Availableonlineatwww.sciencedirect.com dAS SCIENCE DIREC EEs ELSEVIER Materials Letters 57(2003)2919-2926 www.elsevier.com/locate/matlet Near-net-shape fibre-reinforced ceramic matrix composites by the sol infiltration technique A Deya, M. Chatterjee a, * M.K. Naskar K. Basu Sol-Gel Division, Central Glass and Ceramic Research Institute, 196 Raja S.C. Mullick Road, Jadavpur, Kolkata 700 032, India Regional Research Laboratory, Bhopal 462026, India Received 7 October 2002; accepted 15 October 2002 Abstract The infiltration of a ceramic fibre preform by a ceramic sol is a cost-effective method for the manufacture of ceramic matrix mposites(CMCs). The ceramic sol after calcination forms the matrix and the discontinuous fibres act as the reinforcement ents. In this paper, a technique has been presented in which discontinuous mullite preform with 15 vol of fibre content was infiltrated with the stabilized zirconia-yttria(ZrO210 wt Y203)sol to fabricate near-net-shape CMCs. The effects of sol viscosity, number of infiltration, in situ deposition of carbon in the CMC samples, and calcination(in air and nitrogen atmosphere) temperature on physico-mechanical properties of fabricated CMCs were examined. The characterization of the preform and the developed CMCs were performed by X-ray diffraction (XRD) and scanning electron microscopy (SEM). XRD indicated the presence of both cubic (c)and tetragonal(t)zirconia in the CMCs calcined even at 1400C. The flexural strength of the CMCs (calcined in air at 1400C) when determined by the three-point bend test was found to be almost 14 MPa under a given set of experimental conditions, while calcination of the same materials at 1400C in nitrogen atmosphere(carbon-containing CMCs) exhibited a modulus value of almost 51 GPa. SEM indicated multiple fractures of the matrix, which gave rise to pseudo- ductility This is also evident from the load-elongation curve of the three-point bend test. The reinforced fibres acted as crack arresters which prevented the propagation of the cracks. SEM studies also indicated fibre pull-out in the fracture surface of the CMCs C 2003 Elsevier Science B.V. All rights reserved. Keywords: Sol-gel preparation; Composite materials; Vacuum infiltration; Preforms 1. Introduction of reinforcing inorganic fibres into the ceramic matrix develops CMCs that exhibit pseudo-ductility, prevent- Recently, the development of ceramic fibre-rein- ing catastrophic crack growth by such mechanisms as forced ceramic matrix composites(CMCs) has crack bridging, fibre debonding, and fibre pull-out attracted attention as a promising alternative to mono- [1-7]. The characteristics of the CMCs will depend lithic ceramics for obtaining structural materials suit- upon the shape, the size and dispersion of the rein- able for engineering applications. The incorporation forcement, the microstructure of the ceramic matrix, and the physico-chemical nature of the interface 4 Corresponding author. Tel. +91-33-483-8086; tel/fax: +91 between the fibres and the matrix. The interaction 33-473-0957 between the fibre and the matrix is expected to take E-mailaddress:minati33@hotmail.com(M.Chatterjee). place, and a judicious control of processing parame- 0167-577X/03/S.see front matter e 2003 Elsevier Science B V. All rights reserved doi:10.1016/S0167-577X(02)01397-6
Near-net-shape fibre-reinforced ceramic matrix composites by the sol infiltration technique A. Deya , M. Chatterjeea,*, M.K. Naskar a , K. Basub a Sol –Gel Division, Central Glass and Ceramic Research Institute, 196 Raja S.C. Mullick Road, Jadavpur, Kolkata 700 032, India bRegional Research Laboratory, Bhopal 462 026, India Received 7 October 2002; accepted 15 October 2002 Abstract The infiltration of a ceramic fibre preform by a ceramic sol is a cost-effective method for the manufacture of ceramic matrix composites (CMCs). The ceramic sol after calcination forms the matrix and the discontinuous fibres act as the reinforcement agents. In this paper, a technique has been presented in which discontinuous mullite preform with 15 vol.% of fibre content was infiltrated with the stabilized zirconia – yttria (ZrO210 wt.% Y2O3) sol to fabricate near-net-shape CMCs. The effects of sol viscosity, number of infiltration, in situ deposition of carbon in the CMC samples, and calcination (in air and nitrogen atmosphere) temperature on physico-mechanical properties of fabricated CMCs were examined. The characterization of the preform and the developed CMCs were performed by X-ray diffraction (XRD) and scanning electron microscopy (SEM). XRD indicated the presence of both cubic (c) and tetragonal (t) zirconia in the CMCs calcined even at 1400 jC. The flexural strength of the CMCs (calcined in air at 1400 jC) when determined by the three-point bend test was found to be almost 14 MPa under a given set of experimental conditions, while calcination of the same materials at 1400 jC in nitrogen atmosphere (carbon-containing CMCs) exhibited a modulus value of almost 51 GPa. SEM indicated multiple fractures of the matrix, which gave rise to pseudo-ductility. This is also evident from the load – elongation curve of the three-point bend test. The reinforced fibres acted as crack arresters, which prevented the propagation of the cracks. SEM studies also indicated fibre pull-out in the fracture surface of the CMCs. D 2003 Elsevier Science B.V. All rights reserved. Keywords: Sol – gel preparation; Composite materials; Vacuum infiltration; Preforms 1. Introduction Recently, the development of ceramic fibre-reinforced ceramic matrix composites (CMCs) has attracted attention as a promising alternative to monolithic ceramics for obtaining structural materials suitable for engineering applications. The incorporation of reinforcing inorganic fibres into the ceramic matrix develops CMCs that exhibit pseudo-ductility, preventing catastrophic crack growth by such mechanisms as crack bridging, fibre debonding, and fibre pull-out [1– 7]. The characteristics of the CMCs will depend upon the shape, the size and dispersion of the reinforcement, the microstructure of the ceramic matrix, and the physico-chemical nature of the interface between the fibres and the matrix. The interaction between the fibre and the matrix is expected to take place, and a judicious control of processing parame- 0167-577X/03/$ - see front matter D 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0167-577X(02)01397-6 * Corresponding author. Tel.: +91-33-483-8086; tel./fax: +91- 33-473-0957. E-mail address: minati33@hotmail.com (M. Chatterjee). www.elsevier.com/locate/matlet Materials Letters 57 (2003) 2919 – 2926
A. Dey et al. / Materials Letters 57(2003 )2919-2926 ters is necessary to establish the optimum degree of oxychloride octahydrate(ZrOCl28H2O)and hydrous interfacial bonding [1, 7 yttrium nitrate(both from M/s Indian Rare Earths There are several methods of producing CMCs Mumbai), each with a purity of about 99.9%, were (e.g, hot pressing, melt infiltration, polymer pyroly- used as the starting materials [10, 11]- Is, and sol-gel)[4, 5, 8, 9]. The sol-gel method of Precipitates of hydrated zirconia were obtained by fabricating CMCs has several advantages, such as adding 1: I vol/vol aqueous ammonia solution(25 better homogeneity, low processing and sintering wt % G.R., E. Merck, India) to a solution of zirco- temperatures, and near-net-shape fabrication. How- nium oxychloride octahydrate in deionized water ever, the technique requires multiple infiltration with [10, 11]. The washed(almost electrolyte free) precip intermediate heat treatment to overcome matrix crack- itate was peptized with glacial acetic acid (99.8% ing due to excessive shrinkage of matrix during AnalaR, BDH, India) under stirring at 65+1C. The drying and sintering [1, 4, 8]. Keeping in mind the sol thus obtained had a Zr concentration of 1.2 M. aforementioned points, an attempt has been made in To a known volume of the zirconium acetate sol. a the present investigation towards: (i) the development required amount of yttrium nitrate(10 wt. equivalent of a technique for the fabrication of near-net-shape Y2O3) was mixed under stirring. The pH and the CMCs using discontinuous mullite fibre preforms and viscosity of the resulting yttrium containing zirconia stabilized zirconia-yttria(ZrO2 10 wt %Y2O3) sol as sol (parent sol) were found to be 2.53+ 0.02 and the infiltrant; (ii) the optimization of the process 2.8+ 1 mPa s, respectively. No organics were added parameters for the fabrication of CMCs; and (ii the to the sol as the viscosity controlling agent. From the characterization of the developed CMCs by different parent sol, several sols of viscosities ranging from analytical techniques 3±1to90±1mPas( Table I) were prepared by solvent evaporation. The pH of the sols was measured with a Jencons pH meter(Model 3030) while the 2. Experimenta viscosity values were recorded using a Brookfield viscometer(Model LVTDV-D 2. 1. Preparation of precursor fibre preforms 2.3. Preparation of CMCs by the vacuum infiltration Samples of dimensions 40 x 7 6 mm were cut technique Ir from as-received 15% volume fraction discontinuous mullite fibre preforms of 100 mm diameter and 10 mm In the present investigation, the vacuum infiltration height (from M/s Orient Cerlane, Gujrat) for investi- technique, using the zirconia-yttria sol(Section 2.2) gating and establishing the parameters of sol infiltration of various viscosities(Table 1) as the infiltrant, was technique for the fabrication of near-net-shape CMCs. followed [1, 8, 12]. To carry out the infiltration experi- The samples were activated at 200C (at the heating ments, a laboratory-made set-up was used. Activated rate of 1C/min) with a dwell time of I h to remove precursor preforms of dimensions 40 X7X 6 mm most of volatiles, if any, absorbed on the fibre surface were immersed in the sol for 10 min on the bed of a and the intrafibre regions of the sample preform. The specially designed infiltration unit. The sol was then activated precursor fibre preform samples were pre- removed slowly (3-5 ml/min) by a rotary vacuum of 25%. These samples will function as the reinforcing unit. The infiltrated preform samples were placed in material air at ambient temperature to convert the sol penetrat ing into the preform to the corresponding gel. The 2.2. Preparation of infiltrants (liquid matrix precur- samples were further dried in an air-circulating oven sor) for the fabrication of CMCs at100±2°Cfor4 h and subsequently calcined at 500.800 and 1000C in air under static condition to Zirconia(ZrO2)sols, stabilized with 10 wt % Y2O3, remove the volatiles and decomposable materials were used as the infiltrant (i.e, the liquid matrix Calcining led to the formation of voids and cracking ) For the preparation of this sol, zirconium of matrix due to shrinkage. The above infiltration
ters is necessary to establish the optimum degree of interfacial bonding [1,7]. There are several methods of producing CMCs (e.g., hot pressing, melt infiltration, polymer pyrolysis, and sol –gel) [4,5,8,9]. The sol – gel method of fabricating CMCs has several advantages, such as better homogeneity, low processing and sintering temperatures, and near-net-shape fabrication. However, the technique requires multiple infiltration with intermediate heat treatment to overcome matrix cracking due to excessive shrinkage of matrix during drying and sintering [1,4,8]. Keeping in mind the aforementioned points, an attempt has been made in the present investigation towards: (i) the development of a technique for the fabrication of near-net-shape CMCs using discontinuous mullite fibre preforms and stabilized zirconia – yttria (ZrO210 wt.% Y2O3) sol as the infiltrant; (ii) the optimization of the process parameters for the fabrication of CMCs; and (iii) the characterization of the developed CMCs by different analytical techniques. 2. Experimental 2.1. Preparation of precursor fibre preforms Samples of dimensions 40 7 6 mm were cut from as-received 15% volume fraction discontinuous mullite fibre preforms of 100 mm diameter and 10 mm height (from M/s Orient Cerlane, Gujrat) for investigating and establishing the parameters of sol infiltration technique for the fabrication of near-net-shape CMCs. The samples were activated at 200 jC (at the heating rate of 1 jC/min) with a dwell time of 1 h to remove most of volatiles, if any, absorbed on the fibre surface and the intrafibre regions of the sample preform. The activated precursor fibre preform samples were preserved in a desiccator maintained at a relative humidity of 25%. These samples will function as the reinforcing material. 2.2. Preparation of infiltrants (liquid matrix precursor) for the fabrication of CMCs Zirconia (ZrO2) sols, stabilized with 10 wt.% Y2O3, were used as the infiltrant (i.e., the liquid matrix precursor). For the preparation of this sol, zirconium oxychloride octahydrate (ZrOCl28H2O) and hydrous yttrium nitrate (both from M/s Indian Rare Earths, Mumbai), each with a purity of about 99.9%, were used as the starting materials [10,11]. Precipitates of hydrated zirconia were obtained by adding 1:1 vol/vol aqueous ammonia solution (25 wt.%, G.R., E. Merck, India) to a solution of zirconium oxychloride octahydrate in deionized water [10,11]. The washed (almost electrolyte free) precipitate was peptized with glacial acetic acid (99.8%; AnalaR, BDH, India) under stirring at 65 F 1 jC. The sol thus obtained had a Zr4 + concentration of 1.2 M. To a known volume of the zirconium acetate sol, a required amount of yttrium nitrate (10 wt.% equivalent Y2O3) was mixed under stirring. The pH and the viscosity of the resulting yttrium containing zirconia sol (parent sol) were found to be 2.53 F 0.02 and 2.8 F 1 mPa s, respectively. No organics were added to the sol as the viscosity controlling agent. From the parent sol, several sols of viscosities ranging from 3 F 1 to 90 F 1 mPa s (Table 1) were prepared by solvent evaporation. The pH of the sols was measured with a Jencons pH meter (Model 3030) while the viscosity values were recorded using a Brookfield viscometer (Model LVTDV-II). 2.3. Preparation of CMCs by the vacuum infiltration technique (VIT) In the present investigation, the vacuum infiltration technique, using the zirconia –yttria sol (Section 2.2) of various viscosities (Table 1) as the infiltrant, was followed [1,8,12]. To carry out the infiltration experiments, a laboratory-made set-up was used. Activated, precursor preforms of dimensions 40 7 6 mm were immersed in the sol for 10 min on the bed of a specially designed infiltration unit. The sol was then removed slowly (3 –5 ml/min) by a rotary vacuum pump (Model TSRP/100) attached to the infiltration unit. The infiltrated preform samples were placed in air at ambient temperature to convert the sol penetrating into the preform to the corresponding gel. The samples were further dried in an air-circulating oven at 100 F 2 jC for 4 h and subsequently calcined at 500, 800, and 1000 jC in air under static condition to remove the volatiles and decomposable materials. Calcining led to the formation of voids and cracking of matrix due to shrinkage. The above infiltration 2920 A. Dey et al. / Materials Letters 57 (2003) 2919–2926
A. Dey et al. /Materials Letters 57(2003)2919-2926 921 Table Characteristics of the products obtained under different experimental conditions Sol Number of Intermediate Final Flexural Characteristics vIscosIty infiltration sintering sIntering strength of the products (mPa s) temperature(°C) temperature(°C 30±1 0.97 0.93 1.15 Fragile formation Ing 80+1 1000 ormation of surface coating 1000 5.64 Brittle, good surface (ceramic character) 1149 Brittle, good surface (ceramic character) 60±1(1)+40±1(2)3 88 5.21 Good surface seudo-ductility) 60±1(1)+40±1(4) 1000 Brittle, good surface (ceramic character) 60±1(1)+40±1(4) 1200 Brittle, good surface (ceramic character) 60±1(1)+40±1(4)25 13.95 Brittle, good surface 60±1(1)+40±1(2)3 1000b 6.32 Good surface (pseudo-ductility) 60±1(1)+40±1(2)3 1400 6.21 Good surface Figures in the parentheses indicate the number of infiltration. process was repeated to examine the effect of the (SEM; Model Leo 400c) on samples of dimensions number of infiltrations on the characteristics of the 2 X 1 mm; and(iii) wet chemical analysis CMCs. The final sintering of the infiltrated preforms (b) The infiltrated materials(CMCs) were charac was performed at 1000, 1200, and 1400C in terized by(i)XRD as described in Section 2.4(a)(i) under static condition. All the CMCs were white in above; (ii) SEM in which the fracture surfaces and the colour. For the preparation of carbon-containing top surfaces of samples of the CMCs of same dimen- CMCs, the intermediate heating of the infiltrated sions were examined as mentioned in Section 2. 4(a)(ii) preforms was performed at 500C with a dwell time above; and (iii) the flexural strength and modulus of I h in static air followed by the final calcination at measurement on samples of dimensions 40x7 x 6 1000 and 1400C, each for 1 h in nitrogen (N2) mm using a three-point bend test with the help of the atmosphere with a flow rate of N2 at 1 I min(runs 11 Instron Universal Testing Machine(Model 5500 R) and 12 of Table 1). The CMCs obtained in N2 under a cross-head speed of 0.5 mm/min Each strength atmosphere were black in colour. datum is an average over six samples 2. 4. Characterization of the materials 3. Results and discussion (a) The as-received fibre preforms of tensile strength of about 2.5 GPa and modulus of about 3.1. Characteristics of the precursor preforms 100 GPa were characterised by: (i)X-ray diffraction (XRD; Model Philips PW 1730)using Ni-filtered The XRD of the precursor preforms used in the CuKa radiation; (ii) scanning electron microscopy present investigation indicated the presence of mullite
process was repeated to examine the effect of the number of infiltrations on the characteristics of the CMCs. The final sintering of the infiltrated preforms was performed at 1000, 1200, and 1400 jC in air under static condition. All the CMCs were white in colour. For the preparation of carbon-containing CMCs, the intermediate heating of the infiltrated preforms was performed at 500 jC with a dwell time of 1 h in static air followed by the final calcination at 1000 and 1400 jC, each for 1 h in nitrogen (N2) atmosphere with a flow rate of N2 at 1 l/min (runs 11 and 12 of Table 1). The CMCs obtained in N2 atmosphere were black in colour. 2.4. Characterization of the materials (a) The as-received fibre preforms of tensile strength of about 2.5 GPa and modulus of about 100 GPa were characterised by: (i) X-ray diffraction (XRD; Model Philips PW 1730) using Ni-filtered CuKa radiation; (ii) scanning electron microscopy (SEM; Model Leo 400c) on samples of dimensions 2 2 1 mm; and (iii) wet chemical analysis. (b) The infiltrated materials (CMCs) were characterized by (i) XRD as described in Section 2.4 (a)(i) above; (ii) SEM in which the fracture surfaces and the top surfaces of samples of the CMCs of same dimensions were examined as mentioned in Section 2.4 (a)(ii) above; and (iii) the flexural strength and modulus measurement on samples of dimensions 40 7 6 mm using a three-point bend test with the help of the Instron Universal Testing Machine (Model 5500 R) under a cross-head speed of 0.5 mm/min. Each strength datum is an average over six samples. 3. Results and discussion 3.1. Characteristics of the precursor preforms The XRD of the precursor preforms used in the present investigation indicated the presence of mullite Table 1 Characteristics of the products obtained under different experimental conditions Run Sol viscosity (mPa s) Number of infiltration Intermediate sintering temperature (jC) Final sintering temperature (jC) Flexural strength (MPa) Characteristics of the products 1 30 F 1 2 100 1000 0.97 Fragile 2 60 F 1 3 100 1000 0.93 Fragile 3 85 F 1 2 100 1000 1.15 Fragile, formation of surface coating 4 80 F 1 3 1000 1400 5.59 Formation of surface coating 5 60 F 1 3 1000 1400 5.64 Brittle, good surface (ceramic character) 6 60 F 1 5 800 1400 11.49 Brittle, good surface (ceramic character) 7 60 F 1(1)a + 40 F 1(2)a 3 800 1400 5.21 Good surface (pseudo-ductility) 8 60 F 1(1)a + 40 F 1(4)a 5 800 1000 6.25 Brittle, good surface (ceramic character) 9 60 F 1(1)a + 40 F 1(4)a 5 800 1200 7.31 Brittle, good surface (ceramic character) 10 60 F 1(1)a + 40 F 1(4)a 5 800 1400 13.95 Brittle, good surface (ceramic character) 11 60 F 1(1)a + 40 F 1(2)a 3 500 1000b 6.32 Good surface (pseudo-ductility) 12 60 F 1(1)a + 40 F 1(2)a 3 500 1400b 6.21 Good surface (pseudo-ductility) a Figures in the parentheses indicate the number of infiltration. b N2 atmosphere. A. Dey et al. / Materials Letters 57 (2003) 2919–2926 2921
1. Dey et al. Materials Letters 57(2003)2919-2926 as the only phase. The microstructural feature of the low processing temperature are the main advantages preform in Fig. I indicates that the fibres are circular of using a sol as the liquid infiltrant [1, 3,4, 18] diameter with the diameter distribution in the range 3 Moreover, in the infiltration process, the sol wets 8 um. The shot content in the fibre preform was found and encapsulates the fibres [19, 20]. Keeping in mind to be negligible. The presence of considerable amounts the aforementioned points, CMC preparation using of interfibre pores and voids is evident from the micro- bicomponent sols as the infiltrant has been taken into structure. The infiltration of the sol in the preform is consideration in the present investigation expected to fill these interfibre pores and voids, leading The higher the sol viscosity, the more difficult it is to to the formation of the continuous phase (i.e, the achieve good wetting of the fibres by the sol as well as matrix), which is the primary aim of this investigation. the infiltration of the sol into all the interconnected The preforms were found to contain 54.88 wt. SiO2 pores and voids On the other hand, too low a viscosity and 43.05 wt. Al2O3 as the major constituents, with requires long processing time for achieving high den trace impurities of Fe2O3, TiO2, KO, Na2O, and LOL. sity. Therefore, to examine the effect of sol viscosity on the characteristics of the CMCs. several sols of vis 3. 2. Characteristics of the infiltrant (liquid matrix cosities ranging from 3 I to 90+ l mPa s were used precursor Sols with high viscosity(80+ 1 mPa s and above)were difficult to infiltrate due to the formation of (i)consid- As alumina and alumino-silicate/ZrO2 composites erable amounts of air bubbles during operation and (ii) are known to exhibit good mechanical properties [13- fragmented coatings on the surface of the preform after 16], in the present study, yttria-stabilised zirconia sol air drying, which prevented further penetration of sols was used to act as the matrix material for mullite fibre inside the preform during repeated infiltration [10].On preforms. The Y2O3 additive in the sol helped to the other hand, sols of low viscosity(30+ 1 mPa s) retain the t-ZrO2 polymorph even at 1400c by were also found to be unsuitable because in such cases, inhibiting grain growth. The sols (infiltrant)of various even after nine cycles of infiltrations, the samples viscosities used for the infiltration of the samples of exhibited considerably brittle character. Based on the the preforms are given in Table 1. The characteristics above results, two different sols of viscosities 40+ 1 of the infiltrant should be such that its decomposition and 60+ 1 mPa s were used as the infiltrant in the takes place at a temperature below that at which present investigation sintering occurs, thus preventing interference with the densification process [17. The homogenous mix 3.3. Characteristics of the CMC samples prepared by g of the multicomponent oxides, higher purity, and the vT Table I presents the characteristics of sample CMcs obtained in the present investigation. VIT has been followed for the preparation of CMCs in the present study as VIT aids in the removal of entrapped air in the fibre preform, thereby facilitating the infiltrat ess [1, 7]. Further, the application of vacuum causes an increase in capillary pressure difference across the curved fibre surfaces [7 thus leading to enhanced infiltration of the liquid medium into the interfibre regions, thereby causing an increase in flexural strength of the CMCs(Table 1). To overcome a large shrinkage of matrix materials during drying, leading to extensive 40m matrix cracking and residual fine scale porosity [7, 8]. multiple infiltration cycles followed by intermediate heat treatment was adopted in the present study [1, 4, 8 Fig. 1. SEM of the preform showing interfibre voids and pores. Multiple infiltrations in the green state help to improve
as the only phase. The microstructural feature of the preform in Fig. 1 indicates that the fibres are circular in diameter with the diameter distribution in the range 3 – 8 Am. The shot content in the fibre preform was found to be negligible. The presence of considerable amounts of interfibre pores and voids is evident from the microstructure. The infiltration of the sol in the preform is expected to fill these interfibre pores and voids, leading to the formation of the continuous phase (i.e., the matrix), which is the primary aim of this investigation. The preforms were found to contain 54.88 wt.% SiO2 and 43.05 wt.% Al2O3 as the major constituents, with trace impurities of Fe2O3, TiO2, K2O, Na2O, and LOI. 3.2. Characteristics of the infiltrant (liquid matrix precursor) As alumina and alumino-silicate/ZrO2 composites are known to exhibit good mechanical properties [13 – 16], in the present study, yttria-stabilised zirconia sol was used to act as the matrix material for mullite fibre preforms. The Y2O3 additive in the sol helped to retain the t-ZrO2 polymorph even at 1400 jC by inhibiting grain growth. The sols (infiltrant) of various viscosities used for the infiltration of the samples of the preforms are given in Table 1. The characteristics of the infiltrant should be such that its decomposition takes place at a temperature below that at which sintering occurs, thus preventing interference with the densification process [17]. The homogenous mixing of the multicomponent oxides, higher purity, and low processing temperature are the main advantages of using a sol as the liquid infiltrant [1,3,4,18]. Moreover, in the infiltration process, the sol wets and encapsulates the fibres [19,20]. Keeping in mind the aforementioned points, CMC preparation using bicomponent sols as the infiltrant has been taken into consideration in the present investigation. The higher the sol viscosity, the more difficult it is to achieve good wetting of the fibres by the sol as well as the infiltration of the sol into all the interconnected pores and voids. On the other hand, too low a viscosity requires long processing time for achieving high density. Therefore, to examine the effect of sol viscosity on the characteristics of the CMCs, several sols of viscosities ranging from 3 F 1 to 90 F 1 mPa s were used. Sols with high viscosity (80 F 1 mPa s and above) were difficult to infiltrate due to the formation of (i) considerable amounts of air bubbles during operation and (ii) fragmented coatings on the surface of the preform after air drying, which prevented further penetration of sols inside the preform during repeated infiltration [10]. On the other hand, sols of low viscosity (30 F 1 mPa s) were also found to be unsuitable because in such cases, even after nine cycles of infiltrations, the samples exhibited considerably brittle character. Based on the above results, two different sols of viscosities 40 F 1 and 60 F 1 mPa s were used as the infiltrants in the present investigation. 3.3. Characteristics of the CMC samples prepared by the VIT Table 1 presents the characteristics of sample CMCs obtained in the present investigation. VIT has been followed for the preparation of CMCs in the present study as VIT aids in the removal of entrapped air in the fibre preform, thereby facilitating the infiltration process [1,7]. Further, the application of vacuum causes an increase in capillary pressure difference across the curved fibre surfaces [7], thus leading to enhanced infiltration of the liquid medium into the interfibre regions, thereby causing an increase in flexural strength of the CMCs (Table 1). To overcome a large shrinkage of matrix materials during drying, leading to extensive matrix cracking and residual fine scale porosity [7,8], multiple infiltration cycles followed by intermediate heat treatment was adopted in the present study [1,4,8]. Fig. 1. SEM of the preform showing interfibre voids and pores. Multiple infiltrations in the green state help to improve 2922 A. Dey et al. / Materials Letters 57 (2003) 2919–2926
A. Dey et al. /Materials Letters 57(2003)2919-2926 at 1400C. Hence, the strength of the CMCs has been found to depend, to a great extent, on its final sintering temperature, which affects the degree interaction between fibres and matrix. At too high a temperature, the material becomes brittle because of the high interfacial reaction, while at too low a emperature, the matrix does not sinter adequately [1, 3]. This is supported by the strength values of the CMCs sintered at different temperatures (Table 1) Thus, an optimum sintering temperature is needed. 0000040080120160.200.2 Table 1 further shows that the number of infiltration Displacement(mm) cycles also affects the characteristics of the CMCs Comparing the results of runs 7 and 10, it reveals that Fig. 2. Load-displacement curve of the product of run 7 of Table 1. with the decrease in the number of cycles of infiltration keeping other parameters unaltered, the strength of the the green strength and green machinability and han- infiltrated samples decreases(run 7), but it exhibits dling. Unless the decomposable materials present in the pseudo-ductility. This may be due the development of green body (even after multiple infiltration) are weak interaction between the fibre and the matrix removed during intermediate sintering steps, a high materials. Further, the reinforced fibres acted as crack strength of the CMCs is difficult to attain arresters, which prevented the propagation of cracks. The choice of this intermediate sintering temper This is revealed from the load-elongation curve of the ature, however, depends on the system under consid- three-point bend test presented in Fig. 2. Fig 3 repre- eration. Fourier transform infrared(FTIR) spectra of sents the SEM of the fracture surface of the sample of the sol-gel CaO-doped ZrO2 fibres prepared from the run 7 after a three-point bend test, which indicates fibre zirconium acetate sols and calcined at different temper- pull-out in the sample Multiple fractures of the matrix, atures from 30 to 1000C had confirmed the removal which gave rise to pseudo-ductility, are also evident of almost all the decomposable and carbonaceous from the SEM of Fig 4 XRD of the samples sintered materials at 800C [21] with the formation of white- the 1400C confirmed the presence of both cubic(c) coloured fibres. On the other hand, the fibres calcined and tetragonal(t)ZrO2 along with the mullite, without at 500C were black in colour due to the deposition of any formation of monoclinic polymorphs. Obviously, carbon via the decomposition of acetate groups present in the fibres. The acetate groups in the fibres originated from the corresponding zirconium acetate precursor sols. Based on this result, the intermediate sintering temperatures of 500, 800, and 1000C were selected for multiple infiltration of CMC samples in the present investigation. The final sintering of the samples was Table I indicates that the initial infiltration with a sol of high viscosity(60+ I mPa s), followed by intermedi- ate sintering at 800C and further infiltration with a sol of low viscosity (40+ 1 mPa s), improves the flexural trength of the samples The infiltrated sol. in the interfibre sample preforms, aft formation followed by calcination may form agglomerates of particles at Fig. 3. SEM of the fracture surface of the CMC after a three-point the intermediate sintering temperature of 800C. bend test(product of run 7 of Table 1) showing fibre pull-out from which undergoes densification after further sintering the matrix
the green strength and green machinability and handling. Unless the decomposable materials present in the green body (even after multiple infiltration) are removed during intermediate sintering steps, a high strength of the CMCs is difficult to attain. The choice of this intermediate sintering temperature, however, depends on the system under consideration. Fourier transform infrared (FTIR) spectra of the sol –gel CaO-doped ZrO2 fibres prepared from the zirconium acetate sols and calcined at different temperatures from 30 to 1000 jC had confirmed the removal of almost all the decomposable and carbonaceous materials at 800 jC [21] with the formation of whitecoloured fibres. On the other hand, the fibres calcined at 500 jC were black in colour due to the deposition of carbon via the decomposition of acetate groups present in the fibres. The acetate groups in the fibres originated from the corresponding zirconium acetate precursor sols. Based on this result, the intermediate sintering temperatures of 500, 800, and 1000 jC were selected for multiple infiltration of CMC samples in the present investigation. The final sintering of the samples was performed at temperatures of 1000, 1200, and 1400 jC. Table 1 indicates that the initial infiltration with a sol of high viscosity (60 F 1 mPa s), followed by intermediate sintering at 800 jC and further infiltration with a sol of low viscosity (40 F 1 mPa s), improves the flexural strength of the samples. The infiltrated sol, in the interfibre region of the sample preforms, after gel formation followed by calcination may form agglomerates of particles at the intermediate sintering temperature of 800 jC, which undergoes densification after further sintering at 1400 jC. Hence, the strength of the CMCs has been found to depend, to a great extent, on its final sintering temperature, which affects the degree of interaction between fibres and matrix. At too high a temperature, the material becomes brittle because of the high interfacial reaction, while at too low a temperature, the matrix does not sinter adequately [1,3]. This is supported by the strength values of the CMCs sintered at different temperatures (Table 1). Thus, an optimum sintering temperature is needed. Table 1 further shows that the number of infiltration cycles also affects the characteristics of the CMCs. Comparing the results of runs 7 and 10, it reveals that with the decrease in the number of cycles of infiltration, keeping other parameters unaltered, the strength of the infiltrated samples decreases (run 7), but it exhibits pseudo-ductility. This may be due the development of weak interaction between the fibre and the matrix materials. Further, the reinforced fibres acted as crack arresters, which prevented the propagation of cracks. This is revealed from the load –elongation curve of the three-point bend test presented in Fig. 2. Fig. 3 represents the SEM of the fracture surface of the sample of run 7 after a three-point bend test, which indicates fibre pull-out in the sample. Multiple fractures of the matrix, which gave rise to pseudo-ductility, are also evident from the SEM of Fig. 4. XRD of the samples sintered at the 1400 jC confirmed the presence of both cubic (c) and tetragonal (t) ZrO2 along with the mullite, without any formation of monoclinic polymorphs. Obviously, Fig. 2. Load – displacement curve of the product of run 7 of Table 1. Fig. 3. SEM of the fracture surface of the CMC after a three-point bend test (product of run 7 of Table 1) showing fibre pull-out from the matrix. A. Dey et al. / Materials Letters 57 (2003) 2919–2926 2923
2924 A. Dey et al. Materials Letters 57(2003)2919-2926 20 um 15 um Fig. 4. SEM of the top surface of the CMC (product of run 7 of Fig. 6. SEM of the fracture surface of the CMC after a three-point Table 1)showing multiple fractures of the matrix bend test (product of run 10 of Table 1)showing filling up of the interfibre voids and pores by the infiltration of sol. the Y2O3 additive in the matrix materials helped to retain t-ZrO2 at 1400C by inhibiting grain growth Intermediate heat treatment of the products of runs of run 10 of Table ll and 12 of Table I at 500"C for I h in air resulted [10]. In contrast to run 7, the sample of run 10 of Table 1 fails to exhibit pseudo-ductility. The development of in the formation of black-coloured materials due to the strong interaction between the fibre and the matrix in situ deposition of carbon from the decomposable material may be the reason of such failure. This is acetate groups present in the infiltrated preforms [21] reflected from the load-elongation curve of the three- The black colour of the above materials is retained point bend test presented in Fig. 5. Fig. 6 represents the after final calcination at 1000 and 1400C in N2 SEM of the fracture surface after the three-point bend atmosphere. It is to be noted that the presence of test of the product of run 10 of Table 1. Filling up of carbon in the developed products corresponding to interfibre voids and pores by the infiltration of sol is runs ll and 12 of Table I caused an increase in their clearly discernible from the figure. It is to be noted that lexural strengths in comparison with that obtained for fle all the CMC samples corresponding to runs 1-10 wer run 7(free from carbon). Further, comparing the white in colour (almost free from carbon) after heat fter heat results of the products of run 7(modulus value. 3 treatment in air and exhibited modulus values in the GPa)and run 12(modulus value 51 GPa), it is range 1-3 GPa observed that the presence of carbon in the fibre/ matrix composite material significantly increased the modulus values, thereby increasing pseudo-ductility z50 40 0000040080120160.200.24 0000.020040060080.100.120.14 Displacement(mm) Fig. 5. Load-displacement curve of the product of run 10 of Table I Fig. 7. Load-displacement curve of the product of run 12 of Table 1
the Y2O3 additive in the matrix materials helped to retain t-ZrO2 at 1400 jC by inhibiting grain growth [10]. In contrast to run 7, the sample of run 10 of Table 1 fails to exhibit pseudo-ductility. The development of strong interaction between the fibre and the matrix material may be the reason of such failure. This is reflected from the load –elongation curve of the threepoint bend test presented in Fig. 5. Fig. 6 represents the SEM of the fracture surface after the three-point bend test of the product of run 10 of Table 1. Filling up of interfibre voids and pores by the infiltration of sol is clearly discernible from the figure. It is to be noted that all the CMC samples corresponding to runs 1 – 10 were white in colour (almost free from carbon) after heat treatment in air and exhibited modulus values in the range 1 –3 GPa. Intermediate heat treatment of the products of runs 11 and 12 of Table 1 at 500 jC for 1 h in air resulted in the formation of black-coloured materials due to the in situ deposition of carbon from the decomposable acetate groups present in the infiltrated preforms [21]. The black colour of the above materials is retained after final calcination at 1000 and 1400 jC in N2 atmosphere. It is to be noted that the presence of carbon in the developed products corresponding to runs 11 and 12 of Table 1 caused an increase in their flexural strengths in comparison with that obtained for run 7 (free from carbon). Further, comparing the results of the products of run 7 (modulus value, 3 GPa) and run 12 (modulus value 51 GPa), it is observed that the presence of carbon in the fibre/ matrix composite material significantly increased the modulus values, thereby increasing pseudo-ductility Fig. 5. Load – displacement curve of the product of run 10 of Table 1. Fig. 6. SEM of the fracture surface of the CMC after a three-point bend test (product of run 10 of Table 1) showing filling up of the interfibre voids and pores by the infiltration of sol. Fig. 7. Load – displacement curve of the product of run 12 of Table 1. Fig. 4. SEM of the top surface of the CMC (product of run 7 of Table 1) showing multiple fractures of the matrix. 2924 A. Dey et al. / Materials Letters 57 (2003) 2919–2926
A. Dey et al. /Materials Letters 57(2003)2919-2926 for the development of CMCs with pseudo-ductile characteristics 3. XRD indicates the presence of both c-ZrO2 and t- ZrO2 along with mullite in the CMCs calcined at 1400°C. 4. The pseudo-ductile character developed in CMCs is evident from the load-elongation of the three-point bend 5. SEM indicates fibre pull-out in the fracture surface of the cmcs 20 um 能 Acknowledgements Fig 8. SEM of the fracture surface of the carbon-containing Cmc The authors thank dr h.s. maiti. Director Central after a three-point bend test(product of run 12 of Table 1)showing Glass and Ceramic Research Institute(CG and CRD fibre pull-out from the matrix and Dr N Ramakrishnan, Director, Regional Research Laboratory, Bhopal, for their kind permission to in the materials. This may be explained to be due to publish this paper. The authors sincerely thank Dr. the fact that the in situ deposition of carbon in the K.K. Phani, Head, Composite Division, for providing composite material presumably protected the fibre/ valuable suggestions throughout this work. They also matrix interface from strong interaction [22-25). This thank colleagues of the X-ray, SEM, Composite and is discernible from the load-displacement curve of Analytical Chemistry Sections for their kind help in the three-point bend test in Fig. 7 and the fibre pu materials characterization. The financial assistance out from the SEM of Fig 8 of the same sample. provided by the Aeronautics Research and Develop- Therefore, the bonding at the fibre matrix interface ment Board(AR and DB), Ministry of Defence an be tailored by changing the characteristics of the Govemment of India, is also thankfully acknowledged. infiltrant. number of infiltration. intermediate and final sintering temperatures, and finally via in situ Referen deposition of carbon in the fibre/matrix composite materials by a proper choice of carbon-generating [1.S. Russell-Floyd, B. Harris, R.G. Ine. F.W. decomposable groups present in the corresponding ammett,R W. Jones, T. Wang, Soc. 76 precursor sols 993)2635 2]H.-K. Liu, w.S. Kuo, B.-H. Lin, J. Mater. Sci. 33(1998 3]KK Chawla, Ceramic Matrix Composites, Chapman [4] J.R. Strife, JJ. Brennan, K M. Prewo, Ceram. Eng. Sci. 1. A sol-gel vacuum infiltration technique has been l1(1990)871 developed for the fabrication of near-net-shape [5] Anon, Am. Ceram Soc. Bull. 65(1986)297 CMCs using discontinuous mullite fibre preform [6]J. Wu, F.R. Jones, P.F. James, Advances in Ceramic Matrix with 15 vol% fibre content and ZrO, 10 wt% Composite ll, p. 1 07 x Gu, P.A. Trusty, E.G. Butler, C B. Ponton, J. Eur. Ceram. Y2O3 sol as the infiltrant Soc.20(2000675 2. Effects of sol viscosity, number of infiltration, in [8]S. M.Sim,RJ Kerans, Ceram. Eng. Sci. Proc. 13(1992)632. situ deposition of carbon in the composite 9] W.B. Hilling, J. Am. Ceram Soc. 71(1988)C-9 materials, and calcination temperature on phys [10 P.K. Chakra Chatterjee, M.K. Naskar, B. Sil D anguli, J. Eur. Ceram. Soc. 21(2001)355 ico-mechanical properties of CMCs are examined. [11 M.K.Naskar, DGanguli,J Mater. Sci. 31(1996)6263 Multiple infiltration using sols of viscosities 40+ 1 [12] E.H. Moore, T. Mah, K.A. Keller, Ceram. Eng. Sci. Proc. 15 and 60+ 1 mPa s has been found to be effective 994)113
in the materials. This may be explained to be due to the fact that the in situ deposition of carbon in the composite material presumably protected the fibre/ matrix interface from strong interaction [22 – 25]. This is discernible from the load – displacement curve of the three-point bend test in Fig. 7 and the fibre pullout from the SEM of Fig. 8 of the same sample. Therefore, the bonding at the fibre matrix interface can be tailored by changing the characteristics of the infiltrant, number of infiltration, intermediate and final sintering temperatures, and finally via in situ deposition of carbon in the fibre/matrix composite materials by a proper choice of carbon-generating decomposable groups present in the corresponding precursor sols. 4. Conclusions 1. A sol – gel vacuum infiltration technique has been developed for the fabrication of near-net-shape CMCs using discontinuous mullite fibre preform with 15 vol.% fibre content and ZrO210 wt.% Y2O3 sol as the infiltrant. 2. Effects of sol viscosity, number of infiltration, in situ deposition of carbon in the composite materials, and calcination temperature on physico-mechanical properties of CMCs are examined. Multiple infiltration using sols of viscosities 40 F 1 and 60 F 1 mPa s has been found to be effective for the development of CMCs with pseudo-ductile characteristics. 3. XRD indicates the presence of both c-ZrO2 and tZrO2 along with mullite in the CMCs calcined at 1400 jC. 4. The pseudo-ductile character developed in the CMCs is evident from the load –elongation curve of the three-point bend test. 5. SEM indicates fibre pull-out in the fracture surface of the CMCs. Acknowledgements The authors thank Dr. H.S. Maiti, Director, Central Glass and Ceramic Research Institute (CG and CRI) and Dr. N. Ramakrishnan, Director, Regional Research Laboratory, Bhopal, for their kind permission to publish this paper. The authors sincerely thank Dr. K.K. Phani, Head, Composite Division, for providing valuable suggestions throughout this work. They also thank colleagues of the X-ray, SEM, Composite and Analytical Chemistry Sections for their kind help in materials characterization. The financial assistance provided by the Aeronautics Research and Development Board (AR and DB), Ministry of Defence, Government of India, is also thankfully acknowledged. References [1] R.S. Russell-Floyd, B. Harris, R.G. Cooke, J. Laurie, F.W. Hammett, R.W. Jones, T. Wang, J. Am. Ceram. Soc. 76 (1993) 2635. [2] H.-K. Liu, W.-S. Kuo, B.-H. Lin, J. Mater. Sci. 33 (1998) 2095. [3] K.K. Chawla, Ceramic Matrix Composites, Chapman & Hall, London, 1993. [4] J.R. Strife, J.J. Brennan, K.M. Prewo, Ceram. Eng. Sci. Proc. 11 (1990) 871. [5] Anon., Am. Ceram. Soc. Bull. 65 (1986) 297. [6] J. Wu, F.R. Jones, P.F. James, Advances in Ceramic Matrix Composite II, p. 177. [7] X. Gu, P.A. Trusty, E.G. Butler, C.B. Ponton, J. Eur. Ceram. Soc. 20 (2000) 675. [8] S.-M. Sim, R.J. Kerans, Ceram. Eng. Sci. Proc. 13 (1992) 632. [9] W.B. Hilling, J. Am. Ceram. Soc. 71 (1988) C-96. [10] P.K. Chakrabarty, M. Chatterjee, M.K. Naskar, B. Siladitya, D. Ganguli, J. Eur. Ceram. Soc. 21 (2001) 355. [11] M.K. Naskar, D. Ganguli, J. Mater. Sci. 31 (1996) 6263. [12] E.H. Moore, T. Mah, K.A. Keller, Ceram. Eng. Sci. Proc. 15 (1994) 113. Fig. 8. SEM of the fracture surface of the carbon-containing CMC after a three-point bend test (product of run 12 of Table 1) showing fibre pull-out from the matrix. A. Dey et al. / Materials Letters 57 (2003) 2919–2926 2925
A. Dey et al. / Materials Letters 57(2003 )2919-2926 [13E.A. Puger, P E D. Morgan, J. Am. Ceram. Soc. 69(1986) [19 B. Ben-Nissan, D. Martin, J. Sol-Gel Sci. Technol. 6(1996) [ K. Ishida, K. Hirota, O. Yamaguchi, J. Am. Ceram Soc. 77 [20 M.K. Cinibulk, Ceram. Eng. Sci. Proc. 17(1996)241 (21M. Chatterjee, M.K. Naskar, D. Ganguli, Trans. Indian Ceram. [15] S.C.H. Koh, K.K. Aik, R. McPherson, in: S Somiya, N.Yama- Soc.52(1993)51 moto, H. Yanagida(Eds ) Advances in Ceramics, voL. 24A, [22]CA. Doughan, R L. Lehman, V.A. Greenhut, Ceram. Eng The American Ceramic Society, Westerville, OH, 1986, P. 293. Sci.Proc.101989912. [ P. White, H K. Bowen, B Fegley Jr, in: S Somiya, N. Yama- [23]R W. Rice, J.R. Spann, D. Lewis, w. Coblenz, Ceram. Eng moto, H. Yanagida(Eds ) Advances in Ceramics, vol. 24A, Sci.Proc.5(1984)614. The American Ceramic Society, Westerville, OH, 1986, P. 301. [24K.A Keller, T Mah, T.A. Parthasarathy, C M. Cooke, J Am. [7B.R Marple, D.J. Green, J Am Ceram. Soc. 71(1988)C-471 Ceran.Soc.83(2002)329 [18]H.-K.Liu,J. Mater..Sci.31(1996)5093 [25.-K. Liu, B.-H. Lin. Mater. Lett. 48(2001)230
[13] E.A. Puger, P.E.D. Morgan, J. Am. Ceram. Soc. 69 (1986) C-120. [14] K. Ishida, K. Hirota, O. Yamaguchi, J. Am. Ceram. Soc. 77 (1994) 1391. [15] S.C.H. Koh, K.K. Aik, R. McPherson, in: S. Somiya, N. Yamamoto, H. Yanagida (Eds.), Advances in Ceramics, vol. 24A, The American Ceramic Society, Westerville, OH, 1986, p. 293. [16] P. White, H.K. Bowen, B. Fegley Jr., in: S. Somiya, N. Yamamoto, H. Yanagida (Eds.), Advances in Ceramics, vol. 24A, The American Ceramic Society, Westerville, OH, 1986, p. 301. [17] B.R. Marple, D.J. Green, J. Am. Ceram. Soc. 71 (1988) C-471. [18] H.-K. Liu, J. Mater. Sci. 31 (1996) 5093. [19] B. Ben-Nissan, D. Martin, J. Sol-Gel Sci. Technol. 6 (1996) 187. [20] M.K. Cinibulk, Ceram. Eng. Sci. Proc. 17 (1996) 241. [21] M. Chatterjee, M.K. Naskar, D. Ganguli, Trans. Indian Ceram. Soc. 52 (1993) 51. [22] C.A. Doughan, R.L. Lehman, V.A. Greenhut, Ceram. Eng. Sci. Proc. 10 (1989) 912. [23] R.W. Rice, J.R. Spann, D. Lewis, W. Coblenz, Ceram. Eng. Sci. Proc. 5 (1984) 614. [24] K.A. Keller, T. Mah, T.A. Parthasarathy, C.M. Cooke, J. Am. Ceram. Soc. 83 (2002) 329. [25] H.-K. Liu, B.-H. Lin, Mater. Lett. 48 (2001) 230. 2926 A. Dey et al. / Materials Letters 57 (2003) 2919–2926