《复合材料 Composites》课程教学资源（学习资料）第五章 陶瓷基复合材料_3Fabrication and properties of laminated Al2O3

CERAMICS INTERNATIONAL ELSEVIER Ceramics International 27(2001)597-602 www.elsevier.com/locate/ceramint Fabrication and properties of laminated Al,O3 TiC composites Yuping Zeng*, Dongliang Jiang The State Key Laboratory of High Performance Ceramics and Superfine Microstructure, Shanghai Institute of Ceramics Chinese Academy of Sciences, 1295 Ding Xi Road, Shanghai 200050, China Received 2 September 2000: received in revised form 29 October 2000: accepted 20 November 2000 Abstract High solid content AlOr-TiC slurries for tape casting were obtained by selecting the appropriate solvent, dispersant, binder, and other organic additives. Some factors which affected the viscosity of the slurries, such as the TiC: Al2O3 ratio, ph value and the dispersant concentration, were discussed. AlO3-TiC laminated composites with weak interfaces can be obtained after laminating d stacking the tapes, binder removal and hot-pressing the green body. Compared with the monolithic composite, the crack pro- pagation in the Al,OTiC laminated composites showed a large scale deflection and multi-fracture which increased the work of racture of the laminates significantly. The results showed that the laminated structure was an effective way to improve the fracture toughness and the work of fracture of the composites. C 2001 Published by Elsevier Science Ltd and Techna S.r.L. All rights reserved Keywords: A. Tape casting: Laminated composite: Slurry; Viscosity 1. Introduction weak interface, and the work of fracture significantly increases. Matsui et al. [9] used tape casting to fabricate Tape casting is a low cost and useful process for pre- SiC-whisker-reinforced 20 wt Si3N4-matrix compo paring thin ceramic sheets. It has been widely used to sites with a high degree of orientation and Kic values as produce ceramic substrates, multilayer capacitors, solid high as 9.5 MPa m/2. The high-temperature bending oxide fuel cells etc [ 1-5]. In addition to the advantages strength of these composites was 1180 MPa at 1250C which all other slurry processes have, tape casting has twice that of Si3N4 alone, and the fracture toughness several other potential benefits, which are not easily was also higher than that of Si3 N4. Chartier and Rouxel achieved by conventional powder processing. Tape cast- [10 used tape casting to fabricate an Al2O3/ ZrO2 lami- ng allows us to meet specific requirements for structural nated composite with residual surface compressive design by using tapes of different compositions and stresses to increase the bending strength from 380 to 560 hicknesses. Tape casting is a very convenient and useful MPa and the fracture toughness from 3.7 to 8.0 MPa method for preparing a laminated composite. It has been m. Huang et al. 11] designed a Sis N4-TiN/ Si3N4- recently used to design laminated structural composites Si3 N4 trilayered composite with residual surface com- with improved mechanical properties, because the lami- pressive stress which resulted in improved mechanical nated composites, such as Si3N4-BN, SiC-C, and mul- properties. During tape casting, some particles with an lite-Al2O3, offers many advantages for improving the anisotropic morphology(e.g. whiskers and platelets) fracture resistance(high fracture toughness and work of would be aligned as they passed the doctor blade and a fracture, etc )and total fracture toughness [6-8]. Lami- preferential alignment in Sic whiskers [12] or platelet nated composites with weak interfaces are most suitable reinforced AlO3 [13] matrix composites have been for this purpose. Catastrophic failure is avoided in these obtained. The objective of this study was to prepare composites because crack propagation is arrested at the laminated structural Al_ TiC composites to improve the mechanical properties, and some factors which Co effected the slurry properties, such as dispersant con- centration, pH value, etc, are discussed 0272-8842/01/S20.00@ 2001 Published by Elsevier Science Ltd and Techna S.r.l. All rights reserved PII:S0272-8842(01)00007-4

Fabrication and properties of laminated Al2O3/TiC composites Yuping Zeng *, Dongliang Jiang The State Key Laboratory of High Performance Ceramics and Superfine Microstructure, Shanghai Institute of Ceramics, Chinese Academy of Sciences, 1295 Ding Xi Road, Shanghai 200050, China Received 2 September 2000; received in revised form 29 October 2000; accepted 20 November 2000 Abstract High solid content Al2O3–TiC slurries for tape casting were obtained by selecting the appropriate solvent, dispersant, binder, and other organic additives. Some factors which affected the viscosity of the slurries, such as the TiC:Al2O3 ratio, pH value and the dispersant concentration, were discussed. Al2O3–TiC laminated composites with weak interfaces can be obtained after laminating and stacking the tapes, binder removal and hot-pressing the green body. Compared with the monolithic composite, the crack pro￾pagation in the Al2O3–TiC laminated composites showed a large scale deflection and multi-fracture, which increased the work of fracture of the laminates significantly. The results showed that the laminated structure was an effective way to improve the fracture toughness and the work of fracture of the composites. # 2001 Published by Elsevier Science Ltd and Techna S.r.l. All rights reserved. Keywords: A. Tape casting; Laminated composite; Slurry; Viscosity 1. Introduction Tape casting is a low cost and useful process for pre￾paring thin ceramic sheets. It has been widely used to produce ceramic substrates, multilayer capacitors, solid oxide fuel cells etc. [1–5]. In addition to the advantages which all other slurry processes have, tape casting has several other potential benefits, which are not easily achieved by conventional powder processing. Tape cast￾ing allows us to meet specific requirements for structural design by using tapes of different compositions and thicknesses. Tape casting is a very convenient and useful method for preparing a laminated composite. It has been recently used to design laminated structural composites with improved mechanical properties, because the lami￾nated composites, such as Si3N4–BN, SiC–C, and mul￾lite–Al2O3, offers many advantages for improving the fracture resistance (high fracture toughness and work of fracture, etc.) and total fracture toughness [6–8]. Lami￾nated composites with weak interfaces are most suitable for this purpose. Catastrophic failure is avoided in these composites because crack propagation is arrested at the weak interface, and the work of fracture significantly increases. Matsui et al. [9] used tape casting to fabricate SiC-whisker-reinforced 20 wt.% Si3N4-matrix compo￾sites with a high degree of orientation and KIC values as high as 9.5 MPa m1/2. The high-temperature bending strength of these composites was 1180 MPa at 1250
C, twice that of Si3N4 alone, and the fracture toughness was also higher than that of Si3N4. Chartier and Rouxel [10] used tape casting to fabricate an Al2O3/ZrO2 lami￾nated composite with residual surface compressive stresses to increase the bending strength from 380 to 560 MPa and the fracture toughness from 3.7 to 8.0 MPa m1/2. Huang et al. [11] designed a Si3N4–TiN/Si3N4– Si3N4 trilayered composite with residual surface com￾pressive stress which resulted in improved mechanical properties. During tape casting, some particles with an anisotropic morphology (e.g. whiskers and platelets) would be aligned as they passed the doctor blade and a preferential alignment in SiC whiskers [12] or platelet reinforced Al2O3 [13] matrix composites have been obtained. The objective of this study was to prepare laminated structural Al2O3–TiC composites to improve the mechanical properties, and some factors which effected the slurry properties, such as dispersant con￾centration, pH value, etc., are discussed. 0272-8842/01/$20.00 # 2001 Published by Elsevier Science Ltd and Techna S.r.l. All rights reserved. PII: S0272-8842(01)00007-4 Ceramics International 27 (2001) 597–602 www.elsevier.com/locate/ceramint * Corresponding author. E-mail address: zyuping@hotmail.com (Y. Zeng)

Y. Zeng, D Jiang/Ceramics International 27(2001)597-602 nitride d carbides can react with water and for oxidized thin films on the surface of the particles during High-purity powder with an average particle long time ball-milling. Therefore, water is rather diffi size of 0.48 um specific area of 5.7 m/g( Shang- cult to use as a solvent to prepare a non-oxide slurry for hai Wushen Chemical plant, Shanghai China )and Tic tape casting An organic solvent, which has a low boil- powder with an average particle size of 0. I um and a ing point and high vapor pressure, is preferred for the pecific surface area of 10 m/g( Chendu Institute of preparation of a slurry for tape casting. On the other Organic Chemistry, Chinese Academy of Science, hand, most organic solvents have a low dielectric con- Sichuan Chen Du, China) were used in the experiment. stant, and are flammable and toxic. The low dielectric A triglyceride was used as the dispersant and poly- constant makes the ion concentration decrease and the vinylbutyral (PVB) as a binder; dioctylphthalate and electrostatic energy is not large enough to make a stable glycerol were used as plasticizers, and a mixture of eth- slurry. In order to overcome some deficiencies of a sin anol and trichloroethylene was used as the solvent gle solvent, a mixture of solvents or binary azeotropic The Al2O3 and TiC powders were mixed with the ti olvent mixtures are often used to obtain good dielectric glyceride dispersant and the ethanol/trichloroethylene properties, a low boiling point as well as a good solu- mixture( volume ratio of 3: 1)in a polyethylene jar and blility for the organic additives. tCE(trichloroethylene)/ then ball-milled for m 20 h using Al,O, balls as the EtOH(ethanol)[14], MeK (methyl ethyl ketone )/EtOH grinding media. The binder and plasticizers were then [20] mixtures are often selected as solvents in tape cast added to the slurry, and the mixture was ball-milled for ing. Normally, the dispersants can be classified into four another 16 h. The ball-milled slurry was degassed under groups based on their charges on the surface active part vacuum, to remove any gas bubbles, cast onto a glass of the molecule, (1) non-ionic (2)anionic (3) cationic substrate, and naturally dried. The tapes were cut into and (4)zwitterionic. Triglycerides, zwitteric molecules, 4050 mm squares. In order to control the resulting are often used in Al2O3 [15] and Zro2 [16] tape casting stresses between the interfaces, a SiC/C slurry was used processing. In this experiment, a mixture of TCE/EtOH to paint the tape surfaces, the solid content of the slurry and triglycerides were chosen as the solvent and dis- was 30 wt % [50 wt. SiC (particle size 2. 1 um)+50 persant respectively wt%C(particle size 3.5 um) without any other organic addition], the solvent was ethanol. After stacking, lami- 3. 2. Properties of the slurry nating and binder removal, the laminated green body was hot-pressed at 1700C and 30 MPa. The samples According to the dlvo theory [17], van der Waal were machined to 2.5x5x26 mm bars for measuring the forces and repulsive forces originate in the double elec- three-point bending strength with a 20 mm span; trical layers surrounding the particles. Because the 5x2.5x26 mm bars were prepared for measuring the counterions have a strong infuence on the diffuse dou- fracture toughness using the single-edge- notch-beam ble electrical layer, moreover, the surface charge of the nethod(20 mm span, notch beam width 0.25 mm) and particles are dependent on the ph value, the concentra 3x426 mm bars for measuring the work of fracture tion of the other specifically adsorbed ions and the ionic with a V-type notch. The mechanical properties were strength of the slurry; therefore, the adjustments in the measured using an Instron-1195 testing machine. The pH value and the ionic strength are helpful to obtain a microstructure and the crack propagation of the com- high surface charge density which results in a stron posites were analyzed using a SEM and an optical double-layer repulsion; the electrostatic repulsion is microscope. The viscosity of the slurries was measured responsible for particles far from the point of zero charge with a NDJ-7 type viscometer of the powder and result in well-dispersed ceramic slur- ries. Although the electrostatic repulsion energy in non- aqueous media is less than that in water due to the low 3. Results and discussion ionic concentration and the low dielectric constant the electrostatic repulsion energy in a nonaqueous slurry is 3.I. Dispersant and solvent important, even if the nonaqueous ceramic slurry is stabilized by a steric mechanism. In this experiment, the The criterion for the solvent selection depends mainly pH value of the mixture solvent, trichlocthylene and upon the solubility of the binders, plasticizers, and other ethanol (volume rate is 1: 3), was about 8.67. After add organic additives. In order to get a well-dispersed and ing Al,O3 and TiC, the ph value of the mixture solvent stable slurry, the relation between the powder, solvent, has changed. Table I shows the effect of the dispersant dispersant and other additives has to be considered. The and the powder on the ph value of the solvent. It is ceramic powders must be chemically stable and should obvious that the dispersant and powders have an effect not react with the organic additives. Although water is a on the ph value of the slurries in the tce/EtOh mix common solvent, some ceramic powders, such as ture. The electrophoretic results showed that the Al2O3

2. Experimental procedure High-purity Al2O3 powder with an average particle size of 0.48 mm and a specific area of 5.7 m2 /g ( Shang￾hai Wushen Chemical plant, Shanghai China ) and TiC powder with an average particle size of 0.1 mm and a specific surface area of 10 m2 /g (Chendu Institute of Organic Chemistry, Chinese Academy of Science, Sichuan Chen Du, China) were used in the experiment. A triglyceride was used as the dispersant and poly￾vinylbutyral (PVB) as a binder; dioctylphthalate and glycerol were used as plasticizers, and a mixture of eth￾anol and trichloroethylene was used as the solvent. The Al2O3 and TiC powders were mixed with the tri￾glyceride dispersant and the ethanol/trichloroethylene mixture ( volume ratio of 3:1) in a polyethylene jar and then ball-milled for 20 h using Al2O3 balls as the grinding media. The binder and plasticizers were then added to the slurry, and the mixture was ball-milled for another 16 h. The ball-milled slurry was degassed under vacuum, to remove any gas bubbles, cast onto a glass substrate, and naturally dried. The tapes were cut into 4050 mm squares. In order to control the resulting stresses between the interfaces, a SiC/C slurry was used to paint the tape surfaces, the solid content of the slurry was 30 wt.%, [50 wt.% SiC (particle size 2.1 mm)+50 wt.% C (particle size 3.5 mm) without any other organic addition], the solvent was ethanol. After stacking, lami￾nating and binder removal, the laminated green body was hot-pressed at 1700
C and 30 MPa. The samples were machined to 2.5526 mm bars for measuring the three-point bending strength with a 20 mm span; 52.526 mm bars were prepared for measuring the fracture toughness using the single-edge-notch-beam method (20 mm span, notch beam width 0.25 mm) and 3426 mm bars for measuring the work of fracture with a V-type notch. The mechanical properties were measured using an Instron-1195 testing machine. The microstructure and the crack propagation of the com￾posites were analyzed using a SEM and an optical microscope. The viscosity of the slurries was measured with a NDJ-7 type viscometer. 3. Results and discussion 3.1. Dispersant and solvent The criterion for the solvent selection depends mainly upon the solubility of the binders, plasticizers, and other organic additives. In order to get a well-dispersed and stable slurry, the relation between the powder, solvent, dispersant and other additives has to be considered. The ceramic powders must be chemically stable and should not react with the organic additives. Although water is a common solvent, some ceramic powders, such as nitrides, and carbides can react with water and form oxidized thin films on the surface of the particles during long time ball-milling. Therefore, water is rather diffi- cult to use as a solvent to prepare a non-oxide slurry for tape casting. An organic solvent, which has a low boil￾ing point and high vapor pressure, is preferred for the preparation of a slurry for tape casting. On the other hand, most organic solvents have a low dielectric con￾stant, and are flammable and toxic. The low dielectric constant makes the ion concentration decrease and the electrostatic energy is not large enough to make a stable slurry. In order to overcome some deficiencies of a sin￾gle solvent, a mixture of solvents or binary azeotropic solvent mixtures are often used to obtain good dielectric properties, a low boiling point as well as a good solu￾blility for the organic additives. TCE(trichloroethylene)/ EtOH(ethanol) [14], MEK(methyl ethyl ketone)/EtOH [20] mixtures are often selected as solvents in tape cast￾ing. Normally, the dispersants can be classified into four groups based on their charges on the surface active part of the molecule, (1) non-ionic (2) anionic (3) cationic and (4) zwitterionic. Triglycerides, zwitteric molecules, are often used in Al2O3 [15] and ZrO2 [16] tape casting processing. In this experiment, a mixture of TCE/EtOH and triglycerides were chosen as the solvent and dis￾persant respectively. 3.2. Properties of the slurry According to the DLVO theory [17], van der Waals forces and repulsive forces originate in the double elec￾trical layers surrounding the particles. Because the counterions have a strong influence on the diffuse dou￾ble electrical layer, moreover, the surface charge of the particles are dependent on the pH value, the concentra￾tion of the other specifically adsorbed ions and the ionic strength of the slurry; therefore, the adjustments in the pH value and the ionic strength are helpful to obtain a high surface charge density which results in a strong double-layer repulsion; the electrostatic repulsion is responsible for particles far from the point of zero charge of the powder and result in well-dispersed ceramic slur￾ries. Although the electrostatic repulsion energy in non￾aqueous media is less than that in water due to the low ionic concentration and the low dielectric constant, the electrostatic repulsion energy in a nonaqueous slurry is important, even if the nonaqueous ceramic slurry is stabilized by a steric mechanism. In this experiment, the pH value of the mixture solvent, trichlocthylene and ethanol (volume rate is 1:3), was about 8.67. After add￾ing Al2O3 and TiC, the pH value of the mixture solvent has changed. Table 1 shows the effect of the dispersant and the powder on the pH value of the solvent. It is obvious that the dispersant and powders have an effect on the pH value of the slurries in the TCE/EtOH mix￾ture. The electrophoretic results showed that the Al2O3 598 Y. Zeng, D. Jiang / Ceramics International 27 (2001) 597–602

Y. Zeng, D Jiang/Ceramics International 27(2001)597-602 particles and the TiC particles exhibited a positive sur- concentration on the viscosity of the slurries, it was face charge and a negative surface charge, respectively. clear that the slurry was stabilized by an"electrosteric However, after the dispersant is added, the Al2O3 par- mechanism. ticle surface charge changed to negative. The organic molecules absorbed on the surface of the powder chan- 3.3. Tape preparation ged the property of the surface charge, therefore, the electrostatic energy which affected the stabilization of the There are many factors which affect the properties of nonaqueous slurries cannot be ignored. The effect of the green tapes, such as solid content, the amount of organic, dispersant concentration on the viscosity of the slurries is the ratio of the binder and plasticizers, temperature as shown in Fig. 1. With increasing concentration of the well as vapor pressure of the solvent etc. Table 2 shows dispersant, the viscosity of the Al2Oj-TiC slurry goes the formulation of the Al2O3-25 wt TiC slurry used in through a minimum For the different ratios of TiC and the experiment. a green tape with good flexibility, and AlO3, the minimum viscosity value for the slurries was without bubbles and cracks could be obtained. Theore- different. The effect of the pH value on the viscosity of tically, the less the organic addition, the better the densi the Al2O3-TiC slurries are shown in Fig. 2. The viscos- fication and the mechanical properties of the composites; ity of the slurries decreased and then increased with if the amount of binder and plasticizers was too low, it increasing pH, but the value of the viscosity had a close was difficult to acquire a perfect green tape or the green relation to the weight ratio of TiC and Al2O3. The pH tape was very brittle. However, a large amount of value for the minimum viscosity of a 30 vol. solid organic cannot produce a well-dispersed slurry. The air content slurry(weight ratio of TiC and Al2O3 45: 55, and is trapped in slurry and forms bubbles which cannot be 35: 65)was about 5, 6. Because the particle size of the Tic removed by ultrasonifiction and vacuum degassing was much smaller than that of AlO3, and the rheological During slurry drying, the bubbles will form defects and properties of TiC and Al2O3 are different for the same cracks, so the preparation of a uniform slurry is very solids content, the viscosity of all slurries increased as the important for controlling the microstructure of the green ratio of TiC and Al2O3 increased and was sensitive to the tape. Slurry drying is another important process to pH value variation. Since the composition of the slurry is very complex, it is difficult to explain precisely what are 2500 the main factors affecting the viscosity. From the results Shear rate 350 S dispersant(.0005g of the influence of the pH value and the dispersant 2000 Table I 1500 Effects of dispersant and powder on the ph value of the solvent pH(without dispersant) 81000 B 7.31-74 4.54-5.02 AlO/TiC 5.43-5.26 00 h value of slurries 500 Fig. 2. Effect of the pH value of slurries upon viscosity of Al,Or-TiC A:TiC:Al2O3(45:55)30vol.%,B:TiC:Al2O3(25:75)25vol.%,C: A TiC:Al2O3(35:65)30vol.%. 自400 Table 2 0 Formulation of Al2Ox-25 wt %TiC slurry Materi mposition Function Al2O3TiC Mg(NO3h 6H,O Grain growth inhibite PVB oncentration of dispersant(x1000/g-m) Fig. 1. Effect of the dispersant concentration on the viscosity of slur Plasticizer ries A: TiC: Al2O3(35: 65)35 vol%, B: TiC: Al2O3(25: 75)25 vol% Glycerol/dioctylphthalate 3.1

particles and the TiC particles exhibited a positive sur￾face charge and a negative surface charge, respectively. However, after the dispersant is added, the Al2O3 par￾ticle surface charge changed to negative. The organic molecules absorbed on the surface of the powder chan￾ged the property of the surface charge, therefore, the electrostatic energy which affected the stabilization of the nonaqueous slurries cannot be ignored. The effect of the dispersant concentration on the viscosity of the slurries is shown in Fig. 1. With increasing concentration of the dispersant, the viscosity of the Al2O3–TiC slurry goes through a minimum. For the different ratios of TiC and Al2O3, the minimum viscosity value for the slurries was different. The effect of the pH value on the viscosity of the Al2O3–TiC slurries are shown in Fig. 2. The viscos￾ity of the slurries decreased and then increased with increasing pH, but the value of the viscosity had a close relation to the weight ratio of TiC and Al2O3. The pH value for the minimum viscosity of a 30 vol.% solid content slurry (weight ratio of TiC and Al2O3 45:55, and 35:65) was about 5, 6. Because the particle size of the TiC was much smaller than that of Al2O3, and the rheological properties of TiC and Al2O3 are different for the same solids content, the viscosity of all slurries increased as the ratio of TiC and Al2O3 increased and was sensitive to the pH value variation. Since the composition of the slurry is very complex, it is difficult to explain precisely what are the main factors affecting the viscosity. From the results of the influence of the pH value and the dispersant concentration on the viscosity of the slurries, it was clear that the slurry was stabilized by an ‘‘electrosteric’’ mechanism. 3.3. Tape preparation There are many factors which affect the properties of green tapes, such as solid content, the amount of organic, the ratio of the binder and plasticizers, temperature as well as vapor pressure of the solvent etc. Table 2 shows the formulation of the Al2O3–25 wt.% TiC slurry used in the experiment. A green tape with good flexibility, and without bubbles and cracks could be obtained. Theore￾tically, the less the organic addition, the better the densi- fication and the mechanical properties of the composites; if the amount of binder and plasticizers was too low, it was difficult to acquire a perfect green tape or the green tape was very brittle. However, a large amount of organic cannot produce a well-dispersed slurry. The air is trapped in slurry and forms bubbles which cannot be removed by ultrasonifiction and vacuum degassing. During slurry drying, the bubbles will form defects and cracks, so the preparation of a uniform slurry is very important for controlling the microstructure of the green tape. Slurry drying is another important process to Fig. 1. Effect of the dispersant concentration on the viscosity of slur￾ries A: TiC: Al2O3 (35:65) 35 vol.%, B: TiC: Al2O3 (25:75) 25 vol.%. Table 1 Effects of dispersant and powder on the pH value of the solvent Powder pH (without dispersant) pH (with dispersant) Al2O3 7.76 7.31–7.44 TiC 5.08 4.54–5.02 Al2O3/TiC 5.78 5.43–5.26 Table 2 Formulation of Al2O3–25 wt.% TiC slurry Materials Composition (wt.%) Function Al2O3/TiC 52.5 Powder Mg(NO3)2 6H2O 1.5 Grain growth inhibitor PVB 2.9 Binder Triglyceride 2.1 Dispersant Trichloroethylene/ethanol 37.9 Solvent Glycerol/dioctylphthalate 3.1 Plasticizer Fig. 2. Effect of the pH value of slurries upon viscosity of Al2O3–TiC A: TiC: Al2O3 (45:55) 30 vol.%, B: TiC: Al2O3 (25:75) 25 vol.%, C: TiC: Al2O3 (35:65) 30 vol.%. Y. Zeng, D. Jiang / Ceramics International 27 (2001) 597–602 599

Y. Zeng, D Jiang/Ceramics International 27(2001)597-602 control the microstructure of the green body. If the sol- including benzaldehyde, phenol, acetophenone etc. In vents evaporate too quickly, drying shrinkage can cause this experiment, there were several kinds of organic the disappearance of the evaporating channels and leads additives in the green tape; therefore it can be deduced to a polymeric skin on the surface of the tape, while the that the decomposition products are even more complex underlying slurry is still in the liquid state; when the sol- Based on the TGA measurements, the binder removal vent vapor evaporates then from the bottom of the tape, process for the laminated composite green bodies at a cracks can form easily. If the solvents evaporate too heating rate of 1C/min was used and the temperature slowly, a density gradient can form in the tape. In this then kept at 800C for 3 h to insure that all the organic experiment, the slurries were dried in a closed chamber at additives decomposed completely room temperature without any flowing air. The proper solvent vapor pressure can prevent tapes from forming 3. 5. Effect of pressure, pH value and the TiC content on racks by fast drying and a density gradient due to slow the green body density d The density of the green body (after binder removal), 3. 4. Binder removal as a function of the pressure is shown in Fig. 4. The green body density increased with the pressure increase, The binder removal process has an effect on the and the experimental results were consistent with the microstructure of the green body. If the organic additives previous observations of Kevin [20]. The green body decompose too fast, the green tape may be warped density, after binder removal, as a function of the TiC surface craters and bubbles form. In order to find the content is shown in Fig. 5. The green density decreased as optimum heating rate, the binder removal process was the TiC content increased, because TiC has a small par- determined by TGA TGA curves of the Al2O3-TiC tape ticle size and a large specific surface, which absorbed are shown in Fig 3. TGa data were obtained in an argon more organic additives and decrease the density of the atmosphere using rates of 1C/min and 5C/min. The green body. The relation between the pH value and the evaporation of solvent from the tapes occurs from room green density is shown in Fig. 6. In general, the green temperature to about 100C; the organic additives density decreased as the pH value of the slurry increased decomposition start about 100C and was completed at about 600 C. The shapes of the two curves with different heating rates are similar. However, the decomposition of the organic additives occurred at a lower temperature A,O3- 25wt%TiC when the heating rate was reduced. This behavior is similar to that mentioned by Scheiffele [18]in the Al2O3/ PVB system. The shapes of the curves were similar to those in a previous report of the author [19]. The binder removal procedure was very complex. Shih et al. [191 51 reported that during the decomposition of PVB various products can be detected at different temperature ran ges. From 300 to 400C, the main products were water and butyraldehyde. When the temperature was above 500oC, a variety of aromatic compounds were detected, Fig. 4. Green density, after binder removal, as a function of pressure 100 Pressure 50MPa t90 5°min 51 250 750 TiC content( Temperature Cc) Fig. 5. Green density, after binder removal. as a function of TiC Fig 3. TGA curves of Al2O-25 wt TiC tape binder removal

control the microstructure of the green body. If the sol￾vents evaporate too quickly, drying shrinkage can cause the disappearance of the evaporating channels and leads to a polymeric skin on the surface of the tape, while the underlying slurry is still in the liquid state; when the sol￾vent vapor evaporates then from the bottom of the tape, cracks can form easily. If the solvents evaporate too slowly, a density gradient can form in the tape. In this experiment, the slurries were dried in a closed chamber at room temperature without any flowing air. The proper solvent vapor pressure can prevent tapes from forming cracks by fast drying and a density gradient due to slow drying. 3.4. Binder removal The binder removal process has an effect on the microstructure of the green body. If the organic additives decompose too fast, the green tape may be warped, surface craters and bubbles form. In order to find the optimum heating rate, the binder removal process was determined by TGA. TGA curves of the Al2O3–TiC tape are shown in Fig. 3. TGA data were obtained in an argon atmosphere using rates of 1
C/min and 5
C/min. The evaporation of solvent from the tapes occurs from room temperature to about 100
C; the organic additives decomposition start about 100
C and was completed at about 600
C. The shapes of the two curves with different heating rates are similar. However, the decomposition of the organic additives occurred at a lower temperature when the heating rate was reduced. This behavior is similar to that mentioned by Scheiffele [18] in the Al2O3/ PVB system. The shapes of the curves were similar to those in a previous report of the author [19]. The binder removal procedure was very complex. Shih et al. [19] reported that during the decomposition of PVB various products can be detected at different temperature ran￾ges. From 300 to 400
C, the main products were water and butyraldehyde. When the temperature was above 500
C, a variety of aromatic compounds were detected, including benzaldehyde, phenol, acetophenone etc. In this experiment, there were several kinds of organic additives in the green tape; therefore it can be deduced that the decomposition products are even more complex. Based on the TGA measurements, the binder removal process for the laminated composite green bodies at a heating rate of 1
C/min was used and the temperature then kept at 800
C for 3 h to insure that all the organic additives decomposed completely. 3.5. Effect of pressure, pH value and the TiC content on the green body density The density of the green body (after binder removal), as a function of the pressure is shown in Fig. 4. The green body density increased with the pressure increase, and the experimental results were consistent with the previous observations of Kevin [20]. The green body density, after binder removal, as a function of the TiC content is shown in Fig. 5. The green density decreased as the TiC content increased, because TiC has a small par￾ticle size and a large specific surface, which absorbed more organic additives and decrease the density of the green body. The relation between the pH value and the green density is shown in Fig. 6. In general, the green density decreased as the pH value of the slurry increased. Fig. 3. TGA curves of Al2O3–25 wt.% TiC tape binder removal. Fig. 4. Green density, after binder removal, as a function of pressure. Fig. 5. Green density, after binder removal, as a function of TiC content. 600 Y. Zeng, D. Jiang / Ceramics International 27 (2001) 597–602

Y. Zeng, D Jiang/Ceramics International 27(2001)597-602 Pressure: 50MPa Fig. 6. Relationship between the pH value and the green density 50um Table 3 Mechanical properties of Al2Or-25 wt %TiC composites Method Bending strength lmm nated structure When the ph value of the slurry was above 7, the green body density had a rapid decrease, which was in accor- dance with the change in viscosity of the slurry the higher the viscosity of slurry, the lower the green body notch d 3.6. Mechanical properties and the work of fracture In order to compare the differences between the Fig. 7. SEM and optical micrographs of the Al2O-25 wt% TiC monolithic composite and the laminated composite, the laminated composite: (a)sEM micrograph, (b)optical micrograph. monolithic Al2O3-25 wt TiC composite was also prepared with AlO3/TiC mixed powder and hot press ing at 30 MPa and 1700 C. The mechanical properties of the Al2O3-25 wt TiC laminated composite with the SiC/C weak interface and the monolithic Al,O3-25 wt. TiC composites are listed in Table 3. The bending strength and fracture toughness of the laminated com posite Al2O3-25 wt TiC with weak interfaces were 605 40 MPa and 5.78 MPa m-12 respectively. The bending B: Tape casting strength and fracture toughness of the monolithic com- posite were 762 MPa and 5.65 MPa m-1/2 respectively From these results the laminated structure was effective for improving fracture toughness, but the degree was not very large as other paper mentioned [9, 10]. Fig. 7 shows an optical micrograph and a SEM micrograph of Deflection(mm) the Al2Ox-25 wt% TiC laminated composite. The Fig 8. Load-deflection curves for Al2 03-25 wt% TiC composite tes- thickness of the SiC/C interface was about 2 um, it was ted as notch specimens in three- point bending difficult to obtain a high density in the Sic/c at sintering temperature, resulting in weak interfaces in the lami- However, many weak interfaces existed in the laminated nated composite; when cracks propagated and met the material which will lead to a bending strength decrease weak interfaces. the main crack was deflected and the Fig 8 is the load-deflection curves for the Al2Or-25 crack branched under the stresses. Because the main crack wt. TiC monolithic composite and AlO 25 wt %o TiC deflection and branching can absorb a large amount of laminated composite tested as notch specimens in three energy, it was useful to improve the fracture toughness. point bending. In laminated structure non-catastrophic

When the pH value of the slurry was above 7, the green body density had a rapid decrease, which was in accor￾dance with the change in viscosity of the slurry ; the higher the viscosity of slurry, the lower the green body density. 3.6. Mechanical properties and the work of fracture In order to compare the differences between the monolithic composite and the laminated composite, the monolithic Al2O3–25 wt.% TiC composite was also prepared with Al2O3/TiC mixed powder and hot press￾ing at 30 MPa and 1700
C. The mechanical properties of the Al2O3–25 wt.% TiC laminated composite with the SiC/C weak interface and the monolithic Al2O3–25 wt.% TiC composites are listed in Table 3. The bending strength and fracture toughness of the laminated com￾posite Al2O3–25 wt.% TiC with weak interfaces were 605 MPa and 5.78 MPa m1/2 respectively. The bending strength and fracture toughness of the monolithic com￾posite were 762 MPa and 5.65 MPa m1/2 respectively. From these results, the laminated structure was effective for improving fracture toughness, but the degree was not very large as other paper mentioned [9,10]. Fig. 7 shows an optical micrograph and a SEM micrograph of the Al2O3–25 wt.% TiC laminated composite. The thickness of the SiC/C interface was about 2 mm, it was difficult to obtain a high density in the SiC/C at sintering temperature, resulting in weak interfaces in the lami￾nated composite; when cracks propagated and met the weak interfaces, the main crack was deflected and the crack branched under the stresses. Because the main crack deflection and branching can absorb a large amount of energy, it was useful to improve the fracture toughness. However, many weak interfaces existed in the laminated material which will lead to a bending strength decrease. Fig. 8 is the load-deflection curves for the Al2O3–25 wt.% TiC monolithic composite and Al2O3–25 wt.% TiC laminated composite tested as notch specimens in three￾point bending. In laminated structure non-catastrophic Table 3 Mechanical properties of Al2O3–25 wt.% TiC composites Method Fracture toughness ( MPa m1/2) Bending strength (MPa) Monolithic 5.65 762 Laminated structure 5.78 605 Fig. 6. Relationship between the pH value and the green density. Fig. 7. SEM and optical micrographs of the Al2O3–25 wt.% TiC laminated composite: (a) SEM micrograph, (b) optical micrograph. Fig. 8. Load-deflection curves for Al2O3–25 wt.% TiC composite tes￾ted as notch specimens in three-point bending. Y. Zeng, D. Jiang / Ceramics International 27 (2001) 597–602 601

Y. Zeng, D Jiang/ Ceramics International 27(2001)597-602 crack growth occurred as long as the crack reached an Buchanan (Ed). Ceramic Materials for Electronics. Marcel- nterface and the crack was deflected along the interface Dekker. New York, 1986, pp. 125-139 The work of fracture of the laminated composite and 5 M.R. Freedom, M. L Millard consolidation of silicon he monolithic composite determined from the area arbide. Ceram. Eng. Sci. Proc. 986884892 [6w.Clegg, K. Kendall, N.M. Al Button. J D. Birchall under the load-deflection curves were 52.4 and 584 J/m2 A simple way to make tough Nature(London) 347 respectively. The results showed that the work of frac (1990)455457 ture of the laminated structure was almost ll times [7 H. Katsuki, Y. Hirata, Coat of alumina sheet with needle-like higher than that of the monolithic composite mullite, J. Ceram Soc. Jpn 98(10)(1990)1114-1119 [8 H. Liu S.M. Hsu, Fracture behavior of multilayer silicon nitrid boron nitride ceramics, J. Am. Ceram. Soc. 79(9)(1996)2452- 4. Conclusion 9 T. Mastui, O. Komora, M. Miyake, The effects of surface and orienting of whiskers on mechanical properties of SiCw/ AlOrTiC laminated composites with a weak inter SiN4J. Ceram.Soe.Jpn99()(1991)1103-110. ace can by tape casuing a nd hot-pressin [0 T. Chartier, T. Rouxel, Tape-cast alumina-zirconia laminates Although dispersed in an organic solvent, Al2O3 and processing and mechanical properties, J. Eur. Ceram Soc TiC powders can effect significantly the pH value of the [11] J.L. Huang, Y L. Chang, H.H. Lu, Fabrication of mult solvent. For the same solid content, the viscosity of the laminated Si3]N4/TiN composites and its anisotropic frac- slurries increased with an increasing TIC: AlO3 ratio ture behavior, J. Mater Res. 12(9)(1997)2337-2344. and was more sensitive to the change of pH value of the [2]E. Krangness, M. Amateau, G.L. Messing, Processing and naracterization of laminated Sic whisker reinforced AlO3,J. slurry. The green density decreased with an increase in Compos. Mater.25(1991)416432. the ph value and the TiC: Al,O3 ratio. The TGa ana [13 M.F. Amateau, G.L. Messing, Laminated ceramic composites, lyses indicated that all the organic additives in tapes can Penn State Centre Adv Mater. Newsletter 4(1)(1990)75-78. be removed above 600%C. due to the weak interfaces in [4] T. Clap, N. Claussen, Processing of ceramic-matrix/platelet the laminated composite cracks propagated along these (1992)263-271 interfaces and allowed stress relaxation: the fracture [15 P.D. Calvert, E.S. Tomey, R.S. Pober, Fish oil and toughness can be improved; the work of fracture of the ersant for alumina. Am. Ceram. Soc. Bull. that of the monolith composite. These results indicated [6] V.L. Richards. Adsorption of dispersant on zirconia powder in that the laminated structure design might be effective tape casting slip composites, J. Am. Ceram. Soc. 72(2)(1989) 325-327. for advanced ceramics to improve fracture resistance [17 EJ. Werwey, J.T. G. Overrbeeck, Theory of the Stability of lyo. phobic Colloids, Elsevier, Amsterdam, 1948 [18 G.W. Sheiffele, M. D. Sack, in: G.L. Messing, E.R. Fuller, et al. References (Eds ) Ceramic Transations, Vol. I(A). Ceramic Powder Science Il, Pyrolysis of poly( vinyl butyal)binder I, Degradation mechan [D. Cooper, P.G. Newland, Fw Effect of processing variables, We igh quality alumina substrates, Ceramic Society, 1988, pp. 559-566. [9 w.K. Shih, M.D. Sack, in: G L. Messing, E R Fuller, et al. [ E Streicher, T. Chartier, P. Boch ce of organic (Eds ) Ceramic Transations. Vol I(A). Ceramic Powder Science, nents on properties of tape cast aluminum nitride substrate. Pyrolysis of poly( vinyl butyal)binder ll, Effect of processing Ceran.Int.16(4)(1990)247-252. variables, Westerville, OH, American Ceramic Society, 1988, pp 3S. Majumdar, T. Claar, B. Flaudermeyer, Stress and fracture 549-558 havior of monolithic fuel cell tapes, J. Am. Ceram. Soc. 69(8 [20]D. Kevin, Pluckneet, H. Cacers. Tape casting of fine alumina/ (1986)628-633 zirconia powders for composites fabrication J. Am. Ceram Soc. 4G.H. Haertling, Piezoelectric and electrooptic ceramics, in: R.C. 77(8)(1994)2137-2144

crack growth occurred as long as the crack reached an interface and the crack was deflected along the interface. The work of fracture of the laminated composite and the monolithic composite determined from the area under the load-deflection curves were 52.4 and 584 J/m2 respectively. The results showed that the work of frac￾ture of the laminated structure was almost 11 times higher than that of the monolithic composite. 4. Conclusion Al2O3–TiC laminated composites with a weak inter￾face can be obtained by tape casting and hot-pressing. Although dispersed in an organic solvent, Al2O3 and TiC powders can effect significantly the pH value of the solvent . For the same solid content, the viscosity of the slurries increased with an increasing TiC:Al2O3 ratio and was more sensitive to the change of pH value of the slurry. The green density decreased with an increase in the pH value and the TiC:Al2O3 ratio. The TGA ana￾lyses indicated that all the organic additives in tapes can be removed above 600oC. Due to the weak interfaces in the laminated composite cracks propagated along these interfaces and allowed stress relaxation; the fracture toughness can be improved; the work of fracture of the laminated composite was about eleven times higher than that of the monolith composite. These results indicated that the laminated structure design might be effective for advanced ceramics to improve fracture resistance. References [1] D. Cooper, P.G. Newland, F.W. Shapley, The development of high quality alumina substrates, in: P. Vincenzini (Ed.), High Tech. Ceramics, Elsevier, Amsterdam, 1987, pp. 1549–1554. [2] E. Streicher, T. Chartier, P. Boch, Influence of organic compo￾nents on properties of tape cast aluminum nitride substrate, Ceram. Int. 16 (4) (1990) 247–252. [3] S. Majumdar, T. Claar, B. Flaudermeyer, Stress and fracture behavior of monolithic fuel cell tapes, J. Am. Ceram. Soc. 69 (8) (1986) 628–633. [4] G.H. Haertling, Piezoelectric and electrooptic ceramics, in: R.C. Buchanan (Ed.), Ceramic Materials for Electronics, Marcel￾Dekker, New York, 1986, pp. 125–139. [5] M.R. Freedom, M.L. Millard, Improved consolidation of silicon carbide, Ceram. Eng. Sci. Proc. 7 (7–8) (1986) 884–892. [6] W. Jclegg, K. Kendall, N.M. Alford, T.W. Button, J.D. Birchall, A simple way to make tough ceramics, Nature (London) 347 (1990) 455–457. [7] H. Katsuki, Y. Hirata, Coat of alumina sheet with needle-like mullite, J. Ceram. Soc. Jpn 98 (10) (1990) 1114–1119. [8] H. Liu, S.M. Hsu, Fracture behavior of multilayer silicon nitride/ boron nitride ceramics, J. Am. Ceram. Soc. 79 (9) (1996) 2452– 2457. [9] T. Mastui, O. Komora, M. Miyake, The effects of surface coating and orienting of whiskers on mechanical properties of SiCw/ Si3N4, J. Ceram. Soc. Jpn 99 (11) (1991) 1103–1106. [10] T. Chartier, T. Rouxel, Tape-cast alumina–zirconia laminates: processing and mechanical properties, J. Eur. Ceram. Soc. 11 (1997) 299–308. [11] J.L. Huang, Y.L. Chang, H.H. Lu, Fabrication of multi￾laminated Si3N4–Si3N4/TiN composites and its anisotropic frac￾ture behavior, J. Mater. Res. 12 (9) (1997) 2337–2344. [12] E. Krangness, M. Amateau, G.L. Messing, Processing and characterization of laminated SiC whisker reinforced Al2O3, J. Compos. Mater. 25 (1991) 416–432. [13] M.F. Amateau, G.L. Messing, Laminated ceramic composites, Penn. State Centre Adv. Mater. Newsletter 4 (1) (1990) 75–78. [14] T. Claab, N. Claussen, Processing of ceramic-matrix/platelet composites by tape casting and laminated, J. Eur. Ceram. Soc. 10 (1992) 263–271. [15] P.D. Calvert, E.S. Tomey, R.S. Pober, Fish oil and triglycerides as dispersant for alumina, Am. Ceram. Soc. Bull. 65 (4) (1985) 669–672. [16] V.L. Richards, Adsorption of dispersant on zirconia powder in tape casting slip composites, J. Am. Ceram. Soc. 72 (2) (1989) 325–327. [17] E.J. Werwey, J.T.G. Overrbeeck, Theory of the Stability of Lyo￾phobic Colloids, Elsevier, Amsterdam, 1948. [18] G.W. Sheiffele, M.D. Sack, in: G.L. Messing, E.R. Fuller, et al. (Eds.), Ceramic Transations, Vol. 1(A). Ceramic Powder Science II, Pyrolysis of poly(vinyl butyal)binder :I, Degradation mechan￾ism, Effect of processing variables, Westerville, OH, American Ceramic Society, 1988, pp. 559–566. [19] W.K. Shih, M.D. Sack, in: G.L. Messing, E.RFuller, et al. (Eds.), Ceramic Transations, Vol. 1(A). Ceramic Powder Science, II, Pyrolysis of poly(vinyl butyal)binder :II, Effect of processing variables, Westerville, OH, American Ceramic Society, 1988, pp. 549–558. [20] D. Kevin, Pluckneet, H. Cacers. Tape casting of fine alumina/ zirconia powders for composites fabrication J. Am. Ceram. Soc. 77(8) (1994) 2137–2144. 602 Y. Zeng, D. Jiang / Ceramics International 27 (2001) 597–602