An Improving Densication in Spark Plasma Sintering Ultrane-grained Y2O3 Transparent Ceramic by Particle Fracture and Rearrangement

Herein, we report a new strategy on an improving densication in spark plasma sintering (SPS) ultrane-grained Y 2 O 3 transparent ceramic using anisotropic-shaped nanorod powders. At the low temperature stage (<600°C), fracture and rearrangement of nanorod powders will appear which can further improve the initial density of particle compact and reduce the initial particle size. When the temperature reaches beyond 600°C, the sintering eciency of nanorod powder compact will exceed that of the green body packed by spherical powders. The ceramic sample sintered at 1300°C from nanorods is transparent with ultra-ne grain and shows good optical and mechanical properties, while the corresponding ceramic from near-spherical nanocrystalline powders is not dense enough and opaque.


Introduction
Transparent ceramics can nd a wide range of applications, such as transparent armors, infrared (IR) domes/windows, host material for uorescence application, and laser hosts, on account of their processing exibility in fabricating large-sized or complex-shaped products [1][2][3][4][5][6]. In most cases, the preparation process of any ceramics is sintering the consolidated/packed powders at high temperatures [7]. Therefore, the properties of the powder signi cantly in uence the microstructures and performances of transparent ceramics.
As highly transparent ceramics require full density, well-reputed wisdom believes that ceramic powders with high sintering e ciency for obtaining highly transparent ceramics have some important characteristics including small crystallite size, narrow size distribution, spherical/near-spherical shape, and low-degree agglomeration [7]. It is well known that the surface energy of starting powders is serving as the driving force for densi cation during the sintering process [8,9]. Thus, the smaller the crystallite/particle size of the initial powders, the higher the densi cation rate [7]. Besides, spherical nanoparticles with narrow size distribution guarantee homogeneous packing and low pore-to-particle-size ratio in the ceramic green body [10]. Mono-dispersed powders can avoid large pores between agglomerates and forming crack-like voids in the nal ceramics [11,12]. Fallaciously, achieving these aforementioned desirable characteriastics for powders faces a sea of troubles. For example, addressing agglomeration of the powders, especially for nanocrystalline powders, is a remaining challenge as nanoparticles tend to agglomerate to reduce surface energy due to the minimum energy principle [13,14].
Agglomeration will directly result in heterogeneity of the packed green bodies, which will cause a phenomenon referred to as differential sintering [15]. It is inspiring that anisotropic-shaped powders, rather than spherical/near-spherical powders, can improve the densi cation process and provide some new densifying routes [16,17]. For example, with the aid of high pressure (5 GPa), Y 2 O 3 nanorod powder compact can achieve near full density through particle fracture, rearrangement, deformation, and interface sliding at only 500°C without atomic diffusion [16]. However, most sintering methods, such as spark plasma sintering (SPS), vacuum sintering, and hot isostatic pressure (HIP) sintering cannot provide such a high pressure [1,7]. In addition, it is still an unrevealed controversy that whether anisotropicshaped nanorod powders are better than spherical/near-spherical powders supported by conventional wisdom in terms of sintering e ciency in these conventional sintering methods with dominated densi cation mechanisms of atomic diffusion. If they are suitable for preparing transparent ceramics, is there a sintering mechanism different from that of spherical/near spherical powders?
To address these questions, we have tried to use a solvothermal method to synthesize Y 2 O 3 nanocrystalline powders with singularly shaped nanoparticles. As a comparison for checking sintering e ciency, near-spherical Y 2 O 3 nanocrystalline powders were also synthesized by a co-precipitation method. Then, SPS was employed to sinter ceramics at various temperatures from 25 to 1400°C. Speci c surface areas and microstructures of these powders, as well as the relative densities, microstructures, and grain sizes of the sintered ceramics, were analyzed. The sintering e ciency and densi cation mechanism of the nanorods powders were investigated. Furthermore, the optical and mechanical properties of the Y 2 O 3 transparent ceramic obtained at the optimized sintering temperature were analyzed.

Material And Methods
Herein, the Y 2 O 3 nanorod powders were synthesized by a solvothermal method. Y(NO 3 ) 3 ·6H 2 O (> 99.99%, Yutai Qixin Chemical Co. Ltd, China) was dissolved in deionized water to a concentration of 0.1 mol/L. Then, the solution was added drop by drop into the diluted ammonia solution (0.5 mol/L). The mixed solutions were stirred for 6 h and aged for 20 h to get gel-like precipitate. After that, the precipitate was washed by deionized water and ethanol in turn several times using a centrifugal machine. The precipitate after drying and calcination can obtain nanocrystalline powders composed of spherical particles. To prepare singularly-shaped nanocrystalline powders, the washed precipitate was subsequently diluted by ethanol and placed into a Te on reactor to react at 200°C for 24 h. The reactants were washed centrifugally with ethanol and then dried at 60°C. Finally, the as-obtained precursor powders were sieved using a 200-mesh screen and calcined at 800°C to remove the adsorbed water and bound water. Nearspherical powders were fabricated by directly dring the washed precipitate, sieving, and calcinating the precursor powders at 800°C.
The as-prepared powders were placed into a graphite die with an inner diameter of 10 mm. The densi cation processes of the samples were carried out in an SPS system (LABOX-325, Sinter Land, Japan) with an ambient pressure of 6 Pa. A uniaxial pressure of 50 MPa was applied on the Y 2 O 3 samples during the sintering process. The samples are sintered at the temperature range of 25-1400°C with a ramping and cooling rate of 20°C/min, and hold for 5 min at the desired temperatures. Postannealing of the SPSed samples was conducted at 1000°C for 8 h in air using a mu e furnace. The asobtained samples were mirror polished for both two sides for in-line transmittance and Vickers hardness tests.
The crystal structure of the as-synthesized powders was investigated by X-ray diffraction using an X-ray diffractometer (Bruker D8 Advance, Bruker Co., German). The XRD data was recorded over the 2θ range of 10°-70° with a step size of 0.02°. High-magni ed microstructural images of the powders and ceramics were checked by transmission electron microscopy (TEM, F200, JOEL, Tokyo, Japan). The fracture surfaces of the ceramics sintered at different temperatures were investigated by scanning electron microscope (SEM, JIB-4700F, JOEL, Tokyo, Japan). Speci c surface areas of nanocrystalline powders were measured by a multipoint Brunauer Emmett Teller (BET) method (Tristar 3000, Micromeritics, Atlanta) using N 2 as the adsorbate gas. The in-line transmittance test was performed on a UV-VIS-NIR spectrophotometer (Lamda 750, Perkin Elmer, USA). Vickers hardness tests were operated at an applied load of 0.5 N for 15 s (DUH-211s, Shimadzu, Japan). At least 10 measurements were carried out on each sample. phase (PDF#43-1036). Broad diffraction peaks indicate that the crystallite size of the powders may be in nanoscale. After the whole pattern tting on the XRD data using the JADE software (Fitting error <10%), the average crystallite size of the calcined powder is calculated from peak broadening of XRD pattern using the Scherrer's equation [18], D=0.89λ/βcosθ, where D is the crystallite size, λ value is 1.5406 Å, θ is the angle of the peak maximum (in 2θ) and β is the full width at half maximum (in 2θ). From the calculation, the average crystallite size is 19.8±0.6 nm. Fig. 1B displays the microstructure of the asprepared powders after calcination under 800 °C. The Y 2 O 3 powders after solvothermal reaction and calcination present a rod-like morphology with low agglomeration and well-dispersed characteristics (Fig.  1B). The length of the nanorods is about 500 nm. Previous studies have shown that well-dispersed nanocrystalline powders can supply abundant surface energy to promote ceramic densi cation [19][20][21].

Results And Discussion
Thus, the as-obtained nanorod powders may have good sinterability. The formation mechanism of anisotropically shaped nanorods rather than spherical nanoparticles may be induced by the competition between crystallographic anisotropy and crystallite growth kinetics in the solvothermal reaction process [22]. Fig. 1C shows the microstructure for a monodispersed nanorod. The diameter of the nanorod is about 20 nm, which is consistent with the average crystallite size calculated from XRD data. Fig.1D-E present the HRTEM images taken from different areas in Fig. 1C. We can observe that a nanorod is composed of randomly oriented polycrystallines. Fig. 2A and Fig. 3A show that the green body from nanorod powders compacted at 50 MPa possessing a relative density of 36.7% is composed of randomly and loosely distributed nanorods. Fig. 2B-I reveal the microstructural evolutions of sintered ceramics from nanorod powders during sintering. As the sintering temperature reaches 900 °C, the nanorods are completely disappeared and near-spherical grains were observed. When the sintering temperature is in the range of 900-1200 °C, residual pore signi cantly decreases and bulk density gradually increases with the elevation of the sintering temperature (Fig. 2B-E and Fig. 3). When the temperature ≥1250 °C, pores almost disappeared and ceramics with relative density higher than 98.8% exhibit a dense structure (Fig. 2F-I and Fig. 3). It is worth noting that when the sintering temperature is higher than 1300 °C, the density begins to decrease slightly. This phenomenon that the density decreases with the grain growth is common in the pressure-assisted sintering process [23,24]. The process of grain growth easily induces pore growth and non-uniform grains. Few fusion-grown pores lead to a decrease in density [25]. In addition, we can see from Fig. 2 that the grain size of ceramic increases with the increase of sintering temperature. We use Nanomeasure software to statistically count the average grain size of ceramics from Fig. 2. Grains with obvious and distinguishable boundaries for selected statistics. As the grains of ceramics sintered at relatively low temperatures (200-1000 °C) are di cult to distinguish, XRD patterns of these ceramics with no polishing were recorded.
Whole patter tting on XRD patterns with tting errors <7% were conducted for calculating the average grain size. The broaden peaks (Fig. 3B) indicate that the grain sizes of these ceramics are in nanoscale. FWHMs of diffraction peaks of (222), (400), and (440) decrease with the increase of sintering temperature (Fig. 3C), indicating that grain starts to grow slowly after >400 °C. Image-based statistical and XRD calculated results show that the average grain size is less than 100 nm when the sintering temperature is <1200 °C, and is at the sub-micron scale when the sintering temperature is ≤1300 °C. As the temperature rises to 1350 °C, the average grain size will increase to 1.9 μm (Fig. 3D). 1350 °C is the T g (the critical temperature to trigger signi cant grain growth) of the Y 2 O 3 specimen in this work as the average grain size increases from 1.9 to 38 μm when the sintering temperature is elevated from 1350 to 1400 °C. It can be concluded that 1300 °C is the optimized sintering temperature for Y 2 O 3 nanorod powders as the sintered ceramic reaches near-full density relative density with the average grain size is only 0.7 μm. Besides, the corresponding ceramic after polishing is highly transparent. Its optical property is discussed in the next part.
To investigate the powder effect on the sintering, near-spherical nanoparticles with an average crystallite size of about 21 nm (see Fig. S1 in the supplementary le) were also densi ed at various temperatures. It can be seen from Figure 3b that when the temperature is ≤ 400°C, the density of the ceramic from the near-spherical powders is higher than that of the ceramic from nanorod powders. However, as the temperature reached 600°C, the density of the ceramic prepared by the nanorod powders increased signi cantly, which is higher than the density of the ceramic from the near-spherical powders. In the optimal sintering condition (1300 °C) for nanorod powders, the ceramic from near-spherical powders is completely opaque with many submicron-sized pores dwelled between grains (see Fig. S2 in the supplementary le). To elucidate the superiority of the nanorod powders over the near-spherical nanocrystalline powder in terms of sinterability, speci c surface areas of these two kinds of powders were measured. The results show that the BET surface area of the nanorod powders is 67 m 2 /g, higher than that of the spherical nanoparticles (43 m 2 /g). This increase in BET surface area is mainly due to that ethanol can effectively break up the structures of the hydrated hydroxide, effectively eliminating agglomerations caused by hydrogen bond [21]. Thus, the near-full densi cation of nanorod powder achieved at only 1300 °C is promoted by the high speci c surface area based on the well-reputed wisdom that surface energy provides a driving force for the densi cation of ceramic.
Furthermore, we have noticed that 600 °C is a breakthrough point for the increase in the density of nanorod powders. There may be new mechanisms in the sintering process. In the process of preparing transparent ceramics by high-pressure sintering, the nano-particles will yield under high temperature and high-pressure conditions, which will cause the microscopic residual stress and strain of the bulk piece to decrease [26,27]. The experimental results showed that the yttrium oxide nanorods were fractured and rearranged under the conditions of 5 GPa/500 °C [16]. It can be observed in Figure 2B that the nanorods have completely disappeared at 900°C, and the ceramic grains have not grown signi cantly (remaining at the nanometer level). Therefore, the nanorods are likely fractured under the pressure of 50MPa and high temperature. We consider that the fracture of nanorods will also cause the microscopic residual stress from the uncoordinated deformation of particles during consolidation. As microscopic deviatoric strain will lead to the X-ray line broadening in addition to reduced grain size [26,28], we calculate the residual strain through the XRD data after the whole spectrum tting. According to the well-reputed Williamson-Hall model [28,29] and its corresponding variations, the residual micro-strain (ε) of the specimen can be de ned as: where Δd obs is the FWHM, Δd ins the peak width at a stress-free state, Δd size the peak width of grain size and d(P, T) the d spacing of a given lattice plane which is incorporated in the Scherrer formula. During the calculations, we subtract the instrument resolution and don't disjoin the changes in the peak width of the various contributions. Therefore, the micro-strain can be denoted as follows: The calculated results show that the residual micro-strain increases with the increase of temperature and reaches the maximum value at 400 °C. As the temperature continues to increase, the residual strain decreases rapidly and tends to a stable value after 800 °C (Fig. 4). To elucidate the mechanism of microstrain variations, TEM images of ceramic samples sintered at 200-900 °C were recorded. The results show that when the sintering temperature is ≤ 400 °C, the sample is mainly composed of nanorod particles. As the temperature reaches 600°C, most nanorods disappear, and the length of remaining nanorods is greatly reduced. Near-spherical grains with a size of about 20 nm can be observed. Therefore, we can conclude that with increasing temperature, the fracture strength decreases and the nanorod particles fracture. When the temperature reaches 800 °C, the nanorods completely disappear, and the average grain size has grown to about 30 nm. As the temperature continues to increase, the grains grow further. Thus, the SPS process of the nanorod powders can be divided into three stages with the sintering temperature increases, as shown in the schematic in Fig.4.. The rst stage is the deviatoric-stressaccumulation stage. Then, nanorods begin to fracture and rearrange in the second stage. The stress begins to release and the strain decreases signi cantly. After that, grains growth is triggered. This stage is equivalent to the intermediate sintering stage in most ceramic sintering methods as the relative density of the corresponding ceramic sintered at 800 °C is about 77.2% [7]. Based on the aforementioned results, we consider that the fracture and rearrangement of nanorods under modest temperature and pressure also contributes to the high sintering e ciency of nanorods powders as it will further increase the surface area of the powders and consolidated density of ceramics. Y 2 O 3 is an attractive optical material using as infrared dome and laser host [30]. As Y 2 O 3 transparent ceramic was successfully obtained from nanorods powders, the in-line transmittance spectra and the optical image of the ceramic sintered at 1300 °C is presented in Fig. 5. It can be seen from the optical image that the sample sintered under the optimized condition is highly transparent with maximum transmittances of 65.9% and 80.8% in the visible light and infrared wavelength, respectively. The relatively lower transparency in the visible light band than the infrared band indicates the light scattering.
Based on the well-established light-scattering model [31,32], the main source for light scattering may be assigned to a similar size between the average grain and the visible light wavelength. Generally, transparent ceramics prepared by SPS tend to appear yellow or brown colors, which is mainly caused by the absorption of visible light by dislocations or oxygen vacancies formed during the sintering process [33][34][35]. The as-obtained Y 2 O 3 transparent ceramic is colorless, indicating near no dislocations or oxygen vacancies inside the transparent ceramic sample after annealing. Table 1 representatively lists the maximum transmittances of Y 2 O 3 transparent ceramics prepared by SPS in the visible light and infrared band. It can be seen from Table 1 that the transmittances of Y 2 O 3 transparent ceramics via SPS method in the visible light band are di cult to exceed 80%, which are generally lower than the transmittances in the infrared range. The sintering temperatures to prepare Y 2 O 3 transparent ceramics with good optical properties in most cases need to exceed 1400 °C [36][37][38][39]. However, when the nanorod powders were used in the present work, transparent ceramic with a transmittance of more than 80% in the infrared band was obtained at only 1300 °C. Moreover, the transmittance of Y 2 O 3 transparent ceramic obtained at 1300 °C as reported in other work is only 62.1% in the infrared band [39]. Although a higher thickness of the sample will reduce the transparency, it has no transmittance in the visible light band. To sum up, the assynthesized nanorod powders exhibit desirable sinterability and facilitate obtaining Y 2 O 3 transparent ceramic with good optical properties at a relatively low temperature. Furthermore, the mechanical properties are very important in mechanical industrial applications of transparent ceramics [41]. According to the Hall-Petch relationship, the strength of the material is inversely proportional to the square root of the average grain size [42]. Therefore, reducing the grain size is of great signi cance for improving the mechanical properties of transparent ceramics.

Conclusions
In summary, using the lab-made Y 2 O 3 nanorod powders as the starting materials, ultra ne-grained Y 2 O 3 transparent ceramic with good optical and mechanical properties was fabricated successfully at a relatively low temperature via SPS. An improving densi cation mechanism involving particle fracture and rearrangement at 600 °C /50 MPa is revealed. Speci cally, the following results were obtained: 1) Colourless Y 2 O 3 transparent ceramic with suppressed grain growth (average grain size is about 0.7 μm) was obtained successfully from nanorod powders, while the corresponding ceramic from nearspherical nanocrystalline powders is not dense enough and opaque.
2) The transparent ceramic shows maximum transmittances of 65.9% and 80.8% in visible light and infrared wavelength ranges, respectively. The Vickers hardness and fracture toughness are 9.0 GPa (theoretical hardness value of Y 2 O 3 is 7.7 GPa) and 1.30 MPa·m 1/2 , respectively.
3) The high sintering e ciency of nanorod powders originates from its high speci c surface area and unique sintering mechanism involving the fracture and rearrangement of nanorods.
4) The unique sintering process can be divided into three stages: (i) In the primary sintering stage, the deviatoric stress accumulates with the increase of sintering pressure. (ii) The nanorod particles can fracture and rearrange when the sintering temperature reaches 600°C, which associates with the further increase of compact density and decrease of initial particle size.   TEM images of the ceramics from nanorod powders sintered at 200-900 °C. Residual strain calculated from whole-pattern-tted XRD data as a function of sintering temperature was also presente. Combining microstructure evolution and residual stress analysis, a schematic is plotted to show the speci c densi cation mechanism of nanorod powders by fracture/rearrangement.