recycling of printed circuit boards

Milling and Classification of the Printed Circuit Boards (PCBs) for Material Recycling C. Eswaraiaha,*, R.K. Sonib, R. Sakthivela a

CSIR-Institute of Minerals & Materials Technology, Bhubaneswar-751 013, India b Indian School of Mines, Dhanbad-826 004, India

Table of Contents ABSTRACT ..................................................................................................................2 1. Introduction...............................................................................................................2 2. Materials and methods...............................................................................................5 2.1. Materials .............................................................................................................5 2.2. Methods ..............................................................................................................5 3. Results and discussion ...............................................................................................7 3.1. Size reduction & Liberation .................................................................................7 3.2. Air classification .................................................................................................8 3.2.1. Air flow rate measurements ..........................................................................9 3.2.2. Classification process ................................................................................. 10 3.2.3. Species/Components analysis...................................................................... 11 3.2.4. Separation efficiency .................................................................................. 12 4. Conclusions.............................................................................................................. 13 Acknowledgements ...................................................................................................... 14 Nomenclature............................................................................................................... 14 References ................................................................................................................... 14

* Corresponding author. Tel.: +91-674-2379263; Fax: +91-674-2581637. E-mail address: [email protected] (Dr. C. Eswaraiah) 1

ABSTRACT Recycling of waste electric and electronic equipment is an emerging issue due to its hazardous nature as well as valuable materials associated with the waste. It is important to identify the appropriate environment-friendly processes to recover valuables and safe disposal of associated hazardous materials. In this work, two-stage crushing process is employed to liberate the valuables and also to obtain the suitable fragment size distribution. From the size reduction process, it is observed that the decrease in particle size decreases the metal content. Thus, it is important to know the optimum particle size for better liberation. This work deals with the separation of metals and plastics from the milled PCBs using circulating air classifier. Size analysis of milled products and fractions obtained from the air classifier were carried out. The amount of metals and plastics present in each size fraction were estimated. It is concluded that the main metals enriched was found to be in the size fraction of -1.6+0.5 mm. The optimum airflow rate of 60 m3/h or say, optimum superficial airflow velocity 13160 m/h found to be the best to recover the metal-rich fraction from mixture of PCBs. The presented work studies the strategy for separation of metal rich and plastic material from e-waste for their safe and hazardous free disposal, as some of the metals in e-waste such as Pb, As, Cd are poisonous in nature. Keywords: recycling

Printed circuit boards; crushing; air classification; waste treatment;

1. Introduction 2

The robust growth of new technologies is producing short-lived electric/electronic goods, which leads to the generation of obsolete waste in large scale. The generation of e-waste is rapidly growing and therefore leading attraction towards the efficient separation of its components and their safe disposal. These goods are brought destined for landfills, incinerators or hazardous waste exports. (Kang and Schoenung, 2005; Lee et al., 2007; Zhang and Forssberg, 1999; Cui and Forssberg, 2007; Bi et al., 2007; Wang et al., 2005). According to Bertram et al. (2002), the generation of waste electric and electronic equipments (WEEE) is around 7 kg per capita per annum in Europe. It is reported that the amount of waste expected to increase by at least 3–5% per annum (Cui and Forssberg, 2003). The usual compositions of PCBs are non-metals such as epoxy, glass fiber and resin greater than 70%, Copper 17%, solder 4%, iron and ferrite 3%, nickel 2% and other metals less than 1% (Goosey and Krishnan, 2003). The recycling of electric/electronic waste is difficult as the PCBs is hard to pulverize (Kinoshita et al., 2003). Therefore, most of the PCBs waste has been disposed as landfill dumps. Various chemical processes have been investigated with the reagents such as aquaregia, tetrabromo ethane (Zong et al., 2002). Furthermore, the chemical processes generate hazardous waste, which requires additional treatment before disposal. Since, the most of the chemical processes are not environment friendly and economically viable, Some other processes of recycling of printed circuit boards use pyrometallurgical (Felix and Riet, 1994; Shichang et al., 1994) or hydro-metallurgical routes (Gloe et al., 1990; Pozzo et al., 1991; Marca et al., 2002), which again generates atmospheric pollution through the release of dioxins and furans or high volumes of effluents (Menad et al., 1998). Mechanical physical and chemical methods are two traditional recycling processes for waste PCBs. Chemical methods mainly include pyrolysis, combustion, hydration,

3

and electrolysis. The major waste of PCBs is processed in backyards or small workshops using primary methods such as open burning and acid washing (Guo et al., 2010). The pyrometallurgical and combustion route generates atmospheric pollution through the release of dioxins and furans (Leung et al., 2007; Wong et al., 2007). During the recycling process of hydration and electrolysis, large quantity of waste acid liquid is produced, which needs to be treated carefully before disposal (Li et al., 2007). On the other hand, in China the fluid bed separation technology is widely used in recycling of waste PCBs due to the lack of effective technologies (Huang et al., 2009). Huge amounts of wastewater get generated during the process, which may contain heavy metals leading to more serious secondary pollution without proper treatment. However recently, mechanical processing of PCBs gradually focused on to recover the valuables associated therein (Eswaraiah et al., 2008; Das et al., 2009; Hall and Williams, 2007; Veit et al., 2005). Some studies reported that the mechanical processing (Zhang and Forssberg, 1997; Veit et al., 2002; Veit et al., 2006) is as an alternative process to concentrate the metals in one fraction, and the polymers, ceramics in another. Moreover, the metal concentrated fraction can be used to electrochemical processes (Scott et al., 1997; Kekesi et al., 2000; Ubaldini et al., 2003) in order to separate the metals among themselves. Mechanical process is mainly used for recycling of PCBs scrap because it can yield maximum material recovery, including plastics (Jarring and Forssberg, 2003). Recently, the recovery of metals and nonmetals from pulverized PCBs by corona electrostatic separation (CES) based on the extreme differences in density and electric conductivity of metallic and nonmetallic materials, has been introduced (Wu et al., 2008). In addition, studies have been reported that the metallic components from waste printed circuit boards were enriched by using a zig-zag classifier (Yoo et al., 2009).

4

Recent studies have demonstrated the feasibility of copper recovery through vacuum pyrolysis and mechanical processing for recycling of waste printed circuit boards (Song et al., 2010; Guo et al., 2011). The present work focused on to develop an environmentfriendly recycling process for separation of metals and plastics from the PCB scrap and their utilization in appropriate applications.

2 Materials and methods 2.1 Materials The metals and plastics in PCBs are present in an embedded form. Therefore, an efficient size reduction process is necessary for liberating metals from the plastics. Around 50 kg of waste PCBs was separated manually from the old obsolete personal computers. The material is first broken into smaller pieces using power hammer and mechanical shearing machine followed by two stage crushing process. The crusher product was used as feed material for air classification process to separate the metals and plastics.

2.2. Methods In this work, the Hazemag impact crusher, hammer mill and air classifier were used for milling and classification of PCBs. Initially, 50 kg of PCBs were broken into small pieces by power hammers and mechanical shearing machine. As mentioned in preceding paragraph, the amount of sample was subjected to the Hazemag impact crusher that has given the 80% passing size (d80) passing size as 8 mm which was too course to be analyzed for metal liberation. A snapshot of used Hazamag crusher is shown in Fig. 1. Eswaraiah et al. (2008) can be referred for the similar size reduction process. with small

5

changes such as replacement of primary crushing stage of hand cutting with the Hazamag crusher in this work and the variation in sieve sizes that were used for size distribution analysis. In Hazemag impact crusher, the samples are broken by the impact of swinging hammers. The crusher has the provision to change the variables such as gap width, rpm and feed rate. The experiments were carried out at different crusher parameters to obtain the better product at preliminary stage of comminution. Further, the Hazemag crusher product was subjected to hammer mill for size reduction to achieve the better liberation and required size for air classification process to separate metal rich and the metal lean fractions. Metal rich and metal lean products of the hammer mill were analyzed with the sinkfloat test for the amount of metals and non-metals in them collectively, irrespective of the type of metal or non-metal present. Sink-float analysis separates the components of a mixture in multiple fractions according to differences in their densities, a detailed description of Sink-float analysis especially in concurrence with present work therefore with possibly similar densities can be found in Eswaraiah et al. (2008). Each product fractions of sink-float test was dissolved one by one in HNO3 solution to account the weightage of metals present in fraction. Paper also details the procedure of subsequent stage of present work therefore determination of individual metals in the different

Comment [RKS1]: present concentration and other things related to sink float analysis/ What were the durations of sink float test, dense media used in that, and solid liquid ratio (or relative amoun of solid required) for preparation of medium? What were the densities of solution taken?

density fraction with the technique of component/species analysis therefore Atomic Absorption Spectrometry (AAS). Air classification is an important unit operation for the separation of dispersed solid particles based on their difference in size and density. The detailed description of design of air classifier and its methodology can be referred from Eswaraiah et al. (2008). AAS was again used for the determination of total amount of metal present in the

6

Comment [RKS2]: how aas was used. give som detail.

classification product metal rich and metal lean fractions, and the variation in collective metal amount in the metal rich fraction was analyzed against the changes in airflow.

3. Results and discussion 3.1. Size reduction & Liberation

Particle size distributions of the PCBs product resulted from the primary Hazemag

Comment [RKS3]: mention the sieve sizes for each step at appropriate places

impact crusher was analyzed and found distribution is shown in Fig. 2. The sieves used for the sieve analysis were having the sizes of 125, 211, 420, 600, 1000, 1400, 1980,

Comment [R4]: Please check whether all the sizes are available in practice.

2800, 3960, 5600, 7920, 11200 µm. It was observed that the d80 passing product size is insensitive with respect to the rpm and gap width. However, the product size influenced more by the crusher feed rate in comparison with the other variables. The d80 passing

Comment [RKS5]: what were the values of fee rates. and how it varied

size of the product from Hazemag crusher was found to be 8 mm, which indicates that the product size distribution could not have well liberated metals at this stage of crushing. The product from the Hazemag crusher at optimized conditions were subjected to hammer milling for further size reduction to achieve the better liberation characteristics. The product of the Hazemag impact crusher was fed into the hammer mill for maximizing the liberation characteristics of metals and plastics. A schematic diagram of hammer mill is shown in Fig. 3. The hammer mill gives the highest throughput rate due to an optimum number of beaters, large effective sieving surfaces or grate gaps and full utilization of the grinding chamber. The collection chamber is detachable from the metal frame, which was clipped with the bag to the downside of the mill, and serves as the dual purpose of product collector and as well as air filters. The ground product was collected from the hammer mill in the collection chamber. The milled products were

7

Comment [RKS6]: what are this optimized conditions?

subjected to sieve analysis to determine the required size for subsequent air classification process. A comparison of the product size distribution obtained at optimized conditions of Hazemag crusher and hammer mill is shown in Fig. 4. It is observed that the d80 passing size found to be 8 mm and 1.7 mm in Hazemag crusher and hammer mill respectively. The hammer milled product size distribution at different feed rates inferred that, as the milling rate increases the residence time of particles in the mill decreases which in turn reduces the relative weightage of broken particles towards lower sizes, as shown in Fig. 5. The probability of particles being exposed to size reduction process decreases with an increase in milling rate, which This consequently increases the d80 passing size of milled product. In addition, it was observed that the variation in the milling rate is significant up to certain feed rate after that there was no effect of feed rate. Therefore, the variation in d80 of product is very small as evident in Fig. 5. The hammer-milled products obtained at optimised conditions were analysed for size and species distribution. The component analysis indicated that metals are mainly enriched in the size range of -1.6+1 mm and -1+0.5 mm size fractions. The grades of the metal enrichment of the above two size fractions were found to be 29.5% and 61.5% respectively. Though the yield of the crushed product in the size fraction of -2+0.5 mm was 55.38%, it covered the 80.81% metals of the total comminuted product. From this, it is also concluded that the prime metal rich fraction of comminuted product was in size range of -1.6+0.5 mm.

3.2. Air classification

A snapshot as well as the schematic diagram of a circulating air classifier used in

8

Comment [R7]: What are optimized conditions

the present study to classify the metal rich and metal lean fraction is shown in Fig. 6. The air classifier used in the present work is different in its design from the design used in Eswaraiah et al. (2008). In the Eswaraiah et al. (2008), the concentration of study is on the column classifier that requires an external blower to flow the air inside the classifier while in the present study, the fan in combination with vanes in the classifier take the air-in through the air inlet section (as shown in the schematic line diagram of a circulating air classifier) which creates an upward flow for the lighter or plastic materials and subsequently, these lighter particles come out with the air from air outlet section. While the heavier metal particles remain in the inner concentric cylinder and slowly comes out from the air inlet section. An air classifier can be controlled and moderated by several parameters such as airflow rate, vane designs, material feed rate and speed of rotating guiding vanes etc.

3.2.1. Air flow rate measurements Prior to the air classification experiments, measurement of airflow rates at different rotational wheel speeds and different guide vane angles were carried out. Angular and radial vane configurations were considered for the measurement of air flow rate. These were measured with the help of two techniques; a turbine flow meter and a traversing pitot tube connected to the micro-manometer were used. The measurements were carried out at the air outlet section, and then converted into superficial air flow rates according to equation 1. Vs  V f A

….(1)

where Vs is the converted superficial air flow velocity, V f is air flow rate measured from pitot tube or turbine flow meter, and A is the cross sectional area of the classifier

9

Comment [R8]: Put the table regarding air flow rate measurement from the PhD thesis here.

outlet cross-section. In the present study, the classifier used is having the cross-sectional diameter equal to 3 inch; and therefore is calculates the cross-sectional area equal to 4.56e(3) . Velocity profiles obtained from pitot tube was nearly axi-symmetric and a typical

Comment [RKS9]: provide evidence

turbulent-like velocity profile was found. The flow rate measurements were carried out over a range of wheel speeds, and the results obtained and compared are shown in Fig. 7. It is observed that there is a great variation in airflow rate as well as in the recirculation pattern with respect to the stationary bottom vane configurations such as angular and radial vanes. This difference in flow rate and recirculation plays an important role while separating metals and plastics. 3.2.2. Classification process A circulating air classifier separates particulate mixtures by utilizing the differences in terminal falling velocities of the material, in the air. The solid mixture to be separated was fed into a stream of upward moving air; the solid particles were carried along or settled against the air stream with respect to the difference between the superficial air velocity and their settling velocity. The particles experience centrifugal, gravity and drag forces according to the intensity of the force field. The optimum airflow rate required to separate metal lean and metal-rich fraction is studied by using circulating air classifier. To find the optimum airflow rate for the separation of metal-rich fraction, experiments were conducted at five different airflow rates such as 30, 50, 60, 70, and 80 m3/h (or superficial air flow velocities such as 6580, 10970, 13160, 15350 and 17540 m/h, respectively). Further, classification products were subjected to sieve analysis and subsequently determined the percentage metal content. The effect of airflow rate on percent metal content inferred that the metal found

10

Comment [RKS10]: mention size of sieves

maximum for the superficial air flow velocity of 13160 m/h. Further increase in flow rate decreases the metal values as evident in Fig. 8. When the superficial air flow velocity is maintained at 13160 m/h, the metals rich collected at the side bottom contained 78.41% metal and the loss of metal content was around 8%. Further, the increase of superficial air flow velocity to 17540 m/h (or 80 m3/h of air flow rate) caused some of metal rich particles reported to metal lean fraction. At higher flow rate, the metal concentration in the underflow of heavy particles is around 11.5% only; which gives an indication that the optimum superficial air flow velocity could be below 17540 m/h. It is also observed from the experiments that the separation is impossible at very low flow rates as well as at very high flow rates. The misplacements observed is due to the entrainment of smaller size metal particles carried along with the larger size plastic particles. However, the small amount of misplacement of metals in plastics as well as plastics in metals may be there all the time. Entrainment of particles may be mainly due to the bypass of airflow within the classifier.

3.2.3. Species/Components analysis

In this work, the estimation of metals and plastics present in various size fractions were carried out by adopting the sink-float method using HNO3 solution as sink-float media. The differential distribution of the metals in the hammer milled PCBs revealed that 80% of the total liberated metal lies in the population of particle size between 1600 μm and 500 μm, whereas float fractions predominate at lower particle sizes as evident in Fig. 9. It was also observed that the liberation of metals and plastics occurred at particle size below 1200 μm. The visual analysis shows that the sink-float method is an effective method to determine the amount of metals and plastics present in the printed circuit

11

Comment [RKS11]: 8. What were the durations of sink float test, dense media used in that, and solid liquid ratio (or relative amoun of solid required) for prepation of medium? What were the densties of solution taken? 9. What was the concentration of HNO3 How the metals who does not dissolve in HNO3 were handles?

boards. The similar pattern of distribution was observed for all hammer milling conditions. The metal particles did not take part in complete breakage because they were not as brittle as epoxy resin and glass. The results of the acid dissolution analysis of PCBs of various fractions obtained from hammer mill shows that the most of the metal contents were concentrated in the coarser fraction. As mentioned earlier, this may be due to metal particles did not participate in the complete breakage. The metal content presented in below 500 μm fraction was considerably less. It is evident that the metal content presented below 63 m size fraction is almost negligible. And as the particle size decreased up to 150 m, plastic content present in the sample increased and after that the plastic content decreased rapidly. This may be due to the large amount of glass fiber and ceramics present at the finer size fraction. Glass fiber and ceramic powder of this size fraction could be used as filler materials.

3.2.4. Separation efficiency The optimum superficial air flow velocity for maximum metal recovery is found to be 13160 m/h. Further, increase in flow rate leads to an increased amount of metal content in the metal lean fraction as misplacement along with the air. The high-flow rate increases upward air velocity, which results in heavier fraction reported to the other side. The flow rate required for separation of metal rich and lean fraction is very high. This is due to varying metal densities and unequal settling properties of the particles. At a high-superficial air flow velocity (17540 m/h), there was no metal content present in the coarse fraction in the particle size below 500 μm. At this flow rate, large amounts of below 500 μm particles were reported to the fine fraction and remaining

12

particles were collected at the coarse fraction. It was observed that at higher flow rate, the metal concentration was more in the fine fraction of the classifier. This might be explained as smaller size metal particles carried along with the larger size plastic particles. The air classification mainly depends upon the size and density of the particle. The separation is easier for particle size around 1000 μm. Despite some losses in metal content, air classifier is operated at 13160 m/h which led to the optimal efficiency for the metal-rich product. The results quantitatively demonstrate the opposing role played by density and size of the particles in separation.

4. Conclusions The liberation is the core process in mechanical pre-treatment process. From the analysis, it is observed that the decrease in particle size decrease the metal content due to reduced tendency of metals to take part in comminution at lower sizes. Thus, it is important to know the optimum particle size for better liberation. It is concluded that the main metal enrichment was found to be in the size fraction of -1.6+0.5 mm. In the present work, the angular vane configuration at 30˚ as shown in fig. 10, was considered for air classification experiments because it helps preventing the heavier metallic particle to misplace in lighter fraction. The airflow rate is one of the crucial parameter in classification process to separate the metal rich and lean fractions. The optimum superficial air flow velocity of 13160 m/h facilitates the high recovery of metal-rich fraction from the mixture of PCBs. This study successfully demonstrates the strategy for material recycling from PCBs. In this approach, the size reduction followed by air classification resulted in better separation of metal rich and metal lean fraction. The obtained metal-rich fraction could be effectively used as a feed material for selective

13

recovery of valuable metals from the PCBs waste.

Acknowledgements The author wishes to acknowledge the head of department, Mineral Processing Department, Institute of Minerals and Materials Technology for his valuable suggestions while preparation of the manuscript. The authors also wish to acknowledge for the financial support provided by CSIR, Govt. of India.

Nomenclature d80

80% passing size

dp

particle size

rpm

revolutions/minute

PCBs

printed circuit boards

AAS

atomic absorption spectroscopy

References Kang, H.Y., Schoenung, J.M., 2005. Electronic waste recycling: a review of US infrastructure and technology options. Res., Conser. and Recyc. 45, 368-400. Lee, J.C., Song, H.T., Yoo, J.M., 2007. Present status of the recycling of waste electrical and electronic equipment in Korea. Res., Conser. and Recycl. 50, 380–397. Zhang, S., Forssberg, E., 1999. Intelligent liberation and classification of electronic scrap. Powder Tech. 105, 295-301. Cui, J.R., Forssberg, E., 2007. Characterization of shredded television scrap and implications for materials recovery. Waste Management 27, 415–424. Bi, X.H., Thomas, G.O., Jones, K.C., Qu, W.Y., Sheng, G.Y., Martin, F.L., Fu, J.M., 2007. Exposure of electronics dismantling workers to polybrominated diphenyl ethers, polychlorinated biphenyls, and organ-chlorine pesticides in South China. Environmental Science and Tech. 41, 5647–5653.

14

Wang, H., Gu, G.H., Qi, Y.F., 2005. Crushing performance and resource characteristic of printed circuit board scrap. J. of Central South University of Tech. 12, 552–555. Bertran, M., Graedel, T.E., Rechberger, H., Spatari, S., 2002. The contemporary European copper cycle: Waste Management sub system. Ecological Economics 42, 4357. Cui, J., Forssberg, E., 2003. Mechanical recycling of waste electric and electronic equipment: a review. J. of Haz. Mat. B99, 243-263. Goosey, M., Kellner, R., 2003. Recycling technologies for the treatment of end of life printed circuit boards (PCB). Circuit World 29, 33-37. Kinoshita, T., Akita, S., Kobayashi, N., Nii, S., Kawaizumi, F., Takahashi, K., 2003. Metal recovery from non-mounted printed wiring boards via hydrometallurgical processing. Hydrometall. 69, 73-79. Zong, G., Jianzhi, L., Hong, C.Z., 2002. Printed circuit board recycling: a state-of-art survey. In: Proc. of IEEE Int. Sym. on Electronics and the Environment, 234-241. Felix, N., Riet, C., 1994. Recycling of electronic scrap. In: Proc. of the Precious Metals Conf., Vancouver, 159-169. Shichang, Z., Yuquan, Z., Meixuan, C., Ying, W., Xikang, Z., 1994. Recycling of electric materials. Tran. of material research society of Japan 18A, 201-206. Gloe, K., Muhl, P., Knothe, M., 1990. Recovery of precious metals from electronic scrap, in particular from waste products of the thick-layer technique. Hydrometall. 25, 99-110. Pozzo, R.L., Malicsi, A.S., Iwasaki, I., 1991. Removal of lead from printed circuit board scrap by an electro-dissolution-delamination method. Res., Conser. and Recyc. 5, 21-34. Marca. F.L., Massacci, P., Piga, L., 2002. Recovery of precious metals from spent electronic boards. Recyc. and Waste Treatment in Mineral and Metal Processing: Technical and Economic Aspects, Lule´a, Sweden. Menad, N., Bjorkman, B., Allain, E.G., 1998. Combustion of plastics contained in electric and electronic scrap. Res., Conser. and Recyc. 24, 65-85. Guo, Q., Yue, X., Wang, M., Liu, Y., 2010. Pyrolysis of scrap printed circuit board plastic particles in a fluidized bed. Powder Tech. 198, 422–428. Leung, A.O.W., Luksemburg, W.J., Wong, A.S., Wong, M.H., 2007. Spatial distribution of polybrominated diphenyl ethers and polychlorinated dibenzo-p-dioxins and dibenzofurans in soil and combusted residue at Guiyu, an electronic waste recycling site in southeast China. Environmental Science and Tech. 41, 2730–2737.

15

Wong, M.H., Wu, S.C., Deng, W.J., Yu, X.Z., Luo, Q., Leung, A.O.W., Wong, C.S.C., Luksemburg, W.J., Wong, A.S., 2007. Export of toxic chemicals – a review of the case of uncontrolled electronic-waste recycling. Environmental Pollution 149, 131–140. Li, J., Lu, H., Guo, J., Xu, Z.M., Zhou, Y.H., 2007. Recycle technology for recovering resources and products from waste printed circuit boards. Environmental Science and Tech. 41, 1995–2000. Huang, K., Guo, J., Xu, Z.M., 2009. Recycling of waste printed circuit boards: a review of current technologies and treatment status in China. J. of Haz. Mat. 164, 399–408. Eswaraiah, C., Kavitha, T., Vidyasagar, S., Narayanan, S.S., 2008. Classification of metals and plastics from printed circuit boards (PCB) using air classifier. Chem. Engg. and Processing 47, 565–576. Das, A., Vidyadhar, A., Mehrotra, S.P., 2009. A novel flowsheet for the recovery of metal values from waste printed circuit boards. Res., Conser. and Recyc. 53, 464–469. Hall, W.J., Williams, P.T., 2007. Separation and recovery of materials from scrap printed circuit boards. Res., Conser. and Recyc. 51, 691–709. Veit, H.M., Diehl, T.R., Salami, A.P., Rodrigues, J.S., Bernardes, A.M., Tenório, J.A.S., 2005. Utilization of magnetic and electrostatic separation in the recycling of printed circuit boards scrap. Waste Management 25, 67–74. Zhang, S., Forssberg, E., 1997. Mechanical separation-oriented characterization of electronic scrap. Res., Conser. and Recyc. 21, 247-269. Veit, H.M., Pereira, C.C., Bernardes, A.M., 2002. Using mechanical processing in recycling printed circuit boards. JOM 45-47. Veit, H.M., Bernardes, A.M., Ferreira, J.Z., Tenorio, J.A.S., Malfatti, C.D., 2006. Recovery of copper from printed circuit boards scraps by mechanical processing and electrometallurgy. J. of Haz. Mat. 137, 1704–1709. Scott, K., Chen, X., Atkinson, J.W., Todd, M., Arnstrong, R.D., 1997. Electrochemical recycling of tin, lead, copper from stripping solution in the manufacture of circuit board. Res., Conser. and Recyc. 20, 43-55. Kekesi, T., Toro, T., Kabelik, G., 2000. Extraction of steel from scrap by chemical and electrochemical methods in alkaline media. Hydrometall. 55, 213-222. Ubaldini, S., Veglio, F., Quaresima, R., Fornari, P., 2003. Recovery of valuable metals from electronic and galvanic industrial wastes by leaching and electrowinning. Waste Management 23, 245-252. Jarring, C., Forssberg, E., 2003. Mechanical recycling of waste electric and Electronicequipment: a review. J. of Haz. Mat. 99, 243-263.

16

Wu, J., Li, J., Xu, Z., 2008. Optimization of key factors of electrostatic separation for crushed PCB wastes using roll-type separator. J. of Haz. Mat. 154, 161-171. Yoo, J.M., Jeong, J., Yoo, K., Lee, J., Kim, W., 2009. Enrichment of metallic components from waste printed circuit boards by a mechanical separation process using a stamp mill. Waste Management 29, 1132-37. Song, L., Sun, S., Zhong, S., Dai, W., Liu, J., Song, W., 2010. Using vacuum pyrolysis and mechanical processing for recycling waste printed circuit boards. J. of Haz. Mat. 164, 626-632. Guo, C., Wang, H., Liang, W., Fu, J., Yi, X., 2011. Liberation characteristic and physical separation of printed circuit board (PCB). Waste Management 31, 2161–66.

17

Fig. 1: Schematic diagram of Hazemag crusher experimental set-up

18

1

Cumulative fraction

0.8

0.6 0.4 0.2 0 100

1000

10000

Particle size, µm

Fig. 2: Product size analysis of Hazemag crusher at optimal conditions

19

Fig. 3: Schematic diagram of Hammer mill experimental set-up

20

1 Hammer mill

Cumulative fraction

0.8

Hazemag crusher

0.6

0.4

0.2

0 10

100

1000

10000

Particle size, µm

Fig. 4: Comparison of product size analysis of Hazemag crusher and Hammer mill at optimal conditions

21

Cumulative mass fraction

1 0.8 Feed rate

0.6 0.4

Low (0.933 kg/h) Medium (1.000 kg/h)

0.2

High (1.080 kg/h)

0 10

100

1000

10000

Particle size, µm

Fig. 5: Product size distribution of hammer milled PCB scraps at varying feed rates

22

Fig. 6: Schematic line diagram of a circulating air classifier

23

60 50

% metal

40 30 20 10

0 0

500

1000

1500

2000

2500

3000

Particle size, µm

Fig. 7: The measured variation of % metal content with the particle size

24

Metal content, %

100

10

1 6000

8000

10000

12000

14000

16000

Superficial Air Flow Velocity, m/h

Fig. 8: Variation of percent metal content with superficial air flow velocity for the stationary bottom angular vane configurations

25

60 50

% metal

40 30 20 10 0 0

500

1000

1500

2000

2500

Particle size, µm

Fig. 9: Differential distribution of metals in the hammer milled PCB

26

3000

Fan

Guide Vanes Feeder

Fig. 10: Rotating disc and fan section of wheel classifier

27

Figure caption:

Fig. 1: Schematic diagram of Hazemag crusher experimental set-up Fig. 2: Product size analysis of Hazemag crusher at optimal conditions Fig. 3: Schematic diagram of Hammer mill experimental set-up Fig. 4: Comparison of product size analysis of Hazemag crusher and Hammer mill at optimal conditions Fig. 5: Product size distribution of hammer milled PCB scraps at varying feed rates Fig. 6: Schematic line diagram of a circulating air classifier Fig. 7: The measured variation of % metal content with the particle size Fig. 8: Variation of percent metal content with superficial air flow velocity for the stationary bottom angular vane configurations Fig. 9: Differential distribution of metals in the hammer milled PCB Fig. 10: Rotating disc and fan section of wheel classifier

28