Article pubs.acs.org/est
Computer Simulation of the Pneumatic Separator in the Pneumatic− Electrostatic Separation System for Recycling Waste Printed Circuit Boards with Electronic Components Mianqiang Xue and Zhenming Xu* School of Environmental Science and Engineering, Shanghai Jiao Tong University, 800 Dongchuan Road, Shanghai 200240, People’s Republic of China S Supporting Information *
ABSTRACT: Technologies could be integrated in different ways into automatic recycling lines for a certain kind of electronic waste according to practical requirements. In this study, a new kind of pneumatic separator with openings at the dust hooper was applied combing with electrostatic separation for recycling waste printed circuit boards. However, the flow pattern and the particles’ movement behavior could not be obtained by experimental methods. To better control the separation quantity and the material size distribution, computational fluid dynamics was used to model the new pneumatic separator giving a detailed understanding of the mechanisms. Simulated results showed that the tangential velocity direction reversed with a relatively small value. Axial velocity exhibited two sharp decreases at the x axis. It is indicated that the bottom openings at the dust hopper resulted in an enormous change in the velocity profile. A new phenomenon that was named dusting was observed, which would mitigate the effect of particles with small diameter on the following electrostatic separation and avoid materials plugging caused by the waste printed circuit boards special properties effectively. The trapped materials were divided into seven grades. Experimental results showed that the mass fraction of grade 5, grade 6, and grade 7 materials were 27.54%, 15.23%, and 17.38%, respectively. Grade 1 particles’ mass fraction was reduced by 80.30% compared with a traditional separator. Furthermore, the monocrystalline silicon content in silicon element in particles with a diameter of −0.091 mm was 18.9%, higher than that in the mixed materials. This study could serve as guidance for the future material flow control, automation control, waste recycling, and semiconductor storage medium destruction.
1. INTRODUCTION A physical method for recycling electronic waste could be divided into crushing and separation stages. The crushing stage targets liberating metals from nonmetal components. And the following separation stages use screening, shape separation, magnetic separation, cyclone separation, eddy current separation, jigging, and electrostatic separation to upgrade desired materials.1−3 According to practical requirements, technologies could be integrated in different ways into automatic recycling lines for a certain kind of electronic waste.4,5 However, several problems are encountered particularly for particles when these recycling systems are applied on an industrial scale. One problem is that the material flow in the system, usually in negative pressure, is difficult to control. The pressure difference plays the role of materials transportation and dust collection. A second problem is material plugging. The plugged part will be the bottleneck affecting the whole recycling system operation. A third problem is the contradiction between the dust removal efficiency and the metal capture rate. The comminuted waste printed circuit boards were complex in composition and size distribution. Cyclones, characterized by simple design, high reliability, low maintenance cost, and © 2013 American Chemical Society
capability to work in extreme conditions, are used to achieve solid−solid grading according to density difference. During the process of industrial application, three categories of strategies generally are adopted: (i) optimizing the geometrical and operational parameters of the cyclone;6 (ii) designing new geometrical configuration for the cyclone;7 (iii) combining with other technologies.8 To better wipe off the small particles, the cyclone pierced an opening at the dust hopper was employed as a pretreatment unit for the electrostatic separation in this study. Different from traditional cyclones, the new separator tried to remove small particles rather than trap them. The controllable openings could change the flow field and the particles’ behavior, which attempted to cope with the problems analyzed above. To enhance the understanding of the new kind of pneumatic separator, the flow behavior and the particle trajectories were predicted by computational fluid dynamics. There are a number of turbulence models including the standard k−ε model, the Received: Revised: Accepted: Published: 4598
January 11, 2013 March 5, 2013 April 5, 2013 April 5, 2013 dx.doi.org/10.1021/es400154g | Environ. Sci. Technol. 2013, 47, 4598−4604
Environmental Science & Technology
Article
renormalization group (RNG) k−ε model, and the Reynolds stress model (RSM).9 Nevertheless, researches indicated that the RSM model can yield a more accurate prediction on flow pattern, velocity profile, and pressure distribution because of the anisotrotropic nature of the turbulence inside this unit.9−11 However, to calculate the trajectories of particles, a discrete phase model could be used to give important information on particles’ behavior.12 The models were described in detail in the next section. In this work, the flow field characteristics and the particles’ movement behavior of the pneumatic separator were studied by numerical simulation. One engineering application of the pneumatic separator combined with multiple crushing and electrostatic separation was taken as a case study. Size distribution tests were carried out and monocrystalline silicon concentrations were determined to confirm the effect of the bottom openings. The objective of this study is to explore the rules of the continuous and discrete phase behavior of the new pneumatic separator by computer simulation providing the foundation for the improvement of the integrated system for recycling waste printed circuit boards.
crusher by screw conveyor, and the undersize products were separated by the electrostatic separator. The whole operation process in detail was presented in the Supporting Information. 2.2. Turbulence Model. The pneumatic separator operation was regarded as an isothermal and incompressible process. The fluid can be described by the continuity equation and the Reynolds-averaged Navier−Stokes equation as follows: The continuity equation: ∂u i =0 ∂xi
(1)
The Reynolds-averaged Navier−Stokes equation: ρg uj
14
⎡ ∂uj ⎞⎤ ∂τij ∂ui ∂P ∂ ⎢ ⎛⎜ ∂ui ⎟⎥ + + = μ⎜ + ∂xj ∂xi ⎟⎠⎥⎦ ∂xj ∂xi ∂xj ⎣⎢ ⎝ ∂xj
(2)
The Reynolds stress tensor made the equations unclosed. Different turbulent models adopted different closed methods. Under the condition that the turbulence is anisotropic, the Reynolds stress transport equation was written as:15 ∂ ∂ (ρui′u′j ) + (ρukui′u′j ) = Dij + Pij + Φij − εij ∂t ∂xk
2. MATERIALS AND METHODS 2.1. Crushing-Pneumatic Separating-Electrostatic Separating System. Figure 1 presents schematic of the
(3)
The four terms on the right were the turbulent diffusion, stress production, press strain, and dissipation term, which were shown in Table 1. 2.3. Discrete Particle Model. The particles in the separator are subjected by fluid drag force, gravity, buoyancy force, pressure gradient force, and so on. The particle motion can be described in a Lagrangian reference frame as16 du p dt
= FD(u − u p) +
gx (ρp − ρ) ρp
+ Fx (4)
Re = ρDp|up − u|/μ, and the where FD = terms on the right side were fluid drag force, gravity, and other forces per unit particle mass, respectively. The velocity could be obtained by integration of eq 4 along the trajectory of an individual particle, and the particle trajectory was calculated by integration of the following equation: (18μ/ρpD2p)(CDRe/24),
Figure 1. Schematic of the crushing-pneumatic separating-electrostatic separating system.
dx = up dt
crushing-pneumatic separating-electrostatic separating system. Electronic waste was comminuted by the multicrushing system into particles. To guarantee sufficient liberation of metals and nonmetals, fine particles were inevitably produced along with the comminuting process. Unfortunately, study13 showed that nonconductive powder could seriously affect the metal particles’ charging and the following recycling efficiency due to the filling, enfolding, and adhering effects. The pneumatic separation acted as a pretreatment process for the following electrostatic separation. Fine particles were collected by the multipneumatic separation system and the bag-type dust collector. The oversize products were delivered back to the
(5)
2.4. Pneumatic Separator Model. Figure 2 presents the schematic of the pneumatic separator, and its dimensions are given in Table 2. The square inlet is tangential. The continuous phase was air. The inlet air and particle velocity were set both as 18 m/s. The bottom openings inlet velocity was set as 10 m/s. Velocity inlet and outflow boundary conditions were used for inlets and outlets, respectively. When calculating the discrete trajectories, particles were uniformly distributed on the inlet surface, with a concentration of 0.0008 kg/m3. It is assumed that there is no slip condition at the wall. Standard wall functions were used for near wall treatment. The SIMPLE
Table 1. Terms of Eq 3
numerical expression
Pij
Φij
⎛ ∂uj ∂u ⎞ − ρ⎜ui′uk′ + u′j uk′ i ⎟ ∂xk ∂xk ⎠ ⎝
⎛ ∂u ′ ∂u′j ⎞ ⎟ p⎜⎜ i + ∂xi′ ⎟⎠ ⎝ ∂x′j
Dij
component
−
⎤ ∂ ⎡ ⎢ρui′u′j uk′ + p δkjui′ + δiku′j ⎥ ∂xk ⎣ ⎦
(
)
4599
εij
2μ
∂ui′ ∂u′j ∂xk′ ∂xk′
dx.doi.org/10.1021/es400154g | Environ. Sci. Technol. 2013, 47, 4598−4604
Environmental Science & Technology
Article
Figure 3. Waste printed circuit boards with electronic components.
system. The turbulent kinetic energy, worked by Reynolds shear stress, was directly related to the energy loss. There was a large turbulent kinetic energy in the peripheral region of vortex finder followed by the region having a same height with the bottom openings and the joint between the separation main body and the dust hopper, whereas the central and near region had little turbulent kinetic energy. However, the turbulent intensity in the inlets and the vortex finder were large as shown in part b of Figure 4. It is seen that the bottom of the dust hopper had little turbulent intensity. In addition, the velocity vector diagram was presented in part c of Figure 4. There were two layers of rotational flow inside the separator: the external downward flow and the internal upward flow. Taken as a whole, the flow pattern was axisymmetric with small oscillations at the two sides. Remarkably, because of the effect of the bottom openings, the dust hopper flow pattern swirled up and down along with some local second vortex. On the basis of the above analysis, the tangential velocity, axial velocity, and static pressure profile of the pneumatic separator at the yz plane at the position z = 190 mm were compared with the traditional design separator without bottom openings. Figure 5 shows the tangential velocity profile. The tangential velocity producing centripetal force to drive the movement of particles played an important role in the separation process.17 For the traditional design, a maximum tangential velocity was observed at each side although the value in the x axis negative direction was a little bigger than that in the x axis positive direction. However, the tangential velocity direction reversed with a relatively small value for the new pneumatic separator providing more probability for dust removal. Figure 6 presents the axial velocity profile comparison. For the traditional design, the axial velocity increased first and then decreased at the range of x = −0.095 to x = 0. The same trend was observed at the range of x = 0 to x = 0.095. Whereas the new pneumatic separator exhibited two sharp decreases, of which one was at the x axis positive side and the other one was at the x axis negative side. The vertical airflow changed the local velocity direction and magnitude. It is suggested that the bottom openings at the dust hopper resulted in an enormous variation in the velocity profile. The static pressure profile at the position z = 190 mm was demonstrated in Figure 7. A good symmetry was seen for both the units, decreasing from the wall
Figure 2. Schematic of the pneumatic separator.
pressure-velocity coupling algorithm and first-order upwind scheme were employed. 2.5. Sample Screening and Analysis. To study the effect of the bottom openings, experiments were performed using the waste printed circuit boards with electronic components with a feeding rate of 200 kg/h (Figure 3). The trapped materials in 10 s were sieved by a standard screen machine (8411). The size grade is divided into seven levels: 1# (−0.091 mm), 2# (−0.125 + 0.091 mm), 3# (−0.3 + 0.125 mm), 4# (−0.45 + 0.3 mm), 5# (−0.6 + 0.45 mm), 6# (−0.8 + 0.6 mm), 7# (+0.8 mm). A weighted sample was sieved and material on each level was collected and weighted by analytical balance for size distribution investigation. The content of the monocrystalline silicon was determined by the X-ray photoelectron spectroscopy (AXIS UltraDLD, Kratos Analytical-A Shimadzu group company) using hybrid magnification mode. At the condition of ultrahigh vacuum, kinetic energy and the quantity of the electron that escaped from material surface beneath 1 to 10 nm were detected. X-ray photoelectron spectroscopy diagram of relative strength as a function of photoelectron kinetic energy was got. Si peak at 99.24 eV and SiOx peak at 103.3 eV were obtained using the XPSPEAK software.
3. RESULTS AND DISCUSSION 3.1. Flow Field Characteristics. For the new kind of pneumatic separator, the general flow pattern was first examined. Part a of Figure 4 shows the distribution of turbulent kinetic energy at the yz plane of the xyz coordinate Table 2. Geometry of the Pneumatic Separator symbol
D
De
a
b
c
d
H1
H2
H3
H4
B1
B2
S
value (mm)
190
64
95
38
30
30
285
760
100
150
72.5
72.5
95
4600
dx.doi.org/10.1021/es400154g | Environ. Sci. Technol. 2013, 47, 4598−4604
Environmental Science & Technology
Article
Figure 4. Distribution of (a) turbulent kinetic energy, m2/s2; (b) turbulent intensity, dimensionless; (c) velocity vector, m/s.
Figure 5. Tangential velocity profile on the yz plane at z = 190 mm.
Figure 7. Static pressure profile on the yz plane at z = 190 mm.
3.2. Particles Movement Behavior. Apart from the flow pattern characteristics, the particles behavior was also important for understanding the working mechanisms of the new pneumatic separator. Figure 8 provides a visualized result of the resin particle trajectories. For the traditional design separator, after injecting into the cylindrical part, large particles subjected to bigger centrifugal force were pushed to the wall and slipped down. Finally, they were collected in the dust hopper as shown in part a of Figure 8. Small particles presented great randomness. Specifically, some small particles directly escaped from the outlet. Some small particles spiral downward with the outer gaseous phase flow into the dust hooper, and finally they were dragged into the inner upward flow as shown in part c of Figure 8. It is worth noting that, for the particles with a diameter of 0.091 mm, they were harder to trap than large particles but were more easily affected by the gaseous phase than that of the small particles showed in part b of Figure 8. Compared with the particle traces in traditional design separator, the particle traces in the new pneumatic separator had a little depth for all of the particles shown in parts d−f of Figure 8. The reasons were that the bottom inlet flow enhanced
Figure 6. Axial velocity profile on the yz plane at z = 190 mm.
to the center. The difference was that the new pneumatic separator had a relatively bigger pressure loss. This might be attributed to the disturbance of the bottom inlet flow generating energy loss. 4601
dx.doi.org/10.1021/es400154g | Environ. Sci. Technol. 2013, 47, 4598−4604
Environmental Science & Technology
Article
Figure 8. Particle traces colored by residence time of the traditional cyclone: (a) d = 0.6 mm, (b) d = 0.091 mm, (c) d = 0.001 mm; and the pneumatic separator: (d) d = 0.6 mm, (e) d = 0.091 mm, (f) d = 0.001 mm.
the inner upward flow. Importantly, the movement direction of the large particles changed in the dust hooper, and after complex motion they were tapped. From the viewpoint of the continuous phase, it could also be confirmed by the new flow characteristics in Figure 5 and Figure 6. This phenomenon could be named dusting, which was a preferred result in this study. The particles with small diameter and even the dust that would have a strong impact on the following electrostatic separation 13 got more opportunities to be cleaned from particles with large diameter. Moreover, the bottom inlet reinforced the flow in the dust hooper, and for these reasons materials plugging caused by the waste printed circuit boards’ special properties can be effectively avoided. 3.3. Effect of Bottom Openings on Size Distribution. To validate the simulation results, experiments were carried out in industrial scale in the crushing-pneumatic separatingelectrostatic separating system, as shown in Figure 1. The bottom inlet velocity was set as 10 m/s. The size distribution features of the trapped materials were presented in Figure 9. In the traditional design separator, the mass fraction of grade 1
particles accounted for 84.81%. Other grades’ mass fraction only occupied a small part. Correspondingly, the cumulative distribution curve grew sharply from grade 1 to grade 2 particles and then increased gradually as the particle grade increased. Nevertheless, in the new design separator the proportion of large particles was significantly high. Specifically, grade 5, grade 6, and grade 7 mass fractions were 27.54%, 15.23%, and 17.38%, respectively. From the perspective of cumulative distribution, it maintained a rapid growth rate from grade 1 to grade 7. Small particles were wiped off with a positive effect. Taking grade 1 as an example, the mass fraction reduced by 80.30%, which was very helpful for the following electrostatic separation. 3.4. Performance Evaluation of the Recycling System. From the material flow point of view, material characterization was conducted first. Figure 10 presents the characteristics of the materials collected from the four outputs as shown in Figure 1. Output 1, metal production, was mainly spherical and laminated. Output 2 was nonmetal production, and it was 4602
dx.doi.org/10.1021/es400154g | Environ. Sci. Technol. 2013, 47, 4598−4604
Environmental Science & Technology
Article
combined application of the electrostatic separation and pneumatic separation. From waste recycling point of view, each part production was graded and purified. It provided a foundation for the aim of zero waste discharge. On the one hand, the materials in the multipneumatic separation system were graded, as shown in Table S1 of the Supporting Information. Different level of pneumatic separation has different size distribution. On the other hand, because of the dedust effect of the pneumatic separation, the materials in the electrostatic separation system were sorted well. Then, nonmetal products of different sizes could be reused as raw material for degradation applications18 and metal products could be refined by green chemical methods. From an information destruction point of view, the contents of Si in the products were determined considering the distinctiveness of information storage. Figure S1 of the Supporting Information presents the speciation analysis of silicon: (a) mixed materials, (b) materials with diameter of +0.45−0.6 mm, (c) materials with diameter of −0.091 mm. The monocrystalline silicon content in large particles (+0.45− 0.6 mm) was 10.1%, whereas the monocrystalline silicon content in small particles (−0.091 mm) was 18.9%. Therefore, the parts having memory function can be destroyed up to standard and other parts are recycled as resources given the fact that it is impossible to crush all materials into powder especially for the metals. The recycling system could be a reference for the semiconductor storage medium destruction considering feasibility, energy consumption, and system complexity. In summary, the bottom openings at the dust hopper brought about a significant variation in the velocity profile. The movement direction of the large particles changed in the dust hooper, and after complex motion they were tapped, which was favorable for the following electrostatic separation. The material flow could be adjusted by the openings. Moreover, the upward flow could effectively avoid the material plugging. In the new design separator, the mass fractions of grade 5, grade 6, and grade 7 were 27.54%, 15.23%, and 17.38%, respectively. The grade 1 particles mass fraction was reduced by 80.30%, which indicating a good dust remove effect. It is suggested that the integrated electrostatic and new design pneumatic separation have meaningful results for electronic waste disposal from the perspective of dust remove, waste recycling, and semiconductor destruction.
Figure 9. Size distribution of the trapped materials: (a) traditional design, (b) with bottom openings.
■
ASSOCIATED CONTENT
S Supporting Information *
Description of the operation process; table showing size distribution of the outputs from the recycling system when the bottom inlet velocity was 6.3 m/s; table showing the nomenclature; figure showing speciation analysis of silicon: (a) mixed materials, (b) +0.45−0.6, (c) −0.091 mm. This material is available free of charge via the Internet at http:// pubs.acs.org.
■
Figure 10. Characteristics of the outputs from the recycling system.
AUTHOR INFORMATION
Corresponding Author
needle-shaped and laminated. Different from the electrostatic separation product, the pneumatic separation product was mainly particles with small diameter. Dust was first removed by the pneumatic separator and the cleaned particles were sorted by the electrostatic separator. This confirms the value of the
*Tel: +86 21 54747495; fax: +86 21 54747495; e-mail: zmxu@ sjtu.edu.cn. Notes
The authors declare no competing financial interest. 4603
dx.doi.org/10.1021/es400154g | Environ. Sci. Technol. 2013, 47, 4598−4604
Environmental Science & Technology
■
Article
ACKNOWLEDGMENTS This work was supported by the National High Technology Research and Development Program of China (863 program 2009AA06Z318, 2012AA063206) and the National Natural Science Foundation of China (21077071).
■
REFERENCES
(1) Cui, J.; Forssberg, E. Mechanical recycling of waste electric and electronic equipment: A review. J. Hazard. Mater. 2003, 99, 243−263. (2) He, W.; Li, G.; Ma, X.; Wang, H.; Huang, J.; Xu, M.; Huang, C. WEEE recovery strategies and the WEEE treatment status in China. J. Hazard. Mater. 2006, 136, 502−512. (3) Xue, M.; Yang, Y.; Ruan, J.; Xu, Z. Assessment of noise and heavy metals (Cr, Cu, Cd, Pb) in the ambience of the production line for recycling waste printed circuit boards. Environ. Sci. Technol. 2012, 46, 494−499. (4) Zhou, L.; Xu, Z. Response to waste electrical and electronic equipments in China: Legislation, recycling system, and advanced integrated process. Environ. Sci. Technol. 2012, 46, 4713−4724. (5) Duan, C.; Wen, X.; Shi, C.; Zhao, Y.; Wen, B.; He, Y. Recovery of metals from waste printed circuit boards by a mechanical method using a water medium. J. Hazard. Mater. 2009, 166, 478−482. (6) Elsayed, K.; Lacor, C. Optimization of the cyclone separator geometry for minimum pressure drop using mathematical models and CFD simulations. Chem. Eng. Sci. 2010, 65, 6048−6058. (7) Hwang, K. J.; Hwang, Y. W.; Yoshida, H.; Shigemori, K. Improvement of particle separation efficiency by installing conical topplate in hydrocyclone. Powder Technol. 2012, 232, 41−48. (8) Li, J.; Xu, Z. Environmental friendly automatic Line for recovering metal from waste printed circuit boards. Environ. Sci. Technol. 2010, 44, 1418−1423. (9) Guo, J.; Zheng, Y.; Sun, G.; Wang, J. Study of the separation efficiency and the flow field of a dynamic cyclone. Sep. Purif. Technol. 2006, 49, 157−166. (10) Gimbun, J.; Chuah, T. G.; Fakhru’l-Razi, A.; Choong, T. The influence of temperature and inlet velocity on cyclone pressure drop: a CFD study. Chem. Eng. Process. 2004, 44, 7−12. (11) Hoekstra, A. J.; Derksen, J. J.; Van Der Akker, H. E. A. An experimental and numerical study of turbulent swirling flow in gas cyclones. Chem. Eng. Sci. 1999, 54, 2055−2065. (12) Elsayed, K.; Lacor, C. Numerical modeling of the flow field and performance in cyclones of different cone-tip diameters. Comput. Fluids 2011, 51, 48−59. (13) Wu, J.; Qin, Y.; Zhou, Q.; Xu, Z. Impact of nonconductive powder on the electrostatic separation for recycling crushed waste printed circuit board. J. Hazard. Mater. 2009, 164, 1352−1358. (14) Azadi, M.; Azadi, M.; Mohebbi, A. A CFD study of the effect of cyclone size on its performance parameters. J. Hazard. Mater. 2010, 182, 835−841. (15) Wang, B.; Yu, A. B. Numerical study of particle-fluid flow in hydrocyclones with different body dimensions. Miner. Eng. 2006, 19, 1022−1033. (16) Raoufi, A.; Shams, M.; Farzaneh, M.; Ebrahimi, R. Numerical simulation and optimization of fluid flow in cyclone vortex finder. Chem. Eng. Process. 2008, 47, 128−137. (17) Hsieh, K. T.; Rajamani, R. K. Mathematical model of the hydrocyclone based on physics of fluid flow. AIChE J. 1991, 5, 735− 746. (18) Guo, J.; Tang, Y.; Xu, Z. Wood plastic composite produced by nonmetals from pulverized waste printed circuit boards. Environ. Sci. Technol. 2010, 44, 463−468.
4604
dx.doi.org/10.1021/es400154g | Environ. Sci. Technol. 2013, 47, 4598−4604
