- Email: [email protected]
Recovering low-turbidity cutting liquid from silicon slurry waste
Journal of Hazardous Materials 271 (2014) 252–257
Contents lists available at ScienceDirect
Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat
Recovering low-turbidity cutting liquid from silicon slurry waste Tzu-Hsuan Tsai ∗ , Yu-Pei Shih Department of Materials and Mineral Resources Engineering, National Taipei University of Technology, Taipei 10608, Taiwan
h i g h l i g h t s
g r a p h i c a l
• Recovering clean cutting liquids was
The particles in the solution of glycol (PAG) and water were larger than those in PAG–ethanol or PAG–acetone solutions, showing that water interfered with the adsorption of PAG molecules and weakened the steric stabilization. Using water as a diluent for sedimentation facilitated the separation of solids from cutting liquids.
achieved by sedimentation with a diluent. • Water as a diluent was better than ethanol and acetone. • The recovered liquids (<100 NTU) could be reused in the cutting process.
a r t i c l e
i n f o
Article history: Received 25 November 2013 Received in revised form 10 February 2014 Accepted 21 February 2014 Available online 1 March 2014 Keywords: Recovery Turbidity Cutting liquid Silicon slurry waste Sedimentation
a b s t r a c t In order to recover a low-turbidity polyalkylene glycol (PAG) liquid from silicon slurry waste by sedimentation, temperatures were adjusted, and acetone, ethanol or water was used as a diluent. The experimental results show that the particles in the waste would aggregate and settle readily by using water as a diluent. This is because particle surfaces had lower surface potential value and weaker steric stabilization in PAG–water than in PAG–ethanol or PAG–acetone solutions. Therefore, water is the suggested diluent for recovering a low-turbidity PAG (<100 NTU) by sedimentation. After 50 wt.% water-assisted sedimentation for 21 days, the solid content of the upper liquid reduced to 0.122 g/L, and the turbidity decreased to 44 NTU. The obtained upper liquid was then vacuum-distillated to remove water. The final recovered PAG with 0.37 NTU had similar viscosity and density to the unused PAG and could be reused in the cutting process. © 2014 Elsevier B.V. All rights reserved.
1. Introduction The rapid development of photovoltaic industries has increased the need for silicon wafers. However, substantial silicon slurry waste is formed during the cutting process from silicon ingot to wafers [1,2]. More than 50% of this waste comes from the cutting liquid. For a factory with a polysilicon throughput of 2000 tons
∗ Corresponding author. Tel.: +886 2 2771 2171x2775; fax: +886 2 2778 7579. E-mail address: [email protected] (T.-H. Tsai). http://dx.doi.org/10.1016/j.jhazmat.2014.02.032 0304-3894/© 2014 Elsevier B.V. All rights reserved.
a b s t r a c t
annually, the required cutting liquid is about 1600 tons annually. Meanwhile, at least 3200 tons of silicon slurry waste is generated [3]. The cutting liquid generally consists of water-soluble glycol molecules, which can be classified as small molecules or polymers. The former includes ethylene glycol (EG), propylene glycol (PG) or diethylene glycol (DEG); the latter includes polyethylene glycol or copolymer of ethylene oxide and propylene oxide, also known as polyalkylene glycol (PAG). The long chain molecules in polymer-type cutting liquid provide good dispersion of abrasives, high lubricity and heat tolerance, resulting in an excellent cutting yield [4]. However, the polymer-type cutting liquid forms
T.-H. Tsai, Y.-P. Shih / Journal of Hazardous Materials 271 (2014) 252–257
wastewater with high COD, which is difficult to treat and becomes a major environmental issue. Recycling this glycol liquid from sawing waste would greatly reduce the amount of waste, and thereby the total slicing cost can be efficiently reduced [5]. Manufacturers and recyclers currently recover small-molecule cutting liquids by distillation, which heats the waste beyond the boiling point of glycol so that the obtained vapor can be separated from the solid waste. High-purity glycol can then be collected by condensation [6]. The collected liquid can be reused in the cutting process after adding additives. However, PAG cutting liquid dissociates or oxidizes at high temperature and is not recoverable through the collected vapor during distillation. Therefore, some manufacturers and recyclers use low-temperature methods to separate cutting liquids from silicon slurry waste, such as centrifugal separation [7,8] and filtration [5,8]. Furthermore, SiC, Si particles and metal fragments suspended in PAG liquid are so stable that recyclers have difficulty obtaining a low-turbidity PAG liquid from silicon slurry waste. In recent years, many patents proposed the techniques or skills to recovery and renew polymer cutting liquids, including predilution [5,7,8] for enhancing separation of solid and liquid, coagulation [9], membrane filtering [10] or ultrafiltration [11] for separation of particles, ion exchange treatment [12], reverse osmosis treatment [10] or vacuum evaporation [13] for removing water, and decolorizing for adsorbing impurities [13]. However, the suggested methods were only patented, i.e. no practical data could be available for their use in recovering PAG liquid from silicon slurry waste. This study, therefore, used sedimentation process to recover PAG cutting liquid from silicon slurry waste. Sedimentation is the simplest way to separate solids from liquids and is easily performed in factories, such as the cases for recycling lubricant [14] and heavy oil [15]. To obtain a low-turbidity PAG liquid (<100 NTU) from silicon slurry waste, this study adjusted temperatures and employed acetone, ethanol or deionized water as a diluent during sedimentation. Observation of the settling behaviors revealed the separation mechanism of particles during recovery of PAG liquids from silicon slurry waste. Finally, the recovering process to obtain a clean PAG liquid was built up. 2. Experimental 2.1. Analysis of silicon slurry waste Silicon slurry waste was obtained from Sino-American Silicon Products, Inc. (Chu-Nan, Taiwan). After preliminary recycling of large SiC particles with a centrifugal decanter, the slurry waste with low solid content was transferred to our laboratory for the sequential analysis and recovery of cutting liquids. The obtained silicon slurry waste contained kerf Si, SiC abrasives, metal fragments from cutting wire and polyalkylene glycol (PAG) cutting liquid. The average molecular weight of PAG was 750. To analyze its composition, each waste sample was first heated to 500 ◦ C to remove liquids. The solid content percentage (S1 ) was then calculated as the mass of the residual solid powder dividing by the mass of the original obtained waste. The particle size distribution (PSD) in the obtained silicon slurry waste was measured using static light scattering (model: LA300, HORIBA). Viscosity was measured using a Viscometer (model: LVDV-I Prime, BROOKFIELD), and turbidity was measured with a turbidity meter (model: 2100Q, HACH). 2.2. Separation process Silicon slurry waste (100 ml) was transferred to a graduated cylinder for the sedimentation experiment. In order to control the
253
sample temperature during sedimentation, the graduated cylinder was put in an oven set to temperatures of 25, 40 and 80 ◦ C. Moreover, three liquids, acetone, ethanol and deionized water, were used as diluents and mixed with the obtained silicon slurry waste by 5-min stirring. After thoroughly mixing the waste and diluents, 100-ml diluted waste was poured into a graduated cylinder to allow the particles to settle freely. 2.3. Analysis of the recovered liquid After sedimentation for different days, the upper liquid or the recovered liquid, in the cylinder was analyzed. Particle size distribution (PSD) was measured by static light scattering (model: LA300, HORIBA). Viscosity was measured with a viscometer (model: LVDVI Prime, BROOKFIELD). Turbidity was measured with a turbidity meter (model: 2100Q, HACH). In addition, the recovery liquid was heated to 500 ◦ C to remove glycol or diluents, and the solid content percentage (S2 ) was obtained by the mass of the residual solid dividing by the mass of the original recovered liquid. Due to the hygroscopicity of PAG molecules, the obtained waste or recovered liquid would contain water. The water content was measured with a Karl Fischer titrator (model: 870 plus, HACH). When the diluent was deionized water, the measured water content percentage in the obtained waste (W1 ) and in the recovered liquid (W2 ) could determine the recovery percentage of glycol by the following equation Recovery percentage of glycol (%) =
(Mass of the recovered liquid) × (1 − S2 − W2 ) × 100% (Mass of the obtained waste) × (1 − S1 − W1 )
(1)
3. Results and discussion Fig. 1(a) shows that the obtained slurry waste had a black and opaque appearance. The measured density, viscosity at 25 ◦ C, and solid content of the waste were 1.009 g/cm3 , 72 cp and 15.6 wt.%, respectively. The particle size distribution of the waste (Fig. 1(b)) indicated that the particle size of the obtained waste was below 7 m, and the main peaks of the particle sizes were 0.3 and 1.2 m, which showed that most of the large SiC (around 15 m as used originally) was recycled before delivery to our laboratory. These micron particles in spent slurry waste included kerf Si, metal fragments and broken SiC abrasives. Because of their small size, the particles were suspended stably in PAG cutting liquid, which produced an opaque appearance and high nephelometric turbidity units (NTU). The turbidity of the obtained waste exceeded 1000 NTU, which was the detection limit of the turbidity meter (model: 2100Q, HACH) used in this study. To recovery a clean PAG liquid, lowering the solid content and turbidity is essential. Generally, low turbidity indicates a low solid content, but suspension liquids with the same solid content might have different nephelometric turbidity units due to different particle sizes. Therefore, before recovery of a low-turbidity cutting liquid, a standard value corresponding to a clean PAG liquid was examined by adding approximately 1 m silicon particles into PGA. The prepared PAG-Si suspension was used in analyzing the relations among solid content, turbidity and appearance of PAG-Si suspension. Table 1 shows how the solid content and nephelometric turbidity units affected the appearance of the prepared PAG-Si suspension. Notably, the suspension was not transparent, and the turbidity exceeded 1000 NTU when the solid content reached 0.5 g/L. The appearance was clean and resembled an unused cutting liquid through visual inspection when the solid content decreased to 0.0075 g/L and the turbidity reduced to 109 NTU. The smaller particles at the same solid content had a higher turbidity. Since the
254
T.-H. Tsai, Y.-P. Shih / Journal of Hazardous Materials 271 (2014) 252–257
80
1.02
1.01
40
3
Density (g/cm )
Viscosity (cp)
60
1.00 20
20
40
60
80
100
o
Temperature ( C) Fig. 2. The effect of temperature on viscosity and density of the slurry waste.
analyze the settling of particles. The deduced terminal velocity (Vt ) of particles in the waste could be described as [16]
Vt =
Fig. 1. The obtained slurry waste: (a) the appearance and (b) particle size distribution.
waste analyzed in this study consisted mainly of 0.3 and 1.2 m particles, the turbidity requirement of the recovered liquids was set below 100 NTU. To recover a low-turbidity PAG liquid (<100 NTU) from a spent slurry waste by sedimentation, the Stokes law was used to
(p − s )dp2 g
(2)
18s
where dp , p , s and s represent particle diameter, particle density, waste density and waste viscosity, respectively. According to Eq. (2), the settling velocity of 1.2 m Si particles in the obtained slurry waste at 25 ◦ C would be 1.44 × 10−6 cm/s. This creeping motion retarded the recovery rate for the PAG liquid. Reducing the viscosity or density of the slurry waste might increase the settling velocity and enhance the separation of particles from cutting liquid. Therefore, to enhance separation performance, temperatures were adjusted or diluents were used to change the viscosity or density of the slurry waste. Fig. 2 shows the effect of temperature on viscosity and density of the slurry waste. The waste viscosity decreased as temperature increased while the density did not significantly differ. At 40 ◦ C, the waste viscosity was 37 cp; at 80 ◦ C, the viscosity decreased to 15 cp. According to Eq. (2), the settling velocity of 1.2 m Si particles in the obtained slurry waste should be approximately 4.8 times faster at 80 ◦ C than at 25 ◦ C. Therefore, a higher recovery rate was predicated at 80 ◦ C.
Table 1 The solid content, turbidity and appearance of the prepared PAG-Si suspension. Solid content (g/L) Turbidity (NTU)
Appearance a
Unused PAG cutting liquid.
0a 0.5
0.0025 31
0.0075 109
0.1 468
0.5 >1000
T.-H. Tsai, Y.-P. Shih / Journal of Hazardous Materials 271 (2014) 252–257
(a)
o
25 C o 40 C o 80 C
80
Water Ethanol Acetone
60
Viscosity (cp)
Solid content (g/L)
100
255
10
40
20
1 0
5
10
15
20
0 0
Time (day)
20
Fig. 3. The solid content of the recovered liquid after sedimentation at 25 ◦ C, 40 ◦ C and 80 ◦ C for different days.
60
80
(b)
1.2
Water Ethanol Acetone
1.1
3
Density (g/cm )
Fig. 3 shows the solid content of the recovered liquid after sedimentation at 25, 40 and 80 ◦ C for different days. The solid content decreased over time during the first 7 days, and the decrease at 80 ◦ C was larger than that at 25 or 40 ◦ C. The decreasing extent at 80 ◦ C reduced after 7 days, and that at 25 or 40 ◦ C also went down after 14 days. After 21 days, the solid content at 80 ◦ C decreased to 5.13 g/L, and the particle removal fraction reached 99.5%. However, the turbidity of the recovered liquid still exceeded 1000 NTU. The PSDs of the recovered liquids at various temperatures in Fig. 4 showed that the main peak of the particle size approximated 0.2 m after sedimentation for 14 days. These submicron sized particles suspended stably and were difficult to separate from PAG liquid by gravitational force. Therefore, a cutting liquid with low nephelometric turbidity units could not be obtained. Water, ethanol and acetone were then used as diluents to decrease the viscosity and density of the slurry waste. Fig. 5 shows that both the viscosity and density of the slurry waste decreased with the addition of the diluents, especially adding acetone and ethanol. Based on this finding, the settling velocity (Vd ) was calculated at various diluent concentrations by using Eq. (2). Fig. 6 shows the effect of adding diluents on Vd with respect to the original velocity (Vo ). The curves in Fig. 6 predicted that the addition of diluents could certainly increase the settling velocity, and adding ethanol provided the fastest settling. For example, when the
1.0
0.9
0.8
0
20
40
60
80
Diluent (wt.%) Fig. 5. The effect of adding diluents on viscosity and density of the slurry waste.
40 Water Ethanol Acetone
30 20 o
15
80 C Vd/Vo
10 5
Volume (%)
40
Diluent (wt.%)
20 15
20
o
40 C
10
10
5 20 15
o
25 C
10
0
5 0 0.1
0 1
10
100
20
40
60
80
Diluent (wt.%)
Particle size (µm) Fig. 4. Particle size distribution of the recovered liquid after sedimentation at 25 ◦ C, 40 ◦ C and 80 ◦ C for 14 days.
Fig. 6. The settling velocity with respect to the original velocity (Vd /Vo ) after adding various concentrations of diluents.
256
T.-H. Tsai, Y.-P. Shih / Journal of Hazardous Materials 271 (2014) 252–257
100
Solid content (g/L)
Table 2 The solid content, turbidity and appearance of the unused PAG and the upper liquid after sedimentation.
Origin Water Ethanol Acetone
Liquid
Unused PAG
Upper liquid
Solid content (g/L) Turbidity (NTU)
0 0.5
0.122 44
10
1
0.1 0
5
10
15
20
25
Time (day) Fig. 7. The solid content of the upper liquid after sedimentation with 50 wt.% diluents for different days.
Volume (%)
diluent concentration was 50 wt.% for 1.2 m Si particles, Vd /Vo was 18.72 after adding ethanol and 9.92 after adding acetone but only 3.67 after adding water, which forecasted that recovery rate could be increased by using a diluent, ethanol especially. Fig. 7 shows the effect of adding 50 wt.% diluent on the solid content of the upper liquid after sedimentation for different days. The data indicated that all solid content decreased after addition of diluents. Initially, as forecasted by Vd /Vo in Fig. 6, the solid content was lower for adding ethanol or acetone compared to that for adding water. After 3 days, the solid content was below 3 g/L by adding ethanol or acetone, but their decreasing extent reduced. After sedimentation for 21 days, the solid content decreased to 0.304 g/L by adding ethanol and decreased to 0.724 g/L by adding acetone. Despite the low solid content, the turbidity still exceeded 1000 NTU with the addition of ethanol or acetone. Although addition of water initially obtained a smaller decrease in solid content, the effective reduction of solid content was sustained for 14 days. After 14 days, the solid content after addition of water was lower than that after addition of ethanol or acetone. After 21 days of sedimentation, adding water lowered the solid content to 0.122 g/L and decreased turbidity to 44 NTU. Table 2 shows that the upper liquid with 44 NTU had a clean appearance and resembled an unused PAG liquid.
20 15 10 5 20 15 10 5 20 15 10 5 0 0.1
Water
Ethanol
Acetone
1
10
100
Particle size (µm) Fig. 8. Particle size distribution of the upper liquid after sedimentation with 50 wt.% diluents for 14 day.
Appearance
Fig. 8 shows the PSDs of the upper liquids by sedimentation after addition of diluents for 14 days. The data show the main peak of the particle size was around 0.15 m after addition of ethanol or acetone. These submicron particles suspended stably in PAG–ethanol or PAG–acetone solutions and were difficult to separate from PAG liquid by gravitational sedimentation. However, after sedimentation with the addition of water for 14 days, most particles in the upper liquid were 3.4 m (Fig. 8). The particles in PAG–water solution were larger than those in PAG–ethanol and PAG–acetone solutions. It means that the particles aggregated when water was used as a diluent. The dispersion or aggregation of particles in suspension depends on surface potentials and the presence of adsorbed layers that physically limit the approach of particles, i.e. steric stabilization. Assume Si and SiC surfaces in the spent waste produced SiO2 due to oxidation [17]. According to the references, the electrophoretic mobilities of SiO2 in water, ethanol and acetone have been measured, and the data are −1.7 × 10−8 m2 V−1 s−1 [18], −9.5 × 10−9 m2 V−1 s−1 [17], −2.3 × 10−8 m2 V−1 s−1 [19], respectively. Because the particles are small and the medium are non-electrolytic solutions, the electrophoretic mobilities could be converted to surface potentials using Hückel equation [20]. Thus, the surface potentials of SiO2 in water, ethanol and acetone are −38.87 mV, −75.19 mV and −61.00 mV, respectively, showing submicron particles would suspend more stably in ethanol and in acetone than in water. In Fig. 7, because the settling particles were large initially, the gravitational force would dominate the settling velocity. As forecasted by Vd /Vo in Fig. 6, addition of water obtained a smaller decrease in solid content than addition of ethanol or acetone. The surface potentials became important for the settling of submicron particles. Due to the lower surface potential value in water than that in ethanol or acetone, the particles aggregated easily in water as confirmed by the PSDs of Fig. 8. Therefore, after 3 days, the decreasing extent of solid content reduced by adding ethanol or acetone, but it reduced sustainably for 14 days by adding water. Furthermore, the settling medium included PAG molecules. The PAG molecules could adsorb on particles to form steric stabilization. In PAG–ethanol or PAG–acetone solutions, PAG molecules adsorbed on Si or SiC surfaces would prevent particle aggregation
T.-H. Tsai, Y.-P. Shih / Journal of Hazardous Materials 271 (2014) 252–257 Table 3 The properties of the unused PAG and the recovered liquid by sedimentation and vacuum-distillation. Liquid (25 ◦ C)
Unused PAG
Recovered liquid
Viscosity (cp) Density (g/cm3 ) Turbidity (NTU) Water content (wt.%)
36.6 0.994 0.5 0.251
38.6 0.997 0.37 0.246
257
solution because, in PAG–ethanol or PAG–acetone solutions, PAG molecules adsorbed on Si or SiC surfaces prevent particle aggregation due to steric stabilization. In PAG–water solution, Si or SiC particles had hydrophilic surfaces, and the attraction of hydrogen bonding between water–water or between PAG–water would interfere with the adsorption of PAG molecules on particles and weaken the steric stabilization. Thus, the particles would aggregate readily, causing rapid settling and facile separation from PAG–water solution. After sedimentation by adding 50 wt.% water for 21 days, the solid content of the upper liquid decreased to 0.122 g/L, and the turbidity reached 44 NTU. The obtained upper liquid was then vacuum-distillated at 60 torr and 70 ◦ C to remove water. The turbidity of the final recovered PAG with 0.246 wt.% water was only 0.37 NTU. The recovery percentage of glycol was 81%. Acknowledgements The authors would like to thank the National Science Council Taiwan for financially supporting this research under Contract No. NSC 101-2221-E-027-109-MY3, and thank Sino-American Silicon Products, Inc. for material assistance.
Appearance
due to steric stabilization [21]. However, in PAG–water solution, formation of oxide enhanced the hydrophilicity of particles [22]. Additionally, among these three diluents, water has the highest chemical polarity and the largest hydrogen bonding component of Hansen solubility parameters [23], indicating a different interactions by using water as a diluent. In PAG–water solution, the attraction of hydrogen bonding happened between water–water or between PAG–water interfered with the adsorption of PAG molecules on particles, and weakened the steric stabilization. The particles were attracted to each other in the PAG–water solution and aggregated into larger particles. The large particles readily settled and then separated from PAG–water solution. Finally, a low-turbidity liquid was obtained. After sedimentation with the addition of 50 wt.% water for 21 days, the obtained upper liquid was vacuum-distillated at 60 torr and 70 ◦ C to remove water. Table 3 shows that the final recovered product contained only 0.246 wt.% water, and the recovered liquid and the unused PAG had similar viscosity and density. Moreover, the turbidity of the recovered PAG was only 0.37 NTU. The estimated recovery percentage of glycol reached 81% according to Eq. (1). Water is the suggested diluent for mixing with the spent slurry waste, and sedimentation can be performed at room temperature without a complex apparatus. After sedimentation, water removal was performed by vacuum distillation at low temperature, such as 70 ◦ C or below 70 ◦ C. Thus, the obtained liquid had a low turbidity and could be reused in wide applications. 4. Conclusions This study recovered PAG cutting liquids from silicon slurry waste by sedimentation. The settling velocity of particles and separation performance were enhanced by adjusting the temperature and using diluents. Although a particle removal fraction exceeding 99.5% could be achieved by increasing temperature or by adding ethanol or acetone, the recovered liquids were still opaque and turbid with larger than 1000 NTU due to the stable suspension of submicron particles. When water was used as the diluent, the PSD results showed that the particles tended to aggregate in PAG–water
References [1] D. Sarti, R. Einhaus, Silicon feedstock for the multi-crystalline photovoltaic industry, Sol. Energy Mater. Sol. Cells 72 (2002) 27–40. [2] T. Neesse, Review on SiC-recycling in wafer sawing operation, Interceram 55 (2006) 430–432. [3] T.H. Tsai, Y.P. Shih, Y.F. Wu, Recycling silicon wire saw slurries: separation of silicon and silicon carbide in a ramp settling tank under an applied electrical field, J. Air Waste Manag. Assoc. 63 (2013) 512–517. [4] W. Yu, D.A. Aguilar, J.B. Cuthbert, L.Y.P. Zhu, Cutting fluids with improved performance, U.S. Patent (2012) 20120196779. [5] C. Zavattari, G. Fragiacomo, Method for the separation, regeneration and reuse of an exhausted glycol-based slurry, U.S. Patent (2001) 6231628. [6] Y.L. Sun, Y.H. Liu, M.Z. Tsai, Method for recycling cutting fluid, U.S. Patent (2012) 20120073952. [7] K. Toyama, System for reusing oily slurry waste fluid, U.S. Patent (1998) 5830369. [8] G. Gaudet, S. Grumbine, N. Naguib, F. Batllo, Wire saw slurry recycling process, U.S. Patent (2009) 20090293369. [9] K.S. Chang, Separation of waste cutting oil generated from solar wafer slicing process, T.W. Patent (2011) 098142647. [10] J.J. Iang, Method and apparatus for recycling and treating wastes of silicon wafer cutting and polishing processes, U.S. Patent (2012) 20120312747. [11] N. Drouiche, M.W. Naceur, H. Boutoumi, N. Aitmessaoudene, R. Henniche, T. Ouslimane, Assessment of the recovery of photovoltaic cells cutting fluid by chemical pretreatment and ultrafiltration, Desalination Water Treat. 51 (2013) 713–716. [12] L. Liu, Recycling cutting fluid components from crystal silicon cutting waste mortar comprises homogenizing agent into waste mortar of crystal silicon cutting liquid, and preparing homogenizing agent and water in each waste mortar, C.N. Patent (2012) 102399619. [13] T.Y. Yao, Method for recovering water-soluble cutting fluid from silicon wafer cutting fluid, C.N. Patent (2012) 102746934. [14] R.U. Owolabi, K.A. Alabi, O.B. Oke, Reclamation of spent automobile engine lubricating oil, Fount. J. Natl. Appl. Sci. 2 (2013) 11–16. [15] D. Feng, L. Liu, P. Wang, H. Zhang, X. Xie, The experimental study of natural sedimentation in heavy oil, Adv. Mater. Res. 413 (2012) 60–64. [16] P.G.W. Hawksley, The effect of concentration on the settling of suspensions and flow through porous media, Some Aspects of Fluid Flow, Edward Arnold, London, 1951, pp. 114–118. [17] Y. Fukada, P.S. Nicholson, The role of Si-O species in the colloidal stability of silicon-containing ceramic powders, J. Eur. Ceram. Soc. 24 (2004) 17–23. [18] W.J. Masschelein, Unit Processes in Drinking Water Treatment, CRC Press, New York, 1992. [19] H.A. Ketelson, R. Pelton, M.A. Brook, Colloidal stability of stöber silica in acetone–water mixtures, J. Colloid Interface Sci. 179 (1996) 600–607. [20] E. Hückel, Die kataphorese der kugel, Physiol. Zool. 25 (1924) 204–210. [21] D.H. Napper, Polymeric Stabilization of Colloidal Dispersions, Academic Press, London, 1983. [22] A. Feng, B.J. McCoy, Z.A. Munir, D.E. Cagliostro, Water adsorption and desorption kinetics on silica insulation, J. Colloid Interface Sci. 180 (1996) 276–284. [23] C.M. Hansen (Ed.), Hansen Solubility Parameters: A User’s Handbook, 2nd ed., CRC Press, Boca Raton, FL, 2007.
Contents lists available at ScienceDirect
Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat
Recovering low-turbidity cutting liquid from silicon slurry waste Tzu-Hsuan Tsai ∗ , Yu-Pei Shih Department of Materials and Mineral Resources Engineering, National Taipei University of Technology, Taipei 10608, Taiwan
h i g h l i g h t s
g r a p h i c a l
• Recovering clean cutting liquids was
The particles in the solution of glycol (PAG) and water were larger than those in PAG–ethanol or PAG–acetone solutions, showing that water interfered with the adsorption of PAG molecules and weakened the steric stabilization. Using water as a diluent for sedimentation facilitated the separation of solids from cutting liquids.
achieved by sedimentation with a diluent. • Water as a diluent was better than ethanol and acetone. • The recovered liquids (<100 NTU) could be reused in the cutting process.
a r t i c l e
i n f o
Article history: Received 25 November 2013 Received in revised form 10 February 2014 Accepted 21 February 2014 Available online 1 March 2014 Keywords: Recovery Turbidity Cutting liquid Silicon slurry waste Sedimentation
a b s t r a c t In order to recover a low-turbidity polyalkylene glycol (PAG) liquid from silicon slurry waste by sedimentation, temperatures were adjusted, and acetone, ethanol or water was used as a diluent. The experimental results show that the particles in the waste would aggregate and settle readily by using water as a diluent. This is because particle surfaces had lower surface potential value and weaker steric stabilization in PAG–water than in PAG–ethanol or PAG–acetone solutions. Therefore, water is the suggested diluent for recovering a low-turbidity PAG (<100 NTU) by sedimentation. After 50 wt.% water-assisted sedimentation for 21 days, the solid content of the upper liquid reduced to 0.122 g/L, and the turbidity decreased to 44 NTU. The obtained upper liquid was then vacuum-distillated to remove water. The final recovered PAG with 0.37 NTU had similar viscosity and density to the unused PAG and could be reused in the cutting process. © 2014 Elsevier B.V. All rights reserved.
1. Introduction The rapid development of photovoltaic industries has increased the need for silicon wafers. However, substantial silicon slurry waste is formed during the cutting process from silicon ingot to wafers [1,2]. More than 50% of this waste comes from the cutting liquid. For a factory with a polysilicon throughput of 2000 tons
∗ Corresponding author. Tel.: +886 2 2771 2171x2775; fax: +886 2 2778 7579. E-mail address: [email protected] (T.-H. Tsai). http://dx.doi.org/10.1016/j.jhazmat.2014.02.032 0304-3894/© 2014 Elsevier B.V. All rights reserved.
a b s t r a c t
annually, the required cutting liquid is about 1600 tons annually. Meanwhile, at least 3200 tons of silicon slurry waste is generated [3]. The cutting liquid generally consists of water-soluble glycol molecules, which can be classified as small molecules or polymers. The former includes ethylene glycol (EG), propylene glycol (PG) or diethylene glycol (DEG); the latter includes polyethylene glycol or copolymer of ethylene oxide and propylene oxide, also known as polyalkylene glycol (PAG). The long chain molecules in polymer-type cutting liquid provide good dispersion of abrasives, high lubricity and heat tolerance, resulting in an excellent cutting yield [4]. However, the polymer-type cutting liquid forms
T.-H. Tsai, Y.-P. Shih / Journal of Hazardous Materials 271 (2014) 252–257
wastewater with high COD, which is difficult to treat and becomes a major environmental issue. Recycling this glycol liquid from sawing waste would greatly reduce the amount of waste, and thereby the total slicing cost can be efficiently reduced [5]. Manufacturers and recyclers currently recover small-molecule cutting liquids by distillation, which heats the waste beyond the boiling point of glycol so that the obtained vapor can be separated from the solid waste. High-purity glycol can then be collected by condensation [6]. The collected liquid can be reused in the cutting process after adding additives. However, PAG cutting liquid dissociates or oxidizes at high temperature and is not recoverable through the collected vapor during distillation. Therefore, some manufacturers and recyclers use low-temperature methods to separate cutting liquids from silicon slurry waste, such as centrifugal separation [7,8] and filtration [5,8]. Furthermore, SiC, Si particles and metal fragments suspended in PAG liquid are so stable that recyclers have difficulty obtaining a low-turbidity PAG liquid from silicon slurry waste. In recent years, many patents proposed the techniques or skills to recovery and renew polymer cutting liquids, including predilution [5,7,8] for enhancing separation of solid and liquid, coagulation [9], membrane filtering [10] or ultrafiltration [11] for separation of particles, ion exchange treatment [12], reverse osmosis treatment [10] or vacuum evaporation [13] for removing water, and decolorizing for adsorbing impurities [13]. However, the suggested methods were only patented, i.e. no practical data could be available for their use in recovering PAG liquid from silicon slurry waste. This study, therefore, used sedimentation process to recover PAG cutting liquid from silicon slurry waste. Sedimentation is the simplest way to separate solids from liquids and is easily performed in factories, such as the cases for recycling lubricant [14] and heavy oil [15]. To obtain a low-turbidity PAG liquid (<100 NTU) from silicon slurry waste, this study adjusted temperatures and employed acetone, ethanol or deionized water as a diluent during sedimentation. Observation of the settling behaviors revealed the separation mechanism of particles during recovery of PAG liquids from silicon slurry waste. Finally, the recovering process to obtain a clean PAG liquid was built up. 2. Experimental 2.1. Analysis of silicon slurry waste Silicon slurry waste was obtained from Sino-American Silicon Products, Inc. (Chu-Nan, Taiwan). After preliminary recycling of large SiC particles with a centrifugal decanter, the slurry waste with low solid content was transferred to our laboratory for the sequential analysis and recovery of cutting liquids. The obtained silicon slurry waste contained kerf Si, SiC abrasives, metal fragments from cutting wire and polyalkylene glycol (PAG) cutting liquid. The average molecular weight of PAG was 750. To analyze its composition, each waste sample was first heated to 500 ◦ C to remove liquids. The solid content percentage (S1 ) was then calculated as the mass of the residual solid powder dividing by the mass of the original obtained waste. The particle size distribution (PSD) in the obtained silicon slurry waste was measured using static light scattering (model: LA300, HORIBA). Viscosity was measured using a Viscometer (model: LVDV-I Prime, BROOKFIELD), and turbidity was measured with a turbidity meter (model: 2100Q, HACH). 2.2. Separation process Silicon slurry waste (100 ml) was transferred to a graduated cylinder for the sedimentation experiment. In order to control the
253
sample temperature during sedimentation, the graduated cylinder was put in an oven set to temperatures of 25, 40 and 80 ◦ C. Moreover, three liquids, acetone, ethanol and deionized water, were used as diluents and mixed with the obtained silicon slurry waste by 5-min stirring. After thoroughly mixing the waste and diluents, 100-ml diluted waste was poured into a graduated cylinder to allow the particles to settle freely. 2.3. Analysis of the recovered liquid After sedimentation for different days, the upper liquid or the recovered liquid, in the cylinder was analyzed. Particle size distribution (PSD) was measured by static light scattering (model: LA300, HORIBA). Viscosity was measured with a viscometer (model: LVDVI Prime, BROOKFIELD). Turbidity was measured with a turbidity meter (model: 2100Q, HACH). In addition, the recovery liquid was heated to 500 ◦ C to remove glycol or diluents, and the solid content percentage (S2 ) was obtained by the mass of the residual solid dividing by the mass of the original recovered liquid. Due to the hygroscopicity of PAG molecules, the obtained waste or recovered liquid would contain water. The water content was measured with a Karl Fischer titrator (model: 870 plus, HACH). When the diluent was deionized water, the measured water content percentage in the obtained waste (W1 ) and in the recovered liquid (W2 ) could determine the recovery percentage of glycol by the following equation Recovery percentage of glycol (%) =
(Mass of the recovered liquid) × (1 − S2 − W2 ) × 100% (Mass of the obtained waste) × (1 − S1 − W1 )
(1)
3. Results and discussion Fig. 1(a) shows that the obtained slurry waste had a black and opaque appearance. The measured density, viscosity at 25 ◦ C, and solid content of the waste were 1.009 g/cm3 , 72 cp and 15.6 wt.%, respectively. The particle size distribution of the waste (Fig. 1(b)) indicated that the particle size of the obtained waste was below 7 m, and the main peaks of the particle sizes were 0.3 and 1.2 m, which showed that most of the large SiC (around 15 m as used originally) was recycled before delivery to our laboratory. These micron particles in spent slurry waste included kerf Si, metal fragments and broken SiC abrasives. Because of their small size, the particles were suspended stably in PAG cutting liquid, which produced an opaque appearance and high nephelometric turbidity units (NTU). The turbidity of the obtained waste exceeded 1000 NTU, which was the detection limit of the turbidity meter (model: 2100Q, HACH) used in this study. To recovery a clean PAG liquid, lowering the solid content and turbidity is essential. Generally, low turbidity indicates a low solid content, but suspension liquids with the same solid content might have different nephelometric turbidity units due to different particle sizes. Therefore, before recovery of a low-turbidity cutting liquid, a standard value corresponding to a clean PAG liquid was examined by adding approximately 1 m silicon particles into PGA. The prepared PAG-Si suspension was used in analyzing the relations among solid content, turbidity and appearance of PAG-Si suspension. Table 1 shows how the solid content and nephelometric turbidity units affected the appearance of the prepared PAG-Si suspension. Notably, the suspension was not transparent, and the turbidity exceeded 1000 NTU when the solid content reached 0.5 g/L. The appearance was clean and resembled an unused cutting liquid through visual inspection when the solid content decreased to 0.0075 g/L and the turbidity reduced to 109 NTU. The smaller particles at the same solid content had a higher turbidity. Since the
254
T.-H. Tsai, Y.-P. Shih / Journal of Hazardous Materials 271 (2014) 252–257
80
1.02
1.01
40
3
Density (g/cm )
Viscosity (cp)
60
1.00 20
20
40
60
80
100
o
Temperature ( C) Fig. 2. The effect of temperature on viscosity and density of the slurry waste.
analyze the settling of particles. The deduced terminal velocity (Vt ) of particles in the waste could be described as [16]
Vt =
Fig. 1. The obtained slurry waste: (a) the appearance and (b) particle size distribution.
waste analyzed in this study consisted mainly of 0.3 and 1.2 m particles, the turbidity requirement of the recovered liquids was set below 100 NTU. To recover a low-turbidity PAG liquid (<100 NTU) from a spent slurry waste by sedimentation, the Stokes law was used to
(p − s )dp2 g
(2)
18s
where dp , p , s and s represent particle diameter, particle density, waste density and waste viscosity, respectively. According to Eq. (2), the settling velocity of 1.2 m Si particles in the obtained slurry waste at 25 ◦ C would be 1.44 × 10−6 cm/s. This creeping motion retarded the recovery rate for the PAG liquid. Reducing the viscosity or density of the slurry waste might increase the settling velocity and enhance the separation of particles from cutting liquid. Therefore, to enhance separation performance, temperatures were adjusted or diluents were used to change the viscosity or density of the slurry waste. Fig. 2 shows the effect of temperature on viscosity and density of the slurry waste. The waste viscosity decreased as temperature increased while the density did not significantly differ. At 40 ◦ C, the waste viscosity was 37 cp; at 80 ◦ C, the viscosity decreased to 15 cp. According to Eq. (2), the settling velocity of 1.2 m Si particles in the obtained slurry waste should be approximately 4.8 times faster at 80 ◦ C than at 25 ◦ C. Therefore, a higher recovery rate was predicated at 80 ◦ C.
Table 1 The solid content, turbidity and appearance of the prepared PAG-Si suspension. Solid content (g/L) Turbidity (NTU)
Appearance a
Unused PAG cutting liquid.
0a 0.5
0.0025 31
0.0075 109
0.1 468
0.5 >1000
T.-H. Tsai, Y.-P. Shih / Journal of Hazardous Materials 271 (2014) 252–257
(a)
o
25 C o 40 C o 80 C
80
Water Ethanol Acetone
60
Viscosity (cp)
Solid content (g/L)
100
255
10
40
20
1 0
5
10
15
20
0 0
Time (day)
20
Fig. 3. The solid content of the recovered liquid after sedimentation at 25 ◦ C, 40 ◦ C and 80 ◦ C for different days.
60
80
(b)
1.2
Water Ethanol Acetone
1.1
3
Density (g/cm )
Fig. 3 shows the solid content of the recovered liquid after sedimentation at 25, 40 and 80 ◦ C for different days. The solid content decreased over time during the first 7 days, and the decrease at 80 ◦ C was larger than that at 25 or 40 ◦ C. The decreasing extent at 80 ◦ C reduced after 7 days, and that at 25 or 40 ◦ C also went down after 14 days. After 21 days, the solid content at 80 ◦ C decreased to 5.13 g/L, and the particle removal fraction reached 99.5%. However, the turbidity of the recovered liquid still exceeded 1000 NTU. The PSDs of the recovered liquids at various temperatures in Fig. 4 showed that the main peak of the particle size approximated 0.2 m after sedimentation for 14 days. These submicron sized particles suspended stably and were difficult to separate from PAG liquid by gravitational force. Therefore, a cutting liquid with low nephelometric turbidity units could not be obtained. Water, ethanol and acetone were then used as diluents to decrease the viscosity and density of the slurry waste. Fig. 5 shows that both the viscosity and density of the slurry waste decreased with the addition of the diluents, especially adding acetone and ethanol. Based on this finding, the settling velocity (Vd ) was calculated at various diluent concentrations by using Eq. (2). Fig. 6 shows the effect of adding diluents on Vd with respect to the original velocity (Vo ). The curves in Fig. 6 predicted that the addition of diluents could certainly increase the settling velocity, and adding ethanol provided the fastest settling. For example, when the
1.0
0.9
0.8
0
20
40
60
80
Diluent (wt.%) Fig. 5. The effect of adding diluents on viscosity and density of the slurry waste.
40 Water Ethanol Acetone
30 20 o
15
80 C Vd/Vo
10 5
Volume (%)
40
Diluent (wt.%)
20 15
20
o
40 C
10
10
5 20 15
o
25 C
10
0
5 0 0.1
0 1
10
100
20
40
60
80
Diluent (wt.%)
Particle size (µm) Fig. 4. Particle size distribution of the recovered liquid after sedimentation at 25 ◦ C, 40 ◦ C and 80 ◦ C for 14 days.
Fig. 6. The settling velocity with respect to the original velocity (Vd /Vo ) after adding various concentrations of diluents.
256
T.-H. Tsai, Y.-P. Shih / Journal of Hazardous Materials 271 (2014) 252–257
100
Solid content (g/L)
Table 2 The solid content, turbidity and appearance of the unused PAG and the upper liquid after sedimentation.
Origin Water Ethanol Acetone
Liquid
Unused PAG
Upper liquid
Solid content (g/L) Turbidity (NTU)
0 0.5
0.122 44
10
1
0.1 0
5
10
15
20
25
Time (day) Fig. 7. The solid content of the upper liquid after sedimentation with 50 wt.% diluents for different days.
Volume (%)
diluent concentration was 50 wt.% for 1.2 m Si particles, Vd /Vo was 18.72 after adding ethanol and 9.92 after adding acetone but only 3.67 after adding water, which forecasted that recovery rate could be increased by using a diluent, ethanol especially. Fig. 7 shows the effect of adding 50 wt.% diluent on the solid content of the upper liquid after sedimentation for different days. The data indicated that all solid content decreased after addition of diluents. Initially, as forecasted by Vd /Vo in Fig. 6, the solid content was lower for adding ethanol or acetone compared to that for adding water. After 3 days, the solid content was below 3 g/L by adding ethanol or acetone, but their decreasing extent reduced. After sedimentation for 21 days, the solid content decreased to 0.304 g/L by adding ethanol and decreased to 0.724 g/L by adding acetone. Despite the low solid content, the turbidity still exceeded 1000 NTU with the addition of ethanol or acetone. Although addition of water initially obtained a smaller decrease in solid content, the effective reduction of solid content was sustained for 14 days. After 14 days, the solid content after addition of water was lower than that after addition of ethanol or acetone. After 21 days of sedimentation, adding water lowered the solid content to 0.122 g/L and decreased turbidity to 44 NTU. Table 2 shows that the upper liquid with 44 NTU had a clean appearance and resembled an unused PAG liquid.
20 15 10 5 20 15 10 5 20 15 10 5 0 0.1
Water
Ethanol
Acetone
1
10
100
Particle size (µm) Fig. 8. Particle size distribution of the upper liquid after sedimentation with 50 wt.% diluents for 14 day.
Appearance
Fig. 8 shows the PSDs of the upper liquids by sedimentation after addition of diluents for 14 days. The data show the main peak of the particle size was around 0.15 m after addition of ethanol or acetone. These submicron particles suspended stably in PAG–ethanol or PAG–acetone solutions and were difficult to separate from PAG liquid by gravitational sedimentation. However, after sedimentation with the addition of water for 14 days, most particles in the upper liquid were 3.4 m (Fig. 8). The particles in PAG–water solution were larger than those in PAG–ethanol and PAG–acetone solutions. It means that the particles aggregated when water was used as a diluent. The dispersion or aggregation of particles in suspension depends on surface potentials and the presence of adsorbed layers that physically limit the approach of particles, i.e. steric stabilization. Assume Si and SiC surfaces in the spent waste produced SiO2 due to oxidation [17]. According to the references, the electrophoretic mobilities of SiO2 in water, ethanol and acetone have been measured, and the data are −1.7 × 10−8 m2 V−1 s−1 [18], −9.5 × 10−9 m2 V−1 s−1 [17], −2.3 × 10−8 m2 V−1 s−1 [19], respectively. Because the particles are small and the medium are non-electrolytic solutions, the electrophoretic mobilities could be converted to surface potentials using Hückel equation [20]. Thus, the surface potentials of SiO2 in water, ethanol and acetone are −38.87 mV, −75.19 mV and −61.00 mV, respectively, showing submicron particles would suspend more stably in ethanol and in acetone than in water. In Fig. 7, because the settling particles were large initially, the gravitational force would dominate the settling velocity. As forecasted by Vd /Vo in Fig. 6, addition of water obtained a smaller decrease in solid content than addition of ethanol or acetone. The surface potentials became important for the settling of submicron particles. Due to the lower surface potential value in water than that in ethanol or acetone, the particles aggregated easily in water as confirmed by the PSDs of Fig. 8. Therefore, after 3 days, the decreasing extent of solid content reduced by adding ethanol or acetone, but it reduced sustainably for 14 days by adding water. Furthermore, the settling medium included PAG molecules. The PAG molecules could adsorb on particles to form steric stabilization. In PAG–ethanol or PAG–acetone solutions, PAG molecules adsorbed on Si or SiC surfaces would prevent particle aggregation
T.-H. Tsai, Y.-P. Shih / Journal of Hazardous Materials 271 (2014) 252–257 Table 3 The properties of the unused PAG and the recovered liquid by sedimentation and vacuum-distillation. Liquid (25 ◦ C)
Unused PAG
Recovered liquid
Viscosity (cp) Density (g/cm3 ) Turbidity (NTU) Water content (wt.%)
36.6 0.994 0.5 0.251
38.6 0.997 0.37 0.246
257
solution because, in PAG–ethanol or PAG–acetone solutions, PAG molecules adsorbed on Si or SiC surfaces prevent particle aggregation due to steric stabilization. In PAG–water solution, Si or SiC particles had hydrophilic surfaces, and the attraction of hydrogen bonding between water–water or between PAG–water would interfere with the adsorption of PAG molecules on particles and weaken the steric stabilization. Thus, the particles would aggregate readily, causing rapid settling and facile separation from PAG–water solution. After sedimentation by adding 50 wt.% water for 21 days, the solid content of the upper liquid decreased to 0.122 g/L, and the turbidity reached 44 NTU. The obtained upper liquid was then vacuum-distillated at 60 torr and 70 ◦ C to remove water. The turbidity of the final recovered PAG with 0.246 wt.% water was only 0.37 NTU. The recovery percentage of glycol was 81%. Acknowledgements The authors would like to thank the National Science Council Taiwan for financially supporting this research under Contract No. NSC 101-2221-E-027-109-MY3, and thank Sino-American Silicon Products, Inc. for material assistance.
Appearance
due to steric stabilization [21]. However, in PAG–water solution, formation of oxide enhanced the hydrophilicity of particles [22]. Additionally, among these three diluents, water has the highest chemical polarity and the largest hydrogen bonding component of Hansen solubility parameters [23], indicating a different interactions by using water as a diluent. In PAG–water solution, the attraction of hydrogen bonding happened between water–water or between PAG–water interfered with the adsorption of PAG molecules on particles, and weakened the steric stabilization. The particles were attracted to each other in the PAG–water solution and aggregated into larger particles. The large particles readily settled and then separated from PAG–water solution. Finally, a low-turbidity liquid was obtained. After sedimentation with the addition of 50 wt.% water for 21 days, the obtained upper liquid was vacuum-distillated at 60 torr and 70 ◦ C to remove water. Table 3 shows that the final recovered product contained only 0.246 wt.% water, and the recovered liquid and the unused PAG had similar viscosity and density. Moreover, the turbidity of the recovered PAG was only 0.37 NTU. The estimated recovery percentage of glycol reached 81% according to Eq. (1). Water is the suggested diluent for mixing with the spent slurry waste, and sedimentation can be performed at room temperature without a complex apparatus. After sedimentation, water removal was performed by vacuum distillation at low temperature, such as 70 ◦ C or below 70 ◦ C. Thus, the obtained liquid had a low turbidity and could be reused in wide applications. 4. Conclusions This study recovered PAG cutting liquids from silicon slurry waste by sedimentation. The settling velocity of particles and separation performance were enhanced by adjusting the temperature and using diluents. Although a particle removal fraction exceeding 99.5% could be achieved by increasing temperature or by adding ethanol or acetone, the recovered liquids were still opaque and turbid with larger than 1000 NTU due to the stable suspension of submicron particles. When water was used as the diluent, the PSD results showed that the particles tended to aggregate in PAG–water
References [1] D. Sarti, R. Einhaus, Silicon feedstock for the multi-crystalline photovoltaic industry, Sol. Energy Mater. Sol. Cells 72 (2002) 27–40. [2] T. Neesse, Review on SiC-recycling in wafer sawing operation, Interceram 55 (2006) 430–432. [3] T.H. Tsai, Y.P. Shih, Y.F. Wu, Recycling silicon wire saw slurries: separation of silicon and silicon carbide in a ramp settling tank under an applied electrical field, J. Air Waste Manag. Assoc. 63 (2013) 512–517. [4] W. Yu, D.A. Aguilar, J.B. Cuthbert, L.Y.P. Zhu, Cutting fluids with improved performance, U.S. Patent (2012) 20120196779. [5] C. Zavattari, G. Fragiacomo, Method for the separation, regeneration and reuse of an exhausted glycol-based slurry, U.S. Patent (2001) 6231628. [6] Y.L. Sun, Y.H. Liu, M.Z. Tsai, Method for recycling cutting fluid, U.S. Patent (2012) 20120073952. [7] K. Toyama, System for reusing oily slurry waste fluid, U.S. Patent (1998) 5830369. [8] G. Gaudet, S. Grumbine, N. Naguib, F. Batllo, Wire saw slurry recycling process, U.S. Patent (2009) 20090293369. [9] K.S. Chang, Separation of waste cutting oil generated from solar wafer slicing process, T.W. Patent (2011) 098142647. [10] J.J. Iang, Method and apparatus for recycling and treating wastes of silicon wafer cutting and polishing processes, U.S. Patent (2012) 20120312747. [11] N. Drouiche, M.W. Naceur, H. Boutoumi, N. Aitmessaoudene, R. Henniche, T. Ouslimane, Assessment of the recovery of photovoltaic cells cutting fluid by chemical pretreatment and ultrafiltration, Desalination Water Treat. 51 (2013) 713–716. [12] L. Liu, Recycling cutting fluid components from crystal silicon cutting waste mortar comprises homogenizing agent into waste mortar of crystal silicon cutting liquid, and preparing homogenizing agent and water in each waste mortar, C.N. Patent (2012) 102399619. [13] T.Y. Yao, Method for recovering water-soluble cutting fluid from silicon wafer cutting fluid, C.N. Patent (2012) 102746934. [14] R.U. Owolabi, K.A. Alabi, O.B. Oke, Reclamation of spent automobile engine lubricating oil, Fount. J. Natl. Appl. Sci. 2 (2013) 11–16. [15] D. Feng, L. Liu, P. Wang, H. Zhang, X. Xie, The experimental study of natural sedimentation in heavy oil, Adv. Mater. Res. 413 (2012) 60–64. [16] P.G.W. Hawksley, The effect of concentration on the settling of suspensions and flow through porous media, Some Aspects of Fluid Flow, Edward Arnold, London, 1951, pp. 114–118. [17] Y. Fukada, P.S. Nicholson, The role of Si-O species in the colloidal stability of silicon-containing ceramic powders, J. Eur. Ceram. Soc. 24 (2004) 17–23. [18] W.J. Masschelein, Unit Processes in Drinking Water Treatment, CRC Press, New York, 1992. [19] H.A. Ketelson, R. Pelton, M.A. Brook, Colloidal stability of stöber silica in acetone–water mixtures, J. Colloid Interface Sci. 179 (1996) 600–607. [20] E. Hückel, Die kataphorese der kugel, Physiol. Zool. 25 (1924) 204–210. [21] D.H. Napper, Polymeric Stabilization of Colloidal Dispersions, Academic Press, London, 1983. [22] A. Feng, B.J. McCoy, Z.A. Munir, D.E. Cagliostro, Water adsorption and desorption kinetics on silica insulation, J. Colloid Interface Sci. 180 (1996) 276–284. [23] C.M. Hansen (Ed.), Hansen Solubility Parameters: A User’s Handbook, 2nd ed., CRC Press, Boca Raton, FL, 2007.