Availableonlineatwww.sciencedirect.com SCIENCE E噩≈S Journal of the European Ceramic Society 23(2003)613-617 www.elsevier.com/locate/j Pest-resistance in SiC/BN/SiC composites Linus U.j.T. ogbuji* QSS Inc, NASA Glenn Research Center, Cleveland, Ohio, US.A Received 4 October 2001; received in revised form I July 2002; accepted 14 July 2002 Abstract State-of-the-art non-oxide ceramic-matrix composites(consisting of Sic fibers, cvi-BN interphase coating, and mi-SiC matrix) exhibit excellent mechanical properties at room temperature, as well as above 1000C (where oxidation easily seals flaws with silica); but they are prone to pest degradation at intermediate temperatures in an oxidizing environment, and especially so in the fast, moist flame of a jet engine. Two modes of pest may be distinguished in these composites. The more severe promoted by extraneous factors, like a layer of elemental carbon underlying the Bn interphase and undermining dation resistance. It is shown that, when care is taken to exclude such a carbon layer, SiC/BN SiC composite can 100-h exposure in a burner rig without noticeable loss of strength or strain to fracture. C 2003 Elsevier Science Ltd. All rights reserved Keywords: BN interphase: Composites: Oxidation: Pest; SiC/SIC 1. Introduction CMC is a SylramicM/BN/mi-SiC composite. The sali ent points of its development under a succession of service in advanced engines has been known for paper: programs are highlighted in a companion The potential of ceramic-matrix composites(CMCs) decades, but realization of that promise has been fru- While significant progress has been made on the Cmc strated by a slow developmental curve. Most designs of matrix and fibers, development of the CMC interphase advanced turbines for aircraft engines or land-based (material at the fiber/matrix interface) has been slow, power generators assume significantly higher operating though interphase problems were evident from the start temperatures than in current engines, both for gains in Initially, carbon was used as an interphase in SiC/SiC thermodynamic efficiency and for mandated reductions composites due to its excellent compliance, but that in exhaust emissions. Hence, it is assumed that hot sec- advantage was offset by its oxidative volatility, which tions of advanced turbines (especially combustor liners, rendered carbon unsuitable at intermediate tempera- nozzles, and vanes) will be made of ceramic-matrix tures. Above 500C oxidative loss of interphase carbon composites(CMCs). CMC development has been led by becomes catastrophic only above 1000 C is protection progress in the fiber and matrix constituents For non- achieved as Sio from matrix oxidation becomes sub- oxide CMCs, reaction-bonded silicon nitride(rBsn) stantial enough to seal off ingress of oxidants. Boron and chemical-vapor-infiltrated silicon carbide(cvi-SiC) nitride, the currently preferred interphase, also oxidizes proved unsuitable as matrix materials and melt-infiltrated almost as readily as carbon in the same intermediate silicon carbide(mi-SiC) became the matrix of choice; temperature range, so that SiC/BN/ SiC composites can similarly, "fat""fibers(e.g. SCS-0 and SCS-6)have been pest as severely as Sic/C/SiC composites.Still,BN superseded by thin fibers, with preference shifting succes- remains the interphase material of choice because its sively from NicalonM to Hi-NicalonTM and to Sylra- shortcomings seem, in principle, remediable. However micM varieties of increasingly refined chemistry and the remedies have been slow in coming, and pesting microstructure. The current state-of-the-art non-oxide remains a major obstacle to SiC/BN/SiC utilization *Tel:+1-216-433-6463;fax:+1-216-43-554 also involves selective attack of the interphase by ambi ent oxidants. However, whereas pesting of Sic/C/SiC 0955-2219/03/S. see front matter C 2003 Elsevier Science Ltd. All rights reserved. PII:S0955-2219(02)00268-6
Pest-resistance in SiC/BN/SiC composites Linus U.J.T. Ogbuji* QSS Inc., NASA Glenn Research Center, Cleveland, Ohio, USA Received4 October 2001; receivedin revisedform 1 July 2002; accepted14 July 2002 Abstract State-of-the-art non-oxide ceramic-matrix composites (consisting of SiC fibers, cvi-BN interphase coating, and mi-SiC matrix) exhibit excellent mechanical properties at room temperature, as well as above 1000 C (where oxidation easily seals flaws with silica); but they are prone to pest degradation at intermediate temperatures in an oxidizing environment, and especially so in the fast, moist flame of a jet engine. Two modes of pest may be distinguished in these composites. The more severe type of pest is promoted by extraneous factors, like a layer of elemental carbon underlying the BN interphase and undermining its intrinsic oxidation resistance. It is shown that, when care is taken to exclude such a carbon layer, SiC/BN/SiC composite can easily survive a 100-h exposure in a burner rig without noticeable loss of strength or strain to fracture. # 2003 Elsevier Science Ltd. All rights reserved. Keywords: BN interphase; Composites; Oxidation; Pest; SiC/SiC 1. Introduction The potential of ceramic-matrix composites (CMCs) for service in advanced engines has been known for decades, but realization of that promise has been frustrated by a slow developmental curve. Most designs of advanced turbines for aircraft engines or land-based power generators assume significantly higher operating temperatures than in current engines, both for gains in thermodynamic efficiency and for mandated reductions in exhaust emissions. Hence, it is assumedthat hot sections of advanced turbines (especially combustor liners, nozzles, andvanes) will be made of ceramic-matrix composites (CMCs). CMC development has been led by progress in the fiber andmatrix constituents. For nonoxide CMCs, reaction-bonded silicon nitride (RBSN) andchemical-vapor-infiltratedsilicon carbide (cvi-SiC) provedunsuitable as matrix materials andmelt-infiltrated silicon carbide (mi-SiC) became the matrix of choice; similarly, ‘‘fat’’ fibers (e.g. SCS-0 andSCS-6) have been superseded by thin fibers, with preference shifting successively from NicalonTM to Hi-NicalonTM andto SylramicTM varieties of increasingly refinedchemistry and microstructure. The current state-of-the-art non-oxide CMC is a SylramicTM/BN/mi-SiC composite. The salient points of its development under a succession of NASA programs are highlightedin a companion paper.1 While significant progress has been made on the CMC matrix andfibers, development of the CMC interphase (material at the fiber/matrix interface) has been slow, though interphase problems were evident from the start. Initially, carbon was usedas an interphase in SiC/SiC composites due to its excellent compliance, but that advantage was offset by its oxidative volatility, which rendered carbon unsuitable at intermediate temperatures. Above 500 C oxidative loss of interphase carbon becomes catastrophic;2 only above 1000 C is protection achievedas SiO2 from matrix oxidation becomes substantial enough to seal off ingress of oxidants. Boron nitride, the currently preferred interphase, also oxidizes almost as readily as carbon in the same intermediate temperature range, so that SiC/BN/SiC composites can pest as severely as SiC/C/SiC composites.3 Still, BN remains the interphase material of choice because its shortcomings seem, in principle, remediable. However, the remedies have been slow in coming, and pesting remains a major obstacle to SiC/BN/SiC utilization. As in a SiC/C/SiC composite, pesting of SiC/BN/SiC also involves selective attack of the interphase by ambient oxidants. However, whereas pesting of SiC/C/SiC 0955-2219/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved. PII: S0955-2219(02)00268-6 Journal of the European Ceramic Society 23 (2003) 613–617 www.elsevier.com/locate/jeurceramsoc * Tel.: +1-216-433-6463; fax: +1-216-433-5544. E-mail address: thomas-ogbuji@grc.nasa.gov (L.U.J.T. Ogbuji).
L.U.J.T. Ogbuji/ Journal of the European Ceramic Society 23(2003)613-617 composites involves conversion of the interphase to CO leaving no solid residue, SiC/BN/ SIC pesting is more complex. It seems to involve the synergistic oxidation of constituents in the interphase area. Whether Sio2 forms first via SiC oxidation or it is B,O3 via bn oxi- lation(depending on local conditions), the net result is that the compliant bn interphase is removed or eplaced by stiff, brittle silica which "cements"the fibers down and hence destroys composite behavior. The likely reaction sequence in SiC/BN SiC pesting is sum- marized in Table I. Where the Bn interphase is inter ected by a crack open to the surface it is oxidized locally to B2O3. The boria may be volatilized by hydro- lysis with ambient moisture if the temperature is low, brittle fracture zone fiber pull-out leaving a trench around the fiber; or it may react with Fig. 1. Normal or intrinsic mode of pest damage in SiC/BN/SiC. the surrounding SiC to form borosilicate.( Compare, for instance, Fig. 7a and b in ref 6. Hydrolysis of the bor- from a carbon-rich fiber or from sizing char yield In the silicate by ambient moisture results in a cementitious fast flame of a burner rig(which simulates a jet engine plug of silica in place of the BN the carbon layer rapidly burns off, leaving an annular 4 The foregoing describes what we may call the"nor- trench around the bn as a pathway for deep ingress of al"pest mode of SiC/BN/SiC: degradation starts at the ambient. The same reactions occur as before e surface/edge and proceeds inwards at a rate limited(sequence 0-3 in Fig. 2) but in this case it is exacerbated by intrinsic oxidation resistance of the BN. The fracture by the 10% moisture in the combustion product of the surface looks like a picture frame, as shown in Fig. 1, hydrocarbon fuel, by h flame velocity, and by the with a border of pested material(characterized by brit- huge surface area of bn exposed to ambient tle/flat fracture)around a shrinking core of undamaged attack. 12.3 The bn surface area is the biggest facto material (characterized by composite behavior and The interphase may be likened to a thin sheet of pape fibrous fracture, with copious fiber pull-out). Evans and and the effect of an underlying layer of fugitive carbo co-workers reported that mode of pesting in SiC/SiC and Sic/glass composites, and illustrated the picture- frame effect with a SiC/MAS (magnesium aluminosili- aLambientd D cate) minicomposite in which the sample was found to exhibit"a time-dependent(residual)strength". o Mor- scher and co-workers have reported the same phenom enon in SiC/BN/ SiC composites as well as in SiC/BN/ SiC and SiC/C/SiC minicomposites and correlated sample lure to growth of the rested zone at the borders From our work on SiC/BN SiC composites it has become obvious that a second and more pernicious mode of pest occurs in these composites. The details 0 have been presented elsewhere. This mode is driven by SN Carbon cvi-Sic lm SiC Fiber BN extrinsic factors at work in the interphase region as a Borosilicate(B content increases with depth) whole. It occurs in SIC/BN/SiC materials where a con- Fig. 2. Severe/extrinsic mode of Sic/BN/SiC pest, which is assisted by tinuous or skeletal film of carbon forms under the bn. free carbon able i Oxidative degradation of the Sic/BNSiC interphase I Oxidation of bn to BO3 This is favored over SiC oxidation except in very low oxygen partial pressures [Ref. 4. B2O3 dissolves fiber and cvi-SiC to form borosilicate liquid Borosilicate has a very low-melting eutectic composition(near 400C) 3 Borosilicate is hydrolyzed by HO, releasing B(OH)(g)[Ref 7] H2O is from combustion of hydrocarbon Boron loss increases borosilicate viscosity. 4 The residue is Sio(s) The SiO, residue bonds fibers to the cvi-SiC. Holes in the sio indicate emission of gaseous productsi.e boron hydroxides
composites involves conversion of the interphase to CO, leaving no solidresidue, SiC/BN/SiC pesting is more complex. It seems to involve the synergistic oxidation of all constituents in the interphase area. Whether SiO2 forms first via SiC oxidation4 or it is B2O3 via BN oxidation5 (depending on local conditions), the net result is that the compliant BN interphase is removedor replacedby stiff, brittle silica which ‘‘cements’’ the fibers down and hence destroys composite behavior. The likely reaction sequence in SiC/BN/SiC pesting is summarizedin Table 1. Where the BN interphase is intersectedby a crack open to the surface, it is oxidized locally to B2O3. The boria may be volatilizedby hydrolysis with ambient moisture if the temperature is low, leaving a trench aroundthe fiber;6 or it may react with the surrounding SiC to form borosilicate. (Compare, for instance, Fig. 7a andb in ref. 6.) Hydrolysis of the borosilicate by ambient moisture results in a cementitious plug of silica in place of the BN interphase.58 The foregoing describes what we may call the ‘‘normal’’ pest mode of SiC/BN/SiC: degradation starts at the surface/edge and proceeds inwards at a rate limited by intrinsic oxidation resistance of the BN. The fracture surface looks like a picture frame, as shown in Fig. 1, with a border of pested material (characterized by brittle/flat fracture) arounda shrinking core of undamaged material (characterizedby composite behavior and fibrous fracture, with copious fiber pull-out). Evans and co-workers reportedthat mode of pesting in SiC/SiC andSiC/glass composites,9 andillustratedthe pictureframe effect with a SiC/MAS (magnesium aluminosilicate) minicomposite in which the sample was foundto exhibit ‘‘a time-dependent (residual) strength’’.10 Morscher andco-workers have reportedthe same phenomenon in SiC/BN/SiC composites6 as well as in SiC/BN/SiC andSiC/C/SiC minicomposites11 andcorrelatedsample failure to growth of the pestedzone at the borders. From our work on SiC/BN/SiC composites it has become obvious that a secondandmore pernicious mode of pest occurs in these composites. The details have been presentedelsewhere.5 This mode is driven by extrinsic factors at work in the interphase region as a whole. It occurs in SiC/BN/SiC materials where a continuous or skeletal film of carbon forms under the BN, from a carbon-rich fiber or from sizing char yield. In the fast flame of a burner rig (which simulates a jet engine) the carbon layer rapidly burns off, leaving an annular trench aroundthe BN as a pathway for deep ingress of the ambient. The same reactions occur as before (sequence 0–3 in Fig. 2) but in this case it is exacerbated by the 10% moisture in the combustion product of the hydrocarbon fuel, by the high flame velocity, and by the huge surface area of BN exposedto ambient attack.5,12,13 The BN surface area is the biggest factor. The interphase may be likenedto a thin sheet of paper andthe effect of an underlying layer of fugitive carbon Table 1 Oxidative degradation of the SiC/BN/SiC interphase Step Process Comments 1 Oxidation of BN to B2O3 This is favoredover SiC oxidation except in very low oxygen partial pressures [Ref. 4]. 2 B2O3 dissolves fiber and cvi-SiC to form borosilicate liquidBorosilicate has a very low-melting eutectic composition (near 400 C). 3 Borosilicate is hydrolyzed by H2O, releasing B(OH)4(g) [Ref. 7] H2O is from combustion of hydrocarbon. Boron loss increases borosilicate viscosity. 4 The residue is SiO2(s) The SiO2 residue bonds fibers to the cvi-SiC. Holes in the SiO2 indicate emission of gaseous products—i.e. boron hydroxides Fig. 1. Normal or intrinsic mode of pest damage in SiC/BN/SiC. Fig. 2. Severe/extrinsic mode of SiC/BN/SiC pest, which is assisted by free carbon. 614 L.U.J.T. Ogbuji / Journal of the European Ceramic Society 23 (2003) 613–617
L.U.J.T. Ogbuji/ Journal of the European Ceramic Society 23(2003)613-617 is to expose that sheet to extensive attack on its broad fibers. The SylramicTM/BN SiC materials include one surface(instead of its edge as in"normal"pest). That made with fibers that had been covered with oxide- accelerated mode of pest is the main thrust of this based sizing and another with alcohol-based sizing In paper. contrast. all Hi-Nicalon TM fibers had been covered with SiC/BN/ SiC processing occurs in several steps, 4 some alcohol-based g. Scanning electron microscopy of which can degrade interphase integrity: (a)coverin SEM)and Auger electron spectroscopy(AES) of as the fibers with protective"sizing;(b)weaving or braid received samples showed substantial carbon between the ing the fiber tows into a preform; (c)desizing;( d)coating fiber and its bn coating in composites containing either fibers with chemical-vapor-infiltrated BN (cvi-BN) inter- Hi-Nicalon M fibers or oxide-sized SylramicTM fibers phase;(e) protecting the bn by chemical-vapor infiltra- The carbon was in essentially continuous layers in both tion of Sic cladding(cvi-SiC);(f impregnating the cases, albeit skeletal on the oxide-sized SylramicTM preform with a slurry, including carbon to be converted fibers to SiC; and (g) melt-infiltration of Si to form the Sic The samples were made by Honeywell Advanced matrix (mi-SiC). Table 2 outlines the ways in which Composites, Inc, in the approximate processing steps these steps can influence interphase integrity and hence summarized above. Work in our laboratories have impact pest behavior. Some recent remedies aim at shown that a simple and direct way to assess the pest improving the Bn quality: by crystallizing the bn (to behavior of these materials is to expose tensile bar spe densify it), 5 doping it with silicon, or reinforcing the cimens in the 0.3 Mach flame of an atmospheric-pres- normal BN with an adjoining layer of high-temperature sure burner rig(APBR)at 800C, determining their bn generated in-situ. 6 However, it is apparent that residual tensile strength and strain-to-fracture after control of extrinsic factors is still needed to protect the exposure, and analyzing the tensile fracture surfaces by BN layer whether or not its intrinsic properties are micrography and spectrometry. A 100-h APBR expo- ate for scr 2. Materials and methods 3. Results and discussion The four varieties of SiC/BN SiC studied, which are The only SiC/BN SiC varieties that did not undergo of the utmost interest in aerospace applications, are severe(extrinsic) pest in the burner rig(Table 3)are those described in Table 3, along with their respective severity of that featured (stoichiometric) SylramicM fibers with pesting as evidenced in strength loss. They include two PVA sizing; all composites made with(carbon-rich) made with Hi-NicalonTM and Hi-Nicalon(S)TM fibers that Nicalon fibers (i.e. Hi-NicalonTM or Hi-Nicalon(s) ontain 40 and 5% excess carbon, respectively, and as well as those made with PEO-sized SylramicTM fibers wo varieties made with the stoichiometric SylramicM sizing on the fibers exhibited severe pesting in our tests Possible interphase effects of composite processing Process Possible effects Choice of fiber Excess carbon in the fiber can migrate to the fiber surface ome sizing materials can char to carbon on the fiber 23456 izings can leave significant carbon residue on fiber. Interphase deposition Deposition conditions will determine interphase quality cvi-SiC deposition Tecnique can leave residue on interphase. Matrix infiltration High temperatures can cause carbon precipitation on fibers Table 3 Sic/BN/SiC materials tested after 100 h at 800C in the burner rig Fiber/sizing Carbon rength loss Comments under bn in APBr Hi-Nicalon/alcohol-based sizing Continuous 60% fibers became bonded together by SiO2 Hi-Nicalon(S)/alcohol-based sizing Continuous fibers became bonded together by SiO2 Sylramicoxide-based sizing Quasi-continuous -53% b of 405 MPa aaa fibers became bonded together by Sio Sylramic/alcohol-based sizing None None of BN/cvi-SiC interface arbon-rich in some cases(C/Si ratio 1.4) A SiO2 layer grew at that interface during BR exposure
is to expose that sheet to extensive attack on its broad surface (insteadof its edge as in ‘‘normal’’ pest). That acceleratedmode of pest is the main thrust of this paper. SiC/BN/SiC processing occurs in several steps,14 some of which can degrade interphase integrity: (a) covering the fibers with protective ‘‘sizing’’; (b) weaving or braiding the fiber tows into a preform; (c) desizing; (d) coating fibers with chemical-vapor-infiltratedBN (cvi-BN) interphase; (e) protecting the BN by chemical-vapor infiltration of SiC cladding (cvi-SiC); (f) impregnating the preform with a slurry, including carbon to be converted to SiC; and(g) melt-infiltration of Si to form the SiC matrix (mi-SiC). Table 2 outlines the ways in which these steps can influence interphase integrity andhence impact pest behavior. Some recent remedies aim at improving the BN quality: by crystallizing the BN (to densify it),15 doping it with silicon,8 or reinforcing the normal BN with an adjoining layer of high-temperature BN generatedin-situ.16 However, it is apparent that control of extrinsic factors is still needed to protect the BN layer whether or not its intrinsic properties are improved. 2. Materials and methods The four varieties of SiC/BN/SiC studied, which are of the utmost interest in aerospace applications, are described in Table 3, along with their respective severity of pesting as evidenced in strength loss. They include two made with Hi-NicalonTM andHi-Nicalon(S)TM fibers that contain 40 and 5% excess carbon, respectively, and two varieties made with the stoichiometric SylramicTM fibers. The SylramicTM/BN/SiC materials include one made with fibers that had been covered with oxidebasedsizing andanother with alcohol-basedsizing. In contrast, all Hi-NicalonTM fibers hadbeen coveredwith alcohol-basedsizing. Scanning electron microscopy (SEM) andAuger electron spectroscopy (AES) of asreceivedsamples showedsubstantial carbon between the fiber andits BN coating in composites containing either Hi-NicalonTM fibers or oxide-sized SylramicTM fibers. The carbon was in essentially continuous layers in both cases, albeit skeletal on the oxide-sized SylramicTM fibers. The samples were made by Honeywell Advanced Composites, Inc., in the approximate processing steps summarizedabove. Work in our laboratories have shown that a simple anddirect way to assess the pest behavior of these materials is to expose tensile bar specimens in the 0.3 Mach flame of an atmospheric-pressure burner rig (APBR) at 800 C, determining their residual tensile strength and strain-to-fracture after exposure, andanalyzing the tensile fracture surfaces by micrography andspectrometry. A 100-h APBR exposure has been foundadequate for screening purposes. 3. Results and discussion The only SiC/BN/SiC varieties that did not undergo severe (extrinsic) pest in the burner rig (Table 3) are those that featured(stoichiometric) SylramicTM fibers with PVA sizing; all composites made with (carbon-rich) Nicalon fibers (i.e. Hi-NicalonTM or Hi-Nicalon(S)TM) as well as those made with PEO-sized SylramicTM fibers sizing on the fibers exhibitedsevere pesting in our tests. Table 2 Possible interphase effects of composite processing Step Process Possible effects 1 Choice of fiber Excess carbon in the fiber can migrate to the fiber surface. 2 Choice of sizing Some sizing materials can char to carbon on the fiber. 3 De-sizing Sizings can leave significant carbon residue on fiber. 4 Interphase deposition Deposition conditions will determine interphase quality. 5 cvi-SiC deposition Tecnique can leave residue on interphase. 6 Matrix infiltration High temperatures can cause carbon precipitation on fibers. Table 3 SiC/BN/SiC materials testedafter 100 h at 800 C in the burner rig Fiber/sizing Carbon under BN Strength loss in APBR Comments Hi-Nicalon/alcohol-basedsizing Continuous 60% of 380 MPa After BR exposure adjacent fibers became bonded together by SiO2. Hi-Nicalon(S)/alcohol-basedsizing Continuous 56% of 395 MPa After BR exposure adjacent fibers became bonded together by SiO2. Sylramic/oxide-based sizing Quasi-continuous 53% of 405 MPa After BR exposure adjacent fibers became bonded together by SiO2. Sylramic/alcohol-basedsizing None None of 420 MPa BN/cvi-SiC interface was carbon-rich in some cases (C/Si ratio 1.4). A SiO2 layer grew at that interface during BR exposure. L.U.J.T. Ogbuji / Journal of the European Ceramic Society 23 (2003) 613–617 615
L.U.J.T. Ogbuji/ Journal of the European Ceramic Society 23(2003)613-617 The origin of the carbon depends on the fiber or coat ing. With carbon-rich fibers the excess carbon builds up between the fiber surface and Bn interphase; with oxide-sized fibers it comes from residual char yield upon de-sizing. Alcohol sizing seems to avoid the latter prob N lem, reflecting either an intrinsic superiority or a more negligible char yield, since fiber makers tend to make the sizing considerably thicker when it is oxide-based Fig 3 shows that in these composites(which have a high SiC fiber fiber fraction, >40 vol %fiber to maximize load bear ing) nearly all fibers are in contact with their neighbors, the number of contacts averaging 3.3 instead of the ideal zero. The higher-magnification image inset in this Fig. 4. Following burner rig exposure, a layer of silica covers the BN figure reveals a carbon sublayer under the bn inter- terphase in a SylramicM/BN/ SiC material that was carbon rich at phase-which probably links up with the carbon layers the BN cvi-SiC interface. on adjacent fibers. Contact of [0] and [90] tows in the third dimension extends the web of carbon throughout 4. Summary and conclusion the composite and leads to a rigid network of silica fol the l:g burn-out of carbon. This occurs most easily in Pest resistance of SiC/BN composites has been the high flame velocity of the burner rig but the effect studied by burner-rig exposure at intermediate tem has been suspected in Hi-NicalonTM/BN/SiC that pes- peratures, followed by tensile tests and microstructural ted during stress-rupture tests in air at intermediate characterization to determine residual strengths and temperatures. pest-related features. All samples reinforced with In some Sylramic/BN/SiC materials that did not Hi-NicalonM and Hi-Nicalon(S) M fibers exhibited the exhibit extrinsic pest, significant carbon enrichment was severe, extrinsic mode of pesting. Those with SylramicTM detected at the BN/cvi-SiC interface(see the bottom of fiber reinforcements resisted extrinsic pesting except in Table 3). After burner rig exposure that carbon-rich cases where the fibers had been sized with oxide-based layer was replaced by silica, as shown in Fig. 4. The sizing. The common characteristic in all cases where silica may afford protection to the bn and so prove extrinsic pesting occurred is that a film of carbon was beneficial to overall durability. However, its actual observed between the bn interphase and fiber, originat effect is yet to be determined ing from excess carbon in the Nicalon-type fibers and Carbon A85530kv127mm×200kSE(L)081 ig. 3. SEM image showing carbon beneath the BN layer in Hi-Nicalon(S)M/BN/SiC
The origin of the carbon depends on the fiber or coating. With carbon-rich fibers the excess carbon builds up between the fiber surface andBN interphase; with oxide-sized fibers it comes from residual char yield upon de-sizing. Alcohol sizing seems to avoid the latter problem, reflecting either an intrinsic superiority or a more negligible char yield, since fiber makers tend to make the sizing considerably thicker when it is oxide-based. Fig. 3 shows that in these composites (which have a high fiber fraction, 540 vol.% fiber to maximize loadbearing) nearly all fibers are in contact with their neighbors, the number of contacts averaging 3.3 insteadof the ideal zero. The higher-magnification image inset in this figure reveals a carbon sublayer under the BN interphase—which probably links up with the carbon layers on adjacent fibers. Contact of [0] and[90] tows in the third dimension extends the web of carbon throughout the composite andleads to a rigidnetwork of silica following burn-out of carbon. This occurs most easily in the high flame velocity of the burner rig but the effect has been suspectedin Hi-NicalonTM/BN/SiC that pested during stress-rupture tests in air at intermediate temperatures.11 In some Sylramic/BN/SiC materials that did not exhibit extrinsic pest, significant carbon enrichment was detected at the BN/cvi-SiC interface (see the bottom of Table 3). After burner rig exposure that carbon-rich layer was replacedby silica, as shown in Fig. 4. The silica may affordprotection to the BN andso prove beneficial to overall durability. However, its actual effect is yet to be determined. 4. Summary and conclusion Pest resistance of SiC/BN/SiC composites has been studied by burner-rig exposure at intermediate temperatures, followedby tensile tests andmicrostructural characterization to determine residual strengths and pest-relatedfeatures. All samples reinforcedwith Hi-NicalonTM andHi-Nicalon(S)TM fibers exhibitedthe severe, extrinsic mode of pesting. Those with SylramicTM fiber reinforcements resistedextrinsic pesting except in cases where the fibers hadbeen sizedwith oxide-based sizing. The common characteristic in all cases where extrinsic pesting occurredis that a film of carbon was observedbetween the BN interphase andfiber, originating from excess carbon in the Nicalon-type fibers and Fig. 3. SEM image showing carbon beneath the BN layer in Hi-Nicalon(S)TM/BN/SiC. Fig. 4. Following burner rig exposure, a layer of silica covers the BN interphase in a SylramicTM/BN/SiC material that was carbon rich at the BN/cvi-SiC interface. 616 L.U.J.T. Ogbuji / Journal of the European Ceramic Society 23 (2003) 613–617
L.U.J.T. Ogbuji/ Journal of the European Ceramic Society 23(2003)613-617 from sizing residue in the case of Sylramic fibers. The 5. Ogbuji, L U.J. T, A pervasive mode of oxidative degradatic carbon caused degradation by providing an easy route a Sic-SiC composite. J. Am. Ceram. Soc., 1998. 81(11). 2777- for deep penetration of the composite by ambient oxi- dants. The samples that were free of elemental carbon 6. Morscher. G. N, Tensile stress-rupture of SiC/Sic mini- proved to be resistant to pest degradation composite with carbon and boron nitride interphases at elevated mperatures in air. J. Am. Cram. Soc., 1997, 80(8), 2029-2042. Therefore, the severest form of pest degradation in 7. Jacobson, N.S., Morscher, G.N. Bryant, D. R. and Tressler. Sic/BN SIC composites can be suppressed by ensuring R. E. High-temperature oxidation of boron nitride: II BN layers that they are free of elemental carbon and its precursors in composites. J. Am. Ceram. Soc.. 1999, 82(6). 1473-1482. 8. Morscher, G. N, Bryant, D. R. and Tressler, R. E, Environ- in the interphase region mental durability of BN-based interphases(for SiC/SiC com tes) mperatures Ceramic Engineering Science Proceedings, 1997 Acknowledgements 8(3),525-533. 9. Evans, A. G. Zok, F. W, McMeeking. R. M. and Du. Z.Z. This work was done at NASa grc under contract Models of high-temperature, environmentally assisted embrittle. ment in ceramic-matrix composites. J. Am. Ceram. Soc., 1996 NAS3-98008. The author acknowledges the assistand 79(9),2345-2352 of M. Cuy(APBR), T. McCue(S-4700 SEM), D 10. Heredia, F. E, McNulty, J. C, Zok, F. w. and Evans, A. G Wheeler(AES); with sample exposure and examination; Oxidation embrittlement probe for ceramic-matrix composites. helpful discussions with G. Morscher on pesting modes J.Am. Ceran.Soc,1995,78(8),2097-2100. are highly appreciated 11. Morscher. G.N. Hurst. J. and Brewer. D. Intermediate-tem. perature stress rupture of a woven Hi-Nicalon, BN-interphase Sic-matrix composite in air. J. Am. Ceram. Soc., 2000. 83(6) References 12. Ogbuji, L. U. J. T, Identification of carbon sublayer in a Hi- Nicalon/BN/SIC composite. J. Mat. Sci. Lett., 1999, 18, 18 1. DiCarlo, J, Yun, H.M., Morscher, G.N., and Ogbuji, L. U.J.T. 13. Ogbuji, L.U.J. T, Silicon-based ceramic matrix composites for ings of Materials Week 2001, Munich Giemang. D. and Jones, In Proc. PAR- SONS 2000: Advanced Materials for 2lst Century Turbines and R. H. Oxidation of the carbon interface in Nica Power Plants, ed. A. Strang et al. The University Press, Cam- orced SiC composite. J. Am. Ceram. idge. 20 3. Heredia, F. E, McNulty, J C, Zok, F.w. and Evans, A G 14. Brewer, D, HSR/EPM combustor materials development pro- Oxidation embrittlement probe for ceramic-matrix composites gram.MatSci.and Eng,1999,A261,284-291 J.Am. Ceran.Soc.,l995,78(8),2097-210 15. Rebillat F. Gallet. SL. Bourrat.X. and Naslain R. Oxidation 4. Sheldon. B. W. Sun. E. Y. Nutt. S. R. and Brennan. J J. Oxi- sistance of BN coatings with different textures (poster). Mate- dation of BN-coated SiC fibers in ceramic-matrix composites rials Week 2001. Munich J.Am. Ceran.Soc,1996,79(2),539-543 16. Yun, H, and DiCarlo, J, unpublished work
from sizing residue in the case of Sylramic fibers. The carbon caused degradation by providing an easy route for deep penetration of the composite by ambient oxidants. The samples that were free of elemental carbon proved to be resistant to pest degradation. Therefore, the severest form of pest degradation in SiC/BN/SiC composites can be suppressedby ensuring that they are free of elemental carbon andits precursors in the interphase region. Acknowledgements This work was done at NASA GRC under contract # NAS3-98008. The author acknowledges the assistance of M. Cuy (APBR), T. McCue (S-4700 SEM), D. Wheeler (AES); with sample exposure andexamination; helpful discussions with G. Morscher on pesting modes are highly appreciated. References 1. DiCarlo, J., Yun, H.M., Morscher, G.N., andOgbuji, L.U.J.T., Progress in SiC/SiC composites for engine applications. Proceedings of Materials Week 2001, Munich, Germany. 2. Windisch, C. F., Henager, C. H., Springer, G. D. and Jones, R. H., Oxidation of the carbon interface in Nicalon-fiber-reinforcedSiC composite. J. Am. Ceram. Soc., 1997, 80(3), 569–574. 3. Heredia, F. E., McNulty, J. C., Zok, F. W. and Evans, A. G., Oxidation embrittlement probe for ceramic-matrix composites. J. Am. Ceram. Soc., 1995, 78(8), 2097–2100. 4. Sheldon, B. W., Sun, E. Y., Nutt, S. R. and Brennan, J. J., Oxidation of BN-coated SiC fibers in ceramic-matrix composites. J. Am. Ceram. Soc., 1996, 79(2), 539–543. 5. Ogbuji, L. U. J. T., A pervasive mode of oxidative degradation in a SiC-SiC composite. J. Am. Ceram. Soc., 1998, 81(11), 2777– 2784. 6. Morscher, G. N., Tensile stress-rupture of SiC/SiC minicomposite with carbon andboron nitride interphases at elevated temperatures in air. J. Am. Cram. Soc., 1997, 80(8), 2029–2042. 7. Jacobson, N. S., Morscher, G. N., Bryant, D. R. andTressler, R. E., High-temperature oxidation of boron nitride: II BN layers in composites. J. Am. Ceram. Soc., 1999, 82(6), 1473–1482. 8. Morscher, G. N., Bryant, D. R. andTressler, R. E., Environmental durability of BN-based interphases (for SiC/SiC composites) in H2O-containing atmospheres at intermediate temperatures. Ceramic Engineering & Science Proceedings, 1997, 18(3), 525–533. 9. Evans, A. G., Zok, F. W., McMeeking, R. M. andDu, Z. Z., Models of high-temperature, environmentally assisted embrittlement in ceramic-matrix composites. J. Am. Ceram. Soc., 1996, 79(9), 2345–2352. 10. Heredia, F. E., McNulty, J. C., Zok, F. W. and Evans, A. G., Oxidation embrittlement probe for ceramic-matrix composites. J. Am. Ceram. Soc., 1995, 78(8), 2097–2100. 11. Morscher, G. N., Hurst, J. andBrewer, D., Intermediate-temperature stress rupture of a woven Hi-Nicalon, BN-interphase, SiC-matrix composite in air. J. Am. Ceram. Soc., 2000, 83(6), 1441–1449. 12. Ogbuji, L. U. J. T., Identification of carbon sublayer in a HiNicalon/BN/SiC composite. J. Mat. Sci. Lett., 1999, 18, 1825– 1827. 13. Ogbuji, L. U. J. T., Silicon-basedceramic matrix composites for advanced turbine engines: degradation issues. In Proc. PARSONS 2000: Advanced Materials for 21st Century Turbines and Power Plants, ed. A. Strang et al. The University Press, Cambridge, 2000, pp. 767–778. 14. Brewer, D., HSR/EPM combustor materials development program. Mat. Sci. and Eng., 1999, A261, 284–291. 15. Rebillat, F., Gallet, S.L., Bourrat, X., andNaslain, R., Oxidation resistance of BN coatings with different textures (poster). Materials Week 2001, Munich. 16. Yun, H., andDiCarlo, J., unpublishedwork. L.U.J.T. Ogbuji / Journal of the European Ceramic Society 23 (2003) 613–617 617