The role of grain boundary phase on contact damage behavior in silicon nitride ceramics was investigated using Hertzian indentation test. In order to consider only grain boundary effect, all of the microstructure effects except grain boundary must be fixed. The facts that could affect the contact damage behavior in silicon nitride ceramics are grain size, aspect ratio of grain, phase fraction, and interphase. In the prior study the effects of grain size, aspect ratio and phase fraction on contact damage behavior were considered simultaneously. But in those study they omitted the role of grain boundary effect because of the different sintering temperature of specimens. Especially because high toughness of silicon nitride are related with the crack deflection and bridging mechanism due to weak interphase boundary, we could expect that grain boundary phase could be a another factor for affecting the damage mode of Hertzian indentation.
Grain boundary effects were treated as two parts. One was grain boundary phase effect, the other was grain boundary composition effect. In part I, grain boundary phase was changed by additional heat treatment. The sintering additives were 2wt% $Al_2O_3$, 5wt% $Y_2O_3$, and 1wt% MgO. From x-ray diffraction pattern new crystalline phases were made after heat treatment. The microstructures of two samples were regarded as similar size and shape as a result of image analysis, and phase fraction of β was also almost same as x-ray diffraction. And more hardness and toughness were not so different. So the grain boundary of G named specimen could be thought to have a glassy grain boundary phase and that of C named specimen to have a crystallized grain boundary phase. For these two samples damage mode were compared by Hertzian indentation as a like top view, side view by bonded specimen method, and strength degradation. As a result crystallized C sample made ring crack at lower load than that of G and showed longer cone crack and narrower damage zone. The reason they showed the different damage behavior was from the difference of grain boundary strength. After crystallization of grain boundary it has strong interface so that its grain boundaries can't slide easily. From that reason, C sample could show brittle fracture and less damage tolerance.
In part Ⅱ, grain boundary composition effect was considered. The sintering additives of AY sample was 2wt% $Al_2O_3$ and 6wt% $Y_2O_3$, and that of M sample was 6.24wt% MgO. They were sintered at different temperatures so that they had similar microstructures when they were analyzed by x-ray diffractometer and image analysis. Hardness and toughness are also similar from the result of Vickers indentation. When Hertzian indentation test is conducted AY sample made ring crack at higher load than M sample and showed more wide damage zone. So we could think that AY sample was more damage tolerant than M sample. To distinguish the source of different damage behavior, glasses of grain boundary phase were made referring to phase diagrams of Y-Si-Al-O-N and Mg-Si-O-N. Then thermal expansion coefficients of each glass were measured and from this data residual stresses which were induced between grain and grain boundary could be calculated. Both sample had residual compressive stresses at grain and residual tensile stressed at grain boundary. But for residual stresses of AY sample was larger than M, we could think these different residual stresses could be the source of different damage behavior.
At finally a conclusion could be made that grain boundary phase could be another factor for damage behavior as well as grain morphology.
