Residual Stresses and Their Effects on the Durability of Environmental Barrier Coatings for SiC Ceramics Kang N. Lee, w,z, * Jeffrey I. Eldridge,* and Raymond C. Robinson y NASA Glenn Research Center, Cleveland, Ohio 44135 Qualitative residual stresses in current environmental barrier coatings (EBCs) were inferred from the curvature of EBC-coat- ed SiC wafers, and the effects of EBC stresses on the durability of EBC-coated SiC were evaluated. The magnitude of substrate curvature correlated fairly well with the EBC–SiC coefficient of thermal expansion (CTE) mismatch, EBC modulus, and ther- mally induced physical changes in EBC. BSAS (1xBaO . xSrO . Al 2 O 3 . 2SiO 2 , 0rxr1) components in the current EBCs, i.e., Si/mullite or mullite1BSAS/BSAS or yttria-stabi- lized zirconia (YSZ: ZrO 2 –8 wt% Y 2 O 3 ), were the most bene- ficial for reducing the EBC stress in as-sprayed as well as in post-exposure EBCs. The reduced stress was attributed to the low modulus of BSAS. The addition of a YSZ top coat signifi- cantly increased the substrate curvature because of its high CTE and sintering in thermal exposures. There were clear cor- relations between the wafer curvature and the EBC durability. The Si/mullite120 wt% BSAS/BSAS EBC maintained excel- lent adherence, protecting the SiC substrate from oxidation, while the Si/mullite120 wt% BSAS/YSZ EBC suffered de- lamination, leading to severe oxidation of the SiC substrate, after a 100 h 13001C exposure in a high-pressure burner rig. I. Introduction O NE key factor that limits the performance of current gas turbine engines is the temperature capability of hot section structural components. Silicon-based ceramics, such as SiC/SiC composites and monolithic Si 3 N 4 , are leading candidates to re- place superalloy hot section components in next generation gas turbine engines because of their superior high-temperature me- chanical properties, such as strength and creep resistance. A major stumbling block to applying Si-based ceramics in the gas turbine hot section is the recession of these materials in com- bustion environments because of the volatilization of silica scale by water vapor. 1,2 An environmental barrier coating (EBC) is an external coating that protects Si-based ceramics from water vapor attack. 3 Therefore, the EBC development is an enabling technology for realizing SiC/SiC composite and Si 3 N 4 ceramic hot section components in next generation gas turbines. Mullite (3Al 2 O 3  2SiO 2 ) has attracted the most interest as an EBC candidate for SiC and Si 3 N 4 ceramics from the early days of EBC development because of its low coefficient of thermal expansion (CTE), chemical compatibility with Si-based ceram- ics, and good adherence. 4–7 One key issue with plasma-sprayed mullite coatings is phase instability. 7 The mullite processed with conventional plasma spraying contains a significant amount of metastable amorphous phase because of the rapid cooling of molten mullite during the solidification on a cold substrate. A subsequent exposure of the mullite to a temperature above B10001C causes the crystallization of the amorphous phase. Shrinkage accompanies the crystallization, leading to cracking and delamination of the mullite. A second generation, fully crystalline plasma-sprayed mullite coating was developed by spraying the coating while the substrate is heated above the crystallization temperature of the amorphous mullite. 7 Conse- quently, the second-generation mullite coating exhibits dramat- ically enhanced adherence and crack resistance. Another issue with mullite coatings is the relatively high silica activity (B0.4) and the resulting recession in high-velocity combustion environ- ments because of the selective volatilization of silica by water vapor. 8 The selective volatilization of silica from mullite leaves a porous alumina surface layer which readily spalls. Despite the silica volatility, mullite and mullite-based coatings constitute a key component in current EBCs, either as a bond coat or as an intermediate coat in conjunction with a water vapor resistant top coat. 9,10 Besides the water vapor stability, low EBC stress, EBC/sub- strate chemical compatibility, and adherence are key ingredients for durable EBCs. 3 Low thermal conductivity is also desired to maximize the thermal insulation capability. Considering the complex nature of the requirements for successful EBCs, it is unlikely that a single-layer coating will satisfy all the require- ments. Consequently, multilayer EBCs have evolved in which each layer satisfies a subset of the key requirements. The current state-of-the-art EBC is comprised of three layers. 9,10 The first layer, a silicon bond, provides the adherence, the second layer, mullite or mullite1BSAS (1xBaO  xSrO  Al 2 O 3  2SiO 2 ,0r xr1), provides the chemical compatibility, and the top layer, BSAS, provides the water vapor stability. Details of perform- ance and durability issues of the current state-of-the art EBC are described in Lee et al. 10 , and its performance in industrial en- gines is described in Eaton et al. 11 and More et al. 12 One area in the current EBC that has not been investigated is the effect of residual stress. Clear understanding of the residual stress in current EBCs and correlation of this stress to EBC performance can provide valuable insights for the development of advanced EBCs. This paper discusses residual stresses in cur- rent EBCs inferred from the curvature of EBC-coated SiC wa- fers and their effects on the EBC durability. II. Experimental Procedure (1) Estimation of Qualitative Residual Stresses In theory, stresses in EBCs can be determined by applying the Stoney equation 13 in conjunction with the curvature of EBC- coated SiC substrates: s c ¼ E SiC 1  n SiC t 2 SiC 6Rt c (1) where s c is the coating stress, E SiC the SiC modulus, n SiC the SiC Possion’s ratio, t SiC the SiC thickness, t c the coating thickness, and R the radius of curvature. However, the Stoney equation is a good approximation of coating stress only when the coating has a uniform thickness and is much thinner than the substrate, 3483 J ournal J. Am. Ceram. Soc., 88 [12] 3483–3488 (2005) DOI: 10.1111/j.1551-2916.2005.00640.x r 2005 The American Ceramic Society No claims to original US government works E. Lara-Curzio—contributing editor *Member, American Ceramic Society. w Author to whom correspondence should be addressed. e-mail: kang.n.lee@grc.nasa. gov z Senior Research Scientist, Chemical Engineering Department, Cleveland State Univer- sity, Cleveland, OH 44115. y Staff Engineer, QSS Inc., Brook Park, OH 44135. Manuscript No. 20529. Received July 9, 2004; approved May 9, 2005.