Journal of the American Ceramic Society-Ma et al. Vol 87. No. 3 value between 800-1000 MPa. The phase proportions in Fig 4 do in the diffraction patterns. There is no such simple explanation of not clarify the matter greatly, other than to also suggest a value why the I-phase fraction determined from the neutron diffraction between 800-1000 MPa. Close examination of the creep data, or patterns does not exactly mirror the m-phase fraction in Fig. 4. It the other hand, shows that there is already significant transforma is possible that the stress-induced orthorhombic phase occurs, but tion occurring at 800 MPa(Fig. 6(a). The reason for these is obscured by the lower counting statistics and broader lines. This differences may well lie in the complexity of the microstructure cannot account for the entire discrepancy, and we must conclude and the residual stresses induced by processing. Earlier work on that intensity in some of the c- and 8-phase peaks has beer he role of the 8 phase in Mg-PSZ highlighted TEM strain contours mistakenly attributed to the t phase during Rietveld refinement. A ithin the tetragonal precipitates adjacent to 8-phase precipitates slight linear decline in the sum of the c and 8 fractions support and neutron diffraction line broadening supported this observa- this explanation. Unlike the earlier work on Ce-TZP and Y-TZP tion. Figure 5(b) shows that even before the application of the where only two phases were involved, the severe peak overlap in external stress, there is a considerable amount of residual strain, these patterns places some limitations on the analysis 0, 22% rms. Conversion to a residual rms microstress yields a Very similar in situ neutron diffraction studies of mechanical value of 440 MPa. Therefore, even in the absence of an externally response have now been conducted on three zirconia ceramics applied stress. some small fraction of the tetragonal precipitates Ce-TZP, Y-TZP, and Mg-PSZ. in this work. Despite some already lie close to the transformation stress. These local commonality in the general perception of these ceramics and their internal stresses, when supplemented by the externally applied behavior, each one has shown quite different behavior wher ess, can trigger the transformation at far lower stresses than the tudied in this manner. All three showed elastic deformation up to gross critical stress of -1000 MPa. In support of this idea, the low ome critical stress, beyond which some plastic deformation tress range of Figs. 4(a) and (b), from 0-800 MPa, appear to occurred, Table Ill summarizes the critical stresses and plastic show a small slope albeit of only approximately one standard strains observed in the three ceramics deviation, A pragmatic definition of the critical stress is that stress It is clear from Table III that the I phase in 3Y-TZP is the required to trigger the I-m transformation in I phase that least transformable by uniaxial compression and that in 9 4Mg experiences only the average long-range stress field of the ceramic PSZ it is the most transformable. Not only do the numerical body. Both an interpolating curve and linear extrapolation of the values differ, but the mechanisms of the deformation also differ last three points above and below the elastic-plastic transition on widely Mg-PSZ deformed largely by the r-m transformation Fig.4 gave o。=925±20MPa but with a strong time dependence. Ce-TZP deformed primarily This critical stress is appropriate only for uniaxial compression by the f-m transformation but with some associated ferro- and for the long-term loading used here. The majority of the 1- elasticity. The t-m transformation in Ce-TZP was mostly in m transformation in this material occurred slowly by creep, as quite rapid bursts followed by much smaller amounts of previously highlighted by Finlayson et al. Hence, mechanical time-dependent transformation or creep. No r-m transforma ests conducted at more usual laboratory strain rates would give a tion was observed in 3Y-TZP, and the plastic deformation in much higher critical stress and much smaller plastic strains In uniaxial compression was attributed to ferroelasticity alone. In contrast, previous creep testing was conducted in tension and a all cases, there would appear to be a tendency to restore the significant amount of creep was observed at stresses as low as 240 original microstructure(and shape )of the material, albeit with MPa. Unlike the earlier creep tests, where samples were taken widely differing time constants. Mg-PSZ shows evidence of stepwise. Hence, each creep curve represents the transformation of several hours and the 3Y-TZP studied had completely relaxed a different subset of r crystals. There appear to be two types of within 3 h of removal of the load behavior. At 800-1000 MPa, the creep is relatively slow and the It is interesting to consider the chemical and microstructural terminal slope of the curves is low. Above this level, i.e., onditions that make these disparate responses possible. The ceramics ignificantly above the estimated bulk critical stress, the slope is studied are stabilized quite differently. The interplay of stabilizer and much greater and the curves lie almost on top of one another. The transformation behavior is not known. In an earlier publication, we last curve, at 1225 MPa, again has a lower slope(see Fig. 6(b)) highlighted that the shear strains (measured using stress-induced shifts This may be caused by the smaller stress increment of 25 MPa n neutron diffraction peak positions)in Ce-TZP are greater than those nstead of 50 MPa. or it may be that the supply of read in Y-TZP. It is possible that this is caused by stabilizer-induced ransformed t crystallites under uniaxial compression in this softening of the elastic shear constants in the former. No comparable direction is declining In the absence of results at a range of data could be extracted for Mg-PSZ because of the microstructural temperatures and in view of the very complex microstructure, little complexity. The behavior of these three ceramics is so different that additional information concerning the creep mechanism can be the influence of the stabilizer on the I phase cannot be solely derived from the fitted creep constants responsible. In systems other than zirconia(e.g, in shape-memory It is interesting to note that the transverse plastic strains are five alloys), stress-induced Martensitic-phase transformations have first times as great as the longitudinal strains, owing to constraint of the been studied extensively in the temperature domain. In Ca-PSZ and Ce-TZP, the t- m transformation is readily induced by cooling transversely. The I- m transformation is triggered by shear below room temperature: however, in 3Y-TZP. the ceramic remains stresses,and so the orientation of individual t particles with in its original majority I-phase state, and in Mg-PSZ. the respect to the stress(applied and internal) is a critical factor. The phase transforms to an orthorhombic (o) phase on cooling. The t-m initial March coefficient indicates a quasi-random distribution for transformation is driven by shear stresses. it is unknown whether the t phase. Only a subset of the I phase will be oriented for the ferroelasticity is driven by shear or dilatational stresses, and the t- maximum shear stress to lie along a potential transformation o transformation is thought to be driven by dilatational stresses. A direction. Therefore, a degree of transformation selectivity is exercised. This is confirmed by the high degree of preferred orientation in the m phase: a clear case of transformation-induced Table lll. Critical Stresses and Plastic texture as predicted by Bowman and Chen. 42 From the limited Strains Observed for Zirconia Ceramics in pole data that may be inferred from these results, the preferred Uniaxial Compression orientation observed in this work agrees well with the calculated pole figures of Bowman and Chen. 12 Preferred orientation may Critieal Plastic Strain (9) influence the small discrepancy between the m-phase fractio derived from the macroscopic volume change and the neutron diffraction phase analysis at high stress. Then, as the stress 0.1330.589 changes, e. g. during unloading, some of the m phase may undergo 12Ce-TZP 1200 0.14 0.375 ferroelastic reorientation into orientations more readily observable 3Y-TZP 70 0.1160.100