III. RESULTS AND DISCUSSION

Calcium fluoride is a naturally abundant ionic compound with a face-centered cubic structure. The XRD patterns of the raw powder and the 700 °C SPS-sintered ceramic samples are shown in Fig. 1 with every peak indexed to the standard CaF₂ diffractions (JCPDS 89-4794, Fm-3m). It can be clearly observed that the full width at half maximum (FWHM) of the raw powder was much larger than that of the sintered ceramic. From the XRD data, with the help of the Scherrer equation, the crystalline size of the CaF₂ raw powder was estimated to be around 79.3 nm. The SEM images in Fig. 2 showed the micrographs of the raw CaF₂ powders before and after ZrO₂ ball milling. The identifiably well-dispersed powder after ball milling indicated that the raw powders used as a starting material for the SPS sintering were only soft agglomerates. The impurities present in the CaF₂ raw powder were not readily detectable via XRD. Furthermore, the EDS spectrum seen in the inset of Fig. 2(a) also barely reveals any significant impurity elements – only very small peaks are observed.

FIG. 1.

XRD of raw powder and 700 °C SPS-sintered ceramic.

FIG. 2.

SEM image of raw powder (a) and powder after ball milling (b), and EDS spectrum of raw powder (inset of a).

The SEM images of the CaF₂ ceramics, sintered at different temperatures, were exhibited in Fig. 3 to investigate the microstructure of the ceramic fracture surface. As the sintering temperature increased from 500 to 850 °C, it is clear that the most distinctive change was the grain size, which increased from several to hundreds of micrometers. For the larger and smaller grains, respectively, the microstructure image of the sample sintered at 500 °C in Fig. 3(a) showed both transgranular and intergranular fracture features. Both intergranular and intragranular – but predominately intragranular – pores with sizes as large as 0.6 μm can also be observed in Fig. 3(a). Some of the larger void spaces with faceted shapes were actually small grains pulled off by the other side of the ceramic sample. Based on the readily identifiable residual pores at the center of the grains, it is apparent that the rate of grain growth from the nanosized crystallites was actually significantly faster than the densification process. The minuscule flaws present, such as pores inside the grains, had the effect of amplifying the applied stress and ultimately resulting in transgranular fractures. As shown in Figs. 3(b)–3(d), all the small grains grew bigger from sintering temperature increases (550, 600, and 650 °C); only transgranular fracture features were observed. However, despite this observable trend, both intergranular and intragranular porosity significantly decreased – which is not consistent with the pertinent phenomenon describing the general atom diffusion mechanism for matter transport during densification. Pores which lie close to grain boundaries should disappear more rapidly than those which lie in the centers of grains which instead tend to remain after the diffusion densification process.^{Reference Burke13} Therefore, in this stage of SPS sintering CaF₂ ceramic, the predominant sintering mechanism may actually involve other matter-transport mechanisms acting alongside diffusion. Besides the clearly observable quickly grown grain sizes in Figs. 3(e)–3(h), with the temperatures further increased to 700, 750, 800, and 850 °C, the grains actually showed irregularly patterned surface morphologies and ultimately started to strip apart from each other. These morphologies appear as if they were formed via condensation.

FIG. 3.

SEM images of the fracture surface of the CaF₂ ceramics sintered at 500 °C (a), 550 °C (b), 600 °C (c), 650 °C (d), 700 °C (e), 750 °C (f), 800 °C (g), and 850 °C (h).

Different grain growth mechanisms can be represented by the numeric value of the grain growth exponent.^{Reference Rahaman17} The original grain size at all the SPS temperatures and prior to beginning the SPS process altogether, was assumed to be a constant (G ₀ = 1.3 μm) – determined via analyzing the average grain size of a sample heated to 500 °C without dwelling. Using the grain size data over the entire empirical temperature range, the Arrhenius-type plot was generated and is shown in Fig. 4. Then, for low, medium, and high temperatures, respectively, the best linear regressions (R) with an exponent of 4 were found to have activation energies (Q) of 276, 98, and 449 kJ/mol. Therefore, the grain growth was actually controlled by surface diffusion. Because of the multitude of concurrently enacting complicated parameters during SPS sintering, such as temperature, pressure, and plasma, it is very difficult to accurately estimate the activation energy associated with the densification process.^{Reference Lu18}

FIG. 4.

Arrhenius plots of the grain size versus the reciprocal SPS temperature.

The relative density and average grain size for the CaF₂ ceramics sintered between 500 and 850 °C are shown in Fig. 5. Surprisingly, with these samples, the density did increase with increasing sintering temperatures, however only before a certain point (around 600 °C); after this point the density actually decreased, with an even more rapidly plummeting density rate at temperatures 750 °C and higher. However, as the sintering temperature increased, the average grain size continued to grow, regardless of the declining density values. This sintering behavior has not been reported before and is most likely attributed to the starting materials and the particular sintering technology involved. In SPS the large pulse electric current used and apparent discharge-induced plasma formation ultimately produces a nonequlibrium electrothermal effect, effectively stimulating the evaporation/condensation of the surface material.^{Reference Orru, Licheri, Locci, Cincotti and Cao19} The sintering trajectory of CaF₂ ceramics is presented in Fig. 6. For the CaF₂ ceramics sintered at temperatures below 600 °C, the continued relative-density-increasing behavior indicates that the bulk transport mechanism was actually either plastic flow or grain-boundary-diffusion-aided densification – acting as the predominant mechanism for this stage as labeled in Fig. 6 (stage A). As stated previously, when sintering temperatures were increased above 600 °C, the curves actually went along the opposite direction from the general and well-documented classical sintering trajectory. In the literature it is currently believed that an evaporation condensation process cannot lead to the densification of a ceramic compact.^{Reference Kingery and Berg20} However, for the pressure-assisted high-temperature plasma sintering, surface transport with evaporation condensation was actually proposed as the predominant mechanism for the sintering stage B in Fig. 6.

FIG. 5.

Densification and grain growth behavior of different temperature SPS-sintered CaF₂ ceramics.

FIG. 6.

Sintering trajectory of SPS-sintered CaF₂ ceramics.

The IR transmittance spectra of the CaF₂ ceramics sintered at different temperatures are all consolidated and compared in Fig. 7. Porosity/relative-density, secondary phase, and double refractions were identified as the most important factors governing the final transmittance of polycrystalline ceramics. As the sintering temperature increased, the porosity of the ceramics ultimately reached their lowest point at 600 °C; and since pores act as light-scattering centers/sources, decreasing the porosity inexorably improved IR transmittance. Interestingly, despite this correlation, when the sintering temperature was increased above 600 °C and the porosity began to increase again, the IR transmittance did not decrease until the temperature was higher than 750 °C. This could be explained by the role of a secondary phase. As barely detected from the EDS signal in Fig. 2(a), there existed some impurity in the raw CaF₂ powder. As the sintering temperature increased, the amount of the low-melting-point secondary phases was reduced. Therefore, although the relative density of the sintered ceramics dropped a little bit, the transmittance of these CaF₂ ceramics kept increasing. Because of the special mechanism involved in the SPS sintering of CaF₂ ceramics, the transmittance was ultimately much lower than that of other hot pressed^{Reference Lyberis, Stevenson, Suganuma, Ricaud, Druon, Herbst, Vivien, Gredin and Mortier1} and hot isostatic pressed^{Reference Aubry, Bensalah, Gredin, Patriarche, Vivien and Mortier3} CaF₂ ceramics. The digital image of CaF₂ ceramic sintered at 700 °C is presented in Fig. 8. The printed letter under the 2 mm thick ceramic was visible. It is commonly found that the natural CaF₂ mineral is often deeply colored due to the presence of F centers. It is more likely that the SPS-sintered CaF₂ ceramics possessed many color centers due to the lower valence ion doping from the 3% impurity. Therefore, annealing in a fluorine-related atmosphere would be a proper way to improve the transmittance.^{Reference Fedorov, Osiko, Kuznetsov and Garibin21} It is difficult to prove whether original impurity is left in grains or on grain boundaries after SPS sintering and fluorine atmosphere annealing. However, from the image of the sample after annealing, the area at the center showed stronger colorations than that on the edge. The annealing after the densification of ceramic is rather slow in defusing out impurity or defects in the bulk.

FIG. 7.

IR transmittance of different temperature SPS-sintered CaF₂ ceramics.

FIG. 8.

Digital image of SPS-sintered CaF₂ ceramics at 700 °C.