EVALUATION OF MECHANICAL RELIABILITY OF Si 3 N 4 NOZZLES AFTER EXPOSURE IN AN INDUSTRIAL GAS TURBINE

Transcription

1 EVALUATION OF MECHANICAL RELIABILITY OF Si 3 N 4 NOZZLES AFTER EXPOSURE IN AN INDUSTRIAL GAS TURBINE H. T. Lin*,a, M. K. Ferber a, M. van Roode b ( a ) Metals and Ceramics Division, Oak Ridge National Laboratory Oak Ridge, TN ( b ) Solar Turbines Incorporated San Diego, CA ABSTRACT This paper provides a review of a recent study undertaken to evaluate the mechanical reliability of SN88 Si 3 N 4 ceramic nozzles exposed in an industrial gas turbine. Two field tests with exposures time of 10 and 68 h were carried out. The first 10 h field test revealed minor changes in both microstructure and strength of airfoil region. However, nozzles with crack generation, which initiated in low temperature airfoil region (< 1000 C), were observed after 68 h test, resulting in the termination of second engine test. Analyses of the cracked nozzles by scanning electron microscopy and x-ray diffraction revealed changes in microstructure and secondary phase due to environmental exposure. Dynamic fatigue tests on as received SN88 Si 3 N 4 at intermediate temperatures in air suggested that changes in secondary phase could result in the formation of an extensive damage zone, due to the generation of a large residual tensile stress, which substantially reduced the mechanical reliability and probably led to the failure of turbine nozzles. INTRODUCTION Silicon nitride ceramics with reinforcing elongated grain microstructure are leading candidates for use as high-temperature, structural components in advanced gas turbines due to their superior thermomechanical properties [1-5]. Recent ceramic gas turbine programs at both Solar Turbines and Rolls Royce Allison [6-8] have executed many field tests to increase the experience base concerning the behavior of ceramic components in industrial gas turbine environments. A key lesson learned in both programs is that environmental effects may severely limit the long-term reliability of Si 3 N 4 materials. In particular, the high velocities and presence of water vapor in the environment can lead to the volatilization of the normally protective silica layer. Researchers have shown that the presence of water vapor leads to the formation of a gaseous Si(OH) 4 species via a reaction with the silica layer [9-11]. The rate of Si(OH) 4 formation of this species and thus the rate at which the Si 3 N 4 is consumed by its continued oxidation are a function of gas velocity, pressure of the water vapor, and total pressure of the environment. In addition to the volatility issues, which may lead to loss of function due to excessive changes in component dimensions, the effect of the environment upon the long-term mechanical reliability must be understood as well. One approach addressing this issue is to evaluate the component properties directly using small test specimens. For example, the availability of the small dog-bone tensile specimen [12] has provided for the measurement of the tensile stress rupture properties of silicon nitride samples from first stage turbine blades. Results showed that there were significant differences in the creep response characteristics obtained for various parts of the ceramic blade due to variations in grain size and secondary phase content and chemistry. More importantly the data obtained from testing specimens prepared from the production billets greatly over-estimated the lifetime of components with complex shapes. This paper describes the results of a recent component verification study involving SN88 Si 3 N 4 nozzles exposed in an industrial gas turbine. Silicon nitride nozzles after exposure to 10 h and 68 h field tests were examined in this study. Scanning electron microscopy was used to elucidate the changes in the microstructure arising from the oxidation process. In addition, the stability of secondary phase(s) was evaluated using x-ray analysis. The exposed surface strength was measured as a function of field test time using a miniature biaxial specimen.

2 EXPERIMENTAL PROCEDURES The nozzles examined in this study were fabricated from NGK SN88 Si 3 N 4 material (NGK Insulators Ltd., Nagoya, Japan). They were gas-pressure-sintered using rare-earth sintering additives and then post-heat-treated by NGK to form a protective silica surface layer. The predominant secondary phase present after densification is Yb 4 Si 2 O 7 N 2, designated as the J-phase. The first stage ceramic nozzles (Figure 1) were designed for retrofit into a Centaur 50S turbine engine (Solar Turbines Incorporated, San Diego, CA) [13]. The first engine test was carried out with one hour full load and 10 h total run time. The second engine test was planed for 100 h endurance test. However, the test was terminated after 68 h total run time including 15 start/stop cycles due to crack generation in the nozzles (Fig. 1b). (a) Figure 1 Solar SN88 first stage Si 3 N 4 nozzles after engine test time of 10 h (a) and 68 h. X-ray diffraction was used to identify the predominant secondary phases in the airfoil as a function of engine test time. Scanning electron microscopy (SEM) was first used to examine the surfaces of the airfoils and platforms after removal of the nozzles from the engine. Nozzles were subsequently sectioned by making a longitudinal cut parallel and adjacent to the trailing edge. These pieces were subsequently polished and examined with SEM. The biaxial flexure strength [14-16] was measured for samples from selected vanes using the ball-on-ring arrangement. Specimens were machined from both the airfoil and platform surfaces by first diamond core drilling small cylinders having nominal diameters of 5.5 mm. Each cylinder was then machined on one face only until the thickness was 0.4 to 0.5 mm. In this way the tensile face of each specimen always consisted of the exposed surface of either the airfoil or platform. The details of testing fixture and procedures can be found in Ref. 17 RESULTS AND DISCUSSION Macroscopically, the SN88 Si 3 N 4 nozzles after 10 h engine test revealed little change on airfoil surface features except some reddish deposits (Fig.1a). On the other hand, the nozzles after 68 h engine test showed a whitecolored, powder-like scale with crack generation (Fig. 1b). X-ray analysis indicated that the white-colored scale present on airfoil surfaces of 68 h nozzles was mainly the Yb 2 Si 2 O 7 plus Yb 2 SiO 5, which formed due to the oxidation of initial secondary phase (J- Phase) present in the as-received nozzles. Note that the dominant crack always initiated in the low temperature (< 1000 C) airfoil region [18]. SEM examinations of airfoil surfaces of both 10 and 68 h engine tested nozzles showed the presence of corrosion pits associated with Fe-Si-O deposits (Fig. 2). The Fe element detected in the deposits was attributed the erosion of superalloy components in the engine. Also, the 68 h tested nozzles revealed more surface porosity due to oxidation and volatilization of Si 3 N 4 as compared with 10 h tested nozzles. In addition, SEM analysis confirmed that the white-colored, powder-like scale observed in 68 h nozzles was indeed a secondary phase, which consisted of Yb, Si, O and trace amount of Y. The accumulation of the Yb 2 Si 2 O 7 plus Yb 2 SiO 5 on the surface was most likely due to the selective recession of Si 3 N 4 grains, as seen in Fig. 2b. The Si 3 N 4 grains were removed by a combination of oxidation and volatilization processes, resulting in the accumulation of secondary phases on the airfoil surface. The SEM micrographs of the polished cross-sections are shown in Figure 3. The SN88 Si 3 N 4 nozzles showed minor changes in microstructure, e.g., pores in secondary phase and very limited intergranular cracks, in the region adjacent to the airfoil surface after 10 h engine exposure (Fig. 3a). However, extensive generation of intergranular cracks plus pores in the secondary phase was observed in nozzles after 68 h engine exposure (Fig. 3b). The changes in microstructure, i.e., formation of environment-induced damage zone, were attributed to the oxidation of Si 3 N 4 materials in the presence of high-pressure

3 (a) Fe-Si-O (a) Fe-Si-O 2 m 20 m 2 m 10 m Figure 2. SEM of airfoil surface features of SN88 silicon nitride nozzles after (a) 8 h and 68 h engine test. water vapor. A previous study had showed that the presence of high-temperature, highpressure water steam would lead to formation of an extensive subsurface damage zone in Si 3 N 4 materials [19]. Also, results of Vickers indentation in oxidation zone in the vicinity of crack initiation site showed a preferential crack growth along the oxidation zone/bulk material boundary, which was in sharp contrast to the symmetrical indentation feature observed in the bulk material region (Fig. 4). The feature of preferential crack growth in the oxidation zone suggested the presence of a large residual stress. SEM examinations also showed that the airfoil/platform transition region, where the dominant crack normally initiated and led to Figure 3. SEM micrographs of polished crosssection of nozzle airfoil surface region after (a) 10 h and 68 h engine test. failure of nozzles, exhibited a more extensive damage zone formation (~30-40 µm) than the middle airfoil surface region (~10 µm). Note that the temperature and air velocity in the transition region were lower than the middle airfoil surface region. The finite-elementanalysis of temperature distribution at steady state during engine operation indicated that the temperature in the airfoil/platform transition region ranges from 800 to 950 C [18]. The SEM observations just described are in contrast to the experimental results and oxidation models [9-11]. It was possible that in the middle airfoil region the high gas velocity aggravated the removal of silica due to volatilization of the silica layer, resulting in less extensive damage zone evidence. On the other hand, in the transition region little, if

4 (a) h Test As-received Stress (MPa) h Test 1 mm 20 m Z (mm) Figure 5. Stress versus displacement curves of biaxial samples extracted from airfoils of SN88 nozzles after 10 and 68 h engine test. Figure 4. SEM micrographs show indentations in (a) oxidation zone and inner bulk region. The inserted photo shows the presence of an oxidation zone (light region). any, protective silica would form after the initial silica layer present in the as-received nozzle receded and could lead to an extensively accumulated damage zone (Figure 3b). Strength test results at room temperature of samples extracted from nozzle airfoils after engine test are summarized in Figure 5. Note that the disk samples were tested with exposed, pressured sides as tensile surfaces. Results showed that the 10 h tested samples exhibited ~10% decrease in strength, while the 68 h tested samples exhibited ~30% strength degradation. The extensive strength degradation observed in nozzles after 68 h engine exposure is due to the extensive formation of subsurface damage zone (Fig. 3b). The crack initiation in low temperature airfoil/platform transition region has led to a hypothesis that SN88 Si 3 N 4 might exhibit a mechanical instability at intermediate temperatures in air. Note that previous creep studies indicated that SN88 Si 3 N 4 exhibited excellent creep performance at temperature 1038 C in air [20]. Recently, studies of dynamic fatigue by Wereszczak et al. have shown that some commercial Si 3 N 4 materials exhibited substantial strength degradation at temperature 850 C in air [21]. Therefore, a dynamic fatigue test in four-point bending was conducted to evaluate the mechanical reliability of SN88 Si 3 N 4 at 850 C in air [22]. Figure 6 summarizes the earlier dynamic fatigue response of SN88 Si 3 N 4 tested at 850 C in air at stressing rates of 30 MPa/s and MPa/s. Note the NT154 Si 3 N 4, which contained Y 2 Si 2 O 7 as secondary phase, is used as a benchmark in this study. Results showed that the fracture strength of NT154 Si 3 N 4 was similar those obtained at room-temperature and was independent of stressing rates accompanied by a high fatigue exponent of 87. On the other hand, strength of SN88 Si 3 N 4 was very sensitive to stressing rates. For instance, samples tested at 30 MPa/s exhibited similar strength as those obtained at room temperature, while samples tested at MPa/s exhibited ~ 43% strength decrease. In addition, SN88 Si 3 N 4 exhibited a low fatigue exponent of 16, indicative of a high susceptibility to slow crack growth (SCG) process at high temperatures in air. The fracture surface of SN88 samples tested at 850 C/0.003 MPa/s in air revealed a light-colored ring surrounding the bend bar (as shown in Fig. 7), indicating the presence of an environment-affected zone (EAZ). SEM examinations showed that a ~ 30 µm damage zone, containing multiple cracks plus pores in secondary phase, developed in the lightcolored EAZ. Note that the EAZ developed in all four-side surfaces, suggesting the development of EAZ was not a stresspromoted phenomenon. Results of x-ray analysis for MPa/s tested samples also indicated a change in secondary phase from J- phase to Yb 2 Si 2 O 7 plus Yb 2 SiO 5.

5 The phase transformation phase from J- phase to Yb 2 Si 2 O 7 could introduce an extreme high residual tensile stress in the EAZ due to a ~64% decrease in material volume [23], consistent with the indentation results in the oxidation zone of cracked nozzles (Fig. 4). Note that there is a minor material volume increase (~5%) for J-phase changing to Yb 2 SiO 5. The development of a high tensile stress would then lead to fracture of Si 3 N 4 grains and, also, generation of multiple intergranular cracks. Similar mechanical instability due to phase changes was previously reported for Si 3 N 4 materials sintered with Y 2 O 3 additive [23]. Fracture Strength (MPa) The results of dynamic fatigue tests at 850 C in air suggested that an EAZ developed in the airfoil/platform transition region, where the temperature was low and no protective silica layer would form during the exposure, after the initial silica layer present in the asreceived nozzles recessed. The generation of extensive multiple cracking in EAZ would significantly reduce the mechanical reliability and, thus, greatly increase the susceptibility of nozzles to SCG processes. Consequently, a critical crack would readily initiate at the transition region and lead to the failure of nozzles. On the other hand, the reason why the EAZ was not observed in the higher temperature regions; i.e., middle airfoil region, appeared to be due the elimination of the EAZ by oxidation/volatilization processes and thus continuous material recession. ACKNOWLEDGEMENTS The authors thank Drs. and for reviewing the manuscript. Research is sponsored by the U.S. Department of Energy, Office of Power Technologies, Microturbine Technologies Program, under Contract DE-AC05-00OR22725 with UT-Battelle, LLC C in air SN88 NT Stressing Rate (MPa/s) Figure 6. Fracture strength versus stressing rate results of SN88 silicon nitride tested at 850 C in air. SUMMARY Silicon nitride nozzles were designed and field-tested in an industrial gas turbine. Dominant crack developed in the airfoil/platform transition region after 68 h test, resulting in the termination of engine test. Both SEM and x-ray analyses indicated changes in microstructure and secondary phase due to the exposure in engine environment. An extensive damage zone, containing multiple cracks and pores in secondary phase, developed in the subsurface region after 68 h engine test. A supporting dynamic fatigue test at 850 C in air suggested that changes in secondary phase could result in the formation of an extensive damage zone, due to the generation of a large residual tensile stress, which substantially reduced the mechanical reliability and led to the failure of turbine nozzles. Figure 7. Fracture surface of SN88 bend bar tested at 850 C/0.003 MPa in air exhibited an environment-affected zone as indicated by arrows. REFERENCES 1 mm (1) D. Anson, K. S. Ramesh, and M. DeCorso, "Application of Ceramics To Industrial Gas Turbines"; DOE/CE/ &2; Battelle Columbus, March (2) H. E. Helms, R. A. Johnson, and L. E. Groseclose, "AGT 100-Advanced Gas Turbine Technology Development Project," Proc. 23rd Automotive

6 Technology Development Contractors' Coordination Meeting P-165, Warrendale, PA, March (1986) (3) M. van Roode, Ceramic Retrofit Program, Proc. Joint Contractors Meeting: FE/EE Advanced Turbine Systems Conference FE Fuel Cells and Coal-Fired Heat Engines Conference, DOE/METC-93/6132 (1993) (4) D. Carruthers and L. Lindberg, "Critical Issues for Ceramics for Gas Turbines," Proc. 3rd Int. Symp. on Ceramic Components and Materials for Engines, Westerville, Ohio (1988) (5) P. J. Haley, R. L. Holtman, L. E. Groseclose, S. J. Hilpisch, and A. H. Bell III, "Advanced Turbine Technology Applications Project (ATTAP)", Proc. 29th Automotive Technology Development Contractors' Coordination Meeting, Warrendale, Penn. (1992) (6) M. van Roode, W. D. Brentnall, K. O. Smith, B. D. Edwards, L. J. Faulder, and P. F. Norton, Ceramic Stationary Gas Turbine Development - Third Annual Summary, ASME Paper 96-GT-460, to be published in the Proc. of ASME Turbo Expo 96 Conference, Birmingham, UK, June 10-13, (7 ) V. Parthasarathy, M. van Roode, J. Price, S. Gates, S. Waslo, and P. Hoffman, "Review of Solar's Ceramic Stationary Gas Turbine Development Program, Pro. 6th Int. Symp. Ceramic Materials and Components for Engines, Arita, Japan, (1997) (8) R. Wenglarz, S. Ali, and A. Layne, Ceramics for ATS Industrial Turbines, to be published in the Proc. of ASME Turbo Expo 96 Conference, Birmingham, UK, June 10-13, (9) E. J. Opila and R. E. Hann, Jr., Paralinear Oxidation of SiC in Water Vapor, J. Am. Ceram. Soc., Vol. 80 [1] (1997) (10) E. J. Opila, Variation of the Oxidation Rate of Silicon Carbide with Vapor Pressure, J. Am. Ceram. Soc., Vol. 82, No. 3, pp , (1999). (11) J. L. Smialek, R. C. Robinson, E. J. Opila, D. S. Fox, and N. Jacobson, SiC and Si 3 N 4 Recession due to SiO 2 Scale Volatility Under Combustor Conditions, Advanced Composite Materials, Vol. 8, No. 1 (1999) (12) H. T. Lin, P. F. Becher, M. K. Ferber, and V. Partasarathy, Verification of Creep Performance of a Ceramic Turbine Blade, Proc. 2nd Int. Symp. Science of Engineering Ceramics, Osaka, Japan (1998). (13) A. F. Kirstein and R. M. Woolley, "Symmetrical Bending of Thin Circular Elastic Plates on Equally Spaced Point Supports," Journal of Research of the National Bureau of Standards - C. Engineering and Instrumentation, Vol 71C, No 1, Jan-Mar (14) R. Thiruvengadaswamy and R. O. Scattergood, "Biaxial Flexure Testing of Brittle Materials," Scripta Metall. et Mater., Vol 25 (1991) (16) D. K. Shetty, A. R. Rosenfield, P. McGuire, B. K. Bansal, and W. H. Duckworth, "Biaxial Flexure Tests for Ceramics," Ceram. Bull., Vol 59 [12] (1980) (17) M. K. Ferber, H. T. Lin, V. Parthasarathy, and R. A. Wenglarz, Degradation of Silicon Nitrides in High Pressure, Moisture Rich Environments, to be published in the Transactions of ASME, 2000 (IGTI, ASME TURBO EXPO 2000, May 8-11, 2000, Munich, Germany). (18) O. Jimenez, Solar Turbines, private communication. (19) M. K. Ferber, H. T. Lin, and J. Keiser, Oxidation Behavior of Non-Oxide Ceramics in a High-Pressure, High- Temperature Steam Environment, Mechanical, Thermal and Environmental Testing and Performance of Ceramic Composites and Components, ASTM STP 1392, American Society for Testing and Materials, West Conshohocken, PA, (20) A. A. Weresczak and T. P. Kirkland "Creep Performance of Candidate SiC and Si 3 N 4 materials for Land-based, Gas Turbine Engine Components, ASME Paper 96-GT-385, to be published in the Proc. of ASME Turbo Expo 96 Conference, Birmingham, UK, June 10-13, (21) A. A. Weresczak, T. P. Kirkland, H. T. Lin, and S. K. Lee, "Intermediate Temperature Inert Strength and Dynamic Fatigue of Candidate Silicon Nitrides for Diesel Exhaust Valves," to be published in Proc. Ceram. Eng. Sci. (2000). (22) H. T. Lin, M. K. Ferber, and P. F.Becher, "Mechanical Reliability of Yb-containing Silicon Nitride at Intermediate Temperatures," presented at Am. Ceram. Soc. Annual Meeting, St. Louis, April 30- May 3, (23) F. F. Lange, "High Temperature Deformation and Fracture Phenomena of Polyphase Si 3 N 4 Materials," Proc. 2nd NATO Advanced Study Institute of Nitrogen Ceramics (1981)