Figure 1 shows the microstructure of the MgAl2O4 ceramic sintered at 1420 °C for 3 h. Panels include micrographs of fracture surface (a) and thermal etched fracture surfaces (b) and (c). In Fig. 1 (a), the fracture surface is almost flat with no contrast between the grains and grain boundaries. Obviously, the as-sintered MgAl2O4 ceramic fractured in an intergranular fracture mode. Traditionally, polishing and etching processes are required for ceramics fractured in this mode before observing the microstructure. However, in this process, machining through grinding or polishing were not needed due to the previously formed flat surface after fracturing. In Fig. 1 (b) and (c), clear grain boundaries can be distinguished after directly thermally etching the fractured surface. Grain sizes were distributed on a scale of 50–500 nm; all were imaged successfully. The original pores after etching are still similar as that in Fig. 1 (a). In summary, the as-prepared fine-grained MgAl2O4 ceramic described here has grains small enough to be totally imaged on the micrometer scale by SEM. As a result, the machining step was diminished, and the etching process proceeded directly on the fracture surface. The simplified sample preparation method is not only time-saving but can also protect the original microstructure from damage during machining.
On the contrary, if the machining processes such as grinding or polishing were followed, then all destructions such as scratches and pullout grains may occur. Figure 2 shows the three typical destructions on the microstructure of the MgAl2O4 ceramic that resulted from polishing and etching. Surprisingly, Fig. 1 and Fig. 2 were SEM pictures taken from the same sintered MgAl2O4 ceramic, but they appear completely different. In Fig. 2 (a) and (b), many grains as small as 30 nm or less are shown. However, the actual grain sizes are distributed between 50 nm and 500 nm. As against the same ceramic sample shown in Fig. 1, one sees that the machining and etching processes on the ceramic surface in Fig. 2 resulted in some misleading phenomena such as ultra-small grains. The abrasive particles might be left on the ceramic surface after polishing.
Fig. 2 (b) shows obvious scratches that resulted from grinding. These affect the imaging quality and the calculated average grain size. Fig. 2 (c) shows a single pore with a size of about 1 μm. This is several times bigger than the average grain size. During grinding and polishing, some grains may be pulled out due to the weak grain boundaries; the pore-like information is then detailed on the SEM picture. The large pore in Fig. 2 (c) may result from the grain pullout during machining. These types of damage (false ultra-small grains, scratches, etc.) are barriers for researchers and obscure the true features of the ceramic microstructure.
For a ceramic that fractures in the transgranular fracture mode, observing the fracture surface is preferred for obtaining the actual original morphology of the microstructure. No etching process is usually necessary due to the contrast between the grains and the grain boundaries that already exist because the fracturing behavior occurs just along the grain boundaries. Glassy phases and impurities may exist on the grain boundaries of these ceramics. This will also affect the imaging quality of SEM. To increase contrast, one can groove the grain boundaries by an etching process on the fracture surface before observation. The microstructure can be of higher quality after etching.
Figure 3 (a) shows the thermal etched fracture surface of the ZrO2 ceramic. No pore can be identified here because the ZrO2 ceramic was pre-sintered first, and then HIPed at 1260 °C for 3 h with 200 MPa argon pressure. The ceramic was fully densified. The grain boundaries after grooving by thermal etching are more distinct than without etching [Fig. 3 (b)]. Comparing the two micrographs from the different pre-processing steps suggests that although the etched fracture surface shows better contrast, it somehow changes the original state of the fracture surface. Clearly, the edges and corners of the grains in Fig. 3 (a) were erased via thermal etching. Nevertheless, the micrographs obtained from the fracture surface with or without etching processes were better than Fig. 3 (c), which resulted from polishing and etching processes. Although it seems that no obvious destruction occurred on the microstructure of the ZrO2 ceramic, some white impurities do remain on the surface. In summary, the etching process on the fracture surface may not be necessary for ceramic fractures just along the grain boundaries or in cases with clear grain boundaries.
However, not all ceramics fractured in a transgranular mode show clear contrast between the grains and grain boundaries. In this case, proper etching processes are necessary before SEM observation. For example, the alumina ceramic fractures in a transgranular fracture mode, but the contrast between the grains and grain boundaries is low. Figure 4 (a) and (b) show SEM pictures of an as-fractured surface without etching and a thermal etched fracture surface of the alumina ceramic, respectively. Obviously, the pores observed from Fig. 4 (a) and (b) were very similar. Figure 4 (b) shows better contrast between the grains and grain boundaries due to the glassy phases or impurities that were erased by etching.