Environ. Sci. Technol. 2010, 44, 5177–5181
Electrostatic Separation for Recycling Waste Printed Circuit Board: A Study on External Factor and a Robust Design for Optimization SHIBING HOU, JIANG WU, YUFEI QIN, AND ZHENMING XU* School of Environmental Science and Engineering, Shanghai Jiao Tong University, 800 Dong chuan Road, Shanghai, People’s Republic of China
Received December 28, 2009. Revised manuscript received May 19, 2010. Accepted May 20, 2010.
Electrostatic separation is an effective and environmentally friendly method for recycling waste printed circuit board (PCB) by several kinds of electrostatic separators. However, some notable problems have been detected in its applications and cannot be efficiently resolved by optimizing the separation process. Instead of the separator itself, these problems are mainly caused by some external factors such as the nonconductive powder (NP) and the superficial moisture of feeding granule mixture. These problems finally lead to an inefficient separation. In the present research, the impacts of these external factors were investigated and a robust design was built to optimize the process and to weaken the adverse impact. A most robust parameter setting (25 kv, 80 rpm) was concluded from the experimental design. In addition, some theoretical methods, including cyclone separation, were presented to eliminate these problems substantially. This will contribute to efficient electrostatic separation of waste PCB and make remarkable progress for industrial applications.
Introduction Treatment of waste electric and electronic equipment (WEEE) has become a very hot issue, not only about environmental protection, but also resource recycling (1-3). Electrostatic separation (Figure S1 in the Supporting Information), defined as the selective sorting of charged or polarized bodies in an electric field (4-6), presents an effective way for recycling metals and nonmetals from WEEE without negative impact to the environment, especially for the ground waste printed circuit board (PCB) (7). Generally, the efficiency of electrostatic separation is influenced by many kinds of factors. These factors can be divided into two groups. The first group is defined as electrostatic separation system factor (ESS factor) including voltage level, rotational speed, and electrode configuration. These factors affecting the electrostatic separation have been reported in detail (8). Another group is classified as external factor. Through industrial experience, the main external factors are nonconductive powder (NP) and superficial moisture on the feeding materials. These factors can be * Corresponding author phone:+86 21 54747495; fax:+86 21 54747495; e-mail:
[email protected]. 10.1021/es903936m
2010 American Chemical Society
Published on Web 06/02/2010
eliminated in laboratory conditions, but are difficult to control in practical conditions. Some researchers have demonstrated the negative effects caused by NP (9). However, there is little published information aiming at the systemic discussion about these external factors and methods for solution. Therefore, the aim of this paper is to investigate the external factors and their negative impacts. Then, a robust design is developed for optimization, and to weaken the external factors’ adverse impact. In addition, some theoretical methods are presented to eliminate these problems substantially. This is significant for high efficiency of electrostatic separation under industrial applications.
Problems of External Factors Nonconductive Powder (NP). For practical conditions, the most serious problem is due to the nonconductive powder (NP) in the feeding materials. Before the electrostatic separation, a grinding process is indispensable and an appropriate particle size is prerequisite for an effective separation. For the waste PCBs, the optimum particle size is 0.1-0.6 mm (7, 10). However, overgrinding is almost inevitable even if the grinding process is operated under the optimum condition. This brings about considerable NP (11) in the feeding materials (Figure 1). The NP, as part of nonconductors, in this paper consists of about 40 wt % glass fiber and 60 wt % resin after thermal gravity analysis. Under laboratory conditions, a screening process can be used to remove this NP. NP has very fine size, large specific area, and intensive surface free energy. It will bring about obvious surface phenomenon, some attractive forces like van der Waals force, liquid bridge force between NP or other objects, and lead to some negative impacts on electrostatic separation of ground PCB wastes. In addition, it leads to a serious powder accumulation on electrodes’ surfaces. A previous paper reported the NP’s negative effects on electrostatic separation and three phenomena were respectively named “filling effect”, “enfolding effect”, and “adhering effect” (9). In that paper, a sharp decrease of separation efficiency caused by the NP was demonstrated. These negative impacts bring about a poor efficiency, and are very disadvantageous to industrial applications. Superficial Moisture of Feeding Granule Mixture. The second problem comes from the feeding granule mixture’s superficial moisture caused by ambient humidity. On one hand, a moisture film will form on the nonconductive particles’ surfaces for its adsorption of atmospheric moisture. This increases the nonconductive particles’ superficial conductivity, and its charge acquired by “ion bombardment” will transfer to the grounded roll electrode like the conductive particle. On the other hand, there are some attractive forces between the fine particles. Under dry conditions, van der Waals force is the main attractive force between particles. This is just the theoretical basis for the foregoing NP’s enfolding and adhering effects in dry ambient. However, the liquid bridge force becomes the main attractive force in moist ambient (Figure S2). Once the separation process runs in high moisture ambient, the liquid bridge force caused by superficial moisture film will lead to a strong trend of aggregation, not only for NP, but also for the normal particles. Robust Design. A robust design is a statistical method developed by Genichi Taguchi to improve the quality of manufactured goods. It is a form of optimization whereby a system is made less sensitive to the effects of random variability, or noise. It has been used in many industry fields and proves to be a powerful tool in process quality control VOL. 44, NO. 13, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5177
FIGURE 1. Source of nonconductive powder and its SEM photograph.
TABLE 1. Superficial Moisture Experiment Arrangement experiment
group 1
RHa parameter settings a
Relative humidity.
group 2
b
High voltage. c Roll speed.
d
Y ) -10log(φ2 + σ2)
(1)
Experimental Section Impact of Superficial Moisture on Separation Process. To investigate the impact of superficial moisture, this paper uses RH as an indicator for superficial moisture. Generally, superficial moisture is positively correlated with RH. An experiment was performed in 5 groups under different ambient relative humidity levels (Table 1). As shown in Figure S3, a constant-humidity maintaining device was built to simulate ambient relative humidity. The feeding samples of each group were placed in this device for 12 h to ensure certain superficial moisture. After the simulation process, the samples were separated under the optimum settings that have been discussed in previous research (14, 15) and shown in Table 1. Each group consisted of 10 tests. The samples (0.3-0.45 mm) were synthetic and prepared with ground PCB wastes. For each sample, the mass was 200 g and contained 25 wt % metals (copper) and 75 wt % nonmetals (woven glass, reinforced resin). The products of each test were weighed respectively by an electronic balance with resolution of 0.1 g. All electrostatic separation experiments were carried out in ambient air, at a temperature of 20 °C, with RH 50%-70%. Robust Design of the Process. In this paper, high voltage level (U) and rotational speed (N) were used as design factors. According to recommendation of previous research (15), the 9
group 4
group 5 80%
Position parameters of electrodes.
(12). In this paper, RH (relative humidity) and powder have a great negative effect on the quality of the electrostatic process for crushed PCB. In the plant, these two external factors, which can be called noise factors are very difficult to change manually. The objective of robust design in this paper is to minimize the negative effect of these noise factors on the electrostatic separation process. Experimental design of robust design can be called “inner and outer orthogonal array approach”. The design factor and noise factor are arranged as inner array and outer orthogonal array, respectively. Design factors are easy-tocontrol ESS factors, whereas noise factors are hard-to-control external factors (RH and powder). To measure the robustness, a signal and noise ratio (SNR) is used as the objective function. In this paper, a popular SNR function is used as follows, in which φ is the average output and σ is the variance of each point in the inner array (13). When Y attains maximum, the negative effect of noise factors will reduce to bottom line.
5178
group 3
40% 50% 60% 70% Ub ) 30 kV; Nc ) 60 rpm; s1d ) 70 mm; s2d ) 90 mm; a1d ) 25°; a2d ) 75°
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 13, 2010
design factors varied in the following range: Umax ) 30 KV, Umin ) 25 KV, Nmax ) 80 rpm, Nmin ) 60 rpm. To keep accordance with previous research, the paper use particles under 0.045 mm as NP. The NP content and superficial moisture (relative humidity, or RH, is used as indicator for superficial moisture) were used as noise factors. Their value ranges are NPmax ) 30%, NPmin ) 10%, RHmax ) 80%, RHmin ) 40%. Every design point in the inner array was arranged by a square of outer array. Each inner array and outer array had five design points, so the design number was 5*5 ) 25. Additionally, to verify the reproducibility of the experiment, two more experiments were performed in the central point, making the total experimental number 27 (Figure 2). Actually, the five outer array design factors meant five kinds of experiment material sample. They are defined as shown in Table 2. All experiment settings and facilities were the same as the above except RH, and they were carried out randomly.
Results and Discussion Impact of Superficial Moisture on Separation Process. The experimental results are shown in Figure 3a. The masses of
FIGURE 2. Inner and outer orthogonal arrays of an experimental design.
TABLE 2. Experiment Material Sample Group
TABLE 3. Robust Design Experiment Results
group 1 group 2 group 3 group 4 group 5 RH NP content (wt %)
40% 10
40% 30
80% 10
80% 30
60% 20
middling products and conductive products are generally used as a criterion to judge the performance of separation process (16, 17). With the increase of RH, the middling products increase from 1.1 to 2.1 g (increase 91%), while the purity of conductive products decreases from 99.5% to 87.3%. Doubtless, the materials’ superficial moisture caused by ambient humidity leads to obvious negative impacts on the separation process. The separation efficiency becomes worse with the increase of RH. There are two reasons for this. First, a moisture film will form on the nonconductive particle’s surface for its adsorption of atmospheric moisture. This increases the nonconductive particle’s superficial conductivity and leads to a decrease of charge and electric image force. Under the certain humidity (W), the change of electric image force with chargedecrease can be calculated by the eqs 2 and 3 (18), fi(W) ) q2(W)/[4πε0(2a)2] ) Fiexp[-2t/τ(W)]
(2)
τ(W) ) ε(W)/σ(W)
(3)
where fi(W) is the electric image force, q(W) is acquired charge, ε0 is vacuum dielectric constant, a is the radius of the nonconductive grain, τ(W) is time constant, ε(W) and σ(W) are nonconductive particle’s dielectric constant and conductivity, respectively. Once the ambient humidity increases, τ(W) decreases with the increase of σ(W) and fi(W) decreases accordingly. Because of the decrease of electric image force, the nonconductive particle cannot be pinned to the roll surface and will detach from the roll at a much lower departure angle. As a result, the separation efficiency is poor. Second, the liquid bridge force becomes the main attraction force in a moist ambient (Figure S2). For such a fine particle, the liquid bridge force caused by superficial moisture film will lead to a strong trend of aggregation. To confirm this effect on the granule mixture from ground PCB wastes, an atomic force microscope was used to investigate the nonconductive particle’s surface adhesion force in different RH and the result is shown in Figure 3b. This confirms the negative impacts of superficial moisture. Robust Design of the Process. All 27 experiment results of the robust design are listed in Table 3. The classic analysis approach is to calculate the average value (ψ) and the variance (σ2) of each inner array corresponding to outer array. Then according to eq 1, the output/input ratio table is obtained and shown in the Table 4.
test NP (wt%) RH (%) U (KV) N (rpma) Cb (%) Mc (%) NCd (%) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
10 10 10 10 10 30 30 30 30 30 10 10 10 10 10 30 30 30 30 30 20 20 20 20 20 20 20
40 40 40 40 40 40 40 40 40 40 80 80 80 80 80 80 80 80 80 80 60 60 60 60 60 60 60
25 30 25 30 27.5 25 30 25 30 27.5 25 30 25 30 27.5 25 30 25 30 27.5 25 30 25 30 27.5 27.5 27.5
60 60 80 80 70 60 60 80 80 70 60 60 80 80 70 60 60 80 80 70 60 60 80 80 70 70 70
15.7 14.4 20.3 19.2 17.6 9.7 9.1 18.7 15.5 11.6 27.0 24.4 27.5 23.3 23.5 12.4 13.1 19.8 17.1 14.5 17.2 17.2 20.3 19.4 19.0 18.5 18.1
2.4 3.1 2.0 2.9 2.5 3.6 4.6 2.6 2.9 3.4 4.6 5.8 4.2 5.2 5.6 5.0 5.7 4.5 5.3 4.6 5.7 5.8 5.1 5.5 4.3 4.5 5.5
78.7 79.7 75.1 75.4 77.9 84.8 83.0 77.2 78.4 82.3 67.1 67.6 67.5 68.4 69.8 79.1 76.7 74.3 73.9 77.9 78.0 73.8 72.9 71.7 74.5 74.2 73.8
a Round per minute. b Mass percent of conductors. Mass percent of middling products. d Mass percent of nonconductors. c
It is a clear from Table 4 that the increase of voltage level enhances the middling production. In contrast, the increase of the roll speed diminishes the middling products. This can be easily explained by NP effect, for NP decreasea the conductivity of metals. When voltage level increases, the electric attraction force exerted on those “nonconductive metals” strengthens and more middling products fall into middling collector and nonconductive collector. As for the reduction of middling products caused by roll speed, centrifugal force is the main effect. Drop of roll speed can make the centrifugal force decline. Therefore, more and more metals fall into the middling collector because of not enough centrifugal force to detach at the right point. According to Table 4, an equation (eq 4) can be derived after regression analysis using calculation software to characterize the relation between function Y and design factors (U and N). Only the factors having statistical significance are shown in eq 4. Y ) -12.881 - 0.659U∗ + 0.5541N*-0.023U*N*
(4)
FIGURE 3. Separation results under different ambient RH (a), and particle’s surface adhesion force in moisture atmosphere (b). VOL. 44, NO. 13, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5179
TABLE 4. Calculation Results of Output/Noise Ratio U (KV)
N (rpm)
group 1
group 2
group 3
group 4
group 5
Ψa
σ2b
Yc
25 30 25 30 27.5 27.5 27.5
60 60 80 80 70 70 70
2.4 3.1 2.0 2.9 2.5
3.6 4.6 2.6 2.9 3.4
4.6 5.8 4.2 5.2 5.6
5.0 5.7 4.5 5.3 4.6
5.7 5.8 5.1 5.5 4.3 4.5 4.7
4.3 5.0 3.7 4.4 4.1
1.3 1.1 1.4 1.4 1.1
-12.9 -14.2 -11.7 -13.1 -12.5
a
Average value of inner array.
b
Square of variance of inner array. c Function Y.
FIGURE 4. Surface plots (left) and contour plots (right) of the function Y model.
FIGURE 5. Improved flowchart for electrostatic separation process. An asterisk (*) means the variable has been normalized. The surface plots and contour plots are demonstrated in Figure 4. The figure shows clearly that lower voltage level with higher roll speed make the process more robust (25 KV and 80 rpm will be best), while higher voltage levels make the system much more instable.
Solutions for External Factors’ Negative Impacts Although the robust design can elevate the stableness through some ESS factor adjustments, some technology process should be introduced before the ground PCB wastes undergo an electrostatic separation, if a substantial advance in process stability is desired. In the following chart (Figure 5), cyclone separation and moisture removal are added. a. Solution for NP-Cyclone Separation. To weaken NP’s negative impacts, some size-classification process like screening and cyclone separation could be set before the electrostatic separation. As the cyclone presents a strong ability to remove fine particles (19), the paper strongly recommends the pretreatment technology. A cyclone separator can classify particles in different sizes based on the cut-size (d50) at different settings. The cut-size d50 can be calculated by eq 5 (19), which is derived from the classic Barth Model. d50 )
vrCS9µDx 2 FpvθCS
(5)
The cut-size (d50) is like an “air mesh”. The particle would be removed by air flow if it is smaller than the cut-size; particles larger than the cut-size would be caught. Because 5180
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 13, 2010
of the huge size difference between NP and normal particles (more than 2 times), cyclone separation will be very fit for removing NP from the feeding materials. To verify this idea, a primary experiment in a Shanghai plant was set up. Meanwhile, a laboratory experiment about the NP content effect is set up too. Metal recycling efficiency is measured by the ratio between pure metal collected in the metal collector and total metal in PCB. Table 5 shows that the cyclone can remove most of the NP, and the metals recycling efficiency is increased sharply. Basically, the plant experiment keeps accordance with laboratory experiment. A line treating waste PCB employed the cyclone as a pretreatment process and showed a good result (20). However, advanced research work about the cyclone separation pretreatment technology for waste PCB is still up in the air. These problems include the best operating parameters and how to coordinate with the electrostatic separation. Solution for Superficial Moisture. Wet ambient can make electrostatic separation unstable. So, maintaining dry ambient in the plant is the key for reducing the impacts of feeding materials’ superficial moisture, especially during the moist springtime and summertime. Once the PCB wastes are subject to the moist ambient, a drying process is a positive method to pave the way for an effective electrostatic separation. An automatic moisture controller should be installed in electrostatic separation plants. When the RH is up to 60%, the dryer will work immediately to decrease the RH. With this automatic controller, the RH will be controlled under the optimum value.
TABLE 5. NP Impacts on the Recycling Efficiency NP content (wt %)
metal recycling efficiency (%)
laboratory experiment
0 10 20 30
98.5 91.8 84.2 75.4
plant - no cyclone experiment - with cyclone
30-20 10-5
75-80 90-95
Acknowledgments This project was supported by the National High Technology Research and Development Program of China (863 program 2006AA06Z364), Program for Energy Saver and Exhaust Reducer of Shanghai (09dz1204404), Research Fund for the Doctoral Program of Higher Education (20090073120041), and Shanghai Natural Science Foundation (10ZR1415900).
Supporting Information Available Electrostatic separation process for recycling PCB; liquid bridge between particles in a moist ambient; and constant humidity device schematic diagram. This information is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Cui, J.; Forssberg, E. Mechanical recycling of waste electric and electronic equipment: a review. J. Hazard. Mater. 2003, 99, 243– 363. (2) Leung, A. O. W.; Luksemburg, W. J.; Wong, A. S.; Wong, M. H. 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. Environ. Sci. Technol. 2007, 41, 2730– 2737. (3) Li, H. R.; Yu, L. P.; Sheng, G. Y.; Fu, J. M.; Peng, P. A. Severe PCDD/F and PBDD/F Pollution in Air around an Electronic Waste Dismantling Area in China. Environ. Sci. Technol. 2007, 41, 5641–5646. (4) Ralston, O. C. Electrostatic Separation of Mixed Granular Solids; Elsevier: Amsterdam, 1961. (5) Kiewiet, C.; Bergougnou, M. A.; Brown, J. D. Electrostatic Separation of fine Particle in Vibrate Fluidized Beds. IEEE Trans. Ind. Appl. 1978, 6, 526–530.
(6) TaylorJ. B. Dry electrostatic separation of granular materials. IAS Annual Meeting (IEEE Industry Applications Society) 1988; vol 35, pp 1741-1759. (7) Li, J.; Lu, H. Z.; Guo, J.; Xu, Z. M.; Zhou, Y. H. Recycle technology for recovering resources and products from waste printed circuit boards. Environ. Sci. Technol. 2007, 41, 1995–2000. (8) Wu, J.; Li, J.; Xu, Z. M. Electrostatic Separation for Recovering Metal and Nonmetal from Waste Printed Circuit Board: Problems and Improvements. Environ. Sci. Technol. 2008, 42, 5272–5276. (9) Wu, J.; Qin, Y.; Zhou, Q.; Xu, Z. Impact of nonconductive powder on electrostatic separation for recycling crushed waste printed circuit board. J. Hazard. Mater. 2009, 164, 1352–1358. (10) Lu, H. Z.; Li, J.; Guo, J.; Xu, Z. M. Pulverization characteristics and pulverizing of waste printed circuit boards (printed wiring boards) based on resource utilization. J. Shanghai Jiaotong Univ. 2007, 41, 551–556. (11) Guo, J; Rao, Q. L.; Xu, Z, M. Application of glass-nonmetals of waste printed circuit boards to produce phenolic moulding compound. J. Hazard. Mater. 2008, 153, 728–734. (12) Dascalescu, L. Robust corona-electrostatic separation method for solid waste recycling. Int. J. Environ. Waste Manage. 2008, 2, 423–435. (13) Phadke, M. S.; Taguchi., G. Selection of quality characteristics and S/N ratios for robust design Globecom Meeting, IEEE Communication Society, Tokyo, 1987; pp 1002-1007. (14) Li, J.; Lu, H. Z.; Liu, S. S; Xu, Z. M. Optimizing the operating parameters of corona electrostatic separation for recycling waste scraped printed circuit boards by computer simulation of electric field. J. Hazard. Mater. 2008, 153, 269–275. (15) Wu, J.; Li, J.; Xu, Z. M. Optimization of key factors of the electrostatic separation for crushed PCB wastes using roll-type separator. J. Hazard. Mater. 2008, 154, 161–167. (16) Dascalescu, L.; Samuila, A.; Mihalcioiu, A.; Bente, S.; Tilmatine, A. Robust design of electrostatic separation processes. IEEE Trans. Ind. Appl. 2005, 41, 715–720. (17) Medles, K.; Tilmatine, A.; Miloua, F.; Bendaoud, A.; Younes, M.; Rahli, M.; Dascalescu, L. Set point identification and robustness testing of electrostatic separation processes. IEEE Trans. Ind. Appl. 2007, 43, 618–626. (18) Dascalescu, L.; Samuila, A.; Iuga, A.; Morar, R.; Csorvasy, I. Influence of Material Superficial Moisture on Insulation -Metal Electro-separation. IEEE Trans. Ind. Appl. 1994, 30, 844–849. (19) Hoffmann, A. C.; Stein, L. C. Gas Cyclones and Swirl Tubes: Principles, Design and Operation; Beijing, Springer Verlag: Berlin, 2004. (20) Li, J; Xu, Z. M. Environmental friendly automatic line for recovering metal from waste printed circuit boards. Environ. Sci. Technol. 2010, 44, 1418–1423.
ES903936M
VOL. 44, NO. 13, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5181
