J.Am. Ceran.Soc.,88ll362-136502005 DOI:10.l1111551-2916.2005.00320.x urna Effect of Oxidation on Crack Deflection in SiC/Alo Laminated Ceramic Composites R. Krishnamurthy,* .fJ.Rankin, *and BW.Sheldon' Division of Engineering, Brown University, Providence, Rhode Island 02912 omposite tou ant for the laminate mina interphase were exposed to air at 500C to produce a pe tem considered Also. fiber with high friction co- istent, nearly uniform oxidation product layer. Crack deflection efficients can still provide considerable toughness in certai at the interface was then studied -point bend rocedure and interfacial fracture resistances were found to de- The primary objective of this work is to consider the effects of crease with increasing oxidation times. Electron microscopy ob- oxidation product layers on crack deflection at laminate inter- servations of the fractured interface show a complex multi-phas faces. In Sic-based composites, oxidation is typically expected microstructure. These results show that oxidation can produce a to adversely affect crack deflection, as silica reaction layers are sufficiently weak interface in a Sic-porous alumina interphase normally associated with strong bonding/strong interfaces omposite in contrast to most other SiC composites where in- There are some experimental results in the literature that sup- terface oxidation produces a strongly bonded interface which bort this view. However, no detailed study, using a well-con- nhibits crack deflection trolled test geometry, has been conducted to directly investigate the effect reaction layers have on crack deflection. In this study, we have attempted to address this issue. L. Introduction Porous alumina is an excellent interphase material, having the necessary low fracture resistance and good high temperature s well established that the nature of the fiber/matrix inter properties. SiC, with its excellent creep resistance, is a good face is an important criterion in the design of ceramic matrix choice as matrix material for high temperature applications composites for high temperature structural applications. One However, AL2O and Sic react in high temperature oxidizing important role of the interface is to provide low resistance cra environments with accompanying changes in the structure of the paths that facilitate crack deflection at the interface. This can terface and hence, in the fracture behavior. Thus, this is a good achieved by interposing"interphases"with low fracture resist model system to study the effects of interfacial reaction layers on ances between the fiber and the matrix in a fiber composite an the fracture behavior of composites. In this paper, we report between different laminae in a laminate. Traditionally Cand measurements of the interfacial fracture resistance of these lar BN have been used as interphase coatings because they create inates after oxidation for different time periods at 500C. This weak interfaces. However, both C and BN interphases oxidize in choice of the oxidation temperature was based on the finding air to give volatile oxidation products which renders them inef- that a thin, nearly uniform, oxide product layer is produced after ctive for high temperature applications. Furthermore gaps oxidation at this temperature, as reported in another study introduced in the structure because of volatilization and the The details of the experimental methods used to process the strong interfacial bonding resulting from silica glass formation laminates and test their fracture behavior are given in Section II upon oxidation of the SiC fiber/matrix also result in poor me- and results from these experiments are presented in Section Il chanical es.Numerous coating strategies, including the Discussion of the results of this use of fugitive layers with oxide cor orous oxides 9 sented in Section IV and C/SiC/C multilayer interphases for SiC composite have been advocated and tested in the lit Issue The effectiveness of the different coating materials Il. Experimental Procedure a bove in ding oxidation protection is a matter debate with some favoring porous oxides, of much Sic bars(cvD rohm and Haas, Woburn, MA)of dimensions 2 mm x 4 mm x 70 mm were lapped with I um diameter grit to a ides,,as the interphase coating materials of choice. The ox- surface finish of Ra 0.05 um, and coated with a uniform layer of this ystem, ie, SIC/Al2O high purity porous alumina(>99.95%), 2-5 um thick, using an laminates, was addressed by us in another study, where process aerosol spray deposition technique. Porous alumina coated ing strategies designed to achieve oxidation protection were also discussed. The relatively high friction coefficient of porous ox bars to form Sic/porous alumina laminates using a procedure ides as compared with C or bN can be a problem for fiber described by O'Brien and Sheldon and Krishnamurthy and composites, where fiber pull-out contributes significantly to the Sheldon. All samples used in this study were hot-pressed at 1400.C and 10.4 MPa. a schematic of one of these laminates J. Grecn--contnbuting editor along with typical dimensions is shown in Fig. 1. These lami- nates were oxidized for varying time periods at 500C to pro- duce a thin, persistent oxidation product layer. A combined Vickers indentation and three-point bending Manuscript No 11254. Received August 6, 2004; approved December 28, 2004. he MRSEC Program of the cience Foundation under Award procedure was used to introduce a pre-crack in the top sic lamina of the oxidized SiC/alumina laminates(see Fig. 1). Un- der three-point bend loading, the pre-crack in the top Sic lam- cUrrently at Princeton Institute for the science and technology of materials, Pr ina naturally self-arrests on reaching the interface, and the University specimen is immediately unloaded. These pre-cracked laminates
Effect of Oxidation on Crack Deflection in SiC/Al2O3 Laminated Ceramic Composites R. Krishnamurthy,* ,w,z J. Rankin,* and B.W. Sheldon* Division of Engineering, Brown University, Providence, Rhode Island 02912 Laminated composites consisting of SiC and a thin porous alumina interphase were exposed to air at 5001C to produce a persistent, nearly uniform oxidation product layer. Crack deflection at the interface was then studied using a four-point bend testing procedure and interfacial fracture resistances were found to decrease with increasing oxidation times. Electron microscopy observations of the fractured interface show a complex multi-phase microstructure. These results show that oxidation can produce a sufficiently weak interface in a SiC-porous alumina interphase composite, in contrast to most other SiC composites where interface oxidation produces a strongly bonded interface which inhibits crack deflection. I. Introduction I T is well established that the nature of the fiber/matrix interface is an important criterion in the design of ceramic matrix composites for high temperature structural applications. One important role of the interface is to provide low resistance crack paths that facilitate crack deflection at the interface. This can be achieved by interposing ‘‘interphases’’ with low fracture resistances between the fiber and the matrix in a fiber composite and between different laminae in a laminate.1 Traditionally C2 and BN3 have been used as interphase coatings because they create weak interfaces. However, both C and BN interphases oxidize in air to give volatile oxidation products which renders them ineffective for high temperature applications.4–6 Furthermore, gaps introduced in the structure because of volatilization and the strong interfacial bonding resulting from silica glass formation upon oxidation of the SiC fiber/matrix also result in poor mechanical properties.7 Numerous coating strategies, including the use of fugitive layers with oxide composites,8 porous oxides,9 and C/SiC/C multilayer interphases for SiC composites10–13 have been advocated and tested in the literature to resolve this issue. The effectiveness of the different coating materials discussed above in providing oxidation protection is a matter of much debate with some favoring porous oxides,7 and others non-oxides,14,15 as the interphase coating materials of choice. The oxidation resistance of this specific system, i.e., SiC/Al2O3 laminates, was addressed by us in another study, where processing strategies designed to achieve oxidation protection were also discussed.16 The relatively high friction coefficient of porous oxides as compared with C or BN can be a problem for fiber composites, where fiber pull-out contributes significantly to the composite toughness, but is less important for the laminate system considered here. Also, fiber coatings with high friction coefficients can still provide considerable toughness in certain cases.10 The primary objective of this work is to consider the effects of oxidation product layers on crack deflection at laminate interfaces. In SiC-based composites, oxidation is typically expected to adversely affect crack deflection, as silica reaction layers are normally associated with strong bonding/strong interfaces. There are some experimental results in the literature that support this view.17 However, no detailed study, using a well-controlled test geometry, has been conducted to directly investigate the effect reaction layers have on crack deflection. In this study, we have attempted to address this issue. Porous alumina is an excellent interphase material, having the necessary low fracture resistance and good high temperature properties.18 SiC, with its excellent creep resistance, is a good choice as matrix material for high temperature applications. However, Al2O3 and SiC react in high temperature oxidizing environments with accompanying changes in the structure of the interface and hence, in the fracture behavior. Thus, this is a good model system to study the effects of interfacial reaction layers on the fracture behavior of composites. In this paper, we report measurements of the interfacial fracture resistance of these laminates after oxidation for different time periods at 5001C. This choice of the oxidation temperature was based on the finding that a thin, nearly uniform, oxide product layer is produced after oxidation at this temperature, as reported in another study.16 The details of the experimental methods used to process the laminates and test their fracture behavior are given in Section II, and results from these experiments are presented in Section III. Discussion of the results of this work and conclusions are presented in Section IV. II. Experimental Procedure SiC bars (CVD Rohm and Haas, Woburn, MA) of dimensions 2 mm 4 mm 70 mm were lapped with 1 mm diameter grit to a surface finish of Ra 0.05 mm, and coated with a uniform layer of high purity porous alumina (499.95%), 2–5 mm thick, using an aerosol spray deposition technique.9,19 Porous alumina coated SiC bars were hot-pressed with matching, but uncoated, SiC bars to form SiC/porous alumina laminates, using a procedure described by O’Brien and Sheldon9 and Krishnamurthy and Sheldon.19 All samples used in this study were hot-pressed at 14001C and 10.4 MPa. A schematic of one of these laminates along with typical dimensions is shown in Fig. 1. These laminates were oxidized for varying time periods at 5001C to produce a thin, persistent oxidation product layer.16 A combined Vickers indentation and three-point bending procedure9 was used to introduce a pre-crack in the top SiC lamina of the oxidized SiC/alumina laminates (see Fig. 1). Under three-point bend loading, the pre-crack in the top SiC lamina naturally self-arrests on reaching the interface, and the specimen is immediately unloaded. These pre-cracked laminates Journal J. Am. Ceram. Soc., 88 [5] 1362–1365 (2005) DOI: 10.1111/j.1551-2916.2005.00320.x 1362 D. J. Green—contributing editor Supported by the MRSEC Program of the National Science Foundation under Award Number DMR-0079964. *Member, American Ceramic Society. w Author to whom correspondence should be addressed. e-mail: rkrishna@princeton.edu z Currently at Princeton Institute for the science and technology of materials, Princeton University. Manuscript No. 11254. Received August 6, 2004; approved December 28, 2004
May 2005 Communications of the American Ceramic Society 1363 0.4 0.2 4 mm 2-3 um porous alumina interphase oxidation time(hours) Fig 1. A schematic depiction of the SiC/porous Al2O3 laminate used in Fig. 2. Ti versus time of air exposure. are subsequently tested using a four-point bend test procedure were prepared by the procedure outlined in Section II. TEM to obtain interfacial fracture resistance values(implementation examination of the cross-section samples shows that oxidation details are described). 19-2I This test was continued beyond the of the laminates produces an interphase that includes a glassy steady state loading regime until the specimen completely de- phase and alumina grains(see Figs. 3(a)and( b). X-ray diffrac bonded at the interface to allow for microscopic examination of tion and scanning electron microscopy (SEM) also confirmed the fracture surface. The tests were repeated using a second that the oxidation product was amorphous. Considering the sample processed(and oxidized) under the same conditions to low oxidation temperature, the amorphous phase is likely to be confirm the validity of the fracture resistance measurement Af- silica with some Al incorporated into it. The incorporated Al ter the fracture tests, two pieces from the bottom, uncracked can lead to a faster rate of Sic oxidation. SEM examination of lamina(see Fig. 1)were bonded interlayer to interlayer to pre- the oxidized samples revealed that the interlayer also contained pare cross-sectional transmission electron microscopy (TEM some retained porosity(see Krishnamurthy and Sheldon6for specimens with the debond interface (i.e, the uncracked Sic details). The TEM micrographs in Figs. 3(a)and (b)correspond lamina/porous interlayer interface) in the middle. A Au-Pd to a sample oxidized for 6 h; samples oxidized for other time sputter coat placed atop one of the unbonded pieces was used periods also show similar features. Basic fracture mechanics ar as a marker to locate the debond interface guments that do not consider the effect of the porous interlayer microstructure would require this debond interface to be a clean SiC surface. The presence of alumina grains and a glassy phase II Results in the cross-section of the TEM samples is a clear indication that the oxidized porous interlayer has an effect on the location of In a previous study, laminate specimens hot-pressed in vacuum the actual debond interface in this system. In contrast, a com- at 1400C and 10.4 MPa exhibited the maximum measured in- paratively clean SiC surface with some pulled-out alumina terfacial fracture resistance, Ti(units of J/m), where interfacial rains constituted the debond surface for the non-oxidized sam- crack deflection was possible (i.e, higher temperature and/or ple(the pertinent micrograph can be found in an earlier publi- pressure created interfaces that were too strong to allow crack cation, Krishnamurthy and Sheldon). While this does not deflection).In the current study, laminates processed under necessarily eliminate the possibility that a glassy phase was these conditions were oxidized at 500C for different times. In a formed after hot-pressing, any such layer is likely to be thin sample oxidized at 800C, the deflected crack debonded the in- compared with glassy layers formed after oxidation. The pres- terface completely without arresting after the three-point bend ence of the alumina grains in the examined region of the TEM test used in the pre-cracking procedure. This demonstrates that samples can be explained if oxidation produces a complex mul- this interface is sufficiently weak, however, as described in other tiphase interphase rather than just a layer of glassy phase next k, this higher temperature produces non-uniform interface to the SiC lamina. However, possible microstructural inhomo- oxidation(the likely cause of this fracture result), and further geneities produced during processing of the laminates could also mechanical testing at this temperature was not pursued. Figure 2 lead to pulled-out alumina grains nows that Ti decreases with prolonged oxidation exposures of the laminates at 500C. After a 40 h exposure, the laminate de- bonded completely during the pre-cracking stage indicating that the interface is very weak. These observations are in contrast IV. Discussion and conclusions with the traditional view that silicate layers create a strongly An important theoretical consideration in the study of interfa bonded interphase. However, this conventional wisdom strictly cial fracture is the He-Hutchinson criterion. This states that applies only when the matrix and the fiber are strongly bonded crack deflection will occur in preference to crack penetration if together by an oxide plug. It may not necessarily be applicable the ratio of the interface and substrate fracture resistances Ti/Tr, is less than a threshold value(this value is -1/4 when f the interface to form a strong fiber / matrix bond, as is likely there is no elastic mismatch between the two layers). To assess the case for 500C oxidation, where only a thin oxide product the results in Fig. 2, changes in the interface because of oxida layer is expected to be formed tion must be considered. For very thin reaction layers, crack TEM cross-section specimens with the freshly exposed sur- deflection may still occur at the SiC/AlO3 interface and conse- face (i.e, exposed after the fracture test)of the uncracked bot quently, the measured fracture resistance is close to the initial tom SiC lamina and any residual porous interlayer sticking on it SIC/Al2O3 interfacial fracture resistance (i.e, the He-Hutchin (i.e, the debond interface) in the middle of the cross-section n criterion must be applied to this interface to determine
are subsequently tested using a four-point bend test procedure to obtain interfacial fracture resistance values (implementation details are described).9,19–21 This test was continued beyond the steady state loading regime until the specimen completely debonded at the interface to allow for microscopic examination of the fracture surface. The tests were repeated using a second sample processed (and oxidized) under the same conditions to confirm the validity of the fracture resistance measurement. After the fracture tests, two pieces from the bottom, uncracked lamina (see Fig. 1) were bonded interlayer to interlayer to prepare cross-sectional transmission electron microscopy (TEM) specimens with the debond interface (i.e., the uncracked SiC lamina/porous interlayer interface) in the middle. A Au–Pd sputter coat placed atop one of the unbonded pieces was used as a marker to locate the debond interface. III. Results In a previous study, laminate specimens hot-pressed in vacuum at 14001C and 10.4 MPa exhibited the maximum measured interfacial fracture resistance, Gi (units of J/m2 ), where interfacial crack deflection was possible (i.e., higher temperature and/or pressure created interfaces that were too strong to allow crack deflection).19 In the current study, laminates processed under these conditions were oxidized at 5001C for different times. In a sample oxidized at 8001C, the deflected crack debonded the interface completely without arresting after the three-point bend test used in the pre-cracking procedure. This demonstrates that this interface is sufficiently weak, however, as described in other work,16 this higher temperature produces non-uniform interface oxidation (the likely cause of this fracture result), and further mechanical testing at this temperature was not pursued. Figure 2 shows that Gi decreases with prolonged oxidation exposures of the laminates at 5001C. After a 40 h exposure, the laminate debonded completely during the pre-cracking stage indicating that the interface is very weak. These observations are in contrast with the traditional view that silicate layers create a strongly bonded interphase. However, this conventional wisdom strictly applies only when the matrix and the fiber are strongly bonded together by an oxide plug. It may not necessarily be applicable when the oxide layer does not extend across the entire thickness of the interface to form a strong fiber/matrix bond, as is likely the case for 5001C oxidation, where only a thin oxide product layer is expected to be formed. TEM cross-section specimens with the freshly exposed surface (i.e., exposed after the fracture test) of the uncracked, bottom SiC lamina and any residual porous interlayer sticking on it (i.e., the debond interface) in the middle of the cross-section were prepared by the procedure outlined in Section II. TEM examination of the cross-section samples shows that oxidation of the laminates produces an interphase that includes a glassy phase and alumina grains (see Figs. 3(a) and (b)). X-ray diffraction and scanning electron microscopy (SEM) also confirmed that the oxidation product was amorphous.16 Considering the low oxidation temperature, the amorphous phase is likely to be silica with some Al incorporated into it. The incorporated Al can lead to a faster rate of SiC oxidation.6 SEM examination of the oxidized samples revealed that the interlayer also contained some retained porosity (see Krishnamurthy and Sheldon16 for details). The TEM micrographs in Figs. 3(a) and (b) correspond to a sample oxidized for 6 h; samples oxidized for other time periods also show similar features. Basic fracture mechanics arguments that do not consider the effect of the porous interlayer microstructure would require this debond interface to be a clean SiC surface. The presence of alumina grains and a glassy phase in the cross-section of the TEM samples is a clear indication that the oxidized porous interlayer has an effect on the location of the actual debond interface in this system. In contrast, a comparatively clean SiC surface with some pulled-out alumina grains constituted the debond surface for the non-oxidized sample (the pertinent micrograph can be found in an earlier publication, Krishnamurthy and Sheldon19). While this does not necessarily eliminate the possibility that a glassy phase was formed after hot-pressing, any such layer is likely to be thin compared with glassy layers formed after oxidation. The presence of the alumina grains in the examined region of the TEM samples can be explained if oxidation produces a complex multiphase interphase rather than just a layer of glassy phase next to the SiC lamina. However, possible microstructural inhomogeneities produced during processing of the laminates could also lead to pulled-out alumina grains. IV. Discussion and Conclusions An important theoretical consideration in the study of interfacial fracture is the He-Hutchinson criterion.22 This states that crack deflection will occur in preference to crack penetration if the ratio of the interface and substrate fracture resistances, Gi=Gf, is less than a threshold value (this value is B1/4 when there is no elastic mismatch between the two layers). To assess the results in Fig. 2, changes in the interface because of oxidation must be considered. For very thin reaction layers, crack deflection may still occur at the SiC/Al2O3 interface and consequently, the measured fracture resistance is close to the initial SiC/Al2O3 interfacial fracture resistance (i.e., the He-Hutchinson criterion must be applied to this interface to determine 4 mm 2 mm 2 mm 2−3 mµ porous alumina interphase 70 mm Fig. 1. A schematic depiction of the SiC/porous Al2O3 laminate used in this study. oxidation time (hours) Γi (J/m2) Γi / Γf 0 5 10 15 20 0 2 4 6 8 10 12 14 16 0 0.1 0.2 0.3 0.4 0.5 0.6 Fig. 2. Gi versus time of air exposure. May 2005 Communications of the American Ceramic Society 1363
1364 Communications of the American Ceramic Society Vol. 88. No 5 likely to happen and a simple application of the He-Hutchinson criterion may no longer be valid. The experiments suggest that crack deflection in the reaction layer/Al2O3 region leads to a Elastic mismatch between the oxidation-affected interlayer and the sic lamina and residual stresses at the interface also affect crack defection. As oxidation results rt of the space being filled with a glassy phase, the elastic mismatch be- ween the Sic and the interlayer is likely to be reduced, resulting n a reduced threshold for crack deflection. 23 Also. the debond interface will experience a greater tensile stress for the same rea- Sic son(as the effective thermal expansion coefficient of the inter- layer is increased upon oxidation ), and this too produces a lower threshold for crack deflection. Both factors favor crack de flection occurring at"weak"interfaces located away from the bottom uncracked SiC surface. The He-Hutchinson criterion is trictly applicable only for bi-material interfaces and for a small deflected/penetrated crack. Our previous study of interfacial fracture in Al,O/porous AlO3 laminates provides a very cood fit to the He-Hutchinson criterion. However, experimen al results from another study with a non-oxidized SIC/ AlO3 amorphous lave system show crack deflection for interfaces stronger than the hreshold limit set by the He-Hutchinson criterion. 9 with the deviation from this threshold being larger than can be explained by finite interphase thickness effects. The effect of multi-layered interphases on crack defection has been analyzed using isotrop- ic linear elasticity, and cracks are found to be attracted to the weakest interface within the multi-layered interphase. Our results with oxidized sic laminates seem to be in contrast to the general opinion held in the case of fiber-reinforced SiC composites, where crack deflection is observed to be adverse ffected by the formation of Sio, at the fiber-matrix interface While most of this work considers C and Bn interface layers that volatilize during oxidation, our observations also appear to alumina differ from fiber-reinforced SiC composites with alumina inter phases, where fiber pull-out on fracture surfaces is not observed n composites that have been exposed to an oxidizing environ- ment at elevated temperature. This result was attributed to the strong interfacial bonds that are formed because of reaction be- tween the Nicalon fibers and the sic matrix. Note however that the strong fiber/matrix interface bond requires an oxide plug to be formed by reaction, and the traditional view does not apply to situations where the oxide layer does not extend com- glassy pletely across the thickness of the interphase. In considering our current results, it is interesting to note that reaction layers at phase interfaces that are undesirable because of their deleterious effect on the environmental resistance of the fiber may actually be de- sirable with regard to fracture resistance. The oxidation-induced decrease in fracture resistance in Fig. 2 is interesting and unexpected; more detailed information Fig 3. Transmission electron microscopy microgray on these types of effects is desirable. In particular, it is important terface for a laminate oxidized at 500C for 3 h, showing(a)al to consider the effect of higher oxidation temperatures and/ phous reaction layer at the SiC/Al2O3 interface and(b)An or longer oxidation exposures on both oxidation resistance grain and the amorphous layer stuck to the debond interface. The un- labeled areas in(a)and(b) represent void space. and mechanical properties. For good oxidation resistance, rap- d pore/crack sealing at the ends of the interphase is desirable Our results show that the presence of a thin reaction layer is responsible for the deflection behavior seen in these systems. In whether crack deflection or penetration should occur). For view of these two arguments, an effective strategy for producir thicker reaction layers, crack deflection is more likely to occur weak and oxidation resistant interphase in this system might away from the Sic interface. The limited microscopy result nclude limited low temperature oxidation to produce a thin presented in Fig 3 clearly indicate that both alumina grains and oxide layer and sufficiently rapid pore/crack sealing at the sur- rphous glassy phase are stuck to the lower, uncracked Sic face to provide oxidation protection. Such a strategy has been lamina, on fracture. This can be explained if debonding occurred proposed in the literature, and has been explored by us through at an interface within the porous interlayer that is terminated by theoretical calculations presented in another study. Note, how both the porous alumina and the glassy phase. It is also possible ever, that Sic composites are often considered for applications here fatigue resistance and reliability under thermal cycling are tructural inhomogeneities produced during the processing of necessary requirements. More work is required to evaluate the the laminates, and that low temperature oxidation only allows suitability of this strategy when such effects are significant for a finer control of the interface fracture resistance. Neverthe- Sic composites are generally considered for use at tempera- less, either mechanism is consistent with debonding occurring at tures higher than the processing temperatures used here. A an interface other than the uncracked SiC lamina/porous inter- higher temperatures, the pore space in the interlayer may be layer interface. For thicker reaction layers, this is even more rapidly filled up and or reactions between Al2O3 and Sic ca
whether crack deflection or penetration should occur). For thicker reaction layers, crack deflection is more likely to occur away from the SiC interface. The limited microscopy results presented in Fig. 3 clearly indicate that both alumina grains and amorphous glassy phase are stuck to the lower, uncracked SiC lamina, on fracture. This can be explained if debonding occurred at an interface within the porous interlayer that is terminated by both the porous alumina and the glassy phase. It is also possible that the pulled-out porous alumina grains result from microstructural inhomogeneities produced during the processing of the laminates, and that low temperature oxidation only allows for a finer control of the interface fracture resistance. Nevertheless, either mechanism is consistent with debonding occurring at an interface other than the uncracked SiC lamina/porous interlayer interface. For thicker reaction layers, this is even more likely to happen and a simple application of the He-Hutchinson criterion may no longer be valid. The experiments suggest that crack deflection in the reaction layer/Al2O3 region leads to a reduction in the measured value for Gi. Elastic mismatch between the oxidation-affected interlayer and the SiC lamina and residual stresses at the interface also affect crack deflection. As oxidation results in a part of the pore space being filled with a glassy phase, the elastic mismatch between the SiC and the interlayer is likely to be reduced, resulting in a reduced threshold for crack deflection.23 Also, the debond interface will experience a greater tensile stress for the same reason (as the effective thermal expansion coefficient of the interlayer is increased upon oxidation), and this too produces a lower threshold for crack deflection.24 Both factors favor crack de- flection occurring at ‘‘weak’’ interfaces located away from the bottom uncracked SiC surface. The He-Hutchinson criterion is strictly applicable only for bi-material interfaces and for a small deflected/penetrated crack. Our previous study of interfacial fracture in Al2O3/porous Al2O3 laminates provides a very good fit to the He-Hutchinson criterion.9 However, experimental results from another study with a non-oxidized SiC/Al2O3 system show crack deflection for interfaces stronger than the threshold limit set by the He-Hutchinson criterion,19 with the deviation from this threshold being larger than can be explained by finite interphase thickness effects. The effect of multi-layered interphases on crack deflection has been analyzed using isotropic linear elasticity,25 and cracks are found to be attracted to the weakest interface within the multi-layered interphase. Our results with oxidized SiC laminates seem to be in contrast to the general opinion held in the case of fiber-reinforced SiC composites, where crack deflection is observed to be adversely affected by the formation of SiO2 at the fiber–matrix interface.4 While most of this work considers C and BN interface layers that volatilize during oxidation, our observations also appear to differ from fiber-reinforced SiC composites with alumina interphases, where fiber pull-out on fracture surfaces is not observed in composites that have been exposed to an oxidizing environment at elevated temperature.17 This result was attributed to the strong interfacial bonds that are formed because of reaction between the Nicalont fibers and the SiC matrix. Note, however, that the strong fiber/matrix interface bond requires an oxide plug to be formed by reaction, and the traditional view does not apply to situations where the oxide layer does not extend completely across the thickness of the interphase. In considering our current results, it is interesting to note that reaction layers at interfaces that are undesirable because of their deleterious effect on the environmental resistance of the fiber may actually be desirable with regard to fracture resistance. The oxidation-induced decrease in fracture resistance in Fig. 2 is interesting and unexpected; more detailed information on these types of effects is desirable. In particular, it is important to consider the effect of higher oxidation temperatures and/ or longer oxidation exposures on both oxidation resistance and mechanical properties. For good oxidation resistance, rapid pore/crack sealing at the ends of the interphase is desirable. Our results show that the presence of a thin reaction layer is responsible for the deflection behavior seen in these systems. In view of these two arguments, an effective strategy for producing a weak and oxidation resistant interphase in this system might include limited low temperature oxidation to produce a thin oxide layer and sufficiently rapid pore/crack sealing at the surface to provide oxidation protection. Such a strategy has been proposed in the literature,7 and has been explored by us through theoretical calculations presented in another study.16 Note, however, that SiC composites are often considered for applications where fatigue resistance and reliability under thermal cycling are necessary requirements. More work is required to evaluate the suitability of this strategy when such effects are significant. SiC composites are generally considered for use at temperatures higher than the processing temperatures used here. At higher temperatures, the pore space in the interlayer may be rapidly filled up and/or reactions between Al2O3 and SiC can Fig. 3. Transmission electron microscopy micrographs of the debond interface for a laminate oxidized at 5001C for 3 h, showing (a) An amorphous reaction layer at the SiC/Al2O3 interface and (b) An alumina grain and the amorphous layer stuck to the debond interface. The unlabeled areas in (a) and (b) represent void space. 1364 Communications of the American Ceramic Society Vol. 88, No. 5
May 2005 Communications of the American Ceramic Society 1365 ccur. thus altering the interfacial fracture resistance dramati- Naslain,"Fiber-Matrix Interfaces and Interphases in Ceramic Matrix ally. Clearly, other systems and or mechanisms will then be needed to obtain optimal fracture and oxidation properties In conclusion, SiC/porous alumina interphase laminates were trated Composites with Weak and Strong Interfaces, "J Am Ceram Soc.81 oxidized in air at 500C and the measured interfacial fracture resistance of these laminates decreased with prolonged oxidation I4R.J. Kerans and T. A. Parthasarathy. "Crack Deflection in Ceramic Com- exposures. This suggests some interesting processing strategies atings for Non-Oxide Cer for fabricating fracture and oxidation resistant composites opposites, J. An. Ceram. Soc., 80[12] 3253-7 Laminated Ceramic Composites, "J. Am. Ceram. Soc.(2004)(in press). References A. G. Evans and D. B. Marshall. "The Mechanical Behavior of Ceramic Ma- tinuous Fiber Ceramic Composites. Ceram Eng. Sci. Proc., 16[14]389-99(1995). "A J Phillipps, w.J. Clegg, and T. w. Cline. ""Fracture Behavior of Ceramic Tested at High Temperature in Air, "J.Am. Ceram. Soc.83[12] Laminates in Bending-ll. Comparison of Model Predictions with Experimenta I9R. Krishnamurthy and B. W. Sheldon, "Experimental Observations of High H lu ands.M. Hsu. Fractur Behavior o muilt eo socon Nitnde/Boron Fracture Resistances in SiC/AlO3 Laminated Composites. " J. Am. Ceram. Soc. G. Charalambides, J. Lund. A. G. Evans and R. M. McMeekin. " A Test and Kinetics of I-D SiC/C/SiC Composite Materials: I, An Experimental Ap- pecimen for Determining the Fracture Resistance of Bimaterial Interfaces. J.Am. Cera.Soc,77p2459-6601994 1. Appl. Mech.56782(9 H C Cao, J. Lund, and A G. Evans. ""Development of C/SiC Composite Materials: Il, Modeling, "J. Am. Ceram. Soc., 77 [2]467-80 a Test Method for Measuring the Mixed Mode Fracture Resistance, "J.Appl. fech8.26983(1989) N. S. Jacobson, G. N. Morscher. D. R. Bryant, and R. E. Tressler. ""High M. Y. He and J. w. Hutchinson. "Crack Deflection at an Interface Between rature Oxidation of Boron Nitride Part Il: BN Layers in Composites. Dissimilar Elastic Materials, Int J Solids Struct. 25 [911053-67(1989) gam.Soe,82间6l147-82(199 2z. Suo and J. w n,""Sandwich Test Specimens for Measuring In- J. Kerans, " Issues in the Control of Fiber-Matrix Interface Properties in terface Crack Toughne Sc.Eng,A107,1354301989 Fiber coati ris 1. P.A. Lofvander, A. G. Evans, E. Bischof, and M. L. Emiliani, Materials: Role of Residual Stress. IntJ. Sol- ting Concepts for Brittle-Matrix Composites, J. Am. Ceram. Soc., 76 ids Struct.,31口243443-55(1989) N. Carriere, E. Martin, and J. Lamon, The Influence of the Interphase and J. O'Brien and B w. Sheldon,"A Porous Alumina Coating Associated Interfaces on the deflectio latrix Cracks in Ceramic Matrix Fracture Resistance for Alumina Composites, "J. Am. Ceram Soc., 82[12] Kriven and S. J. Lee, ""Mullite/Cordierite Laminates with B-o C. Droillard and J. Lamon. ""Fracture Toughness of 2D Woven SiC/SIC CVI Cristobalite Transformation Weakened Interphases, Ceram. Eng. Sci. Proc., opposites with Multilayered Interphases, J. Am. Ceram. Soc., 7914] 849-58 9.305-16(1998) w.M. Kriven "Displacive Phase Transformations and Their Applications in C. Droillard. J. Lamon, and X. Bourrat, "Strong Interphases in CMCs: A Condition for Efficient Multilayered Interphases. Mater. Res. Soc. Symp. Proc. erface Engineering in Oxide Fib- er/Oxide Matrix Composites, Int Mater. Rev 45[5]165-89(2000)
occur, thus altering the interfacial fracture resistance dramatically. Clearly, other systems and/or mechanisms will then be needed to obtain optimal fracture and oxidation properties.26–28 In conclusion, SiC/porous alumina interphase laminates were oxidized in air at 5001C and the measured interfacial fracture resistance of these laminates decreased with prolonged oxidation exposures. This suggests some interesting processing strategies for fabricating fracture and oxidation resistant composites. References 1 A. G. Evans and D. B. Marshall, ‘‘The Mechanical Behavior of Ceramic Matrix Composites,’’ Acta Metall. Mater., 37 [10] 2567–83 (1989). 2 A. J. Phillipps, W. J. Clegg, and T. W. Cline, ‘‘Fracture Behavior of Ceramic Laminates in Bending-II. Comparison of Model Predictions with Experimental Data,’’ Acta Metall. Mater., 41 [3] 819–27 (1993). 3 H. Liu and S. M. Hsu, ‘‘Fracture Behavior of Multilayer Silicon Nitride/Boron Nitride Ceramics,’’ J. Am. Ceram. Soc., 37 [10] 2567–83 (1989). 4 L. Filipuzzi, G. Camus, R. Naslain, and J. Thebault, ‘‘Oxidation Mechanisms and Kinetics of 1-D SiC/C/SiC Composite Materials: I, An Experimental Approach,’’ J. Am. Ceram. Soc., 77 [2] 459–66 (1994). 5 L. Filipuzzi and R. Naslain, ‘‘Oxidation Mechanisms and Kinetics of 1D-SiC/ C/SiC Composite Materials: II, Modeling,’’ J. Am. Ceram. Soc., 77 [2] 467–80 (1994). 6 N. S. Jacobson, G. N. Morscher, D. R. Bryant, and R. E. Tressler, ‘‘High Temperature Oxidation of Boron Nitride Part II: BN Layers in Composites,’’ J. Am. Ceram. Soc., 82 [6] 1473–82 (1999). 7 R. J. Kerans, ‘‘Issues in the Control of Fiber-Matrix Interface Properties in Ceramic Composites,’’ Scr. Metall. Mater., 31 [8] 1079–84 (1994). 8 J. B. Davis, J. P. A. Lofvander, A. G. Evans, E. Bischoff, and M. L. Emiliani, ‘‘Fiber Coating Concepts for Brittle-Matrix Composites,’’ J. Am. Ceram. Soc., 76 [5] 1249–57 (1993). 9 M. J. O’Brien and B. W. Sheldon, ‘‘A Porous Alumina Coating with Tailored Fracture Resistance for Alumina Composites,’’ J. Am. Ceram. Soc., 82 [12] 3567–74 (1999). 10C. Droillard and J. Lamon, ‘‘Fracture Toughness of 2D Woven SiC/SiC CVI Composites with Multilayered Interphases,’’ J. Am. Ceram. Soc., 79 [4] 849–58 (1996). 11C. Droillard, J. Lamon, and X. Bourrat, ‘‘Strong Interphases in CMCs: A Condition for Efficient Multilayered Interphases,’’ Mater. Res. Soc. Symp. Proc., 365, 371–6 (1995). 12R. Naslain, ‘‘Fiber–Matrix Interfaces and Interphases in Ceramic Matrix Composites Processed by CVI,’’ Compos. Interfaces, 1 [3] 253–86 (1993). 13F. Rebillat, J. Lamon, R. Naslain, E. Lara-Curzio, M. K. Ferber, and T. M. Bessman, ‘‘Properties of Multilayered Interphases in SiC/SiC Chemical-Vapor-In- filtrated Composites with Weak and Strong Interfaces,’’ J. Am. Ceram. Soc., 81 [9] 2315–26 (1998). 14R. J. Kerans and T. A. Parthasarathy, ‘‘Crack Deflection in Ceramic Composites and Fiber Coating Design Criteria,’’ Composites, Part A, 30, 521–4 (1999). 15K. L. Luthra, ‘‘Oxidation-Resistant Fiber Coatings for Non-Oxide Ceramic Composites,’’ J. Am. Ceram. Soc., 80 [12] 3253–7 (1997). 16R. Krishnamurthy and B. W. Sheldon, ‘‘Oxidation of SiC/Porous Al2O3 Laminated Ceramic Composites,’’ J. Am. Ceram. Soc., (2004) (in press). 17S. Shanmugham, D. P. Stinton, F. Rebillat, A. Bleier, T. M. Besmann, E. Lara-Curzio, and P. K. Liaw, ‘‘Oxidation-Resistant Interfacial Coatings for Continuous Fiber Ceramic Composites,’’ Ceram. Eng. Sci. Proc., 16 [4] 389–99 (1995). 18M. J. O’Brien, F. M. Capaldi, and B. W. Sheldon, ‘‘A Layered Alumina Composite Tested at High Temperature in Air,’’ J. Am. Ceram. Soc., 83 [12] 3033–40 (2000). 19R. Krishnamurthy and B. W. Sheldon, ‘‘Experimental Observations of High Fracture Resistances in SiC/Al2O3 Laminated Composites,’’ J. Am. Ceram. Soc., 84 [10] 2451–3 (2001). 20P. G. Charalambides, J. Lund, A. G. Evans, and R. M. McMeeking, ‘‘A Test Specimen for Determining the Fracture Resistance of Bimaterial Interfaces,’’ J. Appl. Mech., 56, 77–82 (1989). 21P. G. Charalambides, H. C. Cao, J. Lund, and A. G. Evans, ‘‘Development of a Test Method for Measuring the Mixed Mode Fracture Resistance,’’ J. Appl. Mech., 8, 269–83 (1989). 22M. Y. He and J. W. Hutchinson, ‘‘Crack Deflection at an Interface Between Dissimilar Elastic Materials,’’ Int. J. Solids Struct., 25 [9] 1053–67 (1989). 23Z. Suo and J. W. Hutchinson, ‘‘Sandwich Test Specimens for Measuring Interface Crack Toughness,’’ Mater. Sci. Eng., A107, 135–43 (1989). 24M. Y. He, A. G. Evans, and J. W. Hutchinson, ‘‘Crack Deflection at an Interface Between Dissimilar Elastic Materials: Role of Residual Stress,’’ Int. J. Solids Struct., 31 [24] 3443–55 (1989). 25N. Carriere, E. Martin, and J. Lamon, ‘‘The Influence of the Interphase and Associated Interfaces on the Deflection of Matrix Cracks in Ceramic Matrix Composites,’’ Composites, Part A, 31, 1179–90 (2000). 26W. M. Kriven and S. J. Lee, ‘‘Mullite/Cordierite Laminates with b-a Cristobalite Transformation Weakened Interphases,’’ Ceram. Eng. Sci. Proc., 19, 305–16 (1998). 27W. M. Kriven, ‘‘Displacive Phase Transformations and Their Applications in Structural Ceramics,’’ J. Phys. IV, Colloq., C8, 101–10 (1995). 28K. K. Chawla, C. Coffin, and Z. R. Xu, ‘‘Interface Engineering in Oxide Fiber/Oxide Matrix Composites,’’ Int. Mater. Rev., 45 [5] 165–89 (2000). & May 2005 Communications of the American Ceramic Society 1365