Tape-cast alumina-circonia laminates: processing and mechanical properties and 9a). This suggests that tape casting leads to tip healing processes(blunting)at relatively low a more homogeneous green state, with small com- testing rates. In conjunction, the work of fracture paction defects becomes larger and larger as the testing rate The fracture resistance (o) of both series decreases increases with the zirconia content(Fig. 5). A fur Works of fracture(WOF)of 1 9 x 104, 2. X ther improvemcnt is achicved by combining AZIo 10+ and 40x 104 J were dctcrmincd by comput ued AZ5 layers, probably due to interfacial ing the area below the load-deflection curves cor resses inl this latter case. Curiously, material lls responding to cross-head speeds of 0.1, 0.01 from the second series show lower fracture resis- 0-001 mm min, respectively. An estimation of tances than pure alumina( Fig 5b). The propagat- the intrinsic fracture surface energy (yi) can be ing crack is deflected several times in or near layer obtained from the griffith/Irwin similarity rela interfaces for composite materials(Fig. 6) tionship:y =2E. From our experimental data y Results of the Cn tests are depicted in Fig. 7. equals 10-7, 19 and 25.9 J m for cross-head Figure 7a shows the load/deflection curves obtained speeds of 0.1, 0-01 and 0-00l mm min, respec on the A, A/A and AZ1O/AZl0 with a deflection tively. This very simple calculation shows that rate of 0 01 mm min- together with results work of fracture is significantly higher than 2y S obtained on the aziO/AZ10 with different deflec-. where s is the minimum fracture surface area as tion rates. A and A/a curves are almost superim- defined by the chevron geometry(S= 4. 7 mm) posed; these materials exhibit unstable crack This would support either the occurrence of vis propagation. A/A fractured specimens show a fat coplastic energy dissipation, i. e. Effective >yi and/or fracture path, essentially transgranular, whereas crack deflection(Sermective S) AZIO/AZ10 specimens fractured with a mixed transgranular/ intergranular path giving rise to a rough fracture surface, as shown in Fig. 8. Stable or semi-stable crack propagations were observed in this latter case, with a significant influence of the cross-head speed on the maximum load E peak, i.e. the lower the displacement rate, the a 04Z1 higher the fracture toughness. This behaviour is tentatively attributed to the occurrence of crack FRACTURE PATH AND FRACTURE SURFACE Deflection'u’(mm) 2 Chevron notch tests performed in bending on(a)spec- from the second series with a cross-head Fig. 6. Fracture path in a 2/2/2/10/2/2/2 specimen fractured in speed of nin and two layers to crack front orienta- ns in the case of the 2/2/2/10/2/2/2 gradTape-cast alumina-zirconia laminates: processing and mechanical properties 305 and 9a). This suggests that tape casting leads to a more homogeneous green state, with small com￾paction defects. The fracture resistance (a,) of both series increases with the zirconia content (Fig. 5). A fur￾ther improvement is achieved by combining AZ10 and AZ5 layers, probably due to interfacial stresses in this latter case. Curiously, materials from the second series show lower fracture resis￾tances than pure alumina (Fig. 5b). The propagat￾ing crack is deflected several times in or near layer interfaces for composite materials (Fig. 6). Results of the CN tests are depicted in Fig. 7. Figure 7a shows the load/deflection curves obtained on the A, A/A and AZlO/AZlO with a deflection rate of 0.01 mm min.’ together with results obtained on the AZlOiAZlO with different deflec￾tion rates. A and A/A curves are almost superim￾posed; these materials exhibit unstable crack propagation. A/A fractured specimens show a flat fracture path, essentially transgranular, whereas AZlO/AZlO specimens fractured with a mixed tramgranular/ intergranular path giving rise to a rough fracture surface, as shown in Fig. 8. Stable or semi-stable crack propagations were observed in this latter case, with a significant influence of the cross-head speed on the maximum load peak, i.e. the lower the displacement rate, the higher the fracture toughness. This behaviour is tentatively attributed to the occurrence of crack FRACTURE PATH AND FRACTURE SURFACE Fig. 6. Fracture path in a 2/212/10/2/2/2 specimen fractured in bending. tip healing processes (blunting) at relatively low testing rates. In conjunction, the work of fracture becomes larger and larger as the testing rate decreases. Works of fracture (WOF) of 1.9 X lOA, 2.2 X lOA and 4-O x lOA J were determined by comput￾ing the area below the load-deflection curves cor￾responding to cross-head speeds of 0.1, 0.01 and 0.001 mm min-‘, respectively. An estimation of the intrinsic fracture surface energy (7;) can be obtained from the Griffith/Irwin similarity rela￾tionship: yi = s+ From our experimental data yi equals 10.7, 19 and 25.9 J mm2 for cross-head speeds of 0.1, 0.01 and 0.001 mm min’, respec￾tively. This very simple calculation shows that work of fracture is significantly higher than 2yiS where S is the minimum fracture surface area as defined by the chevron geometry (S = 4.7 mm’). This would support either the occurrence of vis￾coplastic energy dissipation, i.e. ?/e~~tive >yi and/or crack deflection (Seffectiue > S). 32 "0 0 02 0 04 0.06 0 08 O.! 0.!2 0.14 0.16 Deflection ‘u’ (mm) (a) 32 a 4 0 I,, i 0 0.01 0.02 0.03 0.04 0.05 0.06 Deflection ‘u’ (mm) @I Fig. 7. Chevron notch tests performed in bending on (a) spec￾imens of the AZlOiAZlO grade with different loading rates and (b) specimens from the second series with a cross-head speed of 10 pm min ’ and two layers to crack front orienta￾tions in the case of the 21212/10121212 grade