D.-K. Kim, W.M. Kriven/ Composites: Part B 37(2006)509-514 511 paper. The work of fracture of each sample was obtained from the calculation of the area under the load-displacement curve 3. Results and discussion Table I summarizes the tape cast mixing formulations for the different oxides. The amount of powder was 25.1 vol% except for the alumina platelets, in which case 30 vol% of powder was used. For the alumina matrix, a lower amount of binder of 3.7 vol% and higher amounts of plasticizers, i.e 5.6 vol% of polyethylene glycol and 6.7 vol% of dibutyl pthalate, were used, because of delamination after binder removal For the 3Y-TZP matrix. 30 vol% excess solvent was Fig2 SEM micrograph of the 5-10 um alumina platelets having thickness of added, before and after the first ball milling, respectively, to lower the viscosity. The viscosity of the AlPO4 formulation was lowered by adding 30 vol% excess solvent before the first Witco Chemicals, Houston, TX)was used as a dispersant. The ball milling. To prevent possible change of their shape by binder was polyvinyl butyral( Butvar B90, Solutia Chemicals, breaking during mixing, alumina platelets were mixed with Louis,MS). Dibutyl phthalate(99% purity, Aldrich polymers by stirring without balls. The alumina platelets Chemical Inc, Milwaukee, wI) and polyethylene glycol 300 vol excess solvent, and dispersant were mixed by stirring 300NF, FCC grade, Union Carbide, Danbury, CT) were for 12 h. Another 12 h mixing was carried out afte er adding the used as plasticizers. A conventional tape casting machine with plasticizers and binder into solution. The excess solvent was double doctor blades was used. The first doctor blade openings evaporated before tape casting for the strong matrix materials and crack deflecting materials The morphology of the alumina platelets is seen in the SEM were 600 and 75 um, respectively. The second doctor blade micrograph of Fig. 2. TI hey had a hexagonal platelet shape, an openings were 1200 and 150 um, respectively. The speed of approximate thickness of I um, and size of 5-10 um. The XR casting was I cm/s. The procedures for making laminated results indicated compatibility between the oxide matrix opposites are shown in the flow chart of Fig. 1. De-airing was materials and AlPO4, and are schematically summarized in ig. 3. The mixtures of Al,O3, mullite, 50 vol% alumina. 50 speed. The laminated composite was thermo-compressed into a vol% YAG in situ composite, 3Y-TZP and AlPOA were 80C. The binder removal was achieved by increasing the 1650C/10 h, and 1550C/I h, respectively. The aluminum temperature from room temperature to 150C at a ramp rate of phosphate was compatible with alumina, mullite, and zirconia 1C/min, then from 150 to 600C at a ramp rate of 0.1C/min, However, AlPO4 was not compatible with the 50 vol% and finally by maintaining the sample at 600C for 2 h. Cold alumina. 50 vol% YAG in situ composite matrix. AlPO4 isostatic pressing ( CIP) was carried out at 413.7 MPa. The reacted with YAG in the composite, and formed yttrium sintering conditions differed depending on the particular phosphate (YPO4) materials The bulk density of sintered pellets was measured by Archimedes method(ASTM C373). To study the chemical bility between oxide matrix materials Rigaku X-ray diffractometer (Model D-Max automated diffractometer, Rigaku/USA, Danvers, MA)was used. Two powders were mixed by 24 h ball milling, sintered, and 1600c/0h analyzed for any co-existing phases by XRD. The microstruc- L"""" tures of the laminated composites were studied by scannin 1650c/0h electron microscopy(SEM, Model S-530, Hitachi, Osaka, Japan). A screw-driven universal testing machine (Model 人xs 4502, Instron Corp, Canton, MA)was used to measure flexural 1550c/h strengths in 3-point bend testing. The cross-head speed was 0. 1 mm/min, the supporting span was 30 mm, and the specimen sIze was3mm(H)×4mm(W)×40mm①L).The 2 Theta flexural strength and work of fracture data were determined by Fig 3. X-ray diffraction profiles indicating the compatibility between the four testing 3-5 samples. The final surface polishing of specimens oxide matrix materials and AIPO4(temperature/time represents the sintering for bend testing were conducted by 600 grit SiC polishing condition)Witco Chemicals, Houston, TX) was used as a dispersant. The binder was polyvinyl butyral (Butvar B90, Solutia Chemicals, St Louis, MS). Dibutyl phthalate (99% purity, Aldrich Chemical Inc., Milwaukee, WI) and polyethylene glycol (300NF, FCC grade, Union Carbide, Danbury, CT) were used as plasticizers. A conventional tape casting machine with double doctor blades was used. The first doctor blade openings for the strong matrix materials and crack deflecting materials were 600 and 75 mm, respectively. The second doctor blade openings were 1200 and 150 mm, respectively. The speed of casting was 1 cm/s. The procedures for making laminated composites are shown in the flow chart of Fig. 1. De-airing was carried out by rotating a ball-free suspension at a very slow speed. The laminated composite was thermo-compressed into a rectangular pellet at 34.5 MPa after being maintained for 1 h at 80 8C. The binder removal was achieved by increasing the temperature from room temperature to 150 8C at a ramp rate of 1 8C/min, then from 150 to 600 8C at a ramp rate of 0.1 8C/min, and finally by maintaining the sample at 600 8C for 2 h. Cold isostatic pressing (CIP) was carried out at 413.7 MPa. The sintering conditions differed depending on the particular materials. The bulk density of sintered pellets was measured by Archimedes’ method (ASTM C373). To study the chemical compatibility between oxide matrix materials and AlPO4, a Rigaku X-ray diffractometer (Model D-Max automated diffractometer, Rigaku/USA, Danvers, MA) was used. Two powders were mixed by 24 h ball milling, sintered, and analyzed for any co-existing phases by XRD. The microstruc￾tures of the laminated composites were studied by scanning electron microscopy (SEM, Model S-530, Hitachi, Osaka, Japan). A screw-driven universal testing machine (Model 4502, Instron Corp., Canton, MA) was used to measure flexural strengths in 3-point bend testing. The cross-head speed was 0.1 mm/min, the supporting span was 30 mm, and the specimen size was 3 mm (H)!4 mm (W)!40 mm (L). The flexural strength and work of fracture data were determined by testing 3–5 samples. The final surface polishing of specimens for bend testing were conducted by 600 grit SiC polishing paper. The work of fracture of each sample was obtained from the calculation of the area under the load–displacement curve from bend testing. 3. Results and discussion Table 1 summarizes the tape cast mixing formulations for the different oxides. The amount of powder was 25.1 vol%, except for the alumina platelets, in which case 30 vol% of powder was used. For the alumina matrix, a lower amount of binder of 3.7 vol% and higher amounts of plasticizers, i.e. 5.6 vol% of polyethylene glycol and 6.7 vol% of dibutyl pthalate, were used, because of delamination after binder removal. For the 3Y-TZP matrix, 30 vol% excess solvent was added, before and after the first ball milling, respectively, to lower the viscosity. The viscosity of the AlPO4 formulation was lowered by adding 30 vol% excess solvent before the first ball milling. To prevent possible change of their shape by breaking during mixing, alumina platelets were mixed with polymers by stirring without balls. The alumina platelets, 300 vol% excess solvent, and dispersant were mixed by stirring for 12 h. Another 12 h mixing was carried out after adding the plasticizers and binder into solution. The excess solvent was evaporated before tape casting. The morphology of the alumina platelets is seen in the SEM micrograph of Fig. 2. They had a hexagonal platelet shape, an approximate thickness of 1 mm, and size of 5–10 mm. The XRD results indicated compatibility between the oxide matrix materials and AlPO4, and are schematically summarized in Fig. 3. The mixtures of Al2O3, mullite, 50 vol% alumina$50 - vol% YAG in situ composite, 3Y-TZP and AlPO4 were sintered under the conditions of 1600 8C/3 h, 1600 8C/10 h, 1650 8C/10 h, and 1550 8C/1 h, respectively. The aluminum phosphate was compatible with alumina, mullite, and zirconia. However, AlPO4 was not compatible with the 50 vol% alumina$50 vol% YAG in situ composite matrix. AlPO4 reacted with YAG in the composite, and formed yttrium phosphate (YPO4). Fig. 2. SEM micrograph of the 5–10 mm alumina platelets having thickness of w1 mm. Fig. 3. X-ray diffraction profiles indicating the compatibility between the four oxide matrix materials and AlPO4 (temperature/time represents the sintering condition). D.-K. Kim, W.M. Kriven / Composites: Part B 37 (2006) 509–514 511