Availableonlineatwww.sciencedirect.co °" Science Direct COMPOSITE STRUCTURES ELSEVIER Composite Structures 78(2007)477-485 ww.elsevier. com/locate/compstruct Mechanical property degradation of a Nicalon fiber reinforced SiNC ceramic matrix composite under thermal shock loading Sirajus Salekeen, Justice Nana Amoako, Hassan Mahfuz, Shaik Jeelani Center for Adranced Materials, Tuskegee Unicersity, Tuskegee, AL 36088, United States Available online 27 December 2005 Abstract The degradation of mechanical properties of a Sic fiber reinforced SiNC ceramic matrix composite due to thermal shock by water quenching have been investigated. Post thermal shock tensile tests were performed to determine the degradation of mechanical properties of this composite In situ acoustic emission(AE) tests were also conducted. The tensile tests data and acoustic emission data were cor- elated. The AE signal indicated a sudden increase in aE events at critical points in the stress-strain relationship. The effects of thermal shock temperature and the number of thermal shock cycles on the mechanical properties, and on the ae responses were also evaluated. It observed that an increase in either factor resulted in more AE responses. fracture damage in the tensile test specimens was examined by Scanning Electron Microscopy. It was observed that the failure mechanism changed as the thermal shock temperature increased. The fracture surfaces of the specimens tested without thermal shock indicated an extensive fiber pullout while the thermally shocked speci mens showed reduced fiber pullout o 2005 Published by Elsevier Ltd Keywords: Ceramic matrix composites; Thermal shock; Acoustic emission 1. Introduction ven by the still spinning rotor passes through the hot sec- tion. The rapid cooling of the hot sections leads to severe In the process of developing more efficient industrial gas thermal shock. Since a typical gas turbine undergoes sev- turbines and turbine engines for airplanes to travel at eral emergency shut downs over its lifetime, ceramic com- supersonic speeds(High Speed Civil Transport, HSCT), ponents of the turbine and turbine engine may suffer let(turbine inlet)temperature. This has created the need to radation as a result of thermal shocks. Thermal shocks develop new materials that can withstand very high therefore, possess potential risk since cracks generally lead temperatures, and of the new materials, fiber-reinforced to the end of the service life for these components. Compo- ceramic matrIx compos AC)have attracted consider- nents made of CMCs are also used at the end section of hot able attention for use in industrial gas turbines and turbine exhaust nozzles and are often exposed to rain. After con- engines for the HSCT. idering these facts, it was decided in the current investiga While thermal shock is not a concern during a steady- tion to determine the effect of thermal shock on SiC/siNC state operation of gas turbines, it becomes important dur- composites subjected to a water-quenching technique. A ing emergency shut downs during which the temperature quench test exposes the materials to a more severe thermal of the inlet air drops abruptly by orders of magnitude In shock than they experience in service this case, the fuel supply is abruptly cut off and cool air dri- Extensive investigation on the behavior of CMC materi- als when subjected to abrupt thermal gradients has not yet been performed, as noted in a recent review by Wang and E-mail address: salekeens(@tuskegee. edu(S. Salekeen Singh [1]. Hasselman [2] expressed that the understanding 0263-8223/S.see front matter 2005 Published by Elsevier Ltd. doi: 10. 1016/j. compstruct 2005.11.016
Mechanical property degradation of a Nicalon fiber reinforced SiNC ceramic matrix composite under thermal shock loading Sirajus Salekeen *, Justice Nana Amoako, Hassan Mahfuz, Shaik Jeelani Center for Advanced Materials, Tuskegee University, Tuskegee, AL 36088, United States Available online 27 December 2005 Abstract The degradation of mechanical properties of a SiC fiber reinforced SiNC ceramic matrix composite due to thermal shock by water quenching have been investigated. Post thermal shock tensile tests were performed to determine the degradation of mechanical properties of this composite. In situ acoustic emission (AE) tests were also conducted. The tensile tests data and acoustic emission data were correlated. The AE signal indicated a sudden increase in AE events at critical points in the stress–strain relationship. The effects of thermal shock temperature and the number of thermal shock cycles on the mechanical properties, and on the AE responses were also evaluated. It was observed that an increase in either factor resulted in more AE responses. Fracture damage in the tensile test specimens was examined by Scanning Electron Microscopy. It was observed that the failure mechanism changed as the thermal shock temperature increased. The fracture surfaces of the specimens tested without thermal shock indicated an extensive fiber pullout while the thermally shocked specimens showed reduced fiber pullout. 2005 Published by Elsevier Ltd. Keywords: Ceramic matrix composites; Thermal shock; Acoustic emission 1. Introduction In the process of developing more efficient industrial gas turbines and turbine engines for airplanes to travel at supersonic speeds (High Speed Civil Transport, HSCT), much effort has been directed at raising the combustor outlet (turbine inlet) temperature. This has created the need to develop new materials that can withstand very high temperatures, and of the new materials, fiber-reinforced ceramic matrix composites (CMC) have attracted considerable attention for use in industrial gas turbines and turbine engines for the HSCT. While thermal shock is not a concern during a steadystate operation of gas turbines, it becomes important during emergency shut downs during which the temperature of the inlet air drops abruptly by orders of magnitude. In this case, the fuel supply is abruptly cut off and cool air driven by the still spinning rotor passes through the hot section. The rapid cooling of the hot sections leads to severe thermal shock. Since a typical gas turbine undergoes several emergency shut downs over its lifetime, ceramic components of the turbine and turbine engine may suffer from the formation of cracks and associated strength degradation as a result of thermal shocks. Thermal shocks, therefore, possess potential risk since cracks generally lead to the end of the service life for these components. Components made of CMCs are also used at the end section of hot exhaust nozzles and are often exposed to rain. After considering these facts, it was decided in the current investigation to determine the effect of thermal shock on SiC/SiNC composites subjected to a water-quenching technique. A quench test exposes the materials to a more severe thermal shock than they experience in service. Extensive investigation on the behavior of CMC materials when subjected to abrupt thermal gradients has not yet been performed, as noted in a recent review by Wang and Singh [1]. Hasselman [2] expressed that the understanding 0263-8223/$ - see front matter 2005 Published by Elsevier Ltd. doi:10.1016/j.compstruct.2005.11.016 * Corresponding author. E-mail address: salekeens@tuskegee.edu (S. Salekeen). www.elsevier.com/locate/compstruct Composite Structures 78 (2007) 477–485
S. Salekeen et al /Composite Structures 78(2007)477-485 of the thermal shock behavior lies in the correlation between retained strength and initiation of cracks under thermal stress. In a few published studies [3,4, 10] on this ubject, flexural strength tests were mainly conducted to R297 assess the thermal shock induced damage. Although flex 48.6 ural test are easy to perform, the results generally do not reflect randomly occurring defects in highly stressed part y1.3 4, 5]. During a bend test, only a small volume of the test piece material experiences the maximum load, making the strength measurements primarily dependent upon defects all dimensions are in mm ear the surface. Also, flexural testing may not adequately Fig. 1. Tensile specimen geometry quantify damage because of the combination of tensile, compressive and shear stresses involved. However, in a ten to uniform stress. The larger volume of fully loaded mate- ature. The composite is then pyrola sang at a low temper- sile test the entire gage length of the specimen is subjected lay-up, and it is then cured by process convert the prece rial in a tensile specimen relative to a flexural specimen ramic polymer into a ceramic. Subsequent impregnation increases the probability of finding a large flaw. Consider- and pyrolysis steps are carried out to achieve a final density ing these facts it was decided to use the tensile test as a post of 2.55 g/em. The volume fraction of fiber is 50% and the thermal shock destructive evaluation technique in this average diameter is approximately 14 um. Edge-loaded Investigation specimens having a geometry as shown in Fig. I were used Most studies of thermal shock on Nicalon fiber rein- for this investigation, eliminating the requirement of forced CMCs have been based on destructive techniques mounting external tabbing materials. The reduction in only [6-9]. Wang and Singh [10] are among the few who cross section from the grips to the specimen gage section have performed both destructive and nondestructive tech- was achieved by utilizing a single blend radius of niques to evaluate the thermal shock behavior of Nicalon 297 mm. The maximum specimen width at the specimen fiber reinforced silicon carbide ceramic composites. Their ends is 14 mm. a gage section width of 6.3 mm is used. choices for destructive and nondestructive evaluation the specimen thickness is 2.5 mm, and the overall specimen methods for this composite were the flexural test and length is 148.6 mm. mechanical-forced resonance methods, respectively. In the present investigation, in addition to post thermal tensile 2. 2. Thermal shock behavior testing a nondestructive evaluation (NDE) technique, in this case, the acoustic emission is simultaneously employed The thermal shock tests involved heating the test sam- The acoustic emissions occur when the elastic energy accu- ples in a furnace to an elevated temperature in air and then dynamic microstructural processes such as crack initiation ture. The g(shocking)them into water at room tempera- mulated within a stressed body is released due to different immersir and propagation. Available literature dealing with ae for 20 min to ensure sufficient thermal equilibrium. The analysis over a time domain can be found in a few publica- elapsed time for removing the specimen from furnace and tions [11-13]. Efforts were made in this investigation to cor- immersing it into water is about 3 s. The specimens were relate different fracture processes with the corresponding then kept in a water bath for 5 min. Each heating and AE responses. Finally, a Scanning Electron Microscopic shocking process together constituted one thermal shock (SEM)evaluation of the fracture surfaces was performed cycle. After each thermal shock cycle, the samples were as part of this study dried in an oven at 100C for 20 min. They were inspected visually for the appearance of any macroscopic damage 2. Experimental procedure such as delamination and chipping. Then destructive and nondestructive tests were carried out on all the specimens 2. Material and geometry The mass and dimensions of the specimens were measured with a precision of 0.0001 g and 0.01 mm, respectively. An The material used in this investigation is a ceramic Instron 8502 servohydraulic machine equipped with a self- matrix composite(CMC) comprised of a two-dimensional aligning super grip was used to perform the post thermal eight-harness satin woven SiC fabric impregnated by a sil- tensile test with a crosshead speed of 0.02 mm/s. aE instru- icon nitrogen carbon (SINC) matrix. The fiber is basically mentation from Physical Acoustics Corporation(PAC) B-SiC crystallites and the matrix is an amorphous mixture was used to monitor acoustic emissions. AE events such of silicon, carbon and nitrogen. These composite specimens as energy and counts were recorded simultaneously in real were supplied by Dow Corning Corporation(Midland, time during the mechanical testing. The AE instrumenta Michigan)and were manufactured by a polymer impregna- tion included a piezoelectric transducer(sensor), a pream- tion and pyrolysis(PIP)process. The matrix precursor is plifier (model 1220 A)with an option on a main first impregnated into the eight-harness satin woven fabric control of 40 dB or 60 dB
of the thermal shock behavior lies in the correlation between retained strength and initiation of cracks under thermal stress. In a few published studies [3,4,10] on this subject, flexural strength tests were mainly conducted to assess the thermal shock induced damage. Although flexural test are easy to perform, the results generally do not reflect randomly occurring defects in highly stressed parts [4,5]. During a bend test, only a small volume of the test piece material experiences the maximum load, making the strength measurements primarily dependent upon defects near the surface. Also, flexural testing may not adequately quantify damage because of the combination of tensile, compressive and shear stresses involved. However, in a tensile test the entire gage length of the specimen is subjected to uniform stress. The larger volume of fully loaded material in a tensile specimen relative to a flexural specimen increases the probability of finding a large flaw. Considering these facts it was decided to use the tensile test as a post thermal shock destructive evaluation technique in this investigation. Most studies of thermal shock on Nicalon fiber reinforced CMCs have been based on destructive techniques only [6–9]. Wang and Singh [10] are among the few who have performed both destructive and nondestructive techniques to evaluate the thermal shock behavior of Nicalon fiber reinforced silicon carbide ceramic composites. Their choices for destructive and nondestructive evaluation methods for this composite were the flexural test and mechanical-forced resonance methods, respectively. In the present investigation, in addition to post thermal tensile testing a nondestructive evaluation (NDE) technique, in this case, the acoustic emission is simultaneously employed. The acoustic emissions occur when the elastic energy accumulated within a stressed body is released due to different dynamic microstructural processes such as crack initiation and propagation. Available literature dealing with AE analysis over a time domain can be found in a few publications [11–13]. Efforts were made in this investigation to correlate different fracture processes with the corresponding AE responses. Finally, a Scanning Electron Microscopic (SEM) evaluation of the fracture surfaces was performed as part of this study. 2. Experimental procedure 2.1. Material and geometry The material used in this investigation is a ceramic matrix composite (CMC) comprised of a two-dimensional eight-harness satin woven SiC fabric impregnated by a silicon nitrogen carbon (SiNC) matrix. The fiber is basically b-SiC crystallites and the matrix is an amorphous mixture of silicon, carbon and nitrogen. These composite specimens were supplied by Dow Corning Corporation (Midland, Michigan) and were manufactured by a polymer impregnation and pyrolysis (PIP) process. The matrix precursor is first impregnated into the eight-harness satin woven fabric lay-up, and it is then cured by processing at a low temperature. The composite is then pyrolized to convert the preceramic polymer into a ceramic. Subsequent impregnation and pyrolysis steps are carried out to achieve a final density of 2.55 g/cm3 . The volume fraction of fiber is 50% and the average diameter is approximately 14 lm. Edge-loaded specimens having a geometry as shown in Fig. 1 were used for this investigation, eliminating the requirement of mounting external tabbing materials. The reduction in cross section from the grips to the specimen gage section was achieved by utilizing a single blend radius of 297 mm. The maximum specimen width at the specimen ends is 14 mm. A gage section width of 6.3 mm is used, the specimen thickness is 2.5 mm, and the overall specimen length is 148.6 mm. 2.2. Thermal shock behavior The thermal shock tests involved heating the test samples in a furnace to an elevated temperature in air and then immersing (shocking) them into water at room temperature. The specimens were held at the elevated temperature for 20 min to ensure sufficient thermal equilibrium. The elapsed time for removing the specimen from furnace and immersing it into water is about 3 s. The specimens were then kept in a water bath for 5 min. Each heating and shocking process together constituted one thermal shock cycle. After each thermal shock cycle, the samples were dried in an oven at 100 C for 20 min. They were inspected visually for the appearance of any macroscopic damage such as delamination and chipping. Then destructive and nondestructive tests were carried out on all the specimens. The mass and dimensions of the specimens were measured with a precision of 0.0001 g and 0.01 mm, respectively. An Instron 8502 servohydraulic machine equipped with a selfaligning super grip was used to perform the post thermal tensile test with a crosshead speed of 0.02 mm/s. AE instrumentation from Physical Acoustics Corporation (PAC) was used to monitor acoustic emissions. AE events such as energy and counts were recorded simultaneously in real time during the mechanical testing. The AE instrumentation included a piezoelectric transducer (sensor), a preamplifier (model 1220 A) with an option on a main gain control of 40 dB or 60 dB. 2.5 All dimensions are in mm 148.6 7.0 1.3 6.3 52.3 R 297 8° Fig. 1. Tensile specimen geometry. 478 S. Salekeen et al. / Composite Structures 78 (2007) 477–485
S. Salekeen et aL. Composite Structures 78(2007)477-485 3. Results and discussion 3.I. Thermal shock behavior The degradation of strength due to thermal shock was determined by measuring the retained strength after ther 苏品 mally quencing the specimens from different elevated tem- peratures to room temperature in water. These elevated temperatures are400°C,600°Cand800°C. Experiments were also performed on unshocked specimens at room tem- Fig 3. The effect of shock temperature difference on strength of the SiC/ perature(RT). The stress-strain behaviors of both unshoc- SINC composite ked specimens, and specimens shocked from various elevated temperatures are shown in Fig. 2. From Fig. 2 it is clear that the mechanical properties of damage occurs is called the critical temperature differen nt ference in temperature(An) below which no significa the thermally shocked samples are lower than those of the (AT ). The critical temperature of a material, therefore, unshocked samples. Properties such as ultimate strength, a design parameter that signifies the limitation of working eld strength, and elastic modulus all degraded due to temperature for that material Fig. 3 shows that the first significant drop of material stress-strain behavior of the composites under all testing strength occurs at about 350 C. Thus, the critical thermal conditions consists of two fundamental regions: the initial shock temperature difference of this material can be tenta- linear region and the final nonlinear region. In the linear tively assumed to be 350C. It is also noticeable that there fibers and matrix. In the nonlinear region, which sustains is a gradual decrease in strength beyond this critical tem- perature difference(ATc). Much literatures is available that more strain than the linear region, the load is gradually deals with effect of thermal shock on the mechanical prop- transferred from the matrix to the fibers. The nonlinear erties of materials. A typical property degradation curve region also indicates a larger damage tolerance as a result due to thermal shock on CMC [10 is shown in Fig.4 f fibers bridging the cracks. The transition point from Fig 4 shows that the onset of damage in CMC occurs at the linear to the nonlinear region signifies the initiation a critical temperature difference ATe. The damage pro- of significant matrix cracks. After the matrix cracks, fibers gresses as AT increases and then eventually saturates at act to bridge the matrix cracks and enhance the toughness even some higher AT. It is understood that the beginning thermal shock was measured in order to determine the crit- or both. When the m e. of the material. These cracks, however, permit the harsh of the gradual decrea mechanical pi roberge was environments to move easily into the composite, degrading brought about by matrix cracking or debonding or both the fiber /matrix interface as well as the fiber strength The ultimate retained strength of each specimen after at AT. The nonsteady state represents the continuation and enlargement state for matrix cracking or debonding ical temperature difference (ATc). The effect of thermal the second steady state indicates that the reinforcing fibers shock temperature difference(AT) on the retained strength bridge the cracks, which limits the thermal shock damage of this composite is shown in Fig 3. The shocking temper- From there on the strength of the composite is determined ature difference(AT), in this case, refers to the difference in largely by the load carrying capability of the fibers and is temperature between the furnace and the room tempera- usually not affected by the thermal shock damage The thermal shock degradation behavior is best charac- maximum difference in temperature (An) that a ceramic terized by the normalized retained strength of the material material can sustain without cracking. This maximum dif- which is defined as the ratio of the retained strength after thermal shocking to the initial strength of the unshocked material. The effect of the thermal shock temperature dif- ference and the number of thermal shock cycles on the nor- malized retained strength are shown in Figs. 5 and 6 Damage starts Non-Steady State 是 Strain Fig. 2. Thermal shock effect on the stress-strain behavior of SiC/SiNC Fig. 4. The effect of shocking temperature on composite properties
3. Results and discussion 3.1. Thermal shock behavior The degradation of strength due to thermal shock was determined by measuring the retained strength after thermally quencing the specimens from different elevated temperatures to room temperature in water. These elevated temperatures are 400 C, 600 C and 800 C. Experiments were also performed on unshocked specimens at room temperature (RT). The stress–strain behaviors of both unshocked specimens, and specimens shocked from various elevated temperatures are shown in Fig. 2. From Fig. 2 it is clear that the mechanical properties of the thermally shocked samples are lower than those of the unshocked samples. Properties such as ultimate strength, yield strength, and elastic modulus all degraded due to thermal shock. It is also evident from Fig. 2 that the tensile stress–strain behavior of the composites under all testing conditions consists of two fundamental regions: the initial linear region and the final nonlinear region. In the linear region, it is assumed that the applied load is carried by both fibers and matrix. In the nonlinear region, which sustains more strain than the linear region, the load is gradually transferred from the matrix to the fibers. The nonlinear region also indicates a larger damage tolerance as a result of fibers bridging the cracks. The transition point from the linear to the nonlinear region signifies the initiation of significant matrix cracks. After the matrix cracks, fibers act to bridge the matrix cracks and enhance the toughness of the material. These cracks, however, permit the harsh environments to move easily into the composite, degrading the fiber/matrix interface as well as the fiber strength. The ultimate retained strength of each specimen after thermal shock was measured in order to determine the critical temperature difference (DTc). The effect of thermal shock temperature difference (DT) on the retained strength of this composite is shown in Fig. 3. The shocking temperature difference (DT), in this case, refers to the difference in temperature between the furnace and the room temperature (22 C). A common thermal shock parameter is the maximum difference in temperature (DT) that a ceramic material can sustain without cracking. This maximum difference in temperature (DT) below which no significant damage occurs is called the critical temperature difference (DTc). The critical temperature of a material, therefore, is a design parameter that signifies the limitation of working temperature for that material. Fig. 3 shows that the first significant drop of material strength occurs at about 350 C. Thus, the critical thermal shock temperature difference of this material can be tentatively assumed to be 350 C. It is also noticeable that there is a gradual decrease in strength beyond this critical temperature difference (DTc). Much literatures is available that deals with effect of thermal shock on the mechanical properties of materials. A typical property degradation curve due to thermal shock on CMC [10] is shown in Fig. 4. Fig. 4 shows that the onset of damage in CMC occurs at a critical temperature difference DTc. The damage progresses as DT increases and then eventually saturates at even some higher DT. It is understood that the beginning of the gradual decrease in mechanical properties was brought about by matrix cracking or debonding or both at DTc. The nonsteady state represents the continuation and enlargement state for matrix cracking or debonding or both. When the nonsteady state ends, the strength at the second steady state indicates that the reinforcing fibers bridge the cracks, which limits the thermal shock damage. From there on the strength of the composite is determined largely by the load carrying capability of the fibers and is usually not affected by the thermal shock damage. The thermal shock degradation behavior is best characterized by the normalized retained strength of the material, which is defined as the ratio of the retained strength after thermal shocking to the initial strength of the unshocked material. The effect of the thermal shock temperature difference and the number of thermal shock cycles on the normalized retained strength are shown in Figs. 5 and 6. 0 50 100 150 200 250 0 0.1 0.2 0.3 0.4 0.5 0.6 %Strain Stress (MPa) RT 400 ˚C 600 ˚C 800 ˚C Fig. 2. Thermal shock effect on the stress–strain behavior of SiC/SiNC composite. 180 185 190 195 200 205 210 0 200 400 600 800 1000 Retained Strength (MPa) ΔT (oC) Fig. 3. The effect of shock temperature difference on strength of the SiC/ SiNC composite. Mechanical Properties Damage starts Steady state Non-Steady State ΔTc ΔT Fig. 4. The effect of shocking temperature on composite properties. S. Salekeen et al. / Composite Structures 78 (2007) 477–485 479
S. Salekeen et al / Composite Structures 78(2007)477-485 --3 cycle 兰39>2352 8001000 △TCC 02004006008001000 Fig. 5. The effect of shock temperature difference on strength of SiC/SiNC Fig. 7. The effect of shock temperature difference on the modulus of Sic/ SINC composite. 105 Fig. 6. The effect of thermal shock cycles on the strength of SiC/SiNC No of thermal shock cycles Fig 8. The effect of thermal shock cycles on the modulus of SiC/SiNc Fig 5 indicates that for a particular thermal shock cycle there is a gradual decrease in retained strength with increasing shocking temperature difference. It is clear that it is also clear from both that the coupling effect of for a particular number of thermal shock cycles, the higher temperatures with number of thermal shock strength degradation rate is almost uniform for all temper- cycles on the retained mo is even greater than either ature differences. In Fig. 6. the same trend is observed with individual effect. increases in the number of thermal shock cycles as the tem- Some of the significant factors mentioned by some erature remains constant. It is clear that for a particular researchers that might affect the thermal shock behavior hermal shock temperature, a sharp decrease of strength of continuous fiber reinforced CMCs are: a mismatch of is observed for the first thermal cycle, after which the coefficients of thermal expansion between fiber and matrix strength degradation rate decreases and remains constant materials, the interfacial properties, the hold time in the as the number of thermal shock cycles increases. Further- furnace prior to thermal shock, the composites processing, more, both figures indicate that the combined effect of and the fiber architecture [14-22]. higher temperature with higher number of thermal shock cycles on the strength degradation is even greater than 3. 2. Acoustic emission behavior either individual effect. The marked decline in composite retained strength caused by quenching(shocking) as shown Recently, the acoustic emission(AE)technique has Figs. 5 and 6 can be attributed to further damage to the gained prominence for detecting the onset and propagation fibers in addition to the matrix cracks of microcracking in composite materials [23, 24]. Under The changes in retained modulus of this CMC as a func- load application, several fracture phenomena can occur tion of both the thermal shock temperature difference, and in composite materials, such as matrix microcracking the number of thermal shock cycles are shown in Figs. 7 fiber-matrix debonding, delamination and fibers failure and 8, respectively. From these figures similar trends to Corresponding to each of these physical phenomena those previously observed for the retained strength are also energy is released in various forms. One of these forms is bserved in the retained modulus. There is a gradual acoustic waves [23]. Thus, by monitoring the acoustic emis- decrease in the retained modulus when either the shocking sions, the actual onset of microcracking and other fracture temperature or the number of thermal shock cycles phenomena can be identified. This happens when a sharp increases. The rate of decrease is almost uniform as the increase in the cumulative number of ae events is temperature increases. In Fig. 8 the rate of decrease is observed. In this investigation, damage development in greater for the first thermal shock cycle, thereafter it composites during tensile testing is monitored by observing remains reasonably constant as the number of additional such AE events as AE energy and AE event counts. The thermal shock cycles increases. As in the retained strength, event count indicates the number of AE events above a
Fig. 5 indicates that for a particular thermal shock cycle there is a gradual decrease in retained strength with increasing shocking temperature difference. It is clear that for a particular number of thermal shock cycles, the strength degradation rate is almost uniform for all temperature differences. In Fig. 6, the same trend is observed with increases in the number of thermal shock cycles as the temperature remains constant. It is clear that for a particular thermal shock temperature, a sharp decrease of strength is observed for the first thermal cycle, after which the strength degradation rate decreases and remains constant as the number of thermal shock cycles increases. Furthermore, both figures indicate that the combined effect of higher temperature with higher number of thermal shock cycles on the strength degradation is even greater than either individual effect. The marked decline in composite retained strength caused by quenching (shocking) as shown in Figs. 5 and 6 can be attributed to further damage to the fibers in addition to the matrix cracks. The changes in retained modulus of this CMC as a function of both the thermal shock temperature difference, and the number of thermal shock cycles are shown in Figs. 7 and 8, respectively. From these figures similar trends to those previously observed for the retained strength are also observed in the retained modulus. There is a gradual decrease in the retained modulus when either the shocking temperature or the number of thermal shock cycles increases. The rate of decrease is almost uniform as the temperature increases. In Fig. 8 the rate of decrease is greater for the first thermal shock cycle, thereafter it remains reasonably constant as the number of additional thermal shock cycles increases. As in the retained strength, it is also clear from both figures that the coupling effect of higher temperatures with higher number of thermal shock cycles on the retained modulus is even greater than either individual effect. Some of the significant factors mentioned by some researchers that might affect the thermal shock behavior of continuous fiber reinforced CMCs are: a mismatch of coefficients of thermal expansion between fiber and matrix materials, the interfacial properties, the hold time in the furnace prior to thermal shock, the composites processing, and the fiber architecture [14–22]. 3.2. Acoustic emission behavior Recently, the acoustic emission (AE) technique has gained prominence for detecting the onset and propagation of microcracking in composite materials [23,24]. Under load application, several fracture phenomena can occur in composite materials, such as matrix microcracking, fiber–matrix debonding, delamination and fibers failure. Corresponding to each of these physical phenomena, energy is released in various forms. One of these forms is acoustic waves [23]. Thus, by monitoring the acoustic emissions, the actual onset of microcracking and other fracture phenomena can be identified. This happens when a sharp increase in the cumulative number of AE events is observed. In this investigation, damage development in composites during tensile testing is monitored by observing such AE events as AE energy and AE event counts. The event count indicates the number of AE events above a 75 80 85 90 95 100 105 0 200 400 600 800 1000 Retained Modulus (%) 1 cycle 3 cycles 5 cycles ΔT (o C) Fig. 7. The effect of shock temperature difference on the modulus of SiC/ SiNC composite. 75 80 85 90 95 100 105 0 12 3 45 6 No. of thermal shock cycles Retained Modulus (%) 400 ˚C 600 ˚C 800 ˚C Fig. 8. The effect of thermal shock cycles on the modulus of SiC/SiNC composite. 75 80 85 90 95 100 105 0 200 400 600 800 1000 Retained Strength (%) 1 cycle 3 cycle 5 cycle ΔT (o C) Fig. 5. The effect of shock temperature difference on strength of SiC/SiNC composite. 75 80 85 90 95 100 105 01 234 56 No. of thermal shock cycles Retained Strength (%) 400 ˚C 600 ˚C 800 ˚C Fig. 6. The effect of thermal shock cycles on the strength of SiC/SiNC composite. 480 S. Salekeen et al. / Composite Structures 78 (2007) 477–485
S. Salekeen et aL. Composite Structures 78(2007)477-485 threshold value in terms of amplitude. The threshold value has to be specified during data acquisition, which, in this case, is 40 dB. For AE analysis purposes, the load versus 10000 time curve is compared with the energy vs. time curve and the count vs time curve. For a particular time, changes in the behavior of the load curve can be correlated with corresponding changes in energy and event count curves 2000. As the crosshead rate and the displacement of the specimen during tensile testing are known, the time required to reach a certain displacement can be calculated, and then the load Time(sec) data can be correlated with the energy and count signals ▲600C x800C Fig 9 represents the load vS time curve whereas Figs. 10 and 1l show the aE energy vs. time and the ae counts Fig. Il. AE counts vs time graph at various temperatures time graphs respectively Figs. 10 and ll show that there are marked differences in acoustic emission responses between unshocked(RT)and This is understood that at the beginning of the fracture shocked specimens at critical points such as yield point process, numerous matrix microcracks initiate and then and specimen failure. It is also observed during testing that propagate as the applied load increases. This initiation shocked samples emit fer acoustic emissions at the early and propagation of matrix cracks yield weak AE signals ages of loading, but as failure approaches the emission which are visible before any kind of spike in the signal rate increases rapidly. So overall, the shock samples pro- observed. The presence of these microcracks did not result duce more Ae events to failure than the unshocked sam- in any kind of departure of material behavior from its lin- ples. It is assumed that the increase of AE events for earity as observed in the load-time curve, a phenomenon shocked samples as failure approaches indicates the possi- also exhibited by other fiber reinforced composites [24] bility of the onset of a micro-failure process as a result of This can be attributed to the fact that these microcracks further propagation of preexisting cracks formed or do not significantly affect the load carrying capacity of extended by the thermal shock treatment. The preexisting the composite Fig 9 shows that, at a time period of about crack can also come from pores and reaction damage sites 25-31 s, the load-time curves deviate from linearity. Close on fibers. As the failure approaches, the formation of an observation at the same time period indicates an initial increased number of enlarged cracks results in a corre- spike in the energy and count curves. The physical phe- sponding increase in AE events. nomenon corresponding to this initial spike is attributed to the initiation of significant matrix cracking- essentially when small microcracks coalesce to form a larger matrix crack. Since the initial spike corresponds well to the pro- portional limit of the material, it is therefore feasible to 乙 correlate the observed ae events with the actual damage phenomena occurring in the material. As the load continue to increase, some portions of the specimens experience 800°C fiber-matrix debonding or fiber breaking. As a result a weak area is created and the load is transferred to the remaining unfractured material. This intact region becomes Time(sec) overloaded by the load transfer, and as a result an enlarged Fig. 9. Force-time graph at various temperatures. area of debonding and broken fibers occurs. The corre- sponding AE signals are therefore, very strong at this region, indicating a larger number of activities as final damage approaches. The largest spike of AE signals occurs at about 170 s which corresponds to final failure. The final failure is associated with extensive fiber breakage. Fig. 12 shock cycles. Fig. 13, on the other hand, shows the counts vs. number of thermal shock cycles for different thermal shock temperatures. Similarly Figs. 14 and 15 show the AE energy vs. temperature and AE energy vs Time(sec) er of thermal shock cycles rest ◆RT400°C▲600Cx800 Figs. 12-15 display evidence of a gradual increase in AE energy and aE count events with increase of either the Fig. 10. AE energy vs time graph shocking temperature difference or the number of thermal
threshold value in terms of amplitude. The threshold value has to be specified during data acquisition, which, in this case, is 40 dB. For AE analysis purposes, the load versus time curve is compared with the energy vs. time curve and the count vs. time curve. For a particular time, changes in the behavior of the load curve can be correlated with corresponding changes in energy and event count curves. As the crosshead rate and the displacement of the specimen during tensile testing are known, the time required to reach a certain displacement can be calculated, and then the load data can be correlated with the energy and count signals. Fig. 9 represents the load vs. time curve whereas Figs. 10 and 11 show the AE energy vs. time and the AE counts vs. time graphs respectively. Figs. 10 and 11 show that there are marked differences in acoustic emission responses between unshocked (RT) and shocked specimens at critical points such as yield point and specimen failure. It is also observed during testing that shocked samples emit fewer acoustic emissions at the early stages of loading, but as failure approaches the emission rate increases rapidly. So overall, the shock samples produce more AE events to failure than the unshocked samples. It is assumed that the increase of AE events for shocked samples as failure approaches indicates the possibility of the onset of a micro-failure process as a result of further propagation of preexisting cracks formed or extended by the thermal shock treatment. The preexisting crack can also come from pores and reaction damage sites on fibers. As the failure approaches, the formation of an increased number of enlarged cracks results in a corresponding increase in AE events. This is understood that at the beginning of the fracture process, numerous matrix microcracks initiate and then propagate as the applied load increases. This initiation and propagation of matrix cracks yield weak AE signals which are visible before any kind of spike in the signal is observed. The presence of these microcracks did not result in any kind of departure of material behavior from its linearity as observed in the load–time curve, a phenomenon also exhibited by other fiber reinforced composites [24]. This can be attributed to the fact that these microcracks do not significantly affect the load carrying capacity of the composite. Fig. 9 shows that, at a time period of about 25–31 s, the load–time curves deviate from linearity. Close observation at the same time period indicates an initial spike in the energy and count curves. The physical phenomenon corresponding to this initial spike is attributed to the initiation of significant matrix cracking—essentially when small microcracks coalesce to form a larger matrix crack. Since the initial spike corresponds well to the proportional limit of the material, it is therefore feasible to correlate the observed AE events with the actual damage phenomena occurring in the material. As the load continue to increase, some portions of the specimens experience fiber–matrix debonding or fiber breaking. As a result a weak area is created and the load is transferred to the remaining unfractured material. This intact region becomes overloaded by the load transfer, and as a result an enlarged area of debonding and broken fibers occurs. The corresponding AE signals are therefore, very strong at this region, indicating a larger number of activities as final damage approaches. The largest spike of AE signals occurs at about 170 s which corresponds to final failure. The final failure is associated with extensive fiber breakage. Fig. 12 shows AE counts vs. temperature for different thermal shock cycles. Fig. 13, on the other hand, shows the AE counts vs. number of thermal shock cycles for different thermal shock temperatures. Similarly Figs. 14 and 15 show the AE energy vs. temperature and AE energy vs. number of thermal shock cycles respectively. Figs. 12–15 display evidence of a gradual increase in AE energy and AE count events with increase of either the shocking temperature difference or the number of thermal 0 0.5 1 1.5 2 2.5 3 3.5 0 50 100 150 200 Time (sec) Force (kN) RT 400 ˚C 600 ˚C 800 ˚C Fig. 9. Force–time graph at various temperatures. 0 10 20 30 40 50 60 0 50 100 150 200 Time (sec) AE Energy (V) RT 400 ˚C 600 ˚C 800 ˚C Fig. 10. AE energy vs. time graph at various temperatures. 0 2000 4000 6000 8000 10000 12000 14000 0 50 100 150 200 Time (sec) AE Count RT 400˚C ˚C ˚C 600 800 Fig. 11. AE counts vs. time graph at various temperatures. S. Salekeen et al. / Composite Structures 78 (2007) 477–485 481
S. Salekeen et al /Composite Structures 78(2007)477-485 120000 I cycle cycles increases when thermal shock temperature remains 100000 3 cycles constant. It is also noticeable from Figs. 12 and 14 that anter cycle, there Is no signi 80000 change in the total AE counts and AE energy below the critical temperature of350°C(△T). Beyond350°,the 40000 acoustic response increases slowly but as the loading con- 20000 re, the acoustic emission increases rapid with corresponding decreases in mechanical properties Hasselmann [2] showed that for thermal shocks at temper- atures less than ATc, the strength remains unaffected Fig. 12. The effect of shocking temperature on the ae event counts for because insuficient strain energy is available to cause prop- SiC/SiNC composite agation of existing flaws. However, beyond ATe, the grad- ual decrease in the mechanical properties can be attributed to the increase in both the number and the size of the 120000 cracks in the materials as a result of the thermal shock 0000 This accounts for the corresponding increase in the cumu- 三80000 lative AE energy and AE counts. It has been widely 60000 accepted that in case of fiber reinforced composites, the 600C material properties become stable after reaching a second 20000 800C critical temperature [25] as shown in Fig. 4. No further No of thermal shock cycles the shocking temperature. Beyond this critical temperature when the cracks become large enough to ig. 13. The effect of number of thermal shock cycles on the aE event counts for SiC/SiNC composite quasi-static fashion, only then does any increase in AT result in further decreases in strength. In this investigation, however, all the specimens were tested below this second critical te d to the region only 3. 3. SEM micrographs Failure of the specimens occurred at the midsection for all cases. The fracture analyses of the specimens were car- 0200400600 ried out using a Scanning Electron Microscope(SEM)as △TCC shown in Fig. 16a-g. The fractographs corroborated the presence of a weak fiber/matrix interface at room or lower Fig. 14. The efiect of shocking temperature on the AE energy event for temperature due to the existence of significant fiber pullout at these temperatures. The presence of the comparably weak fiber/matrix interface causes cracks propagating through the matrix to deflect along the fiber/matrix inter face rather than continuing through the fiber. Therefore, fracture sure protrusion of fibers was observed from the ces of the failed specimens tested at room temperature. This is shown in Fig. 16a The protrusion of the fibers indicates that the matrix cracks did not propagate through the fibers, which remained intact behind the crack tip bridging the crack sur No of thermal cycles faces. They were, however, broken in tension within the matrix as the separation between the crack surfaces Fig. 15. The effect of number of thermal shock cycles on the AE energy increased and had to be pulled out prior to complete spec- event for SiCSiNC composite imen fracture. Fiber pullout was observed in all specimens shock cycles. Furthermore, Figs. 12 and 14 show that the from 600C to 800C. Additionally, as the tempera tested at 400C, which progressively decreased in amou AE signals increase nonuniformly as the temperature increased from 400C to 800C, there is relatively less fiber increases for a particular number of thermal shock cycles. pullout and more broken fibers for a particular thermal n the other hand, Figs. 13 and 15 show that the aE sig- shock cycle number. The specimens tested at room temper nals increase uniformly as the number of thermal shock ature have the largest amount of fiber pullout whereas the
shock cycles. Furthermore, Figs. 12 and 14 show that the AE signals increase nonuniformly as the temperature increases for a particular number of thermal shock cycles. On the other hand, Figs. 13 and 15 show that the AE signals increase uniformly as the number of thermal shock cycles increases when thermal shock temperature remains constant. It is also noticeable from Figs. 12 and 14 that after the first thermal shock cycle, there is no significant change in the total AE counts and AE energy below the critical temperature of 350 C (DTc). Beyond 350 C, the acoustic response increases slowly but as the loading continues to failure, the acoustic emission increases rapidly with corresponding decreases in mechanical properties. Hasselmann [2] showed that for thermal shocks at temperatures less than DTc, the strength remains unaffected because insufficient strain energy is available to cause propagation of existing flaws. However, beyond DTc, the gradual decrease in the mechanical properties can be attributed to the increase in both the number and the size of the cracks in the materials as a result of the thermal shock. This accounts for the corresponding increase in the cumulative AE energy and AE counts. It has been widely accepted that in case of fiber reinforced composites, the material properties become stable after reaching a second critical temperature [25] as shown in Fig. 4. No further drop in strength is expected with any further increase of the shocking temperature. Beyond this critical temperature when the cracks become large enough to propagate in quasi-static fashion, only then does any increase in DT result in further decreases in strength. In this investigation, however, all the specimens were tested below this second critical temperature and were confined to the nonsteady region only. 3.3. SEM micrographs Failure of the specimens occurred at the midsection for all cases. The fracture analyses of the specimens were carried out using a Scanning Electron Microscope (SEM) as shown in Fig. 16a–g. The fractographs corroborated the presence of a weak fiber/matrix interface at room or lower temperature due to the existence of significant fiber pullout at these temperatures. The presence of the comparably weak fiber/matrix interface causes cracks propagating through the matrix to deflect along the fiber/matrix interface rather than continuing through the fiber. Therefore, a significant protrusion of fibers was observed from the fracture surfaces of the failed specimens tested at room temperature. This is shown in Fig. 16a. The protrusion of the fibers indicates that the matrix cracks did not propagate through the fibers, which remained intact behind the crack tip bridging the crack surfaces. They were, however, broken in tension within the matrix as the separation between the crack surfaces increased and had to be pulled out prior to complete specimen fracture. Fiber pullout was observed in all specimens tested at 400 C, which progressively decreased in amount from 600 C to 800 C. Additionally, as the temperature increased from 400 C to 800 C, there is relatively less fiber pullout and more broken fibers for a particular thermal shock cycle number. The specimens tested at room temperature have the largest amount of fiber pullout whereas the 0 20000 40000 60000 80000 100000 120000 0 400 800 1200 ΔT (o C) AE Count 1 cycle 3 cycles 5 cycles Fig. 12. The effect of shocking temperature on the AE event counts for SiC/SiNC composite. 0 100 200 300 400 500 0 200 400 600 800 1000 AE Energy (V) 1 cycle 3 cycles 5 cycles ΔT (o C) Fig. 14. The effect of shocking temperature on the AE energy event for SiC/SiNC composite. 0 100 200 300 400 500 0123456 No. of thermal cycles AE Energy (V) 400 ˚C 600 ˚C 800 ˚C Fig. 15. The effect of number of thermal shock cycles on the AE energy event for SiC/SiNC composite. 0 20000 40000 60000 80000 100000 120000 0246 No. of thermal shock cycles AE Count 400 ˚C 600 ˚C 800 ˚C Fig. 13. The effect of number of thermal shock cycles on the AE event counts for SiC/SiNC composite. 482 S. Salekeen et al. / Composite Structures 78 (2007) 477–485
S. Salekeen et aL. Composite Structures 78(2007)477-485 483 TCAM-22 H-22 Fig. 16. Fracture surfaces showing fiber pullout and fiber fracture:(a)at RT or 22C; (b) at 400C, I thermal cycle; (c)at 800C, I thermal cycle; (d)at 00C,3 thermal cycles; (e)at 800C, 3 thermal cycles; (f) at 400C, 5 thermal cycles and (g) at 800C, 5 thermal cycles specimens tested at 800C have the least. Moreover, there hydrogen and methyl groups decompose. Nicalon fiber were more broken fibers at the matrix crack plane as the used in this study contains 58% silicon, 31% carbon and temperature increased for any particular thermal shock 11% oxygen. At this temperature, Sic oxidizes to SiOz cycle. The likely rationale for this scenario can be explained The silica (SiO2) layer thus formed then bonds the fiber as follows. As mentioned earlier, the weak interface and the matrix together. The resulting strong interfacial between the fiber and the matrix contains carbon. How- bonding causes the composite to become rigid and brittle ever, the carbon layer is oxidized in oxygen-rich environ- [33, 34]. One of the other byproducts of the reaction at this ments at temperatures around 400C [26, 27] After the temperature is the presence of significant quantities of oxy- carbon has been oxidized and as the temperature increases, gen Oxygen penetration can occur through matrix micro- the oxygen reacts with the Nicalon fiber to form a silica cracks that pre-existed in the material as a result of the layer on the fiber matrix interface [28-32]. During the fab- fabrication process, or through cracks formed during ther- rication of the composite in the Polymer Impregnation and mal shock loading. The oxidation process is accelerated Pyrolysis(PIP) process, the first stage of conversion occurs whenever the applied tensile stress is high enough to caus around 550C, when cross linking of the polymer chains further matrix cracking. Also, because Si-C-O ceramics occurs. Above this temperature the side chains containing are thermodynamically unstable, fiber decomposition also
specimens tested at 800 C have the least. Moreover, there were more broken fibers at the matrix crack plane as the temperature increased for any particular thermal shock cycle. The likely rationale for this scenario can be explained as follows. As mentioned earlier, the weak interface between the fiber and the matrix contains carbon. However, the carbon layer is oxidized in oxygen-rich environments at temperatures around 400 C [26,27]. After the carbon has been oxidized and as the temperature increases, the oxygen reacts with the Nicalon fiber to form a silica layer on the fiber matrix interface [28–32]. During the fabrication of the composite in the Polymer Impregnation and Pyrolysis (PIP) process, the first stage of conversion occurs around 550 C, when cross linking of the polymer chains occurs. Above this temperature the side chains containing hydrogen and methyl groups decompose. Nicalon fiber used in this study contains 58% silicon, 31% carbon and 11% oxygen. At this temperature, SiC oxidizes to SiO2. The silica (SiO2) layer thus formed then bonds the fiber and the matrix together. The resulting strong interfacial bonding causes the composite to become rigid and brittle [33,34]. One of the other byproducts of the reaction at this temperature is the presence of significant quantities of oxygen. Oxygen penetration can occur through matrix microcracks that pre-existed in the material as a result of the fabrication process, or through cracks formed during thermal shock loading. The oxidation process is accelerated whenever the applied tensile stress is high enough to cause further matrix cracking. Also, because Si–C–O ceramics are thermodynamically unstable, fiber decomposition also Fig. 16. Fracture surfaces showing fiber pullout and fiber fracture: (a) at RT or 22 C; (b) at 400 C, 1 thermal cycle; (c) at 800 C, 1 thermal cycle; (d) at 400 C, 3 thermal cycles; (e) at 800 C, 3 thermal cycles; (f) at 400 C, 5 thermal cycles and (g) at 800 C, 5 thermal cycles. S. Salekeen et al. / Composite Structures 78 (2007) 477–485 483
S. Salekeen et al /Composite Structures 78(2007)477-485 takes place around these temperatures. Furthermore, the at room temperature had relatively the longest and larg- evolution of Co gas as a result of the carbon oxidation est amount of fiber pullout while the specimen tested at causes further fiber decomposition and voids which serve 800° C had the least as flaws and contribute to the enlargement of existing flaws. 8. It was also observed that for a particular thermal shock The cumulative effect of these activities degrades fibers cycle, as the temperature increased the number of bro making them unable to bear any further load, and conse ken fibers at the fracture surface increased fiber pullout quently leading to fiber fracture decreased 4. Conclusions References I. The critic 350°C.Be critical temperature, the mechanical [] Wang H, Singh RN. Thermal shock behavior of ceramics and ceramic properties nsile strength and modulus of elas- omposites. Int Mat Rev 1994, 39: 228-44. ticity of this composite decrease with increases in either 2] Hasselman DPH. Unified theory of thermal shock fracture initiation the thermal shock temperature difference or the number and crack propagation in brittle ceramics. J Am Ceram Soc 1969;52:600-4 of thermal shock cycles. The cumulative effect of both 3] Boccaccini AR, Janczak-Rusch J, Pearce DH, Kern H. Assessment of the high temperature and high number of thermal shock damage by thermal shock in SiC-fiber reinforced borosilicate gla composite Comp Sci Techn 1999: 59: 105-12 This degradation of mechanical properties can be attrib- (4 Wang i, Singh RN, Lowden RA. Thermal shock behavior of two uted to the oxidation of the carbon-rich interfacial layer, socl996:79:1783-92. oxidation of Nicalon fibers, and thermal shock loading. [5] Wiederhorn SM, Hockey BJ. Ceramics Intl 1991; 17: 243 2. For a temperature range from 400C to 800C, when 6 Klug T, Reichert J, Brckner R. Thermal shock behavior of Sic-fiber- objected to one thermal shock cycle reinforced glasses. Glastech Ber 1992: 65: 41-9. etained strength is in the range of about 89-95% of [7]Shin D-W, Auh KH, Knowles KM. SiC fiber/borosilicate glass the initial strength. The corresponding retained modulus composite(part 3).J Ceram Soc Japan 1995: 103: 533-40 [8 Kagawa Y, Kurosawa N, Kishi T. Thermal shock resistance of Sic are in the range of about 84-94% of initial modulus. For fiber-reinforced borosilicate d lithium aluminosilicate matrix three thermal shock cycles these ranges decrease to composites. J Mat Sci 1993: 28: 735-41 about 83-92% and about 80-91% respectively of the ini- [9] Zawada LP, Wetherhold RC. The effects of thermal fatigue on a SiC tial values. For five thermal shock cycles these values ber/aluminosilicate glass composite. J Mat Sci 1991: 26: 648-54 further decrease to about 78-87% and about 78-86% [loJ Wang H, Singh RN. Thermal shock behavior of fber reinforced eramic composites. Mat Res Soc Symp Proc 1995: 365: 441-7 respectively of the initial values [l Harris B Habib FA, Cooke RG. Matrix cracking and the mechanical 3. Ultimate failure of ceramic matrix composites(CMCs) havior of SiC/CAS composites. Proc Roy Soc London A in this experiment and the material still maintains its [2 hen o, karandik a P. Takeda N, k ishi ti tredway Prewo load carrying ability to a significant degree. This is te. Ceram Eng Sci Proc 1991: 1586-98 reflected in the sizable amount of the retained strength [13] Kim R, Pegano N Crack Initiation in unidirectional brittle compos- even after△Te ite. J Am Ceram Soc 199: 74: 1082 4. During the post thermal shock tensile testing, the [4] Evans AG, Marshall DB. Acta Met 1989:37:2567-83 changes in mechanical behavior are accompanied by [I5] Huchinson JW, Jenson HM Mech Mater 1990: 9: 139 [16 Chawla KK. Ceramic matrix composites composites. second ed. Bos- corresponding changes in the acoustic emission response ton: Kluwer Academic Publishers: 2003. p. 81 of the materials. An initial lower numbered AE event [17] Blisset MJ, Smith PA, Yeomans JA. Thermal shock behavior of (both Ae energy and ae counts)at loading eventually unidirectional silicon carbide fiber reinforced calcium aluminosilica ives way to much higher total AE events at failure J Mater sci1997;32:317-25 5. Acoustic emission (AE) monitoring indicates a sudden [191 Sirieix F, Goursat P, Lecomte A, Dauger A. Comp Sci Tech strain relationship(proportional final failure, [20] Sato K, Morozumi H, Tezuka A, Funayama O, Isoda T. Higl etc). The AE vs time data correspond well with the load rature Ceramic-Matrix Composite Il. Am Ceram Soc Wester- vs time data ille,OH,1995:199 6. Acoustic emission (AE)monitoring indicates an [21] Gonon MF, Fantozzi G. Murat M, Disson JP J Eur Ceram Soc increased amount of overall AE activities with the [22] Wang H, Singh RN, Lowden RA. Thermal shock behavior of increase of either the thermal shock temperature or the unidirectional 0/90, and 2-D woven fiber reinforced CVI SiC matrix number of thermal shock cycles, indicating greater amounts of damage in the materials 23 - mposites. J Mater Sci 1997: 32: 3305-13 keda N, Chen O, Kishi T, Tredway W, Prewo K. 7. There was increased fiber pull-out at the room tempera- emission characterization of the fracture mechanism of compliant glass-matrix composite Eng Fract Mech 1991: 40 ture(22C)tested specimen. A gradual decrease in fiber [24] Kageyama K, Enoki M, Kishi T Efect of microfracture process on pullout was also observed as the temperature increased fracture toughness in fine particulate glass composites. In: Proceed- (i.e.400°C,600°Cand800°C. The specimen tested ings of the 3rd Japan SAMPE, Symposium, 1993. p. 2081-86
takes place around these temperatures. Furthermore, the evolution of CO gas as a result of the carbon oxidation causes further fiber decomposition and voids which serve as flaws and contribute to the enlargement of existing flaws. The cumulative effect of these activities degrades fibers, making them unable to bear any further load, and consequently leading to fiber fracture. 4. Conclusions 1. The critical temperature of this composite is about 350 C. Beyond the critical temperature, the mechanical properties such as tensile strength and modulus of elasticity of this composite decrease with increases in either the thermal shock temperature difference or the number of thermal shock cycles. The cumulative effect of both the high temperature and high number of thermal shock cycles on the mechanical properties is more profound. This degradation of mechanical properties can be attributed to the oxidation of the carbon-rich interfacial layer, oxidation of Nicalon fibers, and thermal shock loading. 2. For a temperature range from 400 C to 800 C, when subjected to one thermal shock cycle, the amount of retained strength is in the range of about 89–95% of the initial strength. The corresponding retained modulus are in the range of about 84–94% of initial modulus. For three thermal shock cycles these ranges decrease to about 83–92% and about 80–91% respectively of the initial values. For five thermal shock cycles these values further decrease to about 78–87% and about 78–86% respectively of the initial values. 3. Ultimate failure of ceramic matrix composites (CMCs) does not necessarily occur right after DTc as observed in this experiment and the material still maintains its load carrying ability to a significant degree. This is reflected in the sizable amount of the retained strength even after DTc. 4. During the post thermal shock tensile testing, the changes in mechanical behavior are accompanied by corresponding changes in the acoustic emission response of the materials. An initial lower numbered AE event (both AE energy and AE counts) at loading eventually gives way to much higher total AE events at failure. 5. Acoustic emission (AE) monitoring indicates a sudden increase in AE activities at critical points in the stress strain relationship (proportional limit, final failure, etc). The AE vs. time data correspond well with the load vs. time data. 6. Acoustic emission (AE) monitoring indicates an increased amount of overall AE activities with the increase of either the thermal shock temperature or the number of thermal shock cycles, indicating greater amounts of damage in the materials. 7. There was increased fiber pull-out at the room temperature (22 C) tested specimen. A gradual decrease in fiber pullout was also observed as the temperature increased (i.e. 400 C, 600 C and 800 C). The specimen tested at room temperature had relatively the longest and largest amount of fiber pullout while the specimen tested at 800 C had the least. 8. It was also observed that for a particular thermal shock cycle, as the temperature increased the number of broken fibers at the fracture surface increased fiber pullout decreased. References [1] Wang H, Singh RN. Thermal shock behavior of ceramics and ceramic composites. Int Mat Rev 1994;39:228–44. [2] Hasselman DPH. Unified theory of thermal shock fracture initiation and crack propagation in brittle ceramics. J Am Ceram Soc 1969;52:600–4. [3] Boccaccini AR, Janczak-Rusch J, Pearce DH, Kern H. Assessment of damage by thermal shock in SiC-fiber reinforced borosilicate glass composite. Comp Sci Techn 1999;59:105–12. [4] Wang H, Singh RN, Lowden RA. Thermal shock behavior of two dimensional woven fiber-reinforced ceramic composites. J Am Ceram Soc 1996;79:1783–92. [5] Wiederhorn SM, Hockey BJ. Ceramics Intl 1991;17:243. [6] Klug T, Reichert J, Brckner R. Thermal shock behavior of SiC-fiberreinforced glasses. Glastech Ber 1992;65:41–9. [7] Shin D-W, Auh KH, Knowles KM. SiC fiber/borosilicate glass composite (part 3). J Ceram Soc Japan 1995;103:533–40. [8] Kagawa Y, Kurosawa N, Kishi T. Thermal shock resistance of SiC fiber-reinforced borosilicate glass and lithium aluminosilicate matrix composites. J Mat Sci 1993;28:735–41. [9] Zawada LP, Wetherhold RC. The effects of thermal fatigue on a SiC fiber/aluminosilicate glass composite. J Mat Sci 1991;26:648–54. [10] Wang H, Singh RN. Thermal shock behavior of fiber reinforced ceramic composites. Mat Res Soc Symp Proc 1995;365:441–7. [11] Harris B, Habib FA, Cooke RG. Matrix cracking and the mechanical behavior of SiC/CAS composites. Proc Roy Soc London A 1992;437:109. [12] Chen O, Karandikar P, Takeda N, Kishi T, Tredway R, Prewo K. Acoustic emission characterization of the fracture mechanism of a glass-matrix composite. Ceram Eng Sci Proc 1991:1586–98. [13] Kim R, Pegano N. Crack Initiation in unidirectional brittle composite. J Am Ceram Soc 199;74:1082. [14] Evans AG, Marshall DB. Acta Met 1989;37:2567–83. [15] Huchinson JW, Jenson HM. Mech Mater 1990;9:139. [16] Chawla KK. Ceramic matrix composites composites. second ed. Boston: Kluwer Academic Publishers; 2003. p. 81. [17] Blisset MJ, Smith PA, Yeomans JA. Thermal shock behavior of unidirectional silicon carbide fiber reinforced calcium aluminosilicate. J Mater Sci 1997;32:317–25. [18] Lundberg L, Pompe R, Carlsson R. Comp Sci Tech 1990;37:165. [19] Sirieix F, Goursat P, Lecomte A, Dauger A. Comp Sci Tech 1990;37:7. [20] Sato K, Morozumi H, Tezuka A, Funayama O, Isoda T. High Temperature Ceramic-Matrix Composite II. Am Ceram Soc Westerville, OH, 1995:199. [21] Gonon MF, Fantozzi G, Murat M, Disson JP. J Eur Ceram Soc 1995;15:185. [22] Wang H, Singh RN, Lowden RA. Thermal shock behavior of unidirectional 00 /900 , and 2-D woven fiber reinforced CVI SiC matrix composites. J Mater Sci 1997;32:3305–13. [23] Takeda N, Chen O, Kishi T, Tredway W, Prewo K. Acoustic emission characterization of the fracture mechanism of a high compliant glass–matrix composite. Eng Fract Mech 1991;40:791–9. [24] Kageyama K, Enoki M, Kishi T. Effect of microfracture process on fracture toughness in fine particulate glass composites. In: Proceedings of the 3rd Japan SAMPE, Symposium, 1993. p. 2081–86. 484 S. Salekeen et al. / Composite Structures 78 (2007) 477–485
S. Salekeen et aL. Composite Structures 78(2007)477-485 485 [25]Twitty A, Russell-Floyd Rs, Cooke RG, Harris B. Thermal shock [30]Bonney LA, Cooper RR. Reaction-layer interfaces in SiC-fiber resistance of nextel/silica-zirconia ceramic-matrix composites man reinforced glass-ceraI a high ransmIssIon ufactured by freeze-gelation. J Eur Ceram Soc 1995: 15: 455 J Am oc1990:73(102916-21 [26]Plucknett KP, Cain RL, Lewis MH. Interface degradation in CAS/ [31] Tortorelli PF, Nijhawan S, Reister len RA. Ceram Eng sci Nicalon during elevated temperature aging. Proc Mater Res Soc Proc1993:14(7&8):358 Symp I995:365:42l-6 32] Jones RH, Hanager CH, Tortorelli PF JOM 1993: 45: 26 [27] Daniel AM, Martin-Meizoso A, Plucknett KP. Lewis MH, Braski [33] Evans AG, Zok FW, McMeeking RM, Du ZZ. Models of high- DN. Interface modification during oxidation of a glass-ceramic temperature, environmentally-assisted embrittlement in ceramic- arix/SiC fiber composite. Ceram Eng Sci Proc 1996: 17: 280-7 matrix composites. J Am Cer Soc 1996: 79(9): 2345-52. e8 Brennan JJ. Tailoring multiphase and composite ceramics. Mater Sci [34] Ogbuji LUA. Pervasive mode of oxidative degradation in a SiC-SiC Res1986;20:54960. composite. J Am Cer Soc 1998: 81(11): 2777-84 229Cooper RF, Chung K. Structure and chemistry of fiber-matrix interfaces in silicon carbide fiber-reinforced glass-ceramic composites: an electron microscopy analysis. J Mater Sci 1987: 22: 3148-60
[25] Twitty A, Russell-Floyd RS, Cooke RG, Harris B. Thermal shock resistance of nextel/silica–zirconia ceramic–matrix composites manufactured by freeze-gelation. J Eur Ceram Soc 1995;15:455. [26] Plucknett KP, Cain RL, Lewis MH. Interface degradation in CAS/ Nicalon during elevated temperature aging. Proc Mater Res Soc Symp 1995;365:421–6. [27] Daniel AM, Martin-Meizoso A, Plucknett KP, Lewis MH, Braski DN. Interface modification during oxidation of a glass–ceramic marix/SiC fiber composite. Ceram Eng Sci Proc 1996;17:280–7. [28] Brennan JJ. Tailoring multiphase and composite ceramics. Mater Sci Res 1986;20:549–60. [29] Cooper RF, Chung K. Structure and chemistry of fiber–matrix interfaces in silicon carbide fiber-reinforced glass–ceramic composites: an electron microscopy analysis. J Mater Sci 1987;22:3148–60. [30] Bonney LA, Cooper RR. Reaction-layer interfaces in SiC-fiberreinforced glass–ceramics: a high resolution scanning transmission electron microscopy analysis. J Am Ceram Soc 1990;73(10):2916–21. [31] Tortorelli PF, Nijhawan S, Reister L, Lowden RA. Ceram Eng Sci Proc 1993;14(7&8):358. [32] Jones RH, Hanager CH, Tortorelli PF. JOM 1993;45:26. [33] Evans AG, Zok FW, McMeeking RM, Du ZZ. Models of hightemperature, environmentally-assisted embrittlement in ceramic– matrix composites. J Am Cer Soc 1996;79(9):2345–52. [34] Ogbuji LUA. Pervasive mode of oxidative degradation in a SiC–SiC composite. J Am Cer Soc 1998;81(11):2777–84. S. Salekeen et al. / Composite Structures 78 (2007) 477–485 485