《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_Thermal aging behavior of an Sic-fiber reinforced glass matrix composite in a non-oxidizing atmosphere

MATERIAL CHEM ELSEVIER Materials Chemistry and Physics 53(1998)155-164 Thermal aging behavior of an Sic-fiber reinforced glass matrix composite in a non-oxidizing atmosphere A R Boccaccini, J. Janczak-Rusch, I Dlouhy d Institute for Mechanics <tnd Materials, University of California, Sun Diego, La olla. CA 92093-0404, USA EMPA-Thuen, CH-3602 Thun Switzerlan Institute of Physics of Materials, CZ-61662 Brno, Cech Republi eccived 9 September 1997: accepted 24 November 199 Abstract The thermal aging in argon of a commercially available sic- fber reinforced glass matrix composite was investigated at temperatures in the range 500-700'C for exposure duration of up to 1000 h. An inert atmosphere sed to study the effects of temperature alone, thus minimizing and neglecting the effects of oxidation. The mechanical properties of aged samples were evaluated at room-temperature by using four-point flexure strength and three-point flexure chevron-notch techniques. The interfacial properties were determined by push-out inden tation measurem ing an in-situ SEM indentation apparatus. The fracture toughness values determined by the chevron-notch tests were liule affected b ging conditions and were in the range 19-26 MPay'm. The frictional interfacial shear stress was not affected by the aging condition For the most severe aging conditions investigated( 1000 h at 600oC and 100 h at 700 C), a significant loss of flexure strength and stiffness of the samples was detected, which has been ascribed to microstructural changes that occurred in the material during aging as a consequence of the softening of the glass matrix. At these aging conditions, a lower interfacial shear stress for fiber-matrix debonding initiation was measured, which may be explained also by the occurrence of matrix softening and void formation. @1998 Elsevier Science S.A. All rights reserved Keywords. Glass mats ix tuIlipusiles, Aging: Fracture brtluvior: SiC-liber 1. Introduction through matrix microcracks pre-existing in the material as a result of the fabrication process, or through cracks formed The development of faw-tolerant fiber-reinforced glass, under mechanical or thermomechanical thermal shock glass-ceramic and ceramic matrix composites in the last 20 loading. Oxygen penetration from the free- ends of the fibers years has resulted in a new family of structural materials exposed to the environment is also possible and in this case exhibiting quasi-ductile fracture behavior. oxidation resis- degradation of the structural integrity of the material may nce and high temperature capability [1]. A fundamental occur as a result of long-term exposure at high-temperature requisite for these materials to be useful at high temperatures The thermal stability of silicate matrix composites in oxidiz- is the retention of a weak fiber/matrix interface, which is ing environments has been investigated quite extensively in responsible for their naw-Lolerant behavior 1.2]. In silicate the past by conducting thermal aging experiments overa wide matrix composites containing silicon oxycarbide fibers such range of temperatures [9. 11-20]. A common result of inves s Nicalon and Tyranno as reinforcement [3-9]. the weak tigations conducted at temperatures in the range 500-700C interface is provided by a carbon-rich layer around the fibers is that there is a decrease of tensile and flexural strength of which is grown in-situ during composite fabrication [10] the composites as a result of oxidation of the carbon-rich This carbon-rich interlayer can be severely degraded when interfacial layer. Moreover, push-out tests on the fibers of the material is used in oxidizing environments at temperatures aged samples have revealed al increase in the interfacial shear as low as =400C[9.11. 12]. Oxygen penetration can occur stress which has been attributed to the formation of a strong SiOz interfacial bond [ 11. 12, 21, 22]. Thus in the tentioned Technische Universitat Ilmenau, PF 100565, D-98684 Ilmenau, Germany, temperature range(500-700C), a loss of the 'pseudo-duc- Tel :+493677694401: fax:+49 36/7691597: e-mail: aldo, boccaccini(@ tile composite behavior has been frequently observed, with maschinenbau, tu-ilmenau dc reduced fiber pull-out effects [11, 12, 17, 19, 22 0254-0584/98/$19.000 1998 Elsevier Science S.A. All rights reserved PHIS02540584(97102075-0

MATERIALS CHEM;fR\RtdD ELSEVIER Materials Chemistry and Physics 53 ( 1998) 155-164 Thermal aging behavior of an Sic-fiber reinforced glass matrix composite in a non-oxidizing atmosphere A.R. Boccaccini a.*, J. Janczak-Rusch b, I. Dlouhy ’ Abstract The thermal aging in argon of a commercially available Sic-fiber reinforced glass matrix composite was investigated at temperatures in the range 500-700°C for exposure duration of up to 1000 h. An inert atmosphere was used to study the effects of temperature alone, thus minimizing and neglecting the effects of oxidation. The mrchanical properties of aged samples were evaluated at room-temperature by using four-point flexure strength and three-point flexure chevron-notch techniques. The interfacial properties were determined by push-out inden￾tation measurements using an in-situ SEM indentation apparatus. The fracture toughness values determined by the chevron-notch tests were little affected by the aging conditions and were in the range 19-26 MPaJm The frictional interfacial shear stress was not affected by the aging conditions either. For the most bevere aging conditions invest&ated ( 1000 h at 600°C and 100 h at 7OO”C), a significant loss of flexure strength and stiffness of the samples was detected, which has been ascribed to microstructural changes that occurred in the material during aging as a consequence ofthe softening of the glass matrix. At these aging conditions, a lower interfacial shear stress for fiber-matrixdebonding initiation was measured, which may be explained also by the occurrence of matrix softening and void formation. 0 1998 Elsevier Science S.A. All rights reserved E;e~~r~r-d.r: Glass matrix compobites: A:mg: Fracture behavior; Sic-fiber 1. Introduction The development of flaw-tolerant fiber-reinforced glass, glass-ceramic and ceramic matrix composites in the last 20 years has resulted in a new family of structural materials exhibiting quasi-ductile fracture behavior, oxidation resis￾tance and high temperature capability [ I]. A fundamental requisite for these materials to be useful at high temperatures is the retention of a weak fiber/matrix interface. which is responsible for their flaw-tolerant behavior [ 1.21. In silicate matrix composites containing silicon oxycarbide fibers such as Nicalon and Tyranno as reinforcement [ 3-91, the weak interface is provided by a carbon-rich layer around the fibers which is grown in-situ during composite fabrication [IO], This carbon-rich interlayer can be severely degraded when the material is used in oxidizing environments at temperatures as low as = 400°C [ 9,11,12]. Oxygen penetration can occur a * Corresponding author. Praent ad&s: Fachgebirt Wrrkstofftftrchnik, Technische Universitgt Ilmenau. PF 100565. D-98684 Ilmenau. Germany. Tel.: +49 3677 694401: fax: +19 3677 691597; e-mail: aldo.boccaccini@ maschinenb~u..tu-illncnau.de 0154-0584/98/$19.00 ‘C 1998 Elsevier Science S.A. All righta resewed Plls0254-0584(97)02075-0 through matrix microcracks pre-existing in the material as a result of the fabrication process, or through cracks formed under mechanical or thermomechanical (thermal shock) loading. Oxygen penetration from the free-ends of the fibers exposed to the environment is also possible and in this case degradation of the structural integrity of the material may occur as a result of long-term exposure at high-temperature. The thermal stability of silicate matrix composites in oxidiz￾ing environments has been investigated quite extensively in the past by conducting thermal aging experiments over a wide range of temperatures [ 9.1 l-201. A common result of inves￾tigations conducted at temperatures in the range 500-700°C is that there is a decrease of tensile and flexural strength of the composites as a result of oxidation of the carbon-rich interfacial layer. Moreover, push-out tests on the fibers of aged samples have revealed an increase in the interfacial shear stress, which has been attributed to the formation of a strong SiO-, interfacial bond [ 11.12,2 1,221. Thus. in the mentioned temperature range (500-700°C)) a loss of the ‘pseudo-duc￾tile’ composite behavior has been frequently observed, with reduced fiber pull-out effects [ 11,12,17,19,22]

A R. Boccaccini er al. /Materals chemistr and Physics 33 (1998)155-10 Experimental research has been conducted also on the silicate DURAN) glass matrix composite fabricated by mechanical properties of these composites at high tempera- SchoLl Glaswerke(Mainz. Germany ) Information on th tures in inert atmospheres. The first work on this subject was composite constituents is given in Table 1. The composites probably that carried out as early as in 1969 by Crivelli- were prepared by the sol-gel-slurry method. Details on the Visconti and Cooper [23] un C-fiber reinforced silica Tested processing technique are found in the literature [5,6].The at 800C. A more detailed investigation was conducted by samples were received in the form of rectangular test bars of Prewo et al. [24], showing that the ultimate strength of Sic nominal dimensions (4.5 Im X 3. 8 ImIn X 100 IlIm ).The Nicalon) fiber-lithium aluminosilicate matrix composites density of the composites was 2.4 g cm-3and their fiber did not change significantly from the valuc at room-temper- volume fraction =0.4. An SEM image showing the micros- ature when tested in argon at temperatures as high as 1300C. tructure of an as-received sample is shown in Fig. 1. The However, limited data exist related to the thermal aging Youngs modulus of as-received samples(E=122+2 GPa) behavior in inert atmospheres of this kind of composites, i.e. was determined using a forced resonance frequency tech for unstressed long-tcrm exposures at high temperatures in,. nique, as shown elsewhere [29] for example, vacuum or argon. Certainly, the lack of interest Thermal aging involved encapsulating the as-received bars in the performance of composites in non oxidizing environ- in silica tubes after these had been evacuated and filled with ments has been motivated by the apparent absence of signif- argon. Argon of technical purity was used, the oxygen content icant envisaged potential applications for these materials in in the capsules being of a level lower than 0.001 Lorr(. 133 such environments. Besides the fact that some application Nm). The capsules were introduced in a furnace at given possibilities in non-oxidizing environments do exist [25], aging temperatures and maintained for 100 h. Three the assessment of thermal aging behavior in inert atmospheres temperatures were tried: 500%C, 600oC and 700C presents a convenient way to separate'pure'thermaleffects temperatures were chosen to correspond with those used in from oxidation effects, both of which are likely to occur for separate studies on thermal shock and thermal cycling behav exposures in air. Moreover, knowledge of the materials ior of similar composites [27, 29]. They cover the upper response at high-temperatures in the absence of oxygen will temperature range at which these composites are likely to find aid in assessing the relative e fficacy of strategies developed for inhibiting the rate of oxidation embrittlement (e.g. fiber Table I coatings and use of oxygen getters in the matrix [26]).On operties of the composite constituents [5.61 this basis. the on-going research program, focused on inves- Property Matrix: DURAN Fibre: SiC Nicalon tigating the thermomechanical performance of commercially (borosilicate glass) (NL 202 available SiC-Nicalon fiber reinforced silicate matrix com- posites, has contemplated the study of aging effects in inert Density(g cm - atmospheres in addition to the experiments in oxidizing envi- Young's modulus( GPa) 63 ronments. In this article, the results of the investigations con- Poissons ratio 020 ducted employing non-oxidizing atmospheres are presented Thermal expan 3.25×10 coefticient (K A separate article deals with the results in oxidizing environ Tensile strength(MPa) 2750 ments[27] Fibre volume fraction: 0.4 A commercial borosilicate glass matrix composite rein- Fibre radius: 5-9 um forced with Sic (Nicalon) fibers was investigated. Aging was conducted at temperatures between 500C and 700C in argon. Several techniques were used to detect the possible degradation of the structural integrity and mechanical prop- erties of the materials after the heat-treatments, including standard four-point flexure strength tests, three-point flexor chevron-notch tests. and fiber push-out indentation tests nother objective of this article is to examine the applicability of the chevron-notched specimen technique to monitor changes in the mechanical properties in this class of compos- ite materials. In the literature reviewed by the authors, only one study was found which has considered in the past the u of chevron-notched specimens to evaluate fracture properties in fiber reinforced ceramic matrix composite 2. Experimental The material investigated was a commercially available Fig. 1 SEM image of a polished section of an as-received sample. showing unidirectional SiC Nicalon(NL202)fiber reinforced boro- the fairly regular fiber distribution and absence of porosity

Experimental research has been conducted also on the mechanical properties of these composites at high tempera￾tures in inert atmospheres. The lirst work on this subject was probably that carried out as early as in 1969 by Crivelli￾Visconti and Cooper [ 231 on C-fiber reinforced silica tested at 800°C. A more detailed investigation was conducted by Prewo et al. 1241, showing that the ultimate strength of Sic (Nicalon) fiber-lithium aluminosilicate matrix composites did not change significantly from the value at room-temper￾ature when tested in argon at temperatures as high as 1300°C. However, limited data exist related to the thermal aging behavior in inert atmospheres of this kind of composites, i.e. for unstressed long-term exposures at high temperatures in, for example, vacuum or argon. Certainly, the lack of interest in the performance of composites in non-oxidizing environ￾ments has been motivated by the apparent absence of signif￾icant envisaged potential applications for these materials in such environments. Besides the fact that some application possibilities in non-oxidizing environments do exist [ 2.51, the assessment of thermal aging behaviorin inertatmospheres represents a convenient way to separate ‘pure’ thermal effects from oxidation effects, both of which are likely to occur for exposures in air. Moreover, knowledge of the materials response at high-temperatures in the absence of oxygen will aid in assessing the relative efficacy of strategies developed for inhibiting the rate of oxidation embrittlement (e.g. fiber coatings and use of oxygen getters in the matrix [ 261). On this basis. the on-going research program, focused on inves￾tigating the thermomechanical performance of commercially available Sic-Nicalon fiber reinforced silicate matrix com￾posites, has contemplated the study of aging effects in inert atmospheres in addition to the experiments in oxidizing envi￾ronments. In this article, the results of the investigations con￾ducted employing non-oxidizing atmospheres are presented. A separate article deals with the results in oxidizing environ￾ments [ 271, A commercial borosilicate glass matrix composite rein￾forced with Sic (Nicalon) fibers was investigated. Aging was conducted at temperatures between 500°C and 700°C in argon. Several techniques were used to detect the possible degradation of the structural integrity and mechanical prop￾erties of the materials after the heat-treatments, including standard four-point Rexure strength tests, three-point flexure chevron-notch tests, and fiber push-out indentation tests. Another objective of this article is to examine the applicability of the chevron-notched specimen technique to monitor changes in the mechanical properties in this class of compos￾ite materials. In the literature reviewed by the authors, only one study was found which has considered in the past the use of chevron-notched specimens to evaluate fracture properties in fiber reinforced ceramic matrix composites 128 1. 2. Experimental The material investigated was a commercially available unidirectional Sic Nicalon (NL202) fiber reinforced boro￾silicate (DURAN) glass matrix composite fabricated by Schott Glaswerke (Mainz, Germany). Information on @e composite constituents is given in Table 1. The composites were prepared by the sol-gel-slurry method. Details on the processing technique are found in the literature [5,6]. The samples were received in the form of rectangular test bars of nominal dimensions (4.5 mm X 3.8 mm X 100 mm). De density of the composites was 2.4 g crnm3 and their fiber volume fraction = 0.4. An SEM image showing the micros￾tructure of an as-received sample is shown in Fig. 1. The Young’s modulus of as-received samples ( E = 122 f 2 GPa) was determined using a forced resonance frequency tech￾nique, as shown elsewhere [ 291. Thermal aging involved encapsulatingthe as:received bars in silica tubes after these had been evacuated and filled with argon. Argon of technical purity was used, the oxygen content in the capsules being of a level lower than 0.001 torr [O. 13; Nm-‘). The capsules were introduced in a furnace at given aging temperatures and maintained for 100 h. Three aging temperatures were tried: 500°C. 600°C and 700°C. Thxe temperatures were chosen to correspond with those used in separate studies on thermal shock and thermal cycling behav￾ior of similar composites [ 27,291. They cover the upper temperature range at which these composites are likely to find Table 1 Properties of the composite conbtituentz. [5.6] Property Matrix: DURAN Fibrc: Sic Nicalon (borosilicate glaw) (NL 203) ~ Density ig cm-‘) 2.23 3.55 Young’s modulus ( GPa) 63 196 Poisson‘5 ratio 0.22 0.20 Thermal expansion 3,2s x lo-” 3X10-h ~~ coeflicient I K- I ) Tensile strength (MPa) =60 2750 Fibre volume fraction: O.-l Fibre radius: S-9 pm Fig. 1. SEM image of a polished section of an as-received sample. showing the fairly regular fiber distribution and absence of porosily

A R, Boc'caccint et al. /Materials Chemistr and Physics 53(/998)155-104 applications. For the 600"C tests. aging experiments of 250 and 1000 h were conducted also. After thermal treatment. the mass and dimensions of the samples were measured, and the outer surfaces inspected carefully for the appearance of any macroscopic damage, such as change of color, delamination, fiber debonding/protrusion, etc. X-ray diffraction (XRD) analyses were performed on polished sections of the sample after aging in order to detect any crystallization in the matrix Aged samples together with samples in the as-received con- dition were characterized using several mechanical testing The flexural strength( modulus of rupture)and Youngs modulus were determined in four-point bending using 32 mm outer span and 16 mm inner span. A cross-head speed of 0. 1 mm s was used. The chevron-notched specimen tech nique was employed for fracture toughness determination Fig. 2. Indentor acting on an SiC-Nicalon fiber in an as-received sample during a push -out test Chevron notches with angles of 90 were cut carefully using a thin (0. 15 mm) diamond wheel. A three point bending sure. Similar measurements on as-received samples have configuration at stant cross-head speed of 0. 1 mm been conducted previously [29,35]. The measurement ofthe min"was employed and the test bars were 60 mm in length. interfacial characteristics was conducted using an in-situ Graphs of load versus time were recorded and the maximu SEM indentation apparatus(Touchstone Ltd, West Virginia, force was determined from each trace. The fracture toughness USA). The thickness of the specimens( two samples foreach value was calculated from the maximum load( Fmax )and the condition) was in the range 340-362 um. A diamond inden orresponding minimum value of geometrical compliance tor was used to apply the load to the fibers at a displacement function(Ymin" )according to the equation 30] rate of 0. 18 um s-. The apparatus offers high accuracy for Fm y the indentor positioning, which allows the push-out (1) to be carried out without fiber and /or matrix damage. The load and displacement data are digitally recorded and the where B and w are the thickness and height of the specimens, push-out events are observed in-situ. As an example of the respectively. The calculation of the geometric function Yn technique. Fig. 2 shows the indentor acting on an Sic Nicalon for chevron-notched bend bars was based on the use of fiber in an as-received sample. Between 15 and 45 fibers were Bluhm's slice model [31]. To overcome the associated com- indented at each condition and the average interfacial fric putational complexities of this approach, Shang-Xian [321 tional shear stress(r). as well as the interfacial shear stress and Withey et al. [ 33] have suggested a simplified solution at the debonding initiation( Tinit )and the debonding shear to determine normalized stress intensity factors for chevron- strength (Tb) were determined. A complete description of notched specimens using trigonometric functions. The cal- the indentation push-out technique and details concerning the ulation procedure used for the purposes of this investigation calculation of the stresses are presented elsewhere [351 has been described in detail elsewhere [34]. The chevron notch depth ag was measured after testing from SEM micro- graphs of fractured specimens. Additionally, the acoustic 3. Results emission technique was used during the test. Traces of cumu- lative number of counts(acoustic emission(AE) events 3 1. Mechanical tests were obtained in the same time scale as the load versus time plots. This technique allows for an accurate detection of the Table 2 summarizes the results obtained from the mechan- onset of microcracking at the chevron notch tip which occurs ical tests performed. The data for the Youngs modulus and when a sharp increase in the number of AE events is observed. the ultimate flexural strength for the as-received and heat Valid measurements for computing Kle are those in which treated material correspond closely to values reported in the this increase of AE events coincides with the end of the linear literature for similar composites [5,6], except for the two part of the force versus time trace, as explained below, The more extreme aging conditions of 600oC-1000 h and 700C work of fracture( woF) was evaluated from the area under 00 h. For these two conditions, there is an abrupt decrease the load-displacement curve. At least five measurements for of the flexural strength and a(less pronounced decrease of rere pe d and the results ar he young s modulus. the fracture toughness values deter fracture surfaces of broken samples were observed by SEM. mined by the chevron-notched specimen technique seem to Push-out indentation tests on aged samples were performed be little dependent on the aging condition. The material exhib in order to assess if any change had occurred in the interfacial its average Ki values in the range 19-26 MPay'm, which are properties as a consequence of the high-temperature expo- similar to data quoted in the literature for similar fiber rein

applications. For the 600°C tests, aging experiments of 2.50 and 1000 h were conducted also. After thermal treatment, the mass and dimensions of the samples were measured, and the outer surfaces inspected carefully for the appearance of any macroscopic damage, such as change of color, delamination, fiber debonding/protrusion, etc. X-ray diffraction (XRD) analyses were performed on polished sections of the samples after aging in order to detect any crystallization in the matrix. Aged samples together with samples in the as-received con￾dition were characterized using several mechanical testing techniques. The Aexural strength ( modulus of rupture) and Young’s modulus were determined in four-point bending using 32 ~-II outer span and 16 mm inner span. A cross-head speed of 0.1 mm s- ’ was used. The chevron-notched specimen tech￾nique was employed for fracture toughness determination. Chevron notches with angles of 90” were cut carefully using a thin (0.15 mm) diamond wheel. A three point bending configuration at a constant cross-head speed of 0.1 mm min-’ was employed and the test bars were 60 mm in length. Graphs of load versus time were recorded and the maximum force was determined from each trace. The fracture toughness value was calculated from the maximum load (F,,,,,) and the corresponding minimum value of geometrical compliance function ( Ymln+) according to the equation [ 30 ] : where B and Ware the thickness and height of the specimens, respectively. The calculation of the geometric function Y,,,,“* for chevron-notched bend bars was based on the use of Bluhm’s slice model [ 3 I 1. To overcome the associated com￾putational complexities of this approach, Shang-Xian [ 32 ] and Withey et al. [ 331 have suggested a simplified solution to determine normalized stress intensity factors for chevron￾notched specimens using trigonometric functions. The cal￾culation procedure used for the purposes of this investigation has been described in &tail elsewhere [ 341. The chevron￾notch depth cl0 was measured after testing from SEM micro￾graphs of fractured specimens. Additionally. the acoustic emission technique was used during the test. Traces of cumu￾lative number of counts (acoustic emission ( AE) events) were obtained in the same time scale as the load versus time plots. This technique allows for an accurate detection of the onset of microcracking at the chevron notch tip. which occurs when a sharp increase in the number of AE events is observed. Valid measurements for computing K,, are those in which this increase of AE events coincides with the end of the linear part of the force versus time trace, as explained below. The work of fracture ( WOF) was evaluated from the area under the load-displacement curve. At least five measurements for each condition were performed and the results averaged. The fracture surfaces of broken samples were observed by SEM. Push-out indentation tests on aged samples were performed in order to assess if any change had occurred in the interfacial properties as a consequence of the high-temperature expo￾Fig 2. lndentor acting on an Sic-Nicalon fiber in an as-received si during a push-out test. sure. Similar measurements on as-received samples have been conducted previously [ 29,351. The measurement of the interfacial characteristics was conducted using an in-situ SEM indentation apparatus (Touchstone Ltd., West Virginia, USA). The thickness of the specimens (two samples foreach condition) was in the range 340-362 km. A diamond inden￾tor was used to apply the load to the fibers at a displacement rate of 0.18 km s-‘. The apparatus offers high accuracy for the indentor positioning, which allows the push-out process to be carried out without fiber and/or matrix damage. The load and displacement data are digitally recorded and the push-out events are observed in-situ. As an example of the technique, Fig. 2 shows the indentor acting on an SIC Nicalon fiber in an as-received sample. Between 1.5 and 45 fibers were indented at each condition and the average interfacial fric￾tional shear stress ( Tag), as well as the interfacial shear stress at the debonding initiation (.i;,,,,) and the debonding shear strength ( 7db) were determined. A complete description of the indentation push-out technique and details concerning the calculation of the stresses are presented elsewhere [ 351. 3. Results 3.1. Mec~hnizic~al tests Table 2 summarizes the results obtained from the mechan￾ical tests performed. The data for the Young’s modulus and the ultimate flexural strength for the as-received and heat￾treated material correspond closely to values reported in the literature for similar composites [ 5,6]. except for the two more extreme aging conditions of 6OO”C-1000 h and 7OO”C- 100 h. For these two conditions, there is an abrupt decrease of the flexural strength and a (less pronounced) decrease of the Young’s modulus. The fracture toughness values deter￾mined by the chevron-notched specimen technique seem to be little dependent on the aging condition. The material exhib￾its average K,, values in the range 19-26 MPaJm, which are similar to data quoted in the literature for similar fiber rein-

158 A.R. Boccaceint er al/ Marericis Chemistr and Piivsics 53(7998)155-1o4 Table 2 Mechanical properties of SiC-fiber reinforced DURAN-glass matrix composites after agings at different durations and temperatures in argon Aging condition Y Flexure strength Fracture touches Work of fracture K, (st, dev, Jm-2) Duration(h) (MPa) f MPam) 500:100 240(0.9) 600:100 7550 600:250 25.7!.3 600:1000 0.8(0.7) 700:100 00 9.2(1.1 638 not measured not measured Determined in four-point bending. etermined using three N aged 600"C/100h CE+O 03D ig. 3. Typical load-displacement plots obtained in chevron-notch tests for samples aged for 100 h at different temperatures: (a)500 C, (b)600 C, (c)700C forced glasses and glass-ceramics 16] and for SiC/SiC MPay/m)[16]. The lower work of fracture of the samples itcs [36]. These Kle values are also similar to the aged at 600 C for 1000 h and at 700C for 100 h can be fracture toughness of aluminum alloys( 22 MPa/m)an correlated with the lower flexure strength of these materials 30% lower than those of graphite epoxy composites

Table 2 Mechanical properties of Sic-fiber reinforced DURAN-glass matrix composites after agings at different durations and temperatures in argon Aging condition Young’s modulus * Flexure strength IOE,O 000 0 10 0.20 0.30 0.40 deflection Fig. 3. Typical load-displacement plots obtained in chevron-notch tests for samples aged for 100 h at different temperatures: (a) ~OO@C, (b) 6oo”c, cc) 7~7~. forced glasses and glass-ceramics [ 161 and for Sic/Sic MPaJm) [ 161. The lower work of fracture of the samples composites [ 361. These K,, values are also similar to the aged at 600°C for 1000 h and at 700°C fo? 100 h can-be fracture toughness of aluminum alloys (22 MPaJm) and only correlated with the lower flexure strength of these materials. = 30% lower than those of graphite-epoxy composites (34 Typical load-displacement plots obtained in chevron-notch

R Boccaccini er u. / Materials Chemistry and PhysIcs 53(1998)155-104 3.2. Push-out indentation tests Fig. 5 shows typical stress-displacement curves during aged 600"C/100 h 600/250h oush-out indentation for samples aged for 100 h at 500C 600C and 700C, The interfacial parameters, frictional shear 600e/1000h stress( Tr), debonding initiation shear stress(Tinit )ane debonding shear strength(Tdb), were determined from the data of the push-out experiments using the linear model of Marshall [37]. The calculated values are listed in Table 3 Similar measurements on as-received samples have been con ducted previously [29,35], and the results are also listed in Table 3. These data confirm the existence of a weak fiber/ matrix interfacial bond, which results from the presence of a thin carbon-rich laver at the interface. a characteristic feature of these composite materials [38]. It is observed that for the aged samples, Tr remains nearly constant(Tr=13-18 MPa) and close to the value obtained for the as-received material Thus. these results indicate that the carbon-rich interfacial 8x 0.00 0. 10 deflection (mm) 0.40 0.5 layer at the matrix/fiber interface, which provides thecom- Fig4. Load-displacement curves obtained in chevron-notch tests for three posite'properties to the material, i.e. the pseudo-ductile frac samples aged at 600C for different durations ture behavior, has not been degraded significantly under the heat-treatments in inert environments. The average debond tests are shown in Fig. 3 for samples aged for 100 h at dif- ing shear stress( Tab) was also little influenced by the aging ferent temperatures. Also the acoustic emission events conditions, being also in close agreement with the values recorded during the tests are plotted. Fig 4 shows the load- determined in the as-received material(Tab=29+3 MPa or displacement curves obtained in chevron-notch tests for three 37+4 MPa for sample thicknesses of 340 and 355 um samples aged at 600C for different durations. The poorer respectively). Table 3 also reveals that for agings at 500C mechanical response of the samples aged at the most severe all three characteristic stresses are very close to the conditions, 100 h at 700C and 1000 h at 600C. hecomes received values and in both cases the standard deviation of evident by observing the plots the data are similar. However, starting with the aging at 10 Fig. 5. Typical stress-displacement curves during push-out indentation for samples aged for 100 h at 500C. 600C and 700C Table 3 Interfacial property data measured by push-out indentation technique, Aging of the samples was conducted for 100 h at the temperatures indicated Thickness Average diameter of tcsted fibre shear stress hear stress debonded fues ( MPa Tab(MPa) Tr(MPa) s receive 17.55÷1.23 16.8±3.7 4±3.3 15.6±2.0 As received 18.00±I.9 172±4.5 375±4,4 175±3.6 Aging 500C 362 16.17±.0 0+5,6 17.7±2.6 aging600°C 16.48±97 8.1±1.7 94±8.5 4.6±3.1 A 1625±1.30 8.5±4.4 11.2±3.1 0.0+4. 324±82 13.0±2.6

L59 1600 1111,,~1,,11~,,1111,1111 force PI 1400 - aged 600 “C / 100 h 6OO”C/ 250 h - 1200 - 600”C/1000h - 1000 - 800 - 600 - 0.00 0 10 0.20 0.30 deflection [mm] 0 40 0.3 Fig. 4. Load-displacement curves obtained in chevron-notch tests for three samples aged at 600°C for different durations. tests are shown in Fig. 3 for samples aged for 100 h at dif￾ferent temperatures. Also the acoustic emission events recorded during the tests are plotted. Fig. 4 shows the load￾displacement curves obtained in chevron-notch tests for three samples aged at 600°C for different durations. The poorer mechanical response of the samples aged at the most severe conditions, 100 h at 700°C and 1000 h at 6OO”C, becomes evident by observing the plots. 3.2. Push-out indentation tests Fig. 5 shows typical stress-displacement curves during push-out indentation for samples aged for 100 h at 5OO”C, 600°C and 700°C. The inter-facial parameters, frictional shear stress (rr,), debonding initiation shear stress (Ti”i,) and debonding shear strength ( TV,,), were determined from the data of the push-out experiments using the linear model of Marshall [ 371. The calculated values are listed in Table 3. Similar measurements on as-received samples have been con￾ducted previously [ 29,351, and the results are also listed in Table 3. These data confirm the existence of a weak fiber/ matrix interfacial bond, which results from the presence of a thin carbon-rich layer at the interface, a characteristic feature of these composite materials [ 381. It is observed that for the aged samples, rfTfr remains nearly constant ( 7rr = 13-l 8 MPa) and close to the value obtained for the as-received material. Thus, these results indicate that the carbon-rich interfacial layer at the matrix/fiber interface, which provides the ‘com￾posite’ properties to the material, i.e. the pseudo-ductile frac￾ture behavior, has not been degraded significantly under the heat-treatments in inert environments. The average debond￾ing shear stress (Tag) was also little influenced by the aging conditions, being also in close agreement with the values determined in the as-received material ( rdb = 29 f 3 MPa or 37 i4 MPa for sample thicknesses of 340 and 355 p,m, respectively). Table 3 also reveals that for agings at 500°C all three characteristic stresses are very close to the as￾received values and in both cases the standard deviation of the data are similar. However, starting with the aging at 0 1 2 3 4 5 6 7 8 9 IO Displacement (microns) Fig. 5. Typical stress-displacement curves during push-out indentation for samples aged for 100 h at 500°C 600°C and 700°C. Table 3 Interfacial property data measured by push-out indentation technique. Aging of the samples was conducted for 100 h at the temperatures indicated Specimen Thickness (pm) Average diameter of Debonding initiation Debonding shear Frictional Percentage of teared tibre shear stress strength shear stress debonded fibrea (pm) L,,, ( MPa) 7db (MPa) G (MPa) (S’o) As received 310 17.55 + 1.23 16.5 i 3.7 29.4 + 3.3 15.6k2.0 0 As received 355 IS.00 i I .97 17.2k4.5 37.5 14.4 17.5 f 3.6 0 Aging 500°C 362 16.17+ 1.06 17.3 i2.7 35.Ok5.6 17.7 & 2.6 0 Aging 600°C 318 16.48 k 1.97 8.1 + 1.7 39.4 i 8.5 14.6i3.1 13 Aging 700°C 360 16.25 i 1.30 8.5 * 4.4 27.3 i 10.9 11.2+3.1 25 without debonded libres: 10.0~4.1 32.4 k 8.2 13.012.6 0

A R Boccaccini et al. Materials Chemistry and Physica 53(1998)155-10 600C, the standard deviation of the measured debonding shear strength is increased up to 50% in comparison with the nRs. x 25.8 PHuI. 18 values for as-received material. Furthermore the initiation debonding stress(Tinit of the samples aged at 600C and 700C is decreased nearly by a factor of two compared with the as-received material. In particular. for the sample aged at 700C, a considerable number of partially and completely debonded fibers was found, the fibers being held only by mechanical clamping. 3.3. Microscopy Fracture surfaces of an as-received samplc and of a sample aged at 600 C for 100 h are shown in Fig. 6(a)and (b). Both fracture surfaces exhibit the characteristic extensive fiber pull-out of these composite materials [5,6, 29]. Inspection of che micrographs reveals that the average pull-out lengths in both samples is similar. Similar qualitative behavior was 3a;。命1x1 observed for the rest of the samples independently of the aging condition. This is consistent with the little variation of fracture toughness values measured by the chevron-notch technique, as reported in Table 2. Furthermore, the significant fiber pull-out effect exhibited by the composites is indicative of weak fiber/matrix interfacial bonds, and thus a qualitative confirmation of the results of the push-out indentation measurements 4. Discussion 4.1. The chevron-notch technique coupled with acoustic emission: assessment of its applicability to fiber reinforced. Fig. 6. Fracture surfaces of an as-received sample (a) and of a sample aged glass matrix composite at 600C for 100 h(b), showing similar fiber pull-out lengths. As mentioned in the Introduction, the chevron-notch tech- faces. Besides considering this successful application of the nique has not received wide application to measure the frac hevron-notched specimen technique for fiber reinforced ture toughness of fber reinforced ceramic composite brittle matrix composites by Ha and Chawla[28], the fol- materials. Other specimen configurations have been more lowing reasons were taken into account also for choosing the extensively used, for example notched three-point bending chevron-notch technique in the present study pecimens with notch depth-sample thickness ratios of 0.5 (i) Literature results and own previous experiments have [3, 16]. As pointed out by Mecholsky [391, as long as the shown that chevron-notched specimens can provide highly notch is large enough to cover many fibers at the critical crack reliable and accurate results for fracture toughness determi length, the chevron-notch test should be a useful method to nation in particulate reinforced brittle matrix composites measure toughness and work of fracture in ceramic compos-: 33,41, including glass matrices similar to that of the present ites. Moreover, the chevron-notched specimen technique is composites [41] apted for tests at elevated temperatures 140] and 1)For unidirectionally reinforced materials in which the therefore, it should be considered more frequently for KI fracture behavior is governed not only by the matrix brittle- o. posites t n high-temperature fiber reinforced ceramic ness but also by the fiber/matrix interface, the chevron-notch 1-notched three-point flexure specimens to determine both laminates several years ago (42 ally revealed by testing inlullite composites, Ha and Chawla [28] used chey ck of fracture(wOF)and fracture toughness(Kic). The (iii) The lack of previous studies in the literature dealin technique was capable of detecting the different mechanical with the application of the chevron-notch test to determine behavior of composites fabricated BN-coateu and fracture properties in fiber reinforced brittle matrix compos uncoated fibers. The different toughness and Wof values as menti above, indicating the need for further measured were consistent with obscrvations of fracture su

6OO”C, the standard deviation of the measured debonding shear strength is increased up to 50% in comparison with the values for as-received material. Furthermore, the initiation debonding stress (T,,,~~) of the samples aged at 600°C and 700°C is decreased nearly by a factor of two compared with the as-received material. In particular, for the sample aged at 700°C a considerable number of partially and completely debonded fibers was found, the fibers being held only by mechanical clamping. 3.3. Microscop> Fracture surfaces of an as-received sample and of a sample aged at 600°C for 100 hare shown in Fig. 6(a) and (b). Both fracture surfaces exhibit the characteristic extensive fiber pull-out of these composite materials [ 5,6,29]. Inspection of the micrographs reveals that the average pull-out lengths in both samples is similar. Similar qualitative behavior was observed for the rest of the samples independently of the aging condition. This is consistent with the little variation of fracture toughness values measured by the chevron-notch technique, as reported in Table 2. Furthermore. the significant fiber pull-out effect exhibited by the composites is indicative of weak fiber/matrix interfacial bonds, and thus a qualitative confirmation of the results of the push-out indentation measurements. 4. Discussion 4.1. The chevron-notch rechnique coupled with acoustic emission: nssessment of its npplicnbiliry tojber reirzfkced glass matrix composites As mentioned in the Introduction, the chevron-notch tech￾nique has not received wide application to measure the frac￾ture toughness of fiber reinforced ceramic composite materials. Other specimen configurations have been more extensively used, for example notched three-point bending specimens with notch depth-sample thickness ratios of 0.5 [ 3,161. As pointed out by Mechoisky [39 1, as long as the notch is large enough to cover many fibers at the critical crack length, the chevron-notch test should be a useful method to measure toughness and work of fracture in ceramic compos￾ites. Moreover, the chevron-notched specimen technique is well adapted for tests at elevated temperatures [40] and, therefore, it should be considered more frequently for K,, determination in high-temperature fiber reinforced ceramic composites. In their study on the mechanical properties of mullite-mullite composites, Ha and Chawla [ 28] used chev￾ron-notched three-point Rexure specimens to determine both work of fracture (WOF) and fracture toughness (K,, j . The technique was capable of detecting the different mechanical behavior of composites fabricated using BN-coated and uncoated fibers. The different toughness and WOF values measured were consistent with observations of fracture sur￾Fig. 6. Fracture surfaces of an as-received sample ( a) and of a sample aged at 600°C for 100 h (b), showing similar fiber pull-out length&. faces. Besides considering this successful application of the chevron-notched specimen technique for fiber reinforced brittle matrix composites by Ha and Chawla [2X], the fol￾lowing reasons were taken into account also for choosing the chevron-notch technique in the present study: (i) Literature results and own previous experiments have shown that chevron-notched specimens can provide highly reliable and accurate results for fracture toughness determi￾nation in particulate reinforced brittle matrix cornposit& [ 33,413, including glass matrices similar to that of the present composites [ 411. (ii) For unidirectionally reinforced materials in which the fracture behavior is governed not only by the matrix brit&- ness but also by the fiber/matrix interface, the chevron-notch is a successful crack initiator. Indeed, the functionality of chevron-notch specimens was initially revealed by testing laminates several years ago [ 421. (iii) The lack of previous studies in the literature dealing with the application of the chevron-notch test to determine fracture properties in fiber reinforced brittle matrix compos￾ites, as mentioned above, indicating the need for further research in this area

A R. Bvcccccini et al/ Mcieriuls Citerrcisirv urd! Pltysic3 5311998)235-104 Acoustic emission, a technique that has gained application 1000 h and at 700C for 100 h are analyzed next, Neither the for detecting the onset of microcracking when testing com- Nicalon fibers nor the interfaces should be significantly posite materials [43, 44], was used to assess whether valid degraded by theInal exposures in inert atmospheres al conditions for obtaining fracture toughness data from the T< 1000"C. Extensive literature data support this statement. chevron-notch tests were met. When a mechanical stress is For example, several studies have shown that for aging applied to a composite, several fracture phenomena can occur inert atmospheres at the temperatures used in the present in the material, including matrix microcracking, fiber-matrix study, the Nicalon NL202 fibers remain stable e no micros- debonding, delamination and fiber failure. During these proc ructural changes occur, such as decomposition of the Si-C Cssc crgy is released in various forms, onc of them bcing O phase, crystallization, grain growth, ctc. [ 17, 45-48]. Other acoustic waves [43]. Thus. using the acoustic emission tech- studies coincide in that at least up to =800 C, the Nicalon ique, the actual onset of microcracking initiation at the chev nd related Tyranno fibers remain chemically stable in both on-notch tip is possible to be identified. This is the case when oxidizing and inert atmospheres, even for exposures as long sharp increase in the cumulative number of counts(AE as 500 h[ 15, 16 48-53]. In particular. it was shown that Nica events)is observed. An increase in the number of AE events lon fibers retain a high tensile strength of 2 GPa up to 800C bserved at the end of the linear part of the load-time trace, for short exposures [51.53], while the Weibull modulus and indicates that the actual crack is developed at the chevron- characteristic strength of single Nicalon fibers aged in air at notch tip(see Fig 3).In these cases, the measurements of 800C systematically decreased with increasing exposure Kle can be taken as valid. since the unstable fracture(at time [54]. However when the fibers are surrounded by the maximum force )occurs from a propagating crack perpen- matrix in an inert environment, it is expected that the strength dicular to the fiber axis degradation due to chemical attack be much less severe at The acoustic emission technique also showed that individ- these relatively low temperatures. Indeed, significant strength lal fracture events started at about 1/2 to 2/3 of the maximum degradation was reported to occur only at temperatures higher load. This can be evidenced by a smooth increase in the than =1 100.C in both inert and oxidizing atmospheres due cumulative number of AE events. Microstructural changes to the decomposition of the amorphous constituent of the corresponding to this acoustic emission should be matrix fiber recrystallization cavity formation and/or creep effects microcracking and partial local interfacial decohesion. These [47, 51, 55-58] microfracture events. however, did not result in a departure The results of the push-out technique( Fig. 5. Table 3 rom linearity of the load-displacement trace. a phenomenon have shown that the interfacial frictional properties remain also exhibited by other fiber reinforced glass composites nearly unchanged for the different aging conditions investi 43]. The first non-linearity observed in all samples tested gated. Therefore, the degradation of the mechanical strength occurred at loads close to the maximum load applied Fig. of the material at the most severe exposures cannot be ) and it was associated with a strong increase in the cumu- ascribed strictly to a chemical degradation of the interfacial lative number of acoustic on events. This increase microstructure either. Thus the reason for the deterioration so high that only crack propagation through the glass matrix of the mechanical strength of the samples aged at 600C for and fiber debonding and fracture could be responsible for this 1000 h and at 700C for 100 h should be sought primarily in significant effect. In principle, different fracture mechanisms the evolution of microstructural changes in the matrix, which cause acoustic events with different wave energy or ampli- could, in tum, affect( at least locally)the fiber/matrix inter tude and, therefore it should be possible to identify the type face. XRD analyses conducted on thermally aged samples of fracture process taking place by analyzing AE parameters ruled out the possibility of an uncontrolled matrix crystalli uch as events, event rate and amplitude [43]. This analysis zation as a possible mechanism contributing to the degrada as beyond the scope of the present investigation, however. tion of the material. However softening of the matrix and the he near constancy of the interfacial frictional shear stress consequent microstructural rearrangement, could explain the for the different aging conditions( Fig. 5 and Table 3), which observed loss in strength and stiffness that occurred without confirms the existence of the carbon-rich interface layer, significant changes in the toughness values, as discussed relates well with the similarly little variation of the fracture below. toughness values measured using the chevron- notch tech ble 2) for the different samples. This. in turn, is 43. Effect of the glass matrix softening also verified by the fracture surface examination of as- received and aged samples: significant fiber pull-out and thus The viscosity of the duran borosilicate glass matrix at composite'failure behavior was observed for all samples 500C. 600C and 700C is 10235.3, 10 and 10.Pa s, ndependently of the aging condition espectively [59], indicating the increasing possibility of matrix viscous flow with increasing temperature. The starting 4.2. Analysis of the aging behaviour at 600C and 700C ftening of the matrix in similar borosilicate glass matrix composites at temperatures above 500C has been docu The reasons for the marked decrease of the Youngs mod- mented by other authors[60-62], who also found extensive ulus and flexural strength of the samples aged at 600C for viscous flow of the matrix at 600 C and 700C

Acoustic emission, a technique that has gained application for detecting the onset of microcracking when testing com￾posite materials [ 43,441, was used to assess whether valid conditions for obtaining fracture toughness data from the chevron-notch tests were met. When a mechanical stress is applied to a composite, several fracture phenomena can occur in the material, including matrix microcracking, fiber-matrix debonding, delamination and fiber failure. During these proc￾esses, energy is released in various forms, one of them being acoustic waves [ 431. Thus. using the acoustic emission tech￾nique, the actual onset of microcracking initiation at the chev￾ron-notch tip is possible to be identified. This is the case when a sharp increase in the cumulative number of counts (AE events) is observed. An increase in the number of AE events observed at the end of the linear part of the load-time trace, indicates that the actual crack is developed at the chevron￾notch tip (see Fig. 3). In these cases, the measurements of K,, can be taken as valid, since the unstable fracture (at maximum force) occurs from a propagating crack perpen￾dicular to the fiber axis. The acoustic emission technique also showed that individ￾ual fracture events started at about l/2 to 2/3 of the maximum load. This can be evidenced by a smooth increase in the cumulative number of AE events. Microstructural changes corresponding to this acoustic emission should be matrix microcracking and partial local interfacial decohesion. These microfracture events, however, did not result in a departure from linearity of the load-displacement trace, a phenomenon also exhibited by other fiber reinforced glass composites [43]. The first non-linearity observed in all samples tested occurred at loads close to the maximum load applied (Fig. 3), and it was associated with a strong increase in the cumu￾lative number of acoustic emission events. This increase is so high that only crack propagation through the glass matrix and fiber debonding and fracture could be responsible for this significant effect. In principle, different fracture mechanisms cause acoustic events with different wave energy or ampli￾tude and, therefore, it should be possible to identify the type of fracture process taking place by analyzing AE parameters such as events, event rate and amplitude [ 431. This analysis was beyond the scope of the present investigation. however. The near constancy of the interfacial frictional shear stress for the different aging conditions (Fig. 5 and Table 3)) which confirms the existence of the carbon-rich interface layer, cor￾relates well with the similarly little variation of the fracture toughness values measured using the chevron-notch tech￾nique (Table 2) for the different samples. This. in turn. is also verified by the fracture surface examination of as￾received and aged samples: significant tiber pull-out and thus ‘composite’ failure behavior was observed for all samples independently of the aging condition. The reasons for the marked decrease of the Young’s mod￾ulus and Aexural strength of the samples aged at 600°C for 1000 h and at 700°C for 100 h are analyzed next. Neither the Nicalon fibers nor the interfaces should be significantly degraded by thermal exposures in inert atmospheres at T< 1000°C. Extensive literature data support this statement. For example. several studies have shown that for aging in inert atmospheres at the temperatures used in the present study, the Nicalon NL202 fibers remain stable. i.e. no micros￾tructural changes occur, such as decomposition of the Si-C- 0 phase, crystallization, grain growth, etc. [ 17,45%4&]. Other studies coincide in that at least up to = 8OO”C, the Nicalon (and related Tyranno) fibers remain chemically stable in both oxidizing and inert atmospheres, even for exposures as long as 500 h [ 15.46.48-531. In particular. it was shown that Nica￾Ion fibers retain a high tensile strength of 2 GPa up to 800°C for short exposures [ 5 I .53]. while the Weibull modulus and characteristic strength of single Nicalon fibers aged in air at 800°C systematically decreased with increasing exposure time [ 541. However. when the fibers are surrounded by the matrix in an inert environment. it is expected that the strength degradation due to chemical attack be much less severe at these relatively low temperatures. Indeed, significant strength degradation was reported to occuronly at temperatures higher than = 1100°C in both inert and oxidizing atmospheres due to the decomposition of the amorphous constituent of the fiber, recrystallization, cavity formation and/or creep effects [47.51,55-581. The results of the push-out technique (Fig. 5, Table 3) have shown that the interfacial frictional properties remain nearly unchanged for the different aging conditions investi￾gated. Therefore. the degradation of the mechanical strength of the material at the most severe exposures cannot be ascribed strictly to a chemical degradation of the interfacial microstructure either. Thus, the reason for the deterioration of the mechanical strength of the samples aged at 600°C for 1000 h and at 700°C for 100 h should be sought primarily in the evolution of microstructural changes in the matrix, which could, in turn, affect (at least locally) the fiber/matrix inter￾face. XRD analyses conducted on thermally aged samples ruled out the possibility of an uncontrolled matrix crystalli￾zation as a possible mechanism contributing to the degrada￾tion of the material. However, softening of the matrix and the consequent microstructural rearrangement. could explain the observed loss in strength and stiffness that occurred without significant changes in the toughness values, as discussed below. 4.3. Efect of the glnss matrix softenirlg The viscosity of the DURAN borosilicate glass matrix at 500°C. 600°C and 700°C is 10’““-i, IO’O and 10”.6Pas-‘, respectively [ 591, indicating the increasing possibility of matrix viscous flow with increasing temperature. The starting of softening of the matrix in similar borosilicate glass matrix composites at temperatures above 500°C has been docu￾mented by other authors [ 60-621, who also found extensive viscous how of the matrix at 600°C and 700°C

A.R. BoLcac'cini er al. Materials Chemistry and Physics 53(1998)155-164 SEM observations of polished surfaces of samples aged at 20.0M1 GX769.P3 700C for 100 h confirm porosity formation in the matrix around fibers in the present composites, as Fig. 7 shows Besides porosity formation, the viscous flow of the matrix at 600C and 700C has resulted in local interfacial separation and interfacial cracking, which may be the reason for the ower values of the debonding initiation shear stress and the higher scattering of the measured debonding shear strength in these samples (Table 3). Matrix viscous flow may have also produced reaccommodation or displacement of the fibers within the composite. All mentioned mechanisms are likely to result in the loss of composite strength and stiffness obscrved. For example, extensive fiber-fiber contact as a consequence of the matrix softening and microstructural rear- rang of contact, and thus to reduced composite strength due to premature fiber failure [63,64]. The cavitation of the matrix Fig. 7. SEM image of a polished surface of a sample aged at700-C for [00 may also lead to damage on the fiber surface, thus increasing matrix viscous fow The arrows indicate damage on hber surfaces he flaw population of the fiber, which should, in turn, con ribute to the decrease of the flexural strength of the compos- exposure, since at this temperature range the matrix softens, ites [65, 66]. Damage on fiber surfaces has been detected by leading to microstructural damage. This includes porosity SEM ( see area marked by arrows in Fig. 7). More detailed formation and fber debonding and rearrangement, that results icrostructural investigations would be necessary, however, in the loss of mechanical properties. particularly in very to assess the extent and characteristics of the damage on the fracture strength. Short excursions at these temperatures may fiber surface and its dependence on aging conditions, this. he possible, however. The three-point flexure chevron-notch being the focus of current research. Nevertheless, the present technique used was adequate to determine fracture toughness esults are in broad agreement with literature data for exam- and work of fracture values. the push-out test to determine ple, loss in fiexural strength as a consequence of porosity the interfacial properties. in these materials. The chevron- formation and matrix softening has been observed in ther- notch test measurements for different aging conditions wer mally aged samples of Nicalon-fiber reinforced lithium alu- consistent with the results of interfacial properties determined minosilicate(LAS) glass-ceramic matrix composites [16]. by the push-out indentation method, and with microscopic It must be mentioned. finally, that the large scattering of the observations of fracture surfaces of as-received and aged interfacial properties measured by the push-out technique for samples. The Kie values determined using the chevron-notch he samples aged at 700C is likely the result of the local technique(in the range 19-26 MPay'm)are comparable to damage caused by matrix softening. Moreover, the presence data obtained in similar composite materials using other tech- of pores and crevices at the interfaces in these samples should niques [161 contribute to the low interfacial strength measured, leading to the observed extensive pull-out as also suggested by other authors [67, 68] Acknowledgements The authors gratefully acknowledge Dr. w. Beier and 5 Conclusions Mr R. Liebald of Schott Glaswerke, Mainz, Germany, for supplying the composite samples. The experimental assis The thermal aging of Sic-Nicalon fiber reinforced boro- tance of Mr H P Feuz during the push-out indentation tests licate glass matrix composites in non-oxidizing atmosphere and of Dr. M. Reinisch during fracture toughness tests is for different exposure durations. Overall, the results obtained tute for Mechanics and Materials, University of California are in broad agreement with data in the literature, in that San Diego thermal exposures of silicate matrix composites at tempera tures below 1000C in inert atmospheres do not have direct deleterious consequences on the properties of the SiC-Nica- References lon fibers and on the interfaces. Thus the temperature capa bility of the matrix is the limiting factor for applications I K.K. Chawla. Ceramic Matrix Composites. Chapman and Hall, Lon involving non-oxidizing environments. For the borosilicate glass matrix investigated here, the limiting temperature is in [2]AG. Evans, F.W. Zok, R. M. McMeeking, Z Z Du, Models of higl temperature, environmentally assisted embrittlement in ceramic thie range 600C-700"C, depending on the duration of the atrix composites. J. Am. Ceram Soc. /y(1996)2345-2352

SEM observations of polished surfaces of samples aged at 700°C for 100 h confirm porosity formation in the matrix around fibers in the present composites, as Fig. 7 shows. Besides porosity formation, the viscous flow of the matrix at 600°C and 700°C has resulted in local interfacial separation and interfacial cracking, which may be the reason for the lower values of the debonding initiation shear stress and the higher scattering of the measured debonding shear strength in these samples (Table 3). Matrix viscous flow may have also produced reaccommodation or displacement of the fibers within the composite. All mentioned mechanisms are likely to result in the loss of composite strength and stiffness observed. For example, extensive fiber-fiber contact as a consequence of the matrix softening and microstructural rear￾rangement can lead to local stress concentrations at the points of contact, and thus to reduced composite strength due to premature fiber failure [ 63,641. The cavitation of the matrix may also lead to damage on the fiber surface, thus increasing the flaw population of the fiber, which should, in turn, con￾tribute lo the decrease of the Aexural strength of the compos￾ites [ 65,661. Damage on fiber surfaces has been detected by SEM (see area marked by arrows in Fig. 7). More detailed microstructural investigations would be necessary, however, to assess the extent and characteristics of the damage on the fiber surface and its dependence on aging conditions. this being the focus of current research. Nevertheless. the present results are in broad agreement with literature data: for exam￾ple, loss in flexural strength as a consequence of porosity formation and matrix softening has been observed in ther￾mally aged samples of Nicalon-fiber reinforced lithium alu￾minosilicate (LAS) glass-ceramic matrix composites [ 161. It must be mentioned, finally, that the large scattering of the interfacial properties measured by the push-out technique for the samples aged at 700°C is likely the result of the local damage caused by matrix softening. Moreover, the presence of pores and crevices at the interfaces in these samples should contribute to the low interfacial strength measured, leading to the observed extensive pull-out, as also suggested by other authors [ 67.681. 5. Conclusions The thermal aging of Sic-Nicalon fiber reinforced boro￾silicate glass matrix composites in non-oxidizing atmosphere was investigated at temperatures in the range 5Oo”C-700°C for different exposure durations. Overall, the results obtained are in broad agreement with data in the literature, in that thermal exposures of silicate matrix composites at tempera￾tures below 1000°C in inert atmospheres do not have direct deleterious consequences on the properties of the SiC-Nica￾Ion fibers and on the interfaces. Thus, the temperature capa￾bility of the matrix is the limiting factor for applications involving non-oxidizing environments. For the borosilicate glass matrix investigated here, the limiting temperature is in the range 6OO”C-7OO”C, depending on the duration of the Fig. 7. SEM image of a polished surt’acr of a sampjlc aged ;1t 700°C for i@ h, showing porosity formation in the matrix around Libcrs and indication of matrix viscous Row. The arrows indicate damage on fiber surfaces. exposure, since at this temperature range the matrix softens, leading to microstructural damage, This includes porosity formation and fiber debonding and rearrangement, that results in the loss of mechanical properties, particularly in very low fracture strength. Short excursions at these temperatures may be possible, however. The three-point flexure chevron-notch technique used was adequate to determine fracture toughness and work of fracture values, the push-out test to determine the interfacial properties, in these materials. The chevron￾notch test measurements for different aging conditions were consistent with the results ofinterfacial properties determined by the push-out indentation method, and with microscopic observations of fracture surfaces of as-received and a@d samples. The K,, values determined using the chevron-notch technique (in the range 19-26 MPaJm) are comparable to data obtained in similar composite materials using othertech￾niques [ 161. Acknowledgements The authors gratefully acknowledge Dr. W. Beier and Mr. R. Liebald of Schott Glaswerke, Mainz, Germany, for supplying the composite samples. The experimental assis￾tance of Mr. H.P. Feuz during the push-out indentation tests and of Dr. M. Reinisch during fracture toughness tests is appreciated. ARB acknowledges the support of the NSFInsti￾tute for Mechanics and Materials, University of California, San Diego. References [ I] K.K. Chawia. Ceramic Matrix Composites. Chapman and Hall, Lon￾don, 1993. [2] A.G. Evans, F.W. Zok, R.M. McMeeking, Z.Z. Du, Models oi hiih￾temperature. environmentally assisted rmbrittlemcnt in ceramic￾matrix composites, J. Am. &ram. Sot. 79 ( 1996) 734552352

A R. Boccac'rint er af/ Materials Chemistry and Physics 53(1998)155-164 3]K M. Prewo, J.J. Brennan, High-strength silicon carbide fibre-rei [25] R. Liebald, Schutt Glaswerke, Mainz. Germany, personal commun forced glass-matrix composites J Mater. Sci. 15(1980)463-468 [4] K M. Prewo, Tension and flexural strength of silicon carbide fiber. [26] T.E. Steyer, F W. Zok, D P. Walls, Stress rupture of an enhanced reinforced glass ceramics, J. Mater. Sci. 21(1986)3590-3600 Nicalon/Sic composite at intermediate temperatures, J.Am [5] H. Hegeler. R. Bruckner, Fibre reinforced glasses. J Mater. Sci. 24 Ceram Soc. (1997) submitted for publication (1989)1191-1194 [271 A R. Boccaccini, A., Strull, D.H. Pearce, K.S. Vecchio. Thermal [6]W. Pannhorst, M. Spallek, R Bruckner, et aL. Fiber-reinforced glasses shock and thermal cycling behaviour of fiber reinforced glass matrix and glass-teramics fabricated by a novel proress, Ceram. Eng S posites. Proceedings of the 29th International SAMPE Technic Proe.l1(1990)947-963 Conference. Oct. 29-Nov. 1, 1997, Orlando, FL. Pp 207-219 71 H.H. Shin, R.F. Speyer, Effect of hot-pressing time and posl-heat [28] J.s.Ha, K.K. Chawla. Mechanical behaviour of mullite composites treatment on the microstructure and mechanical properties of SiC- inforced with mullite fibres, Mater. Sci. Eng. A203( 1995)127 L760. 30-3636 I291A. R. Boccaccini, D H, Pearce, J Janczak. W. Beier, C.B. Ponton [8I B.L. Metcalfe, I W. Donald, DJ. Bradley, Preparation and properties Investigation of the cyclic thermal shock behaviour of fibre reinforced of a SiC fibre- reinforced glass-ceramic matrix composite, Br.Ceram glass matrix composites using a non-destructive forced 92(1993 technique. Mater. Sci. Technol, 13( 1997)852-859 [91 S Sutherland, K P Plucknett, M.H. Lewis. High t [30] L Dlouhy, L. Valka, J. Man. M. Holzmann. Experiences with tracture ical and thermai stability of silicate matrix composites. Comp. Eng. 5 toughness detcranination using chevron notch three point bend spec (1995)1367-1378 en, Proceedings of the Euromat 94, vol. 4. Balatonszeplak. 1994. [101 P. M. Benson. K E. Spear. G.C. Pantano. Interfacial characterization pp.1206-121l of glass matrix/nicalon SiC fibre composites: a thermodinamic [31 J.L. Bluhm. Slice synthesis of a three dimensional"work of fracture approach. Ceram. Eng. Sci. Proc. 9(1988)663-670 pecimen". Engng. Fract. Mech. 7(1975)593-60 LIl K P Plucknett, R, L, Cain, M H. Lewis, Interface degradation in CAS [32] W. Silang-Xiall. Fraclure toughness determination of bearing steel Nicalon during clevated temperature aging. Mater, Res. Soc. Symp ing Chevron-notched three point bending specimen. Engng. Fract. Proc.365(1995}421-126 Mech.I9(1984)221-232 12 A M. Daniel. A. Martin-Mcizoso, K P, Plucknett. D N. Braski. Inter- [33]P A. withey, R L BretL, P. Bowen. Use of chevron notches for fracture face moditication during oxidation of a glass-ceramic matrix/SiC hess determination in brittle solids, Mater. Sci. Technol. 8 re composite. Ceram. Eng, Sci, Proc. 17(1996)280-287. [131 G. West. D M.R. Taplin, A R. Boceatcini, K Plucknett, M.H. Lewis [341 I Dlouhy, M. Holzmann, J. Man, L, valka, The use of chevron notched leclianical behaviour and environnental stability of continuous nibre imen for fracture toughness determination, Metal. Mater. 32 reinforced glass-ceramic matrix composites. Int. J. Glass Sci. Tech Glastech. Ber. 69(1996)31-13 [35] J. Janczak, G. Buerki, L Rohr, Interfacial characterization of MMCs [14]S M. Blcay, V D Scott, Effect of heat treatment in air on the structure and CMCs using a SEM-Pushout technique. Key Eng Mater, 127- and properties of barium osumilite reinforced with Nicalon fibres.J 131(1997)623-630 Mater.Scl.27(19y2)825-838 [36] PJ. Lamicq, G.A. Bernhart, M.M. Daucher. J.G. Mace, Sic/Siccom- [151 E.Y. Sun, H.T. Lin J.J. Brennan. Intermediate-temperature environ posite ceramics, Ceram. Bull. 65(1986)336-338. ental effects on boron nitride-coated silicon carbide-tiber-reinforced [37] D.B. Marshall, An indentation method for measuring matrix-fiber glass-ceramic composites, J. Am. Ceram Soc. 80(1997)609-614. frictional stresses in ceramic composites. J Am Ceram Soc. 67(12) (1984)C259-C-260 [16] JJ. Brennan, K.M. Prewo, Silicon carbide fibre reinforced glass- [381 A Hahnel, E. Pippel, J Woltersdorf, Nanostructure of interlayers in ceramic matrix composites exhibiting high strength and toughness, J Mater.Sci.17(1982)2371-2383 different Nicalon fibre/glass and their effect [17] M W. Pharaoh, A M. Daniel, M H. Lewis. Stability of interfaces mechanical properties, J Microsc. I77(1995)264-271 calcium aluminosilicate matrix/Nicalon SiC fibre composite, JMater [391 J.J. Mecholsky, Evaluation of mechanical property testing method Sci.Let.12(1993}998-1001 for ceramic matrix composites Ceram. Bull. 65( 1986)315-3 [181 S. Kooner, C W. Lawrence. B Derby, High ceram. Eng. Sci rature interfacial [401 E. Kasprzuk, Y, Barbaux. The fracture behaviour of Sic-reinforced ear strength testing of ceramic matrix composite SiATYON ceramic matrix composites, J. de Physique IV 3 (1993) 1923-1926 Hroc.14(1993)29-236 [19] S.M. Bleay, V D. Scott, B. Harris, R G. Cooke. F.A. Habib, Interface [ I. Dlouhy, M. Rheinish, A.R. Boccaccini. J F. Knott. Fracture cha aracterizution and fructure of calcium aluminosilicate glass acteristics of borosilicate glasses reinforced by ductile metallic parti ceramic rcinforced with Nicalon fibres. J Mater. Sci. 27(1992)281 cles, Fatigue Fract. Engng Mater. Struct. 20(1997)1235-1253 [421 J. Mencik. Strength and Fracture of Glass and Ceramics (in Czech) SNTL Prague [990 [20] M.J. Blisset, P.A. Smith. J.A. Yeomans. Thermal shock behaviour of 43N. Takeda. O. Chen. T. Kishi, W. T .Mate.Sci.32(1997)317-325 emission characterization of the fracture mechanism of a high cor pliant glass-matrix composite Eng. Fract. Mech. 40( 1991)791 21] H.S. Kim. J.A. Yong. R D. Rawlings. P.S. Rogers Interfacial behavior of fibre reinforced glass ceramic composite at elevated temperature Mater. Sci. Technol. 7(1991)155-157 [4+]K Kageyama, M. Enoki, T. Kishi, Effect of microfracture process on [22]A R. Boccaccini, G. West, J Janczak. M H. Lewis H. Kern. Tensile fracture toughness in fine particulate glass composites. Proceedings of the 3rd Japan International SAMPE Symposium, 1993, pp. 2081 ehavior and cyclic creep of continuous fiber-reinforced glass matrix 086 composites at room and elevated temperatures, J. Mater Eng Perfo [45] T. Cutard M. Huger, D. Fargeot, Ch. Gault, High-temperature ultra- 6(1997)344-348 ic characterization of the mechanical and microstructural behavior [23]I. Crivelli- Visconti, G A. Cooper, Mechanical properties of a new arbon fibre material, Naturc 221(1969)754-75 glass-ceramic matrix, J. Am. Ceram Soc. 79(1996)957-96 24]K M. Prewo, B. Johnson. S. Starrett, Silicon carbide-fibre-reinforced [461 Ph. Schreck. C. Vix-Guterl, P. Ehrburger, J. Lahaye. Reactivity and glass-ceramic composite tensile behaviour at elevated temperature, Mlecular structure of silicon carbide fibres derived from polycarbon Mater.si.24(1989I373-1379 anes,J, Mater.Sci27(1992)4237-4242

[31 141 I51 [61 [71 [81 [ 101 [III 1121 1131 [I41 [ 151 [ 161 [ 171 [ 181 1191 [201 [211 [221 ~231 [241 KM. Prewo, J.J. Brennan, High-atrcngth silicon carbide hbre-rein￾forced glass-matrix composites, J. Mater, Sci. 15 ( 1980) 463-468. KM. Prewo, Tension and Hexural strength of silicon carbide tibcr￾reinforced glass ceramics, J. Mater. Sci. 2 I ( 1986) 3590-3600. H. Hegelcr, R. Bruckner. Fibrc reinforced glasses, J. Mater. Sci. 21 (1989) 1191-1191. W. Pannhorst, M. Spallck, R. Bruckner. et al.. Fiber-reinforccdglasses and glnss+cramics fabricated by a novel process, Ceram. Eng. Sci. Proc. 11 t 1990) 947-963. H.H. Shin, R.F. Speyer, Effect of hot-pressing time and post-heat treatment on the microstructure and mechanical properties of SiC￾tibre-reinforced glaaa-ccramic composites. J. Mater. Sci. 29 i 1991) 3630-3636. B.L. Metcalfc, I.W. Donald, D.J. Bradley, Preparation and properties of a Sic libre-reinforccd glass-ceramic matrix composite, Br. Ceram. Trans. 92 (1993) 13-20. S. Sutherland, K.P. Plucknett, M.H. Lewis. High temperature mechan￾ical and thermal stability of silicate matrix composites. Camp. Eng. 5 (1995) 1367-1378. P.M. Benson, K.E. Spear, G.C. Pantano. Interfacial characterization of glass mntrixinicnlon SIC fibre composites: a thermodinamic approach. Ceram. Eng. Sci. Proc. 9 ( 1988) 663-670. K.P. Plucknett, R.L. Cain, M.H. Lewis, Interface degradation in CAS/ Nicalon during elevated temperature aging, Mater. Rcs. Sot. Symp. Proc. 365 ( 1995 ) 42 1126. A.M. Daniel, A. Martin-Mcizoso, K.P. Plucknctt, D.N. Braski, Inter￾face moditication during oxidation of a glass-ceramic matrix/Sic hbre composite, Ccram. Eng. Sci. Proc. I7 ( 1996) 280-287. G. West, D.M.R. Taplin, A.R. Boccaccini, K. Plucknett. 1M.H. Lewis. Mechanical behaviour and environmental stability of continuous hbre reinforced glans-ceramic matrix composites, Int. J. Glass Sci. Tech. Glastech. Ber. 69 ( 1996) 3U3. SM. Bleay, V.D. Scott, Effect of heat treatment in air on the structure and properties of barium osumiiite reinforced with Nicalon fibrrs. .I. Mater. Sci. 27 ( 1992) 825-838. E.Y. Sun, H.T. Lin. J.J. Brcnnan, Intermediate-temperature environ￾mental effects on boron nitride-coated silicon carbide-fiber-reinforced glass-ceramic composites, J. Am. Ccram. Sot. 80 ( 1997) 609-611. J.J. Brennnn, K.M. Prewo, Silicon carbide librc reinforced glass￾ceramic matrix composites exhibiting high strength and toughness, J. Mater. Sci. 17 ( 1982) 237 l-2383. M.W. Pharaoh, A.M. Daniel, M.H. Lewis. Stability of interfaces in calcium nluminosilicatc matrix/Nicalon Sic libre composite, J. Mater. Sci. Lett. 12 (1993) 998-1001. S. Kooner, C.W. Lawrence, B. Derby, High temperature interfacial shear strength testing of ceramic matrix composites, Ceram. Eng. Sci. Proc. 14 ( 1993) 229-236. S.M. Bleay, V.D. Scott, B. Harris, R.G. Cooke, F.A. Habib, Interface characterization and fracture of calcium aluminosilicate glass￾ceramic reinforced with Nicnlon tibrea, J. Muter. Sci. 27 ( 1992) 28 1 i- 2822. M.J. Blisset, P.A. Smith, J.A. Yeomans, Thermal shock brhaviour of unidirectional silicon carbide tibre reinforced calcium aluminoailicate, J. Mater. Sci. 32 ( 1997) 317-325. H.S. Kim, J.A. Yong. R.D. Rawlings, P.S. Rogers, Interfacial behavior of tibre reinforced glass ceramic composite at elevated temperature, Mater. Sci. Tcchnol. 7 t 1991) 155-157. A.R. Boccaccini. G. West, J. Jancznk. M.H. Lewis, H. Kern, Tensile behavior and cyclic creep of continuous fiber-reinforced glass matrix composites at room and elevated temperatures. J. Mater. Eng. Perfor. 6 (1997) 344-348. I. Crivelli-Visconti, G.A. Cooper, Mechanical properties of a new carbon fibre material, Nature 221 ( 1969) 754-755. K.M. Prewo, B. Johnson. S. Starrett, Silicon carbide-Abre-reinforced glass-ceramic composite tensile behaviour at elevated temperature, J. Mater. Sci. 21 ( 1989) 1373-1379. [251 [261 1271 [?81 [291 1301 [311 [32] [331 [341 [351 [361 1371 1381 [391 1401 i4ll [Ql [431 [4-+1 [45] I361 R. Liebald, Schott Glaswerke, Mainz, Germany, personal communi￾cation, 1997. T.E. Steyer, F.W. Zok. D.P. Walls, Stress rupture of an enhanced NicalonT”/SiC composite at intermediate temperatures, J. Am. &ram. Sot. ( 1997) submitted for publication, A.R. Boccaccini, A.J. Strutt. D.H. Pearce. KS. Vecchio, Thermal shock and thermal cycling behaviour of tiber reinforced glass matrix composites, Proceedings of the 29th International SAMPE Technical Conference. Oct. 29-Nov. I. 1997, Orlando. Fl. pp. 207-219. J.S. Ha, K.K. Chawla. Mechanical bchaviour of mullite composites reinforced with mullite tibres. Mater. Sci. Eng. A203 ( 1995) 1271- 1760. A.R. Boccaccini, D.H. Pearce. J. Janczak, W. Beier, C.B. Ponton, Investigation of the cyclic thermal shock behaviour of tibre reinforced glass matrix composites using a non-destructive forced resonance technique. Mater. Sci. Technol. 13 ( 1997) 852-859. I. Dlouhy, L. VQlka, J. Man, IM. Holzmann, Experiences with fracture toughness determination using chevron notch three point bend spec￾imen, Proceedings of the Euromat.94, vol. 4. Balatonszeplak, 1994, pp. 1206-121 I. J.I. Bluhm, Slice synthesis of a three dimensional “work of fracture specimen”, Engng. Fmct. Mech. 7 ( 1975) 593-604. W Shang-Xian, Fracture toughness determination of bearing steel using Chevron-notched three point bending specimen. Engng. Fract. Mech. 19 (1984) 221-232. P.A. Withey, R.L. Brett, P. Bowen, Use of chevron notches for fracture toughness determination in brittle solids, Mater. Sci. Technol. 8 (1992) 805-809. I. Dlouhp. M. Holzmann, J. Man, L. Valka, The use ofchevron notched specimen for fracture toughness determination, Metal. Mater. 32 (1994) 3-13. J. Janczak. G. Buerki. L. Rohr, Interfacial characterization of MMCs and CMCs using a SEM-Pushout technique. Key Eng. Mater. 127- 131 (1997) 623-630. P.J. Lamicq, G.A. Bemhart, MM. Daucher. J.G. Mace, SiC/SiCcon￾posite ceramics. Ceram. Bull. 65 ( 1986) 336-338. D.B. Marshall. An indentation method for measuring matrix-fiber frictional stresses in ceramic composites, J. Am. Ceram. Sot. 67 i 12) (1984)C-259-C-260. A. Hhhnel. E. Pippel, J. Woltersdorf. Nanostructure of interlayers in different Nicalon fibre/glass matrix composites and their effect on mechanical properties, J. Microsc. 177 ( 1995) 264-271. J.J. Mecholsky, Evaluation of mechanical property testing methods for ceramic matrix composites, Ceram. Bull. 65 ( 1986) 315-322. E. Kasprzak, Y. Barbaux, The fracture behaviour of Sic-reinforced SiAlYON ceramic matrix composites, J. de Physique IV 3 ( 1993) 1923-1926 I. Dlouhy, M. Rheinish, A.R. Boccaccini, J.F. Knott. Fracture char￾acteristics of borosilicate glasses reinforced bg ductile metallic parti￾cles. Fatigue Fract. Engng Mater. Stntct. 20 ( 1997) 1235-1253. J. Mencik. Strength and Fracture of Glass and Ceramics (in Czech), SNTL Prague, 1990. N. Takeda. 0. Chen, T. Kishi, W. Tredway, K. Prewo, Acoustic emission characterization of the fracture mechanism of a high com￾pliant glass-matrix composite , Eng. Fract. Mech. 10 (1991) 791- 799. K. Kageyamn. M. Enoki: T. Kishi. Effect of microfracture process on fracture toughness in fine particulate glass composites, Proceedings of the 3rd Japan International SAMPE Symposium, 1993, pp. 2081- 2086. T. Cutard, M. Huger, D. Fargeot, Ch. Gault, High-temperature ultra￾sonic characterization of the mechanical and microstructural behavior of a fibrous composite with a magnesium lithium aluminum silicate glass-ceramic matrix, J. Am. Ceram. Sot. 79 ( 1996) 957-96 1. Ph. Schreck. C. Vix-Guterl, P. Ehrburger, J. Lahaye. Reactivity and molecular structure of silicon carbide fibres derived from polycarbos￾ilanes, .J. Mater. Sci. 27 ( 1992) 42374242

64 A.R. Boccaccini et al. /Materiais Chemistr and PhysIcs 53(1998)155-164 (47) T Mah N L Hecht. D.E. McCullum, J. R. Hoenigman, H M. Kim, [57] T. Clark, R. M. Arons, J B. Stamatoff, Thermal degradation of Nica A.P. Katz. H A. Lipsitt, Thermal stability of SiC fibers(Nicalon), J lon SiC fibers, Ceram. Eng. Sci. Proc. 6(1985)576-588 Mater.Sci.19(1984)119|-120 [581 C. Labrugere. A. Guette R. Naslain, Effect of ageing treatments 48 P Le Coustumer, M. Monthioux, A Oberlin, Understanding nicalon he microstrueture and mechanical behaviour of fiber, J. Eur. Ceram Soc. 11(1993)95-103 D Nicalon/C/SiC composites. 1: Ageing under vacuum or argon, J. 491 T. Yamamura, T. Hurushima, M. Kimoto, T. Ishikawa, M. Shibuya, Eur. Ceram.Soc,17(1997)623-610 T Iwai. Development of new continuous Si-Ti-C-O Fiber with high 159] P, Manns, R. Bruckner. Eiastischen Verhalten hochviskoser Glassch melzen mit Bezug auf dic Heipbiegefestigkeit, Glastech. Ber. 59 Tech Ceramics, Elsevier, Amsterdam, 1987. pp. 737-746 0J G. Simon, A.R. Bunsell. Mechanical and structural characterization 60] L F. Johnson, D P H Hasselman E Minford, Thermal diffusivity and of the Nicalon silicon carbide libre. J Mater. Sci. 19(1984)3649- conductivity oI a carbon hbre-reintorced borosilicate glass, J. Mat 3657 22(1987)311-3117 [51] H.E. Kill, A J Moorhead, Stength uf Nicalon silicon carbide fibt [61 KM. Prewo, J.A. Batt, The oxidative stability of carbon fibre rein sed to high-temperature gaseous environments, J. Am. Ceram. J. Mater.Sc.33(1988)523-527 [62]KM. Prewo, A compliant. high failure strain fibre-reinforced glass [52] C. Vahlas, P. Rocabois C Bernard, Thermal degradati matrix composite. J. Mater. Sci. 17(1982)3549-350 of Nicalon ibre: thermodynamic simulation, J Mater. Sci. 29(1994) [63]SR. Levitt, High strength graphite fibre/lithium aluminosilicate com- 839-5846. J. Mater.Sci.8(1973)793-806 [53 D.J. Physher, K.C. Goretta, R.S. Hodder Jr, R.E. Tressler Jr [64] A.S. Fareed, M.J. Koczak F Ko Fracture of SiC/LAS ceramic com- Strengths of ceramic fibers at elevated temperatures, J. Am. Ceram posites, in: V, D Frechette. J.R. Varner( Eds ) Advances in Ceramics, Soc.72(1989)284-288 vol. 22: Fractography of Glasses and Ceramics, The American Ceramic Society, Ohio, 1988, pp. 261-278 [54] M.D. Thouless, O. Sbaizero, L.S. Sigl, A.G. Evans, Effect of interface 165] D Singh, J.P. Singh, M.J. Wheeler, Mechanical behaviour of SiC(C)/ anical properties on pullout in a SiC-fiber reinforced lithium aluminum silicate glass-ceramic, J Am Ceram Soc. 72(1989)525- SiC composites and correlation to in situ fiber strength at room and elevated temperature, I. Am. Ceram. Soc. 79(1996)391-596. [55] R. Bodet, J. Lamon. N Jia, R E. Tressler. Microstructural stability and [66]wA. Curtin, Theory of mechanical properties of ceramic-matrix reep behaviour of Si-C-o( Nicalon fibers in carbon monoxide and composites. J Am Ceram Soc. 74(1991)2837-2845 argon environments, J. Am. Ceram. Soc. 79(1996)2673-2686 [671J. Wu, F R. Jones, P F. James, Continuous fibre reinforced mullin matrix composites by sol-gel processing, J. Mater. Sci. 32(1997) 1561 G. Chollon, R. Pailler, R. Naslain, F. Laanani, M. Monthioux, P Olry 3629-3635 T hermal stability of a PCS-derived SiC fibre with a low oxygen cor [68]L M. Sheppard, Enhancing the performance of ceramic composites. ( Hi-Nicalon), J Mater. Sci. 32(1997)327-347. Ceran.Bull71(1992)6i7-631

[501 1511 1521 [531 1541 1551 1561 T. Mah. N.L. Hecht. D.E. McCullum, J.R. Hoenigman, H.M. Kim, A.P. Katz. H.A. Lipsitt, Thermal stability of Sic fibers (Nicalon), J. Mater. Sci. 19 (1984) 1191-1201. P. Le Coustumer, M. Monthioux, A. Oberlin, Understanding Nicalon tiber, J. Eur. Ceram. Sot. 11 ( 1993) 95-103. T. Yamamura, T. Hurushima, M. Kimoto, T. Ishikawa, M. Shibuya, T. Iwai, Development of new continuous Si-T&C-O Fiber with high mechanical strength and heat-resistance, in: P. Vincenzini (Ed.), High Tech Ceramics, Elsevier, Amsterdam, 1987, pp. 737-746. G. Simon, A.R. Bunsell. Mechanical and structural characterization of the Nicalon silicon carbide Libre, .I. Mater. Sci. 19 ( 1984) 3619- 3657. H.E. Kim, A.J. Moorhead, Strength of Nicalon silicon carbide iibers exposed to high-temperature gaseous environments, J. Am. Ceram. Soc.7-i 11991)666-669. C. Vahlas, P. Rocabois. C. Bernard, Thermal degradation mechanisms of Nicalon tibre: thermodynamic simulation, J. Mater. Sci. 29 ( 1994) 5839-5846. D.J. Physher, K.C. Goretta, R.S. Hoddcr Jr., R.E. Tressler Jr., Strength& of ceramic fiberb at elevated temperatures, J. Am. Ceram. Sot. 72 ( 19S9) 284-288. M.D. Thouless; 0. Sbaizero. L.S. Sigi, A.G. Evans, Effect of inierpdace mechanical properties on pullout in a Sic-fiber reinforced lithium aluminum silicate glass-ceramic, J. Am. Ceram. Sot. 72 ( i 989) 525- 532. R. Bodet, J. Lamon, N. Jia, R.E. Tressler, Microstrucrural stability and creep behaviour of Si-C-O (Nicalon) fiber5 in carbon monoxide and argon environments, J. Am. Ceram. Sot. 79 ( 1996) 2673-2686. G. Chollon, R. Pailler, R. Naslain. F. Laanani, M. Monrhioux. P. Olry, Thermal stabiliryofaPCS-derived Sic fibrewithalow oxygencontent I Hi-Nicalon), J. Mater. Sci. 32 i 1997) 327-347. 1571 YSSl 1601 1611 [Ql [631 [441 Lb51 [661 [671 1681 T.J. Clark, R.M. Arons, J.B. Stamaroff, Thermal degradation of Nica￾ion Sic fibers, Ceram. Eng. Sci. Proc. 6 ( 1985) 576-588. C. Labrugere, A. Guette. R. Naslain, Effect of ageing trratments at high temperatures on the microstructure and mechanical behaviourof 2D NicalonlCiSiC composites. 1: Ageing under vacuum or argon, f. Eur. Ceram. Sot. 17 ( 1997) 633-640. P. Manns. R. Briickner, Elartischen Verhalten hochvihkow Glassch￾melzen mit Bezug auf die Heipbiegefestigkeit, Glabtech. Be:. 59 (1986) 77-84. L.F. Johnson, D.P.H. Hassehnan, E. Minford. Thermal diffusivity and conductivity of a carbon Iibre-reinforced borosilicate glass, J. Mater. Sci.32(1987)3111-3117. K.M. Prewo, J.A. Batt. The oxidative stability of carbon fibre rein￾forced glass-matrix compositea, J. Mater. Sci. 23 ( 1988) 523-527. K.M. Prewo, A compliant, high failure strain tibre-reinforced glass matrix composite, J. Mater. Sci. 17 ( 1982) 3549-3563. S.R. Levitt, High strength graphite Iibreilithium aluminosilicate com￾posites, J. Mater. Sci. 8 ( 1973) 793-806. A.S. Fareed, M.J. Koczak, F. Ko, Fracture of Sic/LAS ceramic cam￾posites, in: V.D. Frechette, J.R. Vamer (Eds.), Advances inceramics, vol. 22: Fractography of Glasses and Ceramics, The American Ceramic Society, Ohio, 1988, pp. 26 l-278. D. Sin&h. J.P. Singh, M.J. Wheeler, MechanicZibehaviourofSiC( fj/ Sic composites and correlation to in situ fiber strength al room and elevated temperature, J. Am. Cemm. Sac. 79 j1996) 591-596. W.A. Curtin, Theory of mechanical properties of ceramic-m&x composires, J. Atn. Ceram. Sot. 74 (1991) 2837-2815. J. Wu, F.R. Joneb, P.F. James, Continuous libre reinforced mullite matrix composites by sol-gel processing, J. Mater. Sci. 32 ( 19912 3629-3635. L.M. Sheppard, Enhancing the performance of ceramic compo&, Ceram. Bull. 71 ( 1992) 617-631