CERAMICS INTERNATIONAL ELSEⅤIER Ceramics International 28(2002)565-573 Interphase effects on the bend strength and toughness of an oxide fibre/oxide matrix composite Ramanan venkatesh School of Materials and Mineral Resources Engineering, Universiti Sains Malaysia, 14300 Nibong Tebal, Seberang Perai Setatan Pulau Pinang, Malaysia Received 30 April 2001; received in revised form 12 October 2001; accepted 3 December 2001 Abstract The effect of alumina-20 wt ZrO2(PRD-166) fibre and Sno2 coating properties on the bending strength and toughness of alumina fibre/SnO2/glass matrix composites have been investigated. The mean strength of as-received alumina-20 wt ZrO2 fibres was 1380 MPa for a gage length of 17 mm and decreased with increase in heat treatment temperatures. It was also observed that as the Sno, coating thickness increased, roughness of the coating increased and this decreased the strength of the fibres. This rough ness effect had serious implications on the fracture characteristics of PRD-166/SnO2/glass and Saphikon/SnO2/glass matrix com- posites. PRD-166/SnO,glass matrix composites exhibited non-planar failure with fiber bridging and fibre debonding as major toughening mechanisms. Saphikon/SnO2/glass matrix composites failed in a tough manner with extensive fibre pullout. The differ- nce the failure mode between PRD-166/ SnO2/glass and Saphikon/SnO2/glass matrix composites was attributed to the clamping stress associated with fiber roughness at the PrD-166/SnO, interface as compared to the smoother Saphikon/ Sno, interface C 2002 Elsevier Science Ltd and Techna S.r. l. All rights reserved Keywords: Interface effects; Bend strength; Toughness; Oxide fibre composite Introduction irregularities at the fibre /matrix interface. In composites with strong bonding at the interface, cracks originating Fibre reinfor great potential for in the brittle matrix tensile cut through the fibres. improving strength and toughness of ceramic materials resulting in a planar brittle failure of the composite In [1-5]. Parameters that influence the properties of fibres composite with a weak bonding at the interface, when in ceramic matrix composites(CMCs) include: tensile matrix tensile strength is exceeded, multiple cracking of strength, strain to failure, Weibull modulus, aspect ratio the matrix takes place with the fibres having enough and surface roughness. Alumina, mullite and zirconia strength to bridge the cracks. Further increase in stress are the principal polycrystalline oxide fibres developed causes fibre debonding due to interfacial stress and [6-21]. Oxide fibre/oxide matrix composites are con- Poisson's effect. Continued stressing of the composite sidered for potential use at extremely high temperatures beyond fibre debonding causes the failure of the fibre (1400-1600C)and in severe environments [22-31]. along its length and then depending on residual stress, Failure strength and toughness of CMCs depend on a Poissons ratio of fibre and matrix and interfacial fric multitude of mechanisms involving matrix microcrack- tional stress, fibre pullout occurs. Fibre pullout is the ing, matrix prestressing, fibre debonding and fibre pull- main toughening mechanism in CMCs. Hence for a out. Strength and toughness of CMCs are greatly tough composite, the interface bonding should be influenced by the interfacial bonding at the fibre/matrix strong enough to allow load transfer but weak enough interface. Interfacial strength is a strong function of the to aid crack deflection, fibre debonding and fibre pull- degree of bonding(chemical or mechanical) between out Interfacial strength can be controlled by modifying fibre and matrix and the thermal mismatch between fibre the fibre/matrix reactions at the processing and service and matrix. Mechanical bonding is primarily due to temperatures either through proper selection of materi als or by means of interface gs. For a operate successfully the following conditions should be 0272-8842/02/S22.00C 2002 Elsevier Science Ltd and Techna S.r. l. All rights reserved. PII:S0272-8842(02)00011-1
Interphase effects on the bend strength and toughness of an oxide fibre/oxide matrix composite Ramanan Venkatesh School of Materials and Mineral Resources Engineering, Universiti Sains Malaysia, 14300 Nibong Tebal, Seberang Perai Setatan, Pulau Pinang, Malaysia Received 30 April 2001; received in revised form 12 October 2001; accepted 3 December 2001 Abstract The effect of alumina–20 wt.% ZrO2 (PRD-166) fibre and SnO2 coating properties on the bending strength and toughness of alumina fibre/SnO2/glass matrix composites have been investigated. The mean strength of as-received alumina–20 wt.% ZrO2 fibres was 1380 MPa for a gage length of 17 mm and decreased with increase in heat treatment temperatures. It was also observed that as the SnO2 coating thickness increased, roughness of the coating increased and this decreased the strength of the fibres. This roughness effect had serious implications on the fracture characteristics of PRD-166/SnO2/glass and Saphikon/SnO2/glass matrix composites. PRD-166/SnO2/glass matrix composites exhibited non-planar failure with fiber bridging and fibre debonding as major toughening mechanisms. Saphikon/SnO2/glass matrix composites failed in a tough manner with extensive fibre pullout. The difference in the failure mode between PRD-166/SnO2/glass and Saphikon/SnO2/glass matrix composites was attributed to the clamping stress associated with fiber roughness at the PRD-166/SnO2 interface as compared to the smoother Saphikon/SnO2 interface. # 2002 Elsevier Science Ltd and Techna S.r.l. All rights reserved. Keywords: Interface effects; Bend strength; Toughness; Oxide fibre composite Introduction Fibre reinforcement offers a great potential for improving strength and toughness of ceramic materials [1–5]. Parameters that influence the properties of fibres in ceramic matrix composites (CMCs) include: tensile strength, strain to failure, Weibull modulus, aspect ratio and surface roughness. Alumina, mullite and zirconia are the principal polycrystalline oxide fibres developed [6–21]. Oxide fibre/oxide matrix composites are considered for potential use at extremely high temperatures (1400–1600 C) and in severe environments [22–31]. Failure strength and toughness of CMCs depend on a multitude of mechanisms involving matrix microcracking, matrix prestressing, fibre debonding and fibre pullout. Strength and toughness of CMCs are greatly influenced by the interfacial bonding at the fibre/matrix interface. Interfacial strength is a strong function of the degree of bonding (chemical or mechanical) between fibre and matrix and the thermal mismatch between fibre and matrix. Mechanical bonding is primarily due to irregularities at the fibre/matrix interface. In composites with strong bonding at the interface, cracks originating in the brittle matrix tensile cut through the fibres, resulting in a planar brittle failure of the composite. In a composite with a weak bonding at the interface, when matrix tensile strength is exceeded, multiple cracking of the matrix takes place with the fibres having enough strength to bridge the cracks. Further increase in stress causes fibre debonding due to interfacial stress and Poisson’s effect. Continued stressing of the composite beyond fibre debonding causes the failure of the fibre along its length and then depending on residual stress, Poisson’s ratio of fibre and matrix and interfacial frictional stress, fibre pullout occurs. Fibre pullout is the main toughening mechanism in CMCs. Hence for a tough composite, the interface bonding should be strong enough to allow load transfer but weak enough to aid crack deflection, fibre debonding and fibre pullout. Interfacial strength can be controlled by modifying the fibre/matrix reactions at the processing and service temperatures either through proper selection of materials or by means of interfacial coatings. For a coating to operate successfully the following conditions should be met. 0272-8842/02/$22.00 # 2002 Elsevier Science Ltd and Techna S.r.l. All rights reserved. PII: S0272-8842(02)00011-1 Ceramics International 28 (2002) 565–573 www.elsevier.com/locate/ceramint E-mail address: ram5nan@tm.net.mu (R. Venkatesh)
R Venkatesh/Ceramics International 28 (2002)565-573 1. Chemical compatibility with the matrix The alumina(Prd-166) fibres were coated with SnO 2. Refractoriness by a chemical vapor deposition technique. The alumina 3. Stability in oxidising, water vapor and corrosive fibre tows were placed in the central hot zone of the environments reactor and heated to the deposition temperature of 4. Providing a relatively weak fibre-coating inter- 500C. Dry nitrogen was the carrier gas for SnCl4. The phase allowing for fibre debonding and pullout. flow rate of nitrogen was I I/min. A second bubbler contained water heated to 80C through which oxygen In oxide fibre/oxide matrix composite systems, reac- was passed at a rate of 0.6 I/min. The deposition occur tions at the fibre/matrix interface, e.g. Al_O3/SiO2, red via the chemical reaction [38]. mullite/mullite leads to a strong interfacial bonding with the consequence of brittle behaviour of the composite. SnCl4(g)+ 2H20(g)+ SnO2(S)+4HCI(g) Various interphase coatings have been applied in oxide fibre/oxide matrix composites including, SnO2, TiO2, The microstructures of as-received fibres heat treated ZrO2, HfO2, monazite, magnetoplumbite, perouskite at 500, 600 and 900C for 90 minutes and SnO2 coated structures like BaTiO3, etc. [26-33]. In the present work, fibres were characterised using SEM and XRD. SEM the effect of fibre and coating properties on the bending was used to determine uniformity, morphology and strength and toughness of alumina fibre/glass matrix thickness of the coating. The fracture surfaces of the as- composites have been investigated SnO2 was chosen as received and SnO2 coated fibres were also characterised a coating since it has no reaction with alumina up to by SEM 1400 oC in a partial pressure of oxygen >10-7atm Single fibre tensile tests were carried out on as- [34, 35]. The strength of alumina-zirconia fibres as- received, heat treated and SnO, coated PRD-166 fibres received, heat treated and SnO2 coated were deter- A random selection of single fibres was made from the mined. Glass matrix composites fabricated using slurry impregnation technique and reinforced with two ypes of fibres, namely PrD-166(alumina-20 wt zir- conia) fibres and relatively smooth saphikon fibres were tested to investigate how the properties of the fibres can nfluence the bending strength and toughness of CMCs Experimental procedure The PRD-166 fiber used in the present work is a polycrystalline a-Al2O3 fiber, 20 um in diameter and containing 15-20 wt %Y2O3 partially stabilized zirco- nia particles. The properties of PRD-166 fiber are given in Table 1 [36]. The zirconia particles are dispersed throughout the fiber but primarily along the grain boundaries. Saphikon is a single crystal alumina fila- ment. The c-axis of the filament is oriented parallel to the fibre surface. The mechanical and physical proper ies of the saphikon filaments are given in Table 1 [37] N 5lA, a borosilicate glass, obtained from Owens Illinois Inc, was used as a matrix in the present study. Table Room-temperature properties of PRD-166 fiber and Saphikon fila- ment [ 36,37 Fibre Melting Density Tensile Tensile Thermal ao point(C)(g/cm)strength modulus expansion Mpa)(GPa)(×10-°/°C Saphikon 2053 3931503809.12∥/toc-axis) PRD-1662045 2070380 Fig. I.(a)Chevron notch specimen;(b) geometry of chevron notch
1. Chemical compatibility with the matrix. 2. Refractoriness. 3. Stability in oxidising, water vapor and corrosive environments. 4. Providing a relatively weak fibre-coating interphase allowing for fibre debonding and pullout. In oxide fibre/oxide matrix composite systems, reactions at the fibre/matrix interface, e.g. Al2O3/SiO2, mullite/mullite leads to a strong interfacial bonding with the consequence of brittle behaviour of the composite. Various interphase coatings have been applied in oxide fibre/oxide matrix composites including, SnO2, TiO2, ZrO2, HfO2, monazite, magnetoplumbite, perouskite structures like BaTiO3, etc. [26–33]. In the present work, the effect of fibre and coating properties on the bending strength and toughness of alumina fibre/glass matrix composites have been investigated. SnO2 was chosen as a coating since it has no reaction with alumina up to 1400 C in a partial pressure of oxygen >107 atm. [34,35]. The strength of alumina-zirconia fibres asreceived, heat treated and SnO2 coated were determined. Glass matrix composites fabricated using a slurry impregnation technique and reinforced with two types of fibres, namely PRD-166 (alumina–20 wt.% zirconia) fibres and relatively smooth saphikon fibres were tested to investigate how the properties of the fibres can influence the bending strength and toughness of CMCs. Experimental procedure The PRD-166 fiber used in the present work is a polycrystalline a-Al2O3 fiber, 20 mm in diameter and containing 1520 wt.% Y2O3 partially stabilized zirconia particles. The properties of PRD-166 fiber are given in Table 1[36]. The zirconia particles are dispersed throughout the fiber but primarily along the grain boundaries. Saphikon is a single crystal alumina filament. The c-axis of the filament is oriented parallel to the fibre surface. The mechanical and physical properties of the saphikon filaments are given in Table 1[37]. N 51A, a borosilicate glass, obtained from Owens Illinois Inc., was used as a matrix in the present study. The alumina (PRD-166) fibres were coated with SnO2 by a chemical vapor deposition technique. The alumina fibre tows were placed in the central hot zone of the reactor and heated to the deposition temperature of 500 C. Dry nitrogen was the carrier gas for SnCl4. The flow rate of nitrogen was 1l/min. A second bubbler contained water heated to 80 C through which oxygen was passed at a rate of 0.6 l/min. The deposition occurred via the chemical reaction [38], SnCl4ðgÞ þ 2H2OðgÞ ! SnO2ðsÞ þ 4HClðgÞ ð1Þ The microstructures of as-received fibres heat treated at 500, 600 and 900 C for 90 minutes and SnO2 coated fibres were characterised using SEM and XRD. SEM was used to determine uniformity, morphology and thickness of the coating. The fracture surfaces of the asreceived and SnO2 coated fibres were also characterised by SEM. Single fibre tensile tests were carried out on asreceived, heat treated and SnO2 coated PRD-166 fibres. A random selection of single fibres was made from the Table 1 Room-temperature properties of PRD-166 fiber and Saphikon filament [36,37] Fibre Melting point (C) Density (g/cm3 ) Tensile strength (Mpa) Tensile modulus (GPa) Thermal expansion (106 / C) Saphikon 2053 3.9 3150 380 9.12 (// to c-axis) 7.95 (to c-axis) PRD-166 2045 4.2 2070 380 9.0 Fig. 1. (a) Chevron notch specimen; (b) geometry of chevron notch. 566 R. Venkatesh / Ceramics International 28 (2002) 565–573
R Venkatesh/Ceramics International 28(2002)565-573 material to be tested. The fibres were center-line moun- coeffecient of variation were then evaluated by Weibull ted on a paper frame. The fibres were centered over the analysis [39] frame and lightly stretched. A small amount of adhesive Alumina fiber reinforced glass matrix composites was then carefully placed at each end of the fibre. The were fabricated by a slurry impregnation technique [40] specimen gage length for all the fibres tested was 17 m The slurry consisted of glass frit, 2-propanol and an An Instron tensile testing machine (model 1 120)was organic binder to impart green strength to the tapes and used with a 5 N load cell. Before fibres were loaded onto facilitate their handling. For fabrication of alumina/ the machine, the diameter of the individual fibres was glass composites, a continuous process was employed to measured with an optical microscope. The frame was make unidirectional tapes. For fabrication of alumina, then gripped in the jaws of the testing machine and the SnO2/glass composites, the coated fibers were dipped in mounting frame was burned on the sides. The fibres the slurry and laid on mylar tapes to form prepeg tape were successively stressed to failure with a crosshead These unidirectional tapes were heated to 500oC in air peed of 0.25 mm/min. An average of 80 fibres in each to remove the binder and then hot pressed. The hot group, i.e. as-received, heat treated at 500, 600 and pressing was performed in a graphite lined die in argon 900oC and Sno, coated were tested. The mean tensile atmosphere at 925C and 3 MPa. strength, Weibull modulus, standard deviation and a Optical microscopy was used to evaluate the volume action and fiber distributions in the composites. The fracture surfaces of the composites were characterized using SEM. Three point bending tests on the glass matrix composites were conducted in the longitudinal direction. Bending tests were carried out on specimens having a span length(S) to thickness(W) ratio >8 and thickness(W) to breadth(B)ratio of 0.75. The three point bending tests were conducted on an Instron machine(model 1102)with a crosshead speed of 0.05 10 slope=B 7 slope= 0.10 Tensile Strength, a;( MPa) Tensile Strength, O:(MPa) Fig 3. Weibull plots (a) as-received PRD-166 fibres;(b)SnO2 coated PRD.166 fibres Table 2 Estimated Weibull parameters of as-received and heat treated alumina (PRD. 166)fibres Standard strength(MPa) deviation(MPa) variation(%) As-received 1375 418 1313 Fig. 2. Microstructure of as-received fibres (a)Rough longitudinal 600°C 1283 urface of the fibres:(b) zirconia particles dispersed throughout the 900°C 29
material to be tested. The fibres were center-line mounted on a paper frame. The fibres were centered over the frame and lightly stretched. A small amount of adhesive was then carefully placed at each end of the fibre. The specimen gage length for all the fibres tested was 17 mm. An Instron tensile testing machine (model 1120) was used with a 5 N load cell. Before fibres were loaded onto the machine, the diameter of the individual fibres was measured with an optical microscope. The frame was then gripped in the jaws of the testing machine and the mounting frame was burned on the sides. The fibres were successively stressed to failure with a crosshead speed of 0.25 mm/min. An average of 80 fibres in each group, i.e. as-received, heat treated at 500, 600 and 900 C and SnO2 coated were tested. The mean tensile strength, Weibull modulus, standard deviation and coeffecient of variation were then evaluated by Weibull analysis [39]. Alumina fiber reinforced glass matrix composites were fabricated by a slurry impregnation technique [40]. The slurry consisted of glass frit, 2-propanol and an organic binder to impart green strength to the tapes and facilitate their handling. For fabrication of alumina/ glass composites, a continuous process was employed to make unidirectional tapes. For fabrication of alumina/ SnO2/glass composites, the coated fibers were dipped in the slurry and laid on mylar tapes to form prepeg tapes. These unidirectional tapes were heated to 500 C in air to remove the binder and then hot pressed. The hot pressing was performed in a graphite lined die in argon atmosphere at 925 C and 3 MPa. Optical microscopy was used to evaluate the volume fraction and fiber distributions in the composites. The fracture surfaces of the composites were characterized using SEM. Three point bending tests on the glass matrix composites were conducted in the longitudinal direction. Bending tests were carried out on specimens having a span length (S) to thickness (W) ratio > 8 and thickness (W) to breadth (B) ratio of 0.75. The three point bending tests were conducted on an Instron machine (model 1102) with a crosshead speed of 0.05 Fig. 2. Microstructure of as-received fibres. (a) Rough longitudinal surface of the fibres; (b) zirconia particles dispersed throughout the fibre. Fig. 3. Weibull plots (a) as-received PRD-166 fibres; (b) SnO2 coated PRD-166 fibres. Table 2 Estimated Weibull parameters of as-received and heat treated alumina (PRD- 166) fibres Fibre Mean tensile strength (MPa) Standard deviation (MPa) Coefficient of variation (%) As-received 1375 418 30 500 C 1313 386 29 600 C 1283 440 34 900 C 1083 320 29 R. Venkatesh / Ceramics International 28 (2002) 565–573 567
R Venkatesh/ Ceramics International 28(2002)565-573 mm/min. Fracture toughness of all the composites was Yc=(5.639+27.440o+1893a, determined using chevron notch specimens as shown in Fig(la). A specimen geometry having a span-to-thick 4342a2+3389a) ness ratio of 4 and thickness-to-width ratio of 1.5 was used to evaluate the fracture toughness. The three point where a=o/w and a, is the initial crack length(dis- bending tests were conducted on an Instron machine tance from line of load application to tip of chevron (model 1102) with a crosshead speed of 0.05 mm/min. notch) as shown in Fig. 1(b) The fracture toughness(Klc) evaluated by the ollowing equation Klc=(P/BW)Yc (2) where P is maximum load, and Ye is a dimensionless stress intensity factor. From a slice model [41] for the specimen geometry used, Yc can be evaluated as [42] M Fig. 4. Fracture surface of an as-received fibre showing processing EstimatedWeibull parameters of as-received and SnO, coated alumina(PRD-166)fibre Fibre Mean tensile Standard Coeffcient of anation 40m m 1375 418 SnO, coated (0.4 um) (b) SnO, coated(0.5 um) oated(0.8 um) Fig. 5. Interface morphology (a) Saphikon/SnO2 and;(b) alumi- a(PRD-166)/SnO2. SnO, coated(10 um) 320 Table Table Amplitude of roughness, A with coating thickness of the fibres Radial (or), circumferential(oo), and axial stresses(o,) at the alumina fibre/SnO2 interphase. Subscript f denotes the fibre and s the coating Thickness of SnO, (Hr hickness f SnO2(um) (MPa) (MPa) (MPa) (MPa) 0.5 3 -485
mm/min. Fracture toughness of all the composites was determined using chevron notch specimens as shown in Fig. (1a). A specimen geometry having a span-to-thickness ratio of 4 and thickness–to-width ratio of 1.5 was used to evaluate the fracture toughness. The three point bending tests were conducted on an Instron machine (model 1102) with a crosshead speed of 0.05 mm/min. The fracture toughness (K1c) was evaluated by the following equation K1c¼ ðP=BW1=2 ÞYc ð2Þ where P is maximum load, and Yc is a dimensionless stress intensity factor. From a slice model [41] for the specimen geometry used, Yc can be evaluated as [42] Yc ¼ ð5:639 þ 27:44oþ18:932 o 43:423 oþ338:94 oÞ ð3Þ where o=ao/W and ao is the initial crack length (distance from line of load application to tip of chevron notch) as shown in Fig. 1(b). Fig. 4. Fracture surface of an as-received fibre showing processing voids. Table 3 Estimated Weibull parameters of as-received and SnO2 coated alumina (PRD-166) fibre Fibre Mean tensile strength (MPa) Standard deviation (MPa) Coeffcient of variation (%) As-received 1375 418 30 SnO2 coated (0.4 mm) 1060 386 25 SnO2 coated (0.5 mm) 966 440 28 SnO2 coated (0.8 mm) 851320 33 SnO2 coated (2.0 mm) 702 440 32 SnO2 coated (10 mm) 166 320 34 Table 4 Radial (sr), circumferential (sy), and axial stresses (z) at the alumina fibre/SnO2 interphase. Subscript f denotes the fibre and s the coating Thickness of SnO2 (mm) srf=sqf=srs (MPa) s (MPa) szs (MPa) sqs (MPa) 0.4 22 45 540 526 0.5 28 55 535 518 0.8 42 88 524 497 2.0 53 97 502 485 Fig. 5. Interface morphology (a) Saphikon/SnO2 and; (b) alumina(PRD-166)/SnO2. Table 5 Amplitude of roughness, A with coating thickness of the fibres Thickness of SnO2 (mm) Amplitude, m (mm) 0.0 0.26 0.4 0.45 0.5 0.53 0.8 0.88 2.0 1.8 10.0 4.0 568 R. Venkatesh / Ceramics International 28 (2002) 565–573
R Venkatesh/Ceramics International 28 (2002)565-573 Results and discussion Weibull plots of as-received and Sno, coated fibres are shown in Fig. 3(a)and(b). The straight line plots PRD-166 fibres. The rough cobblestone surface of the and SnO2 coated follow Weibull distribution. Tabe,& Fig 2(a)and(b) shows the microstructure of alumina indicate that the tensile strength data for the as-rece alumina fibres is shown in Fig. 2(a). The zirconia parti- shows the tensile strength, Weibull modulus(m), scale cles are dispersed throughout the fibres but primarily along grain boundaries [Fig. 2(b)]. The dispersion of 20 wt% zirconia in PRD-166 fibre inhibits grain growth Table 7 and thereby improves strength and ughness of these Bend strength, WOF and fracture toughness of AG and ASG com- fibres [36]. The grain size of alumina, as determined by posites the lineal intercept method, was about 0.5 um and that r(%) Work of of zirconia particles was 0.33 um. XRD showed the zir fracture conia particles to be primarily in tetragonal form (MPa) /m-2) (MPa m/) 110 Table 6 AG Roughness strain A ed with thermal mismatch strain of PRD-166/SnO, and nO, interphase 0.026 0.0013 ASG 0602466 215 770 2.6 120 3.3 190 10 Fig. 6.(a, b and c) Fracture surface of PRD-166 alumina fibre/SnO2/glass matrix composites showing partial debonding and fibre pullout. Note the extremely rough PRD-166 fibre
Results and discussion Fig. 2 (a) and (b) shows the microstructure of alumina PRD-166 fibres. The rough cobblestone surface of the alumina fibres is shown in Fig. 2(a). The zirconia particles are dispersed throughout the fibres but primarily along grain boundaries [Fig. 2(b)]. The dispersion of 20 wt.% zirconia in PRD-166 fibre inhibits grain growth and thereby improves strength and toughness of these fibres [36]. The grain size of alumina, as determined by the lineal intercept method, was about 0.5 mm and that of zirconia particles was 0.33 mm. XRD showed the zirconia particles to be primarily in tetragonal form. Weibull plots of as-received and SnO2 coated fibres are shown in Fig. 3(a) and (b). The straight line plots indicate that the tensile strength data for the as-received and SnO2 coated follow Weibull distribution. Table 2 shows the tensile strength, Weibull modulus (m), scale Table 6 Roughness strain A/r compared with thermal mismatch strain of PRD-166/SnO2 and Saphikon/SnO2 interphase A/r T PRD-166/SnO2 0.026 0.0013 Saphikon/SnO2 0.003 0.001 Fig. 6. (a, b and c) Fracture surface of PRD-166 alumina fibre/SnO2/glass matrix composites showing partial debonding and fibre pullout. Note the extremely rough PRD-166 fibre. Table 7 Bend strength, WOF and fracture toughness of AG and ASG composites Vf (%) Bend strength (MPa) Work of fracture (J/m2 ) Fracture toughness (MPa m1/2) 12 110 220 2.0 AG 20 140 – – 26 205 420 2.3 30 215 – – 42 230 770 2.6 24 120 580 2.8 ASG 36 150 900 3.3 46 190 – – R. Venkatesh / Ceramics International 28 (2002) 565–573 569