1224 Journal of the American Ceramic Society--Leoni et al Vol 91. No 4 200 -200 -200 -400 -600 50100150200250300 50200250300 depth(um depth(um) hrough-thickness residual stress profile in the AZ specimen. Fig, 10. Thr residual stress profile in the AMZ speci in longitudinal mode I for alumina(open dots) and zirconia mode I for alumina(open dots), zirconia The average stress is also shown (stars). The continuous line monds). The average stress is also shown nds to calculated macroscopic interlaminar stresses. The line stars). The corresponds to calculated macroscopic in- ng the points serves as a guide for the eye. terlaminar stresses. The line connecting the points serves as a guide for the eye. Local deviations due to the noise in the experimental data surement conducted in longitudinal mode I(cf. Fig. 2)are ( the X-ray beam travels over a long distance within the material shown in Fig. 10. the laminae are well identified(as suggested n order to reach the deepest buried layers) are evident. More- by the large phase contrast shown in Table D) and a stress profile over, large differences appear between alumina and zirconia can be reconstructed. Differences are again evident between zir- tresses,with a slight tendency to separation into tensile and conia, alumina, and mullite in the composite laminae. compressive components. The weighted average of zirconia and To confirm that problems can come from the cosintering alumina stresses appears to be closer to the design value. This process, microscopy observations were made on the very same result is partly in contradiction with the expectations: by choos- lamina sintered in a homogeneous laminate and in a multilayer ing the reference interplanar spacing, similar trends would be one. As an example, Fig. ll shows scanning electron micro- expected for alumina and zirconia in the az20 and A z40 layer graphs(SEM) of polished cross sections of the AM40 lamina in The observed differences suggest that d is insufficient or unable the monolithic and in the AMz specimens. Differences both to properly account for the interlaminar residual stresses due the grain size and in the size and shape of the residual porosity the sintering process are clear. Even if sintering temperature and lamina composition A similar although less noisy situation appears when the re- are the same (laminae were taken from the same sheet), the en- sidual stress is measured on the same specimen in longitudinal vironment in which sintering occurred is clearly different as in the mode I(Fig. 2): the result is proposed in Fig 9, to be compared AMZ laminate the AM40 is constrained in between heteroge- with Fig 8. Only the region near the surface is shown, to high- neous laminae. This difference, not accounted for by the kreher- light fine details in the stress trends. a difference in the stresses Pompe or analogous models for the composite, can be respon of alumina and zirconia still appears: zirconia is in tension sible for the apparently inconsistent results found in the various hereas alumina is in compression. The difference increases in- laminates. In fact, dr takes into account just the contribution of side the laminate after an increase in the expected compressive the residual stresses created in a composite in the absence of ex- laminate, and in particular sintering stresses in the component over the whole AT. The residual stress that starts to build phases, thus seem to be highly affected by the environment in during the sintering and the cooling phase is probably sufficier which the cosintering occurs. to change the grain interaction mechanism in the composite. alumina-zirconia-mullite AMZ specimen. Results for a mea- k It appears that the magnitude of the average stress in the The observation is further confirmed by the more complex minate can be correctly predicted by existing models, but not Fig. 11. Scanning electron microscopy images of the AM40 composite: (a)sintered in a homogeneous laminate: (b)cosintered in the aMz multilayeredLocal deviations due to the noise in the experimental data (the X-ray beam travels over a long distance within the material in order to reach the deepest buried layers) are evident. More￾over, large differences appear between alumina and zirconia stresses, with a slight tendency to separation into tensile and compressive components. The weighted average of zirconia and alumina stresses appears to be closer to the design value. This result is partly in contradiction with the expectations: by choos￾ing the reference interplanar spacing, similar trends would be expected for alumina and zirconia in the AZ20 and AZ40 layers. The observed differences suggest that dr is insufficient or unable to properly account for the interlaminar residual stresses due to the sintering process. A similar although less noisy situation appears when the re￾sidual stress is measured on the same specimen in longitudinal mode I (Fig. 2): the result is proposed in Fig. 9, to be compared with Fig. 8. Only the region near the surface is shown, to high￾light fine details in the stress trends. A difference in the stresses of alumina and zirconia still appears: zirconia is in tension whereas alumina is in compression. The difference increases in￾side the laminate after an increase in the expected compressive residual stress. Residual stresses within each lamina in a laminate, and in particular sintering stresses in the component phases, thus seem to be highly affected by the environment in which the cosintering occurs. The observation is further confirmed by the more complex alumina–zirconia–mullite AMZ specimen. Results for a mea￾surement conducted in longitudinal mode I (cf. Fig. 2) are shown in Fig. 10: the laminae are well identified (as suggested by the large phase contrast shown in Table I) and a stress profile can be reconstructed. Differences are again evident between zir￾conia, alumina, and mullite in the composite laminae. To confirm that problems can come from the cosintering process, microscopy observations were made on the very same lamina sintered in a homogeneous laminate and in a multilayer one. As an example, Fig. 11 shows scanning electron micro￾graphs (SEM) of polished cross sections of the AM40 lamina in the monolithic and in the AMZ specimens. Differences both in the grain size and in the size and shape of the residual porosity are clear. Even if sintering temperature and lamina composition are the same (laminae were taken from the same sheet), the en￾vironment in which sintering occurred is clearly different as in the AMZ laminate the AM40 is constrained in between heteroge￾neous laminae. This difference, not accounted for by the Kreher– Pompe or analogous models for the composite, can be respon￾sible for the apparently inconsistent results found in the various laminates. In fact, dr takes into account just the contribution of the residual stresses created in a composite in the absence of ex￾ternal residual stresses and assumes the material to be ideally stiff over the whole DT. The residual stress that starts to build up during the sintering and the cooling phase is probably sufficient to change the grain interaction mechanism in the composite. It appears that the magnitude of the average stress in the laminate can be correctly predicted by existing models, but not 0 50 100 150 200 250 300 −800 −600 −400 −200 0 200 400 600 σ (MPa) depth (µm) Fig. 9. Through-thickness residual stress profile in the AZ specimen, measured in longitudinal mode I for alumina (open dots) and zirconia (squares). The average stress is also shown (stars). The continuous line corresponds to calculated macroscopic interlaminar stresses. The line connecting the points serves as a guide for the eye. 0 50 100 150 200 250 300 −800 −600 −400 −200 0 200 σ (MPa) depth (µm) Fig. 10. Through-thickness residual stress profile in the AMZ speci￾men, measured in longitudinal mode I for alumina (open dots), zirconia (squares), and mullite (diamonds). The average stress is also shown (stars). The continuous line corresponds to calculated macroscopic in￾terlaminar stresses. The line connecting the points serves as a guide for the eye. Fig. 11. Scanning electron microscopy images of the AM40 composite: (a) sintered in a homogeneous laminate; (b) cosintered in the AMZ multilayered laminate. 1224 Journal of the American Ceramic Society—Leoni et al. Vol. 91, No. 4