urna 人 Aa CrIm se,87|811536-1542(2004) Effect of Environment on the Stress-Rupture Behavior of a Carbon-Fiber-Reinforced Silicon Carbide Ceramic Matrix Composite Michael J. Verrilli, Elizabeth J. Opila, ' Anthony Calomino, and J. Douglas Kiser NASA Glenn Research Center, Cleveland Ohio 44135 Stress-rupture tests were conducted in air, under vacuum, and Opila and Hann investigated the oxidation of Sic in 50% in steam-containing environments to identify the failure modes H,O/50% O, between 1200 and 1400C at I atm. The water and degradation mechanisms of a carbon-fiber-reinforced vapor oxidized the SiC and simultaneously volatilized the silica silicon carbide (C/SiC) composite at two temperatures, 600 (Sio, )scale, leading to paralinear weight change kinetics Rece: and 1200 C. Stress-rupture lives in air and steam-containing sion of SiC materials was also observed in high-pressure combus environments(50-80% steam with argon) are similar for a tion environments. Linear weight loss and surface recession rates stress of 69 MPa at 1200 C Lives of specimens tested in a 20% of Sic were observed in both fuel-lean and fuel-rich combustion steam/argon environment were about twice as long. For tests onducted at 600 C, composite life in 20% steam/argon was 30 gas mixtures during exposures in the temperature range of 1200%to 1450C, under pressures of 4 to 10 atm. This response was shown times longer than life in air. Thermogravimetric analysis of the to result from SiO, scale volatility. Note that the products of carbon fibers was conducted under conditions similar to the stress-rupture tests. The oxidation rate of the fibers in the aircraft turbine engine combustion contain about 10% water vapor independent of fuel/air ratio.' arious environments correlated with the composite stress- For structural components, characterization of the mechanical rupture lives. Examination of the failed specimens indicated behavior of C/SiC in relevant engine environments is required for component design. The behavior of C/SiC under stressed oxidation mode for specimens tested in air and steam environments at both temperatures (i.e, stress-rupture) conditions in air was investigated. -Test specimens exposed to stressed oxidation conditions over a tem- perature range of 350 to 1500C had reduced residual strength I. Introduction As test temperatures and stresses increased, specimen lives de- creased. Oxidation of the pyrocarbon interface and carbon fibers A DVANCED reusable launch vehicles(RLVs) will likely incor was observed in all samples that failed. Similar fiber degradation porate fiber-reinforced ceramic matrix composites(CMCs)in mechanisms were observed in C/SiC specimens tested under critical propulsion components. Use of CMCs is highly desirable stress-rupture and fatigue conditions in air at 5.%.C"For to save weight, improve reuse capability, and increase perfor the conditions used, specimen life was govemed by a combination mance. One of the candidate CMC materials is carbon-fiber of time at temperature and time-averaged stress and was not cycle reinforced silicon carbide(C/SiC) It is important to note that application of stress changes the blisks, turbopump rotors, and nozzle exit ramps for advanced oxidation rate of the carbon fibers in a C/SiC composite, compared rocket engines. In these applications, C/SiC components will be subjected to a service cycle that includes mechanical \oading under mechanical loads due to crack opening in the stressed condition complex environments. As an example. a typical reusable rocket TGA results obtained in laboratory air revealed higher C fiber turbopump rotor environment would include hydrogen, oxidation rates at intermediate temperatures(750C) than at high n, and steam, at high pressure(200 atm). temperatures(1000 -1400C), while the same C/SiC material had ronmental degradation of both the C fibers and the Sic longer rupture lives in air at 800C than at 1200 C. K is possible in an environment containing oxygen and steam A fundamental evaluation of the role of environment on damage of the oxidation beha avior mechanisms and life of material subjected to mechanical loading is and oxygen revealed three distinct composite degradation mecha- needed to assess the applicability of C/SiC for RLV components fibers. 2 At low temperatures (about 400-500 C), carbon fiber specimens in air or oxygen. or the oxidation kinetics of coupons oxidation kinetics are controlled by surface chemical reaction without external loads via furnace exposures in several environ- Above about 750C, the oxidation kinetics depend on transport of ments (air, oxygen, and water vapor"). In the present study the reactants and/or products in the boundary layer to or from the stress-rupture testing was conducted on C/SiC specimens in air under vacuum, and in steam-containing environments Intermedi- the C/SiC composite, changes in oxidation rate and the presence of ate and high temperatures (600 and 1200 C)were used for the localized or global attack as a function of temperature and time vere related to(i) decreasing widths of microcracks with increas- composite, the oxidation kinetics of the T-300 carbon fibers were ing temperature and ( ii) the increasing reactivity of SiC and carbon monitored by thermogravimetric analysis (TGA) in the same with oxygen. environments and at the same temperatures used for the stress rupture tests of the C/sic D. P. Butl--contributing editor IL. Material and Test Specimen The material examined in this investigation was a woven- carbon-fiber-reinforced SiC matrix composite manufactured by 2013: approved March 16, 200. Honeywell Advanced Composites, Inc. (now General Electric Power Systems Composites) using the chemical vapor infiltration
August 2004 Effect of Environment on the Stress-Rupture Behavior of a C/SiC Ceramic Matrix Composite 1537 As Received Twenty plates of C/SiC material nominally 150 mm X 230 mm x 3.0 mm thick were procured as a single lot for character- ization. All panels were processed in the same furnace runs for deposition of the pyrolytic carbon interface, CVI SiC matrix, and CVI seal coating Specimens machined from five of these C/SiC plates were tested in this study, Results obtained by testing and characterizing other specimens from this lot of C/SiC have been reported elsewhere. The as-fabricated specimens were in- spected using radiography. No defective regions were detected 1270 using this NDE technique 368 R TYi The test specimen geometry had a reduced gauge section design. It was 152 mm long, with a grip section width of 12.7 mm. Fig 1. C/SiC specimen used in this study Dimensions are in millimeters. a reduced gauge section width of 10.2 mm, and a thickness of 3.0 mm(Fig. 1). Test specimens were machined from composite plates using diamond grinding and then were seal coated with CVI SiC. The CVI SiC thickness was 17 um on the composite surface and (0/90)two-dimensional plain weave fabric of T-300 carbon fibers about 5 um on the machined edge in I-k filament tows. Fiber volume fraction, as reported by the The microstructure of the typical as-manufactured composite is omposite manufacturer, was 45%, The fiber coating was pyrolytic shown in Fig. 2. The composite contains microcracks carbon, having a mean thickness of 0. 6 um. The composite density matrix-rich regions and the carbon fiber plies and individual tows was 2.06 g/cm and the composite contained open porosity of This type of microcracking in C/SiC has been well document- about 12 In addition, the seal coating contains a regularly space Porosity CVI SIC seal ati Transverse fiber tows Longitudinal fiber tow mm Fig.2. Polished sections of the as-manufactured C/SiC microstructure: (a) cross section of the gauge section of a stress-rupture specimen, (b) details of the
Journal of the American Ceramic Sociery-Verrilli et al Vol. 87. No. 8 Load Specimen Water-cooled induction coils Ceramic steam injector Copper tubing Thermocouple with heater tape (Type R) SiC susceptor Quartz chamber steam from End pump Cap Carrier gas Load Fig 3. Test system configuration used for stress-rupture testing in steam/argon environments (10 ksi) was used for all tests in this study in an attempt to avoid thermal expansion(CTE)between the SiC matrix and the C fibers When open, these cracks allow ingress of the environment. Several environments were used during the testing, including aboratory air, vacuum (5x 10 torr), and steam/Ar mixtures lament tows. These fibers were from the same lot used in the fil TGA was conducted on T-300 carbon fibers in the form of l-k The steam/Ar testing was conducted at ambient pressure C/SiC panels. Bundles of 0.45 g of fiber were used for the three mixtures(each expressed in terms of molar volumes):80% oxidation experiments. steam/20% Ar, 50% steam/50% Ar, and 20% steam/80% Ar. All environments were used for testing at 1200oC. Only air and the 20%o steam/80% Ar mixture were used for the 600oC stress-rupture Il. Test Procedures tests The test configuration for the steam/Ar stress-rupture testing is stress-rupture tests were performed at 600 and shown in Fig 3. The system includes an MTS 250 kN servo- using electromechanical and servo-hydraulic test ma- ydraulic test machine with water-cooled wedge grips and a steam The tests were performed per the recommended proce- chamber. The steam chamber consists of an inductively heated Sic dures in ASTM standard C 1337. In a previous study, results of tube inside a quartz tube. The SiC susceptor heats the composite stress rupture testing of C/SiC in air at 800 C indicated that a specimen. The quartz tube has caps on either end. Steam is stress of 35 MPa did not open the SiC matrix cracks and thus introduced through the quartz end caps via alumina tubes with allow oxygen ingress and carbon oxidation. Since the high slots. The specimen gauge section is located inside the steam temperature behavior of C/SiC in oxidizing environments chamber, and the ends of the specimen pass through slots in both typically depends on fiber oxidation, a higher stress of 69 MP the quartz tube and the SiC susceptor. The grip ends of the Cahn 1000 balance Counterflow gas Sample Furnace- H2o Thermocoupl Fig. 4. Schematic of the thermogravimetric analysis (TGA) system and the water saturator
t2004 Effect of Environment on the Stress-Rupture Behavior of a C/SiC Ceramic Matrix Composite Table L. Summary of Rupture Lives of C/SiC Obtained a1200° Using a Stress of 69 MPa Test environment Average life (h pecimens tested 1200 Air 2.49±0.18 7 1200 0% 1200 80% 20%Ar20l±008 8.42±3.70 60020% stean/80%Ar250.32±5266 ste 50%Ar20%Ar Test Environment sections were held under vacuum to enable pores and cavities to outgas, the metallurgical sampl covered with epoxy. The Fig. 5. Stress-rupture lives for C/SiC at 600 and 1200 C obtained using specimens were then placed in re chamber and held under a stress of 69 MPa 10 MPa nitrogen gas pressure te poxy into sample pores and damage locations. Samples sectioned, then lapped and polished for examination specimen are held by the wedge grips, outside of the steam hamber. A slightly positive pressure of steam/Ar gas introduced into the steam chamber prevents flow of air to the specimen gauge IV. Results section. Gage section temperature is monitored using type R (1) Stress-Rupture Lives of C/Sic thermocouples The stress-rupture lives obtained at 600 and 1200"C are shown The stress-rupture testing in air was performed using MTS 100 in Fig. 5 and given in Table L At least two tests were conducted for kN electromechanical test rigs fitted with environmental chambers each condition. For any test environment at both test temperatures Water-cooled wedge grips were used for specimen gripping. The duplicate tests resulted in specimen lives that differed by a factor hot zone contained MoSi, heating elements. Vacuum tests were of 2X or less. The average specimen life in air at 600oC was 8.5 h performed using a similar system with a graphite hot zone. Load One of the tests conducted in 20% steam/80% Ar at 600C was train alignment of all test machines was verified before and after stopped before specimen failure after 213 h. The other specimen the testing. The maximum bending strain was less than 5% for a tested in steam failed after 288 h. Thus, the average test duration tensile load of 4.5 kN, as verified using a strain-gauged alignment in steam was at least 30 times longer than the lives obtained in air Seven duplicate tests were conducted in air at 1200 C, resulting The oxidation kinetics of T-300 carbon fibers(l-k tow) were in an average life of 2.45+ 0. 18 h Rupture life in steam at 1200C monitored by TGA, Fiber bundles were oxidized to completion appears to increase with decreasing percent steam. An average life and weight change was continuously recorded with a Cahn 1000 of 2 h was obtained in 80% steam/20% Ar In 50% steam/50% Ar thermogravimetric analyzer and data acquisition system. The at 1200"C, the average life was 2.6 h. However, in 20% steam/80% fibers were placed in a slotted high-purity alumina crucible and Ar, life was about twice as long (4.5 h). The two tests conducted suspended from the balance. A schematic of the TGA setup is under vacuum at 1200"C were stopped after 100 h, before failure shown in Fig. 4. Temperatures of 600 and 1200 C were used at a Lives obtained at 1200C are shorter than the 600C lives under total pressure of I atm. Test environments were air, oxygen, Al the same environment. The average life in 20% steam/80% Ar at and 0.2 atm steam in Ar(20% steam/80% Ar). The gas flow rate 1200C was 50 times shorter than the average test duration at was I Umin, with a corresponding gas velocity of 4.4 cm/s. 600C in the same environment Untested composite material was prepared in parallel with Two tests were conducted in the steam chamber at 1200.C tested samples for microstructural examination. After the specimen using an atmosphere of gettered Ar. The specimens failed after 8. 1200°c.A 1200°c,ar °C,ar ti hr Fig. 6. T-300 carbon fiber oxidation kinetics at 600"and 1200.C
Journal of the American Ceramic Society-Verrilli et al Vol 87, No. 8 Fig. 7. Polished cross sections of C/SiC specimens tested under stress-rupture conditions at 600C: (a)in a 20% steam/80% Ar environment, (b)in air. and 9.2 h. Fracture of both specimens occurred about 10 mm but carbon fiber oxidation within the longitudinal and transverse outside of the gauge section, at a location outside of the chamber fiber tows can be seen under higher magnification. Within the 0 where the specimens are exposed to air ows, bands of oxidation of the fibers can be seen, but complete oxidation of fiber tows was not observed In comparison, oxidation (2) Thermogravimetric Analysis Data for T-300 Carbon damage observed in C/SiC tested in air under stress-rupture Fibers conditions at 600C was more extensive( Fig. 7(b). Entire fiber Figure 6 shows TGA data for the oxidation of T-300 carbon tows near the surface were removed by oxidation and fiber ments were first conducted in Ar as a control for comparison with erally localized around preexisting cracksposite thickness. gen- fibers at 600 and 1200oC in the various environments. Experi oxidation was observed throughout the cor other environments. Residual amounts of oxygen in the system at Composite damage in a specimen tested at 1200C in air can be the start of these experiments are responsible for the initial weight seen in Fig. 8. This specimen failed after 2.4 h. Large regions of the cross section adjacent to the surface are damaged. Surface Weight loss is fastest in air followed by water vapor, with all of the longitudinal fiber tows and transverse tows are missing due to fibers being consumed in 1.5 h Negligible weight loss is observed in Ar. At 600oC, weight loss is still rapid in air, but the weight loss oxidation. Similar to that observed for the specimen tested at in water vapor is indistinguishable from the rate observed in Ar 600 C in 20% steam/80% Ar, the oxidation initiates along preex ing cracks within the fiber tows (3) Examination of Tested C/SiC Specimens The same patterns of fiber oxidation occurred in specimens tested under steam/Ar environments at 1200 C(see Fig. 9).Fiber A cross section from the gauge region of a C/SiC tested in 20% steam/80% Ar at 600oC for 213 h and stopp oxidation was often observed to occur at locations of preexisting failure is shown in Fig. 7(a). In this image and in cracks of the SiC seal coating in all specimens tested in air or subsequent figures showing sections taken from other specimens. steam at 1200oC. Similar to the specimens tested in air at 1200oC the loads were applied perpendicular to the plane of the figure Little composite damage can be seen in the overall cross section Fig. 8. Polished cross section of C/SiC specimen tested under stress- Fig. 9. Polished cross section of C/SiC specimen tested under stress rupture conditions at 1200C in air rupture conditions at 1200"C in a 20% steam/80% Ar environment