広島メタル＆マシナリー ケムテック事業部

Original Research Paper

Low-energy bead-milling dispersions of rod-type titania
nanoparticles and their optical properties

Takashi Tahara^a,b, Yuji Imajou^a, Asep Bayu Dani Nandiyanto^b,c, Takashi Ogi^b,*, Toru Iwaki^b, Kikuo Okuyama^b

^aKotobuki Industries Co., Ltd., 1-2-43 Hiroshiratake, Kure, Hiroshima 737-0144, Japan
^bDepartment of Chemical Engineering, Graduate School of Engineering, Hiroshima University, 1-4-1 Kagamiyama, Higashi Hiroshima 739-8527, Japan
^cDepartemen Kimia, Fakultas Pendidikan Matematika dan Ilmu Pengetahuan Alam, Universitas Pendidikan Indonesia, Jl. Dr. Setiabudhi No. 229, Bandung 40154, Indonesia

ARTICLE INFO

Article history:
Received 25 February 2014
Received in revised form 10 April 2014
Accepted 16 April 2014
Available online xxxx

Keywords:
Titania nanoparticle
Dispersion
Agglomeration
Bead milling
Refractive index

ABSTRACT

　The low-energy dispersion of nanomaterials in the bead-milling process is examined. The effect of milling parameters including bead size, rotation speed, and milling time on the dispersibility of fragile rod-type titanium dioxide nanoparticles is investigated. From experimental data obtained for the morphological, optical, and crystalline properties of dispersed nanoparticles, an unbroken primary particle dispersion in colloidal suspension was obtained only by conducting the bead-milling process using the optimum milling parameters. Deviation from the optimum conditions (i.e., higher rotation speed and larger bead size) causes re-agglomeration phenomena, produces smaller and ellipsoidal particles, and worsens crystallinity and physicochemical properties, especially the refractive index, of the dispersed nanoparticles. We also found that decreases in refractive index induced by the milling process are related to collisions forming broken particles and the amorphous phase on the surface of the particles. In addition, the present low-energy dispersion method is prospective for industrial applications, confirming almost no impurity (from breakage of the beads) was apparent in the final product.

©2014 The Society of Powder Technology Japan. Published by Elsevier B.V. and The Society of Powder Technology Japan. All rights reserved.

1. Introduction

　Recently, the dispersion of nanomaterials has attracted tremendous attention because only well-dispersed and undamaged single nanoparticles show great potential for use in electronic, chemical, mechanical, and biological applications [1, 2]. Well-dispersed nanoparticle suspensions are important because they are able to be reformed and reassembled into either larger particles [3-6] or films [5, 7, 8] with controllable structures. Various well-dispersed nanoparticles are now commercially available, such as metals, metal oxides, metal nitrides, metal carbides, and polymers [9].
　Many methods for dispersing nanomaterials have been utilized including ultrasonic-assisted dispersion, jet, ball, bead, and roll milling, and homogenization. However, only partial information about the effects of operating conditions, dispersion media, surface-modification agents, and type of collision and energy involved in current dispersion methods on the dispersion process is available.
　Previously, we studied the dispersion behavior of various nanomaterials such as titanium oxide, aluminum oxide, zinc oxide, boron nitride, titanium nitride, iron oxide, carbon black, and nickel metal in bead milling [1, 10-14]. Bead milling is effective for dispersing nanomaterials without chemical reaction or changing material properties. However, the influences of the processing parameters of bead milling on the optical properties and crystallinity of the dispersed nanomaterials have not yet been studied in detail.
　Here, we investigated the effects of bead-milling dispersion parameters including bead size, rotation speed, and milling time on the dispersibility of nanoparticles and examined the morphological, optical, and crystalline properties of the dispersed nanoparticles. Additionally, to minimize the impact energies during the dispersion process, we conducted bead milling using beads with diameters of several tens of micrometers, whereas current milling processes use beads that are hundreds of micrometers in diameter [15-22]. The use of smaller beads helps to prevent the fragmentation of nanoparticles and maintain their properties (e.g., crystallinity). Rod-type titanium dioxide (TiO²) was used as a model dispersed nanomaterial. TiO² was selected because it is widely used, nontoxic, and inexpensive, but most commercially available TiO² materials are in bulk or aggregated forms [3]. A rod-type material was chosen to examine the ability of our present beadmilling process to disperse a fragile material. To show the effectiveness of our present method for industrial applications, we also added an investigation of the purity of the beads-milling product, especially from the breakage of bead, while information about purity is typically disregarded in the current dispersion papers.

2. Dispersion behavior of agglomerated nanoparticles in the bead-milling process

　Fig. 1 shows illustration of the particle dispersion during the bead-milling process based on the type of energy used in the dispersion process. We used two types of dispersions: high energy and low energy. In this classification of energy, low energy defines the condition of bead-milling process that provides the break of agglomerated particles in the agglomerate position only.
　In the case of high-energy milling process (the first route), the properties of final dispersed materials sometimes changed. High-energy dispersion process allows the radical break up of particles. The break up position can be in both agglomerate and main body of crystal positions. As a result, dispersed slurry contains multisized particles. Further, this condition is incompatible for the case of fragile materials. In addition, milling time is also important. Too long milling process leads the re-agglomeration phenomena.
　To be effective for dispersing material with less properties damage, the second route can be an alternative. Since the nanoparticles typically softly agglomerate, optimization process that is able to break the agglomerated particles in the agglomerate position only is crucial. Indeed, when applying the low-energy dispersion process, the final slurry contains single sized of dispersed nanoparticles with properties that are similar to their original characteristic.
　Based on the current development, to get low-energy dispersion process in the realistic bead-milling process, several operating parameters should be considered, including milling time, temperature, bead and particle sizes, rotating speed, physicochemical properties and composition of dispersed media, and agglomerated particles. In our previous works[13, 14], the low energy in bead-milling dispersion can be obtained when conducting the milling process with a specific condition. This condition can be achieved when the process was conducted with rotational speed and bead size of 10 m/s and 30 lm, respectively. Deviation from this condition results re-agglomeration phenomena, produces smaller and broken particles, and worsens crystallinity and physicochemical properties of the dispersed nanoparticles. Therefore, in this study, influences of less milling time, lower rotational speed, and smaller bead size were investigated, whereas other operating parameters will be discussed in our future study.

Fig.1 Schematic illustration of the particle dispersion process considering the effect of dispersion energy

3. Experimental method

　Commercial rod-type TiO₂ nanoparticles (MT-01, rutile phase; Tayca Co. Ltd., Japan; surface modified with stearic acid and alumina) as a nanoparticle source were dispersed in toluene (Kanto Chemical Co. Ltd., Japan) using Crodafos (oleth-5-phosphate and dioleyl; Croda, Japan). The composition of raw materials was fixed at a mass ratio of toluene/TiO₂/Crodafos of 90/5/5. This suspension is referred to as the TiO₂ slurry.
　The TiO₂ slurry was then added to the bead-milling apparatus. A schematic diagram of the bead-milling apparatus is shown in Fig. 2 and detailed apparatus information is reported in our previous reports [1, 10-12]. In brief, the apparatus was a bead-milling vessel with a volume of 0.15 L equipped with a pump to supply the nanoparticle slurry, a mixer tank, and a centrifuge to separate the beads and TiO₂ slurry. The inner diameter and height of the bead-milling vessel are 50 and 150 mm, respectively. Regarding the configuration in the bead-milling vessel, we used a rotor with a diameter of 44 mm equipped with 11 of rotor pin. The bead filling ratio is 65% of the total vessel volume. Prior to adding the TiO₂ slurry, beads (zirconia beads; Nikkato Corp., Osaka, Japan) were added to the bead-milling system. In this study, milling time, bead size, and rotation speed were varied. During the bead-milling process, each suspension was sampled at specific times to analyze the effect of milling time on the properties of the TiO₂ product, as shown in Fig. 2.
　To measure the particle size distribution, dynamic light scattering analysis (DLS; FPAR-1000, Otsuka Denshi. Co., Japan) was conducted. Particle morphology was examined using a transmission electron microscope (TEM, JEM-3000F; JEOL Ltd., Japan). The crystallinity of the particles was determined by an X-ray diffraction (XRD; RINT 2550 VHF, Rigaku Denki, Japan; operated with Cu Kα radiation, with an angular domain between 3° and 120°). The optical properties of the slurry were investigated by a haze meter (NDH4000; Nippon Denshoku Co. Ltd., Japan) and spectrophotometer (U-2810; Hitachi, Japan). To analyze refractive index, each solution was spin coated on a glass substrate, dried at 60 ℃ for 6 h, and then the refractive index was measured using a prism coupler (Model 2010; Metricon, Japan). To analyze the concentration of slurry, especially for the investigating the existence of zirconia beads, an induced coupled plasma (ICP; Seiko SPS-4000, Seiko Instrument Inc., Japan) was used.
　In addition, regarding the calculation of the crystallinity, we calculated the crystallinity using a software that is equipped with the XRD measurement apparatus. In simplification of the calculation, the percentage of the crystallinity is defined as the ratio of “crystallinity area” and “total of crystalline and non-crystalline area”. Example of the calculation method for gaining the percentage of the crystallinity is shown in Fig. 3. The example of XRD spectra of anatase shown in this figure is adopted from Ref. [3].

Fig.2 A schematic illustration of the bead-milling apparatus.

Fig.3 Definition of crystallinity measurement in the XRD analysis. XRD pattern of anatase shown in this figure is adopted from Ref. [3].

4. Results

4.1. Particle size distribution

　Fig. 4 shows the DLS size distribution of TiO₂ dispersed with various milling times. The size distribution of the particles changes with processing time. The maximum peak of DLS size shifted to a smaller size as processing time extended.
　The median size of TiO₂ dispersed with various bead sizes and rotation speeds obtained from DLS measurements is presented in Fig. 5, where Fig. 5a and b are the results for rotation speeds of 10 and 8 m/s, respectively. For both rotation speeds, the size of the dispersed particles decreased as the dispersing time prolonged. The size decreased more rapidly at first (in the initial 30 min) and then changed gradually during the remainder of the milling process. In this study, we limited the milling time to 480 min because a longer milling time would not be cost effective for industrial applications.
　In addition to rotation speed, we also found that the characteristic of the size decrease depended on the bead size. For both rotation speeds (Fig. 5a and b), the characteristics of size decrease can be classified into two groups. One is when the process uses 50-lm beads (shown by a dashed line), and the other is when using beads of less than 30 lm in diameter (shown by a solid line). From these results, we can conclude that the use of larger beads promotes the formation of smaller dispersed particles. Additionally, for the samples dispersed with a rotation speed of 10 m/s and a bead size of 50 lm (see Fig. 5a), the particle size decreased until 330 min and then gradually increased with a longer milling time. This is possibly caused by re-agglomeration occurring during the milling process.

Fig.4 Particle size distribution obtained after different processing times. The beadmilling process was conducted using a bead size of 15 μm and rotation speed of 10 m/s.

Fig.5 Median size of particles obtained using different bead sizes, rotation speeds, and processing times. (a) and (b) are results for samples prepared using rotation speeds of 10 and 8 m/s, respectively.

4.2. Characterization of particles during the bead-milling process

　Because the milling process could lead to the formation of more thermodynamically stable materials, the crystal structure of the TiO₂ particles was analyzed (Fig. 6). XRD analysis showed that the dispersed particles were TiO₂ (not displayed), confirming the bead-milling process is effective to disperse particles without changing material phase and pattern. However, we found that the process parameters greatly affected the physicochemical properties (i.e., crystallinity) of the dispersed particles. In the case of crystallinity of sample conducted using a rotating speed and a bead size of 10 m/s and 50 μm, respectively, the crystallinity of TiO₂ decreased from 62% to about 48% as the dispersing time extended. We also found that the decrease of crystallinity depended on rotation speed and bead size. Lower crystallinity was obtained when the bead-milling process was conducted with higher rotation speed and larger bead size. In addition, the crystallinity of the titania since the initial condition (t=0) is relatively low. The appearance of low crystallinity is possibly because of the existence of amorphous phase on the surface of nanoparticles (since the measured crystallinity materials are in nanoparticle form). For this reason, further studies for clarifying the effect of bead-milling process of high crystalline material is required.
　TEM images of TiO₂ before and after dispersion process with various rotation speeds and bead sizes are depicted in Figs. 7a and b–g, respectively. Before bead milling, agglomerated TiO₂ nanoparticles were observed. The primary TiO₂ particle was a rod shape with transverse and longitudinal axes of about 8 and 50 nm, respectively. After bead milling of the TiO₂ particles, non-agglomerated particles were obtained (Figs. 7b–g). Different morphologies were observed depending on rotation speed and bead size. When using rotating speed (m/s)/bead size (μm) of 8/15 and 8/30, the rod shape and size of TiO₂ particles was conserved (Figs. 7b, c). However, under other conditions, TiO₂ nanoparticles of smaller size and ellipsoidal shape were obtained (Figs. 7d–g). 　In addition to morphological transformation, TEM can also detect the crystallinity of dispersed particles. Because amorphous phases are easily destroyed by an electron beam [23], the image quality of particle observation can be used to investigate the crystallinity of a particle. A clear image indicates that the particles are highly crystalline, whereas blurred images suggest that the sample contains the amorphous phase. Clear particle shape was observed in Figs. 7a–c. Conversely, in Figs. 7d–g, the surface of particles was blurred.
　The effect of milling time on the turbidity of the samples is illustrated in Fig. 8. Fig. 8a shows ultraviolet-visible (UV-vis) spectra of TiO₂ dispersions after various processing times. The intensity of the absorption spectra increased with processing time. The change in intensity corresponded with the change in color of the samples (shown in Fig. 8b). A longer milling time resulted in a more transparent solution. Light turbidity was observed because light transmittance through the nanoparticles was possible. Thus, the successful dispersion of particles can be observed directly by visible characterization. To confirm the dispersity of the particles, UV-Vis spectral analysis of TiO₂ dispersed with various rotation speeds and bead sizes was conducted (Fig. 9). Figs. 9a and b show the analysis results for samples prepared with rotation speeds of 10 and 8 m/s, respectively. The transmittance intensity increased with milling time. However, we observed a unique phenomenon when conducting the process using beads with a diameter of 50 μm. The transmittance increased and then decreased gradually after a specific milling time. For rotation speeds of 10 and 8 m/s, the intensity of spectra began to decrease after 180 and 420 min, respectively. The light transmittance of a solution depends on the size of the particles in it, so the results in Figs. 8 and 9 confirm that the increase in transmittance is caused by the decrease in the size of TiO₂ particles in the solution during the bead-milling process. In the case of samples prepared using 50-μm beads in Fig. 9, the observed decrease in transmittance confirms that re-agglomeration occurred after a certain time of bead milling.
Fig. 10 depicts the refractive indices of TiO₂ dispersed with various rotation speeds and bead sizes. In the case of samples prepared with a rotation speed of 10 m/s (Fig. 10a), a considerable decrease in refractive index from 1.73 to 1.64 was observed. In contrast, the refractive index of samples prepared with a rotation speed of 8 m/s (Fig. 10b) showed almost no change over time.
　The effects of rotation speed and bead size on the total transmittance (T.T.) and haze of samples are presented in Fig. 11. Similar trends in total transmittance and haze were observed with increasing milling time, except for the samples prepared with a bead size of 50 μm. With extending processing time, the total transmittance increased, whereas haze decreased. These results correlate well with the above analysis, where increasing milling time reduces the particle size in the solution.
　For the case of milling conducted with a bead size of 50 μm and rotation speed of 10 m/s, an unusual trend was found. As shown in Fig. 11a, total transmittance increased up to 200 min. Between 200 and 330 min, the transmittance results were relatively constant. Then, after 330 min, a decrease in transmittance was found. These results are in good agreement with those for the haze of this sample in Fig. 11c.

Fig.6 Crystallinity of TiO₂ dispersed with various milling times, rotation speeds, and bead sizes. Figures (a) and (b) are samples prepared using rotation speeds of 10 and 8 m/s, respectively.

Fig.7 TEM images of particles (a) before and (b–g) after dispersion. (b)–(g) show samples prepared by bead milling with a rotation speed (m/s)/bead size (lm) of 8/15, 8/30, 8/50, 10/15, 10/30, and 10/50, respectively.

Fig.8 Effect of milling time on light absorbance. Figure (a) shows the absorbance spectra of particles during the dispersion process, whereas Figure (b) is a photograph of samples after dispersion. Samples were prepared using a rotation speed of 10 m/s and bead size of 15 μm

Fig.9 Transmittance spectra of nanoparticles during the dispersion process. Figures (a) and (b) show samples prepared using rotation speeds of 10 and 8 m/s, respectively.

Fig.10 Refractive indices of samples prepared with different rotation speeds and bead sizes. Figs. (a) and (b) show the results obtained for samples prepared with rotation speeds of 10 and 8 m/s, respectively.

Fig.11 Total transmittance (T.T.) (a, b) and haze (c, d) of samples during the dispersion process.

4.3. Prospect of the low-energy dispersion process for industrial applications

　In the common bead-milling processes that use large beads, some of the surface of beads can be lost and becoming impurities (wear debris). For this reason, to be effective for industrial applications, the present process must be able to disperse nanoparticles without adding impurities from the wear debris or breakage of beads. Therefore, contamination of beads in the bead-milling product must be investigated.
　Fig. 13 shows the effect of milling time under various bead sizes and rotation speeds on the existence of zirconia concentration in the final product. Fig. 13a and b are results of bead-milling samples prepared with rotation speeds of 10 and 8 m/s, respectively. Almost no zirconia contamination in the final product was apparent, in which the maximum concentration of zirconia contamination is about 0.04 wt% in the dispersed slurry (after 8 h). The possibility for less wear debris is because of the use of micrometer beads, in which this cannot be achieved when using common bead milling processes that use larger beads. This result confirms that the present low-energy dispersion process is effective to disperse particles without adding impurities in the final product.

Fig.12 Proposal physicochemical transformation during the bead-milling dispersion process.

Fig.13 Effect of milling time under various bead sizes and rotation speeds on the concentration of zirconia contamination in the product. (a) and (b) are samples prepared with rotation speeds of 10 and 8 m/s, respectively.

5. Discussion

　The experimental result showed that the bead-milling process leads the agglomerated particles to disperse in a specific solvent. Based on the type of energy used in the dispersion process, there are two main dispersion mechanisms: high energy and low energy.
　A high-energy milling process tends to break the agglomerated particles both from the agglomerate and in the main body of the crystals. As a result, dispersed slurry contains particles of different sizes. Extended milling induces re-agglomeration, resulting in the formation of compact-type aggregated particles. In contrast, low-energy milling causes the agglomerated particles to break between particles only. Therefore, the final slurry contains dispersed nanoparticles of the same size. To determine the optimum conditions for dispersing a nanomaterial using a low-energy dispersion process, milling time, rotation speed, and bead size should be considered.
　Milling time is important because it relates to the amount of energy used to disperse the particles. However, a long milling time leads to re-agglomeration. This observation is in good agreement with our previous work [14]. In addition, for industrial applications, a short processing time is needed to maximize the efficiency of production.
　Rotation speed and bead size both affect how much energy is used and the milling time is required. Higher rotation speed and larger bead size provide greater energy to break up agglomerations. However, collisions between beads and TiO₂ particles become greater and result in a high-energy dispersion process as bead size and rotation speed increase. As a result, TiO₂ can break not only from the agglomeration but also the main body of the particle/crystal (as seen in Figs. 7d–g). In general, the use of a smaller bead size and slower rotation speed are better to maintain the quality of the dispersed nanoparticles because there is less effect on particle morphology (see Fig. 7b and c).
　To be effective for industrial applications, the performance of dispersed particles must be the same as their original properties. However, changes in the physicochemical properties (i.e., crystallinity (Fig. 6) and refractive index (Fig. 10)) of the dispersed material were found, confirming the need to optimize the milling process.
　Decreases in the crystallinity of the particles during the bead-milling process is the main reason for their changes in physicochemical properties (Fig. 12). During milling, collisions between beads and agglomerated TiO₂ particles provide energy that can break the main body of the crystal, resulting in smaller particles and amorphous phase on the crystal surface, as confirmed by the change in the crystallinity of the dispersed particles (Fig. 6) and the blurred appearance of the particles in Figs. 7d–g.
　To confirm the formation of amorphous phase during the bead-milling process, the refractive index, transmittance, and haze of the dispersed particles were evaluated (Figs. 10 and 11). The presence of amorphous phases in a particle lower the refractive index, which was observed in Fig. 10. 　In addition, because particle size is related to surface area, the number of dispersant molecules that are attached to the surface of TiO₂ should be also considered. Smaller particle sizes are able to adsorb comparatively more dispersant molecules on the surface of TiO₂ than larger ones. Therefore, particle size influences refractive index. As a result, the measured refractive index can differ from the estimated value.

6. Conclusions

　In this study, we investigated the influences of dispersion process parameters including rotation speed and bead size on the dispersion of fragile rod-type TiO₂ nanoparticles. To evaluate the performance of the dispersed TiO₂, its morphology, optical properties, and crystallinity were analyzed. Unbroken primary particle dispersions in colloidal suspensions were obtained using the optimum dispersion conditions. However, deviation from these optimum conditions (i.e., higher rotation speed and larger bead size) caused re-agglomeration, produced smaller, ellipsoidal particles, and lowered the crystallinity and physicochemical properties, especially the refractive index, of the dispersed particles. The refractive index decreased during the milling process because the large number of collisions between beads and nanoparticles broke particles and formed amorphous phase on the surface of particles. In addition, the present low-energy dispersion method is prospective for industrial applications, confirming almost no impurity (from the wear debris or breakage of beads) was apparent in the final product. We believe that the present study furthers understanding of dispersion science and technology, especially relating to the dispersion of fragile nanomaterials.

References

1. M. Inkyo, T. Tahara, T. Iwaki, F. Iskandar, C.J. Hogan Jr, K. Okuyama, Experimental investigation of nanoparticle dispersion by beads milling with centrifugal bead separation, Journal of Colloid and Interface Science 304 (2006) 535-540.
2. N. Mandzy, E. Grulke, T. Druffel, Breakage of TiO₂ agglomerates in electrostatically stabilized aqueous dispersions, Powder Technology 160 (2005) 121-126.
3. A.B.D. Nandiyanto, F. Iskandar, K. Okuyama, Macroporous anatase titania particle: Aerosol self-assembly fabrication with photocatalytic performance, Chemical Engineering Journal 152 (2009) 293-296.
4. A.B.D. Nandiyanto, K. Okuyama, Progress in developing spray-drying methods for the production of controlled morphology particles: From the nanometer to submicrometer size ranges, Advanced Powder Technology 22 (2011) 1-19.
5. A.B.D. Nandiyanto, A. Suhendi, O. Arutanti, T. Ogi, K. Okuyama, Influences of surface charge, size, and concentration of colloidal nanoparticles on fabrication of self-organized porous silica in film and particle forms, Langmuir 29 (2013) 6262-6270.
6. A. Suhendi, A.B.D. Nandiyanto, M.M. Munir, T. Ogi, L. Gradon, K. Okuyama, Self-assembly of colloidal nanoparticles inside charged droplets during spray-drying in the fabrication of nanostructured particles, Langmuir 29 (2013) 13152-13161.
7. A.B.D. Nandiyanto, N. Hagura, F. Iskandar, K. Okuyama, Design of a highly ordered and uniform porous structure with multisized pores in film and particle forms using a template-driven self-assembly technique, Acta Materialia 58 (2010) 282-289.
8. A.B.D. Nandiyanto, T. Ogi, F. Iskandar, K. Okuyama, Highly ordered porous monolayer generation by dual-speed spin-coating with colloidal templates, Chemical Engineering Journal 167 (2011) 409-415.
9. T. Ogi, A.B.D. Nandiyanto, K. Okuyama, Nanostructuring strategies in functional fine-particle synthesis towards resource and energy saving applications, Advanced Powder Technology 25 (2014) 3-17.
10. M. Inkyo, Y. Tokunaga, T. Tahara, T. Iwaki, F. Iskandar, C.J. Hogan, K. Okuyama, Beads mill-assisted synthesis of poly methyl methacrylate (PMMA)-TiO₂ nanoparticle composites, Industrial and Engineering Chemistry Research 47 (2008) 2597-2604.
11. I.M. Joni, A. Purwanto, F. Iskandar, K. Okuyama, Dispersion stability enhancement of titania nanoparticles in organic solvent using a bead mill process, Industrial and Engineering Chemistry Research 48 (2009) 6916-6922.
12. I. Joni, T. Nishiwaki, K. Okuyama, S. Isoi, R. Kuribayashi, Enhancement of the thermal stability and mechanical properties of a PMMA/aluminum trihydroxide composite synthesized via bead milling, Powder Technology 204 (2010) 145-153.
13. T. Tahara, M. Inkyo, Y. Imajyo, K. Okuyama, Low energy dispersion of titania nanoparticles by dual axes beads mill, Journal of Society Powder Technology Japan 48 (2011) 198-205.
14. T. Tahara, M. Inkyo, Y. Imajou, T. Ogi, K. Okuyama, Change in characteristics of titania nanoparticles during the process of dispersion, agglomeration and re-dispersion with a dual-axis beads-mill, Kagaku Kogaku Ronbunshu 39 (2013) 426-432.
15. J. Kano, F. Saito, Correlation of powder characteristics of talc during planetary ball milling with the impact energy of the balls simulated by the particle element method, Powder Technology 98 (1998) 166-170.
16. F. Ambrozio Filho, R.M.L. Neto, C.J. da Rocha, V.S. Gonçalves, L. Carvalho, An investigation of the use of stearic acid as a process control agent in high energy ball milling of Nb-Al and Ni-Al powder mixtures, Materials Science Forum 416 (2003) 144-149.
17. L. Damonte, L. Mendoza Zélis, B. Mari Soucase, M. Hernández Fenollosa, Nanoparticles of ZnO obtained by mechanical milling, Powder Technology 148 (2004) 15-19.
18. J. Qiu, H. Shen, M. Gu, Microwave absorption of nanosized barium ferrite particles prepared using high-energy ball milling, Powder Technology 154 (2005) 116-119.
19. T. Alex, R. Kumar, S. Roy, S. Mehrotra, Stirred bead mill grinding of gibbsite: surface and morphological changes, Advanced Powder Technology 19 (2008) 483-491.
20. L. Liu, A. Rashid, I. Marziano, E. White, T. Howes, J. Litster, Flowability of binary mixtures of commercial and reprocessed ibuprofen through high shear wet milling (HSWM) with lactose, Advanced Powder Technology 23 (2012) 454-458.
21. G. Khayati, K. Janghorban, The nanostructure evolution of Ag powder synthesized by high energy ball milling, Advanced Powder Technology 23 (2012) 393-397.
22. E. Zelaya, M. Esquivel, D. Schryvers, Evolution of the phase stability of Ni–Al under low energy ball milling, Advanced Powder Technology 24 (2013) 1063-1069.
23. A.B.D. Nandiyanto, T. Ogi, K. Okuyama, Doughnut magnesium fluoride nanoparticles prepared by an electron-beam irradiation method, Journal of Nanoparticle Research 14 (2012) 1182.