MIATERIAL TENGE ENGMEERIM ELSEVIER Materials Science and Engineering A241(1998)241-250 fracture of multilayer oxide composites Dong- Hau Kuo Waltraud M. Kriven Department of Materials Science and Engineering, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USA Received 3 February 1997; received in revised form 9 June 1997 Abstract Ceramics with high strength and damage tolerance have been realized in the well-recognized non- oxide systems, e.g. silicon arbide/graphite and silicon carbide/boron nitride laminates. All-oxide systems without compliant materials functioning as do graphite and boron nitride are still facing the problem of catastrophic failure. In this study, ten kinds of multilayered, all-oxide laminates fabricated by a low-cost tape casting technique were evaluated These materials were yttrium phosphate (YPO4)-and lanthanum phosphate (LaPO4)-containing zirconia (ZrO2) laminates, and aluminum phosphate(AlPO4)-containing alumina (Al,O3)laminates. A YPOa- containing ZrO, laminate demonstrated excellent strength and improved damage tolerance. A LaPOa-containing ZrO, laminate also displayed a satisfactory result. An AlPO4/Al,O3 laminate was weak. The AlPO,/Al,O minate was strengthened by reinforcing AlPO4 layers with Al,O3, while retaining non-brittle fracture. Different flexural behaviors in different oxide laminates were discussed. @1998 Elsevier Science s.a Keywords: Y ttrium phosphate: Lanthanum phosphate: Aluminum phosphate; Zirconia; Alumina; Laminate; Mechanical property 1. Introduction between fiber and matrix, to weaken the fiber/matrix interface. Without a weak fiber/matrix interface, the Ceramics have many excellent properties that make fiber-reinforced composites demonstrate catastrophic their use as structural materials very attractive. These failure. However, there are still two remaining prob- properties include high strength, high hardness, wear lems: one is the high temperature oxidation of the resistance, high melting temperature, excellent thermal interlayer and the polymer-derived fibers, and the other and chemical stability, low density, and a unique set of is the high fabrication cost electrical, thermal, and other properties. However, their All-oxide(oxide fiber/oxide matrix) composites might brittleness has prevented their use in structural applica- be the ultimate materials as far as the effect of oxidizing tions to date environment is concerned. Nevertheless, the develop- The most promising candidates for improving ce- ment of oxide fibers and oxide interlayers behaving like amic brittleness have been continuous fiber ceramic carbon and boron nitride as weak interlayers in non-ox composites(CFCCs) which have shown high strength ide systems is nt re- and toughness [1]. The development of these advanced search effort. a potential oxide interphase is lanthanum structural materials for high temperature applications phosphate(LaPO4), having a monazite structure, which nportant technological goal. Progress in was proposed by Morgan et al. [2-4]. Yt phos- many areas of technology, such as gas turbines, heat phate (YPO4, with a xenotime structure, expected to exchangers, space re-entry vehicle design and the like, is behave as an analogue to monazite (LaPO4), has also dependent on the development of these high tempera- been evaluated ture structural materials. These basic materials include Tape-casting has been used to fabricate versatile silicon carbide (Sic) fiber-reinforced ceramic com- laminate composites by stacking tapes of different com- posites with a carbon(C) or boron nitride (BN) layer positions, as well as to incorporate fibers and whiskers into the laminates. Material properties can be g author trolled by adjusting the tape compositions, reinforce- address: Department of Materials Science and Engineer. ment orientation, and stacking sequence. Clegg and Dong Hwa University, Hualien, Taiwan, R.O.C coworkers [9, 10 produced a laminar fabric of Sic 0921-5093/98/S19.00 0 1998 Elsevier Science S.A. All rights reserved PIS09215093097100498·X
Materials Science and Engineering A241 (1998) 241–250 Fracture of multilayer oxide composites Dong-Hau Kuo 1 , Waltraud M. Kriven * Department of Materials Science and Engineering, Uni6ersity of Illinois at Urbana-Champaign, Urbana, IL 61801, USA Received 3 February 1997; received in revised form 9 June 1997 Abstract Ceramics with high strength and damage tolerance have been realized in the well-recognized non-oxide systems, e.g. silicon carbide/graphite and silicon carbide/boron nitride laminates. All-oxide systems without compliant materials functioning as do graphite and boron nitride are still facing the problem of catastrophic failure. In this study, ten kinds of multilayered, all-oxide laminates fabricated by a low-cost tape casting technique were evaluated. These materials were yttrium phosphate (YPO4)- and lanthanum phosphate (LaPO4)-containing zirconia (ZrO2) laminates, and aluminum phosphate (AlPO4)-containing alumina (Al2O3) laminates. A YPO4-containing ZrO2 laminate demonstrated excellent strength and improved damage tolerance. A LaPO4-containing ZrO2 laminate also displayed a satisfactory result. An AlPO4/Al2O3 laminate was weak. The AlPO4/Al2O3 laminate was strengthened by reinforcing AlPO4 layers with Al2O3, while retaining non-brittle fracture. Different flexural behaviors in different oxide laminates were discussed. © 1998 Elsevier Science S.A. Keywords: Yttrium phosphate; Lanthanum phosphate; Aluminum phosphate; Zirconia; Alumina; Laminate; Mechanical property 1. Introduction Ceramics have many excellent properties that make their use as structural materials very attractive. These properties include high strength, high hardness, wear resistance, high melting temperature, excellent thermal and chemical stability, low density, and a unique set of electrical, thermal, and other properties. However, their brittleness has prevented their use in structural applications to date. The most promising candidates for improving ceramic brittleness have been continuous fiber ceramic composites (CFCCs) which have shown high strength and toughness [1]. The development of these advanced structural materials for high temperature applications remains an important technological goal. Progress in many areas of technology, such as gas turbines, heat exchangers, space re-entry vehicle design and the like, is dependent on the development of these high temperature structural materials. These basic materials include silicon carbide (SiC) fiber-reinforced ceramic composites with a carbon (C) or boron nitride (BN) layer between fiber and matrix, to weaken the fiber/matrix interface. Without a weak fiber/matrix interface, the fiber-reinforced composites demonstrate catastrophic failure. However, there are still two remaining problems: one is the high temperature oxidation of the interlayer and the polymer-derived fibers, and the other is the high fabrication cost. All-oxide (oxide fiber/oxide matrix) composites might be the ultimate materials as far as the effect of oxidizing environment is concerned. Nevertheless, the development of oxide fibers and oxide interlayers behaving like carbon and boron nitride as weak interlayers in non-oxide systems is ongoing and requires a significant research effort. A potential oxide interphase is lanthanum phosphate (LaPO4), having a monazite structure, which was proposed by Morgan et al. [2–4]. Yttrium phosphate (YPO4), with a xenotime structure, expected to behave as an analogue to monazite (LaPO4), has also been evaluated [5–8]. Tape-casting has been used to fabricate versatile laminate composites by stacking tapes of different compositions, as well as to incorporate fibers and whiskers into the laminates. Material properties can be controlled by adjusting the tape compositions, reinforcement orientation, and stacking sequence. Clegg and coworkers [9,10] produced a laminar fabric of SiC * Corresponding author. 1 Present address: Department of Materials Science and Engineering, National Dong Hwa University, Hualien, Taiwan, R.O.C. 0921-5093/98/$19.00 © 1998 Elsevier Science S.A. All rights reserved. PII S0921-5093(97)00 49 8- X
D -H. Kuo, W.M. Kricen/ Materials Science and Engineering 4241(1998)241-250 interleaved graphite films. Folsom et al. [ll] demon- 2.3. Laminate fabrication strated a laminar ceramic/carbon fiber-reinforced epoxy composite. Baskaran and coworkers [12-14 The procedure for making laminated composites by died fibrous monolithic ceramics of a SiC/C sys- tape casting was similar to those described elsewhere tem [13] and a SiC/BN system [14]. Shaw et al. [15] [19]. The slurry formulation contained 20 vol% oxide and Chen and Mecholsky Jr. [16] incorporated powders, 60 vol% solvent consisting of a mixture of metallic layers with ceramics to increase ceram trichloroethylene and ethanol, as well as a dispersant, a toughness. Liu and Hsu [17 fabricated multilayer sil- binder and plasticizers. Slurries were tape cast to yield icon nitride(Si,N4/BN ceramics. All of these materi- laminae of 100-200 um thickness with a doctor blade als have problems in high temperature oxidizing opening of 250-350 um. 80-Layer laminated com- environments posites were fabricated by periodically stacking two or 9 In this paper, all-oxide ceramics were fabricated by three kinds of oxide laminae having dimensions of 25 low-cost, tape casting technique without incorpo mm x 51 mm. Thermocompression was done by hold rating expensive fibers. The materials were YPO- ing for I h at 50-80 C under a 10 MPa pressure. The and LaPOa-containing zirconia (ZrO2) oxide lami organic additives were removed by heating to 650C at nates, and aluminum phosphate(AlPO4)-containing a rate of 3C/h, followed by a 3-h holding time. Subse alumina(Al,O )laminates. Fractural behaviors of quently the bulk materials were isostatically cold nese laminates were evaluated by 4-point flexural pressed at. -170 MPa for 10 min, then loaded in a testing, indentation and microstructural examination nates or Al,O3 powders for AlPO4 laminates respec- tively surrounding the pressed laminates Consolidation was performed by hot pressing, under an Experimental procedures argon atmosphere at 28 MPa, at a temperature of 1550C for YPO a and LaPO4 laminates and 1600C for AlPO4 laminates, both for 2 h. After hot pressing, the 2.I. Materials aminate was annealed at 1000%C for 6 h YPO4, LaPO4, and a 50/50 vol%(AlPO4+ Al,O3) owders were prepared by the Pechini method asTable was used for LaPO4 earlier [18]: 3 mol% yttria-par Symbol and mechanical response of YPO4, LaPO", and AlPO-con- ially stabilized(TZ-3Y or Y-ZrO2)and un-stabilized taining oxide laminates zirconia(TZ-0 or ZrO2) powders from Tosoh, At- System Symbol of stacking periodStrength/damage lanta. GA. were used for the ZrO, source: 99.8% tolerance Al6-SG(Alcoa, Pittsburgh, PA)alumina powder for the Al,O3 source; mullite (3Al2O3. 2SiO2) from Kyor- YPO -containing Zro2 laminates itsu, Nagoya, Japan; AlPO4 from Aldrich, Milwau Y(a) YPO/Y-Zr02YZ3-A7/Y- Goo kee, Wi as one of the AlPO, sources: and 99.9% Zro cerium(Iv) oxide(Aldrich),99.9% strontium oxide Y(b) YP7-YZ3/(YCeSr)Z7-A3 Medium/ bad Y(c) YP7-YZ3/( CeSr) (Aldrich), and 99.9% yttrium oxide(Molycorp, White Plains, NY) for the additives Z7-A3/A/YP7-YZ3 Y(d) YPO/Y-ZrO Shattered after hot 2. 2. Chemical compatibility and microstructural LaPOa-containing ZrO. laminates aPO/Y-Zro Shattered during speci- (b) LaPO//(YCeSr)Z7-A3 Medium/medium patibility were carried out on pressed pellets composed of YPO4, LaPO4,or Z7-Mu3/A/(Y CeSr)Z7-Mu3 AlPO4 powders as one component and Y-ZrO L(d) LaPO4/Y-Zr02/Z3./Y Shattered after hot lets were cold isostatically pressed at 86 MPa for 5 AlPO-containing ZrO2 laminates pressing AL,O, as the other. After uniaxial pressing, the pel- min. These materials were fired at 1500%C. 1550C or 600oC for 3 h and the phases were identified using A(b)(AlPO4+AL,O3/AL,O Medium/medium X-ray diffractometry(XRD, Model D-Max, Rigaku Danvers, MA). Microstructural characterization was () Mu: mullite(3A103 2SiO2);(i)(YCeSr)Z7-A3 and (YCeSrz7. performed by optical microscopy and Mu3 were doped with ceria and strontium oxide in addition to yttria; (i)YZ3-A7 and Z3A7 represent the ZrO/,O, composites with tron microscopy(SEM, Model DS-130, International nd without stabilized yttria, respectively; (iv)(AlPO4+Al,O3 ) 50 Scientific Instruments, Santa Clara, CA) vol AlPO,+50 vol% AlO
242 D.-H. Kuo, W.M. Kri6en / Materials Science and Engineering A241 (1998) 241–250 interleaved graphite films. Folsom et al. [11] demonstrated a laminar ceramic/carbon fiber-reinforced epoxy composite. Baskaran and coworkers [12–14] studied fibrous monolithic ceramics of a SiC/C system [13] and a SiC/BN system [14]. Shaw et al. [15] and Chen and Mecholsky Jr. [16] incorporated metallic layers with ceramics to increase ceramic toughness. Liu and Hsu [17] fabricated multilayer silicon nitride (Si3N4)/BN ceramics. All of these materials have problems in high temperature oxidizing environments. In this paper, all-oxide ceramics were fabricated by a low-cost, tape casting technique without incorporating expensive fibers. The materials were YPO4- and LaPO4-containing zirconia (ZrO2) oxide laminates, and aluminum phosphate (AlPO4)-containing alumina (Al2O3) laminates. Fractural behaviors of these laminates were evaluated by 4-point flexural testing, indentation and microstructural examination. 2. Experimental procedures 2.1. Materials YPO4, LaPO4, and a 50/50 vol% (AlPO4+Al2O3) powders were prepared by the Pechini method as was used for LaPO4 earlier [18]; 3 mol% yttria-partially stabilized (TZ-3Y or Y-ZrO2) and un-stabilized zirconia (TZ-0 or ZrO2) powders from Tosoh, Atlanta, GA, were used for the ZrO2 source; 99.8% A16-SG (Alcoa, Pittsburgh, PA) alumina powder for the Al2O3 source; mullite (3Al2O3 · 2SiO2) from Kyoritsu, Nagoya, Japan; AlPO4 from Aldrich, Milwaukee, WI as one of the AlPO4 sources; and 99.9% cerium (IV) oxide (Aldrich), 99.9% strontium oxide (Aldrich), and 99.9% yttrium oxide (Molycorp, White Plains, NY) for the additives. 2.2. Chemical compatibility and microstructural characterization Studies of chemical compatibility were carried out on pressed pellets composed of YPO4, LaPO4, or AlPO4 powders as one component and Y-ZrO2 or Al2O3 as the other. After uniaxial pressing, the pellets were cold isostatically pressed at 86 MPa for 5 min. These materials were fired at 1500°C, 1550°C or 1600°C for 3 h and the phases were identified using X-ray diffractometry (XRD, Model D-Max, Rigaku, Danvers, MA). Microstructural characterization was performed by optical microscopy and scanning electron microscopy (SEM, Model DS-130, International Scientific Instruments, Santa Clara, CA). 2.3. Laminate fabrication The procedure for making laminated composites by tape casting was similar to those described elsewhere [19]. The slurry formulation contained 20 vol% oxide powders, 60 vol% solvent consisting of a mixture of trichloroethylene and ethanol, as well as a dispersant, a binder and plasticizers. Slurries were tape cast to yield laminae of 100–200 mm thickness with a doctor blade opening of 250–350 mm. 80-Layer laminated composites were fabricated by periodically stacking two or three kinds of oxide laminae having dimensions of 25 mm×51 mm. Thermocompression was done by holding for 1 h at 50–80°C under a 10 MPa pressure. The organic additives were removed by heating to 650°C at a rate of 3°C/h, followed by a 3-h holding time. Subsequently the bulk materials were isostatically cold pressed at 170 MPa for 10 min, then loaded in a graphite die with Y-ZrO2 for YPO4 and LaPO4 laminates or Al2O3 powders for AlPO4 laminates respectively, surrounding the pressed laminates. Consolidation was performed by hot pressing, under an argon atmosphere at 28 MPa, at a temperature of 1550°C for YPO4 and LaPO4 laminates and 1600°C for AlPO4 laminates, both for 2 h. After hot pressing, the laminate was annealed at 1000°C for 6 h. Table 1 Symbol and mechanical response of YPO4-, LaPO4-, and AlPO4-containing oxide laminates System Symbol of stacking period Strength/damage tolerance YPO4-containing ZrO2 laminates Y(a) YPO4/Y-ZrO2/YZ3-A7/Y- Good/good ZrO2 Y(b) YP7-YZ3/(YCeSr)Z7-A3 Medium/bad Y(c) Medium YP7-YZ3/(YCeSr) /bad Z7-A3/A/YP7-YZ3 Y(d) Shattered after hot YPO4/Y-ZrO2 pressing LaPO4-containing ZrO2 laminates LaPO4 L(a) /Y-ZrO2 Shattered during specimen cutting LaPO4 L(b) /(YCeSr)Z7-A3 Medium/medium LaPO4 L(c) Low /(YCeSr) /low Z7-Mu3/A/(YCeSr)Z7-Mu3 LaPO4/Y-ZrO2 L(d) Shattered after hot /Z3-A7/YZrO2 pressing AlPO4-containing ZrO2 laminates A(a) AlPO4/Al2O3 Low/medium A(b) (AlPO4+Al2O3)/Al2O3 Medium/medium (i) Mu: mullite (3Al2O3 · 2SiO2); (ii) (YCeSr)Z7-A3 and (YCeSr)Z7- Mu3 were doped with ceria and strontium oxide in addition to yttria; (iii) YZ3-A7 and Z3-A7 represent the ZrO2/Al2O3 composites with and without stabilized yttria, respectively; (iv) (AlPO4+Al2O3): 50 vol% AlPO4+50 vol% Al2O3.
D -H. Kuo, W.M. Kricen/ Materials Science and Engineering 4241(1998)241-250 Ten kinds of YPO4, LaPO4, and AlPOa laminat (a) ere fabricated, as listed in Table 1. For example, three aLaP4o￥2 kinds of tape-cast tapes were fabricated: YPO4, 3 mol% Y-ZrO2, and 30 vol% Y-ZrO2-70 vol% Al,O3(YZ3 A7)tapes. An 80-layer laminated composite was formed by stacking these three kinds of tapes in the ·Y1o4oYZO2 repeating sequence of YPO4→Y-ZrO2→YZ3-A7→Y Zro2(labelled as the Y(a) system) The (Y CeSr)Z7-A3 and (YCeSr)Z7-Mu3 layers were omposed of 70 vol% ZrO2 and 30 vol% Al,, or mullite with 2 mol%Y,O3 and 4 mol% CeO, based on O Y-Zr( O2 ZrO, and 9. 1 mol% Sro based on Al,O3(molar ratio of hese additives were used due to concerns of the strength and the humidity sensitivity of he composites (4Y, 4Ce)-ZrO2/Al,O3)fabri cated by using wet chemical methods have shown high resistance to the tetragonal-to-monoclinic(t-m) phase 045505560 transformation during low temperature agi Simultaneous additions of SrO and Al, to ZrOz, as Fig. 1. XRD of(a)LaPO/Y-ZrO2(b)YPO,/Y-ZrO2 and(c)Y-Zro2 proposed by Cutler et al. [21] can lead to the in situ ellets fired at 1550C for 3 h formation of strontium aluminate platelets. This type of zirconia can have high strength and hardness without reaction compound. The Y-ZrO2 phase remains tetrag- loss of toughness Chemical stability between AlPO4 and Al,O3 was 2. 4. Mechanical evaluation of laminated composites studied by XRD of a powder compact(35 vol% AlPO4 and 65 vol% Al,O3) which was fired at 1600C for 3 h The hot pressed slabs were cut into bars with dimen Fig. 2 shows the XRD result which indicates the chem- point flexural tests having an outer span of 20 mm and AlPO, was in the a-cristobalite form nd ions of25mm×2.0-2.5mm×2.0-2.5mm.Four- ical compatibility between AlPO Al,O3. The an inner span of 10 mm were performed at room emperature. The tensile surface was parallel to the 3.2. Mechanical responses and microstructure of laminate and tested in a screw-driven machine (model laminated composites 4502, Instron, Canton, MA)with a crosshead speed of 0.05 mm/min. Apparent work-of-fracture was obtained Fig 3 shows two load-displacement responses of the y dividing the area under the load-displacement curve YPO4 /Y-ZrO2,/YZ3-A7/Y-ZrO2 laminate(the Y(a)sys- by the cross-sectional area of the sample Radial cracks tem)tested in two different specimens. These responses were generated under a 10-kg indentation load on the Y(a) system and a 5-kg load on the A(a) and A(b) laminates in order to study crack propagation profiles and interaction with the microstructure 口APO 3. Results 口 Studies of chemical reactions between YPO4 and 1500C/3h Y-ZrO, and between LaPO, and Y-zrO, were carried out by firing pellets at 1500°C,1550°C,andl600°Cfo 3 h. It was found that the xrd results were the same for the different firing temperatures. XRD results of Y-ZrO2, YPOA/Y-ZrO2(50/50 vol) and LaPO4/Y-ZrO (50/50 vol) pellets fired at 1550oC for 3 h are shown in Fig. 1. These XRD results indicate that there is chemi- cal compatibility between YPOA and Y-ZrO2 and be Fig. 2. XRD of an AIPO ALO, pellet fired at 1600C for 3 h, tween Lapo, and Y-ZrO, without the formation of a pared with that of an Al,O, pellet fired at 1500oC for 3h
D.-H. Kuo, W.M. Kri6en / Materials Science and Engineering A241 (1998) 241–250 243 Ten kinds of YPO4, LaPO4, and AlPO4 laminates were fabricated, as listed in Table 1. For example, three kinds of tape-cast tapes were fabricated: YPO4, 3 mol% Y-ZrO2, and 30 vol% Y-ZrO2 –70 vol% Al2O3 (YZ3- A7) tapes. An 80-layer laminated composite was formed by stacking these three kinds of tapes in the repeating sequence of YPO4Y-ZrO2YZ3-A7YZrO2 (labelled as the Y(a) system). The (YCeSr)Z7-A3 and (YCeSr)Z7-Mu3 layers were composed of 70 vol% ZrO2 and 30 vol% Al2O3 or mullite with 2 mol% Y2O3 and 4 mol% CeO2 based on ZrO2 and 9.1 mol% SrO based on Al2O3 (molar ratio of SrO/Al2O30.1). These additives were used due to concerns of the strength and the humidity sensitivity of ZrO2. The composites ((4Y,4Ce)-ZrO2/Al2O3) fabricated by using wet chemical methods have shown high resistance to the tetragonal-to-monoclinic (tm) phase transformation during low temperature aging [20]. Simultaneous additions of SrO and Al2O3 to ZrO2, as proposed by Cutler et al. [21] can lead to the in situ formation of strontium aluminate platelets. This type of zirconia can have high strength and hardness without loss of toughness. 2.4. Mechanical e6aluation of laminated composites The hot pressed slabs were cut into bars with dimensions of 25 mm×2.0–2.5 mm×2.0–2.5 mm. Fourpoint flexural tests having an outer span of 20 mm and an inner span of 10 mm were performed at room temperature. The tensile surface was parallel to the laminate and tested in a screw-driven machine (Model 4502, Instron, Canton, MA) with a crosshead speed of 0.05 mm/min. Apparent work-of-fracture was obtained by dividing the area under the load-displacement curve by the cross-sectional area of the sample. Radial cracks were generated under a 10-kg indentation load on the Y(a) system and a 5-kg load on the A(a) and A(b) laminates in order to study crack propagation profiles and interaction with the microstructure. 3. Results 3.1. Chemical compatibility Studies of chemical reactions between YPO4 and Y-ZrO2 and between LaPO4 and Y-ZrO2 were carried out by firing pellets at 1500°C, 1550°C, and 1600°C for 3 h. It was found that the XRD results were the same for the different firing temperatures. XRD results of Y-ZrO2, YPO4/Y-ZrO2 (50/50 vol) and LaPO4/Y-ZrO2 (50/50 vol) pellets fired at 1550°C for 3 h are shown in Fig. 1. These XRD results indicate that there is chemical compatibility between YPO4 and Y-ZrO2 and between LaPO4 and Y-ZrO2 without the formation of a Fig. 1. XRD of (a) LaPO4/Y-ZrO2, (b) YPO4/Y-ZrO2 and (c) Y-ZrO2 pellets fired at 1550°C for 3 h. reaction compound. The Y-ZrO2 phase remains tetragonal. Chemical stability between AlPO4 and Al2O3 was studied by XRD of a powder compact (35 vol% AlPO4 and 65 vol% Al2O3) which was fired at 1600°C for 3 h. Fig. 2 shows the XRD result which indicates the chemical compatibility between AlPO4 and Al2O3. The AlPO4 was in the a-cristobalite form. 3.2. Mechanical responses and microstructure of laminated composites Fig. 3 shows two load-displacement responses of the YPO4/Y-ZrO2/YZ3-A7/Y-ZrO2 laminate (the Y(a) system) tested in two different specimens. These responses Fig. 2. XRD of an AlPO4/Al2O3 pellet fired at 1600°C for 3 h, as compared with that of an Al2O3 pellet fired at 1500°C for 3 h.
D -H. Kuo, W.M. Kricen/ Materials Science and Engineering 4241(1998)241-250 The other laminated system which showed moderate Max, g= 392 MPa strength and damage tolerance was the LaPO4/ wOF= 10 kJ/m (YCeSr)Z7-A3 laminate(the L(b) system). Fig. 6 shows WOF =8,2k] two load-displacement responses of this laminate under 4-point flexural tests. The ultimate strengths were 276 and 254 MPa with corresponding work-of-fracture val ues of 3.3 and 2.9 kJ/m. This laminate had the zirconia layers modified by alumina which reduced the mis- match between LaPO4 and doped ZrO2 in the coeffi cient of thermal expansion. The l(b) laminate with less thermal expansion mismatch was fabricated success- 00050.10.150 25030.350 fully, but the LaPO/Y-ZrO2 laminate(the l(a) system) Crosshead Displ nt(mm) disintegrated during bend bar cutting. Fig. 7 is the side-view micrograph of a 4-point fractured specimen Fig. 3. Load versus displacement curves for two YPO /Y-Zro, /30 as seen by SEM. Interfacial delamination occurred to vol% Y-ZrO2 -70 vol% AL,O,/Y-ZrO2 laminated specimens tested in prevent catastrophic fracture. 4-point flexure Not all the YPO4 and LaPO4 laminates displayed high strength and damage tolerance. The LaPo4/ had ultimate 4-point strengths of 358 and 392 MPa, (YCeSr)Z7-Mu3/A/(YCeSr)Z7-Mu3 laminate(the L(c) respectively. The step-wise load drops, beyond the peak system)had strengths, after two measurements, of 124 stress, were characteristic of the non-catastrophic frac nd 91 MPa. with work-of-fracture values of 1. 2 and ture. Before the bend bar broke the flexural test of the laminate with flexural strength of 358 MPa was stopped, and the specimen was examined under optical nd scanning electron microscopes. Fig. 4(a) and Fig. 4(b)are micrographs of this specimen as seen by optical microscopy and SEM, respectively. The optical mi- crograph shows a low-magnification view of the test bar located between the inner loading points. The delani- nated interfaces extended laterally up to the two outer loading points, but did not run to the end of the test bar. The SEM micrograph revealed the detailed nature of the fracture. The tensile(bottom) part of the lami nate displayed extensive interfacial delamination, while le compressive(top) part stayed intact(Fig. 4(b). The 1 mm delaminated interfaces were only located between YPOa and Y-Zo2. The YPO/Y-ZrO2 interface located close to the mid-plane was severely damaged. Interfaces be- b tween Y-ZO2 and Yz3-A7 were strongly bonded with out interfacial delamination. It was observed that a delaminated interface showed up periodically after each four-layer configuration. Apparent work-of-fracture values were measured to be 8.2 and 10 kJ/m, respec- tively This high strength and damage tolerant oxide lam nate was qualitatively examined by the indentation technique. Vickers indentation cracks were introduced to this oxide laminate at an orientation of 45(Fig. 5) relative to the layer length direction. The radial cracks were generated from the Vickers indent. In addition to the indentation-induced radial cracks, a preferred prop- agation path along the YPO4/Y-ZrO2 interface was observed. This observation was consistent with what Fig. 4.(a) Optical and (b) SEM micrographs showing the side surfaces of a YPO/Y-ZrO,/30 vol% Y-ZrO2-70 vol% AL,./Y-ZrO had occurred during the flexural test laminate after 4-point flexural testing
244 D.-H. Kuo, W.M. Kri6en / Materials Science and Engineering A241 (1998) 241–250 Fig. 3. Load versus displacement curves for two YPO4/Y-ZrO2/30 vol% Y-ZrO2 –70 vol% Al2O3/Y-ZrO2 laminated specimens tested in 4-point flexure. The other laminated system which showed moderate strength and damage tolerance was the LaPO4/ (YCeSr)Z7-A3 laminate (the L(b) system). Fig. 6 shows two load-displacement responses of this laminate under 4-point flexural tests. The ultimate strengths were 276 and 254 MPa with corresponding work-of-fracture values of 3.3 and 2.9 kJ/m2 . This laminate had the zirconia layers modified by alumina which reduced the mismatch between LaPO4 and doped ZrO2 in the coeffi- cient of thermal expansion. The L(b) laminate with less thermal expansion mismatch was fabricated successfully, but the LaPO4/Y-ZrO2 laminate (the L(a) system) disintegrated during bend bar cutting. Fig. 7 is the side-view micrograph of a 4-point fractured specimen as seen by SEM. Interfacial delamination occurred to prevent catastrophic fracture. Not all the YPO4 and LaPO4 laminates displayed high strength and damage tolerance. The LaPO4/ (YCeSr)Z7-Mu3/A/(YCeSr)Z7-Mu3 laminate (the L(c) system) had strengths, after two measurements, of 124 and 91 MPa, with work-of-fracture values of 1.2 and had ultimate 4-point strengths of 358 and 392 MPa, respectively. The step-wise load drops, beyond the peak stress, were characteristic of the non-catastrophic fracture. Before the bend bar broke, the flexural test of the laminate with flexural strength of 358 MPa was stopped, and the specimen was examined under optical and scanning electron microscopes. Fig. 4(a) and Fig. 4(b) are micrographs of this specimen as seen by optical microscopy and SEM, respectively. The optical micrograph shows a low-magnification view of the test bar located between the inner loading points. The delaminated interfaces extended laterally up to the two outer loading points, but did not run to the end of the test bar. The SEM micrograph revealed the detailed nature of the fracture. The tensile (bottom) part of the laminate displayed extensive interfacial delamination, while the compressive (top) part stayed intact (Fig. 4(b)). The delaminated interfaces were only located between YPO4 and Y-ZO2. The YPO4/Y-ZrO2 interface located close to the mid-plane was severely damaged. Interfaces between Y-ZO2 and YZ3-A7 were strongly bonded without interfacial delamination. It was observed that a delaminated interface showed up periodically after each four-layer configuration. Apparent work-of-fracture values were measured to be 8.2 and 10 kJ/m2 , respectively. This high strength and damage tolerant oxide laminate was qualitatively examined by the indentation technique. Vickers indentation cracks were introduced to this oxide laminate at an orientation of 45° (Fig. 5) relative to the layer length direction. The radial cracks were generated from the Vickers indent. In addition to the indentation-induced radial cracks, a preferred propagation path along the YPO4/Y-ZrO2 interface was observed. This observation was consistent with what had occurred during the flexural test. Fig. 4. (a) Optical and (b) SEM micrographs showing the side surfaces of a YPO4/Y-ZrO2/30 vol% Y-ZrO2 –70 vol% Al2O3/Y-ZrO2 laminate after 4-point flexural testing
D -H. Kuo, W.M. Kricen/ Materials Science and Engineering 4241(1998)241-250 Y73-A7 Y-ZrO2 w LaPO4 CYCeSr)Z7-A3 50m Fig. 5. SEM micrograph of an indentation crack pattern in a YPO/ Fig. 7. SEM micrograph showing the side surface of a LaPO/70 YZr( 02 30 vol% Y-Zr02-70 vol% AL,O3/Y-LrO, laminate Indents vol%(YCeSr)Zro2 -30 vol% AlO, laminate after 4-point flexural were oriented at an orientation of 45 relative to the layer direction Cesr)Z7-A3 laminate(the Y(b) system) had flexural inate. A flexural strength of 47 MPa was measured for strengths measured from two specimens as 187 and 163 the AlPO/AL,O, laminate and of 225 MPa for the MPa and fractured with a brittle nature. The YP7-YZ3/(AlPO4+ AlO3)/Al,O3 laminate, with a non- (YCeSr)Z7-A3/A/(YCeSr)Z7-A3 laminate(the Y(c)sys- catastrophic fracture from one test for each tem) had a flexural strength of 217 MPa with a view of the non-brittle A(b) bar was examined under catastrophic failure from one test. SEM micrographs SEM and is shown in Fig. 10(b). Cracks propagating showing the side surfaces of a Y(b)laminate and a Y(c) laterally were responsible for this non-brittle fracture laminate after 4-point flexural testing are displayed in behavior. Fig. 9(a) and Fig. 9(b) Vickers indentation cracks were also introduced to Some laminates did not fabricate successfully. YPO./ the A(a) and A(b) laminates, at an orientation of 45 Y-ZrO2(the Y(d)system)and LaPO/Y-ZrO,/Z3-A7/(Fig. 11) relative to the layer length direction. On Y-ZrO2(the L(d) system)shattered after removal from propagating from the(AlPO4+Al2 O3)layers towards the hot press die. The previously mentioned LaPo4/Y- the Al2O3 layers, the radial cracks preferentially ZrO, laminate stayed intact after removal from the die deflected along the(AlPO4+ AlO3)/AlO Interrace and annealing, but shattered during bend bar cutting ... Max o= 276 MPa WOF= 2,9kJ/ 6SA1203xa (YCeSr)Z7-Mu3 mmm 00050.10150.20.250.30.350.4 Crosshead Displacement(mm) Fig. 8. SEM micrograph showing the side surface of a LaPo/70 Fig. 6. Load of a LaPo/70 vol vol%(YCeSr)ZrO, -30 vol% mullite/Al,,/70 vol%(Y CeSr)ZrO2-30 (YCeSrZrO2 -30 vol% Al,O, laminate tested in 4-point flexure. vol% mullite laminate after 4-point flexural testin
D.-H. Kuo, W.M. Kri6en / Materials Science and Engineering A241 (1998) 241–250 245 Fig. 5. SEM micrograph of an indentation crack pattern in a YPO4/ Y-ZrO2/30 vol% Y-ZrO2 –70 vol% Al2O3/Y-ZrO2 laminate. Indents were oriented at an orientation of 45° relative to the layer direction. Fig. 7. SEM micrograph showing the side surface of a LaPO4/70 vol% (YCeSr)ZrO2 –30 vol% Al2O3 laminate after 4-point flexural testing. 1.0 kJ/m2 , respectively. The flexural strengths and the work-of-fracture values were low. Fig. 8 is its SEM micrograph after a 4-point fracture test. The YP7-YZ3/ (YCeSr)Z7-A3 laminate (the Y(b) system) had flexural strengths measured from two specimens as 187 and 163 MPa and fractured with a brittle nature. The YP7-YZ3/ (YCeSr)Z7-A3/A/(YCeSr)Z7-A3 laminate (the Y(c) system) had a flexural strength of 217 MPa with a catastrophic failure from one test. SEM micrographs showing the side surfaces of a Y(b) laminate and a Y(c) laminate after 4-point flexural testing are displayed in Fig. 9(a) and Fig. 9(b). Some laminates did not fabricate successfully. YPO4/ Y-ZrO2 (the Y(d) system) and LaPO4/Y-ZrO2/Z3-A7/ Y-ZrO2 (the L(d) system) shattered after removal from the hot press die. The previously mentioned LaPO4/YZrO2 laminate stayed intact after removal from the die and annealing, but shattered during bend bar cutting. Load-displacement responses of 4-pt flexural tests are presented in Fig. 10(a) for an AlPO4/Al2O3 (A(a)) laminate and for an (AlPO4+Al2O3)/Al2O3 (A(b)) laminate. A flexural strength of 47 MPa was measured for the AlPO4/Al2O3 laminate and of 225 MPa for the (AlPO4+Al2O3)/Al2O3 laminate, with a noncatastrophic fracture from one test for each. A side view of the non-brittle A(b) bar was examined under SEM and is shown in Fig. 10(b). Cracks propagating laterally were responsible for this non-brittle fracture behavior. Vickers indentation cracks were also introduced to the A(a) and A(b) laminates, at an orientation of 45° (Fig. 11) relative to the layer length direction. On propagating from the (AlPO4+Al2O3) layers towards the Al2O3 layers, the radial cracks preferentially deflected along the (AlPO4+Al2O3)/Al2O3 interface. Fig. 8. SEM micrograph showing the side surface of a LaPO4/70 vol% (YCeSr)ZrO2 –30 vol% mullite/Al2O3/70 vol% (YCeSr)ZrO2 –30 vol% mullite laminate after 4-point flexural testing. Fig. 6. Load versus displacement curves of a LaPO4/70 vol% (YCeSr)ZrO2 –30 vol% Al2O3 laminate tested in 4-point flexure
246 D -H. Kuo, W.M. Kricen/ Materials Science and Engineering 4241(1998)241-250 as a LaAl Oi8 product. To adjust the thermal expansion of the YPO/ZrO2 laminates(system Y(d) which shat tered after fabrication. 30 vol% Y-Zro, and 30 vol% (YCeSr)Z7-A3 AlO, were incorporated into the YPO4 layers and YP7-YZ3 Y-ZrO, layers(system Y(b)), respectively, in consider (YCeSr)Z7-A3 ing the chemical and physical stability The Y-ZrO2 layers in the Y(a) laminate acted as a diffusion barrier between the YPO4 and Yz3-A7 layers where a chemical reaction between YPO4 and AlO an form yttrium aluminate (Y3AlsOj2) A symmetrical stacking in this study can prevent the laminates from distorting. As discussed by Hull and Clyne [22], a symmetrical stacking has the effect of preventing shape distortion of the laminates and local microstructural damage and failure ymmetrical laminates, the coupling forces, due to property differ ence at each layer, largely cancel out and hold the laminate without distortion, even though there are still local stresses across the interlaminar boundaries YCeSr)Z7-A3 4.2. Mechanical evaluation of laminated composites YP7-YZ3 4.2.1. The YPO4/Y-ZrO2/YZ3-A7/Y-ZrO, oxide This Y(a) laminate stacking in a four-layer configura ion was demonstrated as a successful design of a ceramic composite with high strength and high appar ent work-of-fracture(Fig. 3). Strong three-layer Y ZrO2/YZ3-A7/Y-ZrO2(ZrO2-containing) matrices and weak YPOa layers(having a low fracture strength of 75 MPa [23] stacked alternatively between the matrix Fig 9. SEM micrographs showing the side surfaces of (a)a 70 vol% YPO2-30 vol% Y-ZrO,70 vol%(YCeSr-ZrO2 -30 vol% Al,O, lami layers, provided the required strength. The interfaces nate and(b)a 70 vol% YPO4-30 vol% Y-ZrO2/70 vol%(YCeSr)- between YPO4 layers and matrix layers supplied the ZrO, -30 vol%AL, O,/AL, O,/70 vol%(YCeSr)-ZrO2-30 vol% AL, O, weak interfaces. On the other hand, the laminate com- laminate after 4-point flexural testing. posed of only YPOA and Y-ZrO2 shattered, due to the large thermal expansion mismatch between YPO.(co- The weak nature of AlPO4 was demonstrated by the efficient of thermal expansion, a=86x10/C)[5 severe damage in the AlPO4 layer after indentation. and Y-ZrO,(a 10.6×10-6/°C[24 The incorpo Mechanical responses of the ten kinds of YPO4, Lapo4, rated YZ3-A7 layers had a coefficient of thermal and AlPO4 laminates are listed in Table 1 expansion of~9.3×10-6/ C based upon a=~8.8× 10/C for Al,O3 [25]. Although the YZ3-A7 layers did not participate in providing a weak interface, their 4. Discussion function was to increase the stiffness and modify the residual stresses in the oxide laminate. which enabled 4.1. Chemical compatibility and structu the fabrication of the Y(a) laminate. From the compari- son with the YPO/Y-ZrO, laminate, there were two The YPO4, LaPO4, and AlPOa laminate structures reasons for this successful oxide laminate with extended were based upon the chemical compatibility between interfacial delamination. One was the nature of the YPO4 and Y-ZrO2(Fig. 1), between LaPO4 and Y- YPO4/Y-ZrO2 interface. Residual stress-assisted delam ZrO,(Fig. 1), and between AlPO4 and Al,O,( Fig. 2), ination was the other with minor modifications to reach the requirement of There was a close relation between load drops an structural stability. For example, the thermal expansion cking/delamination events. From the observation of misfit in system L(a) can be abated by incorporating Fig. 4(b), eight delaminated interfaces were displayed AlO, into Y-ZrO2 layers as shown in system L(b). The which left discrete composite layers composed of YPO4 amount of 30 vol% Al,O3 in Y-ZrO2 can ensure limited Y-ZrO2, and YZ3-A7 laminae between delaminated contacts between Al,O, and LaPO4, which can produce interfaces. Only two of the matrix composite layers
246 D.-H. Kuo, W.M. Kri6en / Materials Science and Engineering A241 (1998) 241–250 Fig. 9. SEM micrographs showing the side surfaces of (a) a 70 vol% YPO4 –30 vol% Y-ZrO2/70 vol% (YCeSr)-ZrO2 –30 vol% Al2O3 laminate and (b) a 70 vol% YPO4 –30 vol% Y-ZrO2/70 vol% (YCeSr)- ZrO2 –30 vol% Al2O3/Al2O3/70 vol% (YCeSr)-ZrO2 –30 vol% Al2O3 laminate after 4-point flexural testing. a LaAl11O18 product. To adjust the thermal expansion of the YPO4/ZrO2 laminates (system Y(d)) which shattered after fabrication, 30 vol% Y-ZrO2 and 30 vol% Al2O3 were incorporated into the YPO4 layers and Y-ZrO2 layers (system Y(b)), respectively, in considering the chemical and physical stability. The Y-ZrO2 layers in the Y(a) laminate acted as a diffusion barrier between the YPO4 and YZ3-A7 layers, where a chemical reaction between YPO4 and Al2O3 can form yttrium aluminate (Y3Al5O12). A symmetrical stacking in this study can prevent the laminates from distorting. As discussed by Hull and Clyne [22], a symmetrical stacking has the effect of preventing shape distortion of the laminates and local microstructural damage and failure. In symmetrical laminates, the coupling forces, due to property difference at each layer, largely cancel out and hold the laminate without distortion, even though there are still local stresses across the interlaminar boundaries. 4.2. Mechanical e6aluation of laminated composites 4.2.1. The YPO4/Y-ZrO2/YZ3-A7/Y-ZrO2 oxide laminate This Y(a) laminate stacking in a four-layer configuration was demonstrated as a successful design of a ceramic composite with high strength and high apparent work-of-fracture (Fig. 3). Strong three-layer YZrO2/YZ3-A7/Y-ZrO2 (ZrO2-containing) matrices and weak YPO4 layers (having a low fracture strength of 75 MPa [23]) stacked alternatively between the matrix layers, provided the required strength. The interfaces between YPO4 layers and matrix layers supplied the weak interfaces. On the other hand, the laminate composed of only YPO4 and Y-ZrO2 shattered, due to the large thermal expansion mismatch between YPO4 (coefficient of thermal expansion, a= 8.6×10−6 /°C) [5] and Y-ZrO2 (a= 10.6×10−6 /°C) [24]. The incorporated YZ3-A7 layers had a coefficient of thermal expansion of 9.3×10−6 /°C based upon a= 8.8× 10−6 /°C for Al2O3 [25]. Although the YZ3-A7 layers did not participate in providing a weak interface, their function was to increase the stiffness and modify the residual stresses in the oxide laminate, which enabled the fabrication of the Y(a) laminate. From the comparison with the YPO4/Y-ZrO2 laminate, there were two reasons for this successful oxide laminate with extended interfacial delamination. One was the nature of the YPO4/Y-ZrO2 interface. Residual stress-assisted delamination was the other. There was a close relation between load drops and cracking/delamination events. From the observation of Fig. 4(b), eight delaminated interfaces were displayed, which left discrete composite layers composed of YPO4, Y-ZrO2, and YZ3-A7 laminae between delaminated interfaces. Only two of the matrix composite layers The weak nature of AlPO4 was demonstrated by the severe damage in the AlPO4 layer after indentation. Mechanical responses of the ten kinds of YPO4, LaPO4, and AlPO4 laminates are listed in Table 1. 4. Discussion 4.1. Chemical compatibility and structural stability The YPO4, LaPO4, and AlPO4 laminate structures were based upon the chemical compatibility between YPO4 and Y-ZrO2 (Fig. 1), between LaPO4 and YZrO2 (Fig. 1), and between AlPO4 and Al2O3 (Fig. 2), with minor modifications to reach the requirement of structural stability. For example, the thermal expansion misfit in system L(a) can be abated by incorporating Al2O3 into Y-ZrO2 layers as shown in system L(b). The amount of 30 vol% Al2O3 in Y-ZrO2 can ensure limited contacts between Al2O3 and LaPO4, which can produce
D -H. Kuo, W.M. Kricen/ Materials Science and Engineering 4241(1998)241-250 o(max)225 MP (AIPO +AlOAlO laminate lamination d(max)=47 MPa AlPO /AlO laminate 0. 15 Crosshead Displacement(mm) ig. 10.(a) Load-displacement curves for the AlPO4/Al,O, and(AlPO4+ Al,O3)ALO, laminates after 4-point flexural tests and(b)the sid SEM micrograph of a fractured(AlPO4+ AL,O3/Al,O, laminate broke. The small load drop in Fig. 3 was related to microcracking, which operate in most particulate and interface delamination, while a large load drop corre- whisker-reinforced ceramic composites sponded to breakage of a composite matrix layer. Eight YPO -containing Y3Als O,2 laminates with different small load drops in Fig 3 corresponded to eight delam- configurations have previously been investigated [7] inated cracks. Two big load drops in Fig. 3 corre- Two specific examples were laminate systems with a sponded to the breakage of two composite layers. A repeating sequence of YPO4/Y3AlsO12/Al_O3/Y3AlsO, small deviation in the responses could be attributed to and of YPO4Y3AlsO12. After 4-point flexural testing the spontaneous events of cracking and delaminati these laminates displayed catastrophic fracture, al During loading, the Y(a) laminate delaminated to though there was limited interfacial delamination in the partially dissipate the strain energy(accumulated from first laminate and severe crack deflection in the second the applied load and residual strain energy) by creating case. Similar results have been observed in the Y(b) free surfaces. After the first interfacial delamination at laminate(Fig. 9(a)), the Y(c)laminate(Fig. 9(b)) and the mid-plane, the load was redistributed and sup he L(c) laminate(Fig. 8). Comparing the fracture ported by the debonded Y-ZrO2/YZ3-A7/Y-ZrO2/ behaviors of these examples with that of the Y(a) YPO4 layers and the unbroken part. At this stage, the laminate, we understand that:(i) limited delamination ZrO2-containing layers played an important role. These and crack deflection did not benefit damage tolerance strong layers could support the applied load, which and (ii) only pronounced interfacial delamination can kept the delaminated oxide laminate from fracturing enhance the ceramic's flaw tolerance catastrophically. As the cracking/delamination events From the qualitative indentation test(Fig 4), we can continued, a non-brittle fracture response was achieved. also obtain important information. It was noted that After going through several cracking/delamination both radial cracks and cracks along the YPO/Y-ZrO2 events, the bar was severely bent At the same time, the interfaces were identified in the y(a)laminate. The first delaminated interface(the mid-plane)was severely generation of YPO4/Y-ZrO2 interfacial cracks, not the damaged(Fig 4(b). This fracture behavior is similar to radial cracks indicated a weak behavior in those inter. that occurring in fiber-reinforced ceramic composites faces. The propagation of radial cracks through lami- having a weak interface viz., interfacial debonding and nae suggested that strength could benefit from these delamination are followed by load redistribution among strong and brittle parts. These results were consistent the unfractured part and the unbroken fibers. Severe with the outcome obtained from flexural testing interfacial delamination as seen in this study is ar Typical non-oxide laminates with good mechanical mportant mechanism operating in flaw-tolerant ceram- properties have been reported (Table 2). The Y(a) oxide ics. This mechanism is not preferred for polymer laminate of this work has demonstrated excellent re- metal composites, where strength is the main concern. sults: a 4-point flexural strength of 358 and 392 MPa Nevertheless, it is an important damage-tolerant mech- and apparent work-of-fracture values of 8.2 and 10 anism for ceramics, which can reduce the brittleness of kJ/m, respectively, in two separate tests. Thus, this ceramics. It is a more effective toughening mechanism oxide laminate had comparable mechanical properties than those of crack deflection, crack branching, and with those of known non-oxide composites. As com
D.-H. Kuo, W.M. Kri6en / Materials Science and Engineering A241 (1998) 241–250 247 Fig. 10. (a) Load-displacement curves for the AlPO4/Al2O3 and (AlPO4+Al2O3)/Al2O3 laminates after 4-point flexural tests and (b) the side-view SEM micrograph of a fractured (AlPO4+Al2O3)/Al2O3 laminate. broke. The small load drop in Fig. 3 was related to interface delamination, while a large load drop corresponded to breakage of a composite matrix layer. Eight small load drops in Fig. 3 corresponded to eight delaminated cracks. Two big load drops in Fig. 3 corresponded to the breakage of two composite layers. A small deviation in the responses could be attributed to the spontaneous events of cracking and delamination. During loading, the Y(a) laminate delaminated to partially dissipate the strain energy (accumulated from the applied load and residual strain energy) by creating free surfaces. After the first interfacial delamination at the mid-plane, the load was redistributed and supported by the debonded Y-ZrO2/YZ3-A7/Y-ZrO2/ YPO4 layers and the unbroken part. At this stage, the ZrO2-containing layers played an important role. These strong layers could support the applied load, which kept the delaminated oxide laminate from fracturing catastrophically. As the cracking/delamination events continued, a non-brittle fracture response was achieved. After going through several cracking/delamination events, the bar was severely bent. At the same time, the first delaminated interface (the mid-plane) was severely damaged (Fig. 4(b)). This fracture behavior is similar to that occurring in fiber-reinforced ceramic composites having a weak interface viz., interfacial debonding and delamination are followed by load redistribution among the unfractured part and the unbroken fibers. Severe interfacial delamination as seen in this study is an important mechanism operating in flaw-tolerant ceramics. This mechanism is not preferred for polymer and metal composites, where strength is the main concern. Nevertheless, it is an important damage-tolerant mechanism for ceramics, which can reduce the brittleness of ceramics. It is a more effective toughening mechanism than those of crack deflection, crack branching, and microcracking, which operate in most particulate and whisker-reinforced ceramic composites. YPO4-containing Y3Al5O12 laminates with different configurations have previously been investigated [7]. Two specific examples were laminate systems with a repeating sequence of YPO4/Y3Al5O12/Al2O3/Y3Al5O12 and of YPO4/Y3Al5O12. After 4-point flexural testing, these laminates displayed catastrophic fracture, although there was limited interfacial delamination in the first laminate and severe crack deflection in the second case. Similar results have been observed in the Y(b) laminate (Fig. 9(a)), the Y(c) laminate (Fig. 9(b)) and the L(c) laminate (Fig. 8). Comparing the fracture behaviors of these examples with that of the Y(a) laminate, we understand that: (i) limited delamination and crack deflection did not benefit damage tolerance and (ii) only pronounced interfacial delamination can enhance the ceramic’s flaw tolerance. From the qualitative indentation test (Fig. 4), we can also obtain important information. It was noted that both radial cracks and cracks along the YPO4/Y-ZrO2 interfaces were identified in the Y(a) laminate. The generation of YPO4/Y-ZrO2 interfacial cracks, not the radial cracks, indicated a weak behavior in those interfaces. The propagation of radial cracks through laminae suggested that strength could benefit from these strong and brittle parts. These results were consistent with the outcome obtained from flexural testing. Typical non-oxide laminates with good mechanical properties have been reported (Table 2). The Y(a) oxide laminate of this work has demonstrated excellent results: a 4-point flexural strength of 358 and 392 MPa and apparent work-of-fracture values of 8.2 and 10 kJ/m2 , respectively, in two separate tests. Thus, this oxide laminate had comparable mechanical properties with those of known non-oxide composites. As com-
D -H. Kuo, W.M. Kricen/ Materials Science and Engineering 4241(1998)241-250 a Work-of-fracture of many types of materials [27] Work of fracture(kJ/m) Dural Teak wood Toughened polystyrene te: apparent work-of-fracture is defined by the load-displacement curve by the fracture surfaces(2 tional area). posites, how the strength of each functional layer could AlPO4+ be increased. For example, the successful yttria stabi- lized Zro, 20 wt% Al,O3 with strengths as high as 2.5 GPa could replace the Y-ZrO2 to enhance the strength A two-phase YPOa-containing composite could replace the YPO4 to enhance its mechanical properties. To strengthen the high temperature properties, creep resis- tant materials, e.g. particulates, whiskers, or fibers, can 50m be incorporated into the laminates. Thus, strong and damage-tolerant materials for high temperature oxidiz ing applications can be achieved Fig. Il. SEM micrographs of indention crack patterns (a) in an AIPOJAlO, laminate and (b) in an(AlPO+ALO3)/Al,O, lami nate Indents were oriented at 45 relative to the layer length direc- 4.2.2. The LaPO/(yCeSr)Z7-A3 oxide laminate The second successful laminate(Fig. 6) had the LaPOa-(Y CeSr )Z7-A3 stacking sequence with two pared with different materials in Table 3 [27]. the Y(a) kinds of layers(system L(b)). This L(b) laminate was laminate has a higher apparent work-of-fracture than fabricated by modifying the disintegrated LaPO4/Y those of wood, polymer, and ceramics. A 100-fold ZrO2 system(system L(a)) with 30 vol% alumina mixed increase is apparent as compared with alumina. Note: with Y-ZrO,. In this way, the L(b) laminate had a apparent work-of-fracture in Table 3 is obtained by lower mismatch in coefficients of thermal expansion.In considering the fracture surfaces(2 x cross-sectional considering the chemical compatibility and thermal ex area), while Table 2 considers the cross-sectional area pansion mismatch, the two-layered configuration had some limits which prevented it from obtaining good This Y(a)laminate is a simple hybrid laminate. One properties. For these reasons, a four-layered configura- may speculate, from studies in ZrO -containing com tion obtained by stacking three kinds of layers,as Property comparisons of an oxide laminate in this study with damage-tolerant non-oxides System Strength (MPa) Work of Fracture (kJ/m-) Reference hite laminate 4.6and6.7 1-1.3 SiC/BN fibrous ceramic 300-375(4pt) [14 196and437(4pt 5.5and6.5 Carbon fiber reinforced gla 00(3-pt) YPO/Y-ZrO,/YZ3-A7/Y-ZrO2 laminate 358and392(4pt) 8.2and10 This study
248 D.-H. Kuo, W.M. Kri6en / Materials Science and Engineering A241 (1998) 241–250 Fig. 11. SEM micrographs of indentaion crack patterns (a) in an AlPO4/Al2O3 laminate and (b) in an (AlPO4+Al2O3)/Al2O3 laminate. Indents were oriented at 45° relative to the layer length direction. Table 3 Work-of-fracture of many types of materials [27] Material Work of fracture (kJ/m2 ) Dural 140 Copper 50 Key steel 50 Brass 30 Teak wood 6 Cast iron 4 Toughened polystyrene 4 Deal wood 2 Cellulose 2 Polystyrene 1 Reactor graphite 0.1 Firebrick 0.02–0.07 Alumina 0.04 Note: apparent work-of-fracture is defined by dividing the area under the load-displacement curve by the fracture surfaces (2×cross-sectional area). posites, how the strength of each functional layer could be increased. For example, the successful yttria stabilized ZrO2/20 wt% Al2O3 with strengths as high as 2.5 GPa could replace the Y-ZrO2 to enhance the strength. A two-phase YPO4-containing composite could replace the YPO4 to enhance its mechanical properties. To strengthen the high temperature properties, creep resistant materials, e.g. particulates, whiskers, or fibers, can be incorporated into the laminates. Thus, strong and damage-tolerant materials for high temperature oxidizing applications can be achieved. 4.2.2. The LaPO4/(YCeSr)Z7-A3 oxide laminate The second successful laminate (Fig. 6) had the LaPO4(YCeSr)Z7-A3 stacking sequence with two kinds of layers (system L(b)). This L(b) laminate was fabricated by modifying the disintegrated LaPO4/YZrO2 system (system L(a)) with 30 vol% alumina mixed with Y-ZrO2. In this way, the L(b) laminate had a lower mismatch in coefficients of thermal expansion. In considering the chemical compatibility and thermal expansion mismatch, the two-layered configuration had some limits which prevented it from obtaining good properties. For these reasons, a four-layered configuration obtained by stacking three kinds of layers, as pared with different materials in Table 3 [27], the Y(a) laminate has a higher apparent work-of-fracture than those of wood, polymer, and ceramics. A 100-fold increase is apparent as compared with alumina. Note: the apparent work-of-fracture in Table 3 is obtained by considering the fracture surfaces (2×cross-sectional area), while Table 2 considers the cross-sectional area only. This Y(a) laminate is a simple hybrid laminate. One may speculate, from studies in ZrO2-containing comTable 2 Property comparisons of an oxide laminate in this study with damage-tolerant non-oxides Work of Fracture (kJ/m2 System Strength (MPa) ) Reference SiC/graphite laminate 640 (3-pt) 4.6 and 6.7 [9,10] SiC/graphite fibrous ceramic 215–235 (4-pt) 1–1.3 [13] SiC/BN fibrous ceramic 2.4 [14] 300–375 (4-pt) Si3N4/BN laminate 5.5 and 6.5 196 and 437 (4-pt) [17] Carbon fiber reinforced glass 500 (3-pt) 8–9 [26] YPO 358 and 392 (4-pt) 4/Y-ZrO2/YZ3-A7/Y-ZrO2 laminate 8.2 and 10 This study
D -H. Kuo, W.M. Kricen/ Materials Science and Engineering 4241(1998)241-250 demonstrated by the Y(a) laminate, can have both that YPO4- containing ZrO2 laminates can survive if excellent strength as well as damage tolerance △x<1.0×10-6/°C, but shattered if△x=2.0×10-6 4.2.3. The AlPOa-containing oxide laminates systems with△x=0.9×10- c have not enough The AlPO/Al,O laminate was too weak to be useful residual shear stress to initiate interfacial delamination. (Fig. 10). The weakness of the AlPO4 in A(a) and A(b) Therefore, brittle fracture dominates. For LaPOa-con laminates was qualitatively demonstrated by the inden- taining ZrO2 systems, the laminates survived if Ax tation technique(Fig. 11). Severe chipping around a 1.0x 10oC, but shattered as Ao=1.0 x 10oC Vickers indent in the AlPO4- containing layers explained The L(b) system could have adequate interfacial shear the weakness of the AlPO4 layer. The AlPOA laminae strength to have good strength and to enhance interfa (Fig. 1l(a)) in the A(a) laminate demonstrated worse cial delamination, while the L(c) system with the weak damage than did the(AlPO4+Al,O3) laminae(Fig LaPO4 layers under residual tensile stresses had a 11(b))in the A(b) laminate. It was also noted that the duced strength and a brittle fracture. Shattering of the generated indent cracks did not penetrate into the four-layer L(d)system, that did not behave as the Al,O3 layers during 45 indents Indent cracks preferred four-layer Y(a) laminate, could originate from the ther propagating along or deflecting into interfaces. As the mal stresses and the low strength of un-stabilized Zro indent cracks propagated from the weak (AlPO4+ It appears that laminates display brittle behavior when Al,O3) laminae toward the strong Al, O3 laminae thermal expansion misfit is small, while large misfit will cracks were deflected into interfaces. There were many cause laminates to disintegrate in the two-layered lateral cracks propagating along interfaces, but no configuration. For the AlPOA-containing Al_O3 lami cracks were propagating perpendicularly through inter- nates, the residual stresses played a minor role, as faces. This weakness was improved in the(AlPO 4 evidenced by the laminate integrity. From the different ALO3/ALO, laminate by having composite laminae of results observed in YPO4- and LaPO-containing ZrO (AlPO4+Al,O3) to strengthen the AlPO4 laminae. As and AlPO4- containing Al,O3 laminates, we have a bet ompared with the AlPO4 laminates, the indent crack ter understanding of oxide laminate design. Ideally attern for a YPO- containing Zro laminate in Fig. 5 strong laminae with weak interfaces are the general demonstrated the case for strong laminae and weak requ interfaces without weak laminae. In this way, this lami- quence as in the YPOa-containing Zro2 laminate can nate could display both high strength and apparent compensate for insufficiently weak interfaces to yield work-of-fracture better strength and damage-tolerance Apparently, a weak interface exists between AlPO4 Although CFCCs are considered to be poter and Al2O3. There are also weak laminae of AlPO and terials for high toughness and high strength applica AlPO4+Al,O3)in AlPO4-containing Al2O3 laminates. tions, the development of oxide CFCCs for high Although the AlPO -containing Al2O3 systems did not temperature oxidizing environments poses considerable exhibit better mechanical properties, these systems have difficulties yet to be overcome. These difficulties include displayed promising interface delamination availability and stability of oxidation-resistant fibers and weak fiber/matrix interfaces. High strength and 4.2.4. Residual stress effect on the fractural behavior of damage-tolerant oxide laminates can provide a solution YPO4 LaPO4 and AlPO, oxide laminates to reach the demands of high temperature applications From Table l, different behaviors occurred in the in an oxidizing environment. YPO4, LaPO4 and AlPO4 laminates. Although the lam In the system Y(a) of a YPOa- containing ZrO2 lami- inate behaviors are determined by very complex factors nate, we have achieved a material with both high such as the elastic and thermal properties of each layer, strength and high work-of-fracture (non-brittleness) layer stacking sequence, layer thickness, etc, the differ- Usually, ceramic strengths need to be compromised if ent laminate behaviors can be qualitatively explained toughness is considered. We have demonstrated pro- by the role of thermal expansion misfit between the gress in oxide composites, but many questions remain YPO4" LaPO4 or AlPO4-containing layers and the to be further investigated. These include the material adjacent layer (Ax=a-x(YPO4), a-a(LaPO) or response under tensile loading and under high tempera -a(AlPO4)). The thermal expansion coefficient of ture loading conditions each layer was obtained from measurements, literature or the rule of mixtures The values that were used are 8.6, 9.6, 10.6, 8.8 and 5.3 x 10/'C for YPO4 [51. 5. Conclusions LaPO4 [4], zirconia [24], AlO3 [25] and mullite [25] Calculation of residual stresses, which involve many In this work, YPO4 and LaPOa-containing Zro difficulties for the four-layer configurations, is not a laminates and AlPO4- containing Al2O3 laminates were main concern in this study. We can conclude generally fabricated and studied. We demonstrate the feasibility
D.-H. Kuo, W.M. Kri6en / Materials Science and Engineering A241 (1998) 241–250 249 demonstrated by the Y(a) laminate, can have both excellent strength as well as damage tolerance. 4.2.3. The AlPO4-containing oxide laminates The AlPO4/Al2O3 laminate was too weak to be useful (Fig. 10). The weakness of the AlPO4 in A(a) and A(b) laminates was qualitatively demonstrated by the indentation technique (Fig. 11). Severe chipping around a Vickers indent in the AlPO4-containing layers explained the weakness of the AlPO4 layer. The AlPO4 laminae (Fig. 11(a)) in the A(a) laminate demonstrated worse damage than did the (AlPO4+Al2O3) laminae (Fig. 11(b)) in the A(b) laminate. It was also noted that the generated indent cracks did not penetrate into the Al2O3 layers during 45° indents. Indent cracks preferred propagating along or deflecting into interfaces. As the indent cracks propagated from the weak (AlPO4+ Al2O3) laminae toward the strong Al2O3 laminae, cracks were deflected into interfaces. There were many lateral cracks propagating along interfaces, but no cracks were propagating perpendicularly through interfaces. This weakness was improved in the (AlPO4+ Al2O3)/Al2O3 laminate by having composite laminae of (AlPO4+Al2O3) to strengthen the AlPO4 laminae. As compared with the AlPO4 laminates, the indent crack pattern for a YPO4-containing ZrO2 laminate in Fig. 5 demonstrated the case for strong laminae and weak interfaces without weak laminae. In this way, this laminate could display both high strength and apparent work-of-fracture. Apparently, a weak interface exists between AlPO4 and Al2O3. There are also weak laminae of AlPO4 and (AlPO4+Al2O3) in AlPO4-containing Al2O3 laminates. Although the AlPO4-containing Al2O3 systems did not exhibit better mechanical properties, these systems have displayed promising interface delamination. 4.2.4. Residual stress effect on the fractural beha6ior of YPO4, LaPO4 and AlPO4 oxide laminates From Table 1, different behaviors occurred in the YPO4, LaPO4 and AlPO4 laminates. Although the laminate behaviors are determined by very complex factors such as the elastic and thermal properties of each layer, layer stacking sequence, layer thickness, etc., the different laminate behaviors can be qualitatively explained by the role of thermal expansion misfit between the YPO4-, LaPO4- or AlPO4-containing layers and the adjacent layer (Da=a−a(YPO4), a−a(LaPO4) or a−a(AlPO4)). The thermal expansion coefficient of each layer was obtained from measurements, literature or the rule of mixtures. The values that were used are 8.6, 9.6, 10.6, 8.8 and 5.3×10−6 /°C for YPO4 [5], LaPO4 [4], zirconia [24], Al2O3 [25] and mullite [25]. Calculation of residual stresses, which involve many difficulties for the four-layer configurations, is not a main concern in this study. We can conclude generally that YPO4-containing ZrO2 laminates can survive if DaB1.0×10−6 /°C, but shattered if Da=2.0×10−6 / °C except for the Y(a) system. The Y(b) and Y(c) systems with Da=0.9×10−6 /°C have not enough residual shear stress to initiate interfacial delamination. Therefore, brittle fracture dominates. For LaPO4-containing ZrO2 systems, the laminates survived if DaB 1.0×10−6 /°C, but shattered as Da=1.0×10−6 /°C. The L(b) system could have adequate interfacial shear strength to have good strength and to enhance interfacial delamination, while the L(c) system with the weak LaPO4 layers under residual tensile stresses had a reduced strength and a brittle fracture. Shattering of the four-layer L(d) system, that did not behave as the four-layer Y(a) laminate, could originate from the thermal stresses and the low strength of un-stabilized ZrO2. It appears that laminates display brittle behavior when thermal expansion misfit is small, while large misfit will cause laminates to disintegrate in the two-layered configuration. For the AlPO4-containing Al2O3 laminates, the residual stresses played a minor role, as evidenced by the laminate integrity. From the different results observed in YPO4- and LaPO4-containing ZrO2 and AlPO4-containing Al2O3 laminates, we have a better understanding of oxide laminate design. Ideally, strong laminae with weak interfaces are the general requirements. Laminates with a complex stacking sequence as in the YPO4-containing ZrO2 laminate can compensate for insufficiently weak interfaces to yield better strength and damage-tolerance. Although CFCCs are considered to be potential materials for high toughness and high strength applications, the development of oxide CFCCs for high temperature oxidizing environments poses considerable difficulties yet to be overcome. These difficulties include availability and stability of oxidation-resistant fibers and weak fiber/matrix interfaces. High strength and damage-tolerant oxide laminates can provide a solution to reach the demands of high temperature applications in an oxidizing environment. In the system Y(a) of a YPO4-containing ZrO2 laminate, we have achieved a material with both high strength and high work-of-fracture (non-brittleness). Usually, ceramic strengths need to be compromised if toughness is considered. We have demonstrated progress in oxide composites, but many questions remain to be further investigated. These include the material response under tensile loading and under high temperature loading conditions. 5. Conclusions In this work, YPO4- and LaPO4-containing ZrO2 laminates and AlPO4-containing Al2O3 laminates were fabricated and studied. We demonstrate the feasibility
250 D -H. Kuo, W.M. Kricen/ Materials Science and Engineering 4241(1998)241-250 of strong and non-brittle oxides without using the 8 D H. Kuo, W.M. Kriven, Interface properties of oxidation-sensitive C and BN, or expensive fibers. The Y3AlsO12 systems: A double-sandwich system and a fiber YPO4/Y-Zro2YZ3-A7/Y-ZrO2 oxide laminate with composite, Ceram. Trans. 74(1996)71-82 high strength and work-of-fracture exhibits extended 9 w.J. Clegg, K. Kendall, N.M. Alford, D. Birchall, T.w. Button, A simple way to make tough ceramics, Nature 347(1990) interfacial delamination rarely seen in ceramics. This omposites, Acta Metall. 40(ID)(1992)3085-309 2(1990) YPO -containing Zro, oxide is a laminate in a four- [o] w.J. Clegg, The fabrication and failure of laminar cer layer configuration rather than a conventional two layer configuration. A mechanism of residual [l] C.A. Folsom, F.W. Zok, FF. Lange, D B. Marshall, Mechanical interlaminar shear stress-enhanced, delamination is pro ehavior of a laminar ceramic/fiber-reinforced epoxy composite, J.Am. Ceram.Soc.75(11)(1992)2969-2975 posed to contribute to this behavior. As compared with [12]S.Baskaran,SD.Nunn,D Popovic,JW.Halloran,Fibrous this promising result, the different results of other oxide onolithic ceramics: I, Fabrication, microstructure. and indenta- laminates plained by the residual stress effect, tion behavior, J. Am. Ceram. Soc. 76(9)(1993)2209-2216 lamina strength and stacking sequence. Insufficient [3]S. Baskaran, S.D. Nunn, D. Popovic, J.w. Halloran, Fibrous monolithic ceramics: Il. Flexural strength and fracture behavior residual stresses cannot initiate the residual shear-en- of the silicon carbide/graphite system, J. Am. Ceram Soc. 76(9) hanced flexural mechanism, but high values can dam age materials [14]S. Baskaran, J.W. Halloran, Fibrous monolithic ceramic si arbide/ boron nitride system, J. Am. Ceram Soc. 77(5)(1994) Acknowledgements [15 M.C. Shaw, D B. Marshall, M.S. Dadkhah, A.G. Evans, Crack. ng and damage mechanisms in ceramic/metal multilayers, Acta Funding for this work was provided by the U.S. Air Metall..4l(11)(1993)33l1-3322. Force Office of Scientific Research through Dr A [] Z. Chen, J.J. Mecholsky Jr, Toughening by metallic lamina in Pechenik under grant number AFOSR-F49620-93-1 nickel/alumina composites, J. Am. Ceram. Soc. 76(5)(1993) 1258-1264 0027. The authors thank Dr David B. Marshall of [17 H Liu, S M. Hsu, Fracture behavior of multilayer silicon ni- Rockwell Science Center for his information regarding tride/ boron nitride ceramics, J. Am. Ceram Soc. 79(9)(1996) a LaPOa-containing ZrO, laminat 2452-2457 [18 M. Pechini, Method of Preparing Lead and Alkaline- Earth Titanates and Coating Method Using the Same to Form a Capacitor, U.S. Patent No. 3, 330,697, July l1, 1967. References [9] D H. Kuo, W.M. Kriven, Chemical stability, microstructure, d mechanical behavior of LaPO- containing ceramics, Mater [A.G. Evans, Perspective on the development of high-toughness Sci.EngA210(1-2)(1996)123-134 ceramics, J. Am. Ceram Soc. 73(2)(1990)187-206 [20]M. Hirano, H. Inada, Fabrication and properties of yttria- and 2 P.E. D. Morgan, D B. Marshall, Functional interfaces for oxide a-doped tetragonal zirconia /alumina composites, Br. Ceram. oxide composites, Mater. Sci. Eng. A162(1993)15-25. Trans.J.90(1991)48-51 3 P.E.D. Morgan, D B. Marshall, R. M. Housley, High tempera- [21]RA. Cutler, R.J. Mayhew, K M. Prettyman, A.v. Virkar, High- ture stability of monazite-alumina composites, Mater. Sci. Eng toughness Ce-TZP/AL2O, ceramics with improved hardness and A195(1995)215-222. strength, J. Am. Ceram. Soc. 74(1)(1991)179-18 4 P.E.D. Morgan, D B. Marshall, Ceramic composites of monazite 22 D. Hull, T.w. Clyne, in: An Introduction to Composite Materi- and alumina, J. Am. Ceram. Soc. 78(6)(1995)1553-1563 Is, 2nd edn, Cambridge University Press, Cambridge, 1996 Erratum: J. Am. Ceram. Soc. 78(9)(1995)2574 [23]DH. Kuo, Investigation of Oxide fiber/Oxide Matrix Com- [D.H. Kuo, W.M. Kriven, Characterization of yttrium phosphate sites with a Weak Interphase, Ph D. thesis, University of Illinois at Urbana-Champaign, Urbana, IL, 1997. [24] D.J. Green, R.H.J. Hannink, M.V. Swain, Transformation 6 D.H. Kuo, W.M. Kriven, Microstructure and mechanical re- oughening of Ceramics, CRC Press, Boca Raton, FL, 1985 sponse of lanthanum phosphate/yttrium aluminate and yttrium [25] W.D. Kingery, H.K. Bowen, D.R. Uhlmann, Introduction to phosphate/yttrium aluminate systems, Ceram. Sci. Eng. Proc. Ceramics, 2nd edn, John Wiley, New York, 1976 17B(1996)233-240 [26D. C. Phillips, The fracture energy of carbon-fiber reinforced [7 D.H. Kuo, W.M. Kriven, Development of yttrium phosphate as glass, J. Mater. Sci. 7(1972)1175-1191 an interphase for oxide/oxide composites, presented at the 2n [27] H.G. Tattersall, G. Tappin, The work of fracture and its mea- Meeting of Pacific Rim Ceramic Societies, Cairns, Australia, Jul urement in metals. ceramics and other materials. J Mater. Sci. 1(1966)296-301
250 D.-H. Kuo, W.M. Kri6en / Materials Science and Engineering A241 (1998) 241–250 of strong and non-brittle oxides without using the oxidation-sensitive C and BN, or expensive fibers. The YPO4/Y-ZrO2/YZ3-A7/Y-ZrO2 oxide laminate with high strength and work-of-fracture exhibits extended interfacial delamination rarely seen in ceramics. This YPO4-containing ZrO2 oxide is a laminate in a fourlayer configuration rather than a conventional twolayer configuration. A mechanism of residual, interlaminar shear stress-enhanced, delamination is proposed to contribute to this behavior. As compared with this promising result, the different results of other oxide laminates are explained by the residual stress effect, lamina strength and stacking sequence. Insufficient residual stresses cannot initiate the residual shear-enhanced flexural mechanism, but high values can damage materials. Acknowledgements Funding for this work was provided by the U.S. Air Force Office of Scientific Research through Dr A. Pechenik under grant number AFOSR-F49620-93-1- 0027. The authors thank Dr David B. Marshall of Rockwell Science Center for his information regarding a LaPO4-containing ZrO2 laminate. References [1] A.G. Evans, Perspective on the development of high-toughness ceramics, J. Am. Ceram. Soc. 73 (2) (1990) 187–206. [2] P.E.D. Morgan, D.B. Marshall, Functional interfaces for oxide/ oxide composites, Mater. Sci. Eng. A162 (1993) 15–25. [3] P.E.D. Morgan, D.B. Marshall, R.M. Housley, High temperature stability of monazite-alumina composites, Mater. Sci. Eng. A195 (1995) 215–222. [4] P.E.D. Morgan, D.B. Marshall, Ceramic composites of monazite and alumina, J. Am. Ceram. Soc. 78 (6) (1995) 1553–1563. Erratum: J. Am. Ceram. Soc. 78 (9) (1995) 2574. [5] D.H. Kuo, W.M. Kriven, Characterization of yttrium phosphate and a yttrium phosphate/yttrium aluminate laminate, J. Am. Ceram. Soc. 78 (11) (1995) 3121–3124. [6] D.H. Kuo, W.M. Kriven, Microstructure and mechanical response of lanthanum phosphate/yttrium aluminate and yttrium phosphate/yttrium aluminate systems, Ceram. Sci. Eng. Proc. 17B (1996) 233–240. [7] D.H. Kuo, W.M. Kriven, Development of yttrium phosphate as an interphase for oxide/oxide composites, presented at the 2nd Meeting of Pacific Rim Ceramic Societies, Cairns, Australia, July 15–17, 1996. [8] D.H. Kuo, W.M. Kriven, Interface properties of YPO4/ Y3Al5O12 systems: A double-sandwich system and a fiber model composite, Ceram. Trans. 74 (1996) 71–82. [9] W.J. Clegg, K. Kendall, N.M. Alford, D. Birchall, T.W. Button, A simple way to make tough ceramics, Nature 347 (1990) 455–457. [10] W.J. Clegg, The fabrication and failure of laminar ceramic composites, Acta Metall. 40 (11) (1992) 3085–3093. [11] C.A. Folsom, F.W. Zok, F.F. Lange, D.B. Marshall, Mechanical behavior of a laminar ceramic/fiber-reinforced epoxy composite, J. Am. Ceram. Soc. 75 (11) (1992) 2969–2975. [12] S. Baskaran, S.D. Nunn, D. Popovic, J.W. Halloran, Fibrous monolithic ceramics: I, Fabrication, microstructure, and indentation behavior, J. Am. Ceram. Soc. 76 (9) (1993) 2209–2216. [13] S. Baskaran, S.D. Nunn, D. Popovic, J.W. Halloran, Fibrous monolithic ceramics: II, Flexural strength and fracture behavior of the silicon carbide/graphite system, J. Am. Ceram. Soc. 76 (9) (1993) 2217–2224. [14] S. Baskaran, J.W. Halloran, Fibrous monolithic ceramics: III, Mechanical properties and oxidation behavior of the silicon carbide/boron nitride system, J. Am. Ceram. Soc. 77 (5) (1994) 1249–1255. [15] M.C. Shaw, D.B. Marshall, M.S. Dadkhah, A.G. Evans, Cracking and damage mechanisms in ceramic/metal multilayers, Acta Metall. 41 (11) (1993) 3311–3322. [16] Z. Chen, J.J. Mecholsky Jr., Toughening by metallic lamina in nickel/alumina composites, J. Am. Ceram. Soc. 76 (5) (1993) 1258–1264. [17] H. Liu, S.M. Hsu, Fracture behavior of multilayer silicon nitride/boron nitride ceramics, J. Am. Ceram. Soc. 79 (9) (1996) 2452–2457. [18] M. Pechini, Method of Preparing Lead and Alkaline-Earth Titanates and Coating Method Using the Same to Form a Capacitor, U.S. Patent No. 3,330,697, July 11, 1967. [19] D.H. Kuo, W.M. Kriven, Chemical stability, microstructure, and mechanical behavior of LaPO4-containing ceramics, Mater. Sci. Eng A210 (1–2) (1996) 123–134. [20] M. Hirano, H. Inada, Fabrication and properties of yttria- and ceria-doped tetragonal zirconia/alumina composites, Br. Ceram. Trans. J. 90 (1991) 48–51. [21] R.A. Cutler, R.J. Mayhew, K.M. Prettyman, A.V. Virkar, Hightoughness Ce-TZP/Al2O3 ceramics with improved hardness and strength, J. Am. Ceram. Soc. 74 (1) (1991) 179–186. [22] D. Hull, T.W. Clyne, in: An Introduction to Composite Materials, 2nd edn, Cambridge University Press, Cambridge, 1996. [23] D.H. Kuo, Investigation of Oxide fiber/Oxide Matrix Composites with a Weak Interphase, Ph.D. thesis, University of Illinois at Urbana-Champaign, Urbana, IL, 1997. [24] D.J. Green, R.H.J. Hannink, M.V. Swain, Transformation Toughening of Ceramics, CRC Press, Boca Raton, FL, 1989. [25] W.D. Kingery, H.K. Bowen, D.R. Uhlmann, Introduction to Ceramics, 2nd edn, John Wiley, New York, 1976. [26] D.C. Phillips, The fracture energy of carbon-fiber reinforced glass, J. Mater. Sci. 7 (1972) 1175–1191. [27] H.G. Tattersall, G. Tappin, The work of fracture and its measurement in metals, ceramics and other materials, J. Mater. Sci. 1 (1966) 296–301. .