rev Lett 1993;70:3728–31. Leaper MC, Smith AJ, Swift MR, King PJ, Webster HE, Miles N, Kingman SW. The behaviour of water-immersed glass–bronze particulate systems under vertical vibration. Granular Matter 2005:7. Nesbitt AB, Breytenbach W, van der Plas PJ. Characterisation of the pulse wave of an inline pressure jig in a near density application; 2004 [available at http://www.gekkos.com/Papers/Pulse wave of an InLine Pressure Jig.pdf (7 October 2004)]. Puckett J, Byster L, Westervelt S, Gutierrez R, Davis S, Hussain A, Dutta M. Puckett J, Smith T, editors. Exporting harm: the high-tech trashing of Asia; 2002 [available at http://www.crra.com/ewaste/ttrash2/ttrash2/ (7 October 2004)]. Rosato A, Strandburg KJ, Prinz F, Swendsen RH. Why the Brazil nuts are on top: size segregation of particulate matter by shaking. Phys Rev Lett 1987;58:038–40. Sch¨afer T, van Looy E, Weingart A, Pretz T. Automatic separation devices in mechanical recycling processes. In: Proceedings of the International Electronics Recycling Congress, 13–15 January; 2003. Shinbrot T, Muzzio FJ. Reverse buoyancy in shaken granular beds. Phys Rev Lett 1998;81:4365–8. Williams JC. The segregation of particulate materials. A review. Powder Technol 1976;15:245–51.