Availableonlineatwww.sciencedirect.com . Science Direct E≈RS ELSEVIER Joumal of the European Ceramic Society 26(2006)3539-3546 www.elsevier.comlocate/jeurceramsoc Effect of the volume ratio of zirconia and alumina on the mechanical properties of fibrous zirconia/alumina bi-phase composites prepared by co-extrusion Hiroyuki Miyazaki", Yu-ichi Yoshizawa, Kiyoshi hira Advanced Manufacturing Research institute, National Institute of Advanced Industrial Science and Technology (AIST 2266-98, Shimo-shidami, Moriyama- ku, Nagoya 463-8560, Japan Received 30 October 2005: received in revised form 16 December 2005 ted 28 December 2005 Available online 20 Febr Fibrous zirconia/alumina composites with different composition were fabricated by piston co-extrusion. After a 3rd extrusion step and sintering at 1600C, crack-free composites with a fibre width of -50 um were obtained for all compositions. The effect of the volume ratio of secondary phase on the mechanical properties was investigated. The Youngs modulus of the composites decreased linearly with increasing the zirconia content. The fracture toughness of the composites was improved by introducing fine second phase filaments into the matrix. The maximum fracture toughness of 6.2 MPa m 2 was attained in the 3rd co-extruded 47/53 vol% zirconia/alumina composite. The improvement in toughness was attributed to bothstress-induced"transformation of zirconia and a crack deflection mechanism due to thermal expansion mismatch between the two phase Bending strength of the composites was almost the same as that of the monolithic alumina regardless of the composition o 2006 Elsevier ltd. all rights reserved. Keywords: Co-extrusion; ZrO2/Al2O3 composite; Mechanical properties 1. Introduction reports on the microstructure of the fibrous composites with a rigid interface,8-14 but the evaluation of their mechanical prop- Continuous fibre-reinforced ceramics have promising poten- erties is limited. 8-11 tial for high-temperature application because of the remarkable In our previous reports, the feasibility of forming fine-scale fracture toughening through the crack bridging mechanism. fibrous microstructures was demonstrated for alumina/zirconia However, they are very expensive, and many problems appear in composites with a rigid interface, by using a repeated co- the fabrication process. Different geometric configurations that extrusion process through a reduction die. 10, I It was revealed incorporate crack deflecting systems, such as ceramic/ceramic that fracture toughness of these composites was increased over lamellar compositesand fibrous monolithic ceramics, are pro- the constituent monoliths. Leeet al clarified the effect of number posed for low-cost alternatives to conventional continuous-fiber of extrusion steps on the mechanical properties of the 50/50 vol% ceramic composites. In the case of such composites, increased alumina-(m-zirconia)/t-zirconia fibrous composite. However, toughness is usually associated with the presence of a weak the examination of the mechanical properties of the fibrous com- interface, which enables crack deflection into the interface. The posites has been limited to a few compositions. It seems possibl weak interface, however, results in a reduction in strength. In to improve the mechanical properties further by optimizing the order to maintain strength and yet obtain high toughness, lamel- composition. The aim of this study is to clarify the effect of the composites with a strong interface were proposed, in which volume fraction of zirconia phase on the mechanical properties crack deflection was generated by thermal residual stress at the of the fibrous zirconia/alumina composites interface. -In the case of fibrous composites, there are some 2. Experimental Corresponding author. Tel: 7367486:fax:+81527367405 Commercial alumina powder(AL-160SG-4, Showa De E-mail address: h-miyazaki ( aist go. jP(H Miyazaki) KK, Japan) and yttria-stabilized zirconia powder (TZ-3Y 0955-2219/S-see front matter o 2006 Elsevier Ltd. All rights reserved. doi: 10. 1016/j-jeurceramsoc 2005.12 019
Journal of the European Ceramic Society 26 (2006) 3539–3546 Effect of the volume ratio of zirconia and alumina on the mechanical properties of fibrous zirconia/alumina bi-phase composites prepared by co-extrusion Hiroyuki Miyazaki ∗, Yu-ichi Yoshizawa, Kiyoshi Hirao Advanced Manufacturing Research Institute, National Institute of Advanced Industrial Science and Technology (AIST), 2266-98, Shimo-shidami, Moriyama-ku, Nagoya 463-8560, Japan Received 30 October 2005; received in revised form 16 December 2005; accepted 28 December 2005 Available online 20 February 2006 Abstract Fibrous zirconia/alumina composites with different composition were fabricated by piston co-extrusion. After a 3rd extrusion step and sintering at 1600 ◦C, crack-free composites with a fibre width of ∼50m were obtained for all compositions. The effect of the volume ratio of secondary phase on the mechanical properties was investigated. The Young’s modulus of the composites decreased linearly with increasing the zirconia content. The fracture toughness of the composites was improved by introducing fine second phase filaments into the matrix. The maximum fracture toughness of 6.2 MPa m1/2 was attained in the 3rd co-extruded 47/53 vol% zirconia/alumina composite. The improvement in toughness was attributed to both “stress-induced” transformation of zirconia and a crack deflection mechanism due to thermal expansion mismatch between the two phases. Bending strength of the composites was almost the same as that of the monolithic alumina regardless of the composition. © 2006 Elsevier Ltd. All rights reserved. Keywords: Co-extrusion; ZrO2/Al2O3 composite; Mechanical properties 1. Introduction Continuous fibre-reinforced ceramics have promising potential for high-temperature application because of the remarkable fracture toughening through the crack bridging mechanism. However, they are very expensive, and many problems appear in the fabrication process. Different geometric configurations that incorporate crack deflecting systems, such as ceramic/ceramic lamellar composites1 and fibrous monolithic ceramics,2 are proposed for low-cost alternatives to conventional continuous-fiber ceramic composites. In the case of such composites, increased toughness is usually associated with the presence of a weak interface, which enables crack deflection into the interface. The weak interface, however, results in a reduction in strength. In order to maintain strength and yet obtain high toughness, lamellar composites with a strong interface were proposed, in which crack deflection was generated by thermal residual stress at the interface.3–7 In the case of fibrous composites, there are some ∗ Corresponding author. Tel.: +81 52 736 7486; fax: +81 52 736 7405. E-mail address: h-miyazaki@aist.go.jp (H. Miyazaki). reports on the microstructure of the fibrous composites with a rigid interface,8–14 but the evaluation of their mechanical properties is limited.8–11 In our previous reports, the feasibility of forming fine-scale fibrous microstructures was demonstrated for alumina/zirconia composites with a rigid interface, by using a repeated coextrusion process through a reduction die.10,11 It was revealed that fracture toughness of these composites was increased over the constituent monoliths. Lee et al. clarified the effect of number of extrusion steps on the mechanical properties of the 50/50 vol% alumina-(m-zirconia)/t-zirconia fibrous composite.9 However, the examination of the mechanical properties of the fibrous composites has been limited to a few compositions. It seems possible to improve the mechanical properties further by optimizing the composition. The aim of this study is to clarify the effect of the volume fraction of zirconia phase on the mechanical properties of the fibrous zirconia/alumina composites. 2. Experimental Commercial alumina powder (AL-160SG-4, Showa Denko K.K., Japan) and yttria-stabilized zirconia powder (TZ-3Y, 0955-2219/$ – see front matter © 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.jeurceramsoc.2005.12.019
3540 H Miyazaki et al. /Journal of the European Ceramic Sociery 26(2006)3539-3546 Tosoh, Japan) were used as the starting powders. The powders measurement was conducted with an inner and outer span were milled individually in ethanol, and the slurry was dried of 10 mm and 30 mm, respectively, and a crosshead speed of in a rotary evaporator. The alumina powder was tempered with 0.5 mm/min. Youngs modulus was measured from the load- distilled water, organic binders and plasticizer using a hook type displacement curves which were compensated for the compli mixing machine and a roller mill. Similarly, the zirconia powder ance of both the testing machine and the jig Fracture toughness was tempered with distilled water, organic binders, plasticizer (Kic) was determined by the single-edged-precracked-beam nd dispersant using the same apparatus. Detailed description of (SEPB)method with a span of 16 mm. Fracture strength and the mixing process for this study was reported in our previous fracture toughness measurements, as well as Youngs modu- papers. 0. I Each of the obtained plastic bodies was extruded lus measurements, were carried out so that tensile stress during into a hexagonal shape by means of a screw auger machine. measurements was parallel to the extruding direction. Fracture A total of 37 monofilaments were bundled into a hexagonal surfaces and polished surfaces of the composites were observed feedrod. In order to fabricate zirconia/alumina composites with using scanning electron microscope and optical microscope different compositions, the numbers of each zirconia and alu- The grain size of sintered samples was determined by the mean mina monofilament in the feedrod were varied in the ratio of linear intercept method using micrographs of polished and ther- 4/33, 12/25, 18/19, 25/12 and 33/4(Fig. 1). The initial feedrod mally etched surface. 5 The average grain size was calculated was extruded through a 6: I hexagonal reduction die using a pis- by multiplying the mean linear intercept by 1.78. In order to ton extruder. After the first co-extrusion, the individual pieces evaluate the"stress-induced"phase transformation of zirconia were bundled to create a second feedrod. The same process was as well as the residual stress, X-ray diffraction analysis was repeated for the second and third co-extrusion. After the third conducted. It is well known that grinding induces transforma Co-extrusion, the extruded green bodies were dried at room tem- tion of zirconia in the surface region o so that measurements of perature under atmospheric pressure. The dried bodies were then monoclinic phase before fracture were taken place on the pol calcined at 600C for 4 h under a flow of nitrogen to remove ished surface from which the grinding-transformed layer was organic substances, followed by calcining at 500C for 3 h in air completely removed. The monoclinic phase after fracture was to completely remove remaining carbon substances. After being measured on the fractured surface. The volume fraction of mo cold isostatically pressed( CIPed)under 300 MPa, they were aclinic zirconia in the zirconia phase, m/(m+t), before and after pressureless sintered at 1600C for 3 h. The final volume frac- fracture was calculated according to Garvie's equation. Then ions fz of the zirconia for each composition were 10, 31, 47, 66 the volume fraction of the transformed zirconia in the compos and 88 vol%, respectively, which were slightly different from the ite was attained by multiplying the volume fraction of zirconia, yalues estimated from the ratio of each monofilament since the fz. In order to estimate the residual stress in both zirconia and net contents of powder in the each monofilament were different. alumina phase caused by thermal expansion mismatch between For the comparison, monolithic alumina, monolithic zirconia the two phases, the full width at half maximum(FWHM) of and powder-mixture composites(10/90, 50/50 and 90/10 vols zirconia(13 3)peak and alumina(1.0.10)peak was measured zirconia/alumina)were also fabricated by the same procedure. on the polished surface after subtraction of the contribution Density measurements were conducted for both the com- from Ko2 peak. The XRD measurements were repeated for osites and monoliths using the Archimedes technique. With 3-4 times to attain both the average and standard deviation of dimensions of 3 mm x 4 mm x 35 mm, 9-12 specimens of each FWHM composition were machined from the sintered samples for mechanical property measurement Four-point bending strength 3. Results and discussion Pushing rod sl 3.. Microstructure and density Fig. 2 shows optical micrographs of the composites for which Paste A fz was 31, 47 and 66 vol%, respectively. The zirconia phase has a lighter contrast, and alumina has a darker contrast. although the width of the phases was not uniform and interfaces were 5 mm corrugated, the width of the most filaments of minor phase was reduced to m50 um and the number of filaments was increased significantly, close to the theoretical value expected from the number of extrusion. It is obvious that the co-extrusion process effectively reduced the width of each filament. Fine scale fibrous microstructures similar to those in Fig. 2 were also fabricated Piston extrusion in the 10/90 vol% zirconia/alumina composite and 88/12 vol% zirconia/alumina composite by the 3rd co-extrusion process, as reported elsewhere. 0, I Most of the filaments were not con- ° 5mm Extrudate tinuous after the 3rd extrusion as seen in Fig. 2. However, each Fig. 1. Schematic illustration for fabrication of fibrous bi-phase composites by phase still maintained unidirectional alignment. Cracks were not observed with an optical microscope in any of the composites
3540 H. Miyazaki et al. / Journal of the European Ceramic Society 26 (2006) 3539–3546 Tosoh, Japan) were used as the starting powders. The powders were milled individually in ethanol, and the slurry was dried in a rotary evaporator. The alumina powder was tempered with distilled water, organic binders and plasticizer using a hook type mixing machine and a roller mill. Similarly, the zirconia powder was tempered with distilled water, organic binders, plasticizer and dispersant using the same apparatus. Detailed description of the mixing process for this study was reported in our previous papers.10,11 Each of the obtained plastic bodies was extruded into a hexagonal shape by means of a screw auger machine. A total of 37 monofilaments were bundled into a hexagonal feedrod. In order to fabricate zirconia/alumina composites with different compositions, the numbers of each zirconia and alumina monofilament in the feedrod were varied in the ratio of 4/33, 12/25, 18/19, 25/12 and 33/4 (Fig. 1). The initial feedrod was extruded through a 6:1 hexagonal reduction die using a piston extruder. After the first co-extrusion, the individual pieces were bundled to create a second feedrod. The same process was repeated for the second and third co-extrusion. After the third co-extrusion, the extruded green bodies were dried at room temperature under atmospheric pressure. The dried bodies were then calcined at 600 ◦C for 4 h under a flow of nitrogen to remove organic substances, followed by calcining at 500 ◦C for 3 h in air to completely remove remaining carbon substances. After being cold isostatically pressed (CIPed) under 300 MPa, they were pressureless sintered at 1600 ◦C for 3 h. The final volume fractions fZ of the zirconia for each composition were 10, 31, 47, 66 and 88 vol%, respectively, which were slightly different from the values estimated from the ratio of each monofilament since the net contents of powder in the each monofilament were different. For the comparison, monolithic alumina, monolithic zirconia and powder-mixture composites (10/90, 50/50 and 90/10 vol% zirconia/alumina) were also fabricated by the same procedure. Density measurements were conducted for both the composites and monoliths using the Archimedes technique. With dimensions of 3 mm × 4 mm × 35 mm, 9–12 specimens of each composition were machined from the sintered samples for mechanical property measurement. Four-point bending strength Fig. 1. Schematic illustration for fabrication of fibrous bi-phase composites by co-extrusion process. measurement was conducted with an inner and outer span of 10 mm and 30 mm, respectively, and a crosshead speed of 0.5 mm/min. Young’s modulus was measured from the loaddisplacement curves which were compensated for the compliance of both the testing machine and the jig. 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, as well as Young’s modulus measurements, were carried out so that tensile stress during measurements was parallel to the extruding direction. Fracture surfaces and polished surfaces of the composites were observed using scanning electron microscope and optical microscope. The grain size of sintered samples was determined by the mean linear intercept method using micrographs of polished and thermally etched surface.15 The average grain size was calculated by multiplying the mean linear intercept by 1.78. In order to evaluate the “stress-induced” phase transformation of zirconia as well as the residual stress, X-ray diffraction analysis was conducted. It is well known that grinding induces transformation of zirconia in the surface region,16 so that measurements of monoclinic phase before fracture were taken place on the polished surface from which the grinding-transformed layer was completely removed. The monoclinic phase after fracture was measured on the fractured surface. The volume fraction of monoclinic zirconia in the zirconia phase, m/(m+t), before and after fracture was calculated according to Garvie’s equation.17 Then the volume fraction of the transformed zirconia in the composite was attained by multiplying the volume fraction of zirconia, fZ. In order to estimate the residual stress in both zirconia and alumina phase caused by thermal expansion mismatch between the two phases, the full width at half maximum (FWHM) of zirconia (1 3 3) peak and alumina (1.0.10) peak was measured on the polished surface after subtraction of the contribution from K2 peak. The XRD measurements were repeated for 3–4 times to attain both the average and standard deviation of FWHM. 3. Results and discussion 3.1. Microstructure and density Fig. 2 shows optical micrographs of the composites for which fZ was 31, 47 and 66 vol%, respectively. The zirconia phase has a lighter contrast, and alumina has a darker contrast. Although the width of the phases was not uniform and interfaces were corrugated, the width of the most filaments of minor phase was reduced to ∼50m and the number of filaments was increased significantly, close to the theoretical value expected from the number of extrusion. It is obvious that the co-extrusion process effectively reduced the width of each filament. Fine scale fibrous microstructures similar to those in Fig. 2 were also fabricated in the 10/90 vol% zirconia/alumina composite and 88/12 vol% zirconia/alumina composite by the 3rd co-extrusion process, as reported elsewhere.10,11 Most of the filaments were not continuous after the 3rd extrusion as seen in Fig. 2. However, each phase still maintained unidirectional alignment. Cracks were not observed with an optical microscope in any of the composites.
H Miyazaki et al Journal of the European Ceramic Society 26(2006)3539-3546 2004m 200um 00 Fig. 2. Optical micrographs of both cross section(upper row) and side view(lower row) of zirconia/alumina fibrous composites sintered at 1600C for 3 h. The olume fraction of zirconia phase is(a)and (d ): 31%;(b)and(e): 47% and(c)and (f): 66%0, respectively. Fig 3. SEM image of (a)alumina monolith;(b)alumina phase in 47/53 vol% zirconia/alumina fibrous composite; (c)zirconia phase in 47/53 vol o zirconia/alumina fibrous composite; (d)zirconia monolith and (e)50/50 vol% zirconia/alumina powder-mixture composite Fig. 3 shows SEM micrographs of the alumina and zirco- age grain sizes of both alumina and zirconia. The average grain nia monoliths, the fibrous composite(47 vol%o-zirconia)and the size of alumina in the fibrous composites was almost the same powder-mixture composite(50 vol-zirconia) In order to eval- regardless of the composition and larger than that of the mono- uate the microstructures of monoliths and alumina or zirconia lithic alumina. Xue et al. reported that y+ ion enhances grain hase in the composites quantitatively, average grain size was growth of alumina in the presence of glassy phase on the grain measured by the intercept method. Table 1 summarizes the aver- boundary which forms from silica contamination. The alumina Table 1 age grain size of each phase and relative density of the constituent monoliths, the fibrous composites and the powder-mixture composites Sample Alumina Fibrous composites Zirconia Powder-mixture composites 66 00 in size of alumina(um) 4.2 98898299197.3 982 a volume fraction of ZrO2(%)
H. Miyazaki et al. / Journal of the European Ceramic Society 26 (2006) 3539–3546 3541 Fig. 2. Optical micrographs of both cross section (upper row) and side view (lower row) of zirconia/alumina fibrous composites sintered at 1600 ◦C for 3 h. The volume fraction of zirconia phase is (a) and (d): 31%; (b) and (e): 47% and (c) and (f):66%, respectively. Fig. 3. SEM image of (a) alumina monolith; (b) alumina phase in 47/53 vol% zirconia/alumina fibrous composite; (c) zirconia phase in 47/53 vol% zirconia/alumina fibrous composite; (d) zirconia monolith and (e) 50/50 vol% zirconia/alumina powder-mixture composite. Fig. 3 shows SEM micrographs of the alumina and zirconia monoliths, the fibrous composite (47 vol%-zirconia) and the powder-mixture composite (50 vol%-zirconia). In order to evaluate the microstructures of monoliths and alumina or zirconia phase in the composites quantitatively, average grain size was measured by the intercept method. Table 1 summarizes the average grain sizes of both alumina and zirconia. The average grain size of alumina in the fibrous composites was almost the same regardless of the composition and larger than that of the monolithic alumina. Xue et al. reported that Y3+ ion enhances grain growth of alumina in the presence of glassy phase on the grain boundary which forms from silica contamination.18 The alumina Table 1 Average grain size of each phase and relative density of the constituent monoliths, the fibrous composites and the powder-mixture composites Sample Alumina Fibrous composites Zirconia Powder-mixture composites 0a 10a 31a 47a 66a 88a 100a 10a 50a 90a Grain size of alumina (m) 4.2 5.8 6.6 6.0 6.0 5.4 – 1.8 1.3 1.1 Grain size of zirconia (m) – 1.6 1.5 1.4 1.5 1.4 1.3 0.7 1.1 1.4 Relative density (%) 98.8 98.8 98.2 99.1 97.3 97.8 100 99.3 99.3 98.2 a Volume fraction of ZrO2 (%)
H Miyazaki et al. / Journal of the European Ceramic Sociery 26(2006)3539-3546 powder used in this study contained 0.02 wt silica It is likely that yttrium in the zirconia phase may have diffused into the alumina phase and affected the grain growth of alumina By con trast, the size of alumina grain in the powder-mixture composites decreased significantly with increasing the zirconia content. The inhibition of grain growth in matrix phase by addition of sec- ndary phase particles is well known as"pinning effect". 19-21 The grain growth of alumina in the powder-mixture compos ites was inhibited by the pining effect of the zirconia particles. The grain size of zirconia in the fibrous composites was almost the same as that of the monolithic zirconia. The size of 4.0 grain in the powder-mixture composites decreased with increas ing the alumina content because of the pinning effect by alumina O Powder- mixture composite particles. Both the monolithic alumina and monolithic zirconia were Volume fraction of ZrO phase fz(vol%) sintered to almost full density (Table 1), whereas the relative density of the composites was slightly lower than that of the Fig. 5. Dependence of the fracture toughness of the composites on the volume monolithic. The insufficient densification in these fibrous com- fraction of zirconia phase fz posites may arise from mismatch in both the total amount of sintering shrinkage and the shrinkage rate between the two 3.3. Fracture toughness 3.3.1. Effect of"stress-induced"transformation of zirconia 3.2. Youngs modulus on the fracture toughness Fig 5 shows the fracture toughness of both the composites Fig 4 shows the dependence of the Youngs modulus on the and the constituent monolithic ceramics. The fracture tough ness of the powder-mixture composites was almost the same as fz for both the fibrous composites and the powder-mixture com- that of the monolithic alumina when the fz was 10 vol %, and posites. The solid line in the Fig. 4 shows calculated value using the well-known Voigt rule-of-mixture: increased with fz, then became saturated at the higher fz. To correlate toughness with the tetragonal-to-monoclinic transfor E= Ezf ea(-f2 mation, the volume fraction of the transformed zirconia phase in (1) the composites due to fracture was measured. The result is shown where Ez and Ea are the Youngs modulus of zirconia and alu- in Fig. 6 along with the data for the fibrous composites.The mina and fz is the volume fraction of the zirconia phase. The volume fraction of transformed zirconia in the powder-mixture Youngs modulus, Ez and Ea, used for the calculation was mea- composites was very small when fz was 10 vol%, and increased sured with the alumina and zirconia monoliths, respectively. with fz. It is obvious that the improvement in fracture toughness From Fig. 4. it is clear that the Young's modulus of both the of the powder-mixture composites was originated mainly from fibrous composites and powder-mixture composites followed the"stress-induced"transformation of zirconia phase the Voigt rule-of-mixture. The measured value was slightly lower than the predicted value, which is attributable to the lower relative densities of the composites. ◆ Co-extruded composite O Powder-mixture composite ◆ Co-extruded comp Powder-mixture comp 59 FE 6080100 Volume fraction of ZrO2 phase fz(vol% Volume fraction of ZrO2 phasefz(vol%) Fig. 6. Volume fraction of zirconia transformed from sites in both co-extruded composites and powder-mixture semamgosites o a function and powder-mixture composites on the volume fraction of zirconia phase fz. of volume fraction of zirconia phase fz
3542 H. Miyazaki et al. / Journal of the European Ceramic Society 26 (2006) 3539–3546 powder used in this study contained 0.02 wt% silica. It is likely that yttrium in the zirconia phase may have diffused into the alumina phase and affected the grain growth of alumina. By contrast, the size of alumina grain in the powder-mixture composites decreased significantly with increasing the zirconia content. The inhibition of grain growth in matrix phase by addition of secondary phase particles is well known as “pinning effect”.19–21 The grain growth of alumina in the powder-mixture composites was inhibited by the pining effect of the zirconia particles. The grain size of zirconia in the fibrous composites was almost the same as that of the monolithic zirconia. The size of zirconia grain in the powder-mixture composites decreased with increasing the alumina content because of the pinning effect by alumina particles. Both the monolithic alumina and monolithic zirconia were sintered to almost full density (Table 1), whereas the relative density of the composites was slightly lower than that of the monolithic. The insufficient densification in these fibrous composites may arise from mismatch in both the total amount of sintering shrinkage and the shrinkage rate between the two phases. 3.2. Young’s modulus Fig. 4 shows the dependence of the Young’s modulus on the fZ for both the fibrous composites and the powder-mixture composites. The solid line in the Fig. 4 shows the calculated value using the well-known Voigt rule-of-mixture: E = Ezfz + Ea(1 − fz) (1) where Ez and Ea are the Young’s modulus of zirconia and alumina and fZ is the volume fraction of the zirconia phase. The Young’s modulus, Ez and Ea, used for the calculation was measured with the alumina and zirconia monoliths, respectively. From Fig. 4, it is clear that the Young’s modulus of both the fibrous composites and powder-mixture composites followed the Voigt rule-of-mixture. The measured value was slightly lower than the predicted value, which is attributable to the lower relative densities of the composites. Fig. 4. Dependence of the Young’s modulus of both co-extruded composites and powder-mixture composites on the volume fraction of zirconia phase fZ. Fig. 5. Dependence of the fracture toughness of the composites on the volume fraction of zirconia phase fZ. 3.3. Fracture toughness 3.3.1. Effect of “stress-induced” transformation of zirconia on the fracture toughness Fig. 5 shows the fracture toughness of both the composites and the constituent monolithic ceramics. The fracture toughness of the powder-mixture composites was almost the same as that of the monolithic alumina when the fZ was 10 vol%, and increased with fZ, then became saturated at the higher fZ. To correlate toughness with the tetragonal-to-monoclinic transformation, the volume fraction of the transformed zirconia phase in the composites due to fracture was measured. The result is shown in Fig. 6 along with the data for the fibrous composites. The volume fraction of transformed zirconia in the powder-mixture composites was very small when fZ was 10 vol%, and increased with fZ. It is obvious that the improvement in fracture toughness of the powder-mixture composites was originated mainly from the “stress-induced” transformation of zirconia phase. Fig. 6. Volume fraction of zirconia transformed from tetragonal to monoclinic in both co-extruded composites and powder-mixture composites as a function of volume fraction of zirconia phase fZ
H Miyazaki et al Journal of the European Ceramic Society 26(2006)3539-3546 3543 By contrast, the fracture toughness of the fibrous ites was increased even by introducing a small amount of fine zirconia filaments into alumina matrix(Fig. 5), and was further increased with fz and reached a maximum(6.2 MPam")where fz was 47 vol%. Then the fracture toughness decreased slightly with fz but still maintained higher value than that of the mono- lithic zirconia. The volume fractions of transformed zirconia in the fibrous composites are shown in Fig. 6(solid square). The volume fraction of transformed zirconia in the fibrous compos- ites increased with fz, which was almost the same rising curve behavior as that of the powder- mixture composites. It is rea- sonable to suppose that contribution from the"stress-induced transformation to the toughness of the fibrous composites was 100m almost the same level as that of the powder-mixture composites Consequently, the higher fracture toughness of the fibrous com- posites over that of powder-mixture composites is attributed Fig. 8. SEM image of fracture surface of the 47/53 vol% zirconia/alumina co. extruded composite. No trace of the pullout of fibers was observed. another toughening mechanism. 3.3.2. Effect of crack deflection mechanism on the fracture deflection. and the effect of a crack deflection at an interface between the fiber and the matrix. The frequency of interaction Fig 7 shows crack propagation of an indenter-induced crack between the crack and the second phase is represented by the on both the 47/53 vol% zirconia/alumina fibrous composite and volume fraction of the second phase providing the width of the the 50/50 vol% zirconia/alumina powder-mixture composite. second phase filaments is constant. Adachi et al. reported that in the fibrous composite, whereas crack deflection was not phase lamer composite increased with increasing the difterence posite. The same result was also obtained at the 10/90 vol% assume that the effect of a crack deflection on the toughness is zirconia/alumina composite 10 It is clear that the residual stress almost proportional to the variation in the residual stress across produced by mismatch of thermal expansion between alumina and zirconia affected the crack propagation more effectively in due to the crack-deflecting mechanism on fz can be analyzed the case of fibrous microstructure further increasing the fracture by the product of the volume fraction of the secondary phase toughness. Although the composites have the fibrous microstru ture, pullout of the fine fibrous second phase was not observed the matrix. The residual thermal stress in the fiber-reinforced (Fig8). The lack of pullout of the fibrous phase is due to a tough composites was analyzed by Budiansky et al., by using the com- bonding at the interface between the two phases posite cylinder model. The average of residual stress of matrix fz over 47 vol%(Fig. 5), despite the increased rot osites with is given by the following equation The decrease in toughness of the fibrous com olume fraction 入2「Er from the"crack-deflection"mechanism caused by residual stress Em-X EJLI-Vml of transformed zirconia(Fig. 6), implies that the contribution decreased. The contribution of the crack-deflection mechanism where g is axial stress. e the youngs modulus. c the volume to the increment in fracture toughness of the fibrous composite fraction, v Poisson's ratio and subscripts m and f refer to matrix an be estimated by the product of the frequency of interaction and fiber, respectively. A1 and A2 are functions of cm, Em/Er, between the crack and the second phase, which causes crack vm and vf shown explicitly in appendix. S2 is the thermal strain 50um 50m Fig. 7. Propagation of crack generated by indentation in(a)47/53 vol% zirconia/alumina co-extruded composite and( b)50/50 vol o zirconia/alumina powder- composite. The indentation load was 196N for(a) and 98N for(b)
H. Miyazaki et al. / Journal of the European Ceramic Society 26 (2006) 3539–3546 3543 By contrast, the fracture toughness of the fibrous composites was increased even by introducing a small amount of fine zirconia filaments into alumina matrix (Fig. 5), and was further increased with fZ and reached a maximum (6.2 MPa m1/2) where fZ was 47 vol%. Then the fracture toughness decreased slightly with fZ but still maintained higher value than that of the monolithic zirconia. The volume fractions of transformed zirconia in the fibrous composites are shown in Fig. 6 (solid square). The volume fraction of transformed zirconia in the fibrous composites increased with fZ, which was almost the same rising curve behavior as that of the powder-mixture composites. It is reasonable to suppose that contribution from the “stress-induced” transformation to the toughness of the fibrous composites was almost the same level as that of the powder-mixture composites. Consequently, the higher fracture toughness of the fibrous composites over that of powder-mixture composites is attributed to another toughening mechanism. 3.3.2. Effect of crack deflection mechanism on the fracture toughness Fig. 7 shows crack propagation of an indenter-induced crack on both the 47/53 vol% zirconia/alumina fibrous composite and the 50/50 vol% zirconia/alumina powder-mixture composite. Crack deflection at the zirconia/alumina interface was observed in the fibrous composite, whereas crack deflection was not observed by an optical microscope in the powder-mixture composite. The same result was also obtained at the 10/90 vol% zirconia/alumina composite.10 It is clear that the residual stress produced by mismatch of thermal expansion between alumina and zirconia affected the crack propagation more effectively in the case of fibrous microstructure further increasing the fracture toughness. Although the composites have the fibrous microstructure, pullout of the fine fibrous second phase was not observed (Fig. 8). The lack of pullout of the fibrous phase is due to a tough bonding at the interface between the two phases. The decrease in toughness of the fibrous composites with fZ over 47 vol% (Fig. 5), despite the increased volume fraction of transformed zirconia (Fig. 6), implies that the contribution from the “crack-deflection” mechanism caused by residual stress decreased. The contribution of the crack-deflection mechanism to the increment in fracture toughness of the fibrous composite can be estimated by the product of the frequency of interaction between the crack and the second phase, which causes crack Fig. 8. SEM image of fracture surface of the 47/53 vol% zirconia/alumina coextruded composite. No trace of the pullout of fibers was observed. deflection, and the effect of a crack deflection at an interface between the fiber and the matrix. The frequency of interaction between the crack and the second phase is represented by the volume fraction of the second phase providing the width of the second phase filaments is constant. Adachi et al. reported that the degree of the crack deflection in the alumina/zirconia biphase lamer composite increased with increasing the difference of residual stress between the two phases.7 It is reasonable to assume that the effect of a crack deflection on the toughness is almost proportional to the variation in the residual stress across the interface. Then the dependence of the increment in toughness due to the crack-deflecting mechanism on fZ can be analyzed by the product of the volume fraction of the secondary phase and the difference of the residual stress between the fiber and the matrix. The residual thermal stress in the fiber-reinforced composites was analyzed by Budiansky et al., by using the composite cylinder model.22 The average of residual stress of matrix is given by the following equation. σm Em = λ2 λ1 Ef E cf 1 − νm Ω (2) where σ is axial stress, E the Young’s modulus, c the volume fraction, ν Poisson’s ratio and subscripts m and f refer to matrix and fiber, respectively. λ1 and λ2 are functions of cm, Em/Ef, νm and νf shown explicitly in appendix. is the thermal strain Fig. 7. Propagation of crack generated by indentation in (a) 47/53 vol% zirconia/alumina co-extruded composite and (b) 50/50 vol% zirconia/alumina powder-mixture composite. The indentation load was 196 N for (a) and 98 N for (b).