- Email: [email protected]
Research on high gradient magnetic separation of pneumatic conveyed powder products: Investigation from the viewpoint of interparticle interactions
Physica C 484 (2013) 329–332
Contents lists available at SciVerse ScienceDirect
Physica C journal homepage: www.elsevier.com/locate/physc
Research on high gradient magnetic separation of pneumatic conveyed powder products: Investigation from the viewpoint of interparticle interactions Kohei Senkawa ⇑, Yuki Nakai, Fumihito Mishima, Yoko Akiyama, Shigehiro Nishijima Graduate School of Engineering, Osaka University, A1 Bldg., 2-1 Yamadaoka, Suita, Osaka 565-0871, Japan
a r t i c l e
i n f o
Article history: Accepted 26 March 2012 Available online 30 March 2012 Keywords: Magnetic separation Powder Superconducting magnet Adhesion force Humidity
a b s t r a c t The separation and removal of the metallic debris originating from pipe of manufacturing line are required in the manufacturing process of the fine particle products. In this study, we develop a high gradient magnetic separation system (HGMS) under a dry process by using a superconducting magnet to remove ferromagnetic particles such as the material stainless steel (SUS). To avoid the obstruction of the separation part by aggregation of the processed material, we develop a magnetic separation system using a pneumatic conveying as a new transportation method of the particles. The magnetic separations were experimented under the same conditions on different days, but the results were different. The reason is considered to be the difference in adhesion force between the particles due to a change of humidity, we have measured the adhesion forces between the ferromagnetic particles and the paramagnetic medium particles using AFM (Atomic Force Microscope) while changing the humidity. As a result, the adhesion force between the particles increased with the increasing of humidity. Furthermore, we saw that the effect of relative humidity was larger in the adhesion force of alumina with larger cohesive property. Based on these results, an appropriate condition of the separation experiment was clarified. And a dehumidification mechanism was introduced. Ó 2012 Elsevier B.V. All rights reserved.
1. Introduction In manufacturing processes of fine powder products such as foods, medicines or industrial materials, there is an issue of interfusion by metallic debris originated from manufacturing lines such as pipes and moving parts, separation and removal of these impurities are required. To solve the problem, we have developed a high gradient magnetic separation system (HGMS) under dry process by using a superconducting magnet to separate the ferromagnetic particles such SUS particles. Generally, dry separation has the advantage that drying process is not required in comparison with wet separation. However, in dry HGMS method, a blockage of magnetic filters due to aggregation and deposition of the powders is a serious problem. Our previous study reported the possibility to reduce the blockage by using filters designed with consideration of a repose angle for the powders [1]. In this study, for the fundamental resolution of the blockage, we have developed a magnetic separation system using a pneumatic conveying as a new transportation method of particles [2–6]. In this paper, we examine the reason of difference in results of magnetic separation efficiency on different days despite of same
experimental conditions. One of the possibilities is considered to be a change in an interparticle interaction depending on relative humidity, thus adhesion forces acting between the ferromagnetic particle and the paramagnetic medium particle are measured using AFM (Atomic Force Microscopy). Since we had already succeeded to measure the adhesion force between single particles [7]. A further investigation for a relationship between relative humidity and the adhesion forces was conducted. Based on the relation between the adhesion forces and relative humidity, appropriate experimental conditions and a design of magnetic separation system are discussed. 2. Theory of magnetic separation The magnetic separation is a method to capture and to separate targeted particles selectively by the difference of magnetic forces acting on the ferromagnetic and the paramagnetic particle. In the magnetic separation from a fluid, forces acting on a ferromagnetic particle are mainly a magnetic force, a drag force and a gravity force. These forces are respectively shown by Eqs. (1)–(3),
FM ¼ ⇑ Corresponding author. Tel.: +81 6 6879 7898; fax: +81 6 6879 7889. E-mail addresses: [email protected] (K. Senkawa), yoko-ak @qb.see.eng.osaka-u.ac.jp (Y. Akiyama). 0921-4534/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.physc.2012.03.057
4 3 4 @ @ @ H prp ðM rÞH ¼ pr3p Mx þ My þ Mz 3 3 @x @y @z
1 p FD ¼ CD qU2 r2p 2 4
ð1Þ
ð2Þ
330
K. Senkawa et al. / Physica C 484 (2013) 329–332
FG ¼ qp
4 3 pr g 3 p
ð3Þ
where M is the magnetization (Wb/m2), H is the external magnetic field intensity (A/m), CD is the drag coefficient (–), q is the medium density (kg/m3), U is the flow velocity (m/s), rp is the particle radius (m), qp is the particle density (kg/m3) and g is the gravity acceleration (m/s2). The resistance coefficient CD is determined by Reynolds number Re(-). The Reynolds number is a ratio of 4 fictitious force and a viscous force in the fluid as shown as follows:
Re ¼
Ud
v
ð4Þ
where U is the flow velocity (m/s), d is the characteristic length (m) (here corresponds to particle diameter), m is the kinetic viscosity (m2/s). When the Reynolds number is small enough (Re < 1) in a laminar flow, the drag force can be calculated by the following Eq. (5) which is indicated as Stokes’ resistance law:
FD ¼ 6plrp U
3.2. Results and discussion
ð5Þ
and l is the viscosity of the medium (Pss). The possibility of magnetic separation can be basically estimated from a balance among the magnetic force, the drag force and the gravity force. In order to achieve the magnetic separation, the magnetic force should be bigger than the resultant of the drag force and the gravity force. When the ferromagnetic and the paramagnetic particles are adhered to each other to form aggregate, the aggregate are attracted by magnetic force FM, and they would be dispersed only when the drag force FD as a shear stress is bigger than the adhesion force FA acting between the ferromagnetic and paramagnetic particles. However, when the drag force is bigger than the magnetic force, the ferromagnetic particle are blown away without being separated to magnetic filters. If the drag force is smaller than the adhesion force, the ferromagnetic particle will be separated together with the medium particles as the aggregate without dispersion. In order to separate the ferromagnetic particles selectively by magnetic separation, the following Eq. (6) must be satisfied.
FM > FD > FA
tus is the same that we published in our previous papers [7]. This system consists of fluidized bed part (200 mm in flow path and 16 mm in pipe diameter) and magnetic separation part (600 mm in flow path and 16 mm in inner diameter). First of all, in the fluidized-bed part, the powder samples and moving mediums (3 g of polystyrene pellets, 1 mm in diameter) were fluidized by blowing compressed air at flow late of 1 L/s. The air flow velocity was decided to 5 m/s considering the adhesion force obtained from adhesion measurements by atomic force microscope (AFM). And then, only dispersed particles through 80 mesh filter were conveyed into the bore of the superconducting magnet. In the magnetic separation part, a magnetic filter (5 meshes, SUS430) was inserted under an external magnetic field of 2.0 T, and ferromagnetic SUS304 beads were separated selectively. The particles which were not captured by the magnetic filter were fed into the upper part of the superconducting magnet, and were collected into a bag which prevents back-flow.
ð6Þ
The drag force FD acting on the aggregates should be larger than the adhesion force between the particles FA, and also smaller than the magnetic force FM. To determine experimental conditions based on the above relation, it is necessary to evaluate the adhesion force FA. In addition, the adhesion force could be strongly influenced by the relative humidity [9]. Therefore, in this study, the magnetic separation experiment was conducted, and differences of results in separation efficiency were discussed considering the changes in the adhesion forces because of the humidity. In order to consider the appropriate separation processing conditions, we have examined the relationship between humidity and the adhesion force acting between the paramagnetic and ferromagnetic particles.
Fig. 1 and Table 1 respectively show the experimental results and the experimental conditions. High separation efficiency of 90% in both Sample A and B were obtained. This result indicates a usefulness of this magnetic separation system. As expected, the impurities included in Sample B were separated completely by magnetic separation on both days. On the other hand, considering the separation efficiency of ferromagnetic impurities in Sample A, the separation efficiency on Day 2 was much higher than that on Day 1. The reason is considered to be the lower relative humidity on Day 2. Extremely large aggregates are not formed due to higher dispersion of alumina particles in the low-humidity atmosphere, which makes the separation efficiency in Day 2 higher. To verify this idea, we have investigated adhesion forces between the particles while changing humidity. For the removal of impurities in food products, separation efficiency of 100% is required. The reason why the separation efficiency do not accomplish 100% could be that the particles were not dispersed sufficiently in the fluidized bed, and were blown away without being captured by magnetic filters because of a bigger drag force than magnetic force. On the basis of these results, it is necessary to improve the magnetic separation equipment and conditions. 4. Adhesion measurements 4.1. Experiment The adhesion forces acting between the paramagnetic and ferromagnetic particles used in this study were measured. Before
3. Magnetic separation experiment 3.1. Experimental method In this experiment, two kinds of the particles with different cohesiveness were used as powder medium. Sample A was alumina particles with high cohesiveness (average particle diameter 5 lm) including ferromagnetic SUS beads (average particle diameter 35 lm, SUS304) at a rate of 0.1 wt.%. Sample B was silica particles with low cohesiveness (average particle diameter 20 lm) including the above-mentioned ferromagnetic SUS beads at the rate of 0.1 wt.%. In this study, a superconducting solenoid magnet (100 mm in bore diameter and 460 mm in length) was used for the magnetic separation apparatus designed for powder separation. The experimental appara-
Fig. 1. Separation efficiency of Sample A and B.
K. Senkawa et al. / Physica C 484 (2013) 329–332
4.2. Results of adhesion measurements
Table 1 Experimental conditions.
Day 1 Day 2
331
Temperature (°C)
Relative humidity (%)
17.9 23.7
72 55
the adhesion force measurements, morphologies of the sample mixtures were observed using an optical microscope and SEM. SUS particles were caught in alumina particles in Sample A with strong cohesive property, and well dispersed in Sample B with weak cohesive property. The adhesion forces between objective particle (SUS) and medium particles were measured with the contact mode of Atomic Force Microscope (AFM; SPA-300HV, SII Nano Technology Inc.). AFM can estimate the atomic force between a cantilever tip and atoms by setting samples within 20 nm from a cantilever. The adhesion forces acting between particles attached to AFM cantilever tip and pelletized particle by pressing fixed on the sample stage were measured by choroid probe AFM method [7–8]. The basic measurement technique is as the same described in previous papers [9]. The spring constant of the cantilever (SI-AF01A, SII) made of silicon was 0.32 N/m, and the height of square pyramid shaped AFM cantilever tip was 12.5 lm. First of all, a small amount of epoxy resin was adhered at the tip of AFM cantilever tip by moving the AFM sample stage up and down, until is uniformly coated with resin by spin coater. Then a SUS particle was adhered on the tip of AFM cantilever tip in the same way. The cantilever attached with a SUS particle was dried for more than 24 h at room temperature. Before adhesion measurements, the attachment morphology of the particle to cantilever was checked by SEM (Scanning Electron Microscope). Fig. 2 shows the SEM photograph of cantilever attached with a SUS stainless steel bead (35 lm of mean particle diameter, SUS304). From the SEM image, it was confirmed that a SUS particle was attached to the cantilever tip. Moreover, it was clear that the SUS particle was not covered with the epoxy resin which was also confirmed by EDX (Energy Dispersive X-ray Spectroscopy). Under the above conditions, the relationship between the adhesion forces and moisture was examined by using air humidity controller (Rigaku Co. Ltd., HUM-1) while changing humidity. The relative humidity was changed from 10% to 90% by 20% steps. The temperature was kept constant at 30 °C. The number of measurements between alumina and SUS, silica and SUS were 40 times for each condition, and the average measurement value of adhesion force was calculated.
As previously noted, it was required to control the fluid velocity according to the interparticle interaction between the particles. We measured interparticle interaction of test sample with AFM in order to check up the appropriate separation conditions. The results are indicated in Fig. 3. At all relative humidity, the adhesion forces between alumina particles and SUS particles were bigger than those between silica particles and SUS particles. The adhesion forces of silica were linear with humidity, the adhesion forces of alumina rapidly increased while relative humidity increased from 40% to 50%. Since alumina has a large contact surface due to its plate-shaped crystal, it is considered that many water molecules adhered to the surface enlarging the adhesion force. On the other hand, silica has small contact surface area compared with alumina. Furthermore, the reason why adhesion forces of silica are small compared with alumina is that there are few water molecules on the surface area since the water molecules are taken into the pores of particles by the dehumidification function of silica. It showed that adhesion forces of both alumina and silica become small when humidity is low. It can be considered that the experimental results on previous magnetic separation (Day 1) were influenced by humidity. From these results, it may be thought that a higher separation efficiency could be achieved under much lower humidity. 4.3. Discussion on the experimental conditions Based on the above results, the flow velocity required for the magnetic separation was examined according to the humidity of the experimental conditions. First of all, when the flow velocities were changed, the drag forces acting on the particle of 35 lm in diameter were calculated. The results are shown in Fig. 4. From the observation by optical microscope, it can be considered that alumina and silica actually exist as the secondary particle as large as a SUS bead, and Fig. 4 can also be adapted to these particles such alumina and silica. The calculation conditions were set as follows: aggregated particle radius r = 1.75 105 m, density of air as the medium fluids q = 1.17 kg/m3, kinetic viscosity m = 15.0 106 m2/s. The drag forces were calculated for a flow velocity of U = 0.6–15 m/s. Based on the results of the magnetic separation experiment, the magnetic separation system should operate maintaining the relative humidity low in order to increase the separation efficiency. For this reason, it is necessary to improve the separation system to supply the sufficiently dried air from the air compressor. For example, a magnetic separation system such the one in Fig. 5 is
Fig. 2. SEM photograph of the cantilever tip attached SUS particle, (a) from under of the cantilever, (b) from side of the cantilever.
332
K. Senkawa et al. / Physica C 484 (2013) 329–332
Fig. 3. Relationship between adhesion force and relative humidity for silica and alumina particle against a SUS particle.
practically possible. In this system, the inflow air passes through the dehumidification system of a unit filled with calcium chloride, and then the dried airflow runs into the fluidized bed. When the humidity in the channel introduced through a dehumidifier mechanism was decreased to about 30%, the adhesion force between SUS particle and alumina and silica were respectively about 2.6 108 N and 1.5 108 N, as shown in Fig. 3. As mentioned above, it is necessary that the drag force FD acting on the aggregates is bigger than the adhesion force between the particles FA, and smaller than the magnetic force FM. Therefore, in order to realize a flow velocity which exceeds the adhesion force, the flow velocity of at least about 8 m/s and 5 m/s is required for Sample A and B respectively. Because the magnetic force acting on ferromagnetic particle of 35 lm is calculated to be 106–105 N, it is expected that the magnetic separation would be achieved with these flow velocities as long as extremely large secondary particles are not generated. In addition, magnetic separation becomes possible with lower flow velocity and under low humidity, which also reduces the yield loss of the medium powder by the outflow of large secondary particles. In the future, we will try to verify the dehumidifying system experimentally. 5. Conclusion
Fig. 4. Drag force of as a function of flow speed (u35 lm) acts on the particle of 35 lm in diameter.
In this study, we tried to separate SUS particle selectively from a mixture of alumina/silica and SUS particle, and the possibility of HGMS with pneumatic convey system was confirmed as a new method. Considering the magnetic separation experiment with superconducting magnet, high separation efficiencies were obtained for both samples, alumina with high cohesive property and silica with low cohesive property. However, the separation efficiency of the sample using alumina as a medium with relatively high cohesive property did not approach 100%, and the reason was considered to be high humidity. It is possible that re-aggregation of the powder during air conveyance by high humidity causes decrease in separation efficiency. To verify this idea, the adhesion forces between alumina and SUS and that between silica and SUS were measured using colloid probe AFM method while changing the relative humidity. The measurement results showed a significant difference between the two samples, and also showed that the relative humidity greatly contributes to the adhesion forces, in particular for the particle with high cohesive property. Considering these data, the conditions of the separation experiment were examined. In order to improve the magnetic separation efficiency of Sample A, dry condition and control of the flow velocity is required. Introduction of a dehumidifier in the system will be effective to reduce the humidity. We will try the magnetic separation in dry conditions, and the simulations of the magnetic separation process will be performed using the data of the humidity dependency with adhesion force. References
Fig. 5. A dry HGMS system with a dehumidification mechanism.
[1] Y. Nakai, F. Mishima, Y. Akiyama, S. Nishijima, Physica C 470 (2010) 1812–1817. [2] F. Mishima, S. Yamazaki, K. Yoshida, H. Nakane, S. Yoshizawa, S. Takeda, Y. Izumi, S. Nishijima, IEEE Trans. Appl. Supercond. 14 (2004) 1561–1564. [3] F. Mishima, T. Terada, Y. Akiyama, Y. Izumi, H. Okazaki, S. Nishijima, IEEE Trans. Appl. Supercond. 14 (2008) 824–827. [4] F. Mishima, S. Yamazaki, K. Yoshida, H. Nakane, S. Takeda, Y. Izumi, S. Nishijima, J. Magn. Soc. Jpn. 28 (2004) 393–396. [5] Y. Nakai, K. Senkawa, F. Mishima, Y. Akiyama, S. Nishijima, IEEE Trans. Appl. Supercond. 21 (2010) 2063–2067. [6] Y. Nakai, K. Senkawa, F. Mishima, Y. Akiyama, S. Nishijima, Physica C 471 (2011) 1533–1537. [7] Y. Mizuguchi, T. Miyamoto, Konica Minolta Technol. Rep. 1 (2004) 19–22. [8] M. Tanaka, M. Komaagata, M. Tsukada, H. Kamiya, Powder Technol. 183 (2008) 273–281. [9] K. Senkawa, Y. Nakai, F. Mishima, Y. Akiyama, S. Nishijima, Physica C 471 (2011) 1525–1529.
Contents lists available at SciVerse ScienceDirect
Physica C journal homepage: www.elsevier.com/locate/physc
Research on high gradient magnetic separation of pneumatic conveyed powder products: Investigation from the viewpoint of interparticle interactions Kohei Senkawa ⇑, Yuki Nakai, Fumihito Mishima, Yoko Akiyama, Shigehiro Nishijima Graduate School of Engineering, Osaka University, A1 Bldg., 2-1 Yamadaoka, Suita, Osaka 565-0871, Japan
a r t i c l e
i n f o
Article history: Accepted 26 March 2012 Available online 30 March 2012 Keywords: Magnetic separation Powder Superconducting magnet Adhesion force Humidity
a b s t r a c t The separation and removal of the metallic debris originating from pipe of manufacturing line are required in the manufacturing process of the fine particle products. In this study, we develop a high gradient magnetic separation system (HGMS) under a dry process by using a superconducting magnet to remove ferromagnetic particles such as the material stainless steel (SUS). To avoid the obstruction of the separation part by aggregation of the processed material, we develop a magnetic separation system using a pneumatic conveying as a new transportation method of the particles. The magnetic separations were experimented under the same conditions on different days, but the results were different. The reason is considered to be the difference in adhesion force between the particles due to a change of humidity, we have measured the adhesion forces between the ferromagnetic particles and the paramagnetic medium particles using AFM (Atomic Force Microscope) while changing the humidity. As a result, the adhesion force between the particles increased with the increasing of humidity. Furthermore, we saw that the effect of relative humidity was larger in the adhesion force of alumina with larger cohesive property. Based on these results, an appropriate condition of the separation experiment was clarified. And a dehumidification mechanism was introduced. Ó 2012 Elsevier B.V. All rights reserved.
1. Introduction In manufacturing processes of fine powder products such as foods, medicines or industrial materials, there is an issue of interfusion by metallic debris originated from manufacturing lines such as pipes and moving parts, separation and removal of these impurities are required. To solve the problem, we have developed a high gradient magnetic separation system (HGMS) under dry process by using a superconducting magnet to separate the ferromagnetic particles such SUS particles. Generally, dry separation has the advantage that drying process is not required in comparison with wet separation. However, in dry HGMS method, a blockage of magnetic filters due to aggregation and deposition of the powders is a serious problem. Our previous study reported the possibility to reduce the blockage by using filters designed with consideration of a repose angle for the powders [1]. In this study, for the fundamental resolution of the blockage, we have developed a magnetic separation system using a pneumatic conveying as a new transportation method of particles [2–6]. In this paper, we examine the reason of difference in results of magnetic separation efficiency on different days despite of same
experimental conditions. One of the possibilities is considered to be a change in an interparticle interaction depending on relative humidity, thus adhesion forces acting between the ferromagnetic particle and the paramagnetic medium particle are measured using AFM (Atomic Force Microscopy). Since we had already succeeded to measure the adhesion force between single particles [7]. A further investigation for a relationship between relative humidity and the adhesion forces was conducted. Based on the relation between the adhesion forces and relative humidity, appropriate experimental conditions and a design of magnetic separation system are discussed. 2. Theory of magnetic separation The magnetic separation is a method to capture and to separate targeted particles selectively by the difference of magnetic forces acting on the ferromagnetic and the paramagnetic particle. In the magnetic separation from a fluid, forces acting on a ferromagnetic particle are mainly a magnetic force, a drag force and a gravity force. These forces are respectively shown by Eqs. (1)–(3),
FM ¼ ⇑ Corresponding author. Tel.: +81 6 6879 7898; fax: +81 6 6879 7889. E-mail addresses: [email protected] (K. Senkawa), yoko-ak @qb.see.eng.osaka-u.ac.jp (Y. Akiyama). 0921-4534/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.physc.2012.03.057
4 3 4 @ @ @ H prp ðM rÞH ¼ pr3p Mx þ My þ Mz 3 3 @x @y @z
1 p FD ¼ CD qU2 r2p 2 4
ð1Þ
ð2Þ
330
K. Senkawa et al. / Physica C 484 (2013) 329–332
FG ¼ qp
4 3 pr g 3 p
ð3Þ
where M is the magnetization (Wb/m2), H is the external magnetic field intensity (A/m), CD is the drag coefficient (–), q is the medium density (kg/m3), U is the flow velocity (m/s), rp is the particle radius (m), qp is the particle density (kg/m3) and g is the gravity acceleration (m/s2). The resistance coefficient CD is determined by Reynolds number Re(-). The Reynolds number is a ratio of 4 fictitious force and a viscous force in the fluid as shown as follows:
Re ¼
Ud
v
ð4Þ
where U is the flow velocity (m/s), d is the characteristic length (m) (here corresponds to particle diameter), m is the kinetic viscosity (m2/s). When the Reynolds number is small enough (Re < 1) in a laminar flow, the drag force can be calculated by the following Eq. (5) which is indicated as Stokes’ resistance law:
FD ¼ 6plrp U
3.2. Results and discussion
ð5Þ
and l is the viscosity of the medium (Pss). The possibility of magnetic separation can be basically estimated from a balance among the magnetic force, the drag force and the gravity force. In order to achieve the magnetic separation, the magnetic force should be bigger than the resultant of the drag force and the gravity force. When the ferromagnetic and the paramagnetic particles are adhered to each other to form aggregate, the aggregate are attracted by magnetic force FM, and they would be dispersed only when the drag force FD as a shear stress is bigger than the adhesion force FA acting between the ferromagnetic and paramagnetic particles. However, when the drag force is bigger than the magnetic force, the ferromagnetic particle are blown away without being separated to magnetic filters. If the drag force is smaller than the adhesion force, the ferromagnetic particle will be separated together with the medium particles as the aggregate without dispersion. In order to separate the ferromagnetic particles selectively by magnetic separation, the following Eq. (6) must be satisfied.
FM > FD > FA
tus is the same that we published in our previous papers [7]. This system consists of fluidized bed part (200 mm in flow path and 16 mm in pipe diameter) and magnetic separation part (600 mm in flow path and 16 mm in inner diameter). First of all, in the fluidized-bed part, the powder samples and moving mediums (3 g of polystyrene pellets, 1 mm in diameter) were fluidized by blowing compressed air at flow late of 1 L/s. The air flow velocity was decided to 5 m/s considering the adhesion force obtained from adhesion measurements by atomic force microscope (AFM). And then, only dispersed particles through 80 mesh filter were conveyed into the bore of the superconducting magnet. In the magnetic separation part, a magnetic filter (5 meshes, SUS430) was inserted under an external magnetic field of 2.0 T, and ferromagnetic SUS304 beads were separated selectively. The particles which were not captured by the magnetic filter were fed into the upper part of the superconducting magnet, and were collected into a bag which prevents back-flow.
ð6Þ
The drag force FD acting on the aggregates should be larger than the adhesion force between the particles FA, and also smaller than the magnetic force FM. To determine experimental conditions based on the above relation, it is necessary to evaluate the adhesion force FA. In addition, the adhesion force could be strongly influenced by the relative humidity [9]. Therefore, in this study, the magnetic separation experiment was conducted, and differences of results in separation efficiency were discussed considering the changes in the adhesion forces because of the humidity. In order to consider the appropriate separation processing conditions, we have examined the relationship between humidity and the adhesion force acting between the paramagnetic and ferromagnetic particles.
Fig. 1 and Table 1 respectively show the experimental results and the experimental conditions. High separation efficiency of 90% in both Sample A and B were obtained. This result indicates a usefulness of this magnetic separation system. As expected, the impurities included in Sample B were separated completely by magnetic separation on both days. On the other hand, considering the separation efficiency of ferromagnetic impurities in Sample A, the separation efficiency on Day 2 was much higher than that on Day 1. The reason is considered to be the lower relative humidity on Day 2. Extremely large aggregates are not formed due to higher dispersion of alumina particles in the low-humidity atmosphere, which makes the separation efficiency in Day 2 higher. To verify this idea, we have investigated adhesion forces between the particles while changing humidity. For the removal of impurities in food products, separation efficiency of 100% is required. The reason why the separation efficiency do not accomplish 100% could be that the particles were not dispersed sufficiently in the fluidized bed, and were blown away without being captured by magnetic filters because of a bigger drag force than magnetic force. On the basis of these results, it is necessary to improve the magnetic separation equipment and conditions. 4. Adhesion measurements 4.1. Experiment The adhesion forces acting between the paramagnetic and ferromagnetic particles used in this study were measured. Before
3. Magnetic separation experiment 3.1. Experimental method In this experiment, two kinds of the particles with different cohesiveness were used as powder medium. Sample A was alumina particles with high cohesiveness (average particle diameter 5 lm) including ferromagnetic SUS beads (average particle diameter 35 lm, SUS304) at a rate of 0.1 wt.%. Sample B was silica particles with low cohesiveness (average particle diameter 20 lm) including the above-mentioned ferromagnetic SUS beads at the rate of 0.1 wt.%. In this study, a superconducting solenoid magnet (100 mm in bore diameter and 460 mm in length) was used for the magnetic separation apparatus designed for powder separation. The experimental appara-
Fig. 1. Separation efficiency of Sample A and B.
K. Senkawa et al. / Physica C 484 (2013) 329–332
4.2. Results of adhesion measurements
Table 1 Experimental conditions.
Day 1 Day 2
331
Temperature (°C)
Relative humidity (%)
17.9 23.7
72 55
the adhesion force measurements, morphologies of the sample mixtures were observed using an optical microscope and SEM. SUS particles were caught in alumina particles in Sample A with strong cohesive property, and well dispersed in Sample B with weak cohesive property. The adhesion forces between objective particle (SUS) and medium particles were measured with the contact mode of Atomic Force Microscope (AFM; SPA-300HV, SII Nano Technology Inc.). AFM can estimate the atomic force between a cantilever tip and atoms by setting samples within 20 nm from a cantilever. The adhesion forces acting between particles attached to AFM cantilever tip and pelletized particle by pressing fixed on the sample stage were measured by choroid probe AFM method [7–8]. The basic measurement technique is as the same described in previous papers [9]. The spring constant of the cantilever (SI-AF01A, SII) made of silicon was 0.32 N/m, and the height of square pyramid shaped AFM cantilever tip was 12.5 lm. First of all, a small amount of epoxy resin was adhered at the tip of AFM cantilever tip by moving the AFM sample stage up and down, until is uniformly coated with resin by spin coater. Then a SUS particle was adhered on the tip of AFM cantilever tip in the same way. The cantilever attached with a SUS particle was dried for more than 24 h at room temperature. Before adhesion measurements, the attachment morphology of the particle to cantilever was checked by SEM (Scanning Electron Microscope). Fig. 2 shows the SEM photograph of cantilever attached with a SUS stainless steel bead (35 lm of mean particle diameter, SUS304). From the SEM image, it was confirmed that a SUS particle was attached to the cantilever tip. Moreover, it was clear that the SUS particle was not covered with the epoxy resin which was also confirmed by EDX (Energy Dispersive X-ray Spectroscopy). Under the above conditions, the relationship between the adhesion forces and moisture was examined by using air humidity controller (Rigaku Co. Ltd., HUM-1) while changing humidity. The relative humidity was changed from 10% to 90% by 20% steps. The temperature was kept constant at 30 °C. The number of measurements between alumina and SUS, silica and SUS were 40 times for each condition, and the average measurement value of adhesion force was calculated.
As previously noted, it was required to control the fluid velocity according to the interparticle interaction between the particles. We measured interparticle interaction of test sample with AFM in order to check up the appropriate separation conditions. The results are indicated in Fig. 3. At all relative humidity, the adhesion forces between alumina particles and SUS particles were bigger than those between silica particles and SUS particles. The adhesion forces of silica were linear with humidity, the adhesion forces of alumina rapidly increased while relative humidity increased from 40% to 50%. Since alumina has a large contact surface due to its plate-shaped crystal, it is considered that many water molecules adhered to the surface enlarging the adhesion force. On the other hand, silica has small contact surface area compared with alumina. Furthermore, the reason why adhesion forces of silica are small compared with alumina is that there are few water molecules on the surface area since the water molecules are taken into the pores of particles by the dehumidification function of silica. It showed that adhesion forces of both alumina and silica become small when humidity is low. It can be considered that the experimental results on previous magnetic separation (Day 1) were influenced by humidity. From these results, it may be thought that a higher separation efficiency could be achieved under much lower humidity. 4.3. Discussion on the experimental conditions Based on the above results, the flow velocity required for the magnetic separation was examined according to the humidity of the experimental conditions. First of all, when the flow velocities were changed, the drag forces acting on the particle of 35 lm in diameter were calculated. The results are shown in Fig. 4. From the observation by optical microscope, it can be considered that alumina and silica actually exist as the secondary particle as large as a SUS bead, and Fig. 4 can also be adapted to these particles such alumina and silica. The calculation conditions were set as follows: aggregated particle radius r = 1.75 105 m, density of air as the medium fluids q = 1.17 kg/m3, kinetic viscosity m = 15.0 106 m2/s. The drag forces were calculated for a flow velocity of U = 0.6–15 m/s. Based on the results of the magnetic separation experiment, the magnetic separation system should operate maintaining the relative humidity low in order to increase the separation efficiency. For this reason, it is necessary to improve the separation system to supply the sufficiently dried air from the air compressor. For example, a magnetic separation system such the one in Fig. 5 is
Fig. 2. SEM photograph of the cantilever tip attached SUS particle, (a) from under of the cantilever, (b) from side of the cantilever.
332
K. Senkawa et al. / Physica C 484 (2013) 329–332
Fig. 3. Relationship between adhesion force and relative humidity for silica and alumina particle against a SUS particle.
practically possible. In this system, the inflow air passes through the dehumidification system of a unit filled with calcium chloride, and then the dried airflow runs into the fluidized bed. When the humidity in the channel introduced through a dehumidifier mechanism was decreased to about 30%, the adhesion force between SUS particle and alumina and silica were respectively about 2.6 108 N and 1.5 108 N, as shown in Fig. 3. As mentioned above, it is necessary that the drag force FD acting on the aggregates is bigger than the adhesion force between the particles FA, and smaller than the magnetic force FM. Therefore, in order to realize a flow velocity which exceeds the adhesion force, the flow velocity of at least about 8 m/s and 5 m/s is required for Sample A and B respectively. Because the magnetic force acting on ferromagnetic particle of 35 lm is calculated to be 106–105 N, it is expected that the magnetic separation would be achieved with these flow velocities as long as extremely large secondary particles are not generated. In addition, magnetic separation becomes possible with lower flow velocity and under low humidity, which also reduces the yield loss of the medium powder by the outflow of large secondary particles. In the future, we will try to verify the dehumidifying system experimentally. 5. Conclusion
Fig. 4. Drag force of as a function of flow speed (u35 lm) acts on the particle of 35 lm in diameter.
In this study, we tried to separate SUS particle selectively from a mixture of alumina/silica and SUS particle, and the possibility of HGMS with pneumatic convey system was confirmed as a new method. Considering the magnetic separation experiment with superconducting magnet, high separation efficiencies were obtained for both samples, alumina with high cohesive property and silica with low cohesive property. However, the separation efficiency of the sample using alumina as a medium with relatively high cohesive property did not approach 100%, and the reason was considered to be high humidity. It is possible that re-aggregation of the powder during air conveyance by high humidity causes decrease in separation efficiency. To verify this idea, the adhesion forces between alumina and SUS and that between silica and SUS were measured using colloid probe AFM method while changing the relative humidity. The measurement results showed a significant difference between the two samples, and also showed that the relative humidity greatly contributes to the adhesion forces, in particular for the particle with high cohesive property. Considering these data, the conditions of the separation experiment were examined. In order to improve the magnetic separation efficiency of Sample A, dry condition and control of the flow velocity is required. Introduction of a dehumidifier in the system will be effective to reduce the humidity. We will try the magnetic separation in dry conditions, and the simulations of the magnetic separation process will be performed using the data of the humidity dependency with adhesion force. References
Fig. 5. A dry HGMS system with a dehumidification mechanism.
[1] Y. Nakai, F. Mishima, Y. Akiyama, S. Nishijima, Physica C 470 (2010) 1812–1817. [2] F. Mishima, S. Yamazaki, K. Yoshida, H. Nakane, S. Yoshizawa, S. Takeda, Y. Izumi, S. Nishijima, IEEE Trans. Appl. Supercond. 14 (2004) 1561–1564. [3] F. Mishima, T. Terada, Y. Akiyama, Y. Izumi, H. Okazaki, S. Nishijima, IEEE Trans. Appl. Supercond. 14 (2008) 824–827. [4] F. Mishima, S. Yamazaki, K. Yoshida, H. Nakane, S. Takeda, Y. Izumi, S. Nishijima, J. Magn. Soc. Jpn. 28 (2004) 393–396. [5] Y. Nakai, K. Senkawa, F. Mishima, Y. Akiyama, S. Nishijima, IEEE Trans. Appl. Supercond. 21 (2010) 2063–2067. [6] Y. Nakai, K. Senkawa, F. Mishima, Y. Akiyama, S. Nishijima, Physica C 471 (2011) 1533–1537. [7] Y. Mizuguchi, T. Miyamoto, Konica Minolta Technol. Rep. 1 (2004) 19–22. [8] M. Tanaka, M. Komaagata, M. Tsukada, H. Kamiya, Powder Technol. 183 (2008) 273–281. [9] K. Senkawa, Y. Nakai, F. Mishima, Y. Akiyama, S. Nishijima, Physica C 471 (2011) 1525–1529.