《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_strength of A-Z

Availableonlineatwww.sciencedirectcom ScienceDirect E噩≈RS ELSEVIER Joumal of the European Ceramic Society 27(2007)1443-1448 www.elsevier.comlocate/jeurceramsoc Threshold strength evaluation on an Al2 O3-ZrO2 multilayered system J. Pascual, M. Angladaa L Llanes a, w erencia b R Bermejo, Y. Torres a, C. Baudin, A.J. Sanchez Departamento de Ciencia de Materiales e Ingenieria Metaluirgica, Universidad Politecnica de cataluia, Avda. Diagonal 647(ETSEIB) 08028 Barcelona, Spain Instituto de Ceramica y Vidrio(CS/C), C/Kelsen 5. 28010 Madrid, Spain e Institut fiir Struktur- und Funktionskeramik der Montanuniversitar Leoben, Peter-Tunner-Strasse 5, 8700 Leoben, austria Available online 15 June 2006 Abstract The fracture behaviour of an alumina-zirconia multilayered ceramic obtained by slip casting has been evaluated and compared to monolithic alumina taken as a reference. It is found that the laminated system exhibits an apparent fracture toughness KapLe, calculated experimentally(SEPB) and analytically with a weight function approach, higher than twice the value determined for the monolithic material. Additionally, the experimental rupture tests accomplished indicate the existence of a threshold strength in the multilayer. The increase in Kapt. e in the laminate, associated with a pronounced R-curve behaviour, is discussed in terms of the high compressive stresses in the internal layers and their effect on the effective driving force for crack extension o 2006 Elsevier Ltd. All rights reserved. Keywords: Slip casting: Composites; Mechanical properties; Toughness 1. Introduction does not occur. Such a scenario has been reported by Raoet al6 for an alumina-alumina mullite multilayered system when a ten- In the last decades, new microstructural design concepts have sile stress was applied parallel to the layers. From this perspec been attempted for improving fracture toughness of structural tive, zirconia-containing laminar ceramics have been employed ceramics.Among them, doping, fibre and/or particle reinforce- to develop compressive stresses in the internal layers by means ment, functional grading and layered architectural design may of the tetragonal to monoclinic phase transformation that takes be highlighted. In particular, ceramic composites with a lay ered structure such as alumina-zirconia -3 and mullite-alumina place when cooling down during sintering among others, have been reported to exhibit an increased appar- condition the residual stress field within the multilayer. Under ent fracture toughness and energy absorption as well as non- certain conditions, these compressive stresses may act as a bar- catastrophic failure behaviour. For strongly bonded multilayers rier to crack propagation. In other cases, crack bifurcation and/or the elastic mismatch during sintering between adjacent lay- deflection phenomena result in an increase of the material frac ers, resulting from the ditterence in Youngs moduli, thermal ture toughness and energy absorption capability. 8 It is the purpose of this investigation to evaluate the fracture formations, generates residual stresses throughout the material haviour of an alumina/zirconia-based multilayered ceramic, These residual stresses can be controlled in order to improve their with highly compressed thin layers, in comparison to the alu mechanical properties On ceramics and glasses, for instance, the mina monolith taken as reference material. In doing so, special compressive stresses on the surface have proven to be useful for attention is paid to evaluate the possible existence of a thresh- ceramics designed with compressive stresses in the bulk may Apparent fracture toughness is experimentally determined using present a threshold strength below which catastrophic failure the SEPB method and the R-curve behaviour evaluated analyti cally by the weight function approach. Finally, work of fra actur Corresponding author. Tel: +34 93401 1083: fax: +34 934016706 is estimated by conducting crack opening displacement( COD) E-mail address: luis. miguel llanes@ upc.edu(L. Llanes). tests on notched laminated samples 0955-2219/S-see front matter o 2006 Elsevier Ltd. All rights reserved. doi: 10. 1016/j-jeurceramsoc. 2006.05.037

Journal of the European Ceramic Society 27 (2007) 1443–1448 Threshold strength evaluation on an Al2O3–ZrO2 multilayered system R. Bermejo a, Y. Torres a, C. Baud´ın b, A.J. Sanchez-Herencia ´ b, J. Pascual c, M. Anglada a, L. Llanes a,∗ a Departamento de Ciencia de Materiales e Ingenier´ıa Metal ´urgica, Universidad Polit´ecnica de Catalu ˜na, Avda. Diagonal 647 (ETSEIB), 08028 Barcelona, Spain b Instituto de Cer ´amica y Vidrio (CSIC), C/Kelsen 5, 28010 Madrid, Spain c Institut f ¨ur Struktur- und Funktionskeramik der Montanuniversit ¨at Leoben, Peter-Tunner-Strasse 5, 8700 Leoben, Austria Available online 15 June 2006 Abstract The fracture behaviour of an alumina–zirconia multilayered ceramic obtained by slip casting has been evaluated and compared to monolithic alumina taken as a reference. It is found that the laminated system exhibits an apparent fracture toughness Kapt,c, calculated experimentally (SEPB) and analytically with a weight function approach, higher than twice the value determined for the monolithic material. Additionally, the experimental rupture tests accomplished indicate the existence of a threshold strength in the multilayer. The increase in Kapt,c in the laminate, associated with a pronounced R-curve behaviour, is discussed in terms of the high compressive stresses in the internal layers and their effect on the effective driving force for crack extension. © 2006 Elsevier Ltd. All rights reserved. Keywords: Slip casting; Composites; Mechanical properties; Toughness 1. Introduction In the last decades, new microstructural design concepts have been attempted for improving fracture toughness of structural ceramics. Among them, doping, fibre and/or particle reinforce￾ment, functional grading and layered architectural design may be highlighted. In particular, ceramic composites with a lay￾ered structure such as alumina–zirconia1–3 and mullite-alumina4 among others, have been reported to exhibit an increased appar￾ent fracture toughness and energy absorption as well as non￾catastrophic failure behaviour. For strongly bonded multilayers the elastic mismatch during sintering between adjacent lay￾ers, resulting from the difference in Young’s moduli, thermal expansion coefficients, chemical reactions and/or phase trans￾formations, generates residual stresses throughout the material. These residual stresses can be controlled in order to improve their mechanical properties. On ceramics and glasses, for instance, the compressive stresses on the surface have proven to be useful for increasing their fracture strength.1,5 On the other hand, laminar ceramics designed with compressive stresses in the bulk may present a threshold strength below which catastrophic failure ∗ Corresponding author. Tel.: +34 934011083; fax: +34 934016706. E-mail address: luis.miguel.llanes@upc.edu (L. Llanes). does not occur. Such a scenario has been reported by Rao et al.6 for an alumina–alumina mullite multilayered system when a ten￾sile stress was applied parallel to the layers. From this perspec￾tive, zirconia-containing laminar ceramics have been employed to develop compressive stresses in the internal layers by means of the tetragonal to monoclinic phase transformation that takes place when cooling down during sintering.1,7 The correspond￾ing volume increase associated with such transformation will condition the residual stress field within the multilayer. Under certain conditions, these compressive stresses may act as a bar￾rier to crack propagation. In other cases, crack bifurcation and/or deflection phenomena result in an increase of the material frac￾ture toughness and energy absorption capability.8 It is the purpose of this investigation to evaluate the fracture behaviour of an alumina/zirconia-based multilayered ceramic, with highly compressed thin layers, in comparison to the alu￾mina monolith taken as reference material. In doing so, special attention is paid to evaluate the possible existence of a thresh￾old strength under flexural loading normal to the layer plane. Apparent fracture toughness is experimentally determined using the SEPB method and the R-curve behaviour evaluated analyti￾cally by the weight function approach. Finally, work of fracture is estimated by conducting crack opening displacement (COD) tests on notched laminated samples. 0955-2219/$ – see front matter © 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.jeurceramsoc.2006.05.037

R. Bermejo et al. /Journal of the European Ceramic Sociery 27(2007)1443-1448 2. Experimental work ment model developed elsewhere. 0 All materials were assumed to be isotropic so that only two independent properties such 2.1. Material processing and basic characterization as the Youngs modulus and the Poissons ratio had to be pro- vided Thermal expansion coefficients(aATZ=9.82 x 10-6K-1 Laminar ceramic composites have been fabricated via and AMz=8.02 x 10-6K-)were discretely introduced in the sequential slip casting. Starting powders were submicron-sized model from data corresponding to the dilatometry curves, con- Alumina HPA O5(Condea, USA)with a particle size of 0. 25 Fr and pure zirconia TZ-0(Tosoh, Japan)with a particle size of sidering 1200 C as the reference temperature 0.3 um Slurries were made by mixing starting powders with 23. Threshold strength evaluation DI water that contained the desired amount of dispersant used for stabilization. A slurry composed of Al2O3/5 vol. Y203- tabilized ZrO(t-ZrO2), referred as ATZ, was used to form the inner layers may exhibit an apparent R-curve behaviour'as the thicker layers. The t-ZrOz was employed to control the well as a threshold strength, regardless of the type and size of grain size of the Al2O3 during densification. In order to form initial flaws. In order to estimate and study the effect of a possi- the thin layers a slurry containing Al2O3/30 vol. ZrO2(m- ble presence of threshold strength on the mechanical behaviour ZrO2), noted as AMZ, was utilized. Details of the processing of the laminate investigated, four specimens were ground and rocedure may be found in a previous report. Cast speci- polished up to 3 um at the surface and at one of the lateral faces mens were then fired at 1550C for 2h with a heating and Different combinations of Vickers indentations were placed lon cooling rate of 5 C/min. Rectangular plates of approximately gitudinally on each specimen surface with an offset separation 60 mm x 60 mm X 4 mm were obtained. Flexural bars were cut distance of 2 mm to avoid any crack interaction:(a)200, 200, from the plates, and the top and lateral surfaces of each speci- 100, 50N; (b)150, 150, 100, 50N; (c)100, 100, 50, 30N;(d) en were polished to a 3-um finish with a diamond abrasive 50, 50, 30, 30N. The same procedure was conducted on the atz for microstructural observation and mechanical characteriza- monoliths for comparison. The crack length was measured by tion. A homogeneous layer thickness was obtained for both means of an optical microscope and by recourse to Nomarski ATZ and AMZ layers, resulting in a symmetrical multilayered interference contrast. Then, all the specimens were tested until architecture(see Fig. 1) fracture under four-point bending on a servo-hydraulic machine, with inner and outer spans of 10 and 20 mm, respectively, at a 2. 2. Residual stress estimation load rate of 100 N/s. The failure stress for the indented speci Dynamic sintering was carried out on the monoliths to cal- EiM culate the residual stresses generated in the laminate system. Oi, y (ED(- yua (1) Transformation strains of the thin layers containing monoclinic ZrOz were estimated by measuring the transformation strains where Ei is the Youngs modulus of the corresponding layer. of monolithic bar specimens with the same compositions as M the moment for the case of four-point bending tests(M=Fa, the thin compressive layers. The dilatational strains produced where F is the applied load and a the distance between inner y these dense AMZ and atz monoliths, during cooling from and outer spans), yna is the position of the neutral axis and el densification temperature, were measured with a dilatometer the flexural rigidity of the composite calculated for bending per- Netzsch,Germany). The residual stress profile pendicular to the layers. 2. 3 throughout the multilayer was determined by a 3D finite ele 2.4. Fracture toughness assessment In the case of the monolithic materials, fracture toughness was evaluated by two different methods: (i) the microfracture indentation(IM)method, proposed by Anstis et al. 4and (i) the single-edge V-notch beam(SEVNB)method. 5 For the lat- ter, a razor blade automatic machine was utilized to create the notch, which was sharpened up to a radius of less than 10 um to minimize the influence of notch radius on Klc. 6 This method was also employed for the evaluation of the fracture toughness on the laminates. In doing so, a notch of the radius as that for the monolith was machined in the outer most ATZlayer. Very interesting, after reaching a certain notch depth a straight, thin, through-thickness crack developed from the notch 500um and arrested at the aTZAMz interface(see Fig 2a), yielding as a result an effective SEPb configuration. within this context Fig 1. SEM micrograph of the Al2O3-ZrO2 multilayered architecture obtained and considering that when fracture occurs the crack continues by slip casting in mode I(see Fig. 2b), it is sustained to employ the expression

1444 R. Bermejo et al. / Journal of the European Ceramic Society 27 (2007) 1443–1448 2. Experimental work 2.1. Material processing and basic characterization Laminar ceramic composites have been fabricated via sequential slip casting. Starting powders were submicron-sized Alumina HPA 0.5 (Condea, USA) with a particle size of 0.25m and pure zirconia TZ-0 (Tosoh, Japan) with a particle size of 0.3m. Slurries were made by mixing starting powders with DI water that contained the desired amount of dispersant used for stabilization. A slurry composed of Al2O3/5 vol.% Y2O3- stabilized ZrO2 (t-ZrO2), referred as ATZ, was used to form the thicker layers. The t-ZrO2 was employed to control the grain size of the Al2O3 during densification. In order to form the thin layers a slurry containing Al2O3/30 vol.% ZrO2 (m￾ZrO2), noted as AMZ, was utilized. Details of the processing procedure may be found in a previous report.9 Cast speci￾mens were then fired at 1550 ◦C for 2 h with a heating and cooling rate of 5 ◦C/min. Rectangular plates of approximately 60 mm × 60 mm × 4 mm were obtained. Flexural bars were cut from the plates, and the top and lateral surfaces of each speci￾men were polished to a 3-m finish with a diamond abrasive for microstructural observation and mechanical characteriza￾tion. A homogeneous layer thickness was obtained for both ATZ and AMZ layers, resulting in a symmetrical multilayered architecture (see Fig. 1). 2.2. Residual stress estimation Dynamic sintering was carried out on the monoliths to cal￾culate the residual stresses generated in the laminate system. Transformation strains of the thin layers containing monoclinic ZrO2 were estimated by measuring the transformation strains of monolithic bar specimens with the same compositions as the thin compressive layers. The dilatational strains produced by these dense AMZ and ATZ monoliths, during cooling from densification temperature, were measured with a dilatometer (DIL 402 E/7, Netzsch, Germany). The residual stress profile throughout the multilayer was determined by a 3D finite ele￾Fig. 1. SEM micrograph of the Al2O3–ZrO2 multilayered architecture obtained by slip casting. ment model developed elsewhere.10 All materials were assumed to be isotropic so that only two independent properties such as the Young’s modulus and the Poisson’s ratio had to be pro￾vided. Thermal expansion coefficients (αATZ = 9.82 × 10−6 K−1 and αAMZ = 8.02 × 10−6 K−1) were discretely introduced in the model from data corresponding to the dilatometry curves, con￾sidering 1200 ◦C as the reference temperature. 2.3. Threshold strength evaluation Laminar ceramics with high compressive residual stresses on the inner layers may exhibit an apparent R-curve behaviour11 as well as a threshold strength, regardless of the type and size of initial flaws. In order to estimate and study the effect of a possi￾ble presence of threshold strength on the mechanical behaviour of the laminate investigated, four specimens were ground and polished up to 3m at the surface and at one of the lateral faces. Different combinations of Vickers indentations were placed lon￾gitudinally on each specimen surface with an offset separation distance of 2 mm to avoid any crack interaction: (a) 200, 200, 100, 50 N; (b) 150, 150, 100, 50 N; (c) 100, 100, 50, 30 N; (d) 50, 50, 30, 30 N. The same procedure was conducted on the ATZ monoliths for comparison. The crack length was measured by means of an optical microscope and by recourse to Nomarski interference contrast. Then, all the specimens were tested until fracture under four-point bending on a servo-hydraulic machine, with inner and outer spans of 10 and 20 mm, respectively, at a load rate of 100 N/s. The failure stress for the indented speci￾mens, σRi, was calculated using Eq. (1): σi,y = EiM (EI) (y − yna) (1) where Ei is the Young’s modulus of the corresponding layer, M the moment for the case of four-point bending tests (M = Fa, where F is the applied load and a the distance between inner and outer spans), yna is the position of the neutral axis and EI the flexural rigidity of the composite calculated for bending per￾pendicular to the layers.12,13 2.4. Fracture toughness assessment In the case of the monolithic materials, fracture toughness was evaluated by two different methods: (i) the microfracture indentation (IM) method, proposed by Anstis et al.14 and (ii) the single-edge V-notch beam (SEVNB) method.15 For the lat￾ter, a razor blade automatic machine was utilized to create the notch, which was sharpened up to a radius of less than 10m to minimize the influence of notch radius on KIc. 16 This method was also employed for the evaluation of the fracture toughness on the laminates. In doing so, a notch of the same depth and radius as that for the monolith was machined in the outer most ATZ layer. Very interesting, after reaching a certain notch depth a straight, thin, through-thickness crack developed from the notch and arrested at the ATZ/AMZ interface (see Fig. 2a), yielding as a result an effective SEPB configuration. Within this context, and considering that when fracture occurs the crack continues in mode I (see Fig. 2b), it is sustained to employ the expression

R Bermejo et al. / Journal of the European Ceramic Society 27 (2007)1443-1448 1445 (a) AMZ 50 um 100 um Fig. 2.(a) SEM micrograph showing a straight, fine crack that arises from the notch and arrests at the first ATZAMZ interlayer; (b) the crack continues propagating in mode I a certain distance before bifurcation;(c)a second notch is made entering the second ATZ layer, where a crack also appears from the at the second ATZAMz layer. given by Tada et al. 7 to determine the apparent fracture tough- 2.5. Work of fracture: COD tests ness, Kla, at the first ATZ/AMZ interlayer. Following the same procedure, another value of the apparent fracture toughness of To determine the work of fracture in the laminate in com- the laminate, KIb, was determined by generating a notch entering parison with that of the ATZ monolith, samples were tested the second ATZ layer(see Fig. 2c)where, once again, a crack in three-point bend under crack opening displacement(COD) of the same characteristics as above arose from the notch and control at a rate of 1 um/s. Identical specimens for mono- stopped at the second ATZAMZ interface. liths and laminates were notched using a razor blade auto In addition, a fracture mechanics weight function analysis matic machine Notches were machined to enter a small depth was effectively used to estimate the crack growth resistance into the aTZ phase to avoid the appearance of a crack out behaviour(R-curve)as a function of the position of an edge crack of the notch tip for the case of the laminates, as described within each multilayered system investigated. The so called above. Thus, depths of 150 um were not surpassed in any case apparent fracture toughness Kap. c (since it is influenced by the A COD-gauge was attached to the specimen surface at the residual stresses) for an arbitrary stress distribution ores(x)nor- notch site to register the crack opening displacement. Addi- mal to the layer plane, may be defined as follows tionally, the extension of the crack in the SEVNB specimens was continuously monitored using a long distance focal opti- cal microscope(Questar QM100) with a final magnification pt, c=Kl of 1000x. Crack opening displacement and load data were recorded by a software (Labview 6.1) coupled to the testing where Kic is the intrinsic fracture toughness of each individual set-up yer calculated by the SEVNB method in the corresponding Assuming that all the work done is consumed in the growth monoliths, x the distance along the crack length measured from of the crack and no significant elastic energy is stored in the surface, a the crack length, and h(a, x)is a weight function the specimen during testing, the total fracture energy may as developed by Fett and Munz for an edge crack in a bar, 8 be attempted through integration of the registered load-COD commonly employed in the evaluation of R-curve behaviour for curve/material volume unit for the Alz monolith and for the multilayered systems. 9-2 lamil

R. Bermejo et al. / Journal of the European Ceramic Society 27 (2007) 1443–1448 1445 Fig. 2. (a) SEM micrograph showing a straight, fine crack that arises from the notch and arrests at the first ATZ/AMZ interlayer; (b) the crack continues propagating in mode I a certain distance before bifurcation; (c) a second notch is made entering the second ATZ layer, where a crack also appears from the notch and stops at the second ATZ/AMZ layer. given by Tada et al.17 to determine the apparent fracture tough￾ness, KIa, at the first ATZ/AMZ interlayer. Following the same procedure, another value of the apparent fracture toughness of the laminate,KIb, was determined by generating a notch entering the second ATZ layer (see Fig. 2c) where, once again, a crack of the same characteristics as above arose from the notch and stopped at the second ATZ/AMZ interface. In addition, a fracture mechanics weight function analysis was effectively used to estimate the crack growth resistance behaviour (R-curve) as a function of the position of an edge crack within each multilayered system investigated. The so called apparent fracture toughness Kapt,c (since it is influenced by the residual stresses) for an arbitrary stress distribution σres(x) nor￾mal to the layer plane, may be defined as follows: Kapt,c = KIc − a 0 h(a, x)σres(x) dx (2) where KIc is the intrinsic fracture toughness of each individual layer calculated by the SEVNB method in the corresponding monoliths, x the distance along the crack length measured from the surface, a the crack length, and h(a, x) is a weight function as developed by Fett and Munz for an edge crack in a bar,18 commonly employed in the evaluation of R-curve behaviour for multilayered systems.19–21 2.5. Work of fracture: COD tests To determine the work of fracture in the laminate in com￾parison with that of the ATZ monolith, samples were tested in three-point bend under crack opening displacement (COD) control at a rate of 1 m/s. Identical specimens for mono￾liths and laminates were notched using a razor blade auto￾matic machine. Notches were machined to enter a small depth into the ATZ phase to avoid the appearance of a crack out of the notch tip for the case of the laminates, as described above. Thus, depths of 150 m were not surpassed in any case. A COD-gauge was attached to the specimen surface at the notch site to register the crack opening displacement. Addi￾tionally, the extension of the crack in the SEVNB specimens was continuously monitored using a long distance focal opti￾cal microscope (Questar QM100) with a final magnification of 1000×. Crack opening displacement and load data were recorded by a software (Labview 6.1) coupled to the testing set-up. Assuming that all the work done is consumed in the growth of the crack and no significant elastic energy is stored in the specimen during testing, the total fracture energy may be attempted through integration of the registered load–COD curve/material volume unit for the ATZ monolith and for the laminate.22

R. Bermejo et al. /Journal of the European Ceramic Sociery 27(2007)1443-1448 100N Fig. 5. Optical micrograph showing the failure of the laminate(indentation load 三苏 of 50N) when the threshold stress is reached The rest of indentation cracks corresponding to different loads placed at the ATZ surface have become through thickness cracks and arrested at the amzatz interface several conclusions may be inferred: (i) for the same inden tation loads, measured crack lengths are always larger for the laminate, corroborating the presence of tensile residual stresses “· Surface in the indented ATZ layer, as reported in Fig. 3, (ii)the failure stress values are higher for the laminate for every indentation Distance(mm) fa out the enhanced mechanical behaviour in Fig 3. Stress distribution through the layers in the bulk and at the surface of the comparison to the aTz monolith, (iii)the constant value of these multilayer, calculated using a 3D finite element model. failure stresses, regardless of the indentation flaw size, allows to indicate the existence of a threshold strength in the lami 3. Results and discussion nate with a value of 167+4 MPa, and (iv)fracture of the atz monolith always occurs from the defect originated by the largest 3.. Residual stress distribution indentation load, as it should be expected. However, in the mul tilayered material fracture takes place randomly from any of the Fig 3 represents the magnitude and distribution profile of indentations performed. This may be explained by the fact that, the residual stresses within the inner and outer layers of the regardless of the indentation load applied in the laminate, the laminate calculated both at the surface and in the bulk of the crack shape, a/c, is initially always greater than one, and rises Is app seem to present a constant value of-690 MPa through the entire tally by grinding and polishing procedures), helped by the tensile thin layer. For the internal atz layers a symmetrical distribu- residual stresses in the ATZ layer, more significant as the inter- ion is obtained reaching a maximum value of 100 MPa at the face is approached(Fig 3).When a is equal to the ATZ layer ATZ/AMZ interfaces. On the other hand, for the outer layers thickness the crack stops, and then c starts to grow until the crack residual stresses decrease near the free surface. These results becomes a through-thickness one. Once that all the indentation are in good agreement with those reported on a previous work cracks have become through-thickness cracks and arrested at the where residual stress distribution was evaluated by the indenta- AMZ layer(Fig. 5), the stress intensity factor in any of them is ion technique and analytical solutions.g9 equivalent, regardless of the indentation load used for inducing them 3. 2. Threshold streng th 3.3. Fracture toughness Fig. 4 rep the variation of the failure stress with the square root initial flaw size. c. for the indented atz fracture toughness for the monolithic materials evaluated monolith and for the laminate investigated. From these results. by the microfracture indentation (IM) and the single-edge V- notch beam (SEVNB) methods are reported in Table 1. It is interesting to note here that these Kle values may be even over estimated because the IM method is affected by the residual stress field inherent to the indentation while the notch radius for the SevnB method is not as sharp as a real crack and is greater than the microstructure scale length KIb, for the multilayered material were 7. 1 and 8.1 MPamlk d gness K determined at the first and second AMZ layer, respectively Table 1 Fracture toughness values for the ATZ and AMZ monoliths as determined by means of the IM and sevnb methods Square root of initial flaw size, c(m) SEVNB Fig 4. Plot of measured four-point bending failure stress vs square root of crack ATZ gth in ATZ monolith()and laminate (O) specimens containing several AMZ ups of Vickers indentations at the surface

1446 R. Bermejo et al. / Journal of the European Ceramic Society 27 (2007) 1443–1448 Fig. 3. Stress distribution through the layers in the bulk and at the surface of the multilayer, calculated using a 3D finite element model. 3. Results and discussion 3.1. Residual stress distribution Fig. 3 represents the magnitude and distribution profile of the residual stresses within the inner and outer layers of the laminate calculated both at the surface and in the bulk of the specimen, using a 3D finite element model. Compressive stresses seem to present a constant value of −690 MPa through the entire thin layer. For the internal ATZ layers a symmetrical distribu￾tion is obtained reaching a maximum value of 100 MPa at the ATZ/AMZ interfaces. On the other hand, for the outer layers residual stresses decrease near the free surface. These results are in good agreement with those reported on a previous work where residual stress distribution was evaluated by the indenta￾tion technique and analytical solutions.9 3.2. Threshold strength Fig. 4 represents the variation of the failure stress with the square root of the initial flaw size, c, for the indented ATZ monolith and for the laminate investigated. From these results, Fig. 4. Plot of measured four-point bending failure stress vs. square root of crack length in ATZ monolith () and laminate () specimens containing several groups of Vickers indentations at the surface. Fig. 5. Optical micrograph showing the failure of the laminate (indentation load of 50 N) when the threshold stress is reached. The rest of indentation cracks corresponding to different loads placed at the ATZ surface have become through￾thickness cracks and arrested at the AMZ/ATZ interface. several conclusions may be inferred: (i) for the same inden￾tation loads, measured crack lengths are always larger for the laminate, corroborating the presence of tensile residual stresses in the indented ATZ layer, as reported in Fig. 3, (ii) the failure stress values are higher for the laminate for every indentation flaw size, pointing out the enhanced mechanical behaviour in comparison to the ATZ monolith, (iii) the constant value of these failure stresses, regardless of the indentation flaw size, allows to indicate the existence of a threshold strength in the lami￾nate with a value of 167 ± 4 MPa, and (iv) fracture of the ATZ monolith always occurs from the defect originated by the largest indentation load, as it should be expected. However, in the mul￾tilayered material fracture takes place randomly from any of the indentations performed. This may be explained by the fact that, regardless of the indentation load applied in the laminate, the crack shape, a/c, is initially always greater than one, and rises when an external load is applied (as demonstrated experimen￾tally by grinding and polishing procedures), helped by the tensile residual stresses in the ATZ layer, more significant as the inter￾face is approached (Fig. 3). When a is equal to the ATZ layer thickness the crack stops, and then c starts to grow until the crack becomes a through-thickness one. Once that all the indentation cracks have become through-thickness cracks and arrested at the AMZ layer (Fig. 5), the stress intensity factor in any of them is equivalent, regardless of the indentation load used for inducing them. 3.3. Fracture toughness Fracture toughness for the monolithic materials evaluated by the microfracture indentation (IM) and the single-edge V￾notch beam (SEVNB) methods are reported in Table 1. It is interesting to note here that these KIc values may be even over￾estimated because the IM method is affected by the residual stress field inherent to the indentation, while the notch radius for the SEVNB method is not as sharp as a real crack and is greater than the microstructure scale length. As a comparison, the apparent fracture toughness, KIa and KIb, for the multilayered material were 7.1 and 8.1 MPa m1/2, as determined at the first and second AMZ layer, respectively. Table 1 Fracture toughness values for the ATZ and AMZ monoliths as determined by means of the IM and SEVNB methods Material KIc (MPa m1/2) IM SEVNB ATZ 3.5 3.2 AMZ 2.9 2.6

R Bermejo et al. / Journal of the European Ceramic Society 27 (2007)1443-1448 1447 0.16 Laminate ATZ AMZ 012A7z 0.00 0.01 003 0.04 COD (mm) function approach, taking into account the contribution of the residual stresses Fig. 7. Load-COD curves for ATZ monolithic and laminate notched specimen nrough the laminate at a COd rate of I um/s Maximum load for the Atz corresponds to the crac nitiation and consequent fracture of the specimen. For the case of the laminat On the other hand, the apparent fracture toughness, Kapt.c, at the first ATZ/AMZ interface(see micrograph A) and an increase in load determined analytically for the first layers using the weight quired to continue with the crack propagation(see micrograph B). Crack function method, is presented in Fig. 6, showing the R-curve bifurcation and crack deflection can be observed in the fractured specimens(see behaviour of this multilayered system. It can be inferred that micrographs B and C). Shaded area shows the higher work of fracture of the Kapt e decreases while the crack propagates through the ATZten- laminate in comparison to the ATZ monolith sile layers. However, there is a significant R-curve raise when entering a highly compressed AMZ layer, which is the responsi- effect is repeated any time the crack faces a new ATZ/AMZ ble for the crack arrest and the subsequent threshold effect. The interface(see Fig. 7) negative values for Kapt c may be explained by the relatively high The total fracture energy, G, for the ATZ monolith residual tensile stresses present in the ATZ layers. This might the laminate was evaluated through integration of the load-COD notches at some point within the ATZ tensile phase, as discussed registered curve/material volume unit(Fig. 7)resulting in val- the curve associated with the first aTzamz interface is of The experimental values for apparent fracture toughness of 123 J/m2, which is in good agreement with the value for the the laminate are in fair agreement with those predicted by the fracture energy estimated from the experimental fracture tough- analytical model, yet more tests should be conducted to better evaluate the r-curve effect in this material ness at the first interface(Gla=128 J/m; for Ka=7.1 MPam, E=373 GPa and v=f 3. 4. fracture energy The enhanced fracture behaviour of the multilayer inves- tigated in comparison with the ATZ monolith is due to the ive stresses in the thin AMZ layers, which depend The COD tests performed on ATZ monoliths and laminates on the thickness ratio of the layers. In this particular case, allowed to evaluate the work required to break the samples crack bifurcation(Fig. 7B)and crack deflection(Fig. 7) In the case of the monolith, the fracture initiated at the notch are observed. As a consequence, a step-wise behaviour at the tip when reaching a certain crack opening displacement Fail- fracture surface of the laminate is observed mpared to ure was catastrophic as it corresponds to a brittle material. On the flat character exhibited by the aTZ monolith. The exis the other hand, failure in the laminate originated also at the tence of these mechanisms depend, for a given layer thick notch tip but for a smaller stress value(see Fig. 7). It can ness ratio, on the AMZ layer thickness, being possible their be inferred from the recorded data that the first drop in load absence for very thin layers. Further studies are in progress on occurred at a load value below the maximum load reached by this kind of multilayers with various layer thickness ratios to the monolithic sample. This difference is explained by the fact study its effect on the threshold strength and crack propagation that tensile stresses are present in the laminate at the notch mechanisms site increasing the stress intensity factor at that point. Never- theless, the crack initiated at the notch tip does not produce catastrophic failure as in the case of the monolith, since it 4. Conclusions arrests at the first ATZ/AMz interface(see Fig. 7A). To main- tain the crack opening displacement rate during the test a further The fracture behaviour of an Al2O3/Zro2 multilayered mate- increase in load is required. When the crack enters the first rial has been characterized and compared to that of a monolithic ATZAMZ interface, which corresponds to the maximum load alumina, taken as reference. From the results, the following con- in the load/COD curve, a second drop in load is observed. This clusions can be drawn

R. Bermejo et al. / Journal of the European Ceramic Society 27 (2007) 1443–1448 1447 Fig. 6. Apparent fracture toughness calculated analytically using the weight function approach, taking into account the contribution of the residual stresses through the laminate. On the other hand, the apparent fracture toughness, Kapt,c, determined analytically for the first layers using the weight function method, is presented in Fig. 6, showing the R-curve behaviour of this multilayered system. It can be inferred that Kapt,c decreases while the crack propagates through the ATZ ten￾sile layers. However, there is a significant R-curve raise when entering a highly compressed AMZ layer, which is the responsi￾ble for the crack arrest and the subsequent threshold effect. The negative values for Kapt,c may be explained by the relatively high residual tensile stresses present in the ATZ layers. This might also be associated with the spontaneous cracks arising from the notches at some point within the ATZ tensile phase, as discussed in Section 2. The experimental values for apparent fracture toughness of the laminate are in fair agreement with those predicted by the analytical model, yet more tests should be conducted to better evaluate the R-curve effect in this material. 3.4. Fracture energy The COD tests performed on ATZ monoliths and laminates allowed to evaluate the work required to break the samples. In the case of the monolith, the fracture initiated at the notch tip when reaching a certain crack opening displacement. Fail￾ure was catastrophic as it corresponds to a brittle material. On the other hand, failure in the laminate originated also at the notch tip but for a smaller stress value (see Fig. 7). It can be inferred from the recorded data that the first drop in load occurred at a load value below the maximum load reached by the monolithic sample. This difference is explained by the fact that tensile stresses are present in the laminate at the notch site increasing the stress intensity factor at that point. Never￾theless, the crack initiated at the notch tip does not produce catastrophic failure as in the case of the monolith, since it arrests at the first ATZ/AMZ interface (see Fig. 7A). To main￾tain the crack opening displacement rate during the test a further increase in load is required. When the crack enters the first ATZ/AMZ interface, which corresponds to the maximum load in the load/COD curve, a second drop in load is observed. This Fig. 7. Load-COD curves for ATZ monolithic and laminate notched specimens at a COD rate of 1m/s. Maximum load for the ATZ corresponds to the crack initiation and consequent fracture of the specimen. For the case of the laminate, although the first drop in load is also related to crack initiation, crack is arrested at the first ATZ/AMZ interface (see micrograph A) and an increase in load is required to continue with the crack propagation (see micrograph B). Crack bifurcation and crack deflection can be observed in the fractured specimens (see micrographs B and C). Shaded area shows the higher work of fracture of the laminate in comparison to the ATZ monolith. effect is repeated any time the crack faces a new ATZ/AMZ interface (see Fig. 7). The total fracture energy, G, for the ATZ monolith and for the laminate was evaluated through integration of the load–COD registered curve/material volume unit (Fig. 7) resulting in val￾ues of 29 and 178 J/m2, respectively. In particular, the area of the curve associated with the first ATZ/AMZ interface is of 123 J/m2, which is in good agreement with the value for the fracture energy estimated from the experimental fracture tough￾ness at the first interface (GIa = 128 J/m2; for KIa = 7.1 MPa m1/2, E = 373 GPa and υ = 0.22). The enhanced fracture behaviour of the multilayer inves￾tigated in comparison with the ATZ monolith is due to the compressive stresses in the thin AMZ layers, which depend on the thickness ratio of the layers. In this particular case, crack bifurcation (Fig. 7B) and crack deflection (Fig. 7C) are observed. As a consequence, a step-wise behaviour at the fracture surface of the laminate is observed, as compared to the flat character exhibited by the ATZ monolith. The exis￾tence of these mechanisms depend, for a given layer thick￾ness ratio, on the AMZ layer thickness, being possible their absence for very thin layers. Further studies are in progress on this kind of multilayers with various layer thickness ratios to study its effect on the threshold strength and crack propagation mechanisms. 4. Conclusions The fracture behaviour of an Al2O3/ZrO2 multilayered mate￾rial has been characterized and compared to that of a monolithic alumina, taken as reference. From the results, the following con￾clusions can be drawn:

R. Bermejo et al. /Journal of the European Ceramic Sociery 27(2007)1443-1448 (i) The residual stresses developed in the multilayer during 4. Katsuki, H and Hirata, Y, Coat of alumina sheet with needle-like mullite sintering reveal a difference in magnitude and distribution J. Ceram. Soc. Jpn, 1990,98,1114-1119. between external and internal ATZ layers. For the former 5. Green, D.J., Tandon, R and Sglavo, V. M, Crack arrest and multiple crack- residual stresses decrease near the free surface. For the latter ing in glass through the use of designed residual stress profiles. Science, 1999,283.1295-1297. a symmetrical distribution is observed, reaching a maxi- 6. Rao, M P, Sanchez-Herencia, A.I., Beltz, GE, McMeeking, R.M. ane Lange, F F Laminar ceramics that exhibit a threshold strength. Science the mean value of these tensile residual stresses is over 1999,286,102-105 100 MPa whilst for the compressive layers reaches a max 7. Sanchez-Herencia, A. J, Pascual, C, He, J. and Lange, F. F, ZrO/ZrOz imum of -690 MPa layered composites for crack bifurcation. J. Am. Ceram. Soc., 1999,8 1512-1518 (ii) In the multilayered material investigated a threshold 8. Clegg, w.J., Kendall, K, Alford, N. M, Button, T. w. and Birchall, strength of 167+4 MPa is encountered for monotonic J. D, A simple way to make tough ceramics. Na loads applied normal to the layer plane, regardless of the ini- tial flaw size. This behaviour is contrary to the AtZ mono- 9. Bermejo, R. Sanchez-Herencia, A J, Baudin, C. Anglada. M. lith where the failure stress diminishes with the faw size and Llanes. L behaviour of an Al2O3-ZrO2 multilayered ceramic with residual (iii) The fracture toughness of the ATZ monolith (SEVNB) due to phase transformations. Fatigue Fract. Eng- Mater struc 71-78. and laminate is of 3.2 and 7.1 MPam", respectively. 10. Bermejo, R, Llanes, L, Anglada, M, Supancic, P and Lube, T,Ther- The apparent fracture toughness of the laminate structure mal shock behavior of an Al2O3/ZrO2 multilayered ceramic with residual determined experimentally with the SEPB method is in tresses due to phase transformations. Key Eng. Mater., 2005, 290, 19 fair agreement with the values predicted by the weight 11. Fett. T and Munz. D. Evaluation of R-curve effects in ceramics. J. Mater. function analytical procedure Sci.,1993,28,742-752. (iv) The fracture energy of the laminated composite is about 12. Gere, J M and Timoshenko, S P, Mechanics of Materials. Stanley Thornes six times higher than that of the monolithic alumina Ltd.,UK,199,pp.391-404. The mechanism responsible for energy dissipation is the 13. Malzbender, J and Steinbrech, R. w, Mechanical properties of coated ma bifurcation and/or deflection of cracks when they enter the rials and multi-layered composites determined using bending methods. Surf. Coa. Technol.,2004,176,165-172. thin compressive layers showing a step-wise fracture 14. Anstis. G. R. Chantikul. P. Lawn. B. R. and Marshall. D.B.A comparison with the flat character encountered for most ical evaluation of indentation techniques for measuring fracture tou of the brittle materials ness. I. Direct crack measurements. J. An. Ceram. Soc., 1981, 64 These results indicate that this multilayered structure design 15. Kuebler, J, Fracture toughness of ceramics using the SEVNB method: from may be effective for improving the resistance to crack propa- a preliminary study to a standardized method Fracture Resistance Testing of Monolithic and Composite Brittle Materials, ASTM STP 1409. American gation as well as enhancing the reliability of structural ceramic ety for Testing and Materials, West Conshohocken, PA, USA, 2002 components due to the presence of a threshold strength under 16. Damani, R, Gstrein, R and Danzer, R, Critical notch-root radius effect loading axis normal to the layer plane in SENB-S fracture toughness testing. J. Eur Ceram. Soc., 1996, 16, 695- Acknowledgments 17. Tada, H, Paris, P C. and Irwin, G. R, The Stress Analysis of Cracks Hand- book. Paris Productions Incorporated(and Del Research Corporation), St Louis,USA,1973.pp.2.13-2.15 WorksupportedbytheSpanishMinistryofScienceandtech-18.Fett,t.andMunz,D.,StressIntensityFactorandWeightFunctions.com- nology, through grant MAT-2002-00368. Some of the authors tational Mechanics Publications, Elsevier, Southampton, UK, 1997, RB and JP acknowledge the financial support provided through the European Community's Human Potential Programme under 19. Rao, M. P, Rodel, and Lange, F. F Residual stress induced R-curves in contract HPRN-CT-2002-00203, SICMACI laminar ceramics that exhibit a threshold strength. J. Am. Ceram Soc. 2001 84.2722-2724. 20. Moon. R. Hoffman. M. Bowman, J. H. Trumble, K. and Rodel References J, A weight function analysis on the R-curve behaviour of multilay ered alumina-zirconia composites. J. A. Ceram Soc., 2002, 85, 1505- 1. Virkar, A. V, Huang, J. L and Cutler, R.A., Strengthening of oxide ceramics by transformation-induced stresses. J. Am. Ceram Soc., 1987, 70, 164-170. 21. Lugovy, M, Slyunyayev, V, Orlovskaya, N, Blugan, G, Kuebler, J and 2. Requena, J, Moreno, R and Moya, J.S., Alumina and alumina/zirconia mul- Lewis, M, Apparent fracture toughness of Si3N4-based laminates with tilayer composites by slip casting. J. Am. Ceram Soc., 1989, 72, 1511-1513 residual compressive or tensile stresses in surface layers. Acta Mater, 2005 3. Marshall, D. B, Ratto. J J and Lange F. F. Enhanced fracture toughness in 53.289-296. layered microcomposites of Ce-ZrOz and Al2 O3. J Am Ceram Soc., 1991, 22. Tattersall, H G and Tappin, G The work of fracture and its measurements

1448 R. Bermejo et al. / Journal of the European Ceramic Society 27 (2007) 1443–1448 (i) The residual stresses developed in the multilayer during sintering reveal a difference in magnitude and distribution between external and internal ATZ layers. For the former, residual stresses decrease near the free surface. For the latter a symmetrical distribution is observed, reaching a maxi￾mum value close to the ATZ/AMZ interfaces. In the bulk, the mean value of these tensile residual stresses is over 100 MPa whilst for the compressive layers reaches a max￾imum of −690 MPa. (ii) In the multilayered material investigated a threshold strength of 167 ± 4 MPa is encountered for monotonic loads applied normal to the layer plane, regardless of the ini￾tial flaw size. This behaviour is contrary to the ATZ mono￾lith where the failure stress diminishes with the flaw size. (iii) The fracture toughness of the ATZ monolith (SEVNB) and laminate is of 3.2 and 7.1 MPa m1/2, respectively. The apparent fracture toughness of the laminate structure determined experimentally with the SEPB method is in fair agreement with the values predicted by the weight function analytical procedure. (iv) The fracture energy of the laminated composite is about six times higher than that of the monolithic alumina. The mechanism responsible for energy dissipation is the bifurcation and/or deflection of cracks when they enter the thin compressive layers showing a step-wise fracture in comparison with the flat character encountered for most of the brittle materials. These results indicate that this multilayered structure design may be effective for improving the resistance to crack propa￾gation as well as enhancing the reliability of structural ceramic components due to the presence of a threshold strength under loading axis normal to the layer plane. Acknowledgments Work supported by the Spanish Ministry of Science and Tech￾nology, through grant MAT-2002-00368. Some of the authors RB and JP acknowledge the financial support provided through the European Community’s Human Potential Programme under contract HPRN-CT-2002-00203, [SICMAC]. References 1. Virkar, A. V., Huang, J. L. and Cutler, R. A., Strengthening of oxide ceramics by transformation-induced stresses. J. Am. Ceram. Soc., 1987, 70, 164–170. 2. Requena, J., Moreno, R. and Moya, J. S., Alumina and alumina/zirconia mul￾tilayer composites by slip casting. J. Am. Ceram. Soc., 1989, 72, 1511–1513. 3. Marshall, D. B., Ratto, J. J. and Lange, F. F., Enhanced fracture toughness in layered microcomposites of Ce-ZrO2 and Al2O3. J. Am. Ceram. Soc., 1991, 74, 2979–2987. 4. Katsuki, H. and Hirata, Y., Coat of alumina sheet with needle-like mullite. J. Ceram. Soc. Jpn., 1990, 98, 1114–1119. 5. Green, D. J., Tandon, R. and Sglavo, V. M., Crack arrest and multiple crack￾ing in glass through the use of designed residual stress profiles. Science, 1999, 283, 1295–1297. 6. Rao, M. P., Sanchez-Herencia, A. J., Beltz, G. E., McMeeking, R. M. and ´ Lange, F. F., Laminar ceramics that exhibit a threshold strength. Science, 1999, 286, 102–105. 7. Sanchez-Herencia, A. J., Pascual, C., He, J. and Lange, F. F., ZrO2/ZrO2 layered composites for crack bifurcation. J. Am. Ceram. Soc., 1999, 82, 1512–1518. 8. Clegg, W. J., Kendall, K., Alford, N. M., Button, T. W. and Birchall, J. D., A simple way to make tough ceramics. Nature, 1990, 347, 455– 461. 9. Bermejo, R., Torres, Y., Sanchez-Herencia, A. J., Baud ´ ´ın, C., Anglada, M. and Llanes, L., Fracture behaviour of an Al2O3–ZrO2 multilayered ceramic with residual stresses due to phase transformations. Fatigue Fract. Eng. Mater. Struct., 2006, 29, 71–78. 10. Bermejo, R., Llanes, L., Anglada, M., Supancic, P. and Lube, T., Ther￾mal shock behavior of an Al2O3/ZrO2 multilayered ceramic with residual stresses due to phase transformations. Key Eng. Mater., 2005, 290, 191– 198. 11. Fett, T. and Munz, D., Evaluation of R-curve effects in ceramics. J. Mater. Sci., 1993, 28, 742–752. 12. Gere, J. M. and Timoshenko, S. P., Mechanics of Materials. Stanley Thornes Ltd., UK, 1999, pp. 391–404. 13. Malzbender, J. and Steinbrech, R. W., Mechanical properties of coated mate￾rials and multi-layered composites determined using bending methods. Surf. Coat. Technol., 2004, 176, 165–172. 14. Anstis, G. R., Chantikul, P., Lawn, B. R. and Marshall, D. B., A crit￾ical evaluation of indentation techniques for measuring fracture tough￾ness. I. Direct crack measurements. J. Am. Ceram. Soc., 1981, 64, 533– 538. 15. Kuebler, J., Fracture toughness of ceramics using the SEVNB method; from a preliminary study to a standardized method. Fracture Resistance Testing of Monolithic and Composite Brittle Materials, ASTM STP 1409. American Society for Testing and Materials, West Conshohocken, PA, USA, 2002. 16. Damani, R., Gstrein, R. and Danzer, R., Critical notch-root radius effect in SENB-S fracture toughness testing. J. Eur. Ceram. Soc., 1996, 16, 695– 702. 17. Tada, H., Paris, P. C. and Irwin, G. R., The Stress Analysis of Cracks Hand￾book. Paris Productions Incorporated (and Del Research Corporation), St. Louis, USA, 1973, pp. 2.13–2.15. 18. Fett, T. and Munz, D., Stress Intensity Factor and Weight Functions. Com￾putational Mechanics Publications, Elsevier, Southampton, UK, 1997, p. 408. 19. Rao, M. P., Rodel, J. and Lange, F. F., Residual stress induced ¨ R-curves in laminar ceramics that exhibit a threshold strength. J. Am. Ceram. Soc., 2001, 84, 2722–2724. 20. Moon, R., Hoffman, M., Bowman, J. H., Trumble, K. and Rodel, ¨ J., A weight function analysis on the R-curve behaviour of multilay￾ered alumina-zirconia composites. J. Am. Ceram. Soc., 2002, 85, 1505– 1511. 21. Lugovy, M., Slyunyayev, V., Orlovskaya, N., Blugan, G., Kuebler, J. and Lewis, M., Apparent fracture toughness of Si3N4-based laminates with residual compressive or tensile stresses in surface layers. Acta Mater., 2005, 53, 289–296. 22. Tattersall, H. G. and Tappin, G., The work of fracture and its measurements in metals, ceramics and other materials. J. Mater. Sci., 1966, 1, 296–301