A Morales-Rodriguez et al / Journal of the European Ceramic Society 29(2009 )1625-1630 1.2 (MPa)= Fig3. Creep curve at constant load plotted as log i-8 for laminated com- Fig. 5. SEM micrograph of a multilayer composite deformed at 1500C up to posite deformed under isostrain condition. Several determinations of the stress a final strain of 30% Controlled cavitation occurred in the Al2O3 layers.The stress axis is vertical exponent n by stress changes are shown. structural integrity, indicating an excellent interlayer adhesion rates. The measured stress exponent is close to unity, indicating(Fig. 5). No measurable change in grain size or shape during test a Newtonian creep process. Fig. 2 also shows the o-E curve ing was observed. The ZTA layers appear almost unchanged-a at Eo=2 x 10-s-of high-purity monolithic Al2 O3 with an consequence of its superplastic behavior, as noted above, indi- average grain size d=1.8 um, similar to that of the alumina lay- cating that this phase deforms preferentially by grain boundary ers. The monolith failed with very little plastic deformation after sliding. On the other hand, the Al2O3 layers show extensive attaining a maximum stress omax of 55 MPa. creep damage consisting of cavities distributed homogeneously Fig 3 displays a constant load creep test conducted at 1500C throughout the layer, initiated at two-grain boundaries parallel to showing several load changes for determining the stress expo- the stress direction(Fig. 5). Except for the most severe testing nent n(Eq(I); the corresponding stresses o were calculated conditions(higher stresses or strain rates, and lower tempera using the original cross-sectional area of the sample. Again, tures), no microcrack development by cavity coalescence was observed, thus resulting in a damage-tolerant regime. This dam (characterized by negative slopes in compression constant load age mode has been reported previously in pure-alumina at low stresses, where the final failure occurs by the coalescence of a controlled degradation of the specimen can be observed. creep damage at large strains 15-17 The measured stress exponent is n=1.1tO. 1, equal to that determined from constant strain rate tests(Fig. 2). The acti- 4. Discussion vation energy for flow Q, measured from temperature changes △T=±50° C during testing(Fig.4) by using Eq.(1),was The creep behavior of the laminates deformed in isostrain Q=710+40kJ/mol, regardless of stress level and temperature. conditions is characterized by a stress exponent n and an acti- After testing, the laminates showed barrelling due to friction vation energy g of about 1 and 700kJ/mol, respectively. In effects, although the layer interfaces still retained their initial order to elucidate the creep mechanism operating in the lam inates, the mechanical behavior of the composites must be compared to that of the constituent phases, alumina and ZTA, which exhibit very different creep behavior. Monolithic ZTA shows very large tensile elongations(>500%)and, correspond ingly, low flow stresses. At 1500C, for example, the flow stress of ZTA composites o, with composition and grain size similar to those in the present laminates is below 10 MPa at 8=2 x 10->5-1. ZTa deforms primarily by grain boundary sliding, as in superplastic metals and metallic alloys, charac- terized by a stress exponent n=2 and an activation energy 1450114001450140011450 0=600-750kJ/mol. The lack of differential microstructural features of the ZTA layers in the strained laminates with respect 0=50 MPa to the as-received ones(Fig. 5)agrees with this behavior. On the other hand. monolithic alumina is much more resistant 60 than ZTA, with a marked brittle behavior, 10, I5(Fig. 2). For grain sizes larger than I um, it shows a stress exponent n=l and an Fig. 4. Creep curve plotted as log i-e for a multilayer showing several deter- activation energy @=500-800kJ/mol, with very limited grain tions of the activation energy g by temperature changes. boundary sliding.20-22The large scatter in the reported values ofA. Morales-Rodríguez et al. / Journal of the European Ceramic Society 29 (2009) 1625–1630 1627 Fig. 3. Creep curve at constant load plotted as log ε˙ − ε for laminated com￾posite deformed under isostrain condition. Several determinations of the stress exponent n by stress changes are shown. rates. The measured stress exponent is close to unity, indicating a Newtonian creep process. Fig. 2 also shows the σ − ε curve at ε˙o = 2 × 10−5 s−1 of high-purity monolithic Al2O3 with an average grain size d = 1.8m, similar to that of the alumina lay￾ers. The monolith failed with very little plastic deformation after attaining a maximum stress σmax of 55 MPa. Fig. 3 displays a constant load creep test conducted at 1500 ◦C showing several load changes for determining the stress expo￾nent n (Eq. (1)); the corresponding stresses σ were calculated using the original cross-sectional area of the sample. Again, steady-states regimes were attained after every load change (characterized by negative slopes in compression constant load tests), except in the last section of the creep curve where a controlled degradation of the specimen can be observed. The measured stress exponent is n = 1.1 ± 0.1, equal to that determined from constant strain rate tests (Fig. 2). The acti￾vation energy for flow Q, measured from temperature changes T = ±50 ◦C during testing (Fig. 4) by using Eq. (1), was Q = 710 ± 40 kJ/mol, regardless of stress level and temperature. After testing, the laminates showed barrelling due to friction effects, although the layer interfaces still retained their initial Fig. 4. Creep curve plotted as log ε˙ − ε for a multilayer showing several deter￾minations of the activation energy Q by temperature changes. Fig. 5. SEM micrograph of a multilayer composite deformed at 1500 ◦C up to a final strain of 30%. Controlled cavitation occurred in the Al2O3 layers. The stress axis is vertical. structural integrity, indicating an excellent interlayer adhesion (Fig. 5). No measurable change in grain size or shape during test￾ing was observed. The ZTA layers appear almost unchanged – a consequence of its superplastic behavior, as noted above, indi￾cating that this phase deforms preferentially by grain boundary sliding. On the other hand, the Al2O3 layers show extensive creep damage consisting of cavities distributed homogeneously throughout the layer, initiated at two-grain boundaries parallel to the stress direction (Fig. 5). Except for the most severe testing conditions (higher stresses or strain rates, and lower tempera￾tures), no microcrack development by cavity coalescence was observed, thus resulting in a damage-tolerant regime. This dam￾age mode has been reported previously in pure-alumina at low stresses, where the final failure occurs by the coalescence of creep damage at large strains.15–17 4. Discussion The creep behavior of the laminates deformed in isostrain conditions is characterized by a stress exponent n and an acti￾vation energy Q of about 1 and 700 kJ/mol, respectively. In order to elucidate the creep mechanism operating in the lam￾inates, the mechanical behavior of the composites must be compared to that of the constituent phases, alumina and ZTA, which exhibit very different creep behavior. Monolithic ZTA shows very large tensile elongations (>500%) and, correspond￾ingly, low flow stresses. At 1500 ◦C, for example, the flow stress of ZTA composites18,19 with composition and grain size similar to those in the present laminates is below 10 MPa at ε˙ = 2 × 10−5 s−1. ZTA deforms primarily by grain boundary sliding, as in superplastic metals and metallic alloys, charac￾terized by a stress exponent n = 2 and an activation energy Q = 600–750 kJ/mol.14 The lack of differential microstructural features of the ZTA layers in the strained laminates with respect to the as-received ones (Fig. 5) agrees with this behavior. On the other hand, monolithic alumina is much more resistant than ZTA, with a marked brittle behavior8,10,15 (Fig. 2). For grain sizes larger than 1m, it shows a stress exponent n = 1 and an activation energy Q = 500–800 kJ/mol, with very limited grain boundary sliding.20–22 The large scatter in the reported values of