Availableonlineatwww.sciencedirect.com DIRECT. COMPOSITES CIENCE AND TECHNOLOGY ELSEVIER Composites Science and Technology 66(2006)2089-2099 www.elsevier.com/locate/compscitech Creep behavior of NextelM610/Monazite/ Alumina composite at elevated temperatures M B. Ruggles-Wrenn ,S.s. Musil, s Mall, K.A. Keller Department of Aeronautics and Astronautics, Air Force Institute of Technology, Wright-Patterson Air Force Base, OH 45433-7765, USA Received 7 June 2005: received in revised form 8 December 2005: accepted 22 December 2005 Available online 6 March 2006 Abstract The creep behavior of a N610/LaPO/Al,O3 composite was evaluated in this work. The composite consists of a porous alumina matrix reinforced with Nextel 610 fibers coated with monazite in a symmetric cross-ply(0%/90/0 90)s orientation. The te stress-strain behavior was investigated and the tensile properties measured at room temperature and in the 900-1200oC range the a tion of monazite coating resulted in 50% improvement in ultimate tensile strength (UTS) at temperatures <1100C, and in improvement in UTS at 1200C. Tensile creep behavior was examined at temperatures in the 900-1100C range for creep stresses rang- ing from 40 to 150 MPa. Primary and secondary creep s were observed in creep tests at 900C. At temperatures above 900C, the composite exhibited primary, secondary and tertiary creep. Minimum creep rate was reached in all tests. Creep rates accelerated with increasing temperature and creep stress. At 900C creep run-out, defined as 100 h at creep stress, was achieved for stress levels <120 MPa. The residual strength and modulus of all specimens that achieved run-out were characterized. Comparison with results obtained for N610/Al2O3(control) specimens revealed that the use of the monazite coating resulted in improved creep resistance at 900C Creep performance deteriorated rapidly as temperatures increased above 900 C. Composite microstructure, as well as damage and failure mechanisms were investigated e 2006 Elsevier Ltd. All rights reserved Keywords: A. Ceramic-matrix composites(CMCs): A Oxides; A. Fibres; A. Coatings: B Creep; B. High-temperature properties; B. Mechanical prop- erties: D. Fractography 1. Introduction and fracture toughness at high temperatures and they therefore continue to attract attention as candidate materi- The severe conditions encountered by aerospace compo als for such applications. Additionally, the lower densities nents require structural materials that have superior long- of CMCs and their higher use temperatures, along with a term mechanical properties and retained properties under reduced need for cooling air, allow for improved high-tem high temperature, high pressure, and varying environmen- perature performance when compared to conventional tal factors, such as moisture [1]. Ceramic-matrix compos- nickel-based superalloys [2]. Alternately, it is recognized ites(CMCs) are capable of maintaining excellent strength that the thermodynamic stability and oxidation resistance of CMCs are vital issues The views expressed are those of the authors and do not reflect the Non-oxide fiber/non-oxide matrix CMCs such as SiC/ official policy or position of the United States Air Force, Department of SiC generally show poor oxidation resistance [3, 4, partic Defense or the US government larly at intermediate temperatures(800C). The degrad Corresponding author. Tel: +1937 255 3636x4641: fax: +1937 656 tion involves the oxidation of the fibers fiber coatings and 4032 aruggles -wrenn(@afit. edu (M. B. Ruggles. matrices and is typically accelerated by the presence of moisture [5-11]. Using a non-oxide fiber/oxide matrix or I Under USAF Contract F33615-01-C-5214 an oxide fiber/non-oxide matrix composite generally does 02663538/S. see front matter 2006 Elsevier Ltd. All rights reserved doi:10.1016j.compscitech.2005.12.026
Creep behavior of NextelTM610/Monazite/Alumina composite at elevated temperatures q M.B. Ruggles-Wrenn a,*, S.S. Musil a , S. Mall a , K.A. Keller b,1 a Department of Aeronautics and Astronautics, Air Force Institute of Technology, Wright-Patterson Air Force Base, OH 45433-7765, USA b UES, Inc., 4401 Dayton Xenia Road, Dayton, OH 45433-7817, USA Received 7 June 2005; received in revised form 8 December 2005; accepted 22 December 2005 Available online 6 March 2006 Abstract The creep behavior of a N610/LaPO4/Al2O3 composite was evaluated in this work. The composite consists of a porous alumina matrix reinforced with Nextel 610 fibers coated with monazite in a symmetric cross-ply (0/90/0/90)S orientation. The tensile stress–strain behavior was investigated and the tensile properties measured at room temperature and in the 900–1200C range. The addition of monazite coating resulted in 50% improvement in ultimate tensile strength (UTS) at temperatures 61100 C, and in 37% improvement in UTS at 1200 C. Tensile creep behavior was examined at temperatures in the 900–1100 C range for creep stresses ranging from 40 to 150 MPa. Primary and secondary creep regimes were observed in creep tests at 900 C. At temperatures above 900 C, the composite exhibited primary, secondary and tertiary creep. Minimum creep rate was reached in all tests. Creep rates accelerated with increasing temperature and creep stress. At 900 C creep run-out, defined as 100 h at creep stress, was achieved for stress levels 6120 MPa. The residual strength and modulus of all specimens that achieved run-out were characterized. Comparison with results obtained for N610/Al2O3 (control) specimens revealed that the use of the monazite coating resulted in improved creep resistance at 900 C. Creep performance deteriorated rapidly as temperatures increased above 900 C. Composite microstructure, as well as damage and failure mechanisms were investigated. 2006 Elsevier Ltd. All rights reserved. Keywords: A. Ceramic-matrix composites (CMCs); A. Oxides; A. Fibres; A. Coatings; B. Creep; B. High-temperature properties; B. Mechanical properties; D. Fractography 1. Introduction The severe conditions encountered by aerospace components require structural materials that have superior longterm mechanical properties and retained properties under high temperature, high pressure, and varying environmental factors, such as moisture [1]. Ceramic–matrix composites (CMCs) are capable of maintaining excellent strength and fracture toughness at high temperatures and they therefore continue to attract attention as candidate materials for such applications. Additionally, the lower densities of CMCs and their higher use temperatures, along with a reduced need for cooling air, allow for improved high-temperature performance when compared to conventional nickel-based superalloys [2]. Alternately, it is recognized that the thermodynamic stability and oxidation resistance of CMCs are vital issues. Non-oxide fiber/non-oxide matrix CMCs such as SiC/ SiC generally show poor oxidation resistance [3,4], particularly at intermediate temperatures (800 C). The degradation involves the oxidation of the fibers, fiber coatings, and matrices and is typically accelerated by the presence of moisture [5–11]. Using a non-oxide fiber/oxide matrix or an oxide fiber/non-oxide matrix composite generally does 0266-3538/$ - see front matter 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.compscitech.2005.12.026 q The views expressed are those of the authors and do not reflect the official policy or position of the United States Air Force, Department of Defense or the US Government. * Corresponding author. Tel.: +1 937 255 3636x4641; fax: +1 937 656 4032. E-mail address: marina.ruggles-wrenn@afit.edu (M.B. RugglesWrenn). 1 Under USAF Contract # F33615-01-C-5214. www.elsevier.com/locate/compscitech Composites Science and Technology 66 (2006) 2089–2099 COMPOSITES SCIENCE AND TECHNOLOGY
M B. Ruggles-Wrenn et al /Composites Science and Technology 66(2006 )2089-2099 not substantially improve the high temperature oxidation objective of this study was to investigate the effectiveness resistance[12]. The need for environmentally stable com- of using monazite coatings in a crossply NextelM610/alu posites led to the development of CMCs based on oxide mina composite to improve the creep resistance of the constituents [13-21 CMC. Keller et al. [39] investigated the effectiveness of It is widely accepted that in order to avoid brittle frac- monazite coatings in Nextel 610/alumina composites after ture behavior in CMCs and improve the damage tolerance, long-term exposure at 1100 and 1200C Coated fiber sam- a weak fiber/matrix interface is needed, which serves to ples exhibited better tensile strength retention after 1000 h deflect matrix cracks and to allow subsequent fiber pull- at 1200C when compared to the control(uncoated fiber out [22-25]. A matrix consisting of finely distributed poros- material. This effort investigates the creep-rupture behavior o y can provide a similar crack-deflecting behavior instead of the Nextel 610/Monazite/Alumina(N610/M/A)com- of a separate interface between matrix and fibers [26]. In posite developed at the Air Force Research Laboratory this case, there is strong bonding between the fiber and(AFRL/MLLN), Materials and Manufacturing Director matrix; consequently, there is a minimum matrix porosity ate. Creep-rupture tests were conducted at temperatures needed for this concept to work [27]. For a dense (90%) in the 900-1 100C range for creep stresses ranging from matrix composite, however, an interfacial coating is needed 40 to 150 MPa Resulting creep performance imposes lim oxide coatings for oxide and non-oxide composites has applications. The composite microstructure, as wlle for crack deflection. An extensive review of the research on itations on the use of this material in high-temperature been given by Kerans et al. [28]. damage and failure mechanisms are discussed The development of oxide-oxide composites that rely on a weak fiber/matrix interface for crack deflection prompte 2. Material and experimental arrangements research into oxidation resistant fiber coatings that are hemically stable with the composite constituents. Various Nextel 610 fiber tows were desized in air at 1100C and oxidation-resistant coating materials have been investi- then coated with a monazite precursor solution at 1100oC gated, including monazite(LaPO4) and scheelite Of these, As described elsewhere [35] this monazite sol was washed a significant amount of work has been completed on com- multiple times to remove any residual ions, which have been posites containing monazite coatings. The initial works of associated with fiber strength loss. After coating, the fibers Morgan and Marshall [20, 29, 30] showed the potential of were filament wound using an alumina slurry consisting of a monazite as a weak interface material, particularly for alu- polymeric alumina sol and alumina powder (AKP-53 mina-based composites, due to the chemical compatibility Sumitomo Corp. ) In addition, uncoated fibers were wound of monazite with alumina at high temperature. Since then, for control samples. As described by Keller et al. [39], dur- numerous studies have examined the production of mona- ing the winding process, the number of wheel revolutions zite coatings and its use with different fiber/matrix combi- was recorded to determine the number of fiber tows within nations [31-35]. Fiber strength degradation caused by the a given tape width. This value was then used to calculate the coating and long-term, high-temperature heat treatments volume percentage of fibers in the composite. The resultant was identified as the key problem with monazite coatings unidirectional tapes were cut from the spool, stacked into a Boakye et al. [35] investigated the effects of different metal mold in an 8-layer symmetric cross-ply orientation of liquid precursors on coating characteristics and tensile [(0%/90%)]s, and warm vacuum-molded to form green cross- strength of coated fibers, and developed monazite coating ply composites. After drying, the billets were sintered at that did not cause significant loss of fiber strength 1200C for 5 h in air. During the heating cycle, a one hour Porous-matrix oxide/oxide CMCs exhibit several behav- hold at low temperature was performed to remove any or trends that are distinctly different from those exhibited residual organics in the matrix. Control billets, N610/Alu by traditional CMCs with a fiber-matrix interface. Most mina(N610/A), containing uncoated fibers were produced Sic-fiber-containing CMCs exhibit longer life under static with the same procedure. Billet properties, namely fiber loading and shorter life under cyclic loading [36]. For these volume fraction and density, are summarized in Table I materials, fatigue is significantly more damaging than Micrographs of the as-processed material shown in Fig. I creep. Zawada et al. [37] examined the high-temperature reveal shrinkage and sintering cracks that occurred during mechanical behavior of a porous matrix Nextel 610/alumi- the cooling stage of the composite processing. Extensive nosilicate composite Results revealed excellent fatigue per- surface microcracking is seen in Fig. 1(a), while Fig. I(b) formance at 1000C, the material exhibited high fatigue shows interlaminar matrix cracks. limit, long fatigue life and near 100% strength retention The N610/M/A specimens were cut from seven different Conversely, creep lives were short, indicating low creep billets, and N610/A specimens from three different billets. resistance and limiting the use of that CMC to tempera- Specimen numbers contain reference to the billet number. tures below 1000 For example, number Bl-I refers to the specimen I from Because creep is shown to be considerably more damag- billet 1. Billets were cut into flat rectangular coupons, which ing to porous-matrix oxide/oxide CMCs [37,38) creep test- were machined into dog bone-shaped specimens shown in ing is well suited for assessing the long term durability an nd Fig. 2. Diamond-grit grinding was used for billets Bl-B8 gh-temperature erformance of these materials. The and the abrasive water-jet machine, for billets B9-Bll
not substantially improve the high temperature oxidation resistance [12]. The need for environmentally stable composites led to the development of CMCs based on oxide constituents [13–21]. It is widely accepted that in order to avoid brittle fracture behavior in CMCs and improve the damage tolerance, a weak fiber/matrix interface is needed, which serves to deflect matrix cracks and to allow subsequent fiber pullout [22–25]. A matrix consisting of finely distributed porosity can provide a similar crack-deflecting behavior instead of a separate interface between matrix and fibers [26]. In this case, there is strong bonding between the fiber and matrix; consequently, there is a minimum matrix porosity needed for this concept to work [27]. For a dense (>90%) matrix composite, however, an interfacial coating is needed for crack deflection. An extensive review of the research on oxide coatings for oxide and non-oxide composites has been given by Kerans et al. [28]. The development of oxide–oxide composites that rely on a weak fiber/matrix interface for crack deflection prompted research into oxidation resistant fiber coatings that are chemically stable with the composite constituents. Various oxidation-resistant coating materials have been investigated, including monazite (LaPO4) and scheelite. Of these, a significant amount of work has been completed on composites containing monazite coatings. The initial works of Morgan and Marshall [20,29,30] showed the potential of monazite as a weak interface material, particularly for alumina-based composites, due to the chemical compatibility of monazite with alumina at high temperature. Since then, numerous studies have examined the production of monazite coatings and its use with different fiber/matrix combinations [31–35]. Fiber strength degradation caused by the coating and long-term, high-temperature heat treatments was identified as the key problem with monazite coatings [34]. Boakye et al. [35] investigated the effects of different liquid precursors on coating characteristics and tensile strength of coated fibers, and developed monazite coating that did not cause significant loss of fiber strength. Porous-matrix oxide/oxide CMCs exhibit several behavior trends that are distinctly different from those exhibited by traditional CMCs with a fiber–matrix interface. Most SiC-fiber-containing CMCs exhibit longer life under static loading and shorter life under cyclic loading [36]. For these materials, fatigue is significantly more damaging than creep. Zawada et al. [37] examined the high-temperature mechanical behavior of a porous matrix Nextel 610/aluminosilicate composite. Results revealed excellent fatigue performance at 1000 C, the material exhibited high fatigue limit, long fatigue life and near 100% strength retention. Conversely, creep lives were short, indicating low creep resistance and limiting the use of that CMC to temperatures below 1000 C. Because creep is shown to be considerably more damaging to porous-matrix oxide/oxide CMCs [37,38], creep testing is well suited for assessing the long term durability and high-temperature performance of these materials. The objective of this study was to investigate the effectiveness of using monazite coatings in a crossply NextelTM610/alumina composite to improve the creep resistance of the CMC. Keller et al. [39] investigated the effectiveness of monazite coatings in Nextel 610/alumina composites after long-term exposure at 1100 and 1200 C. Coated fiber samples exhibited better tensile strength retention after 1000 h at 1200 C when compared to the control (uncoated fiber) material. This effort investigates the creep-rupture behavior of the Nextel 610/Monazite/Alumina (N610/M/A) composite developed at the Air Force Research Laboratory (AFRL/MLLN), Materials and Manufacturing Directorate. Creep-rupture tests were conducted at temperatures in the 900–1100 C range for creep stresses ranging from 40 to 150 MPa. Resulting creep performance imposes limitations on the use of this material in high-temperature applications. The composite microstructure, as well as damage and failure mechanisms are discussed. 2. Material and experimental arrangements Nextel 610 fiber tows were desized in air at 1100 C and then coated with a monazite precursor solution at 1100 C. As described elsewhere [35], this monazite sol was washed multiple times to remove any residual ions, which have been associated with fiber strength loss. After coating, the fibers were filament wound using an alumina slurry consisting of a polymeric alumina sol and alumina powder (AKP-53, Sumitomo Corp.). In addition, uncoated fibers were wound for control samples. As described by Keller et al. [39], during the winding process, the number of wheel revolutions was recorded to determine the number of fiber tows within a given tape width. This value was then used to calculate the volume percentage of fibers in the composite. The resultant unidirectional tapes were cut from the spool, stacked into a metal mold in an 8-layer symmetric cross-ply orientation of [(0/90)]2S, and warm vacuum-molded to form green crossply composites. After drying, the billets were sintered at 1200 C for 5 h in air. During the heating cycle, a one hour hold at low temperature was performed to remove any residual organics in the matrix. Control billets, N610/Alumina (N610/A), containing uncoated fibers were produced with the same procedure. Billet properties, namely fiber volume fraction and density, are summarized in Table 1. Micrographs of the as-processed material shown in Fig. 1 reveal shrinkage and sintering cracks that occurred during the cooling stage of the composite processing. Extensive surface microcracking is seen in Fig. 1(a), while Fig. 1(b) shows interlaminar matrix cracks. The N610/M/A specimens were cut from seven different billets, and N610/A specimens from three different billets. Specimen numbers contain reference to the billet number. For example, number B1-1 refers to the specimen 1 from billet 1. Billets were cut into flat rectangular coupons, which were machined into dog bone-shaped specimens shown in Fig. 2. Diamond-grit grinding was used for billets B1–B8, and the abrasive water-jet machine, for billets B9–B11. 2090 M.B. Ruggles-Wrenn et al. / Composites Science and Technology 66 (2006) 2089–2099
M B. Ruggles-Wrenn et al/ Composites Science and Technology 66(2006)2089-2099 Table I 10.0 ummary of billet properties for the N610/M/A and the N610/A 36.0 fraction (a) Density (g/cc) MonazitelAlumina co …--1 B2 63.0 44. 9088 Fig. 2. Test specimen, dimensions in mm. B10 ment rate of 0.05 mm/s at 23. 900, 1000, 1100 and 1200oC /610/Almina composi Creep-rupture tests were conducted in load control in 595 accordance with Astm standard c 1337 at 900 and 1100C. Specimens were loaded to the creep stress level at the rate of 20 MPa/s. In each test, stress-strain data were recorded during the loading to the creep stress level and the A servocontrolled MTS mechanical testing actual creep period. Thus both total strain and creep strain equipped with hydraulic water-cooled collet grips could be calculated and examined. Creep run-out was pact single-zone resistance-heated furnace, and a defined as 100 h at a given creep stress. To determine the ture controller were used in all tests. An MTS TestStar II retained strength and modulus, specimens that achieved digital controller was employed for input signal generation run-out were subjected to tensile tests to failure at the tem- and data acquisition. Strain measurement was accom- perature of the creep test. One specimen was tested per test plished with an MTS high-temperature air-cooled uniaxial condition. The authors recognize that this is a limited set of extensometer. For elevated temperature testing, thermo- data. However, this scoping research serves to identify the couples were bonded to the specimens using alumina temperature range where the use of monazite coating cement( Zircar)to calibrate the furnace on a periodic basis. results in improved creep resistance. Furthermore, results The furnace controller(using a non-contacting thermocou- of this exploratory effort can be used to determine whether ple exposed to the ambient environment near the test spec- a more rigorous investigation of the effectiveness of mona imen) was adjusted to determine the power setting needed zite coating in this CMC or in a different material system to achieve the desired temperature of the test specimen. should be undertaken The determined power setting was then used in actual tests Fracture surfaces of failed specimens were gold coated 4. Results and discussion and examined using SEM(Model 360FE, Leica)as well as optical microscopy. 4. Monotonic tension 3. Test procedures The N610/M/A specimens were tested at 900, 1000,1100 and 1200C, and the N610/A specimens, at 23, 1100 and All tests were conducted in laboratory air environment. 1200C. Tensile test results are summarized in Table 2 In elevated temperature tests, a specimen was heated to the where elastic modulus, ultimate tensile strength (UTS) test temperature at a rate of lC/s, and held at tempera- and failure strain are presented for each test temperature. ture for additional 15 min prior to testing. Tensile tests The tensile stress-strain curves for both N610/A and were performed in stroke control with a constant displace- N610/M/A composites are shown in Fig 3 10 mm Fig. I. Micrographs of the as-processed material showing shrinkage cracks: (a) extensive surface microcracking and (b) interlaminar matrix cracks
A servocontrolled MTS mechanical testing machine equipped with hydraulic water-cooled collet grips, a compact single-zone resistance-heated furnace, and a temperature controller were used in all tests. An MTS TestStar II digital controller was employed for input signal generation and data acquisition. Strain measurement was accomplished with an MTS high-temperature air-cooled uniaxial extensometer. For elevated temperature testing, thermocouples were bonded to the specimens using alumina cement (Zircar) to calibrate the furnace on a periodic basis. The furnace controller (using a non-contacting thermocouple exposed to the ambient environment near the test specimen) was adjusted to determine the power setting needed to achieve the desired temperature of the test specimen. The determined power setting was then used in actual tests. Fracture surfaces of failed specimens were gold coated and examined using SEM (Model 360FE, Leica) as well as optical microscopy. 3. Test procedures All tests were conducted in laboratory air environment. In elevated temperature tests, a specimen was heated to the test temperature at a rate of 1 C/s, and held at temperature for additional 15 min prior to testing. Tensile tests were performed in stroke control with a constant displacement rate of 0.05 mm/s at 23, 900, 1000, 1100 and 1200 C. Creep-rupture tests were conducted in load control in accordance with ASTM standard C 1337 at 900 and 1100 C. Specimens were loaded to the creep stress level at the rate of 20 MPa/s. In each test, stress–strain data were recorded during the loading to the creep stress level and the actual creep period. Thus both total strain and creep strain could be calculated and examined. Creep run-out was defined as 100 h at a given creep stress. To determine the retained strength and modulus, specimens that achieved run-out were subjected to tensile tests to failure at the temperature of the creep test. One specimen was tested per test condition. The authors recognize that this is a limited set of data. However, this scoping research serves to identify the temperature range where the use of monazite coating results in improved creep resistance. Furthermore, results of this exploratory effort can be used to determine whether a more rigorous investigation of the effectiveness of monazite coating in this CMC or in a different material system should be undertaken. 4. Results and discussion 4.1. Monotonic tension The N610/M/A specimens were tested at 900, 1000, 1100 and 1200 C, and the N610/A specimens, at 23, 1100 and 1200 C. Tensile test results are summarized in Table 2, where elastic modulus, ultimate tensile strength (UTS), and failure strain are presented for each test temperature. The tensile stress–strain curves for both N610/A and N610/M/A composites are shown in Fig. 3. Fig. 1. Micrographs of the as-processed material showing shrinkage cracks: (a) extensive surface microcracking and (b) interlaminar matrix cracks. Table 1 Summary of billet properties for the N610/M/A and the N610/A composites Billet Fiber volume fraction (%) Density (g/cc) N610/Monazite/Alumina composite B1 44.3 2.55 B2 39.7 2.48 B4 40.2 2.49 B5 43.0 2.50 B9 44.7 2.58 B10 42.9 2.58 B11 41.0 2.53 N610/Alumina composite B3 51.6 2.95 B6 54.8 2.99 B8 46.7 2.85 R=50 36.0 63.0 8.0 10.0 5.0 Fig. 2. Test specimen, dimensions in mm. M.B. Ruggles-Wrenn et al. / Composites Science and Technology 66 (2006) 2089–2099 2091
M.B. Ruggles-Wrenn et al /Composites Science and Technology 66(2006)2089-2099 failure strain at 1200C is approximately 4.5 times that at ummary of tensile properties for the N610/M/A and the N6IO/A 1100 %( composites It is important to note that in all tension tests, as well as Specimen Temperature Elastic modulus UTS Failure strain in all other tests reported herein, the failure occurred within GPa) the gage section of the extensometer N10/ Monazite/Alumina composite 0.3l 4. 2. Creep rupture B2-1 15 1200 130 142 Results of the creep-rupture tests are summarized in N610/Almina composite Table 3, where creep strain accumulation and time to rup l17 0.09 ture are shown for each test temperature and applied stress 105 level. Creep curves obtained for the coated fiber composite at 1100C are presented in Fig. 4. Creep curves obtained at 900C for N610/M/A and N610/A are shown in Figs. 5(a) The stress-strain curves obtained for the uncoated fiber and(b), respectively. The time scale in Figs. 4(a)and 5 is composite at 23 and 1100C are nearly linear to failure. reduced in order to clearly show the creep curves produced all suggest a monolithic fracture behavior with no toughen- 900 C the uncoated fiber composite did not achieve a ing from fibers. For temperatures 75 MPa. and was therefore not The subsequent non-linear behavior, indicative of progres- tested at that temperature sive matrix cracking and crack deflection, continues until Creep curves produced by N610/M/A at 1100C exhibit failure. The fracture surfaces show considerable fiber pull primary, secondary and tertiary creep regimes. Transition decrease with increasing temperature, while the failure ately. For stresses 100 MPa, secondary creep coating results in a nearly 50% increase in UTS. However, transitions to tertiary creep during the first third of the this increase in strength is accompanied by a 34% decrease creep life. Creep strain accumulation decreases witl in modulus and a threefold increase in failure strain increasing applied stress. While at 40 MPa creep strain The stress-strain behavior of both composites changes a significant 7.66%, at 120 MPa creep strain is only 0.7% It is noteworthy that at 1100C, for stress levels in the gely nonlinear, the strength and modulus decrease signif 40-120 MPa range, the accumulated creep strains signifi- cantly,while failure strains increase. The increase in cantly exceed the failure strain obtained in the tension test temperature from 1100 to 1200C causes a 10% loss in The creep run-out, defined as survival of 100 h at a given UTS and a 57% loss in modulus for the uncoated fiber composite, and a 17 loss in UTS and a 34% loss in mod. creep stress, was not achieved ulus for the coated fiber composite. For both materials the Table 3 Summary of creep-rupture results for the N610/M/A and the N610/A composites Specimen Temperature stress Creep strain Time to 900c rupture(s Ng10/Monazite/Alumina cor 15023c 1100°c 7.66 1200°c B4-2 1100c 1.58 B4-3 l100 1200°c B9. 0.05 63.060 Bl0-1 0.04 522,3652 Bl0-2 432,1752 N610/Monazite/ Alumina B51 0.05 BIl-I l40 0.03 BIl-2 150 0.75 STRAIN (% N610/Almina composite B6-1 Fig. 3. Tensile stress-strain curves for N610/A and N610/M/A ceram B6-2 0.03 19,995 composites at various temperatures
The stress–strain curves obtained for the uncoated fiber composite at 23 and 1100 C are nearly linear to failure. The linear behavior, low UTS, and flat fracture surface all suggest a monolithic fracture behavior with no toughening from fibers. For temperatures 61100C and strains 60.1%, the stress–strain behavior of the N610/M/A composite is also nearly linear elastic. However, as the strain exceeds 0.1% the stress–strain curves depart from linearity. The subsequent non-linear behavior, indicative of progressive matrix cracking and crack deflection, continues until failure. The fracture surfaces show considerable fiber pull out. For both composites, the elastic modulus and UTS decrease with increasing temperature, while the failure strain remains fairly temperature-independent. Results in Table 2 reveal that at 1100 C, the addition of the monazite coating results in a nearly 50% increase in UTS. However, this increase in strength is accompanied by a 34% decrease in modulus and a threefold increase in failure strain. The stress–strain behavior of both composites changes dramatically at 1200 C. The stress–strain curves are largely nonlinear, the strength and modulus decrease signifi- cantly, while failure strains increase. The increase in temperature from 1100 to 1200 C causes a 10% loss in UTS and a 57% loss in modulus for the uncoated fiber composite, and a 17% loss in UTS and a 34% loss in modulus for the coated fiber composite. For both materials the failure strain at 1200 C is approximately 4.5 times that at 1100 C. It is important to note that in all tension tests, as well as in all other tests reported herein, the failure occurred within the gage section of the extensometer. 4.2. Creep rupture Results of the creep-rupture tests are summarized in Table 3, where creep strain accumulation and time to rupture are shown for each test temperature and applied stress level. Creep curves obtained for the coated fiber composite at 1100 C are presented in Fig. 4. Creep curves obtained at 900 C for N610/M/A and N610/A are shown in Figs. 5(a) and (b), respectively. The time scale in Figs. 4(a) and 5 is reduced in order to clearly show the creep curves produced at higher stress levels. Results in Table 3 reveal that at 900 C the uncoated fiber composite did not achieve a run-out even for a low stress of 73 MPa. Recognizing that increase in temperature would result in reduced creep life, the N610/A was not expected to achieve a run-out at 1100 C for stresses P75 MPa, and was therefore not tested at that temperature. Creep curves produced by N610/M/A at 1100 C exhibit primary, secondary and tertiary creep regimes. Transition from primary to secondary creep occurs almost immediately. For stresses 680 MPa, secondary creep persists for 70% of the creep life before transitioning to tertiary creep. However, for stresses P100 MPa, secondary creep transitions to tertiary creep during the first third of the creep life. Creep strain accumulation decreases with increasing applied stress. While at 40 MPa creep strain is a significant 7.66%, at 120 MPa creep strain is only 0.7%. It is noteworthy that at 1100 C, for stress levels in the 40–120 MPa range, the accumulated creep strains signifi- cantly exceed the failure strain obtained in the tension test. The creep run-out, defined as survival of 100 h at a given creep stress, was not achieved. Table 2 Summary of tensile properties for the N610/M/A and the N610/A composites Specimen Temperature (C) Elastic modulus (GPa) UTS (MPa) Failure strain (%) N610/Monazite/Alumina composite B9-1 900 83 180 0.31 B5-1 1000 78 162 0.28 B2-1 1100 76 157 0.34 B1-1 1200 50 130 1.42 N610/Alumina composite B3-1 23 129 117 0.09 B3-2 1100 116 105 0.11 B8-1 1200 49 95 0.46 0 50 100 150 200 0.00 0.25 0.50 0.75 1.00 1.25 1.50 STRAIN (%) STRESS (MPa) 1200˚C 1200˚C N610/Monazite/Alumina N610/Alumina 1100˚C 1100˚C 900˚C 23˚C 1000˚C Fig. 3. Tensile stress–strain curves for N610/A and N610/M/A ceramic composites at various temperatures. Table 3 Summary of creep-rupture results for the N610/M/A and the N610/A composites Specimen Temperature (C) Creep stress (MPa) Creep strain (%) Time to rupture (s) N610/Monazite/Alumina composite B4-1 1100 40 7.66 50,432 B4-2 1100 80 3.36 1,452 B2-2 1100 100 1.58 360 B4-3 1100 120 0.70 75 B9-2 1000 80 0.05 63,060 B10-1 900 80 0.04 522,365a B10-2 900 120 0.04 432,175a B5-1 900 130 0.05 40,655 B11-1 900 140 0.03 54,075 B11-2 900 150 0.03 805 N610/Alumina composite B6-1 900 73 0.06 350,055 B6-2 900 80 0.03 19,995 a Run-out. 2092 M.B. Ruggles-Wrenn et al. / Composites Science and Technology 66 (2006) 2089–2099
M.B. Ruggles-Wrenn et al. Composites Science and Technology 66(2006)2089-2099 0.25 N610/Monazite/Alur 100c N610/Monazite/Alumina e0.15 120 MPa 0.10 140 MPa 150 MPa 0.05 120 MPa 40 MPa 100 0.00 2000040000 80000100000 Time(s) 610/Alumina (b) R5.0 e0.1 80 MPa 方0.10 0.05 80 MPa 73 MPa 100002000030000400005000060000 0.00 4000060000 Time(s Fig 4. Creep strain vs time curves for the N610/M/A ceramic composite at 1100C: (a)time scale is reduced to clearly show creep curves obtained Fig. 5. Creep strain vs time curves at 900C for:(a)N610/Monazite/ at stress levels >100 MPa and(b) time scale chosen to show creep strai Alumina and(b) N610/Alumin ccumulated at 40 MPa ble differences between the creep strains accumulated at dif Because creep rupture of this CMC is likely dominated ferent stress levels together with the inherent data scatter by creep of Nextel 610 fibers, it is useful to compare results account for an atypical order of the creep curves in presented here with those reported by Zawada et al. [37]for Fig. 5(a)(i.e 130 MPa creep curve showing slightly higher N610/AS. At 1100C and creep stress levels of 50 and strains than the 140 MPa curve). Note that all creep strains 75 MPa, N610/AS did not exhibit tertiary creep and conse- are an order of magnitude lower than the failure strain quently accumulated considerably lower creep strains. obtained in the tension test at 900C. Creep behavior of However, rupture times were comparable to those obtained the uncoated fiber composite is qualitatively similar to that for N610/M/A. Furthermore, N610/AS did not achieve a of the coated fiber CMC. Likewise, creep strains are similar run-out even for stress as low as 50 MPa. At 1000C, to those accumulated by the coated fiber composite N610/AS exhibited improved creep resistance, achieving Minimum creep rate was reached in all tests. Creep a run-out at 75 MPa. Conversely, creep resistance of the strain rate as a function of applied stress is shown in N610/M/A composite remained poor at 1000C. A scop- Fig. 6, where results of the present investigation are plotted ing creep test conducted at 1000C and 80 MPa failed after together with the data from Wilson and Visser [40]for 17.5h. Short creep lives at 1000 and 1100C, and exces- Nextel 610 fibers, Zawada et al. [37] for a N610/AS com- a ely large creep strains at 1100C reveal poor creep per- posite, and Casas and Martinez-Esnaola [41] for a Nextel mance, making this material unacceptable for 610/Mox composite. To further facilitate comparison applications invol sustained loading at temperatures between the creep properties of the fibers and the compos- >1000C. Therefore, creep performance at 900C was ites, the Nextel 610 fiber data adjusted for V=0.2 and examined Vr=0. 15 (volume fractions of the on-axis fibers in the A It is seen in Fig. 5(a) that the creep curves obtained for N610/M/A and N610/AS composites, respectively), are 10/M/A at 900C exhibit primary and secondary creep also shown regimes. Transition from primary to secondary creep As expected, the creep strain rates increase with increas- ure.Increasing creep stress appears to have little effect on pplied stress as well as with increasing temperature.At occurs early in creep life. Secondary creep continues to fail- ing creep rates of N610/M/A are what may creep strain(see Table 3), which remains <0.05%. Negligi- expected from Nextel 610 fibers alone. Experimental resu
Because creep rupture of this CMC is likely dominated by creep of Nextel 610 fibers, it is useful to compare results presented here with those reported by Zawada et al. [37] for N610/AS. At 1100 C and creep stress levels of 50 and 75 MPa, N610/AS did not exhibit tertiary creep and consequently accumulated considerably lower creep strains. However, rupture times were comparable to those obtained for N610/M/A. Furthermore, N610/AS did not achieve a run-out even for stress as low as 50 MPa. At 1000 C, N610/AS exhibited improved creep resistance, achieving a run-out at 75 MPa. Conversely, creep resistance of the N610/M/A composite remained poor at 1000 C. A scoping creep test conducted at 1000 C and 80 MPa failed after 17.5 h. Short creep lives at 1000 and 1100 C, and excessively large creep strains at 1100 C reveal poor creep performance, making this material unacceptable for applications involving sustained loading at temperatures P1000 C. Therefore, creep performance at 900 C was examined. It is seen in Fig. 5(a) that the creep curves obtained for N610/M/A at 900 C exhibit primary and secondary creep regimes. Transition from primary to secondary creep occurs early in creep life. Secondary creep continues to failure. Increasing creep stress appears to have little effect on creep strain (see Table 3), which remains 60.05%. Negligible differences between the creep strains accumulated at different stress levels together with the inherent data scatter, account for an atypical order of the creep curves in Fig. 5(a) (i.e. 130 MPa creep curve showing slightly higher strains than the 140 MPa curve). Note that all creep strains are an order of magnitude lower than the failure strain obtained in the tension test at 900 C. Creep behavior of the uncoated fiber composite is qualitatively similar to that of the coated fiber CMC. Likewise, creep strains are similar to those accumulated by the coated fiber composite. Minimum creep rate was reached in all tests. Creep strain rate as a function of applied stress is shown in Fig. 6, where results of the present investigation are plotted together with the data from Wilson and Visser [40] for Nextel 610 fibers, Zawada et al. [37] for a N610/AS composite, and Casas and Martinez-Esnaola [41] for a Nextel 610/Umox composite. To further facilitate comparison between the creep properties of the fibers and the composites, the Nextel 610 fiber data adjusted for Vf = 0.2 and Vf = 0.15 (volume fractions of the on-axis fibers in the N610/M/A and N610/AS composites, respectively), are also shown. As expected, the creep strain rates increase with increasing applied stress as well as with increasing temperature. At 1100 C, creep rates of N610/M/A are what may be expected from Nextel 610 fibers alone. Experimental results 0.0 0.5 1.0 1.5 2.0 0 100 200 300 400 Time (s) Strain (%) N610/Monazite/Alumina T = 1100˚C 80 MPa 100 MPa 120 MPa 40 MPa 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 0 10000 20000 30000 40000 50000 60000 Time (s) Strain (%) N610/Monazite/Alumina T = 1100˚C 80 MPa 40 MPa (a) (b) Fig. 4. Creep strain vs time curves for the N610/M/A ceramic composite at 1100 C: (a) time scale is reduced to clearly show creep curves obtained at stress levels P100 MPa and (b) time scale chosen to show creep strain accumulated at 40 MPa. 0.00 0.05 0.10 0.15 0.20 0.25 0 20000 40000 60000 80000 100000 Time (s) Strain (%) N610/Monazite/Alumina T = 900˚C 80 MPa 120 MPa 140 MPa 130 MPa 150 MPa 0.00 0.05 0.10 0.15 0.20 0.25 0 20000 40000 60000 80000 100000 Time (s) Strain (%) N610/Alumina T = 900˚C 80 MPa 73 MPa (a) (b) Fig. 5. Creep strain vs time curves at 900 C for: (a) N610/Monazite/ Alumina and (b) N610/Alumina. M.B. Ruggles-Wrenn et al. / Composites Science and Technology 66 (2006) 2089–2099 2093
M.B. Ruggles-Wrenn et al /Composites Science and Technology 66(2006)2089-2099 1E02 neous linkage of creep-nucleated cracks occurred just prior 1E-03 to failure N310 Fiber at v,a0.15. 1100C Stress-rupture behavior is summarized in Fig. 7, where 1.E04 110o’c creep stress is plotted vs time to rupture at 900 and 1E-05 1100C for both composites. Creep-rupture life of the 1E-06 coated fiber composite increases considerably with decreas- 1E-07 ing temperature. At 1100C, the time to rupture was 14 h for the low creep stress of 40 MPa, while the 80 MPa test survived a mere 0. 4 h. At 900C, the run-out stress was Creep strain rate is< 10st 120 MPa. While the coated fiber CMc achieved a run- E10 out at stresses <120 MPa, the uncoated fiber composite 1000 failed after 97 h at 73 MPa, approaching but not achieving Creep Stress(MPa) a run-out, and survived only 5.5 h at 80 MPa In the case of Fig. 6. Minimum creep rate as a function of applied stress for N610/A the monazite containing composite, the 80 MPa creep test and N610/M/A ceramic composites at 900 and 1100C. Data for Nextel was interrupted after 145 h of creep, producing the creep 610 fibers(Wilson [40), and Nextel 610-reinforced composites(Zawada life at least 25 times that of the uncoated fiber CMc at 37]and Casas [41]are also shown. the same creep stress. The addition of the monazite coating significantly improved the creep life at 900C. obtained for N610/M/A agree well with the N610 fiber Retained strength and modulus of the two specimens data adjusted for V=0. 2. Furthermore, fitting the that reached run-out in creep tests are summarized in Table N610/M/A creep results with a temperature-independent 4. Both specimens retained over 90% of their tensile Norton-Bailey equation of the form strength; the modulus loss was limited to 10%. Considering that these specimens experienced only small creep strains (.4%), better than 90% retention of strength and modu- ields a stress exponent n x3.5, which is approximately lus is not very surprising. Prior creep had minimal effect on equal to that reported for the Nextel 610 fiber. (Here e is the failure strain. Failure strains for the specimens sub the minimum creep rate, A is a temperature-dependent jected to prior creep were only slightly lower than those coefficient that accounts for the activation energy and other for the as-processed material. The excellent strength reten- variables in the full form of the power law, and o is the ap- tion produced in the present study is consistent with the plied stress). Conversely, both N610/AS and N610/Mox results reported by Zawada et al. [37] who observed no composites exhibit lower creep rates compared to the decrease in tensile strength of the N610/AS composite N610/M/A cross-ply. Creep rates of the N610/AS [37] are approximately an order of magnitude lower than what ted from N610 fibers(see the fiber da adjusted for v=0. 15). The lower creep rates of the N610/AS(a composite reinforced with an eight-harness sa tin weave of Nextel 610 fibers) suggest that creep may be Alumina ffected by fiber weave architecture. At 900C, the creep strain rates of both the uncoated fiber composite and 150F N610/onazite/Alumina N610/M/A(with the exception of the creep strain rate at 150 MPa)are <10-8s-I. The creep rupture of N610/M/A at both test temperatures is likely dominated by rupture of the Nextel 610 fibers. Previous studies suggest that the mechanism controlling the stead creep of the Nextel 610 fibers is interface-reaction controlled diffusion creep with fine intergranular crack formation. Cracks continue to nucleate throughout the 1000100001000001000000 creep process until a critical crack density is reached, caus- ing spontaneous crack linka ge and Fig. 7. Creep stress vs time to rupture for N610/A and N610/M/A noticeable tertiary creep at 900C suggests that a sponta- ceramic composites at 900 and 1100C Table 4 Retained properties of the N610/M/A specimens subjected to prior creep at 900C Specimen Creep stress(MPa) Retained strength(MPa) Strength retention (% Retained modulus(GPa) Modulus retention(%) Strain at 80 173 l20 164
obtained for N610/M/A agree well with the N610 fiber data adjusted for Vf = 0.2. Furthermore, fitting the N610/M/A creep results with a temperature-independent Norton–Bailey equation of the form e_ ¼ Arn yields a stress exponent n 3.5, which is approximately equal to that reported for the Nextel 610 fiber. (Here e_ is the minimum creep rate, A is a temperature-dependent coefficient that accounts for the activation energy and other variables in the full form of the power law, and r is the applied stress). Conversely, both N610/AS and N610/Umox composites exhibit lower creep rates compared to the N610/M/A cross-ply. Creep rates of the N610/AS [37] are approximately an order of magnitude lower than what would be expected from N610 fibers (see the fiber data adjusted for Vf = 0.15). The lower creep rates of the N610/AS (a composite reinforced with an eight-harness satin weave of Nextel 610 fibers) suggest that creep may be affected by fiber weave architecture. At 900 C, the creep strain rates of both the uncoated fiber composite and N610/M/A (with the exception of the creep strain rate at 150 MPa) are 6108 s 1 . The creep rupture of N610/M/A at both test temperatures is likely dominated by creep rupture of the Nextel 610 fibers. Previous studies [42–44] suggest that the mechanism controlling the steady-state creep of the Nextel 610 fibers is interface-reactioncontrolled diffusion creep with fine intergranular crack formation. Cracks continue to nucleate throughout the creep process until a critical crack density is reached, causing spontaneous crack linkage and failure. Lack of a noticeable tertiary creep at 900 C suggests that a spontaneous linkage of creep-nucleated cracks occurred just prior to failure. Stress-rupture behavior is summarized in Fig. 7, where creep stress is plotted vs time to rupture at 900 and 1100 C for both composites. Creep-rupture life of the coated fiber composite increases considerably with decreasing temperature. At 1100 C, the time to rupture was 14 h for the low creep stress of 40 MPa, while the 80 MPa test survived a mere 0.4 h. At 900 C, the run-out stress was 120 MPa. While the coated fiber CMC achieved a runout at stresses 6120 MPa, the uncoated fiber composite failed after 97 h at 73 MPa, approaching but not achieving a run-out, and survived only 5.5 h at 80 MPa. In the case of the monazite containing composite, the 80 MPa creep test was interrupted after 145 h of creep, producing the creep life at least 25 times that of the uncoated fiber CMC at the same creep stress. The addition of the monazite coating significantly improved the creep life at 900 C. Retained strength and modulus of the two specimens that reached run-out in creep tests are summarized in Table 4. Both specimens retained over 90% of their tensile strength; the modulus loss was limited to 10%. Considering that these specimens experienced only small creep strains (0.4%), better than 90% retention of strength and modulus is not very surprising. Prior creep had minimal effect on the failure strain. Failure strains for the specimens subjected to prior creep were only slightly lower than those for the as-processed material. The excellent strength retention produced in the present study is consistent with the results reported by Zawada et al. [37], who observed no decrease in tensile strength of the N610/AS composite 1.E-10 1.E-09 1.E-08 1.E-07 1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 10 100 1000 Creep Stress (MPa) Creep Strain Rate (s-1) N610 Fiber, 1100˚C Wilson, 2001 N610 Fiber at Vf = 0.2, 1100˚C N610/Monazite/Alumina, 1100˚C N610/Monazite/Alumina, 900˚C N610/Alumina, 900˚C Creep strain rate is < 10-8 s-1 N610/AS, 1100˚C Zawada, 2003 N610/Umox, 1100˚C Casas, 2004 N610 Fiber at Vf = 0.15, 1100˚C Fig. 6. Minimum creep rate as a function of applied stress for N610/A and N610/M/A ceramic composites at 900 and 1100 C. Data for Nextel 610 fibers (Wilson [40]), and Nextel 610-reinforced composites (Zawada [37] and Casas [41]) are also shown. 0 50 100 150 200 250 1 10 100 1000 10000 100000 1000000 Stress (MPa) N610/Monazite/Alumina UTS at 900˚ C N610/Monazite/Alumina UTS at 1100˚ C N610/Monazite/Alumina 1100˚ C N610/Monazite/Alumina 900˚ C N610/Alumina 900˚ C 0 50 100 150 200 250 1 10 100 1000 10000 100000 1000000 Time (s) Stress (MPa) N610/Monazite/Alumina UTS at 900˚ N610/Monazite/Alumina UTS at 1100˚ N610/Monazite/Alumina 1100˚ C N610/Monazite/Alumina 900˚ C N610/Alumina 900˚ C Fig. 7. Creep stress vs time to rupture for N610/A and N610/M/A ceramic composites at 900 and 1100 C. Table 4 Retained properties of the N610/M/A specimens subjected to prior creep at 900 C Specimen Creep stress (MPa) Retained strength (MPa) Strength retention (%) Retained modulus (GPa) Modulus retention (%) Strain at failure (%) B10-1 80 173 96 80 96 0.28 B10-2 120 164 91 75 90 0.29 2094 M.B. Ruggles-Wrenn et al. / Composites Science and Technology 66 (2006) 2089–2099
M.B. Ruggles-Wrenn et al. Composites Science and Technology 66(2006)2089-2099 previously subjected to 100 h at 75 MPa at 1000C. Tensile in the 0o plies break over a wide range of axial locations, stress-strain curves obtained for the N610/M/A specimens in general spanning the entire width of the specimen. In a subjected to prior creep are presented in Fig.& together given specimen, slightly different pullout lengths are pro- with the tensile stress-strain curve for the as-processed duced in different 0o plies. It is noteworthy that the average material. It is seen that prior creep had no qualitative effect fiber pullout length increases with increasing creep stress on tensile stress-strain behavior The fracture surfaces produced in monotonic tension tests at 1100C display similar"brushy'-type failure with exten 4.3. Composite microstructure sive pullout of individual fibers Fracture surfaces of the N610/M/A ns tested in Fracture surfaces of the N610/M/A specimens tested in creep at 900 C are shown in Figs. 10(ahe)for creep stress creep at 1100C are shown in Figs. 9(aHd), for creep levels of 80, 120, 130, 140 and 150 MPa, respectively. As stress levels of 40, 80, 100 and 120 MPa, respectively. The was the case at 1100C, the fracture planes obtained at fracture planes are not well defined; the fibers in the 0o plies 900C show randomly distributed fiber failure and pullout exhibit random failure producing fiber pullout. The fibers of both individual fibers and fiber tows. However, the aver age length of fiber pullout produced at 900C appears to be independent of creep stress level. Extensive fiber pullout of approximately the same average length is seen in all As-Processed 900C fracture surfaces, except in the fracture surface obtained at 150 MPa, which shows a shorter average length. Recall that this specimen also exhibited the shortest creep life, failing after 805 s. The fast fracture failure sur- faces produced at 900C also have a"brushy"appearance exhibiting significant amount of fiber pull-out Fracture surfaces of the N610/A specimens tested reep at 900C are shown in Figs. ll(a) and(b), for creep stress levels of 73 and 80 MPa, respectively. The contrast between the 80 MPa fracture surfaces of the two compos- ites(Figs. 10(a)and ll(b)is striking. The fracture surface of the N610/M/A is"brushy", showing extensive pullout Fig8. Effects of prior creep at 900C on tensile stress-strain behavior of with pullout lengths reaching al0 mm. Conversely, the the N610/M/A ceramic compo N610/A exhibits no fiber pullout, with a brittle-type 10 mm 10 mm 10 mm 10 mn e surfaces of imens tested at 1100C at creep stress levels of:(a)40. (b)80. (c)100 and (d)120 MPa. Fiber pullout with increasing creep stres 10 101 10m mI 10m Fig. 10. Fracture surfaces of the N610/M/A specimens tested at 900C at creep stress levels of: (a)80, (b)120, (c)130. (d)140 and(e)150 MPa
previously subjected to 100 h at 75 MPa at 1000 C. Tensile stress–strain curves obtained for the N610/M/A specimens subjected to prior creep are presented in Fig. 8 together with the tensile stress–strain curve for the as-processed material. It is seen that prior creep had no qualitative effect on tensile stress–strain behavior. 4.3. Composite microstructure Fracture surfaces of the N610/M/A specimens tested in creep at 1100 C are shown in Figs. 9(a)–(d), for creep stress levels of 40, 80, 100 and 120 MPa, respectively. The fracture planes are not well defined; the fibers in the 0 plies exhibit random failure producing fiber pullout. The fibers in the 0 plies break over a wide range of axial locations, in general spanning the entire width of the specimen. In a given specimen, slightly different pullout lengths are produced in different 0 plies. It is noteworthy that the average fiber pullout length increases with increasing creep stress. The fracture surfaces produced in monotonic tension tests at 1100 C display similar ‘‘brushy’’-type failure with extensive pullout of individual fibers. Fracture surfaces of the N610/M/A specimens tested in creep at 900 C are shown in Figs. 10(a)–(e) for creep stress levels of 80, 120, 130, 140 and 150 MPa, respectively. As was the case at 1100 C, the fracture planes obtained at 900 C show randomly distributed fiber failure and pullout of both individual fibers and fiber tows. However, the average length of fiber pullout produced at 900 C appears to be independent of creep stress level. Extensive fiber pullout of approximately the same average length is seen in all 900 C fracture surfaces, except in the fracture surface obtained at 150 MPa, which shows a shorter average length. Recall that this specimen also exhibited the shortest creep life, failing after 805 s. The fast fracture failure surfaces produced at 900 C also have a ‘‘brushy’’ appearance, exhibiting significant amount of fiber pull-out. Fracture surfaces of the N610/A specimens tested in creep at 900 C are shown in Figs. 11(a) and (b), for creep stress levels of 73 and 80 MPa, respectively. The contrast between the 80 MPa fracture surfaces of the two composites (Figs. 10(a) and 11(b)) is striking. The fracture surface of the N610/M/A is ‘‘brushy’’, showing extensive pullout with pullout lengths reaching 10 mm. Conversely, the N610/A exhibits no fiber pullout, with a brittle-type 0 50 100 150 200 0.0 0.1 0.2 0.3 0.4 Strain (%) Stress (MPa) 900˚C As-Processed 145 h at 80 MPa 120 h at 120 MPa Fig. 8. Effects of prior creep at 900 C on tensile stress–strain behavior of the N610/M/A ceramic composite. Fig. 9. Fracture surfaces of the N610/M/A specimens tested at 1100 C at creep stress levels of: (a) 40, (b) 80, (c) 100 and (d) 120 MPa. Fiber pullout length increases with increasing creep stress. Fig. 10. Fracture surfaces of the N610/M/A specimens tested at 900 C at creep stress levels of: (a) 80, (b) 120, (c) 130, (d) 140 and (e) 150 MPa. M.B. Ruggles-Wrenn et al. / Composites Science and Technology 66 (2006) 2089–2099 2095
M.B. Ruggles-Wrenn et al /Composites Science and Technology 66(2006)2089-2099 10 mm 10 mm Fig. ll. Fracture surfaces of the N610/A specimens tested at 900C at creep stress levels of:(a)73 MPa and(b)80 MPa. fracture surface associated with catastrophic failure. These Views of the coated fiber composite microstructure at dramatically different fracture surfaces reflect the very dif- different scales are shown in Figs. 12 and 13. Fig. 12(a) ferent creep lives of the two specimens. The N610/A spec- shows the random failure of the fibers in the 0o plies and imen suffered brittle failure after 5.5 h. In the case of the resulting"brushes "or fiber pullout. A region of extensive N610/M/A specimen, the monazite coating effectively pre- fiber pullout with pullout of both individual fibers and fiber vented fiber/matrix bonding, allowing fiber pullout, and bundles is shown in Fig. 12(b). It is seen that the locations consequently extended the creep life beyond 145 h. While of the fiber breaks within an individual tow, and conse- the fracture surface of the N610/A specimen tested at quently the lengths of fiber pullout exhibit a broad distribu- 73 MPa shows some pullout of fiber tows, the pullout of tion. a considerable amount of individual fiber pullout is individual fibers, such as seen for the coated fiber compos- seen in Fig. 12(b), where individual fibers are clearly dis- te, is not observed. Fast fracture failure surfaces of the cernable, which demonstrates that the monazite coating N10/A CMC are nearly planar and similar in appearance has prevented the fibers from sintering together at the to those produced in creep elevated test temperatures. As seen in Fig. 12(c), fracture I(c) 100pm ply and the nearly planar fracture of the 90 ply
fracture surface associated with catastrophic failure. These dramatically different fracture surfaces reflect the very different creep lives of the two specimens. The N610/A specimen suffered brittle failure after 5.5 h. In the case of the N610/M/A specimen, the monazite coating effectively prevented fiber/matrix bonding, allowing fiber pullout, and consequently extended the creep life beyond 145 h. While the fracture surface of the N610/A specimen tested at 73 MPa shows some pullout of fiber tows, the pullout of individual fibers, such as seen for the coated fiber composite, is not observed. Fast fracture failure surfaces of the N610/A CMC are nearly planar and similar in appearance to those produced in creep. Views of the coated fiber composite microstructure at different scales are shown in Figs. 12 and 13. Fig. 12(a) shows the random failure of the fibers in the 0 plies and resulting ‘‘brushes’’ or fiber pullout. A region of extensive fiber pullout with pullout of both individual fibers and fiber bundles is shown in Fig. 12(b). It is seen that the locations of the fiber breaks within an individual tow, and consequently the lengths of fiber pullout exhibit a broad distribution. A considerable amount of individual fiber pullout is seen in Fig. 12(b), where individual fibers are clearly discernable, which demonstrates that the monazite coating has prevented the fibers from sintering together at the elevated test temperatures. As seen in Fig. 12(c), fracture Fig. 11. Fracture surfaces of the N610/A specimens tested at 900 C at creep stress levels of: (a) 73 MPa and (b) 80 MPa. Fig. 12. Fracture surfaces of a Nextel 610/Monazite/Alumina specimen tested in creep at 900 C (creep stress = 120 MPa, time to rupture = 432,175 s) showing: (a) extensive fiber pullout, (b) individual fiber pullout, (c) matrix holes left by fiber pullout and (d) region of coordinated fiber fracture in the 0 ply and the nearly planar fracture of the 90 ply. 2096 M.B. Ruggles-Wrenn et al. / Composites Science and Technology 66 (2006) 2089–2099
M B. Ruggles-Wrenn et al/ Composites Science and Technology 66(2006)2089-2099 Fig. 13. Fracture surfaces of a Nextel610/Monazite/Alumina specimen tested in creep at 900C(creep stress= 120 MPa, time to rupture=432, 175 s) showing:(a) Fibers debonding from the matrix and a matrix crack propagating around the fibers and (b) residue of the monazite coating and small pieces of matrix attached to the exposed fibers surfaces also exhibit matrix sockets left by pullout of single brittle-type fracture dominated the fracture surfaces. while filaments as well as of small bundles. Along with the some pullout of fiber bundles was present, the extensive regions of extensive pullout, fracture surfaces contain small pullout of individual fibers was not observed regions of flatter, more coordinated fracture topography with little or no fiber pullout, as shown in Fig. 12(d). Close 5. Concluding remarks examination reveals that most of the fibers fracture on dif- ferent planes, suggesting that a single crack front did not 5.1. Monotonic tension cause this fracture topography. The nearly planar fracture surface topography of the 90 plies is also seen in The tensile stress-strain behavior of N610/A and N610/ ig. 12(d). This micrograph also reveals a higher matrix M/A composites was investigated and the tensile properties volume and sparse fiber distribution at the edge of the measured at room and elevated temperatures. At tempera ply. Fig 13 demonstrates the efficacy of monazite coating tures <1100C, the stress-strain behavior of the uncoated in providing crack deflection. The matrix crack seen in fiber composite is nearly linear elastic until failure, while Fig 13(a)is effectively deflected and the crack front does the behavior of the N610/M/A composite becomes nonlin not cut across. but rather meanders around the fibers. a ear as the strain exceeds 0%. The addition of monazite higher magnification view in Fig. 13(b) shows residue of coating results in near 50% improvement in strength; how- the monazite coating and small pieces of matrix attached ever, it also causes a 34% decrease in modulus and a three- to the exposed fibers fold increase in failure strain. At 1200C, both composites In contrast to the " brushy" fracture surfaces of the exhibit highly nonlinear stress-strain behavior and a signif- N610/M/A, the fracture surfaces of the N610/A composite icant decrease in strength and modulus. The use of mona were considerably more planar(see Fig 14). Flat regions of zite coating results in 37% improvement in UTS Fig. 14. Fracture surface of a Nextel 610/alumina specimen tested in creep at 900C(creep stress=80 MPa, time to rupture= 19,995 s).(a) Regions of planar fracture with no individual fiber pullout dominate the fracture surface and (b) fiber bundle pullout and planar fracture of the 90 plies are visible
surfaces also exhibit matrix sockets left by pullout of single filaments as well as of small bundles. Along with the regions of extensive pullout, fracture surfaces contain small regions of flatter, more coordinated fracture topography with little or no fiber pullout, as shown in Fig. 12(d). Close examination reveals that most of the fibers fracture on different planes, suggesting that a single crack front did not cause this fracture topography. The nearly planar fracture surface topography of the 90 plies is also seen in Fig. 12(d). This micrograph also reveals a higher matrix volume and sparse fiber distribution at the edge of the ply. Fig. 13 demonstrates the efficacy of monazite coating in providing crack deflection. The matrix crack seen in Fig. 13(a) is effectively deflected and the crack front does not cut across, but rather meanders around the fibers. A higher magnification view in Fig. 13(b) shows residue of the monazite coating and small pieces of matrix attached to the exposed fibers. In contrast to the ‘‘brushy’’ fracture surfaces of the N610/M/A, the fracture surfaces of the N610/A composite were considerably more planar (see Fig. 14). Flat regions of brittle-type fracture dominated the fracture surfaces. While some pullout of fiber bundles was present, the extensive pullout of individual fibers was not observed. 5. Concluding remarks 5.1. Monotonic tension The tensile stress–strain behavior of N610/A and N610/ M/A composites was investigated and the tensile properties measured at room and elevated temperatures. At temperatures 61100 C, the stress–strain behavior of the uncoated fiber composite is nearly linear elastic until failure, while the behavior of the N610/M/A composite becomes nonlinear as the strain exceeds 0.1%. The addition of monazite coating results in near 50% improvement in strength; however, it also causes a 34% decrease in modulus and a threefold increase in failure strain. At 1200 C, both composites exhibit highly nonlinear stress–strain behavior and a significant decrease in strength and modulus. The use of monazite coating results in 37% improvement in UTS. Fig. 13. Fracture surfaces of a Nextel610/Monazite/Alumina specimen tested in creep at 900 C (creep stress = 120 MPa, time to rupture = 432,175 s) showing: (a) Fibers debonding from the matrix and a matrix crack propagating around the fibers and (b) residue of the monazite coating and small pieces of matrix attached to the exposed fibers. Fig. 14. Fracture surface of a Nextel 610/alumina specimen tested in creep at 900 C (creep stress = 80 MPa, time to rupture = 19,995 s). (a) Regions of planar fracture with no individual fiber pullout dominate the fracture surface and (b) fiber bundle pullout and planar fracture of the 90 plies are visible. M.B. Ruggles-Wrenn et al. / Composites Science and Technology 66 (2006) 2089–2099 2097
M.B. Ruggles-Wrenn et al /Composites Science and Technology 66(2006)2089-2099 5.2. Creep-rupture behavior everal ceramic-matrix composites. J Am Ceram Soc 2003;86(8:1282-91 The creep-rupture behavior of the N610/M/A ceramic 3]Prewo KM, Batt JA. The oxidative stability of carbon fibre reinforced composite was characterized for stress levels ranging from trix composites. J Mater Sci 1988: 23: 523-7 60 to 150 MPa at 900C. and for stress levels ranging from [4]Mah T Hecht NL, McCullum DE, Hoenigman JR,Kim HM. Katz AP, et al. Thermal stability of sic fibres(Nicalon). J Mater Sci 40 to 120 MPa at 1100C Creep tests of the uncoated fiber 984:19:1191-201. Nextel 610/Alumina composite were conducted at 900oC []More KL, Tortorelli PF, Ferber MK, Keiser JR. Observations of for 73 and 80 MPa accelerated silicon carbide recession by oxidation at high water-vapor pressures. J Am Ceram Soc 2000: 83(1): 211-3. At 1100C, the N610/M/A composite exhibits primary, 6 More KL, Tortorelli PF, Ferber MK, Walker LR, Keiser JR, condary and tertiary creep regimes. Creep strain accumu- Brentnall WD, Miralya N, Price JB. Exposure of ceramic and lation decreases with increasing applied stress. For all stress ceramic-matrix composites in simulated and actual combustor envi- levels investigated, the accumulated creep strain signifi ronments. In: Proceedings of international gas Turbine and aero. cantly exceeds the failure strain obtained in the tension test space Congress, 1999. Paper No. 99-GT-292 At 900C, short primary creep rapidly transitions into sec- [7 Ferber MK, Lin HT, Keiser JR. Oxidation behavior of non-oxide ceramics in a high-pressure, high-temperature steam environment. In ondary creep, no tertiary creep is observed. Creep strains Mechanical, thermal, and environmental testing and performance of are an order of magnitude lower than the failure strain eramic composites and components. In: Jenkins MG. Lara-Curzio obtained in tension test. Creep strains produced at 900C E, Gonczy ST, editors. ASTM STP, Vol 1392. American Society for are at least an order of magnitude lower than those Testing and Materials: 2000. P. 210-5 obtained at ll00°C [8] Haynes JA, Lance MJ, Cooley KM, Ferber MK, Lowden RA, Stinton DP CVD mullite coatings in high-temperature, high-press At 1100C, creep strain rates of N610/M/A range from air-H,O. J Am Ceram Soc 2000: 83(3): 657-9. 1. 4x 10 to 7.9x 10s. Furthermore, creep rates of [9] Opila EJ. Hann Jr RE Paralinear oxidation of sic in water vapor. J N610/M/A are what may be expected from Nextel 610 Am Ceram Soc1997;80(1):197-205 fibers alone. The relationship between the minimum creep [10] Opila EJ. Oxidation kinetics of chemically vapor deposited silicon carbide in wet oxygen. J Am Ceram Soc 1994: 77(3): 730-6 rate and applied stress can be represented by a power [11] Opila EJ Variation of the oxidation rate of silicon carbide with water law. The stress exponent (n a 3.5)is approximately equal sure. J Am Ceram Soc 1999: 82(3): 625-36 to that of the Nextel 610 fibers. The creep rates of the [12] Hermes EE, Kerans RJ. Degradation of non-oxide reinforcement and N610/M/A cross-ply are about an order of magnitude oxide matrix composites. Mat Res Soc, Symp Proc 1988: 125: 73-8 higher than those reported for N610/AS, a CMC rein [13] Szweda A. Millard ML, Harrison MG. Fiber-reinforced ceramic- forced with an eight-harness satin weave of Nextel 610 matrix composite member and method for making, US Pat. No 5 601 674.1997 fibers. Fiber weave architecture may be one of the factors [14] Sim SM, Kerans RJ. Slurry infiltration and 3-D woven composites affecting the creep response of the composite. Creep strain Ceram Eng Sci Proc 1992: 13(9-10): 632-41 rates of N610/M/a decrease significantly with decreasing [15] Moore EH, Mah T, Keller KA. 3D composite fabrication through At 900C, for creep stress levels belo matrix slurry pressure infiltration. Ceram Eng Sci Proc 150 MPa, creep rates of both N610/A and N610/M/A are [16 Lewis MH, Cain MG Doleman P Razzell AG, Gent Development At 150 MPa, creep rate of N610/M/A of interfaces in oxide and silicate matrix composites. In: Evans AG increases to 14×10-7s Naslain RG, editors. High-temperature ceramic-matrix composites At 900C the uncoated fiber composite did not Il: manufacturing and materials development. American Ceramic a run-out For N610/M/A the run-out stress was I Society:1995.p.4-52. (67% UTS). The addition of a monazite coating [17] Lange FF, Tu wC, Evans AG. Processing of damage-toler oxidation-resistant ceramic matrix composites by a precursor cantly improved the creep resistance, presumably by pre- tration and pyrolysis method. Mater Sci Eng A 1995: A195: 145-50 venting fiber/matrix bonding and allowing fiber pullout. [18] Underberg R, Eckerbom L. Design and processing of all-oxide The run-out N610/M/A specimens retained over 90% of composites. In: Evans AG, Naslain RG, editors. High-temperature tensile strength. At 1100oC. run-out was not achieved cramic-matrix composites i: manufacturing and materials develop- ment. American Ceramic Society: 1995. p 95-104 [19] Mouchon E, Colomban P. Oxide ceramic matrix/oxide fiber woven Acknowledgements fabric composites exhibiting dissipative fracture behavior. Compos ites I995:26:175-82 The authors would like to thank Dr. R.A. Kerans and [0] Morgan PED, Marshall DB. Ceramic composites of monazite and Dr. T. Parthasarathy for many valuable discussie 21]Tu wC, Lange FF, Evans AG. Concept for a dama ceramic composite with strong interfaces. J Am Ce References [22]Kerans R, Hay RS, Pagano NJ, Parthasarathy TA. The role of the [] Verrilli M, Opila EJ, Calomino A, Kiser JD. Effect of environment 1989:68(2):429-42. on the stress-rupture behavior of a carbon-fiber-reinforced silicon [23] Evans AG, Zok FW. Review: the physic carbide ceramic m 2004;87(8):1536-42. [24] Kerans RJ, Parthasarathy TA Crack def [2] Zawada LP, Staehler J, Steel S. Consequence of intermittent exposure ites and fiber coating design criteria. Composites: Part A to moisture and salt fog on the high-temperature fatigue durability of 199930:521-4
5.2. Creep-rupture behavior The creep-rupture behavior of the N610/M/A ceramic composite was characterized for stress levels ranging from 80 to 150 MPa at 900 C, and for stress levels ranging from 40 to 120 MPa at 1100 C. Creep tests of the uncoated fiber Nextel 610/Alumina composite were conducted at 900 C for 73 and 80 MPa. At 1100 C, the N610/M/A composite exhibits primary, secondary and tertiary creep regimes. Creep strain accumulation decreases with increasing applied stress. For all stress levels investigated, the accumulated creep strain signifi- cantly exceeds the failure strain obtained in the tension test. At 900 C, short primary creep rapidly transitions into secondary creep, no tertiary creep is observed. Creep strains are an order of magnitude lower than the failure strain obtained in tension test. Creep strains produced at 900 C are at least an order of magnitude lower than those obtained at 1100 C. At 1100 C, creep strain rates of N610/M/A range from 1.4 · 106 to 7.9 · 105 s 1 . Furthermore, creep rates of N610/M/A are what may be expected from Nextel 610 fibers alone. The relationship between the minimum creep rate and applied stress can be represented by a power law. The stress exponent (n 3.5) is approximately equal to that of the Nextel 610 fibers. The creep rates of the N610/M/A cross-ply are about an order of magnitude higher than those reported for N610/AS, a CMC reinforced with an eight-harness satin weave of Nextel 610 fibers. Fiber weave architecture may be one of the factors affecting the creep response of the composite. Creep strain rates of N610/M/A decrease significantly with decreasing temperature. At 900 C, for creep stress levels below 150 MPa, creep rates of both N610/A and N610/M/A are less than 108 s 1 . At 150 MPa, creep rate of N610/M/A increases to 1.4 · 107 s 1 . At 900 C the uncoated fiber composite did not achieve a run-out. For N610/M/A the run-out stress was 120 MPa (67% UTS). The addition of a monazite coating signifi- cantly improved the creep resistance, presumably by preventing fiber/matrix bonding and allowing fiber pullout. The run-out N610/M/A specimens retained over 90% of tensile strength. At 1100 C, run-out was not achieved. Acknowledgements The authors would like to thank Dr. R.A. Kerans and Dr. T. Parthasarathy for many valuable discussions. References [1] Verrilli MJ, Opila EJ, Calomino A, Kiser JD. Effect of environment on the stress-rupture behavior of a carbon-fiber-reinforced silicon carbide ceramic matrix composite. J Am Ceram Soc 2004;87(8):1536–42. [2] Zawada LP, Staehler J, Steel S. Consequence of intermittent exposure to moisture and salt fog on the high-temperature fatigue durability of several ceramic–matrix composites. J Am Ceram Soc 2003;86(8):1282–91. [3] Prewo KM, Batt JA. The oxidative stability of carbon fibre reinforced glass–matrix composites. J Mater Sci 1988;23:523–7. [4] Mah T, Hecht NL, McCullum DE, Hoenigman JR, Kim HM, Katz AP, et al. Thermal stability of sic fibres (Nicalon). J Mater Sci 1984;19:1191–201. [5] More KL, Tortorelli PF, Ferber MK, Keiser JR. Observations of accelerated silicon carbide recession by oxidation at high water-vapor pressures. J Am Ceram Soc 2000;83(1):211–3. [6] More KL, Tortorelli PF, Ferber MK, Walker LR, Keiser JR, Brentnall WD, Miralya N, Price JB. Exposure of ceramic and ceramic-matrix composites in simulated and actual combustor environments. In: Proceedings of International Gas Turbine and Aerospace Congress, 1999. Paper No. 99-GT-292. [7] Ferber MK, Lin HT, Keiser JR. Oxidation behavior of non-oxide ceramics in a high-pressure, high-temperature steam environment. In: Mechanical, thermal, and environmental testing and performance of ceramic composites and components. In: Jenkins MG, Lara-Curzio E, Gonczy ST, editors. ASTM STP, Vol. 1392. American Society for Testing and Materials; 2000. p. 210–5. [8] Haynes JA, Lance MJ, Cooley KM, Ferber MK, Lowden RA, Stinton DP. CVD mullite coatings in high-temperature, high-pressure air–H2O. J Am Ceram Soc 2000;83(3):657–9. [9] Opila EJ, Hann Jr RE. Paralinear oxidation of sic in water vapor. J Am Ceram Soc 1997;80(1):197–205. [10] Opila EJ. Oxidation kinetics of chemically vapor deposited silicon carbide in wet oxygen. J Am Ceram Soc 1994;77(3):730–6. [11] Opila EJ. Variation of the oxidation rate of silicon carbide with water vapor pressure. J Am Ceram Soc 1999;82(3):625–36. [12] Hermes EE, Kerans RJ. Degradation of non-oxide reinforcement and oxide matrix composites. Mat Res Soc, Symp Proc 1988;125:73–8. [13] Szweda A, Millard ML, Harrison MG. Fiber-reinforced ceramic– matrix composite member and method for making, US Pat. No. 5 601 674, 1997. [14] Sim SM, Kerans RJ. Slurry infiltration and 3-D woven composites. Ceram Eng Sci Proc 1992;13(9–10):632–41. [15] Moore EH, Mah T, Keller KA. 3D composite fabrication through matrix slurry pressure infiltration. Ceram Eng Sci Proc 1994;15(4):113–20. [16] Lewis MH, Cain MG, Doleman P, Razzell AG, Gent J. Development of interfaces in oxide and silicate matrix composites. In: Evans AG, Naslain RG, editors. High-temperature ceramic–matrix composites II: manufacturing and materials development. American Ceramic Society; 1995. p. 41–52. [17] Lange FF, Tu WC, Evans AG. Processing of damage-tolerant, oxidation-resistant ceramic matrix composites by a precursor infiltration and pyrolysis method. Mater Sci Eng A 1995;A195:145–50. [18] Lunderberg R, Eckerbom L. Design and processing of all-oxide composites. In: Evans AG, Naslain RG, editors. High-temperature ceramic–matrix composites ii: manufacturing and materials development. American Ceramic Society; 1995. p. 95–104. [19] Mouchon E, Colomban P. Oxide ceramic matrix/oxide fiber woven fabric composites exhibiting dissipative fracture behavior. Composites 1995;26:175–82. [20] Morgan PED, Marshall DB. Ceramic composites of monazite and alumina. J Am Ceram Soc 1995;78(6):1553–63. [21] Tu WC, Lange FF, Evans AG. Concept for a damage-tolerant ceramic composite with strong interfaces. J Am Ceram Soc 1996;79(2):417–24. [22] Kerans RJ, Hay RS, Pagano NJ, Parthasarathy TA. The role of the fiber–matrix interface in ceramic composites. Am Ceram Soc Bull 1989;68(2):429–42. [23] Evans AG, Zok FW. Review: the physics and mechanics of fiberreinforced brittle matrix composites. J Mater Sci 1994;29:3857–96. [24] Kerans RJ, Parthasarathy TA. Crack deflection in ceramic composites and fiber coating design criteria. Composites: Part A 1999;30:521–4. 2098 M.B. Ruggles-Wrenn et al. / Composites Science and Technology 66 (2006) 2089–2099