Manufacturing Conditions for Non-Melting Polycrystalline Translucency Ceramics Using Composite Quartz Materials

Abstract

Verification of the manufacturing conditions for translucency ceramics are showed using powder forming and sintering method in order to establish the manufacturing technology for optical devices with arbitrary complex shape and shape retention, and optical excitation materials. However, this method is extremely difficult to obtain translucency ceramics because the combination of complex factors such as “the shape and particle size of the raw materials, forming conditions and sintering” adversely affect the translucency of the sintered material, which is caused by the transmitted light scattering of the sintered material. Therefore, we have succeeded in realizing transparent ceramics by examining these complicated factors (materials, molding, sintering conditions, crystallization) and specifying the amorphous sintering conditions and the densification-molding conditions necessary to eliminate the light scattering factors inherent in sintered materials. It is necessary to show the correlation of the volume change with respect to T_g-temperature and T_m-temperature, which are important regarding amorphization, considering the sintering treatment conditions that depend on the particle size of the raw material fine powder, and to specify the molding body-pressurization (pressure and time) conditions. These optimizations will make it possible to develop the expected high-functional-photoelectric optical devices with microstructures and, high-power laser light sources independent of materials.

 

This Paper was Originally Published in Japanese in J. Jpn. Soc. Powder Powder Metallurgy 68 (2021) 409–414.

1. Introduction

In recent years, technological innovations in micro-optical waveguide applications have enabled the development of optoelectronic fusion and computational devices that utilize optical properties, the speeding up of optical communication devices,¹⁾ the application of high-power broadband light sources to non-destructive inspection technologies for public infrastructure such as bridges and concrete materials,^2,3) and less invasive diagnosis using optical interference technology developed through collaborations between medical researchers and engineers.⁴⁾ Optical properties have a diverse array of applications, and manufacturing techniques for next-generation optical devices are crucial for the development of these applications.

Photonic crystal structures, such as those utilized in advanced optical devices,^5–7) are based on the optical waveguide principle consisting of a second-order to three-dimensional core and cladding, and they are composed of a microstructure with an array of periodic air holes in the cladding region (corresponding to the wavelength of the guided light), where the transmission characteristics depend on the arrangement of the air holes. This enables arbitrary control of the transmitted light, such as the nonlinear optical properties, chromatic dispersion, confinement effects, and saturable absorption effects.^8,9) These devices are expected to facilitate the development of optical fiber lasers (optical short-pulse and ultra-broadband laser light sources) and contribute to further advances in the fields of medicine and engineering such as optical interference and minimally invasive optical diagnostics. Furthermore, optoelectronic fusion devices such as integrated optical microcircuits are expected to enable high-speed processing for optical arithmetic units.^1,6,7)

However, it is difficult to reproduce and retain fine shapes such as integrated optical circuits and photonic crystals in conventional manufacturing methods based on high-temperature melting for quartz materials.^10–14) Thus, it is necessary to develop a new manufacturing technique for non-melting polycrystalline ceramics. The establishment of such a manufacturing technique is expected to allow the fabrication of microstructures and optical devices of arbitrary shapes with even higher precision.¹⁵⁾

This technique, also called the powder sintering method, produces an inorganic-nonmetallic hybrid material from a highly refined synthetic raw material by a precisely controlled molding-manufacturing process. It is important to specify the forming and sintering treatment conditions strictly according to the particle size and shape of the powder raw materials.

Furthermore, the realization of transparent optical devices using manufacturing methods based on powder sintering naturally requires high translucency and a reduction of the light scattering and attenuation factors of the sintered body. However, overcoming the complex relationships between the numerous material properties and phenomena (gaps, air bubbles, contamination, crystallization, reduced permeability, surface diffusion, viscous flow, etc.) can be extremely difficult.^14,15)

For example, it is difficult to directly observe and analyze the grain boundaries, atomic diffusion, and crystallization factors of the (quartz) raw material at various sintering temperatures. Consequently, it is common for sintered bodies to be opaque or milky white. Therefore, in many reports, it was considered difficult to fabricate transparent optical devices using the powder sintering method.^14,16,17)

In an effort to overcome these problems, there have been reports in recent years of new technologies focusing on the dense forming and sintering of materials, such as hot isostatic pressing and spark plasma sintering (SPS), to realize transparent ceramics.^18,19) Further developments include a high-efficiency ceramic sintering method using improved SPS (flash sintering) and a direct conversion method capable of forming and sintering at ultra-high pressures (>10 GPa).^18,19) In particular, the possible use of oxide, carbide, and nitride nanoparticles rather than quartz materials to obtain nano-polycrystalline ceramics by a direct conversion method has been reported.¹⁸⁾

In recent years, amorphous-polycrystalline-transparent ceramic sintering under ultra-high pressure using nanoparticles (<100 nm) has become mainstream with the introduction of dedicated apparatus. Despite the high cost and need for specialized equipment, this approach is able to effectively produce cylindrical (diameter:height = 1:1) sintered bodies using graphite crucibles.

However, it would be desirable to develop simpler manufacturing methods for preparing non-melting polycrystalline transparent ceramic sintered bodies of arbitrarily complex shapes using general-purpose components and optimal starting materials and sintering conditions.

In this paper, we investigate the manufacturing conditions under which arbitrarily shaped transparent ceramics can be obtained using general molding and sintering equipment in combination with a quartz-based composite material to verify the molding and powder sintering conditions. This composite quartz material is characterized by its ability to be sintered in a typical sintering furnace (1100°C or lower). It is possible to shift the glass transition and softening temperatures to lower values by adding small amounts of various metals to the pure quartz material.

However, because it is difficult to determine the sintering temperature that suppresses crystal nucleation and growth for such composite quartz materials, we verified the temperature at which these changes occur.

The interactions between various aspects of the sintering process (raw materials, pressure forming, sintering, and cooling) are largely responsible for reducing the opaque/milky color of the sintered body, and the crystalline/amorphous properties and densification were investigated to successfully sinter polycrystalline transparent ceramics. The variation of the crystalline/amorphous nature (X-ray diffraction (XRD)), transmittance (Fourier-transform infrared (FT-IR) spectroscopy), and density (Archimedes method) with respect to the processing conditions used for molding and sintering was comprehensively evaluated.

2. Relationship between Transparent Ceramics and Manufacturing Requirements

In general, the fabrication of quartz-based composites using the powder sintering method requires the removal of impurities and air bubbles, as well as densification, during the forming and sintering processes according to the characteristics of the raw material. It is also important to maintain the sintering temperature below the melting point by optimizing the temperature control, to make the sintered body amorphous by suppressing the formation of crystal nuclei, and to maintain the shape.

Manufacturing conditions are identified with the primary objective of eliminating light scattering factors by optimizing the sintering process in consideration of the relationships between sintering, molding conditions, and material properties such as particle size and shape.^14–17)

However, the composite material used in this study has low glass transition and softening temperatures, and although it can be sintered in a conventional electric furnace, the sintering process is difficult to regulate because the crude temperature control induces crystallization inside the sintered body. Therefore, we identified the optimal temperature for obtaining an amorphous sintered body by XRD analysis.

Here, we investigated the effects of particle flowability and uniformity during compression molding (cold isostatic pressing (CIP)) on the molding, sintering, and cooling processes of the raw material, taking into account the effects of surface diffusion and intergranular diffusion.^15,18,19)

In addition, optional molding – in combination with primary molding (uniaxial pressurization) and secondary molding (CIP pressurization)^20,21) – was used, depending on the formwork.^20,21)

To minimize the influence of the thickness of the sintered body sample on the transparency evaluation, the sample shape was standardized (diameter ≅ 10 mm, thickness ≅ 5 mm) and the transmission characteristics were evaluated by FT-IR.¹²⁾

The relationship between the thermal energy provided during the sintering process and the quartz powder surface is known to cause surface diffusion and alter the atomic arrangement, and the complex atomic arrangement inside the quartz material is thought to exert a significant effect on the crystallization/amorphization of the sintered body.

It is known that during the heating and cooling processes used for the fabrication of transparent ceramics, crystallization/amorphization changes are dominant, depending on the regular/irregular structure of the molecular network inside the ceramics.

Therefore, as shown by the temperature–volume correlation diagram in Fig. 1, when the quartz material is subjected to heat treatment (a → b → c → d) followed by rapid cooling (d → c → b → a) from the molten state above the melting point (T_m), the sintered body enters a solid state (vitrification) while maintaining an irregular atomic arrangement below the glass transition temperature (T_g) due to the rapid decrease in volume and molecular motion.^16,17)

Fig. 1

Schematic of the relationship between sintered samples volume with temperature. (T_g: Glass-transition temperature and T_m: Glass-melting temperature)

However, in the case of slow cooling (Fig. 1, d → c → c1 → a1), the long holding time between T_m and T_g promotes crystal nucleation and growth. This non-melting powder sintering method is relatively conducive to nucleus growth, such that the cooling conditions are important for obtaining transparent ceramics.

The sintering conditions used in this paper (maximum temperature = T_x, holding time = from t1 to t2, cooling time = from t2 to t3, t4, or t5) are defined as shown in Fig. 2. On the basis of Fig. 1 and Fig. 2, the amorphization conditions were investigated, using XRD analysis to evaluate the crystalline/amorphous nature and the effects of the sintering conditions.^16,17)

Fig. 2

Schematic of sintering conditions. (T_x: maximum temperature, t1–t2: steady times, t2–t3∼t5: cooling times)

Next, based on the amorphous sintering conditions and results reported in the literature,^22,23) we examined the effects of the grain size distribution and pressure molding conditions on the density and bubble inclusion of the sintered bodies. The results revealed the effect of the particle size ratio of the raw powder (coarse vs. fine particles) on the densification of the sintered body. The grain size of the raw material (0.6 µm, density: >95%) contributed to the densification of the sintered body, and a smaller grain size, shorter sintering time, lower sintering temperature, and narrower grain size distribution led to a more stable sintered body shape.²²⁾

To our knowledge, there have been few reports describing the effect of the grain size on the transmittance and density of transparent ceramics. Therefore, in this paper, the effects of transmittance, density, and air bubbles in the sintered body on the transmittance of quartz materials with two different grain size distributions (D₅₀ and D₉₀) were examined in relation to the pressure applied to the compact (CIP: 70 MPa and CIP time (0.5 to 3.0 h)). High-vacuum sintering, which is highly effective for removing air bubbles from between particles during the sintering process, was also used.¹⁴⁾

3. Sintering Experiments

3.1 Sintering conditions and their effect on crystallization

To obtain amorphous transparent ceramics using the powder sintering method, it is important to suppress crystal nucleation and growth between T_m and T_g in Fig. 1 by carefully controlling the temperature. However, it is generally difficult to observe the sintering process inside the furnace, making it difficult to monitor nucleus growth.¹⁷⁾

Therefore, we evaluated various combinations of sintering conditions (Fig. 2, T_x, t1–t2, and t2–t3 to t5). First, powder sintering was performed using a sample with a rough particle shape, as shown by the SEM image and particle size distribution (D₉₀ ≅ 188 µm (uniform)) in Fig. 3(a) and (b), respectively. In this case, the sample was subjected to CIP molding (70 MPa) and sintering with a temperature increase rate of 5°C/min, T_x ≅ 700–950°C, holding time (t1–t2) ≅ 1–5 h, and cooling rate (from t2 to t3/t4/t5) ≅ 50–180°C/min. A granulation process to promote the filling effect during the molding process was not used in this work. Sintered bodies prepared in low vacuum were also used to remove air bubbles.

Fig. 3

Rough crush powder material SEM-image and particle distribution. ((a): SEM image and (b): particle size D₉₀ ≅ 200 [um])

Figure 4(a)–(c) show the XRD results for various sintering temperatures, holding times, and cooling rates. Figure 4(d) shows the translucent sintered sample obtained after quenching outside of the furnace (T_x ≅ 700°C, t1–t2 ≅ 1 h). The cooling rates in Fig. 2 and Fig. 4(c) were as follows: approximately 180°C/min (outside furnace quenching) for cooling type 1, approximately 130°C/min (outside furnace cooling) for cooling type 2, and approximately 50°C/min (inside furnace cooling) for cooling type 3.

Fig. 4

XRD results for sintering condition. ((a): Changing the T_x-conditions, (b): Changing the t1–t2, (c): Changing the cooling-conditions, and (d): Sintered-sample of cooling-type 1)

Figure 4(a) shows the XRD patterns for five sintered body samples processed at various T_x values (700–950°C at intervals of approximately 50°C) with a holding time of 1 h and a cooling rate of 180°C/min (i.e., the sintered bodies removed from the sintering furnace were rapidly cooled with a strong flow of cold air). The results revealed a tendency toward a smaller crystalline peak (i.e., amorphization) at relatively low sintering temperatures (approximately 700°C).

Figure 4(b) shows the XRD patterns for four sintered body samples processed at T_x ≅ 700°C for various holding times (1–5 h at 1-h intervals). A shorter holding time was found to be associated with a smaller crystalline peak, with no crystalline peak observed for the 1 h sample. Finally, Fig. 4(c) shows the XRD patterns for three sintered body samples processed at T_x ≅ 700°C for 1 h at various cooling rates (50–180°C/min), where rapid quenching outside of the furnace at 180°C/min was found to suppress the crystalline peak.

The above results indicated that the optimal sintering conditions for obtaining amorphous transparent ceramics included a relatively low temperature of T_x ≅ 700°C, a short holding time of 1 h, and rapid quenching outside of the furnace (180°C/min). However, the sintered body sample obtained under the sintering conditions shown in Fig. 4(d) exhibited an FT-IR transmittance of approximately 20% at 1500 nm and a density of 88%. This was ascribed to the voids between the angular particles and the grain boundaries present on their surfaces, which strongly affected the three-dimensional structure obtained after sintering, as well as the limited interstitial-bubble removal capability when sintering under low vacuum.

Based on these results, the particle shape and size in the initial raw material strongly affected the obtained transparent ceramics. Therefore, it is necessary for the material to have properties of small size, sharp particle distribution, and non-angular morphology, and furthermore a high-vacuum sintering system is effective in removing internal air bubbles.

3.2 Effect of grain size and shape

Next, we examined the effect of the particle size distribution (grain size) of the raw material on the transparent ceramics. Sintering was performed under the optimal conditions described above for obtaining amorphous samples (T_x ≅ 700°C, holding time ≅ 1 h, and cooling rate ≅ 180°C/min). To suppress the influence of powder porosity due to the irregular shape of the raw material, CIP (70 MPa for 0.5 h) and high-vacuum sintering (10⁻² Pa) were performed.

In the CIP molding process, the particle shape and particle size distribution of the initial raw material are important factors for obtaining a high-density molded product. Fine particles and a uniform particle size distribution are preferred.^22,23) To determine the ideal particle properties, the particle size was first reduced by milling and the particle size distribution was evaluated by dynamic light scattering.

The obtained particle size distributions are shown in Fig. 5(α) (D₅₀ ≅ 356.4 nm, D₉₀ ≅ 5559.1 nm) and Fig. 5(β) (D₅₀ ≅ 1978.3 nm, D₉₀ ≅ 2761.1 nm). Figure 6(α) and (β) show SEM images of each raw material. Figure 5(α) and (β) show that there was highly uniform grain size of about 1 µm, while Fig. 6(α) and (β) clearly differ from Fig. 3(a) in terms of grain shape and size, demonstrating a remarkable refinement.

Fig. 5

Particle size distribution using dynamic scattering method. ((α) Particle size (D₅₀ ≅ 356 nm, D₉₀ ≅ 5559 nm), (β) Particle size (D₅₀ ≅ 1978 nm, D₉₀ ≅ 2761 nm))

Fig. 6

SEM image of crushed powder using bead mill. ((α) Particle size (D₅₀ ≅ 356 nm, D₉₀ ≅ 5559 nm), (β) Particle size (D₅₀ ≅ 1978 nm, D₉₀ ≅ 2761 nm))

However, the shape of the fine particles in these samples was difficult to discern, suggesting that agglomeration occurred. The two samples shown in Fig. 6 were obtained by drying a colloidal liquid prepared by bead milling, which facilitates the formation of small particles. Figure 7(a) and (b) present photographs of the obtained sintered samples, and Fig. 8(a) and (b) show SEM images of the internal structures and the locations of foreign particles (circles).

Fig. 7

Appearance of sintering-ceramics. ((a): Sample used the (α) (D₅₀ ≅ 356 nm, D₉₀ ≅ 5559 nm), (b): Sample used the (β) (D₅₀ ≅ 1978 nm, D₉₀ ≅ 2761 nm))

Fig. 8

SEM image internal matter rate. ((a): 0.03[%] and (b): 0.21[%])

The sintered specimens shown in Fig. 7(a) and (b) were prepared from the raw materials shown in Fig. 6(α) and (β), respectively, and their transparency was visually confirmed. The foreign particles in each SEM image in Fig. 8(a) and (b) were equally visible throughout the sample, and their percentage was calculated using image analysis software (ImageJ). The amounts of foreign particles indicated by the circles were determined to be approximately 0.03% of the total area for the image shown in Fig. 8(a) and approximately 0.21% of the total area for the image shown in Fig. 8(b).

The transmittance of the sintered sample shown in Fig. 7(a) was approximately 40% at 1500 nm. This improved transmittance can be ascribed to the reduction of foreign particles including bubbles after performing vacuum sintering using a raw material with a grain size of 1 µm or less.

The shape and particle size distribution of the raw material suggest that the uniform fine powder of about 0.6 µm contributed to the high density of the compact and the elimination of air bubbles inside the sintered compact.

3.3 Effects of CIP conditions on transparency of sintered materials

In the experiments described in the preceding subsections, we used constant CIP parameters (70 MPa for 0.5 h) followed by sintering under various conditions to obtain transparent ceramics and examine the influence of temperature, holding time, cooling rate, and grain size and shape. The maximum transmittance obtained was approximately 40%. This was the result of temperature control to suppress crystal nucleus growth with respect to sintering temperature (T_x) and time from T_m to T_g, which are strongly correlated with the state of quartz raw materials. The particle shape and size distribution of the raw materials, which affect the high density of the sintered sample, were also evaluated.

To further improve the translucency, we next investigated the effect of the CIP time on the density and transparency of the sintered samples. Here, CIP was performed at a constant pressure of 70 MPa for between 0.5 and 3.0 h, and the sintering process was operated under the aforementioned conditions (i.e., T_x ≅ 700°C, holding time ≅ 1 h, cooling rate ≅ 180°C/min, and particle size D₅₀ ≅ 356 nm, D₉₀ ≅ 5559 nm).

The transparency was evaluated from the FT-IR spectra. Figure 9 shows the transmittance at CIP molding times ranging from 0.5 to 3.0 h for five sintered samples, and Fig. 10 shows the corresponding FT-IR spectra. Table 1 summarizes the composite quartz material particle size, CIP time, transmittance (at 1500 and 1700 nm, considering potential applications in optical devices operating in the IR band), and density for the five sintered samples. Sintered samples (a) and (B)–(D) were prepared using the raw material shown in Fig. 5(α) (D₅₀ ≅ 356.4 nm and D₉₀ ≅ 5559.1 nm), while sample (b) was prepared using the raw material shown in Fig. 5(β) (D₅₀ ≅ 1978.3 nm and D₉₀ ≅ 2761.1 nm) for the purposes of comparison. Figure 11 shows photographs of the sintered samples (a) and (B)–(D).

Fig. 9

Relationship between CIP-pressurization times [hours] and transparency [%] as transmitted wavelength.

Fig. 10

Relationship between wavelength [nm] and transparency [%] as CIP-pressurization times.

Table 1 Relationship between particle-size, CIP-times, and transmittance, density as characteristics of the sintered four-samples.

Fig. 11

Appearance of sintering-ceramics as (a), (B) to (D) by changing the CIP-pressurization times.

From Fig. 9, it can be seen that the transmittance of the sintered samples was strongly dependent on the CIP time. Comparison of the samples prepared with CIP times of 0.5 and 3.0 h reveals that increasing the CIP time led to a maximum improvement in the transmittance of approximately 10%. Figure 9 also confirms that the grain size of the raw material affected the transparency of the sintered sample, with the transmittance of sample (a) prepared using the smaller grain size being approximately 10% higher than that of sample (b) prepared using the larger grain size.

This was also confirmed by Fig. 10, where the transmittance characteristics of the sintered samples (a) and (B)–(D) prepared from the material shown in Fig. 5(α) exhibited a gradual increase toward longer wavelengths, while sample (b) prepared from the material shown in Fig. 5(β) displayed lower transmittance across the entire IR region.

The conditions necessary for improving the transparency after sintering included a particle diameter of 1 µm or less, while materials composed of larger and more angular particles (Fig. 3(a) and Fig. 5(β)) displayed poor flowability and a lower filling ratio during the CIP process. Moreover, the transmittance decreased owing to the isotropic structural change of the particles during the sintering process and a tendency for lower translucency at shorter wavelengths was confirmed.

These findings are verified by the photographs shown in Fig. 11(a)–(D), which reveal a trend from milky white to transparent depending on the CIP time. Performing CIP at 70 MPa for 3 h afforded the best results. This can also be seen from the densities of the four sintered samples (a) and (B)–(D) listed in Table 1, which ranged from 98.9% to 99.3%.

4. Conclusions

In this work, we have investigated the preparation of amorphous transparent ceramics under various molding and sintering conditions, with potential applications in fabricating samples with the complex shapes required for high-performance next-generation optical devices. We found that sintering the quartz material at a relatively low temperature (T_x) for a relatively short holding time between T_m and T_g followed by rapid cooling was crucial for preventing low transmittance due to crystallization (devitrification) during sintering, which is a major issue.

The flexibility of the atomic arrangement of the atoms in a sintered amorphous ceramic body is determined by the time taken for the cooling process from the molten state to the glassy state. Reduction of the particle size of the raw material to less than 1 µm and CIP molding at 70 MPa for 3 h were important for obtaining amorphous transparent ceramics with improved transmittance and high density, which enabled isotropic and low-temperature sintering with a high filling ratio during molding.

The above conditions allowed the difficulties associated with crystallization, contamination with foreign matter, and bubble formation during powder sintering to be avoided. The most important of these are the suppression of crystal nucleus growth during cooling by controlling the sintering temperature (T_x) and time (from T_m to T_g), which depend on the quartz composite raw material, and the optimization of the shape and particle (grain) size distribution of the raw material, as well as the suppression of foreign matter contamination.

REFERENCES

• 1)   K.  Kitayama,  M.  Notomi,  M.  Naruse,  K.  Inoue,  S.  Kawakami and  A.  Uchida: APL Photon. 4 (2019) 090901. doi:10.1063/1.5108912
• 2)   S.  Kogure and  K.  Sakamoto: Artif. Intell. Data Sci. 1 (2020) 48–56. doi:10.11532/jsceiii.1.J1_48
• 3)   K.  Muramatsu,  H.  Yagi and  T.  Yanagitani: Optronics (2012) 1–4.
• 4)   M.  Haruna and  M.  Ohmi: Rev. Laser Eng. 31 (2003) 654–662. doi:10.2184/lsj.31.654
• 5)   M.  Koshiba and  K.  Saitoh: Rev. Laser Eng. 34 (2005) 31–36. doi:10.2184/lsj.34.31
• 6)   M.  Nakazawa: Rev. Laser Eng. 30 (2002) 426–434. doi:10.2184/lsj.30.426
• 7)   M.  Imada and  S.  Noda: J. Surf. Sci. Soc. Jpn. 22 (2001) 715–722. doi:10.1380/jsssj.22.715
• 8)   K.  Miyagi,  Y.  Namihira,  S.M.A.  Razzak,  S.F.  Kaijage and  F.  Begum: Opt. Rev. 17 (2010) 388–392. doi:10.1007/s10043-010-0072-x
• 9)   K.  Miyagi,  Y.  Namihira,  Y.  Kasamatsu and  A.  Hossain: Opt. Rev. 20 (2013) 327–331. doi:10.1007/s10043-013-0059-5
• 10)   K.  Morinaga and  S.  Fujino: Rev. Laser Eng. 30 (2002) 297–302. doi:10.2184/lsj.30.297
• 11)   S.  Fujino,  H.  Ijiri,  F.  Shimizu and  K.  Morinaga: J. JIMM 62 (1998) 106–110. doi:10.2320/jinstmet1952.62.1_106
• 12)   S.  Yoshida,  S.  Fujino and  K.  Morinaga: J. MMIJ 120 (2004) 451–454. doi:10.2473/shigentosozai.120.451
• 13)   Y.-T.  O,  S.  Fujino and  K.  Morinaga: J. MMIJ 115 (1999) 471–474. doi:10.2473/shigentosozai.115.471
• 14)   K.  Ueda,  J.  Lu,  K.  Takaichi,  H.  Yagi,  T.  Yanagitani and  A.  Kaminskii: Rev. Laser Eng. 31 (2002) 465–470. doi:10.2184/lsj.31.465
• 15)   M.  Takahashi: Micromeritics 53 (2010) 3–14. doi:10.24611/micromeritics.2010003
• 16)   R.  Brückner: J. Non-Cryst. Solids 5 (1970) 123–175. doi:10.1016/0022-3093(70)90190-0
• 17)   S.  Sakka: J. Soc. Mater. Sci., Japan 34 (1985) 763–771. doi:10.2472/jsms.34.763
• 18)   T.  Irifune: Rev. High Pressure Sci. Technol. 28 (2018) 162–169. doi:10.4131/jshpreview.28.162
• 19)   T.  Yamamoto,  T.  Kobayasi and  H.  Yoshida: J. Jpn. Soc. Powder Powder Metallurgy 65 (2018) 646–653. doi:10.2497/jjspm.65.646
• 20)   T.  Maejima: Rev. High Pressure Sci. Technol. 28 (2018) 28–30. doi:10.4131/jshpreview.28.28
• 21)   T.  Yajima,  J.  Yamamoto,  F.  Ishii,  T.  Hirooka,  M.  Yoshida and  M.  Nakazawa: Opt. Express 21 (2013) 30500–30506. doi:10.1364/OE.21.030500
• 22)   S.  Taruta,  K.  Okada and  N.  Otsuka: J. Ceram. Soc. Jpn. 96 (1988) 146–151. doi:10.2109/jcersj.96.146
• 23)   S.  Taruta,  N.  Takasugawa,  K.  Okada and  N.  Otsuka: J. Ceram. Soc. Jpn. 104 (1996) 447–450. doi:10.2109/jcersj.104.447