《复合材料 Composites》课程教学资源（学习资料）第五章 陶瓷基复合材料_Layered materials with high strength and flaw tolerance based on alumina and aluminium titanate

Availableonlineatwww.sciencedirectcom ScienceDirect E噩≈RS ELSEVIER Joumal of the European Ceramic Society 27(2007)1455-1462 www.elsevier.comlocate/jeurceramsoc Layered materials with high strength and flaw tolerance based on alumina and aluminium titanate Salvador bueno. Carmen baudin Instituto de Ceramica y Vidrio, CSIC, Campus de Cantoblanco, C/Kelsen 5, 28049 Madrid, spain Available online g June 2006 Laminates in which high strength external layers and flaw tolerant internal layers with similar compositions are mechanical behaviour in relation to monolithic materials with the same composition as the layers. The limitation of n in which no residual stresses are present, is the difficulty in co-sintering layers with large microstructural differences in the green state. escribes a new method to obtain laminates constituted by layers with large differences in terms of grain size starting from green bodies microstructures. The approach is based on the effect of small amounts of titania as agents for alumina grain growth enhancement. Starting from fine grained green bodies that combined alumina layers with composite layers made of mixtures of alumina and titania, additional"in situ "formed layers constituted by large (20-30 um) alumina grains were found after sintering contiguous to the composite layers. The thickness of the"in situ"formed layers reached up to 200 um, depending on the thermal treatment (1450-1550C). The fracture behaviour of the laminates and the monoliths was studied, using stable Single Edge V Notched Beam(SEVNB)tests, in terms of work of fracture and the critical stress intensity factor in mode l, Kic. The large grain sized alumina layers reinforced the laminates by crack branching and bridging 2006 Elsevier Ltd. All rights reserved. Keywords: Laminates; Al2O3; Al2TiOs: Grain size; Toughness and toughening Introduction as a way to overcome the low strength values of the flaw tolerant alumina(Al2O3)-aluminium titanate(Al2 TiOs)composites 4-7 Alumina materials are widely used in applications where Major limitation is the presence of tensile residual stresses in hardness, wear and/or chemical resistance are required but tra- the external layers since the high strength compositions in this lications as structural components have been system usually present larger thermal expansions and Youngs limited due to the lack of reliability associated to the brittle frac- modulus than the flaw tolerant ones. 5,6 The combination of ture mode Structures found in nature such as biological hard homogeneous external layers with highly heterogeneous layers tissues, shells and teeth are made of layered architectures com- with similar composition has been proposed as means to avoid bining materials with different properties that lead to laminates the development of significant residual stresses. The limit of with mechanical behaviour superior than that of the individual this approach is the difficulty that involves the co-sintering of constituents.-3In this sense, much research is being devoted layers with such microstructural differences. One solution is the to the development of laminates to improve the performance of fabrication of graded materials in which transitional microstruc- brittle materials. Laminates emerge as a new strategy to achieve tures are tailored between both surfaces of the samples through faw tolerance"in opposition to the traditional"flaw elimina- a green processing in several steps, as it allows reaching specific tion"approach of monolithic ceramics. surface properties different than those of the bulk.8.9 In this work, a way to obtain laminates with large microstru strength external layers and internal flaw tolerant layers are tural differences between contiguous layers, based on the effect combined might provide fracture resistance keeping the high of small amounts of titania(TiO2) as agent for alumina grain strength of the surface layers. This approach has been proposed growth enhancement,,I is analysed. The designed structure constituted of high strength external layers of small grain sized alumina combined with flaw tolerant internal layers.(Fig. 1) Corresponding author. Tel: +34917 355 840: fax: +34917 355 843 In the green state, alumina layers are combined with compos- E-mail address: cbaudin @icv csices(C. Baudin) ite layers made of mixtures of alumina and titania The effect 0955-2219/S-see front matter o 2006 Elsevier Ltd. All rights reserved. doi: 10. 1016/j-jeurceramsoc. 2006.05.054

Journal of the European Ceramic Society 27 (2007) 1455–1462 Layered materials with high strength and flaw tolerance based on alumina and aluminium titanate Salvador Bueno, Carmen Baud´ın ∗ Instituto de Cer ´amica y Vidrio, CSIC, Campus de Cantoblanco, C/Kelsen 5, 28049 Madrid, Spain Available online 9 June 2006 Abstract Laminates in which high strength external layers and flaw tolerant internal layers with similar compositions are combined, can provide improved mechanical behaviour in relation to monolithic materials with the same composition as the layers. The limitation of this design, in which no residual stresses are present, is the difficulty in co-sintering layers with large microstructural differences in the green state. This work describes a new method to obtain laminates constituted by layers with large differences in terms of grain size starting from green bodies with similar microstructures. The approach is based on the effect of small amounts of titania as agents for alumina grain growth enhancement. Starting from fine grained green bodies that combined alumina layers with composite layers made of mixtures of alumina and titania, additional “in situ” formed layers constituted by large (∼=20–30m) alumina grains were found after sintering contiguous to the composite layers. The thickness of the “in situ” formed layers reached up to 200m, depending on the thermal treatment (1450–1550 ◦C). The fracture behaviour of the laminates and the monoliths was studied, using stable Single Edge V Notched Beam (SEVNB) tests, in terms of work of fracture and the critical stress intensity factor in mode I, KIC. The large grain sized alumina layers reinforced the laminates by crack branching and bridging. © 2006 Elsevier Ltd. All rights reserved. Keywords: Laminates; Al2O3; Al2TiO5; Grain size; Toughness and toughening 1. Introduction Alumina materials are widely used in applications where hardness, wear and/or chemical resistance are required but tra￾ditionally the applications as structural components have been limited due to the lack of reliability associated to the brittle frac￾ture mode. Structures found in nature such as biological hard tissues, shells and teeth are made of layered architectures com￾bining materials with different properties that lead to laminates with mechanical behaviour superior than that of the individual constituents.1–3 In this sense, much research is being devoted to the development of laminates to improve the performance of brittle materials. Laminates emerge as a new strategy to achieve “flaw tolerance” in opposition to the traditional “flaw elimina￾tion” approach of monolithic ceramics. In particular, laminated structures where alternating high￾strength external layers and internal flaw tolerant layers are combined might provide fracture resistance keeping the high strength of the surface layers. This approach has been proposed ∗ Corresponding author. Tel.: +34 917 355 840; fax: +34 917 355 843. E-mail address: cbaudin@icv.csic.es (C. Baud´ın). as a way to overcome the low strength values of the flaw tolerant alumina (Al2O3)–aluminium titanate (Al2TiO5) composites.4–7 Major limitation is the presence of tensile residual stresses in the external layers since the high strength compositions in this system usually present larger thermal expansions and Young’s modulus than the flaw tolerant ones.5,6 The combination of homogeneous external layers with highly heterogeneous layers with similar composition has been proposed as means to avoid the development of significant residual stresses.4 The limit of this approach is the difficulty that involves the co-sintering of layers with such microstructural differences. One solution is the fabrication of graded materials in which transitional microstruc￾tures are tailored between both surfaces of the samples through a green processing in several steps, as it allows reaching specific surface properties different than those of the bulk.8,9 In this work, a way to obtain laminates with large microstruc￾tural differences between contiguous layers, based on the effect of small amounts of titania (TiO2) as agent for alumina grain growth enhancement,10,11 is analysed. The designed structure is constituted of high strength external layers of small grain sized alumina combined with flaw tolerant internal layers12,13 (Fig. 1). In the green state, alumina layers are combined with compos￾ite layers made of mixtures of alumina and titania. The effect 0955-2219/$ – see front matter © 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.jeurceramsoc.2006.05.054

S. Bueno, C. Baudin/Journal of the European Ceramic Sociery 27(2007)1455-1462 whole thermal cycle, it would be more extensive at high temper- atures once initial co-sintering of the layers has taken place, so the thickness of the large grain sized alumina layer formed"in situ"could be controlled through the adequate selection of the thermal treatment In order to evaluate quantitatively the effect of the"in situ formed layers, toughness and work of fracture of the laminates and of monoliths of the same compositions as those of the ini- Fig 1. Schematic illustration of the designed five layered laminated structure tial layers have been compared. The work of fracture, ywoF, is lumina layers are represented with grey colour and thin alumina+ 10 vol. defined by the mean external work which is consumed to pro- aluminium titanate composite layers are represented with white colour. Dashed duce a unit of fracture surface area during quasi-static failure lines indicate the zones where the development of large grained alumina layers and is determined experimentally from the total area under the due to titania diffusion occurred. Two different sizes of notches are shown with load-load point displacement curve in stable tests. This param- eter provides significant toughness values because no spare energy is involved in the test and, therefore, the whole energy of titania leads to interlayers of large grain sized alumina to be given to the system is employed in creating new surfaces. Nev- formed between the alumina and the composite layers(Fig. 1). ertheless, the difficulty to get stable tests in brittle materials has As the microstructural heterogeneity is developed during sin- usually restricted the use of work of fracture to the characteriza tering, no decohesion of the layers due to differential sintering tion of the reinforcement in materials with non-linear behaviour occurs. Although titania diffusion might take place during the where numerous energy consuming processes can occur during 5 um 5 um uctures of the studied monolithic materials. Alumina grains appear with dark grey colour and aluminium titanate of an intermediate grey shade. FE-SEM (a) Alumina sintered at (b)Alumina+10 vol. aluminium titanate sintered at 1450C, A1OAT-1450 (c) Alumina sintered at 1550C, A-1550. (d) Alumina+10 vol %o aluminium titanate sintered at 1550C, A10AT-1550

1456 S. Bueno, C. Baud´ın / Journal of the European Ceramic Society 27 (2007) 1455–1462 Fig. 1. Schematic illustration of the designed five layered laminated structure showing a bend bar and the notch orientation with respect to the layers. Thick alumina layers are represented with grey colour and thin alumina + 10 vol.% aluminium titanate composite layers are represented with white colour. Dashed lines indicate the zones where the development of large grained alumina layers due to titania diffusion occurred. Two different sizes of notches are shown with a continuous and a dotted line. of titania leads to interlayers of large grain sized alumina to be formed between the alumina and the composite layers (Fig. 1). As the microstructural heterogeneity is developed during sin￾tering, no decohesion of the layers due to differential sintering occurs. Although titania diffusion might take place during the whole thermal cycle, it would be more extensive at high temper￾atures once initial co-sintering of the layers has taken place, so the thickness of the large grain sized alumina layer formed “in situ” could be controlled through the adequate selection of the thermal treatment. In order to evaluate quantitatively the effect of the “in situ” formed layers, toughness and work of fracture of the laminates and of monoliths of the same compositions as those of the ini￾tial layers have been compared. The work of fracture, γWOF, is defined by the mean external work which is consumed to pro￾duce a unit of fracture surface area during quasi-static failure and is determined experimentally from the total area under the load–load point displacement curve in stable tests. This param￾eter provides significant toughness values because no spare energy is involved in the test and, therefore, the whole energy given to the system is employed in creating new surfaces. Nev￾ertheless, the difficulty to get stable tests in brittle materials has usually restricted the use of work of fracture to the characteriza￾tion of the reinforcement in materials with non-linear behaviour where numerous energy consuming processes can occur during Fig. 2. Characteristic microstructures of the studied monolithic materials. Alumina grains appear with dark grey colour and aluminium titanate of an intermediate grey shade. FE-SEM micrographs of polished and thermally etched surfaces: (a) Alumina sintered at 1450 ◦C, A-1450. (b) Alumina + 10 vol.% aluminium titanate sintered at 1450 ◦C, A10AT-1450. (c) Alumina sintered at 1550 ◦C, A-1550. (d) Alumina + 10 vol.% aluminium titanate sintered at 1550 ◦C, A10AT-1550

S. Bueno, C. Baudin Journal of the European Ceramic Sociery 27(2007)1455-1462 1457 h erties of the monolithic materials (G: grain size, Kic: fracture toughness, ywoF: work of fracture, A: alumina, AT: aluminium titanate) GA (S.D. )(um) GAT (S.D. )(um) KiC(SD)(MPam) ywoF(S D Om) All 3.2(0.4) 3.5(0.1) 34.7(1.3) A-1550 5.50 3.0(0.3) 04(28) A10AT1550 3.9(0.3) 24(02) 3.3(0.1) 40.6(1.2) S D. standard deviation fracture. The energy principle is based on macroscopic thermo- layers(Fig. 1)were manufactured by a colloidal route from lynamics and does not require any assumptions regarding the aqueous AlO3 and TiO2 suspensions using the optimum green onstitutive equation of the cracked body for discussing crack- processing conditions previously established .>The starting growth problems. This feature enables the application of the materials were commercial a-Al2O3( Condea, HPAO5, USA) energy principle to characterize non-linear deformation and frac- and TiO2-anatase(Merck, 808, Germany)powders. The single ture behaviours as well as oriented structures such as laminates. oxide(Al2O3)and the mixture(Al2O3/TiO2)were dispersed in deionised water by adding 0.5 wt %(on a dry solid basis)of a 2. Experimental carbonic acid based polyelectrolyte(Dolapix CE64, Zschimmer- Schwarz, Germany). Suspensions were prepared to a solid load- Monoliths of monophase alumina+ ing of 50 vol. and ball milled with Al2O3 jar and balls during 10 vol% aluminium titanate(A ered struc ture combining two external and Plates of monolithic and laminated materials with two internal alumina+10 vol. aluminium titanate composite 70 mm x 70 mm x 10 mm dimensions were obtained by slip 0.1 0.05 300300 d lum] Fig. 3. Characteristic microstructures of the laminated materials. FE-SEM micrographs of polished and chemically (a and b)or thermally (c and d)etched surfaces (a) Laminate sintered at 1450 C with the composite layer(intermediate grey) in the centre of the micrograph, two large grained alumina layers at both sides(clearest grey)and part of the fine grained alumina layers( dark grey ) The profile of TiO2 content in the alumina layers determined by WDS analysis is show (b) Laminate sintered at 1550'C with the composite layer(intermediate grey)in the centre of the micrograph, two large grained alumina layers at both sides(clearest grey)and part of the fine grained alumina layers(dark grey). The profile of TiO2 content in the alumina layers determined by WDS analysis is shown. )Detail of interface between a composite layer(right)and the contiguous relatively large grained alumina layer(left)in the laminated sintered at 1450C (d) Detail of interface between a composite layer(right) and the contiguous relatively large grained alumina layer(left) in the laminated sintered at 1550oC

S. Bueno, C. Baud´ın / Journal of the European Ceramic Society 27 (2007) 1455–1462 1457 Table 1 Properties of the monolithic materials (G: grain size, KIC: fracture toughness, γWOF: work of fracture, A: alumina, AT: aluminium titanate) GA (S.D.) (m) GAT (S.D.) (m) KIC (S.D.) (MPa m1/2) γWOF (S.D.) (J m−2) A-1450 3.5 (0.3) – 2.9 (0.2) 14.7 (1.9) A10AT-1450 3.2 (0.4) 2.2 (0.1) 3.5 (0.1) 34.7 (1.3) A-1550 5.5 (0.5) – 3.0 (0.3) 20.4 (2.8) A10AT-1550 3.9 (0.3) 2.4 (0.2) 3.3 (0.1) 40.6 (1.2) S.D.: standard deviation. fracture. The energy principle is based on macroscopic thermo￾dynamics and does not require any assumptions regarding the constitutive equation of the cracked body for discussing crack￾growth problems.14 This feature enables the application of the energy principle to characterize non-linear deformation and frac￾ture behaviours as well as oriented structures such as laminates. 2. Experimental Monoliths of monophase alumina (A) and alumina + 10 vol.% aluminium titanate (A10AT), and one layered struc￾ture combining two external and one central alumina layers with two internal alumina + 10 vol.% aluminium titanate composite layers (Fig. 1) were manufactured by a colloidal route from aqueous Al2O3 and TiO2 suspensions using the optimum green processing conditions previously established.5,15 The starting materials were commercial -Al2O3 (Condea, HPA05, USA) and TiO2–anatase (Merck, 808, Germany) powders. The single oxide (Al2O3) and the mixture (Al2O3/TiO2) were dispersed in deionised water by adding 0.5 wt.% (on a dry solid basis) of a carbonic acid based polyelectrolyte (Dolapix CE64, Zschimmer￾Schwarz, Germany). Suspensions were prepared to a solid load￾ing of 50 vol.% and ball milled with Al2O3 jar and balls during 4 h. Plates of monolithic and laminated materials with 70 mm × 70 mm × 10 mm dimensions were obtained by slip Fig. 3. Characteristic microstructures of the laminated materials. FE-SEM micrographs of polished and chemically (a and b) or thermally (c and d) etched surfaces: (a) Laminate sintered at 1450 ◦C with the composite layer (intermediate grey) in the centre of the micrograph, two large grained alumina layers at both sides (clearest grey) and part of the fine grained alumina layers (dark grey). The profile of TiO2 content in the alumina layers determined by WDS analysis is shown. (b) Laminate sintered at 1550 ◦C with the composite layer (intermediate grey) in the centre of the micrograph, two large grained alumina layers at both sides (clearest grey) and part of the fine grained alumina layers (dark grey). The profile of TiO2 content in the alumina layers determined by WDS analysis is shown. (c) Detail of interface between a composite layer (right) and the contiguous relatively large grained alumina layer (left) in the laminated sintered at 1450 ◦C. (d) Detail of interface between a composite layer (right) and the contiguous relatively large grained alumina layer (left) in the laminated sintered at 1550 ◦C.

S. Bueno, C. Baudin/Journal of the European Ceramic Sociery 27(2007)1455-1462 electrical box furnace(Termiber, Spain) at heating and cooling rates of 2 Cmin- with 4h dwell at 1200C during heating a/a=0.8 and two different treatments at the maximum temperature: 2h dwell at 1450C and 3 h dwell at 1550C For all tests, samples were diamond machined from the sintered blocks Microstructural characterization was performed by field 80 emissIon scanning electron microscopy(FE-SEM: Hitachi, S- 4700, Japan)on polished and thermally etched (20C below the sintering temperature during I min) or chemically etched (HF 10 vol %, I min) surfaces. The average grain size was deter mined by the linear intercept method considering at least 200 The profiles of titania in the laminates were determined by wavelength dispersive X-ray spectrometer, WDS JEOL, Super- Displacement [mm] probe JXA-8900M, Japan), on polished cross-surfaces of the laminate, operating at 15 kv, 20nA and 10s in the peak posi tion. The k-factors in the quan --a/Wa=0.4 the atomic number-absorption-fluorescence(ZAF)correction. a/Wa=0,8 The analysis was made along three straight lines perpendicular to the layers, taking spots with 5 um diameter, and the average of the three determinations was associated to each corresponding localization across the polished surface of the specimen 乙 Single Edge V Notched Beams (SEVNB)of 4 mm x 6mm x 50 mm of the monoliths and the laminates were tested in a three point bending device using a span of 40 mm and a cross- head speed of 0.005 mm min(Microtest, Spain). The notches were initially cut with a 150 um wide diamond wheel to a depth of about 70% of the final depth. Using this slot as a guide, the remaining part of the notch was done with a razor blade sprin- kled successively with diamond paste of 6 and 1 um. The depth Displacement [mm] of the notches, a, was approximately 0.5 of the thickness of the Fig 4. Characteristic load-displacement curves of notched samples of the lam- monolithic samples(W)and 0. 14 and 0. 28 of the thickness of the nates with a relationship between the notch depth and the height of the sample laminated samples, which resulted in a relation a/WA =0.4 and of 0.14 and 0.29 (corresponding to relative ratios of 0.4 and 0.8 of the width of 0.8, respectively, for the width of the external alumina layer, WA. the extemal alumina layer, respectively) in the laminated samples(Fig. 1). The tip radii of the notches (a) Laminates sintered at 1450C. Unstable test for notch of 0.4 is shown. (b)Laminates sintered at 1550C were optically observed to check that they were below 30 um The curves load versus displacement of the loading frame were recorded during three tests for each material. Fracture tough- casting, removed from the moulds and dried in air at room ness, KIC, values were calculated according to a general stress temperature for at least 24 h. The layered plates, constituted by intensity formulation valid for any crack length6 and the work five layers, with thick external and central layers of alumina of fracture ywoF, values were obtained by dividing the area (1300 um) and two thin intermediate layers of the composite under the stable load-displacement curves by twice the area of (300 um), were fabricated by alternately casting the suspen- the unnotched part of the cross-section of the samples. FE-SEM sions. Casting times were fixed to reach the desired layer thick- was performed on the fracture surfaces ness considering the casting kinetics and sintering shrinkage of each composition. The dried blocks were sintered in air in an optical microscopy(H-Pl, Zeiss, Germany)of the polished ioy The crack path in the notched samples was characterized eral faces Table 2 Properties of the laminates(KiC: fracture toughness, ywoF: work of fracture) 3. Results KIC(SD)(MPam) ywoF(SD)(m-2) Characteristic microstructures of the monolithic materials are shown in Fig. 2 and the microstructural parameters are sum 3.1(0.1)3.1(0.2) 23.7(1.1) marised in Table I together with the values of fracture toughness 3.5(0.2)2.9(0.1) 47.6(1.2) 32.5(1.3) and work of fracture. The alumina materials(Fig. 2a and c)pr 0.4 and 0.8 refer to the notch depth relative to the width of the first alumina layer sented different levels of grain growth(Table 1), the microstruc (a/WA). S.D. standard deviation. ture of the alumina sintered at the highest temperature(1550( Unstable tests Fig. 2c)being bimodal with some grains larger than 10 um and

1458 S. Bueno, C. Baud´ın / Journal of the European Ceramic Society 27 (2007) 1455–1462 Fig. 4. Characteristic load–displacement curves of notched samples of the lam￾inates with a relationship between the notch depth and the height of the sample of 0.14 and 0.29 (corresponding to relative ratios of 0.4 and 0.8 of the width of the external alumina layer, respectively): (a) Laminates sintered at 1450 ◦C. Unstable test for notch of 0.4 is shown. (b) Laminates sintered at 1550 ◦C. casting, removed from the moulds and dried in air at room temperature for at least 24 h. The layered plates, constituted by five layers, with thick external and central layers of alumina (1300m) and two thin intermediate layers of the composite (300m), were fabricated by alternately casting the suspen￾sions. Casting times were fixed to reach the desired layer thick￾ness considering the casting kinetics and sintering shrinkage of each composition.5 The dried blocks were sintered in air in an Table 2 Properties of the laminates (KIC: fracture toughness, γWOF: work of fracture) KIC (S.D.) (MPa m1/2) γWOF (S.D.) (J m−2) 0.4 0.8 0.4 0.8 1450 3.1 (0.1)a 3.1 (0.2) –a 23.7 (1.1) 1550 3.5 (0.2) 2.9 (0.1) 47.6 (1.2) 32.5 (1.3) 0.4 and 0.8 refer to the notch depth relative to the width of the first alumina layer (a/WA). S.D.: standard deviation. a Unstable tests. electrical box furnace (Termiber, Spain) at heating and cooling rates of 2 ◦C min−1, with 4 h dwell at 1200 ◦C during heating and two different treatments at the maximum temperature: 2 h dwell at 1450 ◦C and 3 h dwell at 1550 ◦C. For all tests, samples were diamond machined from the sintered blocks. Microstructural characterization was performed by field emission scanning electron microscopy (FE-SEM; Hitachi, S- 4700, Japan) on polished and thermally etched (20 ◦C below the sintering temperature during 1 min) or chemically etched (HF 10 vol.%, 1 min) surfaces. The average grain size was deter￾mined by the linear intercept method considering at least 200 grains for each phase. The profiles of titania in the laminates were determined by wavelength dispersive X-ray spectrometer, WDS (JEOL, Super￾probe JXA-8900M, Japan), on polished cross-surfaces of the laminate, operating at 15 kV, 20 nA and 10 s in the peak posi￾tion. The k-factors in the quantification were calculated using the atomic number–absorption–fluorescence (ZAF) correction. The analysis was made along three straight lines perpendicular to the layers, taking spots with 5m diameter, and the average of the three determinations was associated to each corresponding localization across the polished surface of the specimen. Single Edge V Notched Beams (SEVNB) of 4 mm × 6 mm × 50 mm of the monoliths and the laminates were tested in a three point bending device using a span of 40 mm and a cross￾head speed of 0.005 mm min−1 (Microtest, Spain). The notches were initially cut with a 150 m wide diamond wheel to a depth of about 70% of the final depth. Using this slot as a guide, the remaining part of the notch was done with a razor blade sprin￾kled successively with diamond paste of 6 and 1 m. The depth of the notches, a, was approximately 0.5 of the thickness of the monolithic samples (W) and 0.14 and 0.28 of the thickness of the laminated samples, which resulted in a relation a/WA ∼= 0.4 and 0.8, respectively, for the width of the external alumina layer, WA, in the laminated samples (Fig. 1). The tip radii of the notches were optically observed to check that they were below 30m. The curves load versus displacement of the loading frame were recorded during three tests for each material. Fracture tough￾ness, KIC, values were calculated according to a general stress intensity formulation valid for any crack length16 and the work of fracture, γWOF, values were obtained by dividing the area under the stable load–displacement curves by twice the area of the unnotched part of the cross-section of the samples. FE-SEM was performed on the fracture surfaces. The crack path in the notched samples was characterized by optical microscopy (H-P1, Zeiss, Germany) of the polished lat￾eral faces. 3. Results Characteristic microstructures of the monolithic materials are shown in Fig. 2 and the microstructural parameters are sum￾marised in Table 1 together with the values of fracture toughness and work of fracture. The alumina materials (Fig. 2a and c) pre￾sented different levels of grain growth (Table 1), the microstruc￾ture of the alumina sintered at the highest temperature (1550 ◦C; Fig. 2c) being bimodal with some grains larger than 10 m and

S. Bueno, C Baudin / Journal of the European Ceramic Sociery 27(2007)1455-1462 1459 the rest of the grains in the range described by the average size doped aluminas, 0, I constituted of groups of small (2-3 um) determined by the linear intercept method (Table 1). In the com- alumina grains surrounded by very large(>20-30 um; Fig 3d posites, aluminium titanate was homogeneously distributed and ones. For both laminates, the initial external and central alumina mainly located at alumina triple points and grain boundaries, layers as well as the internal composite layers(Fig. 3c and d) and no titania was observed by this method (FE-SEM presented microstructures similar to those of the corresponding There were no significant differences in the values of frac- monoliths(Fig. 2; Table 1)with similar average grain sizes ture toughness (Table 1) for the alumina materials, whereas the 3.2+0.2 and 2.9+0.2 um for alumina grains in the fine alumina work of fracture values were higher for the specimens sintered and composite layers, respectively, in the laminate sintered at t1550°C. In the composites, considerably higher values of1450°C;and5.2±0.3and3.9±0.2μ m in the fine alumina work of fracture were achieved in relation to monophase alu- and composite layers, respectively, in the laminate sintered at minas and an increase with the sintering temperature was also 1550C. bserved. The fracture toughness values of the composites were ig 3c and d presents the profiles of titania content in the alu- slightly higher than those of alumina and similar for both sin- mina layers. Significant TiO2 amounts, up to 0.08 andO. 15 mol% in materials sintered at 1450 and 1550C, respectively, were The microstructures of the obtained laminated structures are found at both sides of the composite layers with a decreasing shown in Fig 3. In the samples sintered at 1450C the thickness trend through the widths of the contiguous large grained alu- of the"in situ"developed large grain sized alumina layers was less than 80 um(Fig 3a)and was formed by grains with sizes Characteristic load-displacement curves of notched samples up to 20 um(Fig. 3c). In the specimens sintered at 1550C of the laminates are shown in Fig. 4. The tests of specimens with the thickness was about 150 pm(Fig. 3b)and the layers had relative notch size of 0. 4 of the width of the external alumina extremely bimodal microstructures, as those reported for TiO2- layer(Fig. 1)were unstable for the materials sintered at 1450C Fig. 5. Characteristic fract es of notched samples of the laminates sintered at 1450C. FE-SEM micrographs: in the fine grained alumina layer from the tip of the notch Specimen with a relative notch depth of 0. 4 of the external alumina o) Transition between the fine and large grained alumina layer with mostly intergranular fracture. Specimen with a relative notch depth of 0.8 of the external alumina (c)Intergranular fracture in the large grained alumina layer showing high levels of intergranular porosity. The composite AlOAT layer is shown at the top of the

S. Bueno, C. Baud´ın / Journal of the European Ceramic Society 27 (2007) 1455–1462 1459 the rest of the grains in the range described by the average size determined by the linear intercept method (Table 1). In the com￾posites, aluminium titanate was homogeneously distributed and mainly located at alumina triple points and grain boundaries, and no titania was observed by this method (FE-SEM). There were no significant differences in the values of frac￾ture toughness (Table 1) for the alumina materials, whereas the work of fracture values were higher for the specimens sintered at 1550 ◦C. In the composites, considerably higher values of work of fracture were achieved in relation to monophase alu￾minas and an increase with the sintering temperature was also observed. The fracture toughness values of the composites were slightly higher than those of alumina and similar for both sin￾tering temperatures. The microstructures of the obtained laminated structures are shown in Fig. 3. In the samples sintered at 1450 ◦C the thickness of the “in situ” developed large grain sized alumina layers was less than 80m (Fig. 3a) and was formed by grains with sizes up to 20m (Fig. 3c). In the specimens sintered at 1550 ◦C the thickness was about 150 m (Fig. 3b) and the layers had extremely bimodal microstructures, as those reported for TiO2- doped aluminas,10,11 constituted of groups of small (∼=2–3m) alumina grains surrounded by very large (>20–30 m; Fig. 3d) ones. For both laminates, the initial external and central alumina layers as well as the internal composite layers (Fig. 3c and d) presented microstructures similar to those of the corresponding monoliths (Fig. 2; Table 1) with similar average grain sizes: 3.2 ± 0.2 and 2.9 ± 0.2m for alumina grains in the fine alumina and composite layers, respectively, in the laminate sintered at 1450 ◦C; and 5.2 ± 0.3 and 3.9 ± 0.2m in the fine alumina and composite layers, respectively, in the laminate sintered at 1550 ◦C. Fig. 3c and d presents the profiles of titania content in the alu￾mina layers. Significant TiO2 amounts, up to 0.08 and 0.15 mol% in materials sintered at 1450 and 1550 ◦C, respectively, were found at both sides of the composite layers with a decreasing trend through the widths of the contiguous large grained alu￾mina layers. Characteristic load–displacement curves of notched samples of the laminates are shown in Fig. 4. The tests of specimens with relative notch size of 0.4 of the width of the external alumina layer (Fig. 1) were unstable for the materials sintered at 1450 ◦C Fig. 5. Characteristic fracture surfaces of notched samples of the laminates sintered at 1450 ◦C. FE-SEM micrographs: (a) Mixed trans/intergranular fracture in the fine grained alumina layer from the tip of the notch. Specimen with a relative notch depth of 0.4 of the external alumina layer. (b) Transition between the fine and large grained alumina layer with mostly intergranular fracture. Specimen with a relative notch depth of 0.8 of the external alumina layer. (c) Intergranular fracture in the large grained alumina layer showing high levels of intergranular porosity. The composite A10AT layer is shown at the top of the micrograph.

S. Bueno, C. Baudin/Journal of the European Ceramic Sociery 27(2007)1455-1462 and semi-stable for the materials sintered at 1550C. For larger largest notch, fracture started from the large grain sized internal notch depths(0.8 of the width of the external alumina layer; layer( Fig. 6c), whereas in the other specimens, fracture started ig. 1)all tests were semi-stable and, therefore, suitable for the from the surface fine alumina(Figs. 5 and 6a and b) determination of work of fracture Both parts of the specimens remained together after testing The fracture toughness and work of fracture values for the (Fig. 7) when stable fracture was attained. The crack path was laminates are included in Table 2. In the laminate sintered at low quite straight in the laminated specimens sintered at low temper- temperature, fracture toughness was independent of the notch ature(Fig. 7a), whereas it was extremely tortuous in the materials depth. This trend could not be checked for the work of fracture sintered at 1550"C In these latter, large alumina grains of the due to the unstable behaviour of the samples with the shortest"in situ"formed coarse layers acting as sites for crack branching notch. Conversely, in the laminate with thicker"in situ"formed and bridges in the wake of the propagating crack were observed large grain sized alumina layer there was a decrease in fracture(Fig. 7b) toughness and work of fracture when the tests were performe with the longest notch de 4. Discussion In Figs. 5 and 6, characteristic fracture surfaces of the layered pecimens are shown. For both materials, fracture was mixe Both laminates presented the expected microstructural devel trans/intergranular in the fine grained alumina layers, the largest opment(Fig 3), with"in situ"formed layers of large grain sized grains being the ones most frequently traversed by the crack alumina at both sides of the composites layers(Fig 3a and b) (Figs. 5a and 6a and b), and was intergranularin the large grained There was a complete agreement between maximum sintering alumina layers(Figs. 5b and c, and 6b and c). In these latter, temperature, thickness of the"in situ"formed large grained alu- numerous pores located at the grain boundaries were revealed mina layers and extension of titania diffusion, which verifies the y the fracture. In the specimens sintered at 1550"C and with the suitability of the proposed processing method to obtain large 20 um Fig. 6. Characteristic fracture surfaces of notched f the laminates sintered at 1550C FE-SEM micrograph (a)Mixed trans/intergranular fracture in the fine gra mina layer. Specimen with a relative notch depth of 0.4 of the extemal alumina layer. (b) Intergranular fracture in the large grained alumin in a specimen with relative notch depth of 0. 4 of the external alumina layer. Fine alumina layer is at the bottom and the composite layer at the (c) Intergranular fracture in the specimen with the largest notch(0. 8). Fracture started from the large grained layer and mostly intergranular porosity was revealed by the fracture

1460 S. Bueno, C. Baud´ın / Journal of the European Ceramic Society 27 (2007) 1455–1462 and semi-stable for the materials sintered at 1550 ◦C. For larger notch depths (0.8 of the width of the external alumina layer; Fig. 1) all tests were semi-stable and, therefore, suitable for the determination of work of fracture. The fracture toughness and work of fracture values for the laminates are included in Table 2. In the laminate sintered at low temperature, fracture toughness was independent of the notch depth. This trend could not be checked for the work of fracture due to the unstable behaviour of the samples with the shortest notch. Conversely, in the laminate with thicker “in situ” formed large grain sized alumina layer there was a decrease in fracture toughness and work of fracture when the tests were performed with the longest notch depth. In Figs. 5 and 6, characteristic fracture surfaces of the layered specimens are shown. For both materials, fracture was mixed trans/intergranular in the fine grained alumina layers, the largest grains being the ones most frequently traversed by the crack (Figs. 5a and 6a and b), and was intergranular in the large grained alumina layers (Figs. 5b and c, and 6b and c). In these latter, numerous pores located at the grain boundaries were revealed by the fracture. In the specimens sintered at 1550 ◦C and with the largest notch, fracture started from the large grain sized internal layer (Fig. 6c), whereas in the other specimens, fracture started from the surface fine alumina (Figs. 5 and 6a and b). Both parts of the specimens remained together after testing (Fig. 7) when stable fracture was attained. The crack path was quite straight in the laminated specimens sintered at low temper￾ature (Fig. 7a), whereas it was extremely tortuous in the materials sintered at 1550 ◦C. In these latter, large alumina grains of the “in situ” formed coarse layers acting as sites for crack branching and bridges in the wake of the propagating crack were observed (Fig. 7b). 4. Discussion Both laminates presented the expected microstructural devel￾opment (Fig. 3), with “in situ” formed layers of large grain sized alumina at both sides of the composites layers (Fig. 3a and b). There was a complete agreement between maximum sintering temperature, thickness of the “in situ” formed large grained alu￾mina layers and extension of titania diffusion, which verifies the suitability of the proposed processing method to obtain large Fig. 6. Characteristic fracture surfaces of notched samples of the laminates sintered at 1550 ◦C. FE-SEM micrographs: (a) Mixed trans/intergranular fracture in the fine grained alumina layer. Specimen with a relative notch depth of 0.4 of the external alumina layer. (b) Intergranular fracture in the large grained alumina layer in a specimen with relative notch depth of 0.4 of the external alumina layer. Fine alumina layer is at the bottom and the composite layer at the top. (c) Intergranular fracture in the specimen with the largest notch (0.8). Fracture started from the large grained layer and mostly intergranular porosity was revealed by the fracture

S. Bueno, C. Baudin Journal of the European Ceramic Sociery 27(2007)1455-1462 1461 decreased rapidly with increasing grain sizes, so the dominant fracture mode was transgranular for mean grain sizes larger than 10-20 um. 8 In alumina materials with mean grain size of 20 um and R-curve behaviour, intergranular fracture was also … observed for very large grains, in the order of 100 um, ac ing as bridges. The fractographic observations in this work ( Figs. 5 and 6) show that effectively the fine alumina layers (Figs. 5a and 6a), with relatively small alumina mean grain size (3.2 and 5.2 um at 1450 and 1550C, respectively ), presented a mixed inter/transgranular fracture, whereas the "in situ"formed large grained alumina layers presented intergranular fracture, even for grains with sizes in the range of or larger than 20-30 um (Fig. 6b and c). This mode of fracture is explained by the pres- ence of the extensive intergranular porosity(Fig. 5c and 6b and c) that provides grain boundaries weak enough to change the frac ture fronts These laminated structures developed low residual during cooling from sintering due to the similarity between the thermal expansions of the layers. Approximately constant com- pressive stresses of about 20 MPa and tensile stresses from about 20 MPa at the interface to 2 MPa at the interior in the alumina and composite layers, respectively, have been determined the material sintered at 1550C.> This distribution of residual stresses explains the similarity of the fracture toughness values of the laminates (Table 2)and the monolithic alumina materi- als obtained at the same conditions (Table 1). In the absence of residual stresses the mechanical behaviour of ceramic lami- nates formed by non-interacting layers, such as those fabricated from micrometer-sized particles of compatible phases, should be derived from the mechanical properties of monoliths with the same compositions as those of the layers and processed in the the lar In order to analyse the work of fracture values of the laminates (Table 2)the combination of alumina with composite layers, with different mechanical responses, has to be considered. As terms of comparison, the linear combinations of the work of fracture values of the monoliths using as relative weight the vol laminated samples showing characteristic crack paths. Crack growth propagation ume fraction of each composition in the laminate (UAIOAT =0.1 from the bottom to the top Dashed lines mark the composite internal layers: UA=0.9)are 17.0 and 224J m-2 for the laminates sintered at (a) Laminate sintered at 1450"C. Relative notch size of 0.8 of the extemal 1450 and 1550C, respectively. In both cases, the actual work alumina layer. inate sintered at 1550C. Relative notch size of 0.4 of the external of fracture values of the laminates were higher than the linear mina layer. The large alumina grains of the in situ formed large grain combination, revealing additional energy consuming process mina layers act as sites for crack branching and as bridges in the wake during fracture which should be associated to the presence of the of the propagating crack. large grained alumina layer In the laminate sintered at the lowest temperature, for which less differences between the calculated microstructural differences from similar green bodies. The unex- and determined work of fracture existed (39%), there was not pected high levels of porosity observed at the grain boundaries a macroscopic effect of the coarse grained layer(Fig. 7a) in the large grained alumina layers(Figs. 5c and 6c)have to be crack propagation. In this material, the only reinforcing mecha related to the effect of TiO2 during sintering as occurs in TiO2- nism would be the crack diverting effect of the larger grains, as doped mullite, and their origin has to be further analysed. compared to those of the finer grained layers(Fig5b),which Nevertheless, what is clear is the strong effect of this intergral- would lead to a slightly larger crack front and, consequently, ular porosity on the fracture characteristics of the considered work of fracture values. Conversely, in the laminate sintered at the highest temperature, there was a large scale effect of the In general, the fracture mode of alumina materials has a coarse layer in which significant crack branching and bridging strong dependence on grain size. Materials with mean grain occurred (Fig. 7b). As additional evidence of the reinforcement size in the range 3-20 um showed both transgranular and inter- produced by this heterogeneous layer, the work of fracture val- granular behaviour and the proportion of intergranular fracture ues of specimens of this material with large notches, in which

S. Bueno, C. Baud´ın / Journal of the European Ceramic Society 27 (2007) 1455–1462 1461 Fig. 7. Optical microscopy micrographs of polished lateral faces of fractured laminated samples showing characteristic crack paths. Crack growth propagation is from the bottom to the top. Dashed lines mark the composite internal layers: (a) Laminate sintered at 1450 ◦C. Relative notch size of 0.8 of the external alumina layer. (b) Laminate sintered at 1550 ◦C. Relative notch size of 0.4 of the external alumina layer. The large alumina grains of the in situ formed large grained alumina layers act as sites for crack branching and as bridges in the wake of the propagating crack. microstructural differences from similar green bodies. The unex￾pected high levels of porosity observed at the grain boundaries in the large grained alumina layers (Figs. 5c and 6c) have to be related to the effect of TiO2 during sintering as occurs in TiO2- doped mullite,17 and their origin has to be further analysed. Nevertheless, what is clear is the strong effect of this intergran￾ular porosity on the fracture characteristics of the considered layers. In general, the fracture mode of alumina materials has a strong dependence on grain size. Materials with mean grain size in the range 3–20m showed both transgranular and inter￾granular behaviour and the proportion of intergranular fracture decreased rapidly with increasing grain sizes, so the dominant fracture mode was transgranular for mean grain sizes larger than 10–20m.18 In alumina materials with mean grain size of 20m and R-curve behaviour, intergranular fracture was also observed for very large grains, in the order of 100 m, act￾ing as bridges.19 The fractographic observations in this work (Figs. 5 and 6) show that effectively the fine alumina layers (Figs. 5a and 6a), with relatively small alumina mean grain size (3.2 and 5.2m at 1450 and 1550 ◦C, respectively), presented a mixed inter/transgranular fracture, whereas the “in situ” formed large grained alumina layers presented intergranular fracture, even for grains with sizes in the range of or larger than 20–30 m (Fig. 6b and c). This mode of fracture is explained by the pres￾ence of the extensive intergranular porosity (Fig. 5c and 6b and c) that provides grain boundaries weak enough to change the frac￾ture fronts. These laminated structures developed low residual stresses during cooling from sintering due to the similarity between the thermal expansions of the layers. Approximately constant com￾pressive stresses of about 20 MPa and tensile stresses from about 20 MPa at the interface to 2 MPa at the interior in the alumina and composite layers, respectively, have been determined for the material sintered at 1550 ◦C.13 This distribution of residual stresses explains the similarity of the fracture toughness values of the laminates (Table 2) and the monolithic alumina materi￾als obtained at the same conditions (Table 1). In the absence of residual stresses, the mechanical behaviour of ceramic lami￾nates formed by non-interacting layers, such as those fabricated from micrometer-sized particles of compatible phases, should be derived from the mechanical properties of monoliths with the same compositions as those of the layers and processed in the same way as the laminates. In order to analyse the work of fracture values of the laminates (Table 2) the combination of alumina with composite layers, with different mechanical responses, has to be considered. As terms of comparison, the linear combinations of the work of fracture values of the monoliths using as relative weight the vol￾ume fraction of each composition in the laminate (vA10AT = 0.1, vA = 0.9) are 17.0 and 22.4 J m−2 for the laminates sintered at 1450 and 1550 ◦C, respectively. In both cases, the actual work of fracture values of the laminates were higher than the linear combination, revealing additional energy consuming processes during fracture which should be associated to the presence of the large grained alumina layer. In the laminate sintered at the lowest temperature, for which less differences between the calculated and determined work of fracture existed (∼=39%), there was not a macroscopic effect of the coarse grained layer (Fig. 7a) on crack propagation. In this material, the only reinforcing mecha￾nism would be the crack diverting effect of the larger grains, as compared to those of the finer grained layers (Fig. 5b), which would lead to a slightly larger crack front and, consequently, work of fracture values. Conversely, in the laminate sintered at the highest temperature, there was a large scale effect of the coarse layer in which significant crack branching and bridging occurred (Fig. 7b). As additional evidence of the reinforcement produced by this heterogeneous layer, the work of fracture val￾ues of specimens of this material with large notches, in which

S Bueno, C Bandin /Journal of the European Ceramic Sociery 27(2007)1455-1462 the effect of the first heterogeneous layer was partially removed 5. Bueno, S, Moreno, R. and Baudin, C, Design and processing of because it was partially traversed by the notch(Figs. I an Al2O3-Al2TiOs layered structures. J. Eur. Ceram Soc., 2005, 25, 847-856 were lower than those for specimens with the shortest notches 6. Bueno, S, Moreno, R and Baudin, C, Colloidal processing of laminates in Table 2 the system alumina-titania Key Eng Mater, 2004, 264-268, 61-64. 7. Bueno, S. and Baudin, C, Fracture mechanisms in laminates in the alumina-titania system. Key Eng Mater, 2005, 290. 208-213 5. Conclusions titanate/alumina composites. Mater. Res. Bull., 1998, 33, 1475-1482. Al2O3-Al2TiO5 laminates with"in situdeveloped inter 9. Morsi, K, Keshavan, H and Bal,S, Hot pressing of graded ultrafine-grained alumina bioceramics. Mater Sci. Eng, 2004, 386, 384-389 nal alumina layers with large grains were obtained from small 10. Powers, I.D. and Glaeser,A M, Titanium effects on sin and grain grained green compacts and using the effect of TiO2 on owth of alumina. In Sintering Technology, ed. R. M. German, G.L. Mess- the microstructural development of alumina. These laminates ing and R. G. Cornwall. Marcel Dekker Inc, USA, 1996, pp 33-40. resent large work of fracture values, as compared to those 11. Chi, M, Gu, H. Wang, X and Wang, P. Evidence of bilevel solubility in the of monoliths of the same composition as the layers, which are bimodal microstructure of TiO2-doped alumina. J. Am. Ceram Soc., 2003 86,1953-1955 determined by the size and fracture behaviour of the"in situ 12. Bueno, S. and Baudin, C, In situ developed alumina-aluminium titanate developed alumina layers. Crack branching and bridging were laminates with large microstructural differences between the layers. J Mater identified as the main toughening mechanisms Sci, Published on line 5 April 2006. 13. De Portu, G, Bueno, S, Micele, L, Baudin. C. and Pezzotti, G. Piezo. Acknowledgements Eur Ceram.Soc.,2006,26,2699-2705 14. Sakai, M, Yoshimura, J, Goto, Y and Inagaki, M, R-Curve behavior of Work supported in part by the European Community polycrystalline graphite: microcracking and grain bridging in the wake Human Potential Programme under contract HPRN-CT-2002- region.J Am Ceram Soc., 1988, 71, 609-616 00203 [SICMAC], by the projects CICYT MAT2003-00836, 15. Bueno, S, Moreno, R and Baudin, C, Reaction sintered AlO3/Al2TiOs CAMGRMATO707-2004 and by the grant CSIC I3P-BPD2001 microcrack free composites obtained by colloidal filtration. J. Eur Ceram. Soc,2004,24,2785-2791 I(Spain) 16. Guinea, G.V., Pastor, J. Y, Planas, J and Elices, M, Stress intensity factor, compliance and CMOD for a general three-point-bend beam. Int J. Fract. References 1998,89,103-1 17. Baudin, C. and Moya, J.S., Influence of titanium dioxide on the sintering 1. Chan, H M, Layered and microstructural evolution of mullite bodies. J. Am. Ceram Soc. 1984 Rev mater sci,197,27,249-282. 67,C134C136 Clegg, W.J. Controlling cracks in ceramics. Science, 1999, 286, 1097-1099. 18. Mussler, B- Swain, M and Claussen, N- Dependence of fracture tou 3. Harmer, M. P, Chan, H M. and Miller, G. A, Unique opportunities for of alumina on grain size and test technique. J. Am. Ceram. Soc. I microstructural engineering with duplex and laminar ceramics composites. 566-572. JAm.Cerm.Soc.,1992,75,1715-1728 19. Swanson, P, Fairbanks, C, Lawn, B, Mai, Y and Hockey, B 4. Russo. C.J. Harmer, M. P. Chan, H. M. and Miller, G. A, Design of a interface grain bridging as a fracture resistance mechanism in ceramics: laminated ceramic composite for improved strength and toughness. J. Am I. Experimental study on alumina. J. Am. Ceram. Soc., 1987, 70, 279. Ceran.Soc.,1992,75,3396-4000

1462 S. Bueno, C. Baud´ın / Journal of the European Ceramic Society 27 (2007) 1455–1462 the effect of the first heterogeneous layer was partially removed because it was partially traversed by the notch (Figs. 1 and 6c), were lower than those for specimens with the shortest notches (Table 2). 5. Conclusions Al2O3–Al2TiO5 laminates with “in situ” developed inter￾nal alumina layers with large grains were obtained from small grained green compacts and using the effect of TiO2 on the microstructural development of alumina. These laminates present large work of fracture values, as compared to those of monoliths of the same composition as the layers, which are determined by the size and fracture behaviour of the “in situ” developed alumina layers. Crack branching and bridging were identified as the main toughening mechanisms. Acknowledgements Work supported in part by the European Community’s Human Potential Programme under contract HPRN-CT-2002- 00203 [SICMAC], by the projects CICYT MAT2003-00836, CAM GRMAT0707-2004 and by the grant CSIC I3P-BPD2001- 1 (Spain). References 1. Chan, H. M., Layered ceramics: processing and mechanical behavior. Annu. Rev. Mater. Sci., 1997, 27, 249–282. 2. Clegg, W. J., Controlling cracks in ceramics. Science, 1999, 286, 1097–1099. 3. Harmer, M. P., Chan, H. M. and Miller, G. A., Unique opportunities for microstructural engineering with duplex and laminar ceramics composites. J. Am. Ceram. Soc., 1992, 75, 1715–1728. 4. Russo, C. J., Harmer, M. P., Chan, H. M. and Miller, G. A., Design of a laminated ceramic composite for improved strength and toughness. J. Am. Ceram. Soc., 1992, 75, 3396–4000. 5. Bueno, S., Moreno, R. and Baud´ın, C., Design and processing of Al2O3–Al2TiO5 layered structures. J. Eur. Ceram. Soc., 2005, 25, 847–856. 6. Bueno, S., Moreno, R. and Baud´ın, C., Colloidal processing of laminates in the system alumina–titania. Key Eng. Mater., 2004, 264–268, 61–64. 7. Bueno, S. and Baud´ın, C., Fracture mechanisms in laminates in the alumina–titania system. Key Eng. Mater., 2005, 290, 208–213. 8. Low, I. M., Synthesis and properties of in situ layered and graded aluminium titanate/alumina composites. Mater. Res. Bull., 1998, 33, 1475–1482. 9. Morsi, K., Keshavan, H. and Bal, S., Hot pressing of graded ultrafine-grained alumina bioceramics. Mater. Sci. Eng., 2004, 386, 384–389. 10. Powers, J. D. and Glaeser, A. M., Titanium effects on sintering and grain growth of alumina. In Sintering Technology, ed. R. M. German, G. L. Mess￾ing and R. G. Cornwall. Marcel Dekker Inc., USA, 1996, pp. 33–40. 11. Chi, M., Gu, H., Wang, X. and Wang, P., Evidence of bilevel solubility in the bimodal microstructure of TiO2-doped alumina. J. Am. Ceram. Soc., 2003, 86, 1953–1955. 12. Bueno, S. and Baud´ın, C., In situ developed alumina–aluminium titanate laminates with large microstructural differences between the layers. J. Mater. Sci., Published on line 5 April 2006. 13. De Portu, G., Bueno, S., Micele, L., Baud´ın, C. and Pezzotti, G., Piezo￾spectroscopic characterization of alumina–aluminium titanate laminates. J. Eur. Ceram. Soc., 2006, 26, 2699–2705. 14. Sakai, M., Yoshimura, J., Goto, Y. and Inagaki, M., R-Curve behavior of polycrystalline graphite: microcracking and grain bridging in the wake region. J. Am. Ceram. Soc., 1988, 71, 609–616. 15. Bueno, S., Moreno, R. and Baud´ın, C., Reaction sintered Al2O3/Al2TiO5 microcrack free composites obtained by colloidal filtration. J. Eur. Ceram. Soc., 2004, 24, 2785–2791. 16. Guinea, G. V., Pastor, J. Y., Planas, J. and Elices, M., Stress intensity factor, compliance and CMOD for a general three-point-bend beam. Int. J. Fract., 1998, 89, 103–118. 17. Baud´ın, C. and Moya, J. S., Influence of titanium dioxide on the sintering and microstructural evolution of mullite bodies. J. Am. Ceram. Soc., 1984, 67, C134–C136. 18. Mussler, B., Swain, M. and Claussen, N., Dependence of fracture toughness of alumina on grain size and test technique. J. Am. Ceram. Soc., 1982, 65, 566–572. 19. Swanson, P., Fairbanks, C., Lawn, B., Mai, Y. and Hockey, B., Crack￾interface grain bridging as a fracture resistance mechanism in ceramics: I. Experimental study on alumina. J. Am. Ceram. Soc., 1987, 70, 279– 289