Sintering and Densification of Transparent Ceramics

Abstract

There are various sintering techniques that can be used to fabricate transparent ceramics. Conventional sintering techniques include vacuum sintering, hot pressing (HP), and hot isostatic pressing (HIP), while spark plasma sintering (SPS) is more popular than microwave sintering in the new sintering technique category. Every method has its own advantages and disadvantages. Different methods can be combined to offer higher sintering efficiency. The selection of sintering technique is also dependent on materials.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

7.1 Introduction

There are various sintering techniques that can be used to fabricate transparent ceramics. Conventional sintering techniques include vacuum sintering, hot pressing (HP), and hot isostatic pressing (HIP), while spark plasma sintering (SPS) is more popular than microwave sintering in the new sintering technique category. Every method has its own advantages and disadvantages. Different methods can be combined to offer higher sintering efficiency. The selection of sintering technique is also dependent on materials.

7.2 Vacuum Sintering

Vacuum sintering refers to a sintering process technique carried out with a vacuum equipment at a high vacuum level to achieve desired densification. Vacuum sintering has been widely used to fabricate various transparent ceramics, including garnet [1–7], spinel [8–11], alumina [12, 13], and rare earth sesquioxides ceramics, such as Y₂O₃ [14–18], Sc₂O₃ [19], pyrochlore oxides [20, 21], and so on.

When using vacuum sintering to fabricate YAG ceramics, the procedure usually consists of vacuum sintering followed by air annealing [22–38]. For example, transparent polycrystalline Nd:YAG ceramics can be synthesized by using solid-state reactive sintering the mixture of Al₂O₃, Y₂O_3, and Nd₂O₃ in a tungsten mesh-heated vacuum furnace (M60, Centorr Vacuum) at 5 × 10⁻⁶ Torr during holding [35]. Fully dense transparent samples with average grain sizes of ~50 μm were obtained at 1800 °C. The sintering temperature was slightly affected by the concentration of Nd. A small amount of SiO₂ doping lowered the sintering temperature by ~100 °C. SiO₂ is usually introduced by using tetraethyl orthosilicate (TEOS). After sintering, the specimens were annealed at 1400 °C for 2 h in air for decoloring.

The first YAG transparent ceramics for laser applications was prepared by using vacuum sintering [5]. High-purity fine Al₂O₃ and Y₂O₃ powders presynthesized by using wet chemical methods, with the composition of YAG, were mixed with ethyl silicate as a solution of 0.5 wt% in ethyl alcohol, by using ball milling. After milling, the dried powder mixture was compacted by using isostatic pressing at 140 MPa. The pressed pellets were sintered at 1600‒1850 °C at 1.3 × 10⁻³ Pa. Optical properties, microhardness, and thermal conductivity of the YAG ceramics sintered at 1800 °C were comparable with those of YAG single crystals.

Highly transparent polycrystalline Er³⁺:Y₃Al₅O₁₂ (Er:YAG) ceramics with Er³⁺ ion concentrations of 1–90 % were prepared by using solid-state reaction combined with vacuum sintering technique [22]. Commercial powders of α-Al₂O₃, Y₂O_3, and Er₂O₃ with purity of >99.99 % were thoroughly mixed by using ball milling, according to Er:YAG compositions with different contents of Er, together with a trace of TEOS as sintering aid. The powder mixtures were dry-pressed at 3T to form pellets with a diameter of 25 mm, which were isostatically cold-pressed at 250 MPa. The pellets were vacuum sintered at temperatures of up to 1800 °C for different time durations at 10⁻³ Pa, after which they were annealed at 1450 °C for 20 h in air.

Figure 7.1 shows photographs of the mirror-polished Er:YAG transparent ceramics with different Er contents of from 1 to 90 %. With increasing content of Er, the specimens appeared pinker and pinker, due to the light absorption of Er³⁺ at the visible band. All the Er:YAG ceramics exhibited excellent optical properties.

Fig. 7.1

Photographs of the mirror-polished Er:YAG ceramics with different contents of Er (from left to right): a 1, 5, 10, and 15 at.% and b 30, 50, 70, and 90 at.%. Reproduced with permission from [22]. Copyright © 2011, Elsevier

EPMA images of the ceramics with 1 and 90 % Er are shown in Fig. 7.2. Both samples were very compact with very few pores. No secondary phases were observed both at the grain boundaries and in the inner grains. Average grain size of the 1 % Er:YAG ceramics was about 30 μm. The 90 % Er:YAG sample had a larger average grain size, probably due to the large particle size of starting Er₂O₃ powder. The high-resolution TEM (HRTEM) image indicated that the grain boundaries of the ceramics were clean and narrow, with a width of about 1 nm, thus leading to much negligible optical scattering.

Fig. 7.2

Electron probe micro-analyzer (EPMA) images of the Er:YAG ceramics: a 1 % and b 90 %. Reproduced with permission from [22]. Copyright © 2011, Elsevier

It was found that there was a strong dependence of grain sizes on the location of the samples with green bodies formed by using uniaxial pressing with die, which actually indicated the presence of distributions in density of the green bodies. As mentioned before, cold isostatic pressing can only increase the density of the green body by <10 % and the density distribution of the green bodies would not be affected significantly. For two-directional pressing, the middle part of the green body possessed the lowest density, thus leading to largest grains after sintering. The grain size increased from the center part to the ends of the cylinder sample. For one-directional pressing, the density of the green body decreased from the top end to the bottom end. Figure 7.3 shows the scenario of one-directional pressing [22]. This observation is an agreement with previous discussion on ceramic powder packing.

Fig. 7.3

Demonstration of the relationship between density and grain size distribution in a representative sample with green body formed by using one-directional axial press. Reproduced with permission from [22]. Copyright © 2011, Elsevier

Microstructural evolution during vacuum sintering of Nd:YAG transparent ceramics has been systematically studied [1]. Through microstructural observations, the microstructural maps and grain size density and grain size–pore size sintering trajectories of Nd:YAG ceramics, as a function of silica content, were established. For densities higher than 99.7 %, the occurrence of intragranular porosity was correlated to a critical pore radius of 0.16 μm. The presence of silica favored the formation of intragranular pores, which was attributed to the increase in the grain growth rate, as compared with the densification. An analytical model was established by coupling the analytical laws derived from sintering trajectories and the classical theory of light diffusion, which could be used to correlate the microstructural features of the transparent Nd:YAG ceramics to their optical properties.

Submicrometer α-Al₂O₃ (purity >99.99 %), Y₂O₃ (purity >99.99 %), and Nd₂O₃ (purity >99.99 %) powders were mixed to form the Y_2.94Nd_0.06Al₅O₁₂ (2 at.% Nd:YAG), by using ball milling in water with an organic dispersant and SiO₂ (purity >98 %) as sintering aid with contents of up to 0.3 wt%. Reactive sintering was conducted in a tungsten mesh-heated furnace at vacuum of ≤10⁻² Pa and temperatures of 1000–1800 °C for different times durations, at heating/cooling rate of 5 °C min⁻¹.

Figure 7.4 shows SEM images of the undoped and 0.1 wt% SiO₂-doped Nd:YAG ceramics sintered at vacuum for 5 h at temperatures in the range of 1580–1730 °C [1]. Three types of microstructures could be observed, depending on the experimental parameters and the doping. Type one was porous microstructure, with the presence of open pores, having relative density (ρ) ranging from 60 to 90 % and submicrometer grain sizes of about 500 nm, as shown in Fig. 7.4a. Type two was dense microstructure, containing some closed intergranular pores, with relative density ranging from 90 to 99 % and micrometer grain sizes of about 1 μm, as illustrated in Fig. 7.4b. Type three was fully dense microstructure, with an average grain size of >1 μm, as seen in Fig. 7.4c–f. In the third-type microstructure, intragranular pores could occasionally be observed, as demonstrated in Fig. 7.4d. It was shown that significant grain growth occurred at the final stage of the densification, i.e., with ρ > 85 %. The comparisons, between Fig. 7.4c–e, and d–f, indicated that the doping of 0.1 wt% of silica had no significant effect on overall microstructural profiles of the Nd:YAG ceramics, but could decrease the sintering temperature by about 50 °C correspondingly.

Fig. 7.4

SEM images of the undoped Nd:YAG ceramics sintered at vacuum for 5 h at temperatures of 1580 °C (a), 1630 °C (b), 1680 °C (c), and 1730 °C (d), as well as the 0.1 wt% SiO₂-doped Nd:YAG ceramics sintered for 5 h at 1630 °C (e) and 1680 °C (f). Reproduced with permission from [1]. Copyright © 2013, John Wiley & Sons

Sintering trajectories, i.e., grain size (G) versus relative density (ρ), of the undoped and silica-doped samples sintered for 2 h at 1450–1800 °C, are shown in Fig. 7.5 [11]. It was observed that grain size increased at a very slow rate until the relative density approached 99 %. However, after that, there was a sharp increase in grain size. The presence of silica could promote grain growth and densification.

Fig. 7.5

G–ρ sintering trajectories of the undoped and silica-doped Nd:YAG ceramics sintered at vacuum for 2 h at temperatures ranging from 1450 to 1800 °C. Reproduced with permission from [1]. Copyright © 2013, John Wiley & Sons

Transmittance spectra of the samples with a thickness of 2.5 mm, labeled with their corresponding pore volume densities, are shown in Fig. 7.6a [11]. Obviously, the baseline of transmittance increased with decreasing volume density of the pores, demonstrating the optical scattering effect of pores in the Nd:YAG ceramics. Photographs of the corresponding Nd:YAG ceramic samples are shown in Fig. 7.6b. Analytical model indicated that the attenuation coefficient could be linked to the volume density of pores. It was concluded that once having similar values in pore volume density, e.g., <3 × 10³ cm³, Nd:YAG ceramics could have similar optical properties of their single-crystal counterparts. This study provided a useful reference to other transparent ceramic materials.

Fig. 7.6

a Optical transmittance spectra of the 0.3 wt% SiO₂-doped Nd:YAG transparent ceramics with different pore volume densities (C _n) after sintering at vacuum (samples thickness was 2.5 mm). b Photographs of the Nd:YAG ceramics corresponding to the samples in (a). Reproduced with permission from [1]. Copyright © 2013, John Wiley & Sons

Transparent alumina ceramics with high transmission have been developed by using a rapid vacuum sintering processing at a lower sintering temperature and shorter time duration, as compared with the conventional pressureless sintering [13]. The alumina sintered at 1670 °C for 5 min exhibited a residual porosity of as low as 0.002 %, thus showing a real in-line transmission of 64 % at of 1100 nm. The sintering time had an effect on the densification and transparency of the alumina ceramics sintered at 1650 °C.

A high-purity commercial α-Al₂O₃ (99.99 %) was mixed with 0.1 wt% MgO (99.99 %, 30 nm) as sintering aid by using ball milling with high-purity alumina balls. The milled and dried mixture was made into pellets with diameter of 15 mm and thickness of 2 mm at a uniaxial pressure of 200 MPa. The green pellets were presintered at 1200 °C for 2 h in muffle furnace and finally sintered in vacuum at 1500–1700 °C for 0–30 min, at a heating rate of 100 °C min⁻¹, which was followed by a natural cooling. All sintered samples were mirror-polished on both surfaces and thermally etched at 1400 °C for 1 h in air.

SEM images of the ceramics sintered at different temperatures are shown in Fig. 7.7 [13]. Porosities of the samples sintered at 1570, 1600, 1650, and 1670 °C for 5 min were 1.256, 0.705, 0.075, and 0.002 %, with corresponding average grain sizes of 3.54, 5.47, 8.55, and 11.14 μm, respectively. Both the grain size and density of the samples were increased with increasing sintering temperature. A total of 1600 °C was the critical temperature for complete densification, above which all the samples demonstrated a homogeneous microstructure with clear grain boundaries and without abnormal grain growth.

Fig. 7.7

SEM images of the alumina ceramics sintered for 5 min at different temperatures: a 1570 °C, b 1600 °C, c 1650 °C, and d 1670 °C. Reproduced with permission from [13]. Copyright © 2012, John Wiley & Sons

Figure 7.8 shows SEM images of the samples sintered at 1650 °C for different time durations [13]. All samples exhibited very high density, homogeneous microstructure, and clear grain boundary, although residual pores at triple junction points of grain boundary were occasionally observed. It was found that prolonged sintering time resulted in an increase in pore size.

Fig. 7.8

SEM images of the alumina ceramics sintered at 1650 °C for different time durations: a 0 min, b 10 min, c 20 min, and d 30 min. Reproduced with permission from [13]. Copyright © 2012, John Wiley & Sons

During the rapid vacuum sintering, the densification process of the transparent alumina can be attributed to the presence of a temperature gradient in the sample and the formation of nonequilibrium grain boundaries [39, 40]. The nonequilibrium grain boundaries had higher energy, thus providing a stronger driving force for grain growth. At the same time, due to the large thickness of the nonequilibrium grain boundaries, the grain-boundary mobility was decreased. Therefore, densification was maintained, while grain growth was suppressed.

The mechanism for the improvement in transparency of vacuum sintered Y₂O₃ ceramics due to the addition of ZrO₂ has been analyzed [41]. Y₂O₃ (99.999 %) and ZrO₂ (99.99 %,) with an average particle size of 1 μm were mixed, according to the formula of (Y_1−x Zr _x )₂O₃ (x = 0–0.10), by using a ball milled in absolute ethyl alcohol with agate balls. After milling and drying, pellets with 20 mm in diameter were formed with stainless steel at 15 MPa and then isostatically pressed at 250 MPa. Sintering was conducted at 1800 °C for 15 h in vacuum.

Combined with sintering kinetics, thermodynamics, and experimental results, the mechanism of ZrO₂ in improving the transparency of Y₂O₃ ceramic was clarified. The addition of ZrO₂ led to the formation of ZrO₂–Y₂O₃ binary system with decreased melting point, thus reducing the sintering temperature. The radius of Zr⁴⁺ is 0.80 Å, which is close to that of Y³⁺ (0.90 Å), so that there was only very small lattice distortion when Y³⁺ was substituted with Zr⁴⁺.

At the final stage of sintering, grains with large size and low surface energy tend to grow at the expense of the surrounding smaller grains, which is known as discontinuous or exaggerated grain growth. During this process, due to the high mobility of grain boundaries, pores are easily trapped in grains, which is harmful to the transparency. Due to its high melting point (2715 °C), ZrO₂ is prone to form solid solution at the grain boundaries, so as to reduce the grain-boundary mobility. At the same time, secondary recrystallization could occur, due to the solute drag mechanism, which would be beneficial to the elimination of pores and the homogenization of grain sizes, two critical factors to ensure high optical transmittance of the Y₂O₃ ceramics.

It is well known that the diffusion coefficient of O²⁻ is much higher than that of Y³⁺ in pure Y₂O₃. Therefore, cation diffusion is the rate-controlling step of grain-boundary migration, i.e., the diffusion of $[{\text{Y}}_{\text{i}}^{ \cdots } ]$ dominates the grain-boundary mobility [42]. There are following defects and defect reactions in Y₂O₃.

For Schottky defects, there is:

$${\text{Y}}_{2} {\text{O}}_{3} \Leftrightarrow 2{\text{V}}_{\text{Y}}^{\prime \prime \prime } + 3{\text{V}}_{\text{O}}^{ \cdot \cdot } .$$

(7.1)

For Frenkel defects, there are:

$${\text{O}}_{\text{O}} \Leftrightarrow {\text{O}}_{\text{i}}^{{\prime \prime }} + {\text{V}}_{\text{O}}^{ \cdot \cdot } ,$$

(7.2)

$${\text{YY}} \Leftrightarrow {\text{V}}_{\text{Y}}^{{{\prime \prime \prime }}} + {\text{Y}}_{\text{i}}^{ \cdots } .$$

(7.3)

After doping with ZrO₂, there is:

$$2{\text{ZrO}}_{2} \mathop{\longrightarrow}\limits{{{\text{Y}}_{2} {\text{O}}_{3} }}2{\text{Zr}}_{Y}^{ \cdot } + {\text{O}}_{\text{i}}^{{\prime \prime }} + 3{\text{O}}_{\text{O}} .$$

(7.4)

As a result, when every two Zr⁴⁺ ions were substituted, one ${\text{O}}_{\text{i}}^{''}$ was produced, so that $[{\text{V}}_{\text{Y}}^{'''} ]$ was increased, while $[{\text{Y}}_{\text{i}}^{ \cdots } ]$ was decreased. The decrease in $[{\text{Y}}_{\text{i}}^{ \cdots } ]$ resulted in a reduction in grain-boundary mobility and sintering rate, which ensured the escape of pores and densification of the ceramics. For example, average grain size of the ZrO₂-doped ceramics was about 15 μm, which was much smaller than those widely reported in the open literature. In terms of optical performance of the transparent ceramics, 3 at.% ZrO₂ was the optimized doping concentration.

7.3 Hot Pressure (HP) Sintering

Hot pressure (HP) sintering is a high-pressure low-strain-rate powder metallurgy process, which is used to form a powder or powder compact at a temperature that is sufficiently high to induce sintering and creep processes. This is achieved by the simultaneous application of heat and pressure. Densification at high pressure proceeds through the rearrangement of particles and plastic flow at the particle contacts. HP is mainly used to fabricate hard and brittle materials. As a result, HP has become an important technique to obtain transparent ceramics.

In 1977, transparent spinel with small grain size (1–2 μm) was prepared by using low-temperature (1400 °C) and high-pressure (70 MPa) hot pressing [43]. Transparent magnesia–alumina spinel ceramics were prepared from equimolar mixture of the oxides, which were derived from Mg(OH)₂ and Al(OH)₃ after calcining at 900 °C for 2 h and 1400 °C for 2 h, respectively. Formation of spinel from the oxide mixture was accompanied with a volume expansion of about 7.9 %, which was suppressed by applying a pressure of >330 kg cm⁻² at 950 °C, so as to promote the densification of the spinel ceramics. A step heating schedule was used, i.e., hot-pressing technique at 770 kg cm⁻² at 1300 and 1350 °C for 0.5 h and 1400 °C for 1 h. The transparent spinel ceramics exhibited optical transmittance of 35–75 % in the visible range and 75‒85 % in the IR range.

Since then, HP became a widely used technique to prepare transparent spinel ceramics. For example, HP process was used to fabricate MgAl₂O₄ transparent ceramics, with a nearly 100 % densification [44]. Spinel powders for the production of transparent polycrystalline ceramic windows were produced by using both the traditional ceramic solid-state reaction and solgel methods. Among them, the powders produced from the reaction of organomagnesium compounds with surface modified boehmite precursors were best to produce high-quality transparent spinel ceramics. The powder synthesis method allowed fine control over the particle size, size distribution, purity, and stoichiometry. The process involved the formation of a boehmite solgel from the hydrolysis of aluminum alkoxides. The boehmite nanoparticles were treated through a surface modification with carboxylic acids. The surface modified boehmite nanoparticles were then used to make a precursor powder of pure phase spinel through metal exchange at room temperature with magnesium acetylacetonate.

Different from the conventional sintering process which usually requires high sintering temperatures (>1600 °C), HP process is able to achieve desired densification at relatively low temperature, which is therefore also called low-temperature high-pressure (LTHP) process. Due to its low-temperature requirement, it can be used to develop nanosized ceramics [45–48]. The sintering mechanism at high pressures is entirely different from that at ambient pressure [45]. High pressure can suppress the grain growth and initiate plastic deformation to eliminate pores and/or additional phases existing in triple junctions of the grains. However, the conventional sintering process is controlled by grain growth to avoid imperfections between grain boundaries when preparing transparent ceramics.

High-resolution transmission electron microscopy (HRTEM) and selected area electron diffraction (SAED) of two selected samples, sintered at 4.0 GPa at 600 °C and 4.0 GPa at 1100 °C, are shown in Fig. 7.9 [45]. The average grain size observed from Fig. 7.9a was about 40 nm. The nanostructured profile of the sample was confirmed by the corresponding SAED pattern. The HRTEM image showed a clear grain boundary without any secondary phase and/or in-between phases, as demonstrated in Fig. 7.9b. The clean grain boundaries had no contribution to scattering, which was responsible for the high transparency of the MgAl₂O₄ ceramics sintered at 4.0 GPa at 600 °C.

Fig. 7.9

a TEM image showing microstructure of the MgAl₂O₄ nanocrystalline transparent ceramics sintered at 4.0 GPa and 600 °C for 30 min. The corresponding SAED pattern in the inset proved the characteristics of the polycrystalline nanograins. b HRTEM image of the MgAl₂O₄ nanoceramics. c TEM image showing microstructure of the MgAl₂O₄ ceramics sintered at 4.0 GPa at 1100 °C for 30 min, with an average grain size of the smaller crystallites to be ∼500 nm. d HRTEM image of the MgAl₂O₄ ceramics sintered at 4.0 GPa/1100 °C taken along the [332] direction, with the corresponding SAED pattern shown as the inset. Reproduced with permission from [45]. Copyright © 2010, Elsevier

As the sintering temperature was increased to 1100 °C, the grain/crystallite size increased significantly, so that the nanostructured profile was lost due to the grain growth, as illustrated in Fig. 7.9c. It is also found that the crystals have a wider particle size distribution, probably due to the presence of a temperature gradient in the sample. Consequently, imperfections, nonuniform crystallites and irregular grain boundaries, were present, which would pose negative effects on optical transparency of the ceramics. The HRTEM image of MgAl₂O₄ ceramics sintered at 4.0 GPa at 1100 °C taken along the [332] zone axis, shown in Fig. 7.9d, together with the corresponding SAED pattern, indicated a highly crystallized feature of the sample, demonstrating the grain growth of the materials at 1100 °C.

Therefore, the formation temperature of the transparent MgAl₂O₄ spinel ceramics could be significantly reduced, if sufficiently high pressure is available. As the pressure is in the range of 2–5 GPa, very low sintering temperatures of 500–700 °C can lead to nanosized transparent spinel ceramics [46]. The nanosized ceramics were highly transparent even though their relative densities are less than 99 %, due to the low or negligible light scattering from the nanosized grains and pores. The LTHP process is able to densify nanosized powders without significant grain growth, which becomes a key advantage for the development of transparent ceramics, as discussed above. It is of special interest when mechanical strength is an important requirement.

Another example is the development of a nanocrystalline spinel ceramics with 50 % increase in hardness by using hot pressing [49]. It is well accepted that mechanical strength and hardness of a material can be increased by decreasing its grain size, as described by the empirical Hall–Petch relationship [50, 51]. Based on this principle, an integrated approach was developed by using nanosized spinel powder, combined with high-pressure and low-temperature sintering to fabricate fully dense and high-purity nanocrystalline ceramics with nanometer-sized grains. The Hall–Petch relationship was confirmed by the hardness of the transparent spinel ceramics with grain sizes down 28 nm. As a result, the nanosized spinel ceramics exhibited a 50 % increase in hardness without a decline in fracture resistance. More importantly, the nanocrystalline spinel ceramics had near theoretical optical transparency.

Figure 7.10 shows a schematic diagram of the high-pressure high-temperature sample holder [49]. Spinel powders were pressed into green compacts without binder within glove box in an evacuable die at 1900 kg cm⁻². The ejected green compacts were mechanically sealed in a metal capsule. The sealed sample capsule was inserted into the high-pressure assembly, which consisted of a partially sintered ZrO₂ container saturated with CsCl in the shape of a rectangular cuboid with truncated edges, a graphite heater, mica, and molybdenum heater contacts. K-type thermocouples were inserted through both high-pressure cell assembly lids. A pressed boron nitride tablet isolated the exposed thermocouple junction from the metal sample capsule. High-pressure experiments were performed in a pressless split-sphere apparatus (BARS) equipped with an 8–6-type multi-anvil system. After about 25 min pressurization to 2 GPa, samples were sintered at temperatures from 740 to 845 °C for 15 min. The nanosized spinel ceramics are transparent in the visible spectrum, without major absorption bands from ultraviolet to short-infrared wavelengths. The maximum transmission was about 81 % at infrared wavelengths of 1200–1350 nm.

Fig. 7.10

Schematic of the high-pressure cell assembly. Reproduced with permission from [49]. Copyright © 2014, Elsevier

The fact that the ceramics were fully dense and of high quality, without residual porosity or obvious grain growth, was further confirmed by the SEM and TEM images, as shown in Figs. 7.11 and 7.12 [49]. SEM images of the polished surface of the ceramics derived from the 200 nm powder and the same 30.8 nm nanocrystalline ceramics are shown in Fig. 7.11a, b, respectively. Any potential variability in the sintering environment has been eliminated, because the samples were sintered in the same sample capsule at 2 GPa and 795 °C. Fine pores were present in the submicron sample, while very few features, which included a surface scratch and dust particles, were present at similar magnification for the nanocrystalline ceramics. At higher magnifications, a submicron grain structure was observed in the coarse ceramics and abnormally large grains were present along with the porous structure. The grain structures of the nanocrystalline ceramics are clearly demonstrated by the TEM images, as shown in Figs. 7.11d and 7.12.

Fig. 7.11

Microstructures of the porous and fully dense spinel ceramics produced by sintering at 2 GPa and 795 °C. a and c SEM micrographs of the porous submicron structure of spinel ceramic derived from 200-nm-sized spinel powder. b SEM and d TEM images of the microstructure of the fully dense nanosized spinel ceramics. Reproduced with permission from [49]. Copyright © 2014, Elsevier

Fig. 7.12

TEM images of the fully dense nanosized spinel ceramics produced by sintering at 2 GPa and 795 °C. a and b Different areas revealing that some grain boundaries contain ~1-nm amorphous phase regions. c An amorphous triple junction with the inset showing the fast Fourier transform of the main figure. d Image from a region containing nanotwins. Reproduced with permission from [49]. Copyright © 2014, Elsevier

Figure 7.12 shows TEM images of different grain-boundary regions of the nanosized spinel ceramics [49]. There were nanoscale grains where the crystalline lattice extended into the boundary, as shown in Fig. 7.12a, d. Also, grain boundaries decorated with a small, ~1 nm, amorphous phase region were observed, as shown in Fig. 7.12b. Such amorphous phases were also present at the junctions of three grains, as shown in Fig. 7.11c. Noticing the highly transparent characteristics of the nanosized sample, the slight fraction of amorphous phase at the grain boundaries had negligible influence on its optical properties and density.

Similar process was also used to develop yttrium aluminum garnet (Y₃Al₅O₁₂, YAG) transparent ceramics with nanosized grains [52]. In this study, a high-pressure cell was used, where disk samples were placed in a capsule made of NaCl, both to ensure the quasi-hydrostatic compression of the sample and prevent its contact with the graphite heater. The whole structure was then contained in the pyrophyllite high-pressure cell. The sample temperature in the high-pressure cell was measured directly by using a standard Ni–Cr/Ni–Si thermocouple. The cell pressure was monitored by using a calibrated oil pressure meter. Desired pressures were first applied to the samples before raising the temperature. The pressure range was 2.0‒5.0 GPa, and temperature range was 300‒500 °C, with holding time of 30 min. After that, the pressure was released first, and then, the samples were cooled down to room temperature at a rate of 15 °C min⁻¹. By doing in this way, the residual stress of the samples could be released in order to avoid the happening of cracking. Optimized processing parameters for transparent YAG ceramics were 5 GPa and 450 °C.

One of the characteristics of HP is freedom to optimize the sintering parameters, which has been demonstrated when HP sintering was used to fabricate transparent Y₂O₃ ceramics [53, 54]. For instance, stepwise process has been reported to fabricate Eu:Y₂O₃ transparent ceramics [54]. The program included a high-temperature ramp at constant pressure and a pressure ramp at constant high temperature. After a dwell at 1150 °C to outgas the samples, a pressure of 10 MPa was applied, followed by a temperature ramp to 1580 °C at a heating rate of 8 °C min⁻¹. After a 30 min dwell at constant temperature and pressure, the pressure was raised to 40 MPa at a rate of 0.4 MPa min⁻¹. With this stepwise application of pressure, optimized temperature was 1580 °C.

HP sintering process is a very unique and feasible technique to prepare transparent electro-optic ceramics, including PLZT [55–57], PZN–PLZT [58] and PMN–PT [56, 59–61], and other ferroelectric ceramics [62], which contain volatile element Pb. In the early 1970s, (Pb, La)(Zr, Ti)O₃ (PLZT) transparent ferroelectric ceramics were synthesized by using HP [43]. The transparent PLZT ferroelectric ceramic materials have found a variety of electrooptic applications. They were prepared from mixed oxides by using hot-pressing sintering at 1100 °C for 16 h at 2000 psi. Transmission measurements in the visible and infrared showed that these materials exhibited a nearly constant response from the absorption edge of 0.37 μm to ∼6 μm. The highest transmission values were observed for compositions containing >8 at.% La. Specific compositions exhibited electro-optic memory, conventional linear, and quadratic electrooptic effects, with performances comparable with those of single crystals.

Highly transparent ceramics of La-doped 0.75Pb(Mg_1/3Nb_2/3)O₃–0.25PbTiO₃ (PMN–PT) were fabricated by using a two-stage sintering method [59]. Green pellets were first sintered in an oxygen atmosphere (OA) at different temperatures, which were then hot pressed (HP) at temperatures of >1000 °C for more than 8 h at pressures of 50‒100 MPa. The ceramics prepared in this way exhibited a transparency of as high as 65 % at infrared wavelength. Large quadratic electro-optic coefficient of 66 × 10⁻¹⁶ (m/v)² was obtained, which was the highest value reported in the literature for the ceramics with similar compositions.

The precursor powder with a composition of La-doped PMN–PT 3/75/25 was synthesized by using the columbite precursor method, starting with PbO, MgO, Nb₂O₅, TiO₂, and La₂O₃, in which small amount of excessive PbO was used to promote the densification through liquid-phase sintering. The powder was made into pellets by using uniaxial pressing. The pellets were first sintered in an OA at different temperatures. The sintered pellets were then hot pressed (HP) at a temperature of >1000 °C for more than 8 h at pressures of 50–100 MPa. In the OA–HP two-stage sintering method, the OA sintering temperatures were different: (A) 1150 °C, OA–HP; (B) 1200 °C, OA–HP; and (C) 1230 °C, OA–HP.

The PMN‒PT ceramics exhibited a transparency of as high as 65 % in infrared region, as shown in Fig. 7.13a. SEM images of PMN‒PT 3/75/25 ceramics sintered with different schedules are shown in Fig. 7.13b, c. All samples showed transgranular fracture character, which was attributed to the increased homogeneity between grain boundary and grain, as well as firm conjunct strength of grains caused by the liquid phase that was formed during the sintering. It was found that the sample with more transgranular fracture had higher transparency, which suggested that grain boundary with an increased homogeneity and firm strength would have reduced reflection and scattering of light. This observation once again indicates that the quality of grain boundary plays a more important role in determining optical properties of transparent ceramics, as discussed previously.

Fig. 7.13

Transmission curves of the PMN‒PT 3/75/25 ceramics with a thickness of 0.5 mm (a) and SEM images of the samples sintered with different schedules: b sample A, c sample B, and d sample C. Reproduced with permission from [59]. Copyright © 2010, John Wiley & Sons

HP process was found to be the most feasible method to prepare transparent or translucent α-sialon ceramics [63, 64], which could not be achieved by using the conventional pressless sintering. Single-phase α-sialon ceramics with high optical transmittance have been prepared by using hot pressing [63]. HP was also used to tailor grain morphologies and preferential texturing microstructures of α-sialon ceramics [64–66]. Other examples include translucent MgO ceramics fabricated by using hot-pressing, with nanopowder of MgO containing 2–4 % LiF as a sintering aid [67]. HP process has even been used to prepare fluoride ceramics (CaF₂), such as transparent Yb:CaF₂ ceramics [68, 69].

7.4 Hot Isostatic Pressure (HIP) Sintering

Hot isostatic pressing (HIP) is a fabrication process that is used to reduce the porosity of metals and increase the density of various ceramic materials [70–74]. The HIP process subjects a component to both elevated temperature and isostatic gas pressure in a high-pressure containment vessel. The chamber is heated, causing the pressure inside the vessel to increase. Many systems use associated gas pumping to achieve the desired pressure levels. Pressure is applied to the material from all directions.

Materials are hot isostatically pressed (HIPed) in order to achieve the maximum possible densification, which has become a key technique to achieve high optical transmittance. To reduce the fabrication cost, hot isostatic pressing is usually used as the last step, although it has been proved to be a critical step to prepare high-quality transparent ceramics. This is simply because HIP involves sophisticates equipment and critical experimental conditions.

HIP process has been widely used to synthesize transparent armor ceramics, such as alumina (Al₂O₃) [75–79], YAG [80–83], and spinel (MgAl₂O₄) [84–88]. HIP process can be combined with HP. Hot pressing followed by hot isostatic pressing (HIP) proved to be more feasible to fabricate transparent MgAl₂O₄ ceramics was widely used [84, 85, 88]. Nowadays, HIP process is also used to synthesize cubic sesquioxide ceramics, including Y₂O₃ [89], Sc₂O₃ [90], and Lu₂O₃ [91]. In these cases, vacuum sintering to remove closed pores with a subsequent HIP step was usually employed, which provided an alternate processing route to the fabrication of fully dense sesquioxide ceramics with less possibility for contamination and reduction than hot pressing.

A simple method has been established to achieve nearly 100 % relative density of α‒alumina ceramics by using hot isostatic pressing (HIP) [75]. In this study, transparencies of the samples were correlated to their grain size and residual porosity. A commercially available α-alumina powder was wet milled in order to realize de-agglomeration. After that, the average particle size of the powder was 100–150 nm. Suspensions with ∼75 wt% solid with any organic additive were cast with a nylon filter (pore diameter = 0.2 μm) to form green bodies. The wet bodies were dried at 65 °C for 15 h and then calcined at 600 °C for 30 min in vacuum of <1 mbar. Then, the samples were sintered naturally in vacuum of ~10⁻⁶ mbar at temperatures of about 1200 °C for 2–17 h. The samples with porosities of <10 % were subject to HIP at about 1200 °C at 170 MPa in argon for 3–5 h. The HIP furnace was made of molybdenum. The densities of the HIP processed samples were measured with a very precise method, by measuring mass of the ceramics outside and inside water, which were then compared with the density of alumina single crystal to obtain relative densities.

Figure 7.14 shows porosity values of the samples after natural sintering in vacuum versus the treatment temperature and time [75]. With a constant treatment time, the porosity always increases as the sintering temperature is decreased. It was found in this work that if the porosity was higher than 8 %, the HIP treatment was not workable. The samples sintered at 1200 and 1220 °C; time duration of 2 h was sufficient. However, the samples sintered at 1170 and 1185 °C should be kept for 17 h to decrease the porosity to be <4 %.

Fig. 7.14

Total porosity versus the natural sintering temperature, for different treatment times, at the pressure of ∼10⁻⁶ mbar. Reproduced with permission from [75]. Copyright © 2011, Elsevier

In addition, the treatment temperature and time also showed their effects on grain size of the samples, as shown in Fig. 7.15 [75]. As stated earlier, the grain size of the starting powder was 100–150 nm. After the treatments at 1200 and 1220 °C for 2 h, the grain size was <400 nm, which was close to that of the sample sintered at 1170 °C for 17 h. However, after sintering at 1185 °C for 17 h, the grain size was >600 nm. The HIP treatment led to an increase in grain size. The increase in grain size after the HIP processing was simply based on the grain size of the samples after the natural sintering. For a given natural sintering temperature, the variation in the HIP condition had relatively less effect on the grain size as compared with that of the natural sintering.

Fig. 7.15

Grain sizes of the ceramics versus the natural sintering (NS) temperature: (square) after natural sintering, (triangle) after HIP at 1180 °C and 170 MPa for 3 h, (disc) after HIP at 1200 °C and 170 MPa for 3 h, and (star) after HIP at 1180 °C and 170 MPa for 5 h. The HIPed ceramic values are placed versus the temperature of the corresponding natural sintering. Reproduced with permission from [75]. Copyright © 2011, Elsevier

Figure 7.16 shows relative densities of the samples [75]. It was observed that the effect of the natural sintering temperature was also quite obvious. At the same HIP condition, the higher the natural sintering temperature, the weaker the effect of HIP on the relative density would be. The samples naturally sintered at 1170 and 1185 °C followed by HIP at 1180 °C for 3 h and the sample sintered at 1200 °C followed by HIP at 1180 °C for 5 h exhibited almost the same relative density of 0.9985, while the different in their grain sizes was quite large, i.e., between 580 and 1300 nm. Therefore, this achievement implies that high relative densities can be realized, while grain growth can be well controlled at the same time. It is believed to be also applicable to other transparent ceramic materials.

Fig. 7.16

Relative density (τ) after HIP treatment versus the natural sintering temperature. The HIP parameters are indicated for each data. Reproduced with permission from [75]. Copyright © 2011, Elsevier

Highly transparent 8 mol% Y₂O₃–ZrO₂ (8YSZ) ceramics were fabricated by an improved hot isostatic pressing method [92]. Combined with presintering treatment, microstructures, such as positional difference in residual pores, inter-, or intragranular pores, of the 8YSZ ceramics could be well controlled. Commercially available 8 mol% Y₂O₃–ZrO₂ powder was used as a starting material. Green bodies were formed by using cold isostatic pressing method. Pellets with diameter of 20 mm and thickness of 3 mm were made by die pressing at 50 MPa and then isostatically cold-pressed at 200 MPa. The green bodies were presintered in an alumina tube furnace in air. The heating rate was 100 °C h⁻¹, and the holding time was 2 h at temperatures of <1650 °C. The presintered samples were finally treated by using hot isostatic pressing, at temperatures of 1350–1750 °C for 1 h in argon gas at pressure of 150 MPa. The temperature and pressure were simultaneously raised at a heating rate of 500 °C h⁻¹.

According to the experimental results, a microstructure model for pore elimination during the hot isostatic pressing has been proposed to illustrate microstructural evolution of a green body, as shown in Fig. 7.17 [92]. At low presintering temperatures, small intergranular pores are surrounded with fine grains. The intergranular pores transform to intragranular ones when the presintering temperatures are sufficiently high to facilitate grain growth. It is suggested that such intergranular pores are easily removed during the hot isostatic pressing. On the one hand, fine grains have large plastic deformability, which is favorite for pores to shrink at the hot isostatic pressures [93]. On the other hand, it is well known that high-temperature plastic deformability is increased with decreasing grain size [94, 95]. Also, the rapid migration closed gas is through grain boundaries. Small grain size means more grain boundaries, thus being beneficial to the shrinkage of pores. The model has been found to be applicable to many other oxide ceramics [96, 97].

Fig. 7.17

Microstructure model for pore elimination during the sintering of hot isostatic pressing. Reproduced with permission from [92]. Copyright © 2008, John Wiley & Sons

The feasibility of HIP can be further demonstrated by the following examples. Fully dense transparent garnet ceramics were derived cold-pressed green bodies that were subsequently vacuum sintered, with residual porosity being removed by hot isostatic pressing [98]. It was reported that hot isostatic pressing (HIPing) could be used to enhance the electrical and optical properties of PNNZT (50.0/15.5/34.5) [99]. After hot isostatic pressing at 1100 °C, the density was increased to >99 %. Transparent mullite ceramics could be obtained by pressureless sintered combined with hot isostatically pressed (HIP) [100]. In the HIPed mullite, with porosity of <1 %, a transmittance of 40 % was observed in the VIS range and up to 80 % in NIR. Transparent Yb:CaF₂ ceramics could be fabricated by sintering and hot-pressing powders that were derived by a soft chemistry process [101]. There are also reports to use HIP process to consolidate hydroxyapatite (HAp) filter-cakes [102]. HAp ceramics were fully densified at 800 °C after 2 h when a hot isostatic pressure of 100 MPa was used. Besides, transparent α-sialon ceramics [103], transparent magnesium oxide (MgO) ceramics [104], and transparent La₂Hf₂O₇ (LHO) ceramics [105] were also fabricated by using HIP process.

7.5 SPS Processed Transparent Ceramics

Various transparent ceramics, such as alumina (Al₂O₃) [106–112], yttrium aluminum garnet (YAG) [113, 114], sesquioxides [115–120], ZrO₂ [121–123], MgO [124], MgAl₂O₄ [125–127], mullite [128, 129], hydroxyapatite (HAp) [130, 131], other complex oxide compunds [132], and nonoxides (e.g., AlN) [133], have been processed by using SPS. Also, SPS has been demonstrated to be an effect method to fabricate electro-optic ceramics, including lanthanum-doped lead zirconate stannate titanate (PLZSnT) antiferroelectric ceramics, PbZrO₃–PbTiO₃–Pb(Zn_1/3Nb_2/3)O₃ (PZ–PT–PZN) ferroelectric ceramics [134], and lanthanum-doped lead zirconate titanate (Pb_0.92La_0.08(Zr_0.65Ti_0.35)_0.98O₃, PLZT) ceramics [135]. Processing of these transparent ceramics by using the conventional sintering is a significant challenge, due to the content of volatile element (Pb). More recently, transparent BaTiO₃ and SrTiO₃ ceramics, as well as their solid solutions, Ba_1-xSr_xTiO₃ (or BST in short), which can never be obtained by using the conventional sintering techniques, have been achieved by using SPS [136, 137].

The effects of SPS processing parameters on microstructure and optical performance of Al₂O₃ ceramics have been studied [138]. Commercial α-Al₂O₃ powder, with a purity of 99.99 % and an average particle size of 0.15 μm, was used directly in this study. Compacted samples were heated to 1150 °C at a uniaxial pressure of 80 MPa by using a SPS machine (SPS-1050, Sumitomo) with a pulse duration of 3.4 ms. Heating was conducted in a sequent way, with twelve DC pulses (40.8 ms) followed by zero current for 6.8 ms. The heating rate from 600 to 1150 °C was varied between 2 and 100 °C min⁻¹. Temperature was measured with an optical pyrometer focused on a nonthrough hole (1 mm in diameter and 2 mm in depth) of the graphite die. After holding for 20 min at the sintering temperature and subsequent annealing at 1000 °C for 10 min, the sintered disk was about 30 mm in diameter and 3 mm in thickness. In addition, for the heating rates of 8 and 50 °C min⁻¹, sintering time was varied between 0 and 5 h, in order to examine the grain growth behavior. The mechanical pressure was unloaded before annealing.

It was found that, with the sintering parameters stated above, grain size of the final products increased with increasing heating rate. For example, the sample heated at 10 °C min⁻¹ had an average grain size of 0.29 μm, while the one heated at 100 °C min⁻¹ showed a grain size of 0.55 μm. SEM images of representative samples are shown in Fig. 7.18 [138]. This observation is in contrast to common trend in conventional furnace sintering. It was attributed to the high defect concentration produced by rapid heating and the associated rapid deformation during densification. The defects produced by both rapid heating and deformation would be responsible for the accelerated grain growth, which was further demonstrated in Fig. 7.19. As the temperature just reached 1150 °C, the two heating rates led to samples with almost the same grain size. However, as the sintering time was increased, the grain size of the sample sintered at 50 °C min⁻¹ became much larger than that of the one sintered at 8 °C min⁻¹. The effect of heating rate on porosity was normal, i.e., porosity increased with increasing heating rate.

Fig. 7.18

Microstructures of the alumina sintered at a heating rate of a 2 °C min⁻¹ and b 100 °C min⁻¹. Reproduced with permission from [138]. Copyright © 2009, Elsevier

Fig. 7.19

Grain growth behaviors at 1150 °C for two different heating rates. Reproduced with permission from [138]. Copyright © 2009, Elsevier

Figure 7.20 shows photographs of the Al₂O₃ ceramics sintered at 1150 °C for 20 m with different heating rates, while tier in-line transmission curves are shown in Fig. 7.21 [138]. It is clearly demonstrated that the heating rate should be as low as possible in order to achieve high optical transparency. However, in practice, a trade-off is necessary, because low heating rate means high energy consumption.

Fig. 7.20

Alumina ceramics sintered by SPS at a heating rate of a 2 °C min⁻¹, b 5 °C min⁻¹, c 10 °C min⁻¹, d 25 °C min⁻¹, e 50 °C min⁻¹, and f 100 °C min⁻¹. All samples are 0.9 mm thick and are put on a transparent polystyrene plate 24 mm above the text. Reproduced with permission from [138]. Copyright © 2009, Elsevier

Fig. 7.21

In-line transmission of the alumina sintered at 1150 °C for 20 min. The sample thickness is about 0.9 mm. Reproduced with permission from [138]. Copyright © 2009, Elsevier

SPS technique even can be used to fabricate samples with special shapes, such as hemispherical domes. For instance, transparent polycrystalline alumina domes have been obtained by combining sintering and forming into one step in minutes [139]. In contrast, fabrication of such items by using the convention sintering technology requires hours. Therefore, SPS provides an unprecedented opportunity to make optically transparent domes at much lower cost, in this respect. In addition, alumina transparent ceramics processed by using SPS have relatively finer grains, due to the rapid heating rate and short sintering time duration.

It is possible to further optimize the sintering effectiveness and efficiency when using SPS. For example, a two-step pressure profile was used to demonstrate the feasibility of SPS in processing MgAl₂O₄ [140]. At a low preload pressure, 5 MPa, with a normal fast heating rate of 100 °C min⁻¹, samples could have high in-line transmittance of 51 % at 550 nm and 85 % at 2000 nm. Sintering was carried out in vacuum (4–6 Pa) with a Dr. Sinter 2050 SPS apparatus (Sumitomo Coal Mining Company Ltd., Japan). The temperature was increased to 600 °C within 3 min and then further increased to 1250 °C in 6 min, while the final sintering temperature of 1300 °C was reached in 1 min. The sintering time at the final stage for all experiments was 3 min. Five preload pressures, i.e., P ₁ = 5, 10, 20, 30, 50, and 100 MPa, were used in the experiment. Photographs of the samples are shown in Fig. 7.22. Obviously, the samples prepared at P ₁ of 30, 50, and 100 MPa have a darker discoloration core of about 8 mm in size at their center [140]. The discoloration was attributed to the dislocations formed during the fast densification. Figure 7.23 shows in-line transmittance curves of the samples. It was concluded that a preload pressure P ₁ should not be >20 MPa, so as to obtain acceptable optical transmittance approaching the theoretical limit in the wavelength range of >1000 nm.

Fig. 7.22

Photographs of the transparent MgAl₂O₄ ceramics sintered at 1300 °C for 3 min at different preloaded pressures (P₁): 5 MPa (1.79 mm), 10 MPa (1.99 mm), 20 MPa (1.82 mm), 30 MPa (1.89 mm), 50 MPa (1.82 mm), and 100 MPa (1.60 mm). The samples were 15 mm above the paper, and thicknesses of the samples are indicated in parentheses. Reproduced with permission from [140]. Copyright © 2009, Elsevier

Fig. 7.23

In-line transmittance spectra of the polycrystalline MgAl₂O₄ sintered by SPS at different preload pressures P1. The theoretical transmittance of a single-crystal MgAl₂O₄ was calculated based on the refraction index as a function of wavelength. Reproduced with permission from [140]. Copyright © 2009, Elsevier

Figure 7.24 shows SEM image of the samples processed at 5 MPa and 100 MPa [140]. It was found that both samples exhibited uniform microstructures, which implied that the application of a prepressure had no significant effect on microstructure of the final product. No residual pores were observed on surface the thermally etched sample, while only a very few pores were found by using TEM observation. Average grain sizes of the 5 and 100 MPa samples were 600–700 nm. No significant difference in grain size was observed between the discoloration center and the clean rim. It was attributed to the fact that the temperature gradient was <30 °C in the 12 mm graphite die in the experiment, which was not sufficiently large to lead to energy difference, so that the grain size was not affected.

Fig. 7.24

The microstructure of transparent MgAl₂O₄ sintered by SPS at 1300 °C for 3 min at different pressures of P1: a and b 5 MPa and c and d 100 MPa. a and c SEM images taken near edge of the samples. b and d SEM images taken at the discolored center. Reproduced with permission from [140]. Copyright © 2009, Elsevier

Transparent Y₂O₃ ceramics have been fabricated by using SPS at moderate temperature and pressure [119]. It was found that sintering temperature had a significant effect on densification, microstructure, optical, and mechanical properties of the Y₂O₃ ceramics. Very high-purity Y₂O₃ powder (99.999 %) was used as the starting material. Before SPS processing, the powder was ball milled in ethanol with ZrO₂ balls for 12 h and then calcined at 1273 K in air for 7.2 ks. SPS-210 LX model SPS setup was used to densify the Y₂O₃ powder. The temperature was first increased to 873 K in 180 s and to 1373 K in 300 s and then held for 300 s. After that, the temperature was further increased to 1373–1823 K at a heating rate of 0.17 K s⁻¹ and maintained at each temperature for 2.7 ks. A pressure of 10 MPa was preloaded between room temperature and 1373 K, which was increased to 100 MPa at >1373 K for final sintering. Post-annealing was carried out at 1123–1423 K for 21.6 ks in air.

Relative density of the sample was 98 % after sintering at 1373 K, while fully dense ceramics with relative density of above 99 % were achieved after sintering at 1473–1823 K. The average grain size was slightly increased from 0.24 to 0.32 μm as the sintering temperature was increased from 1473 to 1573 K. After sintering at 1823 K, the grain size was increased to 1.97 μm [119].

Transmittance spectra of the Y₂O₃ ceramics SPS sintered at 1573 K before and after annealing at 1123–1423 K for 21.6 ks are shown in Fig. 7.25 [119]. As the annealing temperature was increased, transmittance of the samples was increased gradually. The optimized annealing temperature was 1323 K, as shown in Fig. 7.25b. The ultraviolet absorption edge was blue-shifted from 308 to 256 nm, as the annealing temperature was increased from 1123 to 1223 K. No significant change was observed after annealing at higher temperatures of up to 1423 K.

Fig. 7.25

a Transmittance spectra of the Y₂O₃ ceramics before and after annealing at 1123–1423 K for 21.6 ks using the samples sintered at 1573 K. b Transmittance at λ = 550 and 2000 nm as a function of annealing temperature. Reproduced with permission from [119]. Copyright © 2012, Elsevier

Figure 7.26 shows transmittance spectra of Y₂O₃ bodies sintered at 1473–1773 K after annealing at 1323 K for 21.6 ks in air, together with those that were calculated with the refractive index of Y₂O₃ single crystal [119]. Transmittance of the Y₂O₃ ceramics was increased greatly after annealing. The sample sintered at 1573 K exhibited highest transmittance of 55.0 and 81.7 % at λ = 550 and 2000 nm. The one sintered at 1573 K after annealing reached nearly 99 % of the theoretical value in the infrared region, as shown in Fig. 7.26a. This temperature was much lower than those required by other sintering techniques.

Fig. 7.26

a Transmittance spectra of the Y₂O₃ ceramics sintered at 1473–1773 K after annealing at 1323 K. b Transmittance at λ = 550 and 2000 nm as a function of sintering temperature. The dashed line indicates transmittance calculated from the refractive index of Y₂O₃ single crystal. Reproduced with permission from [119]. Copyright © 2012, Elsevier

Vickers hardness (H _V) and fracture toughness of the annealed Y₂O₃ ceramics as a function of sintering temperature are shown in Fig. 7.27a, while relationship between H _V and grain size is illustrated in Fig. 7.27b [119]. H _V of the Y₂O₃ ceramics sintered at 1373 K was 8.3 GPa. It was sharply increased to ~17.2 GPa after sintering at 1573 K. A maximum value of ~17.5 GPa was achieved at the sintering temperature of 1623 K. After that, the value of H _V steadily decreased to ~14 GPa after sintering at 1923 K. Fracture toughness was in the range of 1.0–1.5 MPa m^1/2, which was not significantly influenced by sintering temperature, especially over 1573–1823 K. H _V of the Y₂O₃ ceramics showed a linear relationship with d ^−1/2, i.e., the Hall–Petch relation. The fact that H _V increases with d ^−1/2 in ceramics has been explained in terms of the reduced free path for dislocations as the grain size is decreased [141].

Fig. 7.27

a Vickers hardness and fracture toughness of the Y₂O₃ ceramics annealed at 1323 K as a function of sintering temperature. b Vickers hardness as a function of the inverse square root of grain size. Reproduced with permission from [119]. Copyright © 2012, Elsevier

Transparent YAG ceramics have been fabricated by using SPS (Dr. Sinter 2050) and commercial YAG powder as starting material [114]. The YAG powder was of nonagglomerated spherical shaped particles with average particle diameter of 34 ± 17 nm and lognormal particle size distribution. Graphite foils (Grafoil) were used as spacers between the sample to be sintered and the graphite die and the punches. A preloaded pressure of 100 MPa was applied before the heating was started. The pressure was maintained during the whole heating process and at the SPS temperatures during the sintering. The temperature was increased to 600 °C within 3 min, after which it was raised to sintering temperatures at a heating rate of 100 °C min⁻¹.

During the densification of nanocrystalline powder compacts by using cold pressing, it is very common that large pores could be entrapped, due to the agglomeration of the fine particles. Remove of such pores during pressureless sintering and densification is quite difficult, due to the presence of the junctions with high dihedral angles at the pore–grain boundaries. However, the SPS compact of YAG at 900 °C exhibited particles with homogeneous distribution and pores with comparable size with the particle sizes, as shown in Fig. 7.28 [114]. This was mainly because the YAG powder had no agglomeration. Such homogeneous and small pore size could not result in full densification when using pressureless sintering, because many pores had high coordination numbers. However, this problem could be addressed by using SPS, due to the application of pressures. Pressures would assist the closure of such pores through grain sliding.

Fig. 7.28

Low (a) and high (b) magnification SEM images of the samples derived from the YAG compact subjected to SPS for 3 min at 900 °C and 100 MPa, which were further annealed at 900 °C for 5 h. Homogeneous distribution of the particles and pores with comparable sizes could be observed. Reproduced with permission from [114]. Copyright © 2007, Elsevier

Figure 7.29 shows SEM images of the samples sintered at higher temperatures [114]. The sample sintered at 1250 °C was still opaque, while the one sintered at 1400 °C had been transparent. Both samples showed homogeneous microstructure. Relative density of the YAG samples was increased significantly from 83 to 99.6 %, although the temperature range was as narrow as 150 °C. Grain size in this temperature range was increased in a parabolic manner. Significant grain growth was observed at the temperatures of above 1400 °C. The grain growth behaviors below and above 1400 °C were different [114].

Fig. 7.29

SEM images of fracture surfaces of YAG ceramics after SPS for 3 min at 100 MPa at different temperatures: a 1250 °C and b 1450 °C. Reproduced with permission from [114]. Copyright © 2007, Elsevier

SPS time duration had effect on densification and grain growth of YAG ceramics. For example, as the SPS duration was increased from 3 to 6 min, the relative density was increased tremendously, especially at lower temperatures, e.g., <1300 °C, as shown in Fig. 7.30 [114]. Therefore, if the sintering duration was increased from 3 to 6 min, the sintering temperature could be reduced by about 75 °C, for a given relative density. After 6 min sintering at 1350 °C, transparent YAG ceramics could be obtained. It was worth mentioning that the sample sintered for 6 min at 1300 °C had higher density than the one sintered to 3 min at 1350 °C, which indicated the negative effect of grain growth on densification during SPS. In this case, the densification was governed by the atomic transport processes at surface of the particles and particle–particle interfaces. Densification and grain growth by grain rotation and sliding would convert the originally high-angle grain boundaries to low-angle grain boundaries that in turn are annealed out by short-circuit diffusion. TEM observations verified the presence of such low-angle grain boundaries in dense ceramics was confirmed by TEM analysis, as shown in Fig. 7.31 [114]. In addition, Fig. 7.30 also indicated the effect of pressure.

Fig. 7.30

Relative density of the YAG powder compacts after SPS for 3 and 6 min at 50 and 100 MPa, as a function sintering temperature. Reproduced with permission from [114]. Copyright © 2007, Elsevier

Fig. 7.31

TEM image of the YAG ceramics after SPS for 3 min at 1400 °C and 100 MPa. The high-angle grain boundaries appeared as straight lines. The low-angle grain boundaries (arrowed) appeared to be curvy due to the strain field caused by dislocations. Nanometer-sized residual pores were present at the grain-boundary junctions. Reproduced with permission from [114]. Copyright © 2007, Elsevier

In another study, it was found that the way to treat the precursor powder had significant effect on performance of YAG ceramics by using SPS [142]. When sintering YAG synthesised by using reverse-strike co-precipitation with SPS, the precursor powder was calcined at 1000 °C and dispersed by using ball milling with α-alumina (BM_A) or zirconia (BMz) balls or ultrasonication (US). All the three dispersed powders could be consolidated by using SPS to nearly theoretical density, but only the US powder was optically transparent, as shown in Fig. 7.32. The 1-mm-thick transparent sample possessed a transmittance of about 60 % at 600 nm. According to Raman spectroscopy, it was found that there were some phonon vibrational shifts, which implied that minor secondary phases could be have introduced due to the contamination from the milling media. Such minor secondary phases could not be detectable by using XRD, because of their very small quantity. This result further indicates that special carefulness is important when processing transparent ceramics.

Fig. 7.32

Photographs of the samples derived from powders of BM_Z (a), BM_A (b), and US (c) after thermal annealing. Reproduced with permission from [142]. Copyright © 2013, Elsevier

Recently, SPS has been used to prepare pyrochlore transparent ceramics, including Lu₂Ti₂O₇ [143, 144], Lu₂Hf₂O₇ [145], and Lu₃NbO₇ [146, 147], which is still a challenge by using other sintering techniques. Figure 7.33 shows photograph and transmittance curves of transparent Lu₂Ti₂O₇ pyrochlore ceramics, which was synthesized by using SPS through reactive sintering directly from Lu₂O₃ and TiO₂ oxide precursors at 1723 K for 45 min [143]. The annealed ceramics demonstrated a transmittance of 72 % at wavelength of 2000 nm and 40 % transmittance at 550 nm. The average grain size was 14.5 μm with uniform microstructure. Representative results of Lu₃NbO₇ transparent ceramics, fabricated by using SPS, are shown in Figs. 7.34 and 7.35 [147].

Fig. 7.33

Transmittance spectra of Lu₂Ti₂O₇ sintered at 1723 K for 45 min in different ranges: a 190–2500 nm and b 2.5–8.5 μm. The inset is photograph of the transparent ceramics. Reproduced with permission from [143]. Copyright © 2011, Elsevier

Fig. 7.34

Photographs of the Lu₃NbO₇ ceramics SPS sintered at 1500 °C and annealed at various temperatures in air for 6 h: a 750 °C, b 850 °C, c 950 °C, and d 1050 °C. Reproduced with permission from [147]. Copyright © 2013, Elsevier

Fig. 7.35

Transmittance curves of the Lu₃NbO₇ ceramics SPS sintered at 1450 °C and annealed at various temperatures in air for 6 h: a 750 °C, b 850 °C, c 950 °C, d 1050 °C, e 1150 °C, and f 1250 °C. Reproduced with permission from [147]. Copyright © 2013, Elsevier

Another distinctive advantage of SPS is its ability to processing the transparent ceramics containing elements that are unstable in air, such as nitrogen (N), phosphorus (P), and fluorine (F). Examples include AlON [148] and Yb³⁺-doped Sr₅(PO₄)₃F [149]. More significantly, Sr₅(PO₄)₃F is a hexagonal crystal, instead of cubic ones as required according to the general understanding on transparent ceramics, as discussed previously. 20 mol% Yb³⁺-doped Sr₅(PO₄)₃F nanopowder was synthesized by using a co-precipitation method, with Sr(NO₃)₂, Yb(NO₃)₃, (NH4)₂HPO₄, and NH₄F as starting materials [149]. Solution containing 0.03 M Sr(NO₃)₂ and 0.0013 M Yb(NO₃)₃ was mixed with solution of 0.013 M NH₄F and 0.04 M (NH₄)₂HPO₄ to precipitate the precursor particles. The powder was ball milled for 20 h using ethanol as a solvent first and then dried at 70 °C and sieved with a 200 mesh screen. The powder was finally put into graphite mold and then separated by foil and BN spraying coatings from the mold. The sample was sintered in a FCT System SPS facility, at a heating rate of 50 °C min⁻¹ from RT to 1050 °C at a loading pressure of 100 MPa for 8 min. A trance of second phase, Sr₉Yb(PO₄)₇, was found in the sintered ceramics, which could be suppressed by using a small amount of SrF₂.

Photographs of representative Yb³⁺-doped Sr₅(PO₄)₃F transparent ceramics are shown in Fig. 7.36 [149]. A TEM image of the ceramics is also included in the figure. It was found that almost no porosity could be observed in the SPS-sintered sample. The grain sizes were ranged from 40 to 200 nm, with an average value of about 150 nm. In-line optical transmittance spectrum indicated that the Sr₅(PO₄)₃F:Yb ceramics had a strong absorption in the range of 850–980 nm, showing high potential to be high-power laser material. The transmittance near the emission wavelength was above 74 % at 1000 nm.

Fig. 7.36

Photographs (top) and TEM image (bottom) of the Yb³⁺-doped Sr₅(PO₄)₃F transparent ceramics processed by using SPS. Reproduced with permission from [149]. Copyright © 2012, John Wiley & Sons

7.6 MW-Processed Transparent Ceramics

The use of MW sintering to fabricate transparent ceramics is relatively less popular currently. Transparent ceramics that have been processed by using MW sintering include Al₂O₃ [150–152], YAG [153], AlON [154], and mullite [155]. MW sintering has been shown to be an effective technique to produce transparent ceramics as compared with the conventional sintering technologies.

First example discussed in this section is MW-processed Al₂O₃ transparent ceramics reported by Cheng et al. [150]. High-purity (99.99 %) commercial Al₂O₃ powder with average particle size of 0.15 μm was used as precursor. A total of 0.05 wt% of MgO was incorporated through a wet-chemical route with Mg(NO₃)₂·5H₂O as sintering promoter. Green pellets, with relative densities of 52–54 %, were prepared by dry pressing uniaxially, followed by cold isostatic pressing at 280 MPa. The compacted pellets were preheated at 1100 °C for 2 h in a resistance furnace to burn out the binder. Microwave sintering was carried out using a TE103 single-mode microwave applicator that was coupled with a 1.5-kW microwave generator operating at 2.45 GHz for small samples with diameters of <12.7 mm and a multi-mode microwave applicator with a 6-kW microwave power source for large samples with diameters of up to 25.4 mm. Ultrahigh-purity hydrogen under ambient pressure was used as sintering atmosphere.

Figure 7.37 shows photograph of a representative transparent Al₂O₃ ceramics that was microwave sintered at 1750 °C for 45 min [150]. In the reported experimental range, optical transmittance of the samples was slightly increased with increasing sintering time. Figure 7.38 shows SEM images of the samples microwave sintered at 1750 °C for different time durations. All samples exhibited uniform microstructure and grain size, with almost 100 % relative density, i.e., 3.97 g cm⁻³. As the sintering time was increased from 15 to 45 min, the average grain size was increased from 20 to 40 μm.

Fig. 7.37

Photograph of a representative highly transparent Al₂O₃ ceramics microwave sintered at 1750 °C for 45 min. Reproduced with permission from [150]. Copyright © 2002, Elsevier

Fig. 7.38

SEM images of the Al₂O₃ ceramics with 0.05 wt% MgO microwave sintered at 1750 °C at a heating rate of 100 °C min⁻¹ for different time durations: a 15, b 30, and c 45 min. Reproduced with permission from [150]. Copyright © 2002, Elsevier

Some microwave sintered transparent Al₂O₃ samples were post-sintered in microwave field at 1850–1880 °C for 2 h under ultrahigh-purity hydrogen (H₂) atmosphere, in order to convert the polycrystalline ceramics to single crystals. The processing is schematically shown in Fig. 7.39 [150]. A 9.5-mm-diameter Al₂O₃ disk, which was an as-sintered transparent sample by using MW sintering at 1750 °C for 30 min and supported by a high-purity Al₂O₃ tube, was placed in the single-mode microwave cavity to apply microwave post-sintering treatment. It was found that the temperatures at the center and periphery of the Al₂O₃ disk could be slightly different, due to the cooling effect of the flowing H₂ on the peripheral area.

Fig. 7.39

Schematic of the microwave setup for solid-state conversion of polycrystalline Al₂O₃ sample to single-crystal sapphire. Reproduced with permission from [150]. Copyright © 2002, Elsevier

As shown in Fig. 7.40, significant change in microstructure was observed in the post-sintered Al₂O₃ sample that was microwave heated at 1880 °C for 30 min [150]. It was demonstrated that the conversion of the polycrystalline ceramics to single crystal could be achieved by using the microwave post-sintering process. The peripheral area of the sample was still polycrystalline with average grain size of 30–40 μm, as shown at the top panel in Fig. 7.40, whereas the center part of the post-sintered sample was converted into single crystal without the presence of grain boundaries, i.e., the bottom panel of the figure. The converted sample exhibited an enhanced transmittance by about 20 % as compared with the as-sintered ones, showing the feasibility and advantage of MW sintering technique.

Fig. 7.40

SEM image of the Al₂O₃ sample microwave post-sintered at 1880 °C for 120 min, showing microstructure development due to the conversion of polycrystalline to single-crystal structure. Reproduced with permission from [150]. Copyright © 2002, Elsevier

Another example of transparent Al₂O₃ was fabricated by using microwave sintering at 2.54 GHz [151]. Figure 7.41 shows a schematic diagram of the microwave sintering with a special designed insulating casket. The hole on the top cover was used for infrared pyrometer. SiC powder was used as a microwave absorber, acting as a low-temperature microwave susceptor to preheat the Al₂O₃ samples, because Al₂O₃ is a poor microwave absorber at room temperature.

Fig. 7.41

Schematic diagram of a microwave sintering with specially designed insulating casket. Reproduced with permission from [151]. Copyright © 2003, John Wiley & Sons

The precursor alumina powder was doped with 350 ppm Y₂O₃ and 500 ppm MgO. Binder was removed by heating to 650 °C at 5 °C min⁻¹ with by using a conventional furnace. The average green density of the tubes after binder removal was 1.63 ± 0.01 g cm⁻³, i.e., 40.85 ± 0.35 % relative density. A multi-mode 6-kW, 2.45-GHz microwave furnace was used to sintering the samples. The sintering temperature and heating rate of microwave sintering were controlled by adjusting the input power. The sintering temperatures in the microwave furnace were measured by using digital infrared pyrometer.

Figure 7.42 shows density of the samples, for microwave and conventional sintering, as a function of sintering temperature [151]. Without dwelling at the sintering temperatures, microwave sintered Al₂O₃ samples reached relative density of 95 % after sintering at 1350 °C, which was much lower than 1600 °C for the conventionally heated samples. Also, the microwave sintered sample was almost fully densified after sintering at 1400 °C, while the density of the sample sintered through the conventional sintering was only about 50 %. Microwave sintering reduced the temperature for the samples start to densify by about 300 °C.

Fig. 7.42

Density of the samples, sintered by using microwave sintering and conventional sintering without holding at the sintering temperature, as a function of sintering temperature. Reproduced with permission from [151]. Copyright © 2003, John Wiley & Sons

It was found that both the microwave sintering and the conventional sintering exhibited a similar grain growth behavior, as shown in Fig. 7.43 [151]. Therefore, the enhancement in densification of Al₂O₃ by the microwave sintering was not due to the fast heating rate. However, the microwave densification corresponded to apparent activation energy of 85 ± 10 kJ mol⁻¹, much lower than the value of 520 ± 14 kJ mol⁻¹ for the conventional sintering, meaning that the there was a difference in densification mechanism between the two sintering techniques, which deserves for further investigation.

Fig. 7.43

Grain growth trajectories of the microwave sintered and conventionally sintered samples. Reproduced with permission from [151]. Copyright © 2003, John Wiley & Sons

A gyrotron-based system operating at a frequency of 24 GHz with microwave power up to 6 kW has been used to develop YAG transparent ceramics through reactive sintering from oxide precursors [153]. TEOS with contents of 0.3 and 0.5 wt% was used to provide SiO₂ as sintering aid.

Figure 7.44 shows photograph and microstructure of representative sample [153]. The sample contained 1.0 % Yb and 0.5 wt% TEOS and sintered at 1730 °C for 10 h. As shown in Fig. 7.44b, on the fracture surface, few pores were observed, most of which were preferentially within the grains. The sample exhibited grain sizes ranging from 5 to 20 μm, as shown by the polished and thermally etched surface in Fig. 7.44c. It was also found that the content of TEOS had no significant effect on grain size of the YAG transparent ceramics. Figure 7.44d indicated that amorphous or crystalline phase enriched with SiO₂ could be found at some triple junctions or grain boundaries. Such phases might have been formed during the thermal etching process, due to the diffusion of silicon through the grains toward the grain boundaries.

Fig. 7.44

Photograph a and SEM images of the sample 1.0 % Yb:YAG with 0.5 wt% TOES sintered at 1730 °C for 10 h: b fracture surface and b, c polished and thermally etched surface. Reproduced with permission from [153]. Copyright © 2013, Elsevier

As stated previously, the sintering behavior of various transparent ceramics, including YAG, can be improved by using silica-containing sintering aid. In conventional sintering, silica functions to form a liquid phase at about 1390 °C. The liquid phase makes it easier for the particles to rearrange and enhances mass transport by grain-boundary diffusion, so as to promote densification. During cooling from the sintering temperature, secondary phases of Al₂O₃–Y₂O₃–SiO₂, either amorphous or crystalline, would be formed. The presence of such phases is harmful to optical properties due to their light scattering. Therefore, a trade-off is necessary to optimize the content of silica, i.e., promoting densification yet without compromising optical performances.

One of the most distinctive features of the microwave sintering is that there is a special behavior in microwave power in the initial stage of heating. For example, in that study, there was a drop in microwave power as the temperature was above 700 °C [153]. It was attributed to the increase in microwave absorption of the sample, due to the increase in their ionic conductivity. The microwave power was also influenced by the exothermal reactions of the intermediate Y₄Al₂O₉ and YAlO₃.

The formation of YAG through solid-state reaction is controlled by the diffusion of Al³⁺ [156]. Such diffusion of Al³⁺ was enhanced by the microwave field, so that the reaction temperature of YAG and its sintering temperature were decreased, as evidenced by the pores entrapped in grains, as shown in Fig. 7.44b. Similarly, the formation of the liquid phase was also promoted. In other words, the content of sintering aids could be reduced when using microwave sintering, which however needs further clarification.

The enhancement in mass transport of polycrystalline materials has been explained by the so-called ponderomotive action of electromagnetic field on charged particles [157, 158]. The theory suggested that the ponderomotive effect is more effective in ionic crystalline materials with high microwave absorption and it is maximized when the imaginary part and real parts of the complex dielectric permittivity of the materials have very close values. It can be expected the microwave processing parameters could be further optimized in such a way that all diffusion processes would be operated concurrently during the densification of powder compacts. In this case, the amount of sintering aids could be greatly reduced or even completely avoided.

Besides, Roy et al. found that single-phase ALON could be synthesized from Al₂O₃ and AlN by using microwave heating at 1650 °C for 1 h, while fully dense and transparent ALON ceramics were obtained through microwave sintering at 1800 °C for 1 h [154]. A total transmission of 60 % was achieved for the polished sample with a thickness of 0.6 mm. The microwave sintering was carried out by using a homemade 1.5-kW, 2.45-GHz single-mode microwave applicator in nitrogen (N₂) flowing at ambient pressure. The heating rate was about 100 °C min⁻¹ by controlling the incident microwave power.

Transparent mullite ceramics were developed by using microwave sintering at about 1300 °C [155]. The microwave sintering was conducted by using a 2.45-GHz, 900-W microwave furnace. The samples were heated in the microwave field directly from room temperature to 1300 °C and held at the temperature for 10 min, then allowed to cool down naturally. The results indicated that aerogel of mullite stoichiometry, free of agglomeration and highly active, was key factor to achieving transparent mullite ceramics.

7.7 Concluding Remarks

Vacuum sintering is the most widely used technique to fabricate transparent ceramics, due to its various advantages, such as simple processing, large size product, and relatively cost-effectiveness. HP and HIP offer lower sintering temperatures, so that transparent ceramics with fine-grained microstructures could be obtained, which make them attractive in some applications requiring strong mechanical performances.