Availableonlineatwww.sciencedirect.com DIRECTo LETTERS ELSEVIER Materials Letters 58(2004)1410-1414 reparation and mechanical properties of 10 vol. zirconia/alumina composite with fine-scale fibrous microstructure by co-extrusion process Hiroyuki Miyazaki", Yu-ichi Yoshizawa, Kiyoshi Hirao nergy Materials Research Center, National Institute of Advanced Industrial Science and Technology 2268-1 Shimo-shidami, Nagoya 463-8687, Japan red 9 July 2003; accepted 29 September 2003 Abstract Ten volume-percent zirconia/alumina bi-phase composite with fine-scale aligned microstructure could be fabricated by repeated co- extrusion through a reduction die. Crack-free composites with a fiber diameter of 20 um were obtained after 3rd extrusion. The fracture toughness was increased by introducing the fine zirconia filaments into alumina matrix. C 2003 Elsevier B V. All rights reserved. Keywords: Ceramics; Composite materials; Mechanical properties; Aligned microstructure; Extrusion; Alumina; Zirconia 1. Introduction reduce the diameter of the structures and the effect of the reduced diameter on the mechanical properties Monolithic ceramics are intrinsically brittle, which is one Here, the authors focused on co-extrusion process to of the reasons to limit their engineering applications. Much realize fine-scale fibrous microstructure. Although this effort has been directed to improve their toughness by process has been used to obtain fine-scale objects [21, various concepts. Reinforcement by continuous fibers is macrochannelled structure [3 and continuous fibrillate one solution for the given problem. However, ceramic fibers microstructure [4], there is no report on the mechanical are very expensive, and it is also difficult to fully densify the properties of the composite. The aim of this study is to composites because of the hindrance to sintering by a woven demonstrate the feasibility of forming fine-scale fibrous fiber preform. Another candidate is fibrous microstructure microstructures by using co-extrusion process and to fabricated from coated green fibers. Baskaran et al. [1 investigate the effect of microstructure on the mechanical reported a novel and powerful method to fabricate fibrous properties ceramics. The fibrous ceramics consisted of high aspect ratio polycrystalline regions(cells) of a primary phase separated by thin second-phase regions(cell boundaries). 2. Experimental procedure The cells are the remnants of the green fiber which consists of ceramic powder and a polymer binder. The coating The starting powders in this work were low-soda alumina pplied on the green fiber formed the cell boundaries. They(AL-160SG-4, Showa Denkou, Japan) and yttria-stabilized showed that the fibrous ceramics showed a potentiality to zirconia (TZ-3Y, Tosoh, Japan). Since alumina-zirconia improve the mechanical properties. However, the diameter system does not have any solid solution or compound, it of fibrous microstructure was around 200 um so that their is suitable to examine the formation of fine-fibrous micro- microstructure was too coarse to maximize the mechanical structure. The average particle size of alumina was 0.6 um properties. There is still room to investigate a possibility to and the specific surface area of the zirconia was 16 m2/g The powders were milled individually in ethanol, and then the slurry was dried in rotary evaporator. Alumina powder 4 Correspondin ng author. Tel:+81-52-736-7486: fax:+81-52-739-0136. was tempered with 18 mass% of distilled water and 20 E-mail address: h-miyazaki(@aist.go jP(H. Miyazaki) mass%of organic binders and plasticizer (VL-El, VL-E6, 0167-577X/S- see front matter c 2003 Elsevier B.V. All rights reserved. doi:10.1016 j. mallet2003.09.037
Preparation and mechanical properties of 10 vol.% zirconia/alumina composite with fine-scale fibrous microstructure by co-extrusion process Hiroyuki Miyazaki*, Yu-ichi Yoshizawa, Kiyoshi Hirao Synergy Materials Research Center, National Institute of Advanced Industrial Science and Technology, 2268-1 Shimo-shidami, Nagoya 463-8687, Japan Received 9 July 2003; accepted 29 September 2003 Abstract Ten volume-percent zirconia/alumina bi-phase composite with fine-scale aligned microstructure could be fabricated by repeated coextrusion through a reduction die. Crack-free composites with a fiber diameter of 20 Am were obtained after 3rd extrusion. The fracture toughness was increased by introducing the fine zirconia filaments into alumina matrix. D 2003 Elsevier B.V. All rights reserved. Keywords: Ceramics; Composite materials; Mechanical properties; Aligned microstructure; Extrusion; Alumina; Zirconia 1. Introduction Monolithic ceramics are intrinsically brittle, which is one of the reasons to limit their engineering applications. Much effort has been directed to improve their toughness by various concepts. Reinforcement by continuous fibers is one solution for the given problem. However, ceramic fibers are very expensive, and it is also difficult to fully densify the composites because of the hindrance to sintering by a woven fiber preform. Another candidate is fibrous microstructure fabricated from coated green fibers. Baskaran et al. [1] reported a novel and powerful method to fabricate fibrous ceramics. The fibrous ceramics consisted of high aspect ratio polycrystalline regions (cells) of a primary phase separated by thin second-phase regions (cell boundaries). The cells are the remnants of the green fiber which consists of ceramic powder and a polymer binder. The coating applied on the green fiber formed the cell boundaries. They showed that the fibrous ceramics showed a potentiality to improve the mechanical properties. However, the diameter of fibrous microstructure was around 200 Am so that their microstructure was too coarse to maximize the mechanical properties. There is still room to investigate a possibility to reduce the diameter of the structures and the effect of the reduced diameter on the mechanical properties. Here, the authors focused on co-extrusion process to realize fine-scale fibrous microstructure. Although this process has been used to obtain fine-scale objects [2], macrochannelled structure [3] and continuous fibrillate microstructure [4], there is no report on the mechanical properties of the composite. The aim of this study is to demonstrate the feasibility of forming fine-scale fibrous microstructures by using co-extrusion process and to investigate the effect of microstructure on the mechanical properties. 2. Experimental procedure The starting powders in this work were low-soda alumina (AL-160SG-4, Showa Denkou, Japan) and yttria-stabilized zirconia (TZ-3Y, Tosoh, Japan). Since alumina –zirconia system does not have any solid solution or compound, it is suitable to examine the formation of fine-fibrous microstructure. The average particle size of alumina was 0.6 Am and the specific surface area of the zirconia was 16 m2 /g. The powders were milled individually in ethanol, and then the slurry was dried in rotary evaporator. Alumina powder was tempered with 18 mass% of distilled water and 20 mass% of organic binders and plasticizer (VL-E1, VL-E6, 0167-577X/$ - see front matter D 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.matlet.2003.09.037 * Corresponding author. Tel.: +81-52-736-7486; fax: +81-52-739-0136. E-mail address: h-miyazaki@aist.go.jp (H. Miyazaki). www.elsevier.com/locate/matlet Materials Letters 58 (2004) 1410 – 1414
H. Miyazaki et al./Materials Letters 58(2004)1410-1414 VL-N2, VL-N3 and VL-A(trial production), Yuken Indus- expected to have 5476(148 X 37)zirconia fine-scaled tries, Japan) using a hook type of mixing machine and a filaments(a filament size of around 20 um) roller mill. Zirconia powder was also tempered with 1 The extruded green bodies were dried at room tempera- nass% of distilled water and 16 mass% of organic binders, ture under atmospheric pressure. The dried bodies were then plasticizer and dispersant (VL-El, VL-E6, VL-N2, VL-N3, calcined at 600C for 4 h under a flow of nitrogen, VL-A(trial production) and Cl, Yuken Industries)using the followed by calcining at 500C for 3 h in air to remove same apparatus. The volume fraction of the binders and remaining organic substances. After being cold isostatically plasticizer in each powder was almost the same. In contrast, pressed(CIPed)under 300 MPa, they were pressureless the amount of distilled water was adjusted between the sintered at 1600C for 3 h. The heating and cooling rate powders so as to maintain the viscosity of plastic body to be was 5C/min. For the sake of comparison, monolithic same. Each of the obtained plastic body was extruded into a alumina, without zirconia filaments, was also fabricated he spacing between opposite side: 5 mm) by the same procedure y means of a screw auger machine(FM-P30-1, Miyazaki- For bending strength measurements, test bars with Tekko, Japan). Then a total of 37 monofilaments(4 zirconia dimensions of 3 x 4 X 35 mm were machined from the and 33 alumina)were bundled into a hexagonal feedrod sintered body and polished on the tensile face. Four point (Fig. 1). The initial feedrod was extruded through 6: 1 bending strength measurements were conducted with an hexagonal reduction die using a piston extruder. After the inner span of 10 mm, outer span of 30 mm, and a crosshead first co-extrusion, the individual pieces were bundled to speed of 0.5 mm/min Fracture toughness(Kic) was deter- create second feedrod. The second feedrod was expected to mined by the single-edged-precracked-beam(SEPB) meth- have a total of 148(4 X 37) zirconia filaments. For the od with a span of 16 mm. Fracture strength and fracture second and third co-extrusion, the same process was repea toughness measurements were carried out so that tensile ed. After the third co-extrusion, the extruded filament was stress during measurements was parallel to the extruding direction. The reported mechanical properties are the aver age of a minimum of three separate tests In order to clarify the phase transformation of zirconia Zirconia due to fracture. the volume fraction of monoclinic zirconia (m) on the fracture surface and the polished surface was First Feedrod measured by means of XRD analysis(Rigaku, Japan), and was calculated according to the following the equation [5] (111+lm(1 lm(111+(111)+h(111) where I is the integral intensity and the subscripts m and t refer to the monoclinic and the tetragonal phases, 3. Results and discussion Fig. 2(a)shows a cross-sectional view of lst stage co- extruded composite after sintering at 1600C for 3 h. The light phase corresponds to zirconia phase. The diameter of zirconia filament was reduced from 5 mm to <600 um Second Feedrod through the reduction die. It should be noted that the linear shrinkage in radial direction of the composite was about 33% after CIP and sintering at 1600C for 3 h Observed diameter of zirconia filament was in good agreement with the value that was calculated from the diameter of initial monofilament. the reduction rate of the die and the linear shrinkage after CIP and sintering. Significant crack forma tion was observed within both zirconia phase(light phase) Second extrudate and alumina phase(dark phase). Fig. 2(b)shows a side view of the above sample. Some cracks are seen in the zirconia filament and also in the alumina matrix. There are two Fig. 1 Schematic illustration for microfabrication by co-extrusion process. possibilities of crack formation; the difference in sintering
VL-N2, VL-N3 and VL-A (trial production), Yuken Industries, Japan) using a hook type of mixing machine and a roller mill. Zirconia powder was also tempered with 16 mass% of distilled water and 16 mass% of organic binders, plasticizer and dispersant (VL-E1, VL-E6, VL-N2, VL-N3, VL-A (trial production) and C1, Yuken Industries) using the same apparatus. The volume fraction of the binders and plasticizer in each powder was almost the same. In contrast, the amount of distilled water was adjusted between the powders so as to maintain the viscosity of plastic body to be same. Each of the obtained plastic body was extruded into a hexagonal shape (the spacing between opposite side: 5 mm) by means of a screw auger machine (FM-P30-1, MiyazakiTekko, Japan). Then a total of 37 monofilaments (4 zirconia and 33 alumina) were bundled into a hexagonal feedrod (Fig. 1). The initial feedrod was extruded through 6:1 hexagonal reduction die using a piston extruder. After the first co-extrusion, the individual pieces were bundled to create second feedrod. The second feedrod was expected to have a total of 148 (4 37) zirconia filaments. For the second and third co-extrusion, the same process was repeated. After the third co-extrusion, the extruded filament was expected to have 5476 (148 37) zirconia fine-scaled filaments (a filament size of around 20 Am). The extruded green bodies were dried at room temperature under atmospheric pressure. The dried bodies were then calcined at 600 jC for 4 h under a flow of nitrogen, followed by calcining at 500 jC for 3 h in air to remove remaining organic substances. After being cold isostatically pressed (CIPed) under 300 MPa, they were pressureless sintered at 1600 jC for 3 h. The heating and cooling rate was 5 jC/min. For the sake of comparison, monolithic alumina, without zirconia filaments, was also fabricated by the same procedure. For bending strength measurements, test bars with dimensions of 3 4 35 mm were machined from the sintered body and polished on the tensile face. Four point bending strength measurements were conducted with an inner span of 10 mm, outer span of 30 mm, and a crosshead speed of 0.5 mm/min. Fracture toughness (KIC) was determined by the single-edged-precracked-beam (SEPB) method with a span of 16 mm. Fracture strength and fracture toughness measurements were carried out so that tensile stress during measurements was parallel to the extruding direction. The reported mechanical properties are the average of a minimum of three separate tests. In order to clarify the phase transformation of zirconia due to fracture, the volume fraction of monoclinic zirconia (Vm) on the fracture surface and the polished surface was measured by means of XRD analysis (Rigaku, Japan), and was calculated according to the following the equation [5]: Vm ¼ Imð111 ¯ Þ þ Imð111Þ Imð111 ¯ Þ þ Imð111Þ þ Itð111Þ ð1Þ where I is the integral intensity and the subscripts m and t refer to the monoclinic and the tetragonal phases, respectively. 3. Results and discussion Fig. 2(a) shows a cross-sectional view of 1st stage coextruded composite after sintering at 1600 jC for 3 h. The light phase corresponds to zirconia phase. The diameter of zirconia filament was reduced from 5 mm to < 600 Am through the reduction die. It should be noted that the linear shrinkage in radial direction of the composite was about 33% after CIP and sintering at 1600 jC for 3 h. Observed diameter of zirconia filament was in good agreement with the value that was calculated from the diameter of initial monofilament, the reduction rate of the die and the linear shrinkage after CIP and sintering. Significant crack formation was observed within both zirconia phase (light phase) and alumina phase (dark phase). Fig. 2(b) shows a side view of the above sample. Some cracks are seen in the zirconia filament and also in the alumina matrix. There are two Fig. 1. Schematic illustration for microfabrication by co-extrusion process. possibilities of crack formation; the difference in sintering H. Miyazaki et al. / Materials Letters 58 (2004) 1410–1414 1411
1412 H. Miyazaki et al. / Materials Letters 58(2004)1410-1414 from the difference in the flow rate between at center of the extrudate and near the die wall during the initial stage of extrusion where the paste was not filled in the top of the reduction die. The total number of zirconia filaments in the which was almost the same as the calculated value of 148 (4 X 37). The loss in the number of zirconia filaments was attributed to the severe distortion of their shape at rim of the composite. Cracks were not observed in the cross-sectional view(Fig 3(a)). However, in the side view(Fig 3(b)some cracks perpendicular to the extrusion direction was observed within the thick zirconia filament though there was no crack in alumina phase. The absence of cracks that run parallel to 500um the extrusion direction may be attributed to reduced diam- eter of the zirconia filament since the thinner filament could not store sufficient radial strain energy for the crack forma- tion In the case of the thick filament, because the length of (b) filament along with extrusion direction unchanged, the 500um 500um Fig. 2. Optical micrographs of the first stage co-extruded 10 vol. zirconia/ alumina extrudate. (a)Cross-sectional and (b)side view of composite sintered at 1600.C for 3 h (b) shrinkage and thermal expansion mismatch between two phases. According to the Menon and Chens [6] report on laminate biomaterial composites, cracks formed during sintering have a relatively wide crack-opening displacement and rounded edges. In this study, the crack-opening dis- placement was not as large as that of sintering crack reported in Menon's study and the rounded edges of cracks were not found. Then, it seems reasonable to suppose that the cracks observed in this study were cooling cracks Fig. 3(a)and(b) show the cross-sectional and side view of 2nd stage co-extruded composite after sintering at 1600 c for 3 h. the diameter of the zirconia filament was 500um reduced to 50-150 um and the number of filaments in- eased significantly. Significant shape distortion of zirconia phase was observed in Fig 3(a). The shape distortion of filaments in the co-extruded bi-phase composite was also 3. Optical micrographs of the second stage co-extruded a/alumina extrudate. (a) Cross-sectional and (b) side reported by Kaya et al. [4. The shape distortion may arise composite sintered at 1600C for 3h
shrinkage and thermal expansion mismatch between two phases. According to the Menon and Chen’s [6] report on laminate biomaterial composites, cracks formed during sintering have a relatively wide crack-opening displacement and rounded edges. In this study, the crack-opening displacement was not as large as that of sintering crack reported in Menon’s study and the rounded edges of cracks were not found. Then, it seems reasonable to suppose that the cracks observed in this study were cooling cracks. Fig. 3(a) and (b) show the cross-sectional and side view of 2nd stage co-extruded composite after sintering at 1600 jC for 3 h. The diameter of the zirconia filament was reduced to 50 –150 Am and the number of filaments increased significantly. Significant shape distortion of zirconia phase was observed in Fig. 3(a). The shape distortion of filaments in the co-extruded bi-phase composite was also reported by Kaya et al. [4]. The shape distortion may arise from the difference in the flow rate between at center of the extrudate and near the die wall during the initial stage of extrusion where the paste was not filled in the top of the reduction die. The total number of zirconia filaments in the composite observed with a optical microscope was 127, which was almost the same as the calculated value of 148 (4 37). The loss in the number of zirconia filaments was attributed to the severe distortion of their shape at rim of the composite. Cracks were not observed in the cross-sectional view (Fig. 3(a)). However, in the side view (Fig. 3(b)) some cracks perpendicular to the extrusion direction was observed within the thick zirconia filament though there was no crack in alumina phase. The absence of cracks that run parallel to the extrusion direction may be attributed to reduced diameter of the zirconia filament since the thinner filament could not store sufficient radial strain energy for the crack formation. In the case of the thick filament, because the length of filament along with extrusion direction unchanged, the Fig. 2. Optical micrographs of the first stage co-extruded 10 vol.% zirconia/ alumina extrudate. (a) Cross-sectional and (b) side view of composite sintered at 1600 jC for 3 h. Fig. 3. Optical micrographs of the second stage co-extruded 10 vol.% zirconia/alumina extrudate. (a) Cross-sectional and (b) side view of composite sintered at 1600 jC for 3 h. 1412 H. Miyazaki et al. / Materials Letters 58 (2004) 1410–1414
H. Miyazaki et al. /Materials Letters 58(2004)1410-1414 1413 volume of filament was large enough to accumulate the Table I longitudinal strain energy which promotes crack formation Mechanical pr fter 3rd extrusion step normal to extrusion direction 4Pt bending Toughness(SEPB). Fig. 4(a)and (b) shows the cross-sectional and side view of 3rd stage co-extruded composite after sintering at 1600 5.3±0 c for 3 h. the diameter of zirconia filament further reduced to about 20 um, which was near the theoretical 462±134 4.3±00 value expected from both the diameter of zirconia filament in 2nd stage composite and the reduction rate. It is obvious cross-sectional view was attributed to this discontinuity of that the co-extrusion process effectively reduced the diam- the filaments since we could not count the filaments which eter of zirconia filaments. The number of zirconia filaments broke near the observed cross-section. No cracks were per unit area was measured from Fig. 4(a). The estimated observed in both zirconia and alumina phases. The absence total number of the filaments within the 3rd extruded body of cracks in zirconia phase may be attributed to both the was about 2500, which was less than half of the theoretical reduced diameter and the limited length of filaments since number of 5476(148 x 37). Most of zirconia filaments have such thin and short filaments cannot retain large strain broken down and were not continuous as seen in Fig. 4(b). energy for crack formation. Thus, we could get crack-free However, the zirconia phase still maintained unidirectional composite by refining the microstructure with repeated co alignment. The reduced number of filaments estimated in extrusion process. The disappearance of cracks by refining the microstructure sometimes can be seen in this kind of experiments. For example, Menon and Chen [6] also reported that crack-free alumina/zirconia laminate compo- sites were obtained when the thickness of the laminate was less than50μm Mechanical properties were evaluated for the composites after the 3rd extrusion step. Table I summarizes mechanical properties of the composite and monolithic alumina obtained from the same extrusion process(after 3rd extru- sion step). The fracture toughness was increased by intro- ducing the fine zirconia filaments into alumina matrix. The volume fraction of monolithic zirconia phase on the fracture surface was 22% and that of as-sintered surface was 6%. It is clear that tetragonal phase transformed into monoclinic phase by applied stress. The result suggests that"stress 500um induced"transformation is one of the toughening mecha nisms for the improvement of the toughness of the co- extruded composite. The propagation of an indenter induced crack on a 3rd co-extruded composite is shown in Fig. 5 Crack deflection at zirconia filament was observed. It is clear that this crack deflection resulted from the residual 50 50um Fig. 4. Optical micrographs of the third stage co-extruded 10 vol% zirconia/alumina extrudate. (a)Cross-sectional and (b) side view of Fig. 5. Propagation of crack generated by indentation in 10 vol. zircon composite sintered at 1 600C for 3 h. alumina composite sintered at 1600C for 3 h
volume of filament was large enough to accumulate the longitudinal strain energy which promotes crack formation normal to extrusion direction. Fig. 4(a) and (b) shows the cross-sectional and side view of 3rd stage co-extruded composite after sintering at 1600 jC for 3 h. The diameter of zirconia filament further reduced to about 20 Am, which was near the theoretical value expected from both the diameter of zirconia filament in 2nd stage composite and the reduction rate. It is obvious that the co-extrusion process effectively reduced the diameter of zirconia filaments. The number of zirconia filaments per unit area was measured from Fig. 4(a). The estimated total number of the filaments within the 3rd extruded body was about 2500, which was less than half of the theoretical number of 5476 (148 37). Most of zirconia filaments have broken down and were not continuous as seen in Fig. 4(b). However, the zirconia phase still maintained unidirectional alignment. The reduced number of filaments estimated in cross-sectional view was attributed to this discontinuity of the filaments since we could not count the filaments which broke near the observed cross-section. No cracks were observed in both zirconia and alumina phases. The absence of cracks in zirconia phase may be attributed to both the reduced diameter and the limited length of filaments since such thin and short filaments cannot retain large strain energy for crack formation. Thus, we could get crack-free composite by refining the microstructure with repeated coextrusion process. The disappearance of cracks by refining the microstructure sometimes can be seen in this kind of experiments. For example, Menon and Chen [6] also reported that crack-free alumina/zirconia laminate composites were obtained when the thickness of the laminate was less than 50 Am. Mechanical properties were evaluated for the composites after the 3rd extrusion step. Table 1 summarizes mechanical properties of the composite and monolithic alumina obtained from the same extrusion process (after 3rd extrusion step). The fracture toughness was increased by introducing the fine zirconia filaments into alumina matrix. The volume fraction of monolithic zirconia phase on the fracture surface was 22% and that of as-sintered surface was 6%. It is clear that tetragonal phase transformed into monoclinic phase by applied stress. The result suggests that ‘‘stressinduced’’ transformation is one of the toughening mechanisms for the improvement of the toughness of the coextruded composite. The propagation of an indenter induced crack on a 3rd co-extruded composite is shown in Fig. 5. Crack deflection at zirconia filament was observed. It is clear that this crack deflection resulted from the residual Fig. 4. Optical micrographs of the third stage co-extruded 10 vol.% zirconia/alumina extrudate. (a) Cross-sectional and (b) side view of composite sintered at 1600 jC for 3 h. Table 1 Mechanical properties of the composite after 3rd extrusion step Sample 4Pt bending strength, MPa Toughness (SEPB), MPa m1/2 10 vol.% ZrO2/Al2O3 312 F 6 5.3 F 0.3 Al2O3 462 F 135 4.3 F 0.0 Fig. 5. Propagation of crack generated by indentation in 10 vol.% zirconia/ alumina composite sintered at 1600 jC for 3 h. H. Miyazaki et al. / Materials Letters 58 (2004) 1410–1414 1413
1414 H. Miyazaki et al. / Materials Letters 58(2004)1410-1414 this study show a potentiality to obtain fine-scale fibrous microstructure by extrusion process and excellent mechan- ical properties when the reduction ratio of the die and sintering temperature are optimized further. 4. Conclusion Ten volume-percent zirconia/alumina bi-phase composite with fine-scale aligned microstructure could be fabricated by piston co-extrusion through a reduction die. First and 200 second extrusion steps produced composites with cracks due to the mismatch in thermal expansion coefficient between Fig. 6. Fracture surface of the third stage co-extruded 10 voL% zirconia/ alumina and zirconia ceramics. However, a crack-free 10% alumina composite sintered at 1600C for 3 h zirconia/alumina composite was obtained after 3rd extrusion step. The toughness of the 3rd extruded composite was 5.3 ress produced by mismatch of thermal expansion coeffi- MPa m, which was larger than that of the monolithic cients between alumina and zirconia. It seems reasonable to alumina. Although the zirconia filament lost continuity after suppose that the crack deflection at zirconia filaments the 3rd extrusion process, effectiveness of the aligned another reason for the increment of the fracture toughness. fibrous microstructure on the toughness is emerged. The strength decreased by introducing the fine zirconia filaments into alumina matrix. Fig. 6 shows the fracture surface of the composite. A large zirconia filament with diameter of around 200 um was observed. We believe that this large zirconia filament made stress distribution in the [1] S. Baskaran, S.D. Nunn, D. Popovic, J.W. Halloran, J. Am. Ceram. composite more complicated and introduced a fracture Soc.76(1993)220 origin around the filament, which lowered the strength. [2] C.V. Hoy, A Barda, M. Griffith, J W.Halloran, J Am Ceram Soc. 81 Therefore, we suppose that it would be possible to fabricate 1998)152 zirconia/alumina composite with higher strength and tough 3]YH Koh, H W. Kim, H E. Kim, J Am Ceram Soc. 85(2002)2578 ness than that of the monolithic alumina by strictly control- (] C. Kaya, E.G. Butler, M.H. Lewis, J. Eur. Ceram Soc. 23(2003)935. []R.C. Garvie, P.S. Nicholson, J. Am. Ceram. Soc. 55(1972)30 ling the diameter of the zirconia filaments. The results in [6] M Menon, L.W. Chen, J Am Ceram Soc. 82(1999)3422
stress produced by mismatch of thermal expansion coefficients between alumina and zirconia. It seems reasonable to suppose that the crack deflection at zirconia filaments is another reason for the increment of the fracture toughness. The strength decreased by introducing the fine zirconia filaments into alumina matrix. Fig. 6 shows the fracture surface of the composite. A large zirconia filament with diameter of around 200 Am was observed. We believe that this large zirconia filament made stress distribution in the composite more complicated and introduced a fracture origin around the filament, which lowered the strength. Therefore, we suppose that it would be possible to fabricate zirconia/alumina composite with higher strength and toughness than that of the monolithic alumina by strictly controlling the diameter of the zirconia filaments. The results in this study show a potentiality to obtain fine-scale fibrous microstructure by extrusion process and excellent mechanical properties when the reduction ratio of the die and sintering temperature are optimized further. 4. Conclusion Ten volume-percent zirconia/alumina bi-phase composite with fine-scale aligned microstructure could be fabricated by piston co-extrusion through a reduction die. First and second extrusion steps produced composites with cracks due to the mismatch in thermal expansion coefficient between alumina and zirconia ceramics. However, a crack-free 10% zirconia/alumina composite was obtained after 3rd extrusion step. The toughness of the 3rd extruded composite was 5.3 MPam1/2, which was larger than that of the monolithic alumina. Although the zirconia filament lost continuity after the 3rd extrusion process, effectiveness of the aligned fibrous microstructure on the toughness is emerged. References [1] S. Baskaran, S.D. Nunn, D. Popovic, J.W. Halloran, J. Am. Ceram. Soc. 76 (1993) 2209. [2] C.V. Hoy, A. Barda, M. Griffith, J.W. Halloran, J. Am. Ceram. Soc. 81 (1998) 152. [3] Y.H. Koh, H.W. Kim, H.E. Kim, J. Am. Ceram. Soc. 85 (2002) 2578. [4] C. Kaya, E.G. Butler, M.H. Lewis, J. Eur. Ceram. Soc. 23 (2003) 935. [5] R.C. Garvie, P.S. Nicholson, J. Am. Ceram. Soc. 55 (1972) 303. [6] M. Menon, I.W. Chen, J. Am. Ceram. Soc. 82 (1999) 3422. Fig. 6. Fracture surface of the third stage co-extruded 10 vol.% zirconia/ alumina composite sintered at 1600 jC for 3 h. 1414 H. Miyazaki et al. / Materials Letters 58 (2004) 1410–1414