Microstructure/Properties Relations of Advanced Materials ournal . Am Ceram. So, 80 [7] 1677-83 (1997) Debonding in Multilayered Composites of Zirconia and Lapo4 David B. Marshall, Peter E D Morgan, and Robert M. Housley Rockwell Science Center, Thousand Oaks, Califomia 91360 Multilayered composites consisting of LaPO. (La-monazite) I. Experimental Procedure layers alternating with various Zro,based materials were fabricated to investigate whether Lapo, provides a weakly Laminar composites consisting of alternating layers of bonded interface suitable for promoting toughening, as pre- LaPO, and zirconia were fabricated by colloidal techniques viously observed in the system LaPO./Al,O,. The following Four types of composites containing different compositions in Zro,based materials were assessed: Y-zrO2, Y-Zro the"zirconia"layers were fabricated: one with 3 mol%Y,O AL,O,, Ce-ZrO2, and Ce-ZrO, / AL,O3. Debonding was addition (Tosoh 3Y); one with 12 mol%CeO2 addition(Tosoh observed in all cases. The composites containing Yzro, 12Ce); one with a mixture(50% by volume)of Al2 0, (Sumi and Y-Zro2 /Al,O, were stable, with no reactions, at tem. tomo AKP30)and Y-ZrO2; and one with a mixture of Al2O3 (50% by volume)and Ce-zrO2. The layer thicknesses were in peratures up to at least 1600C. However, in the composites the range 5 to 50 um for the LaPO layers and 20 to 200 um for containing Ce-ZrO2, interdiffusion of Ce and La occurred, resulting in formation of a pyrochlore like phase and, in the Zro2 layers The composites were consolidated by sequential centrifuging the case of the Ce-rO, AL,O, composite, a(Ce, LaAl Ois or vacuum slip casting of colloidal suspensions of the various se was made of a colloidal technique developed by Velamakanni et d velamakannia in which an I. Introduction aqueous electrolyte (NH.NO,) was used (after dispersing the wders at pH 2) to produce short-range repulsive hydration SF EVERAL mechanisms have been used to achieve toughen forces and reduce the magnitudes of the longer range electro- ing in multilayered ceramic composites. One involves static forces between the suspended particles, Such conditions enhanced transformation toughening in systems containing zir weakly attractive network of particles, which pre conia with strongly bonded interfaces, While large increases vents mass segregation during centrifugation and slip casting in fracture toughness een achieved in such composites but because of the lubricating action of the short-range repul failure in tensile loading involves growth of a single crack: sive forces, allows particles to pack to high green density the composites do not exhibit a large nonlinear behavior of The prevention of mass segregation is especially important for distributed damage. Other mechanisms involve deflection of forming uniform two-phase layers such as Al,O, Zro, from cracks or secondary cracking caused by the presence of weak suspensions containing mixtures of the two types of powder ral defects. -If the degree of crack deflection is sufficient to 350 MPa, then sintered in air (in the case of Y-containing cause splitting between layers, a damage-tolerant response and composites)or oxygen( Ce-containing composites )at 1600oC distributed damage can be achieved in flexural loading, as dem- for 2 h. Several specimens containing Y-ZrO2 were packed in onstrated initially in systems containing layers of carbon alumina powder after the cold-pressing step and hot-pressed in BN -However, the usefulness of systems containing carbon graphite dies at 1400'C for 1 h. Sections were cut, polished and bn is limited by their sensitivity to oxidation. and thermally etched at 1400'C for microstructural analysis Attempts have been made recently to avoid this limitation by Microstructural characteristics of the multilayered compos- developing analogous oxide-based systems, using either layers ites were assessed using X-ray diffraction, electron microprobe SEM, secondary electrons, backscattered electrons, and cathod (La-monazite)and YPO, (xenotime)to cause debonding. In the luminescence were used for imaging, and elements were identi ystem LaPO.,O3, the interfacial toughness is sufficiently fied by energy-dispersive spectroscopy (EDS)of fluorescence X-rays. The response of the cathodoluminescence detector was mally incident crack to deflect along the interface rather than strongly peaked in the blue wavelengths. The electron micro- cross it.This system is stable for long periods in air at temper- probe measurements were taken at an operating voltage of 15kV atures at least as high as 1600oC. 3. I4 In this paper, we preser preliminary study of the stability and debonding characteristics cerium oxide (Electron microprobe measurements were done at logy Department at Caltech, using a JEOL JXA-733 of achieving toughening from both interlaminar cracking and instrument. In the various imaging modes the following phase transformations is also explored ndary electrons, alumina was darker than zirconia and mona- Ite. while zirconia in cathodoluminescence, ( La, Ce)PO, was bright, while Al David J. Green-contributing editor and zro, did not luminesce Debonding at the lapo layers was tested using two methods One involved forming cracks in a polished surface(normal to the ayers by indenting with a vickers indenter. the other involved loading rectangular beams in four-point bending and monitoring crack growth in situ using an optical microscope. The beams Numbers were oriented with the layers normal to the loading direction, so Member, American Ceramic Soc that initial crack growth was normal to the layers. Some beams 1677
f.:' .. ':,:.,);,,?,yii&.is ., 1,:... ._ Microstructure/Properties Relations of Advanced Materials J ~m c.mnr soc, M) 171 1677-83 (im Debonding in Multilayered Composites of Zirconia and LaPO, David B. Marshall,. Peter E. D. Morgan,' and Robert M. Housley. Rockwell Science Center, Thousand Oaks, California 91360 Multilayered composites consisting of LaPO, (La-monazite) layers alternating with various Zr0,-based materials were fabricated to investigate whether LaPO, provides a weakly bonded interface suitable for promoting toughening, as previously observed in the system LaPO, /Al,O,. The following Zr0,-based materials were assessed: Y-ZrO,, Y-ZrO,/ A1203, Ce-ZrO,, and Ce-ZrO,/Al,O,. Debonding was observed in all cases. The composites containing Y-ZrO, and Y-ZrO,/Al,O, were stable, with no reactions, at temperatures up to at least 1600°C. However, in the composites containing Ce-ZrO,, interdiffusion of Ce and La occurred, resulting in formation of a pyrochlore-like phase and, in the case of the Ce-ZrO,/Al,O, composite, a (Ce,La)Al,,O,, magnetoplumbite phase. I. Introduction EVERAL mechanisms have been used to achieve toughen- S ing in multilayered ceramic composites.' One involves enhanced transformation toughening in systems containing zirconia with strongly bonded interfaces." While large increases in fracture toughness have been achieved in such composites, failure in tensile loading involves growth of a single crack: the composites do not exhibit a large nonlinear behavior or distributed damage. Other mechanisms involve deflection of cracks or secondary cracking caused by the presence of weak layers, weak interfaces, residual stresses, or other microstructural If the degree of crack deflection is sufficient to cause splitting between layers, a damage-tolerant response and distributed damage can be achieved in flexural loading, as demonstrated initially in systems containing layers of carbon or BN."' However, the usefulness of systems containing carbon and BN is limited by their sensitivity to oxidation. Attempts have been made recently to avoid this limitation by developing analogous oxide-based systems, using either layers of magnetoplumbite/P-alumina cornpo~nds,'-''-'~ porous oxide layers,I2 or rare-earth orthophosphates'3-16 such as LaPO, (La-monazite) and YPO, (xenotime) to cause debonding. In the system LaPo,-Al,O,, the interfacial toughness is sufficiently low to satisfy the criterion of He and Hutchinson" for a normally incident crack to deflect along the interface rather than cross it.I3 This system is stable for long periods in air at temperatures at least as high as 1600°C.13*'4 Inthis paper, we present a preliminary study of the stability and debonding characteristics of LaPO, with several layered zirconia systems. The possibility of achieving toughening from both interlaminar cracking and transformations is also explored. David J. Green-contributing editor Manuscript No. 191602.,Received August 15,1996; approved February 10,1997. Resented at a Symposium on Microstruc~Ro rty Relations of Advanced Materials, a symposium in honor of Professor Arthur R"uefs 60th birthday, held at Max-Planck-Institut fllr Metallforschun , Stuttgart. Germany. April 29-30,1996. Suppomd by the Air Force Office ot@Scientific Research under Coneact Numbers F49620-92-C-0028 and F49620-96-C-0026 monitored by Dr. A. Pechenik. 'Member. American Ceramic Society. II. Experimental Procedure Laminar composites consisting of alternating layers of LaPO, and zirconia were fabricated by colloidal techniques. Four types of composites containing different compositions in the "zirconia" layers were fabricated: one with 3 mol% Y203 addition (Tosoh 3Y); one with 12 mol% CeO, addition (Tosoh 12Ce); one with a mixture (50% by volume) of A1203 (Sumitom0 AKP30) and Y-Zro,; and one with a mixture of A1,0, (50% by volume) and Ce-ZrO,. The layer thicknesses were in the range 5 to 50 p,m for the LaPO, layers and 20 to 200 pm for the Zro, layers. The composites were consolidated by sequential centrifuging or vacuum slip casting of colloidal suspensions of the various layers. Use was made of a colloidal technique developed by Velamakanni er al." and Chang and VelamakanniZ2 in which an aqueous electrolyte (NH,NO,) was used (after dispersing the powders at pH 2) to produce short-range repulsive hydration forces and reduce the magnitudes of the longer range electrostatic forces between the suspended particles. Such conditions produce a weakly attractive network of particles, which prevents mass segregation during centrifugation and slip casting, but because of the lubricating action of the short-range repulsive forces, allows particles to pack to high green density. The prevention of mass segregation is especially important for forming uniform two-phase layers such as Al,03-Zro, from suspensions containing mixtures of the two types of powder. The consolidated bodies were cold isostatically pressed at 350 MPa, then sintered in air (in the case of Y-containing composites) or oxygen (Ce-containing composites) at 1600OC for 2 h. Several specimens containing Y-ZrO, were packed in alumina powder after the cold-pressing step and hot-pressed in graphite dies at 1400°C for 1 h. Sections were cut, polished, and thermally etched at 1400OC for microstructural analysis. Microstructural characteristics of the multilayered composites were assessed using X-ray diffraction, electron microprobe analysis, and scanning electron microscopy (SEM). In the SEM, secondary electrons, backscattered electrons, and cathodoluminescence were used for imaging, and elements were identified by energy-dispersive spectroscopy (EDS) of fluorescence X-rays. The response of the cathodoluminescence detector was strongly peaked in the blue wavelengths. The electron microprobe measurements were taken at an operating voltage of 15 kV using standards of lanthanum phosphate, zirconium silicate, and cerium oxide. (Electron microprobe measurements were done at the Geology Department at Caltech, using a JEOL JXA-733 instrument.) In the various imaging modes the following phase distinctions could be readily made: with backscattered and secondary electrons, alumina was darker than zirconia and monazite, while zirconia and monazite were almost indistinguishable; in cathodoluminescence, (La,Ce)PO, was bright, while A1203 and ZrO, did not luminesce in the blue and were indistinguishable. Debonding at the LaPO, layers was tested using two methods. One involved forming cracks 1z1 a polished surface (normal to the layers) by indenting with a Vickers indenter. The other involved loading rectangular beams in four-point bending and monitoring crack growth in sifu using an optical microscope.. The beams were oriented with the layers normal to the loading direction, so that initial crack growth was normal to the layers. Some beams 1677
1678 Journal of the American Ceramic Society--Marshallet al Vo.80,No.7 LaPO4 a my 1. Layers of LaPO.(La-monazite)and Y-ZrO2/Al,, Backscat 10 um electron image Thermally etched surface (La, ce)PO4 were tested with a notch of width 200 um, cut with a diamond saw normal to the layers on the tensile side of the beam (b) III. Results ()Microstructures The composites containing Y-ZrO2 and Y-ZrO2/Al,O3 layers did not exhibit any reactions or interdiffusion with the Lapo layers. In composites that were hot-pressed at 1400.C, both layers appeared fully dense on polished cross sections. a back- scattered electron image showing 5 um layers of LaPO4 and Y-ZrO2/AlO, in a hot-pressed composite is shown in Fig. 1 within the ZrO,/Al,O, layers the grain size is. 7 um, with the ZrO2 and AlO, being well dispersed, while the grain size in the LapO, layer is -3 um. In the composites that were sintered at 1600.C, the ZrO2-containing layers were fully dense, although the lapO4 layers contained isolated pores with diameters up to 2 um(see Fig. 7). The grain sizes in both layers are larger than in the hot-pressed composites: 1.5 um for the AlO -4 to 5 um In the composite containing Ce-zrOz layers, there was inter diffusion of La and Ce and precipitation of a second phase within the Ce-Zro2 layers(Fig. 2). However, X-ray diffraction measurements indicated that the composite was primarily zirco- nia and monazite, with only a few small additional diffraction Fig. 2.(a)Cathod image of Ce-ZrO,/LaPO. com- aks that were consistent with a few percent of a pyrochlore. ike structure. The electron microprobe results(Fig. 3)indicate (b) Backscattered rie tayers are(la,Ce)PO. that the Lapo, layers had lost% to 10% of the La which Porous layers are single-phase(La, Ce)PO, had been replaced by Ce, forming the monazite solid solution a-ce, PO4. In the zirconia layers, the la was in precipitates which also contained much higher concentration of Ce than the (2) Fracture and Debonding Fig. 2(b)and are possibly a pyrochlore-like phase. The l0 o host Ce-ZrO (A) Flexural loading In all of the composites, fexural loading of notched Canerain ing tayers are uly sens hah iw tse obs ona zihe causee de lamieaeine tid respentia stich as sho th Ia yes La-monazite layers of the Y-ZrO2 composite(Fig. 2(b) similar response was observed with notch-free beams(Fig. 5(b)) The composite containing Ce-ZrO2/Al2O, layers showed sub- in all cases except for the composite containing Ce-ZrO2/Al203 stantially more reaction associated with the diffusion of La and layers: in that case failure was catastrophic without debonding Ce. In the centers of the Ce-ZrO2/Al,0, layers there were iso- The typical delamination damage can be seen in Fig. 6, which lated elongated grains of Ce, La)Al, O,B magnetoplumbite con- shows a broken beam of the Ce-zrO2/LaPO4 composite, alon taining approximately equal amounts of Ce and La(Fig. 4) with an in situ optical micrograph taken during loading and SEM Near the edges of the layers there were larger concentrations of micrographs of the stepped fracture surface after failure. In the magnetoplumbite and some LaAIO3 grain in situ micrograph of Fig. 6(b), extensive damage is evident in
1678 Journal of the American Ceramic Society-Marshall et al. Vol. 80, No. 7 Fig. 1. Layers of LaPO, (La-monazite) and Y-ZrO,/AI,O,. Backscattered electron image. Thermally etched surface. were tested with a notch of width 200 pm, cut with a diamond saw normal to the layers on the tensile side of the beam. III. Results (I) Microstructures The composites containing Y-Zro, and Y-ZrO,/Al,O, layers did not exhibit any reactions or interdiffusion with the LaPo, layers. In composites that were hot-pressed at 1400°C, both layers appeared fully dense on polished cross sections. A backscattered electron image showing 5 pm layers of LaPo, and Y-ZrO,/AI,O, in a hot-pressed composite is shown in Fig. 1: within the ZroJAl,O, layers the grain size is -0.7 pm, with the Zro, and A1,0, being well dispersed, while the grain size in the LaPO, layer is -3 pm. In the composites that were sintered at 1600°C, the ZrO,-containing layers were fully dense, although the LaPo, layers contained isolated pores with diameters up to 2 pm (see Fig. 7). The grain sizes in both layers are larger than in the hot-pressed composites: 1.5 pm for the A1203/Zr02 and -4 to 5 pm for the monazite. In the composite containing Ce-ZrO, layers, there was interdiffusion of La and Ce and precipitation of a second phase within the Ce-ZrO, layers (Fig. 2). However, X-ray diffraction measurements indicated that the composite was primarily zirconia and monazite, with only a few small additional diffraction peaks that were consistent with a few percent of a pyrochlorelike structure. The electron microprobe results (Fig. 3) indicate that the LaPo, layers had lost -5% to 10% of the La which had been replaced by Ce, forming the monazite solid solution La,-,Ce,PO,. In the zirconia layers, the La was in precipitates, which also contained much higher concentration of Ce than the host Ce-ZrO, (Fig. 3). These precipitates show as bright spots in Fig. 2(b) and are possibly a pyrochlore-like phase. The ZrO,- containing layers are fully dense, while the (La,Ce)-monazite layers contain more porosity than was observed in the La-monazite layers of the Y-Zro, composite (Fig. 2(b)). The composite containing Ce-Zro,/Al,O, layers showed substantially more reaction associated with the diffusion of La and Ce. In the centers of the Ce-Zr02/A1,0, layers there were isolated elongated grains of (Ce,La)Al, ,O,, magnetoplumbite containing approximately equal amounts of Ce and La (Fig. 4). Near the edges of the layers there were larger concentrations of magnetoplumbite and some LaAlO, grains. Fig. 2. (a) Cathodoluminescence image of Ce-ZrOJLaPO, composite showing overall structure. Bright layers are (La,Ce)PO,. (b) Backscattered electron image of Ce-ZrOJLaPO, composite. Porous layers are single-phase (La,Ce)PO,. (2) Fracture and Debonding (A) Flexural Loading In all of the composites, flexural loading of notched beams caused delamination and sequential fracture of the layers, with a nonlinear load-deflection response such as shown in Fig. 5(a). A similar response was observed with notch-free beams (Fig. 5(b)) in all cases except for the composite containing Ce-ZrO2/A1,0, layers: in that case failure was catastrophic without debonding. The typical delamination damage can be seen in Fig. 6, which shows a broken beam of the Ce-ZrO,/LaFQ, composite, along with an in situ optical micrograph taken during loading and SEM micrographs of the stepped fracture surface after failure. In the in situ micrograph of Fig. 6(b), extensive damage is evident in
July 1997 Debonding in Multilayered Composites of zirconia and LaPO, 1679 beeeeboogoogoo0x sition (um) e-ZrO2/LaPO, comp e first six LapO, layers past the notch, whereas the ZrO2 layers has also been observed in LapO/Al,O in this region are fractured in only one or two places. At this in other material under shear stage of loading there was no damage in any layers beyond the glass matrix 品 sixth Lapo layer On the stepped fracture surface( Figs. 6(c)and adhesive joints (d )) cracking in the LaPO4 layers is seen primarily at or near the nterface between the LaPO4 and ce-ZrO, layers A preliminary observation suggestive of possible combined milar results were obtained for the composite consisting of toughening by ic transformation and interlaminar layers of Y-zrO2/Al,O, and LaPO4. Although on the fracture cracking in the PO, composite is shown in Fig 8 surfaces it appeared that separation of the layers occurred mainl The nomarski micrograph of Fig 8(a)shows a zone at or near the interfaces between the ZrO, and LaPO4, there was of uplifted material surrounding the main crack. The uplift is also substantial cracking within the LaPO layers( Fig. 7). Usu- due to the volume increase associated with the tetragonal-to- ally delamination began with simple deflection of the main crack. monoclinic phase transformation in the Ce-zrO, layers. The as in Fig. 7(a)but then continued by forming an array of echelon width of this zone increases as the crack extends from the notch cracks as in Figs. 7(b)and(c). This mechanism of delamination root, as observed previously in multilayered composites of Al2 O3 ind Ce-zr02 in which the layers were strongly bonded. How- ever, in this case there is also cracking in the LapOa layers within the transformation zone, as shown in Fig 8(b) all of the multilayered composites, the LaPO lay effective in confining the cracking due to adjacent vickers inden- tations, as shown in Figs. 9(a)and(b). The damage in the layers echelon cracks, similar to the damge observed in notched beams Cracks produced by vickers indentations located directly on LapO4 layers were used to obtain a measure of the interlaminar toughness, as shown in Fig. 9(c). The area of Fig. 9(c)lies within a beam of 4 mm thickness that consisted of Y-ZrO2/A1,O, with three layers of LapO, in the center. The lengths of the indentation times the length of the crack growing normal to the layer into the Y-ZrO2/Al, matrix. The corresponding fracture toughnesses calculated from 10 indentations using the analysis of Anstis l26are=8±4Jm2 for the interlaminar crack and T 80 +5J/m for the Y-ZrO2/AlO, matrix(using a value of 300 GPa for the elastic modulus). The matrix toughness is consistent with values reported in the literature(corresponding to a critical stress intensity factor of K. =5 MPa"2); while the interlami- nar toughness is similar to values reported for monazite itself. 3 For the composite with Y-zrO, matrix, the fracture energies were 1 um In=110±10J/m This toughness calculation is based on an analysis for a homog Fig. 4. Backscattered electron image from composite containi eneous isotropic material, whereas the measurements were obtained from composites containing layers of differing elastic Al, O, layer, showing presence of elongated Ce-La magnetoplumbite moduli (133 GPa for LapO4, 200 GPa for ZrO2, and 400 GP Thermally etched surface for AL2O3). In the configuration of Fig. 9, with a thin layer of
July 1997 Debonding in Multilayered Composites of Zirconia and LaPO, 1679 0 50 100 Position (pm) 150 Fig. 3. Electron microprobe measurements along a line traversing several layers in Ce-ZrOJLaPO, composite. Atomic proportions are normalized to four oxygen atoms. the first six LaPO, layers past the notch, whereas the ZrOz layers in this region are fractured in only one or two places. At this stage of loading there was no damage in any layers beyond the sixth LaPO, layer. On the stepped fracture surface (Figs. 6(c) and (d)), cracking in the LaPO, layers is seen primarily at or near the interface between the LaPO, and Ce-ZrO, layers. Similar results were obtained for the composite consisting of layers of Y-ZrO,/Al,O, and LaPo,. Although on the fracture surfaces it appeared that separation of the layers occurred mainly at or near the interfaces between the ZrO, and LaPO,, there was also substantial cracking within the LaPo, layers (Fig. 7). Usually delamination began with simple deflection of the main crack, as in Fig. 7(a) but then continued by forming an array of echelon cracks as in Figs. 7(b) and (c). This mechanism of delamination Fig. 4. Backscattered electron image from composite containing Ce-ZrOJAI20, and LaPO, layers, from region in center of Ce-ZrO,/ A1,0, layer, showing presence of elongated Ce-La magnetoplumbite. Thermally etched surface. for A1,0,). In the configuration of Fig. 9, with-a thin layer of has also been observed in LaP04/A1,03 composite^,'^ as well as in other material systems under shear loading, including unidirectionally reinforced glass matrix cornp0sites,2~~~~ brittle adhesive joints between rigid and graphite-epoxy composite~.~~.~’ A preliminary observation suggestive of possible combined toughening by martensitic transformation and interlaminar cracking in the Ce-ZrO,/LaPO, composite is shown in Fig. 8. The Nomarski interference micrograph of Fig. 8(a) shows a zone of uplifted material surrounding the main crack. The uplift is due to the volume increase associated with the tetragonal-tomonoclinic phase transformation in the Ce-Zro, layers. The width of this zone increases as the crack extends from the notch root, as observed previously in multilayered composites of A1,0, and Ce-Zr0,24 in which the layers were strongly bonded. However, in this case there is also cracking in the LaPO, layers within the transformation zone, as shown in Fig. 8(b). (B) Indentation Fracture In all of the multilayered composites, the LaPo, layers were effective in confining the cracking due to adjacent Vickers indentations, as shown in Figs. 9(a) and (b). The damage in the layers at either side of the indentation consists mostly of arrays of echelon cracks, similar to the damge observed in notched beams. Cracks produced by Vickers indentations located directly on LaPo, layers were used to obtain a measure of the interlaminar toughness, as shown in Fig. 9(c). The area of Fig. 9(c) lies within a beam of 4 mm thickness that consisted of Y-ZrO,/Al,O, with three layers of LaPo, in the center. The lengths of the indentation cracks growing along the LaPo, layer in Fig. 9(c) are about 3 times the length of the crack growing normal to the layer into the Y-Zro, /A1,03 matrix. The corresponding fracture toughnesses calculated from 10 indentations using the analysis of Anstis et aLZ8 are I: = 8 ? 4 J/m2 for the interlaminar crack and r, = 80 ? 5 J/mz for the Y-ZrO,/Al,O, matrix (using a value of 300 GPa for the elastic modulus). The matrix toughness is consistent with values reported in the literature (corresponding to a critical stress intensity factor of K, = 5 MPa.m”2); while the interlaminar toughness is similar to values reported for monazite itself,’3 For the composite with Y-Zro, matrix, the fracture energies were I: = 6 2 3 J/m2, and r, = 110 2 10 J/m2. This toughness calculation is based on an analysis for a homogeneous isotropic material, whereas the measurements were obtained from composites containing layers of differing elastic moduli (133 GPa for LaPO,, 200 GPa for Zro,, and 400 GPa
1680 Journal of the American Ceramic Society-Marshall et al Vol. 80. No. 7 DISPLACEMENT (um) DISPLACEMENT (um) B (a)Load-deflection response of notched beam of Y-Zr02 containing three layers of LaPO. near its midplane. (b) Flexural loading response notched beams of multilayered Ce-ZrO2/LaPO4.(Nominal stress is calculated for undamaged linear elastic beam. LaPO, sandwiched between relatively thick layers of ZrO2, the The calculated toughnesses are also by residual calculation of the crack driving force for a crack growing along stresses due to thermal expansion mismat layered com- the layer(either in the layer or at the interface)in terms of the posites. Since the thermal expansion co residual indentation stress field is not affected by the lower ZrO O2 are approximately affect the crack driving force only through its influence on the those two phase sibly small in composites containing only elastic modulus of the lapo,29 The lower agnitude of the indentation stresses, as determined by the Al2O3/Zr02 layers, the lower thermal expansion coefficient of (b) mm 200mm Ce-ZrO Fig. 6. Fracture of notched beam of multilayered composite of Ce-ZrO2 and LaPO.(a)Overall view after failure. (b)In situ optical micrograph during loading. (c)and (d) fracture surface after failure(arrows indicate cracks
400 , . , . , , . . , , . . . . . . . . . , 300. \ \ \ w: \\\\/: \ \ '. ~'""'...' LaPO, sandwiched between relatively thick layers of Zro,, the calculation of the crack driving force for a crack growing along the layer (either in the layer or at the interface) in terms of the residual indentation stress field is not affected by the lower elastic modulus of the LaPO,?g The lower modulus would affect the crack driving force only through its influence on the magnitude of the indentation stresses, as determined by the inelastic deformation around the indentation. This effect must be small if the indentation is large compared with the monazite layer thickness, as in Fig. 9(c). 200 il i3 1 B 0 loo The calculated toughnesses are also affected by residual stresses due to thermal expansion mismatch in the layered composites. Since the thermal expansion coefficients for LaPo, and Zro, are approximately equal (-10 X 0C-1),'3 residual stresses are negligibly small in composites containing only those two phases. However, in composites containing mixed Al,O,/ZrO, layers, the lower thermal expansion coefficient of Al,O, (-8 X "C-') leads to tensile stresses within the LaPO, layers in the direction parallel to the layers and balancing compressive stresses in the Al,O,/ZrO, (the latter stress Fig. 6. Fracture of notched beam of multilayered composite of Ce-Zro, and LaPo,. (a) Overall view after failure. (b) In siru optical micrograph during loading. (c) and (d) Fracture surface after failure (arrows indicate cracks)
July 1997 Debonding in Multilayered Composites of zirconia and LaPO, 81 Y:Zro/Al,O 20 pm. 500gm ALo,Zro LaPo Ce-zro2 Cohesive 20 Fig. 8. Broken notched beam of multilayered composite of Ce-ZrO Fracture of notched beam of multilayered composite of Y-ZrO2 / Al,O, and LaPO4(a) Initial debonding after crack from notch interference micrograph show- was grown to first LaPo position -200 um to the right of (a).(c)Schematic of debonding (b)SEM micrograph showing mechanism cracking within LaPO, layers in uplifted region adjacent to main crack being negligibly small in the specimen of Fig. 9(c), which contains a small volume fraction of LaPO, layers). In regions near the polished cross-sectional surface, there are also residual The results of the previous section indicate that, when stresses in the direction normal to the \a y tme magnitudes of th Ce-stabilized zirconia and LapO a were in contact at 1600"C, to the corresponding parallel stresses. 0 some of the Ce diffused into the LaPO4, displacing La. The normal stresses are maximum at the surface (equal to the mag- displaced La reacted with the zirconia to form a precipitate nitudes of the parallel stresses) and decay within a depth phase containing La, Ce, and Zr, possibly a pyrochlore-like approximately equal to the layer thickness. Therefore, the structure. When Al2O, was present also in the mixed Al,O, growth of the cracks along the LaPo. layer of Fig. 9(c)is Ce- ZrO2 layers, the displaced La formed a(La, Ce) opposed by the compressive near-surface stress within the plumbite as well as LaAlO, In view of these results, LaPO layer. The magnitude of this stress is approximately 130 that CePOa, which has the same monazite structure MPa. However, since the depth over which the stress acts is and has now been found to be phase-compatib small compared with the crack depth, the influence on the crack Ce-ZrO2, would be a more suitable debonding phase for ngth is not large: application of the analysis of Lawn and CeO2-stabilized zrO, nding overestimate of the In contrast, the results indicate that Y, which forms a different interlaminar toughness for the Al,,/ZrO, composites is no (but related) phosphate structure, xenotime, does not diffuse more than 10%0 nto laPo. at 1600.C, and La does not diffuse out of the lapo
July 1997 Debonding in Multilayered Composites of Zirconia and LdO, 1681 Fig. 7. Fracture of notched beam of multilayered composite of Y-ZrO,/AI,O, and LaPO,. (a) Initial debonding after crack from notch was grown to first LaPO., layer. (b) Cracking within LaPO, layer at position -200 km to the right of (a). (c) Schematic of debonding mechanism. being negligibly small in the specimen of Fig. 9(c), which contains a small volume fraction of LaPO, layers). In regions near the polished cross-sectional surface, there are also residual stresses in the direction normal to the layers with opposite signs to the corresponding parallel stre~ses.~' The magnitudes of the normal stresses are maximum at the surface (equal to the magnitudes of the parallel stresses) and decay within a depth approximately equal to the layer thickness. Therefore, the growth of the cracks along the LaPO, layer of Fig. 9(c) is opposed by the compressive near-surface stress within the LaPO, layer. The magnitude of this stress is approximately 130 MPa. However, since the depth over which the stress acts is small compared with the crack depth, the influence on the crack length is not large: application of the analysis of Lawn and FulleP' indicates that the corresponding overestimate of the interlaminar toughness for the Al,O,/ZrO, composites is no more than 10%. Fig. 8. Broken notched beam of multilayered composite of Ce-ZrO, and LaP04 showing evidence for both transformation toughening and interlaminar debonding. (a) Nomarski interference micrograph showing surface uplift adjacent to main crack due to tetragonal-tomonoclinic phase transformation in the ZrO, layers (boundary of uplifted zone indicated by arrows). (b) SEM micrograph showing cracking within LaPO, layers in uplifted region adjacent to main crack. IV. Discussion The results of the previous section indicate that, when Ce-stabilized zirconia and LaPO, were in contact at 1600°C, some of the Ce diffused into the LaPO,, displacing La. The displaced La reacted with the zirconia to form a precipitate phase containing La, Ce, and Zr, possibly a pyrochlore-like structure. When A1,0, was present also in the mixed A1,0,/ Ce-ZrO, layers, the displaced La formed a (La,Ce) magnetoplumbite as well as LaAlO,. In view of these results, it is likely that CePO,, which has the same monazite structure as LaPO, and has now been found to be phase-compatible with Ce-Z10,,3~ would be a more suitable debonding phase for Ce0,-stabilized ZrO,. In contrast, the results indicate that Y, which forms a different (but related) phosphate ~tructure,~' xenotime, does not diffuse into LaPO, at 160O0C, and La does not diffuse out of the LaPO
Journalof the American Ceramic Society-Marshall et al. No. 7 The fracture behavior of the ZrO, /LaPO layered composites grow from LapO, to ZrO2, a=0. 2(E,= 133 GPa(Ref 9). ly similar to that of Al, o,/LaPo composites E2= 200 GPa)and T/T2=0.05; and (2)for a crack about to investigated previously. Specifically, normally incident cracks grow from LapO4 to Al2O3/ZrO2, a=0.35(E2=300 GPa) in a ZrO, layer appear to penetrate the ZrO2-LaPO. interface and T/T-0. 1. These values fall below the critical condition and to debond either at the lapo4- zro2 interface or within the in Fig 10 so that debonding is expected, as observed LapO. layer. The specific location of debonding and dam within the Lapo, layers is influenced by the morphology of the LaPO,, a different response is predicted. In that case, T,is the of the grain size(see Fig. 1). Although debonding rs near changed, corresponding to interchanging the two materials the interface, the crack path lies both at the interface and in corresponding value of r/T,(-1)falls above the critical con- the LapO. This observation is consistent with the measure terlaminar toughness being approximately equal to the tough dition, where the crack is predicted to gro w into the Lapo. as observed ness of LaPO4. Despite this complication, useful insight may be ained by comparing the measured toughnesses with the analy The debonding criterion in Fig. 10 is also infuenced by sis of interfacial debonding by he and Hutchinson, 20 residual stresses parallel and normal to the interface. In these Whether a crack approaching an interface between two mate composites there are two sources of residual stresses (i) Transformation Stresses: If the ZrO, layers were to the second material depends on the ratio I/Ti of the fracture either during cooling or in the tensile stress field ahead of the elastic mismatch parameter,a, given by effect of these stresses can be estimated using the analysis of a=(E2-E1)(E2+E) (1) He et aL., 4 who define a normalized residual stress parameter where e is the plane strain Young,s modulus. The critical alues of T/T, calculated by He and Hutchinson are shown in m= orva/K where o is the residual stress either normal to the interface or parallel to the interface in material 2, a is a characteristic fiaw 5 Fig. 9. Vickers indentations in composite of Y-Zro2/Al,O, and LaPO4.(a)and(b)SEM micrographs showing damage within LaPO. layers (c)Comparison of interlaminar crack length and crack length in Y-ZrO2/Al2O, matri
1682 Journal of the American Ceramic Society-Marshall et al. Vol. 80, No. 7 The fracture behavior of the ZrO,/LaPO, layered composites is qualitatively similar to that of Al,O,/LaPO, composites investigated previou~ly.~~ Specifically, normally incident cracks in a ZrO, layer appear to penetrate the zrO2-LaPo4 interface and to debond either at the LaP0,-ZrO, interface or within the LaPO, layer. The specific location of debonding and damage within the LaPO, layers is influenced by the morphology of the interface, which in these composites has roughness on the scale of the grain size (see Fig. 1). Although debonding occurs near the interface, the crack path lies both at the interface and in the LaPO,. This observation is consistent with the measured interlaminar toughness being approximately equal to the toughness of LaPO,. Despite this complication, useful insight may be gained by comparing the measured toughnesses with the analysis of interfacial debonding by He and Hutchinson?o Whether a crack approaching an interface between two materials will debond along the interface rather than penetrate into the second material depends on the ratio &/r2 of the fracture energies of the interface and the second material, as well as the elastic mismatch parameter,m a, given by (1) where E‘ is the plane strain Young’s modulus. The critical values of r/r2 calculated by He and Hutchinson” are shown in Fig. 10. With the toughnesses measured in the previous section, the following parameters are obtained. (1) for a crack about to a = (E; - E;)/(E; + E:) grow from LaPO, to ZrO,, a = 0.2 (El = 133 GPa (Ref. 9), E, = 200 GPa) and rl/r2 = 0.05; and (2) for a crack about to grow from LaPO, to Al2O3/ZrO2, a = 0.35 (E, = 300 GPa) and &/r2 - 0.1. These values fall below the critical condition in Fig. 10 so that debonding is expected, as observed. For a crack growing in the reverse direction, from zirconia to LaPo,, a different response is predicted. In that case, r, is the fracture energy of LaPO, (-7 J/m2) and the sign of a is changed, corresponding to interchanging the two materials. The corresponding value of &/I-, (-1) falls above the critical condition, where the crack is predicted to grow into the LaPO,, as observed. The debonding criterion in Fig. 10 is also influenced by residual stresses parallel and normal to the interface. In these composites there are two sources of residual stresses. (i) Transformation Stresses: If the Zro, layers were to undergo the tetragonal-to-monoclinic phase transformation, either during cooling or in the tensile stress field ahead of the incident crack, large residual stresses would be generated. The effect of these stresses can be estimated using the analysis of He et ~l.,~ who define a normalized residual stress parameter q = u,&K (2) where u, is the residual stress, either normal to the interface or parallel to the interface in material 2, a is a characteristic flaw Fig. 9. Vickers indentations in composite of Y-Zr0,/AI2O3 and LaPO,. (a) and (b) SEM micrographs showing damage within LaPo, layers. (c) Comparison of interlaminar crack length and crack length in Y-ZrO,/AI,O, matrix
July 1997 1683 2D. B. Marshall, J. J. Ratto, and F.F. "Enhanced Fracture Composites of Ce-Zro, and Z/onAgo ite.D.B. Marshall. "The Design of High Toughness Laminar Zirconia Compos- esistance Curves in Layered T/Ta annister andR J H. Hannink, Techn 1993 C J. Russo, M. P. Harme, H. M. Chan, and g. A. Miller. " Design of a ved Strength and tough rakash, P. Sarkar, and P. S. Nicholson, "Crack Deflection in Ceramic/ 28 Ak203 rO/A1,O, Micro-Laminate Ceramic/Ceramic Composites J. Mater. Sci. J S. Moya, " Layered Ceramic, "Adv Mater., 7. 185-89(1995). w.J. Clegg, K. Kendall, N. M. Alford, T. w. Button, and J. D. Birchall,"A =(E2·E'1(E2+E A J. Phillipps, w.J. Clegg, of fracture energies with debond He and Hutchinson. 2 Solid curve is debond criterion for tetragonal Liu and S. M. Hsu, Fracture Behavior of Multilayer Silicon Nitride/ zirconia layers(negligible residual stresses). Dashed curves are approxi JAm. Ceram.Soc,7992452-57(1996 mate criteria in the presence of residual stresses: (1)m=0. 15, corre- sponding to ZrO, layers that are transformed to the monoclinic structure Structural Applications" pre (residual compression in ZrO, layers, tension in LaPO, layers)and(2)m Ceramic Society, Cincinnati, OH, May 3, 1995, Joint Engineering the american 0.04 for ZrO2/Al,O, layers. Symbols represent measured properties for LaPO/Y-ZrO, and LaPO, /(Y-Zro2, Al,O3)systems P. E. D. Morgan and D. B Ma ize, and K is the applied stress intensity factor for the incident crack. For a composite with equal thickness layers of Lapo D H Kuo and w. m. Kriven, "Characterization of Yttrium Phosphate and O2 layers fully transformed, l stresses are as follows: zero normal to the layers: +800 D. H. Kuo and w. M. Kriven, ""Chemical Stability, Microstructure and MPa parallel to the layers within the ZrO2. Taking the charac- 23-34 19 avior of LaPo, Containing Ceramics,"Mater.SciEng.A,210 MPa parallel to the layers within the lapO. layers; and-800 teristic faw size to be equal to the grain size(-1 um)and K as P. E. D Morgan and D, B. Marshall, "Functional Interfaces in Oxide-Oxid the toughness of zirconia(6 MPa.m), we obtain m=+o15 ay, " Textured Magnetoplumbite Fiber-Matrix for a crack growing from transformed Zro2 to LaPO, and m Interphase Derived Sol-Gel Fiber Coatings, J. Am. Ceram Soc. 79 [5] 0.15 for a crack growing in the reverse direction. The corre- sponding positions of the critical debond conditions from the ment of Int M. H. Lewis, M. G. Cain, P Doleman, A.G. Razzell, and J Gent, Develop- results of He et al. 4 are shown as broken lines in Fig. 10. For Ceramic Transactions, Vol 58, High-Temperature Ceramic-Matrix Composites both directions of crack growth, the effect of residual stress l: Design, Durability, and Performance. American Ceramic Society, Westerville, makes the response already described more likely. (ii) Thermal Expansion Mismatch Stresses: As discussed He and J. w. hutchinson, "Crack Defection at an Interface between lids struci,25,1053-67(1989) sion mismatch are present in composites consisting of layers of Method for efficient colloidal Panic e Eckiange and D s Pearson. "New in the previous section, residual stresses due to thermal expan- Al2 O,ZrO2 and LapO,(these stresses are negligibly small in composites consisting only of zro2 and Lapo4). The sign of the and AL,O, /Zro, Composite Slurries vs Inter Al, O, to ZrO, ratio is-200 MPa, giving m=0.04. The corre sponding shift in the critical debond condition is shown in te" Pp. 543-53 in Proceedings of 5th Intemational Conference on Composite Fig. 10; the shift does not change the behavior described above WrA. G. Evans, "Engineering Property Requirements for High Performance materials. Metallurgical Society, Warrendale, PA, 1985. Ceramics, "Mater Sci Eng, 71, 3-21(1985) V. Conclusions hear Fracture, "Int. J fract. 37, 137-59(1988) Y-stabilized zirconia and Y-zro,/AL, O which is stable at tem- ed ineering Materials, ASTM STP 948 dited by J. E. Masters, American Society for Testing and Materials, West peratures at least as high as 1600'C. However, with Ce-stabilized ZrO,, counter diffusion of Ce and La produces a new pyrochlore- like phase. In the presence of two-phase Ce-ZrO2/AL,o3, the R. Anstis, P. Chantikul, B. R. Lawn, and D. B formation of a( La, Ce)magnetoplumbite was observed. Despite these reactions, multilayered composites exhibited debonding in the LapO4 layers when loaded in bending 2Z. Suo and J W. Hutchinson, "On Sandwich Test Specimens for Measuring Interface Crack Toughness, "Mater. Sci. Eng, A, 107, 135-43(1989) Acknowledgment: We wish to thank Dr. John Armstrong and Paul MS. Ho, C. Hillman, F. F. Lange, and Z. Suo, "Surface Cracking Carpenter for their assistance with the electron microprobe ler Biaxial Residual Compressive Stress "J. Am. Ceram. Soc. 78 [9]2353- Refere B.R. Lawn and E. R J. Fuller, ""Measurement of Thin-Layer Surface er, H. M. Chan, and G. A. Miller, "Unique Op Microstructural Engineering with Duplex and Laminar Ceramic P Compost s r Stres,ss b avis. pt. R. M. Housley, " Machin- JAm. Ceram.Soc,75]175-28(1992). 00. Muller and R. Roy. The Major Ternary Structural Families. Springer- MM.-Y He, A. G. Evans. and J. w, Hutchinson . Crack Deflection at Equation(2)is an approximate expression, valid as long as the elastic mismatch is similar elastic Materials: role of residual stresses ,l J. Solids struct.,31[24]3443-55(1994)
July 1997 Debonding in Multilayered Composites of Zirconia and LaPO, 1683 ’D. B. Marshall, J. J. Ratto. and F. F. Lange. “Enhanced Fracture Toughness in Layered Composites of Ce-ZrO, and A1,0,,” J. Am. Ceram. Soc., 74 [12] ’D. B. Marshall, “The Design of High Toughness Laminar Zirconia Composites,” Am. Ceram. SOC. Bull., 71 [6] 969-73 (1992). ,D. B. Marshall and J. J. Ratto, “Crack Resistance Curves in Layered Ce-ZrO,/Al,O, Ceramics”; pp. 517-23 in Science and Technology of Zirconia V. Edited by S. P. S. Badwal, M. I. Bannister, and R. J. H. Hannink. Technomic, Lancaster, PA, 1993. ’C. J. Russo, M. P. Harmer, H. M. Chan, and G. A. Miller, “Design of a Laminated Ceramic Composite for Improved Strength and Toughness,” J. Am. Ceram. Soc., 75 [I21 3396-400 (1992). 60. Prakash. P. Sarkar. and P. S. Nicholson, “Crack Deflection in Ceramic/ Ceramic Laminates with Strong Interfaces,’’ J. Am. Ceram. Soc., 78 [ 141 1125- 27 (1995). ’P. S. Nicholson, P. Sarkar, and X. Haung, “Electrophoretic Deposition and Its Use to Synthesize ZrO,/Al,O, Micro-Laminate Ceramic/Ceramic Composites,’’ 2979-87 (1991). Fig. 10. Comparison of fracture energies with debonding criterion of He and Hutchinson.’” Solid curve is debond criterion for tetragonal zirconia layers (negligible residual stresses). Dashed curves are approximate criteria in the presence of residual stresses:” (1) q = 0.15. corresponding to ZrO, layers that are transformed to the monoclinic structure (residual compression in ZrO, layers, tension in LaPo, layers) and (2) q = 0.04 for zrO,/Al,O, layers. Symbols represent measured properties for LaP04/Y-Zr0, and LaFQ4/(Y-Zr0,, AI,O,) systems. size, and K is the applied stress intensity factor for the incident crack.+ For a composite with equal thickness layers of LaPO, and ZrO,, and with the ZrO, layers fully transformed, the residual stresses are as follows: zero normal to the layers; +800 MPa parallel to the layers within the LaPO, layers; and -800 Mpa parallel to the layers within the ZrO,. Taking the characteristic flaw size to be equal to the grain size (- 1 pm) and K as the toughness of zirconia (-6 MPa-m’”), we obtain q = +0.15 for a crack growing from transformed Zro, to LaPO, and q = -0.15 for a crack growing in the reverse direction. The corresponding positions of the critical debond conditions from the results of He et aL3, are shown as broken lines in Fig. 10. For both directions of crack growth, the effect of residual stress makes the response already described more likely. (ii) Thermal Expansion Mismatch Stresses: As discussed in the previous section, residual stresses due to thermal expansion mismatch are present in composites consisting of layers of Al,O,-ZrO, and LaPO, (these stresses are negligibly small in composites consisting only of ZrO, and LaPO,). The sign of the mismatch is the same as that of the tetragonal-to-monoclinic transformation. The magnitude of these stresses for the 1:l A1,0, to ZrO, ratio is -200 MPa, giving -q = 0.04. The corresponding shift in the critical debond condition is shown in Fig. 10; the shift does not change the behavior described above. V. Conclusions Lanthanum phosphate forms a weakly bonded interface with Y-stabilized zirconia and Y-Zro,/Al,O,, which is stable at temperatures at least as high as 1600°C. However, with Ce-stabilized Zro,, counter diffusion of Ce and La produces a new pyrochlorelike phase. In the presence of two-phase Ce-ZrO,/Al,O,, the formation of a (La,Ce) magnetoplumbite was observed. Despite these reactions, multilayered composites exhibited debonding in the LaPO, layers when loaded in bending. Acknowledgment: We wish to thank Dr. John Armstrong and Paul Carpenter for their assistance with the electron microprobe measurements. References ‘M. P. Harmer, H. M. Chan, and G. A. Miller, “Unique Opportunities for Microstructural Engineering with Duplex and Laminar Ceramic- Composites,” J.Am. Ceram. SOC., 75 [7] 1715-28 (1992). ‘Equation (2) is an approximate expression, valid as long as the elastic mismatch is not too large.u J. Marer.-Sci., 28,6274178 (1993). ‘5. S. Moya, “Layered Ceramic,” Adv. Mazer., 7,185-89 (1995). W. J. Clegg, K. Kendall, N. M. Alford, T. W. Button, and J. D. Birchall. “A Simple Way to Make Tough Ceramics,”Nature (London), 347,455-57 (1990). ‘‘A. J. Phillipps, W. J. Clegg, and T. W. Clyne, “The Correlation of Interfacial and Macroscopic Toughness in SIC Laminates,” Composites, 24 [2] 166-76 (1993). “H. Liu and S. M. Hsu, “Fracture Behavior of Multilayer Silicon Nitride/ Boron Nitride Ceramics:’J. Am. Ceram. Soc., 79 [9] 2452-57 (1996). ”J. B. Davis and W. J. Clegg, “Ceramic Laminates for High Temperature Structural Applications”; presented at the 97th Annual Meeting of the American Ceramic Society, Cincinnati, OH, May 3, 1995. Joint Engineering Ceramics and Basic Science Divisions (Paper No. C-37-95). ”P. E. D. Morgan and D. B. Marshall, “Ceramic Composites of Monazite and Alumina,”J. Am. Cerum. Soc., 78 [6] 1553-63 (1995). I4P. E. D. Morgan, D. B. Marshall, and R. M. Housley, “High Temperature Stability of MonaziteAlumina Composites,” J. Mater. Sci. Eng. A, 195, 215- 22 (1995). I5D.-H. Kuo and W. M. Kriven, “Characterization of Yttrium Phosphate and a Yttrium Phosphateflttrium Aluminate Laminate,” J. Am. Cerum. SOC.. 78 [ 111 3121-24 (1995). I6D.-H. Kuo and W. M. Kriven, “Chemical Stability, Microstructure and Mechanical Behavior of LaFQ-Containing Ceramics,” Mater. Sci. Eng. A, 210, 123-34 (1996). ”F! E. D. Morgan and D. B. Marshall, “Functional Interfaces in Oxide-Oxide Composites,”J. Marer. Sci. Eng. A, 162 [l-21 15-25 (1993). I’M. K. Cinibulk and R. S. Hay, “Textured Magnetoplumbite Fiber-Matrix Interphase Derived Sol-Gel Fiber Coatings,’’ J. Am. Ceram. Soc., 79 [5] 1233- 46 (1996). I’M. H. Lewis, M. G. Cain, P. Doleman, A. G. Razzell, and J. Gent, “Development of Interfaces in Oxide and Silicate-Matrix Composites”; pp. 41-52 in Ceramic Transactions, Vol. 58, High-Temperature Ceramic-Matrix Composites I: Design, Durabiliry. and Petformanre. American Ceramic Society, Westerville, OH, 1995. ”M.-Y. He and J. W. Hutchinson, “Crack Deflection at an Interface between Dissimilar Materials,” Inr. J. Solids Strucr., 25, 1053-67 (1989). 21B. V. Velamakanni, J. C. Chang, F. F. Lange. and D. S. Pearson, “New Method for Efficient Colloidal Particle Packing via Modulation of Repulsive Lubricating Hydration Forces,” Langmuir, 6 [7] 1323-25 (1990). 22J. C. Chang and B. V. Velamakanni, “Centrifugal Consolidation of A1,0, and Al,O,/ZrO, Composite Slurries vs Interparticle Potentials: Particle Packing and Mass Segre.gation,”J. Am. Ceram. Soc., 74 [9] 2201-204 (1991). 23A. G. Evans, M. D. Thouless, D. P. Johnson-Walls, E. Y. Luh. and D. B. Marshall, “Some Structural Properties of Ceramic Matrix Fiber Composite”; pp. 543-53 in Proceedings of 5th International Conference on Composite Materials. Metallurgical Society. Warrendale, PA, 1985. ”A. G. Evans, “Engineering Property Requirements for High Performance Ceramics,”Marer. Sci. Eng., 71.3-21 (1985). 25H. Chai, “Shear Fracture,” Inr. J. Frucr.. 37, 137-59 (1988). z6M. F. Hibbs and W. L. Bradley, “Correlations between Micromechanical Failure. Processes and the Delamination Toughness of Graphite Epoxy Systems”; pp. 68-97 in Fracrography of Modern Engineering Materials, ASTM STP 948. Edited by J. E. Masters. American Society for Testing and Materials, West Conshohoken, PA, 1987. ,’B. W. Smith and R. A. Grove, “Determination of Crack Propagation Directions in Graphite Epoxy Structures”; see Ref. 26, pp. 68-97. %. R. Anstis, P. Chantikul, B. R. Lawn, and D. B. Marshall, “A Critical Evaluation of Indentation Techniques for Measuring Fracture Toughness: I. Direct Crack Measurements,”J. Am. Ceram. Soc., 64 [9] 533-38 (1981). 29Z. Suo and J W. Hutchinson. “On Sandwich Test Specimens for Measuring Interface Crack Toughness,”Murer. Sci. Eng. A, 107, 135-43 (1989). ’OS. Ho, C. Hillman, F. F. Lange, and 2. SUO, “Surface Cracking in Layers under Biaxial Residual Compressive Stress,” J. Am. Ceram. Soc., 78 [9] 2353- 59 (1995). ”B. R. Lawn and E. R. J. Fuller, “Measurement of Thin-Layer Surface Stresses by Indentation Fracture:’J. Marer. Sci.. 19,4061-67 (1984). ’3. B. Davis, D. B. Marshall, P. E. D. Morgan, and R. M. Housley, “Machinable Ceramics Based on LaPO, and CePO,,” unpublished work. ”0. Muller and R. Roy, The Major Ternary Srrucrural Families. SpringerVerlag, New York, 1974. ’,M.-Y. He, A. G. Evans, and J. W. Hutchinson, “Crack Deflection at an Interface between Dissimilar Elastic Materials: Role of Residual Stresses,” Inr. J. Solids Strucr., 31 [24] 3443-55 (1994). 0