《复合材料 Composites》课程教学资源（学习资料）第五章 陶瓷基复合材料_electrophoretic deposition- pressless sintering

Availableonlineatwww.sciencedirect.com Science Direct E噩≈RS ELSEVIER Joumal of the European Ceramic Society 29(2009)1631-1639 www.elsevier.com/locate/jeurceramsoc Development and characterisation of high-density oxide fibre-reinforced oxide ceramic matrix composites with improved mechanical properties C Kaya.*, F. Kaya, E.G. Butler, A.R. Boccaccini d, KKChawla a yildiz Technical University, Faculty of Chemistry and Metallurgy, Department of Metallurgical and Materials Engineering, avutpasa Campus, Esenler, L. Tirkey b Zonguldak Karaelmas University, Engineering Faculty, Department of Metallurgical Engineering, Zonguldak, Turkey The University of Birmingham, IRC in Materials Processing and School of Metallurgy and Materials, Edgbaston, Birmingham B15 2TT, UK d Imperial College London, Department of Materials, Prince Consort Road, London SW7 2BP e University of Alabama at Birmingham, Department of Materials, Birmingham, AL 35294, USA Received I August 2008: received in revised form 23 September 2008; accepted 27 September 2008 Available online 31 December 2008 Abstract A low cost and reliable ceramic matrix composite fabrication route has been developed. It involves the coating of 2D woven ceramic fibres (Nextel M 720)with oxide nano-size ceramic particles by electrophoretic deposition(EPD) followed by impregnation of the coated fibres with ceramic matrix and warm pressing at 180C to produce the"green"component ready for pressureless sintering. The effects of two different weak interface materials, NdPOa and ZrOz, on the thermomechanical properties of the composites are also examined. Damage mechanisms, such a bonding, fibre fracture, delamination and matrix cracking within the composite plates subjected to tensile loading are analysed using acoustic emission technique and correlated with microstructure. It is shown that the composites with NdPO4 interface, 10%o porosity and 40 vol %o fibre loading have superior thermomechanical properties in terms of strength and damage-tolerant behaviour in multilayer plate form. The improved strength damage-tolerant composites. The final components produced are considered to be suitable for use as shroud seals and insulating plates for ombustor chambers in aircraft engines. C 2008 Elsevier Ltd. All rights reserved. Keywords: Ceramic composite; AL2O3; Nextel M 720: Interphase: Electrophoretic deposition 1. Introduction performance:(a) reinforcement fibres with high strength and modulus, (b)a high temperature-resistant ceramic matrix and Ceramic matrix composites( CMCs)can display quasi- (c) an interface that prevents occurrence of any chemical reac ductile deformation behaviour when mechanisms such as crack tion between fibre and matrix. 4,5 The ceramic matrix selected deflection, fibre pull-out, crack bridging and debonding can be should be chemically compatible with the ceramic fibres and made to operate by optimising the strength of the interface the interface should also be compatible with both to improve between fibre and matrix. -Oxide/oxide CMCs are excellent the overall performance. In recent years, oxide-based compos candidate components for high temperature applications due to ites have been the centre of many research activities as CMCs their increased toughness, inherent oxidation resistance, good are required to be compatible with their service atmospheres thermal shock resistance, decreased flaw sensitivity and damage- which are generally oxidising, such as combustion gases. 3.6-11 tolerant behaviour. -When CMCs are designed for high-risk After selecting the ideal matrix, interface and fibres, the most and high-temperature applications, three constituents should important task is to choose a viable and cost-effective process- carefully be selected and controlled for obtaining the required ing technique for the manufacture of the final component. In the present work we have investigated the use of a novel multi step technology that involves electrophoretic deposition(EPD) Corresponding author. Tel: +90 2124491866: fax: +902124491554 coating of woven ceramic fibres followed by impregnation of E-mail address: cngzky ( yahoo co uk(C Kaya) he ceramic matrix onto individual fibre mats and finally warm 0955-2219 front matter@ 2008 Elsevier Ltd. All rights reserved. doi: 10. 1016/j-jeurceramsoc. 2008.09.027

Available online at www.sciencedirect.com Journal of the European Ceramic Society 29 (2009) 1631–1639 Development and characterisation of high-density oxide fibre-reinforced oxide ceramic matrix composites with improved mechanical properties C. Kaya a,∗, F. Kaya b, E.G. Butler c, A.R. Boccaccini d, K.K. Chawla e a Yildiz Technical University, Faculty of Chemistry and Metallurgy, Department of Metallurgical and Materials Engineering, Davutpasa Campus, Esenler, Istanbul, Turkey b Zonguldak Karaelmas University, Engineering Faculty, Department of Metallurgical Engineering, Zonguldak, Turkey c The University of Birmingham, IRC in Materials Processing and School of Metallurgy and Materials, Edgbaston, Bırmingham B15 2TT, UK d Imperial College London, Department of Materials, Prince Consort Road, London SW7 2BP, UK e University of Alabama at Birmingham, Department of Materials, Birmingham, AL 35294, USA Received 1 August 2008; received in revised form 23 September 2008; accepted 27 September 2008 Available online 31 December 2008 Abstract A low cost and reliable ceramic matrix composite fabrication route has been developed. It involves the coating of 2D woven ceramic fibres (NextelTM 720) with oxide nano-size ceramic particles by electrophoretic deposition (EPD) followed by impregnation of the coated fibres with ceramic matrix and warm pressing at 180 ◦C to produce the “green” component ready for pressureless sintering. The effects of two different weak interface materials, NdPO4 and ZrO2, on the thermomechanical properties of the composites are also examined. Damage mechanisms, such as debonding, fibre fracture, delamination and matrix cracking within the composite plates subjected to tensile loading are analysed using acoustic emission technique and correlated with microstructure. It is shown that the composites with NdPO4 interface, 10% porosity and 40 vol.% fibre loading have superior themomechanical properties in terms of strength and damage-tolerant behaviour in multilayer plate form. The improved sinterability and microstructure stability at moderate temperatures ensure both the fibre integrity and load transfer efficiency resulting in high strength damage-tolerant composites. The final components produced are considered to be suitable for use as shroud seals and insulating plates for combustor chambers in aircraft engines. © 2008 Elsevier Ltd. All rights reserved. Keywords: Ceramic composite; AL2O3; NextelTM 720; Interphase; Electrophoretic deposition 1. Introduction Ceramic matrix composites (CMCs) can display quasi￾ductile deformation behaviour when mechanisms such as crack deflection, fibre pull-out, crack bridging and debonding can be made to operate by optimising the strength of the interface between fibre and matrix.1–3 Oxide/oxide CMCs are excellent candidate components for high temperature applications due to their increased toughness, inherent oxidation resistance, good thermal shock resistance, decreased flaw sensitivity and damage￾tolerant behaviour.1–3 When CMCs are designed for high-risk and high-temperature applications, three constituents should carefully be selected and controlled for obtaining the required ∗ Corresponding author. Tel.: +90 2124491866; fax: +90 2124491554. E-mail address: cngzky@yahoo.co.uk (C. Kaya). performance: (a) reinforcement fibres with high strength and modulus, (b) a high temperature-resistant ceramic matrix and (c) an interface that prevents occurrence of any chemical reac￾tion between fibre and matrix.4,5 The ceramic matrix selected should be chemically compatible with the ceramic fibres and the interface should also be compatible with both to improve the overall performance. In recent years, oxide-based compos￾ites have been the centre of many research activities as CMCs are required to be compatible with their service atmospheres which are generally oxidising, such as combustion gases.3,6–11 After selecting the ideal matrix, interface and fibres, the most important task is to choose a viable and cost-effective process￾ing technique for the manufacture of the final component. In the present work we have investigated the use of a novel multi￾step technology that involves electrophoretic deposition (EPD) coating of woven ceramic fibres followed by impregnation of the ceramic matrix onto individual fibre mats and finally warm 0955-2219/$ – see front matter © 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jeurceramsoc.2008.09.027

1632 C. Kaya et al. Joumal of the European Ceramic Sociery 29(2009)1631-1639 pressing of the coated and impregnated fibre mats for consoli ation of the composites in multilayer form. This process leads to the reduction of the overall porosity and hence to improved mechanical properties. Damage mechanisms of the optimised CMCs during tensile strength tests were analysed using in situ acoustic emission technique 2. Experimental work 2. Matrix material 000 High purity a-alumina powder (Tai-Micron, Japan) with an - average particle size of 160 nm and surface area of 14.3 m/g was used as the matrix material. Alumina powder was first dis- persed in distilled water with the additions of dispersant, liquid binder, which is a water based acrylic polymer(Duramax B1014, Chesham Chemicals Ltd, UK), and 0.5 wt. colloidal Y203 Fig. 1. The EPD coating unit connected to the power supply with an average particle size of 10 nm(Nyacol Corp, USA)as sintering aid. The suspension was magnetically stirred for 2h fibre surface. The EPD coating cell shown in Fig. I was used followed by mechanical ball-milling with alumina balls for 8h to coat the fibre mats using an electrode separation of 20 mm and finally ultrasonication for 0.5h. The solid-loading of the and a deposition voltage of 10V d.c. for 3 min. The coating suspension was 85 wt. and the pH was adjusted to be around chamber made of Perspex contains two stainless steel electrodes (30 cm x 20cm) which are connected to a power supply. The coating suspensions(NdPO4 and ZrO2) were used separately. 2. 2. Interface materials Coated fibre mats were then heat treated at 600oC for 0.5h to increase the adhesion between the fibre and the coating layer NdPO4 and Zro were used as interface materials to coat the woven fibre mats using electrophoretic deposition. The tech- 2. 4. Composite processing niques for the synthesis of these powders were reported in detail elsewhere., 8 Briefly, NdPO4 was prepared by the neu- It is very important to obtain ceramic composites with tral reaction of neodymium nitrate with ammonium di-hydrogen the lowest possible porosity to achieve superior mechanical phosphate(ADPH) at room temperature. Equimolar amounts of properties. The approach followed here to improve composite Nd(NO3 )3 and ADPH were dissolved in water to make 0. 25 M densification was to impregnate the electrophoretically coated solutions. The two solutions were mixed by vigorous stirring fibres with nano-sized AlO3 particles in suspension. A device and heating, followed by filtration. The resultant gel filtrate was was developed in which suspension flow onto the fibre mat was en dried and calcined at 1000C for 3 h to yield stoichiometric automatically controlled, as shown in Fig. 2a. Each coated layer NdPO4 monazite powder with an average particle size of 60 nm. of fibres is taken in turn and placed on a clean sheet of glass. A A 15 wt %o aqueous suspension was prepared by ball milling for controlled amount of suspension is then poured onto the fibre 4 h with the pH value adjusted to be 3. Hydrothermal process- mat, where upon a roller is used to ensure an even and thor- ing was used to synthesise Zroz powders. Zirconium acetate ough impregnation. The fibre mat is then carefully turned over precursor was processed in an autoclave at 220C under auto- and the process repeated. Once one layer has been so treated geneous pressure for 2 h. The resultant sol was dried at 110C it was carefully laid on a sheet of filter paper. The next layer for I h and calcined at 600C for 4 h to remove excess acetates of fibres was subjected to the same treatment, and then neatly and obtain zirconia particles(with an average particle size of stacked on top of the first layer in parallel condition. In this way 20 nm). Zirconia particles were re-dispersed in distilled water the desired number of layers(12 layers in the present work to with solid-loading of 15 wt %o and a pH value of 2. obtain 40 vol. fibre loading in the composite)was built up. The green bodies"containing 12 impregnated fibre mats were then 2.3. Fibre coating by EPD stacked in a warm pressing device shown in Fig 2b to compress and consolidate the multilayer structure at 180C for 2 h. To Eight-harness satin woven mullite fibre mats with 1200 sin- ensure that the layers of fibres would not adhere to the supports, gle filaments in a bundle(Nextel 720, 3M, USA) were used aluminium foil was introduced below the layer of filter paper as reinforcement. The fibre mats were pre-treated by desiz- already supporting the fibres. Once the sample was fully con- ing at 500C for 1 h to remove the organic protection layer structed, spacers were placed at either side in order to obtain from the fibre surface. The fibre mats were then immersed in a composite thickness of 3 mm. The warm pressed specimens an ammonia based solution, consisting of an ammonium salt were finally pressureless sintered at 1200C for 2 h in air. The of polymethacrylic acid ( Versical KA21, pH: 9, Allied colloids, sintered composite plates were then cut to fabricate specimens UK)in order to create a strong negative surface charge on the for mechanical testing and microstructural observations. Fig 3

1632 C. Kaya et al. / Journal of the European Ceramic Society 29 (2009) 1631–1639 pressing of the coated and impregnated fibre mats for consoli￾dation of the composites in multilayer form. This process leads to the reduction of the overall porosity and hence to improved mechanical properties. Damage mechanisms of the optimised CMCs during tensile strength tests were analysed using in situ acoustic emission technique. 2. Experimental work 2.1. Matrix material High purity -alumina powder (Tai-Micron, Japan) with an average particle size of 160 nm and surface area of 14.3 m2/g was used as the matrix material. Alumina powder was first dis￾persed in distilled water with the additions of dispersant, liquid binder, which is a water based acrylic polymer (Duramax B1014, Chesham Chemicals Ltd., UK), and 0.5 wt.% colloidal Y2O3 with an average particle size of 10 nm (Nyacol Corp., USA) as sintering aid. The suspension was magnetically stirred for 2 h followed by mechanical ball-milling with alumina balls for 8 h and finally ultrasonication for 0.5 h. The solid-loading of the suspension was 85 wt.% and the pH was adjusted to be around 4. 2.2. Interface materials NdPO4 and ZrO2 were used as interface materials to coat the woven fibre mats using electrophoretic deposition. The tech￾niques for the synthesis of these powders were reported in detail elsewhere.2,8 Briefly, NdPO4 was prepared by the neu￾tral reaction of neodymium nitrate with ammonium di-hydrogen phosphate (ADPH) at room temperature. Equimolar amounts of Nd(NO3)3 and ADPH were dissolved in water to make 0.25 M solutions. The two solutions were mixed by vigorous stirring and heating, followed by filtration. The resultant gel filtrate was then dried and calcined at 1000 ◦C for 3 h to yield stoichiometric NdPO4 monazite powder with an average particle size of 60 nm. A 15 wt.% aqueous suspension was prepared by ball milling for 4 h with the pH value adjusted to be 3. Hydrothermal process￾ing was used to synthesise ZrO2 powders. Zirconium acetate precursor was processed in an autoclave at 220 ◦C under auto￾geneous pressure for 2 h. The resultant sol was dried at 110 ◦C for 1 h and calcined at 600 ◦C for 4 h to remove excess acetates and obtain zirconia particles (with an average particle size of 20 nm). Zirconia particles were re-dispersed in distilled water with solid-loading of 15 wt.% and a pH value of 2. 2.3. Fibre coating by EPD Eight-harness satin woven mullite fibre mats with 1200 sin￾gle filaments in a bundle (NextelTM 720, 3M, USA) were used as reinforcement. The fibre mats were pre-treated by desiz￾ing at 500 ◦C for 1 h to remove the organic protection layer from the fibre surface. The fibre mats were then immersed in an ammonia based solution, consisting of an ammonium salt of polymethacrylic acid (Versical KA21, pH:9, Allied colloids, UK) in order to create a strong negative surface charge on the Fig. 1. The EPD coating unit connected to the power supply. fibre surface. The EPD coating cell shown in Fig. 1 was used to coat the fibre mats using an electrode separation of 20 mm and a deposition voltage of 10 V d.c. for 3 min. The coating chamber made of Perspex contains two stainless steel electrodes (30 cm × 20 cm) which are connected to a power supply. The coating suspensions (NdPO4 and ZrO2) were used separately. Coated fibre mats were then heat treated at 600 ◦C for 0.5 h to increase the adhesion between the fibre and the coating layer. 2.4. Composite processing It is very important to obtain ceramic composites with the lowest possible porosity to achieve superior mechanical properties. The approach followed here to improve composite densification was to impregnate the electrophoretically coated fibres with nano-sized Al2O3 particles in suspension. A device was developed in which suspension flow onto the fibre mat was automatically controlled, as shown in Fig. 2a. Each coated layer of fibres is taken in turn and placed on a clean sheet of glass. A controlled amount of suspension is then poured onto the fibre mat, where upon a roller is used to ensure an even and thor￾ough impregnation. The fibre mat is then carefully turned over and the process repeated. Once one layer has been so treated, it was carefully laid on a sheet of filter paper. The next layer of fibres was subjected to the same treatment, and then neatly stacked on top of the first layer in parallel condition. In this way the desired number of layers (12 layers in the present work to obtain 40 vol.% fibre loading in the composite) was built up. The “green bodies” containing 12 impregnated fibre mats were then stacked in a warm pressing device shown in Fig. 2b to compress and consolidate the multilayer structure at 180 ◦C for 2 h. To ensure that the layers of fibres would not adhere to the supports, aluminium foil was introduced below the layer of filter paper already supporting the fibres. Once the sample was fully con￾structed, spacers were placed at either side in order to obtain a composite thickness of 3 mm. The warm pressed specimens were finally pressureless sintered at 1200 ◦C for 2 h in air. The sintered composite plates were then cut to fabricate specimens for mechanical testing and microstructural observations. Fig. 3

C. Kaya et al. Journal of the European Ceramic Society 29(2009)1631-1639 shows the manufacturing steps of the fibre-reinforced compos- Matrix suspension ites and Fig. 4 shows the final sintered composite plate and its microstructure Fiber 2.5. Mechanical testing Tensile and flexural strength test bars (50 mm x 15 mm x 3 mm and 100mmx x3 mm in dimensions for flexu ral and tensile test bars, respectively) were cut from sintered composite panels using an Accutom 5 high-speed, precision dia Filter Paper mond saw, and then both surfaces were ground to parallel using Top hot plate a 40 um diamond resin bonded disc. All sharp edges were bev- elled using 6 and 3 um diamond pastes on soft cloths in order to Spacer minimise possible flaw sources. The tensile surfaces were then Bottom hot plate polished to a 0. 25 um finish using diamond paste. A typical ten- sile test specimen is shown in Fig 5a. To enable samples to be Aluminium Foil assembled accurately, a jig was designed and manufactured as EPD coated fiber shown in Fig 5b. It is important that the samples be fixed and aligned correctly as the load must be applied along the centre- Schematic representation of (a) impregnation process and (b)wa line, otherwise shear forces will distort the results(see Fig 5c) g for consolidation of the multilayer composite plates The jig uses two steel pins to align the plates, with a small macralon block to centre the sample. Spacers are also required to keep the steel plates apart the correct distance. Room-and high-temperature four-point bend tests(test bars with dimen- sions 50 mm x 15 mm x 3 mm) were performed on an Instron Testing machine fitted with a furnace which had tungsten mesh Colloidal Y203 elements, which enabled the tests to be carried out at temper a-AlyO3 powder Liquid binder atures up to 1500C. The pushrods and fixtures are all made from a tungsten-zirconium-molybdenum alloy Specimens to I Distilled water be tested at 1300C were held at the test temperature for at least I h prior to the testing to allow the system to equilibrate and to ensure that the specimen was at this temperature. Tensile ultrasonic agitation Desizing the fibers tests were performed at room temperature. For flexural and ten- sile tests, a constant cross-head speed of 0.5 mm/min. was used The interlaminar shear strength of the coated and impregnate yers was also determined using the test specimen geome- Matrix me Coating of individual high solids-loading try shown in Fig. 6. In order to measure the bonding strength fiber mat by EPD between fibre mats, individually coated and impregnated mats were sintered together using exactly the same cycle used for the composite itself and a tensile test specimen, as shown in Fig. 6, was prepared. The Perspex plates used for alignment were strongly glued(Epoxy adhesive, Good Fellow, UK) to the fibre and the aluminium tap surfaces using adhesives that provide coated and impregnated fibres a very strong adhesion so that only the bonding strength could measured during the tensile test. For all mechanical prop- Formation of erty results reported, seven samples were used multilayer green value by eliminating the highest and the lowest values(to obtain CMC plates more reliable data and eliminate big variations and taking the fiv Pressureless sIntering 2.6. Acoustic emission monitorIng The location of sensors and the experimental design for Composite plate acoustic emission monitoring are shown in Fig. 7. Acoustic (AE)is Fig3. Flow chart showing the manufacturing of woven mullite fibre-reinforced on elastic waves generated by the rapid release of localised alumina matrix composites. sources like transient relaxation of stress and strain fields plas-

C. Kaya et al. / Journal of the European Ceramic Society 29 (2009) 1631–1639 1633 Fig. 2. Schematic representation of (a) impregnation process and (b) warm pressing for consolidation of the multilayer composite plates. Fig. 3. Flow chart showing the manufacturing of woven mullite fibre-reinforced alumina matrix composites. shows the manufacturing steps of the fibre-reinforced compos￾ites and Fig. 4 shows the final sintered composite plate and its microstructure. 2.5. Mechanical testing Tensile and flexural strength test bars (50 mm × 15 mm × 3 mm and 100 mm × 10 mm × 3 mm in dimensions for flexu￾ral and tensile test bars, respectively) were cut from sintered composite panels using an Accutom 5 high-speed, precision dia￾mond saw, and then both surfaces were ground to parallel using a 40m diamond resin bonded disc. All sharp edges were bev￾elled using 6 and 3m diamond pastes on soft cloths in order to minimise possible flaw sources. The tensile surfaces were then polished to a 0.25 m finish using diamond paste. A typical ten￾sile test specimen is shown in Fig. 5a. To enable samples to be assembled accurately, a jig was designed and manufactured as shown in Fig. 5b. It is important that the samples be fixed and aligned correctly as the load must be applied along the centre￾line, otherwise shear forces will distort the results (see Fig. 5c). The jig uses two steel pins to align the plates, with a small macralon block to centre the sample. Spacers are also required to keep the steel plates apart the correct distance. Room- and high-temperature four-point bend tests (test bars with dimen￾sions 50 mm × 15 mm × 3 mm) were performed on an Instron Testing machine fitted with a furnace which had tungsten mesh elements, which enabled the tests to be carried out at temper￾atures up to 1500 ◦C. The pushrods and fixtures are all made from a tungsten–zirconium–molybdenum alloy. Specimens to be tested at 1300 ◦C were held at the test temperature for at least 1 h prior to the testing to allow the system to equilibrate and to ensure that the specimen was at this temperature. Tensile tests were performed at room temperature. For flexural and ten￾sile tests, a constant cross-head speed of 0.5 mm/min. was used. The interlaminar shear strength of the coated and impregnated layers was also determined using the test specimen geome￾try shown in Fig. 6. In order to measure the bonding strength between fibre mats, individually coated and impregnated mats were sintered together using exactly the same cycle used for the composite itself and a tensile test specimen, as shown in Fig. 6, was prepared. The Perspex plates used for alignment were strongly glued (Epoxy adhesive, Good Fellow, UK) to the fibre and the aluminium tap surfaces using adhesives that provide a very strong adhesion so that only the bonding strength could be measured during the tensile test. For all mechanical prop￾erty results reported, seven samples were used for each reported value by eliminating the highest and the lowest values (to obtain more reliable data and eliminate big variations) and taking the average value of the remaining five values. 2.6. Acoustic emission monitoring The location of sensors and the experimental design for acoustic emission monitoring are shown in Fig. 7. Acoustic emission (AE) is a non-destructive technique (NDT) based on elastic waves generated by the rapid release of localised sources like transient relaxation of stress and strain fields. Plas-

C. Kaya et al. Joumal of the European Ceramic Sociery 29(2009)1631-1639 Fig 4.(a) An optical microscopy photo of the woven mullite fibre(Nextel 720)reinforced alumina matrix composite with NdPO4 interface and(b) SEM image of the same plate indicating the detailed microstructure. tic deformation, crack initiation and growth as well as fracture while burst type AE signals are characterised by high amplitude events result in AE signals. Depending on the nature of energy pulses. Continuous type signals are caused by plastic defor- release, two types of AE response are observed; continuous and mation, diffusion controlled phase transformations, while burst burst. Continuous AE signals are characterised by low amplitude type signals are caused by diffusionless phase transformations, crack initiation and propagation. However it is difficult to use a single all-encompassing parameter to describe an experimental result uniquely. Some of the parameters used for identifica tion of source are; amplitude, energy level, duration and rise time. 3 AE has been used extensively in composite materials characterisation in order to identify the possible mechanism of fracture.4, I5 It is reported that AE sources such as matrix crack ing, fibre/matrix interfacial separation and fibre failure could be evaluated simultaneously in both metallic and ceramic matrix composites. 4-19 The elastic wave generated by the propaga tion of a crack through the solid to the surface can be recorded by one or more sensors. The sensor is a transducer that con verts the mechanical wave into an electric signal. In this way information about the existence and location of a crack can be obtained. A variety of damage mechanisms including matrix cracking, fibre debonding and fibre fracture which are common events when a ceramic composite is subjected to loading, can be identified%As damage progresses, different forms of AE sig- Fig. 5. Images of (a)tensile test specimen, (b) jig used for alignment and (c) nals are detected depending on the type of fracture mechanism aligned test specimen. acting

1634 C. Kaya et al. / Journal of the European Ceramic Society 29 (2009) 1631–1639 Fig. 4. (a) An optical microscopy photo of the woven mullite fibre (Nextel 720TM) reinforced alumina matrix composite with NdPO4 interface and (b) SEM image of the same plate indicating the detailed microstructure. tic deformation, crack initiation and growth as well as fracture events result in AE signals. Depending on the nature of energy release, two types of AE response are observed; continuous and burst. Continuous AE signals are characterised by low amplitude Fig. 5. Images of (a) tensile test specimen, (b) jig used for alignment and (c) aligned test specimen. while burst type AE signals are characterised by high amplitude pulses. Continuous type signals are caused by plastic defor￾mation, diffusion controlled phase transformations, while burst type signals are caused by diffusionless phase transformations, crack initiation and propagation. However it is difficult to use a single all-encompassing parameter to describe an experimental result uniquely. Some of the parameters used for identifica￾tion of source are; amplitude, energy level, duration and rise time.13 AE has been used extensively in composite materials characterisation in order to identify the possible mechanism of fracture.14,15 It is reported that AE sources such as matrix crack￾ing, fibre/matrix interfacial separation and fibre failure could be evaluated simultaneously in both metallic and ceramic matrix composites.14–19 The elastic wave generated by the propaga￾tion of a crack through the solid to the surface can be recorded by one or more sensors. The sensor is a transducer that con￾verts the mechanical wave into an electric signal. In this way information about the existence and location of a crack can be obtained. A variety of damage mechanisms including matrix cracking, fibre debonding and fibre fracture, which are common events when a ceramic composite is subjected to loading, can be identified19 As damage progresses, different forms of AE sig￾nals are detected depending on the type of fracture mechanism acting

C. Kaya et al. /Journal of the European Ceramic Society 29(2009)1631-1639 1635 (FEG) SEM(FX-4000, Jeol Ltd, Japan). More detailed obser vations were made using transmission electron microscopy TEM, Jeol Ltd, Japan, 4000 FX TEM) operated at 400kV, and equipped with an energy dispersive X-ray(EDX) analysis unit. Porosity and pore size were measured using a mercury pl porosimeter(Hg Por, Micromechanics Instrument Corp, USA) using a penetrometer weight of 62.79 g, and penetration volume of 6. 188 mL. The Archimedes' technique was used for density Aluminium measurements 3. Results and discussion Interface An optical microscopy image of the composite plates pro- duced after pressureless sintering at 1200C for 2 h is shown in Fig. 4a indicating that a homogeneous and macro-flaw free Coated and surface structure of the plate was obtained. The final component impregnated shown in Fig. 4a contains 12 layers with a fibre volume fraction fiber layer of 40%0. The overall thickness of the plate was x3 mm and no delamination of the layers was observed after sintering. The final sintered microstructure of the composite shown in Fig 4b con- tains 10-14 vol %o porosity which is quite low for this type of sinterable matrix precursors which are silca comp.. omposite systems. The development of complex and hi ing nano-scale alumina powders dispersed in aqueous solution mixed with nano-size colloidal Y2O3, led to full infiltration by the impregnation technique, as shown in Fig. 4b. It can be con- cluded that the matrix precursors developed display favourable rheological properties during both the impregnation and warm Fig.6. Schematic representation of the test sample arrangement to measure the pressing stages resulting in enhanced matrix infiltration and interlaminar shear strength(bonding strength) between two fibre mats. formability. In our previous research,using a combination of EPD and pressure filtration for different composite systems, for example mullite/mullite CMCs, a successful deposition was 2.7 Microstructural characterisation also achieved leading to manufacture of sintered samples with Microstructural examinations were carried out on sintered porosity levels from 15 to 20%. The present work represents an advance in the development of CMCs with lower porosity composite samples using a high-resolution Field Emission Gun using an impregnation technique combined with warm press- ing. It is expected from these results that improved sinterability and stability at moderate temperatures will ensure continued fibre integrity and matrix load translation efficiency resulting in high strength, damage-tolerant CMCs. Selected properties of the composite plates produced are given in Table 1. As shown in this table, using two different interface materials, namely NdPO4 and Zroz, quite similar mechanical properties were obtained at room temperature(four-point bending strength values of 279 and 260 MPa and tensile strength of 142 and 136 MPa, respec AE Pre amp tively). However, when the samples were subjected to bend (regain: 40 dB) strength test at 1300C, only a 5% reduction in strength was recorded for composites with NdPO4 interfaces(from 279 to 266 MPa)whereas over a 10% decrease was measured for com- Test posites with ZrO2 interfaces(from 260 to 232 MPa). Although sPecimen these two interface materials are known to be compatible with the matrix and fibres used, the difference in strength is most probably related to the possible different microstructure of the interface at high temperature. Load-displacement curves of the composite samples with the two interfaces were also recorded 中 nonitoring shown in Fis. 8. As Shown in Fig. 8a. both composites exhib

C. Kaya et al. / Journal of the European Ceramic Society 29 (2009) 1631–1639 1635 Fig. 6. Schematic representation of the test sample arrangement to measure the interlaminar shear strength (bonding strength) between two fibre mats. 2.7. Microstructural characterisation Microstructural examinations were carried out on sintered composite samples using a high-resolution Field Emission Gun Fig. 7. Schematic representation of experimental set up for acoustic monitoring during tensile test. (FEG) SEM (FX-4000, Jeol Ltd., Japan). More detailed obser￾vations were made using transmission electron microscopy (TEM, Jeol Ltd., Japan, 4000 FX TEM) operated at 400 kV, and equipped with an energy dispersive X-ray (EDX) analysis unit. Porosity and pore size were measured using a mercury porosimeter (Hg Por, Micromechanics Instrument Corp., USA) using a penetrometer weight of 62.79 g, and penetration volume of 6.188 mL. The Archimedes’ technique was used for density measurements. 3. Results and discussion An optical microscopy image of the composite plates pro￾duced after pressureless sintering at 1200 ◦C for 2 h is shown in Fig. 4a indicating that a homogeneous and macro-flaw free surface structure of the plate was obtained. The final component shown in Fig. 4a contains 12 layers with a fibre volume fraction of 40%. The overall thickness of the plate was ∼3 mm and no delamination of the layers was observed after sintering. The final sintered microstructure of the composite shown in Fig. 4b con￾tains 10–14 vol.% porosity which is quite low for this type of composite systems. The development of complex and highly sinterable matrix precursors which are silica free, compris￾ing nano-scale alumina powders dispersed in aqueous solution mixed with nano-size colloidal Y2O3, led to full infiltration by the impregnation technique, as shown in Fig. 4b. It can be con￾cluded that the matrix precursors developed display favourable rheological properties during both the impregnation and warm pressing stages resulting in enhanced matrix infiltration and formability. In our previous research2,7 using a combination of EPD and pressure filtration for different composite systems, for example mullite/mullite CMCs, a successful deposition was also achieved leading to manufacture of sintered samples with porosity levels from 15 to 20%. The present work represents an advance in the development of CMCs with lower porosity using an impregnation technique combined with warm press￾ing. It is expected from these results that improved sinterability and stability at moderate temperatures will ensure continued fibre integrity and matrix load translation efficiency resulting in high strength, damage-tolerant CMCs. Selected properties of the composite plates produced are given in Table 1. As shown in this table, using two different interface materials, namely NdPO4 and ZrO2, quite similar mechanical properties were obtained at room temperature (four-point bending strength values of 279 and 260 MPa and tensile strength of 142 and 136 MPa, respec￾tively). However, when the samples were subjected to bend strength test at 1300 ◦C, only a 5% reduction in strength was recorded for composites with NdPO4 interfaces (from 279 to 266 MPa) whereas over a 10% decrease was measured for com￾posites with ZrO2 interfaces (from 260 to 232 MPa). Although these two interface materials are known to be compatible with the matrix and fibres used, the difference in strength is most probably related to the possible different microstructure of the interface at high temperature. Load–displacement curves of the composite samples with the two interfaces were also recorded for comparison at room and high temperatures (1300 ◦C), as shown in Fig. 8. As shown in Fig. 8a, both composites exhibit

C. Kaya et al. Joumal of the European Ceramic Sociery 29(2009)1631-1639 Table I The effects of interface materials on the properties of ceramic composite tested Interface material Four-point bending strength, MPa Tensile strength, (RT) MPa Interlaminar shear strength, MPa average pore size, nm ZrO? 266(RT,232(1300°C) a damage-tolerant behaviour at room temperature and compos- detail transmission electron microscopy observations were car- ite failure started with fibre failure followed by multiple matrix ried out. It is seen in Fig 9 that there is no reaction zone between cracking(the first and second peaks as arrowed in the graphs) NdPO4 and alumina matrix, the interface region is clean as evi- and eventually delamination and fibre pull-out took place. Sim- dent by the absence of reaction products, which was confirmed ilar results were observed on the samples tested at 1300C, as also by TEM EDX. It is also seen from the TEM micrograph in shown in Fig. 8b, confirming that the composite plates produced Fig 9 that the grain size of the NdPOa interface is about 200 nm in the present work showed damage-tolerant behaviour at both which should act as a dense barrier preventing oxygen diffusion room temperature and at 1300C. The greater loss in strength from the matrix through the fibres at high temperature in composites with zirconia interface was Table 1 also shows that tensile strength values of the com- attributed to the lower oxidation protection at high temperature posite plates with zirconia and NdPO4 interphases are 136 and offered by zirconia in comparison to NdPO4, which does not 142 MPa, respectively. It is also shown in Table I that interlami- prevent oxygen diffusion through the interface layer. For exam- nar shear strength values are 8.6 and 9. 4 MPa for the coated and ple, the density of the coating layer may play a role at high impregnated fibre samples with zirconia and NdPO4 interphase temperatures. Further experiments are now underway to clarify respectively. These results represent a significant improvement this behaviour and determine the possible reasons for it. It is of the present woven fibre-reinforced CMCs, which exhibit bet important to avoid the occurrence of any reaction between the ter mechanical properties in terms of room temperature tensile coating layer and the reinforcement fibres in order to obtain ideal strength and interfacial bonding strength than previously devel damage-tolerant behaviour by activation of crack deflection and oped composites. The tensile strength and interlaminar shear fibre debonding/pull-out mechanisms. To observe the interface strength are both improved by the addition of liquid binders region of a composite sample containing NdPO4 interface in Fiber failure Matrix Nextel 720M fiber pull-out 00 Displacement mm ZrO NdPO4 interfa 100 nm Displacement mm Fig 8. Load-displacement curves of composites with zirconia and NdPOa inter- containing NdPOa interface indicating that there is no reaction between the aces:(a) at room temperature and (b) at 1300C interface and the fibre

1636 C. Kaya et al. / Journal of the European Ceramic Society 29 (2009) 1631–1639 Table 1 The effects of interface materials on the properties of ceramic composite tested. Interface material Four-point bending strength, MPa Tensile strength, (RT) MPa Interlaminar shear strength, MPa Average pore size, nm NdPO4 279 (RT), 266 (1300 ◦C) 142 9.4 80 ZrO2 266 (RT), 232 (1300 ◦C) 136 8.6 70 a damage-tolerant behaviour at room temperature and compos￾ite failure started with fibre failure followed by multiple matrix cracking (the first and second peaks as arrowed in the graphs) and eventually delamination and fibre pull-out took place. Sim￾ilar results were observed on the samples tested at 1300 ◦C, as shown in Fig. 8b, confirming that the composite plates produced in the present work showed damage-tolerant behaviour at both room temperature and at 1300 ◦C. The greater loss in strength at high temperature in composites with zirconia interface was attributed to the lower oxidation protection at high temperature offered by zirconia in comparison to NdPO4, which does not prevent oxygen diffusion through the interface layer. For exam￾ple, the density of the coating layer may play a role at high temperatures. Further experiments are now underway to clarify this behaviour and determine the possible reasons for it. It is important to avoid the occurrence of any reaction between the coating layer and the reinforcement fibres in order to obtain ideal damage-tolerant behaviour by activation of crack deflection and fibre debonding/pull-out mechanisms. To observe the interface region of a composite sample containing NdPO4 interface in Fig. 8. Load–displacement curves of composites with zirconia and NdPO4 inter￾faces: (a) at room temperature and (b) at 1300 ◦C. detail transmission electron microscopy observations were car￾ried out. It is seen in Fig. 9 that there is no reaction zone between NdPO4 and alumina matrix, the interface region is clean as evi￾dent by the absence of reaction products, which was confirmed also by TEM EDX. It is also seen from the TEM micrograph in Fig. 9 that the grain size of the NdPO4 interface is about 200 nm which should act as a dense barrier preventing oxygen diffusion from the matrix through the fibres. Table 1 also shows that tensile strength values of the com￾posite plates with zirconia and NdPO4 interphases are 136 and 142 MPa, respectively. It is also shown in Table 1 that interlami￾nar shear strength values are 8.6 and 9.4 MPa for the coated and impregnated fibre samples with zirconia and NdPO4 interphase, respectively. These results represent a significant improvement of the present woven fibre-reinforced CMCs, which exhibit bet￾ter mechanical properties in terms of room temperature tensile strength and interfacial bonding strength than previously devel￾oped composites.2 The tensile strength and interlaminar shear strength are both improved by the addition of liquid binders Fig. 9. Transmission electron microscopy (TEM) image of the composite sample containing NdPO4 interface indicating that there is no reaction between the interface and the fibre

C. Kaya et al. Journal of the European Ceramic Society 29(2009)1631-1639 10000 Absolute Energy, (Jx10-6) Fig. 12. AE events of the composite subjected to tensile test showing the rela- tionship between amplitude and absolute energy Fig 10. FEG SEM micrograph of a composite specimen with NdPO4 interface tested under tensile loading. to be higher than the energy generated by the ceramic matrix. A closer examination of AE events shows that these high energy and nano-size colloidal yttria that have promoted the sintering aE events also have amplitude higher than 70 dB, as shown in kinetics of the matrix resulting in higher density values Fig. 12. Both energy and amplitude level of the AE signal indi- In our previous studies on CMCs with different oxide cate how much energy is released during the fracture process, matrices,8 the mechanical properties of the final components hence pointing to high strength-high stiffness materials such as have been linked to microstructural variations without consid- ceramic fibres. In total fourteen events have been detected with a ering in detail the possible effects of matrix pore size on the combination of high energy-amplitude and duration values(see properties. In the present work, the average pore size was mea- Figs. 11 and 12). Two of these events significantly differ from sured and found to be smaller than 100 nm for composites with the others as they have 100 times higher energy and duration val- two different interphases, as shown in Table 1. This fine poros- ues with the highest amplitude(78 dB)recorded during the test ity is related to the presence of ultrafine yttria particles(10nm) These two AE events correspond to final failure. Other 12 AE within the matrix that act as sintering additives and provide a events with a combination of high energy-duration and ampli more homogeneous matrix structure in terms of density and pore tude levels could be interpreted as the fracture of 12 fibre mats size distribution within the composite A fractured surface of a composite sample subjected to ten- Further analysis of high amplitude(70 dB) events shows sile test at room temperature is shown in Fig. 10 indicating the that there is a group of AE events with high amplitude and low presence of extensive fibre pull-out due to mechanisms, such as fibre/matrix debonding and crack deflection both initiated by the values of these ae events range between 20 and 60J 10- ature of the NdPO4 weak interface with event duration between 300 and 600 us. These groups Fig. 11 shows a detailed examination of AE parameters of of events could be deduced to be individual fibre fractures energy versus event duration. a distinct tail at energy val- within the fibre mats. The ultimate tensile strength of these ues higher than 100J x 10-8 and event duration higher than fibres could have been decreased(compared to the average 1000 us can be seen clearly. The AE events giving rise to the tail tensile strength) due to damage occurred during processing could be interpreted as fibre fractures because the fractureenergy Other AE events are categorised as matrix related events that generated by high strength and high modulus fibres is expected possibly arise from matrix cracking but also from interfacial When all AE activity is plotted against the total time of test (i.e, the extension of sample), it can be clearly seen that fibre mats fail towards the end of the tensile loading whereas matrix related events occur throughout the test, as shown in Fig. 13 This is a clear indication of damage-tolerant behaviour of the ceramic composites. In this composite system, the fibre mat has been treated to obtain a weak interfacial strength between the ceramic fibre and matrix. AE response shows that while the matrix crack propagates fibre mats stay intact in the crack wake leading to fibre bridging which leads to increased frac ture toughness. Although some individual fibre failures were observed, these did not compromise the overall integrity of the Absolute Energy, (Jx10") composite. It was also seen that when one fibre mat failed due to Fig. 11. AE events of the composite subjected to tensile test showing the rela- increased load transfer from the fractured ceramic matrix,other tionship between duration and absolute energy mats also failed simultaneously within a fraction of a second

C. Kaya et al. / Journal of the European Ceramic Society 29 (2009) 1631–1639 1637 Fig. 10. FEG SEM micrograph of a composite specimen with NdPO4 interface tested under tensile loading. and nano-size colloidal yttria that have promoted the sintering kinetics of the matrix resulting in higher density values. In our previous studies on CMCs with different oxide matrices,2,8 the mechanical properties of the final components have been linked to microstructural variations without consid￾ering in detail the possible effects of matrix pore size on the properties. In the present work, the average pore size was mea￾sured and found to be smaller than 100 nm for composites with two different interphases, as shown in Table 1. This fine poros￾ity is related to the presence of ultrafine yttria particles (10 nm) within the matrix that act as sintering additives and provide a more homogeneous matrix structure in terms of density and pore size distribution. A fractured surface of a composite sample subjected to ten￾sile test at room temperature is shown in Fig. 10 indicating the presence of extensive fibre pull-out due to mechanisms, such as fibre/matrix debonding and crack deflection both initiated by the nature of the NdPO4 weak interface. Fig. 11 shows a detailed examination of AE parameters of energy versus event duration. A distinct tail at energy val￾ues higher than 100 J × 10−18 and event duration higher than 1000s can be seen clearly. The AE events giving rise to the tail could be interpreted as fibre fractures because the fracture energy generated by high strength and high modulus fibres is expected Fig. 11. AE events of the composite subjected to tensile test showing the rela￾tionship between duration and absolute energy. Fig. 12. AE events of the composite subjected to tensile test showing the rela￾tionship between amplitude and absolute energy. to be higher than the energy generated by the ceramic matrix. A closer examination of AE events shows that these high energy AE events also have amplitude higher than 70 dB, as shown in Fig. 12. Both energy and amplitude level of the AE signal indi￾cate how much energy is released during the fracture process, hence pointing to high strength-high stiffness materials such as ceramic fibres. In total fourteen events have been detected with a combination of high energy-amplitude and duration values (see Figs. 11 and 12). Two of these events significantly differ from the others as they have 100 times higher energy and duration val￾ues with the highest amplitude (78 dB) recorded during the test. These two AE events correspond to final failure. Other 12 AE events with a combination of high energy-duration and ampli￾tude levels could be interpreted as the fracture of 12 fibre mats within the composite. Further analysis of high amplitude (≥70 dB) events shows that there is a group of AE events with high amplitude and low energy and duration values as shown in Fig. 12. The energy values of these AE events range between 20 and 60 J × 10−18 with event duration between 300 and 600 s. These groups of events could be deduced to be individual fibre fractures within the fibre mats. The ultimate tensile strength of these fibres could have been decreased (compared to the average tensile strength) due to damage occurred during processing. Other AE events are categorised as matrix related events that possibly arise from matrix cracking but also from interfacial failure. When all AE activity is plotted against the total time of test (i.e., the extension of sample), it can be clearly seen that fibre mats fail towards the end of the tensile loading, whereas matrix related events occur throughout the test, as shown in Fig. 13. This is a clear indication of damage-tolerant behaviour of the ceramic composites. In this composite system, the fibre mat has been treated to obtain a weak interfacial strength between the ceramic fibre and matrix. AE response shows that while the matrix crack propagates fibre mats stay intact in the crack wake leading to fibre bridging, which leads to increased frac￾ture toughness. Although some individual fibre failures were observed, these did not compromise the overall integrity of the composite. It was also seen that when one fibre mat failed due to increased load transfer from the fractured ceramic matrix, other mats also failed simultaneously within a fraction of a second.

1638 C. Kaya et al. Joumal of the European Ceramic Sociery 29(2009)1631-1639 100000 temperatures(1100.C). 2 The innovative approach to matrix formulation adopted in the present work results in a sinter- active alumina with the addition of ultra-fine particles to promote 10000 sintering kinetics and avoiding the use of other detrimental sintering aids like silica. Electrophoretic deposition coating of 2D fibre mats with nano-sized particles was successfully applied and coated samples were fully impregnated with alu- 1000 mina matrix using low temperature warm pressing leading to formation of damage-tolerant CMCs with superior properties. The effectiveness of this new approach is evidenced in the outstanding properties achieved(flexural strengths of 279 and 266 MPa at room temperature and at 1300C, respectively, in four-point bending test and a room temperature tensile stre Fig. 13. AE events of the composite subjected to tensile test showing the rela of 142 MPa for the composite samples containing 40 vol. fibre tionship between duration and extension loading, NdPO4 interface and an overall porosity of 10%0) In addition, as a significant improvement, interlaminar shear strength of the multilayer composites was also improved to over Further proof of fibre bridging can be seen in the fractogra- 9 MPa compared to the value of 4 MPa obtained in our previous phy where exposed fibre ends indicate interfacial debonding studies. 2 Currently, thermomechanical properties in terms ind crack front deviation due to weak interfacial strength(see creep and fatigue behaviour as well as ballistic impact and frac ture toughness are under investigation. Based on the properties As a conclusion from the results presented Figs. 11-13, AE obtained, these composites could be candidate for use as shroud is confirmed to be a powerful NDT technique for ceramic com- seals or insulating layers for combustor chambers in aircraft posite characterisation providing that it is used in conjunction engines with other evaluation techniques, such as fractography, in order to draw a more complete picture of the relevant fracture mech nisms. Different fracture mechanisms such as matrix fracture 4. Conclusions and fibre failure could be identified using AE event parameters of energy, amplitude and duration. The identification of failure Nextel 720 M fibre-reinforced alumina matrix composites mechanisms such as interfacial separation could be rather diffi- with two different interface materials(NdPO4 and ZrO2)were cult as the low energy released by such a weak source could be manufactured using electrophoretic deposition, colloidal infil- easily masked by the ae burst caused by fracturing fibres. ae tration and subsequent warm pressing at 180C to consolidate results indicate that the fracture of these oxide matrix composites a green body in multilayer form. After pressureless sintering started with matrix crack initiation and propagation. Although at 1200C for 2 h, the final composite had 40 vol. fibre and a few individual fibre fractures were detected the majority of 10% porosity with an average pore size smaller than 100 nm. fibres did not fail until the final failure. Particularly, the fibre mats The samples with dense NdPO4 coating layer showed damage did not fail until the peak load was reached indicating the fibre tolerant fracture behaviour at room temperature and at 1300C bundles within the fibre mats stayed intact in the crack wake. during flexural strength test. Interlaminar shear strength of the This behaviour is attributed to the weak interfacial strength CMC samples was determined to be over 9 MPa with a ten- and fibre bridging, which in turn increases the toughness of the sile strength of 142 MPa. AE results indicated that fracture of composite. the composite started with matrix crack initiation and propaga Overall, the present investigation has shown the devel- tion Although a few individual fibre fractures were detect opment of an optimised and reproducible CMC fabrication majority of fibres did not fail until the final composite route through understanding the relationship between matrix Specifically, the fibre mats did not fail until the peak load was precursor chemistry, process parameters and final mechanical attained indicating that the fibre bundles within the fibre mats properties. The key to attain optimum mechanical properties stayed intact in the crack wake. This is due to the low interfacial in Nextel 720M fibre-reinforced alumina ceramic matrix com- strength and results in fibre bridging, which should increase the posites with NdPO4 interface is the development of a strong toughness of the composite and dense matrix capable of efficient load transfer between rein- forcement fibres in conjunction with retention of fibre properties after processing. The requirement for high integrity matrices Acknowledgements an lead to fibre degradation due to high sintering tempera ures necessary to achieve high densification overcoming back TUBITAK (Turkish Science and Technological Research stresses imposed by the fibres In the past, the addition of sil- Council), EU Commission and YildIz Technical University ica to alumina matrices as sintering aid has resulted in loss of Istanbul are acknowledged for financial support. Richard of high temperature mechanical properties, in particular time Brown is acknowledged for some parts of the experimental dependant properties (creep and fatigue), even at moderate work

1638 C. Kaya et al. / Journal of the European Ceramic Society 29 (2009) 1631–1639 Fig. 13. AE events of the composite subjected to tensile test showing the rela￾tionship between duration and extension. Further proof of fibre bridging can be seen in the fractogra￾phy where exposed fibre ends indicate interfacial debonding and crack front deviation due to weak interfacial strength (see Fig. 10). As a conclusion from the results presented Figs. 11–13, AE is confirmed to be a powerful NDT technique for ceramic com￾posite characterisation providing that it is used in conjunction with other evaluation techniques, such as fractography, in order to draw a more complete picture of the relevant fracture mech￾anisms. Different fracture mechanisms such as matrix fracture and fibre failure could be identified using AE event parameters of energy, amplitude and duration. The identification of failure mechanisms such as interfacial separation could be rather diffi- cult as the low energy released by such a weak source could be easily masked by the AE burst caused by fracturing fibres. AE results indicate that the fracture of these oxide matrix composites started with matrix crack initiation and propagation. Although a few individual fibre fractures were detected the majority of fibres did not fail until the final failure. Particularly, the fibre mats did not fail until the peak load was reached indicating the fibre bundles within the fibre mats stayed intact in the crack wake. This behaviour is attributed to the weak interfacial strength and fibre bridging, which in turn increases the toughness of the composite. Overall, the present investigation has shown the devel￾opment of an optimised and reproducible CMC fabrication route through understanding the relationship between matrix precursor chemistry, process parameters and final mechanical properties. The key to attain optimum mechanical properties in Nextel 720TM fibre-reinforced alumina ceramic matrix com￾posites with NdPO4 interface is the development of a strong and dense matrix capable of efficient load transfer between rein￾forcement fibres in conjunction with retention of fibre properties after processing. The requirement for high integrity matrices can lead to fibre degradation due to high sintering tempera￾tures necessary to achieve high densification overcoming back stresses imposed by the fibres. In the past, the addition of sil￾ica to alumina matrices as sintering aid has resulted in loss of high temperature mechanical properties, in particular time dependant properties (creep and fatigue), even at moderate temperatures (1100 ◦C).12 The innovative approach to matrix formulation adopted in the present work results in a sinter￾active alumina with the addition of ultra-fine particles to promote sintering kinetics and avoiding the use of other detrimental sintering aids like silica. Electrophoretic deposition coating of 2D fibre mats with nano-sized particles was successfully applied and coated samples were fully impregnated with alu￾mina matrix using low temperature warm pressing leading to formation of damage-tolerant CMCs with superior properties. The effectiveness of this new approach is evidenced in the outstanding properties achieved (flexural strengths of 279 and 266 MPa at room temperature and at 1300 ◦C, respectively, in four-point bending test and a room temperature tensile strength of 142 MPa for the composite samples containing 40 vol.% fibre loading, NdPO4 interface and an overall porosity of 10%). In addition, as a significant improvement, interlaminar shear strength of the multilayer composites was also improved to over 9 MPa compared to the value of 4 MPa obtained in our previous studies.12 Currently, thermomechanical properties in terms of creep and fatigue behaviour as well as ballistic impact and frac￾ture toughness are under investigation. Based on the properties obtained, these composites could be candidate for use as shroud seals or insulating layers for combustor chambers in aircraft engines. 4. Conclusions Nextel 720TM fibre-reinforced alumina matrix composites with two different interface materials (NdPO4 and ZrO2) were manufactured using electrophoretic deposition, colloidal infil￾tration and subsequent warm pressing at 180 ◦C to consolidate a green body in multilayer form. After pressureless sintering at 1200 ◦C for 2 h, the final composite had 40 vol.% fibre and 10% porosity with an average pore size smaller than 100 nm. The samples with dense NdPO4 coating layer showed damage￾tolerant fracture behaviour at room temperature and at 1300 ◦C during flexural strength test. Interlaminar shear strength of the CMC samples was determined to be over 9 MPa with a ten￾sile strength of 142 MPa. AE results indicated that fracture of the composite started with matrix crack initiation and propaga￾tion. Although a few individual fibre fractures were detected the majority of fibres did not fail until the final composite failure. Specifically, the fibre mats did not fail until the peak load was attained indicating that the fibre bundles within the fibre mats stayed intact in the crack wake. This is due to the low interfacial strength and results in fibre bridging, which should increase the toughness of the composite. Acknowledgements TUBITAK (Turkish Science and Technological Research Council), EU Commission and Yıldız Technical University of Istanbul are acknowledged for financial support. Richard Brown is acknowledged for some parts of the experimental work

C. Kaya et al. /Journal of the European Ceramic Society 29(2009)1631-1639 References fiber orientation at 1200 degrees C. Comp. Sci. TechnoL, 2008, 6 1588-1595 1.Chawla,K.K, Ceramic Matrix Composites(2nd ed. ) Kluwer Academic 11. Boccaccini, A. R, Kaya, C. and Chawla, K. K Use of electrophoretic Press, Boston, USA. 2003. deposition in the processing of fibre reinforced ceramic and glass matrix 2. Kaya, C, Butler, E. G, Selcuk, A Boccaccini, A.R. and Lewis, M.H., Mul- mposites.A review Compos. Part A, 2001, 32, 997-1006 lite(NextelTM 720) fibre-reinforced mullite matrix composites exhibiting 12. Kaya, C and Butler, E.G. IRC Intemal Report on Development of CMC. favourable thermomechanical properties. J. Eur Ceram. Soc., 2002, 22(13), The University of Birmingham, February 2002 2333-2342. 13. Raj, B, Jayakumar, T and Thavasimuthu, M, Practical Nondestructive 3. Chawla, K. K, Coffin, C and Xu, Z.R. Interface engineering in oxide Testing. Woodhead Publishing, 2002, Pp 100-105 fibre/oxide matrix composites. Int Mater Rev., 2000, 45, 165-189 14. Kaya, C, Kaya, F. and Mori, H, Non-destructive damage evaluation of 4. Lewis, M.H., Tye, E, Butler, E. G and Doleman, P. A, Oxide CMCs: cyclic-fatigued alumina fiber-reinforced mullite ceramic terphase synthesis and novel fibre development. J. Eur. Ceram. So using forced resonance and acoustic emission techniques. J. Mater. Sci. Lett., 20,639644 2002.21 5. Warren, R and Deng, S, Continuous fibre reinforced ceramic composites 15. Kaya, C and Mori. H. Damage as nt of alumina fibre. for very high temperatures. Silicates Ind, 1996, 5(6), 96-107. reinforced mullite ceramic matrix composites subjected to cyclic fatigue 6. Kanka, B and Schneider, H, Aluminosilicate fiber/mullite matrix compos- at ambient and elevated temperatures. J. Eur Ceram. Soc., 2002, 22(4). ites with favorable high-temperature properties. J. Eur. Ceram. Soc., 2000 447-452. 20,619-623 16. Kaya, F, Damage assessment of oxide fibre reinforced oxide ceram 7. Kaya, C, Gu, X, Al-Dawery, I and Butler, E. G, Microstructural develop- matrix composites using acoustic emission. Ceram. Inter, 2007, 33, ment of woven mullite fibre-reinforced mullite ceramic matrix composites 279-284 by infiltration processing. Sci. TechnoL. Adv. Mater., 2002, 3(1), 35-44. 17. Kaya, F and Bowen, P, Effect of increased interfacial strength on the fatigue 8. Kaya, C, He, J. Y, Gu, X and Butler, E. G, Nanostructured ceramic pow crack growth resistance and crack opening displacement of B Ti 21S/SCS 6 lers by hydrothermal synthesis and their applications. Micropor. Mesopor. composites. Mater. Sci. Eng. A, 2008, 476, 301-307. Mater,2002,54(1-2),37-49 18. Kaya, F, Bowen, P and Liu, J, In-situ observation of crack opening dis- 9. Stoll, E, Mahr, P, Kruger, H G. Kern, H, Thomas, B J C and boccacci placement(COD)in a metastable B titanium composites. J Mater. So A.R. Fabrication technologies for oxide-oxide nic matrx compos- 2008,43,270-280 ites based on electrophoretic deposition. J. Eur: Ceram Soc., 2006, 26(9), 19. Boccaccini, A.R, Kern, H. and Dlouhy, I, Determining the fracture resistance of fibre reinforced glass matrix composites by means of the Chevron-Notched Flexural Technique. Mater. Sci. Eng. A, 2001, 308 ior of Nextel(TM) 720Valumina ceramic composite with +/-45 degre l11-117

C. Kaya et al. / Journal of the European Ceramic Society 29 (2009) 1631–1639 1639 References 1. Chawla, K. K., Ceramic Matrix Composites (2nd ed.). Kluwer Academic Press, Boston, USA, 2003. 2. Kaya, C., Butler, E. G., Selcuk, A., Boccaccini, A. R. and Lewis, M. H., Mul￾lite (NextelTM 720) fibre-reinforced mullite matrix composites exhibiting favourable thermomechanical properties. J. Eur. Ceram. Soc., 2002, 22(13), 2333–2342. 3. Chawla, K. K., Coffin, C. and Xu, Z. R., Interface engineering in oxide fibre/oxide matrix composites. Int. Mater. Rev., 2000, 45, 165–189. 4. Lewis, M. H., Tye, E., Butler, E. G. and Doleman, P. A., Oxide CMCs: interphase synthesis and novel fibre development. J. Eur. Ceram. Soc., 2000, 20, 639644. 5. Warren, R. and Deng, S., Continuous fibre reinforced ceramic composites for very high temperatures. Silicates Ind., 1996, 5(6), 96–107. 6. Kanka, B. and Schneider, H., Aluminosilicate fiber/mullite matrix compos￾ites with favorable high-temperature properties. J. Eur. Ceram. Soc., 2000, 20, 619–623. 7. Kaya, C., Gu, X., Al-Dawery, I. and Butler, E. G., Microstructural develop￾ment of woven mullite fibre-reinforced mullite ceramic matrix composites by infiltration processing. Sci. Technol. Adv. Mater., 2002, 3(1), 35–44. 8. Kaya, C., He, J. Y., Gu, X. and Butler, E. G., Nanostructured ceramic pow￾ders by hydrothermal synthesis and their applications. Micropor. Mesopor. Mater., 2002, 54(1–2), 37–49. 9. Stoll, E., Mahr, P., Kruger, H. G., Kern, H., Thomas, B. J. C. and Boccaccini, A. R., Fabrication technologies for oxide-oxide ceramic matrix compos￾ites based on electrophoretic deposition. J. Eur. Ceram. Soc., 2006, 26(9), 1567–1576. 10. Ruggles-Wrenn, M. B., Siegert, G. T. and Baek, S. S., Creep behav￾ior of Nextel(TM) 720/alumina ceramic composite with +/− 45 degrees fiber orientation at 1200 degrees C. Comp. Sci. Technol., 2008, 6, 1588–1595. 11. Boccaccini, A. R., Kaya, C. and Chawla, K. K., Use of electrophoretic deposition in the processing of fibre reinforced ceramic and glass matrix composites. A review. Compos. Part A, 2001, 32, 997–1006. 12. Kaya, C. and Butler, E. G., IRC Internal Report on Development of CMC. The University of Birmingham, February 2002. 13. Raj, B., Jayakumar, T. and Thavasimuthu, M., Practical Nondestructive Testing. Woodhead Publishing, 2002, pp. 100–105. 14. Kaya, C., Kaya, F. and Mori, H., Non-destructive damage evaluation of cyclic-fatigued alumina fiber-reinforced mullite ceramic matrix composites using forced resonance and acoustic emission techniques. J. Mater. Sci. Lett., 2002, 21, 1333–1335. 15. Kaya, C., Kaya, F. and Mori, H., Damage assessment of alumina fibre￾reinforced mullite ceramic matrix composites subjected to cyclic fatigue at ambient and elevated temperatures. J. Eur. Ceram. Soc., 2002, 22(4), 447–452. 16. Kaya, F., Damage assessment of oxide fibre reinforced oxide ceramic matrix composites using acoustic emission. Ceram. Inter., 2007, 33, 279–284. 17. Kaya, F. and Bowen, P., Effect of increased interfacial strength on the fatigue crack growth resistance and crack opening displacement of  Ti 21S/SCS 6 composites. Mater. Sci. Eng. A., 2008, 476, 301–307. 18. Kaya, F., Bowen, P. and Liu, J., In-situ observation of crack opening dis￾placement (COD) in a metastable  titanium composites. J. Mater. Sci., 2008, 43, 270–280. 19. Boccaccini, A. R., Kern, H. and Dlouhy, I., Determining the fracture resistance of fibre reinforced glass matrix composites by means of the Chevron-Notched Flexural Technique. Mater. Sci. Eng. A, 2001, 308, 111–117