Andrew C. E. Reid & Andrew R. Roosen,. CTCMS, MSEL, NIST. ⢠Andrew J. Allen, Jan Ilavsky, Jay S. Wallace,. Ceramics Div., NIST .... Modulus (G. Pa). 0. 20. 40.
Using OOF to Model Mechanical Behavior of Thermal Barrier Coatings Mark R. Locatelli and Edwin R. Fuller, Jr. National Institute of Standards and Technology Gaithersburg , MD 20899-8521, U.S.A. Symposium on Advanced Ceramic Coatings Intl. Conf. on Advanced Ceramics & Composites Cocoa Beach, FL - January 15, 2002
Acknowledgments • Stephen A. Langer, Information Tech. Lab., NIST • Andrew C. E. Reid & Andrew R. Roosen, CTCMS, MSEL, NIST • Andrew J. Allen, Jan Ilavsky, Jay S. Wallace, Ceramics Div., NIST • Scott Terry & Carlos Levi, Univ. of Calif. at Santa Barbara
Technical Issues for TBC’s • Correlate properties with microstructure Ø Ø Ø
to shorten materials development cycle to improve materials & processing to enable more reliable design
• Increase thermal protection • Increase life • Increase reliability,
i.e., predict life, or coating spallation APPROACH: Develop computational tools for elucidating influences of stochastic microstructural features (e.g., porosity) on physical properties; and provide insights into mechanisms that lead to TBC spallation via predictive micro-mechanical models of reliability.
PPM2OOF Tool
Micrograph Binary Image
§ § § §
Mesh
Convert micrograph to “.ppm” (portable pixel map) file Select & identify phases to create binary image Assign constitutive physical properties to each phase Mesh in PPM2OOF via “Simple Mesh” or “Adaptive Mesh” – multiple algorithms that allow elements to adapt to the microstructure
OOF Tool Virtual Experiments: constrained cooling
Visualize & Quantify: normal residual stresses
dT
Perform virtual experiments on finite-element mesh: § To determine effective macroscopic properties § To elucidate parametric influences § To visualize microstructural physics
Young’s Modulus of EB-PVD TBC’s Intercolumnar Gap Porosity: essentially constant through TBC thickness via area fraction measurements
5 µm
t = 60 µm
Ts = 1100 °C 8 RPM 0.9 µm/min
t = 40 µm 66 mm
t = 20 µm
TBC
0.75 ° t = 1-2 µm
FeCrAlY t = 0 µm Courtesy of Scott Terry & Carlos Levi, University of California, Santa Barbara
EB-PVD TBC Cross Sections: 11716
11708
E÷÷ = 126 GPa
E÷÷ = 173 GPa E^ = 156 GPa 1 mm
E^ = 59 GPa 1 mm
Near interface microstructure
Surface microstructure
1.3 m m from TGO/TBC interface
60 m m from TGO/TBC interface
Courtesy of Scott Terry & Carlos Levi, University of California, Santa Barbara
Porosity: key key to to Porosity: TBC Performance Performance TBC Intercolumnar î Compliance
Intracolumnar î Insulation
200 nm
Different scales of shadowing
10 µm
100 nm
TEM courtesy of E. Sommer and M. Rühle MPI-Stuttgart
Courtesy of Scott Terry & Carlos Levi, University of California, Santa Barbara
Young’s Modulus of EB-PVD TBC’s Young’s Modulus (GPa)
Modulus is anisotropic and position dependent E÷÷
E^ Ts = 1100 °C
Height above TGO (mm)
Effect of of Temperature Temperature on on Surface Surface Effect Morphology and and Texture Texture Morphology Rotating Substrates
10 µm
Ts = 1100 °C Ts = 1000 °C
8 RPM, ~0.9 µm/min (Flux ~ 3.2 µm/min)
Ts = 900 °C Courtesy of Scott Terry & Carlos Levi, University of California, Santa Barbara
Young’s Modulus of EB-PVD TBC’s Young’s Modulus (GPa)
Modulus is anisotropic and position dependent E÷÷
¢ £
Ts = 900 °C Ts = 1100 °C
E^
Height above TGO (mm)
Section View of a ZrO2 – 8 wt% Y2O3 Plasma Sprayed Thermal Barrier Coating Free Standing Monolith more than 50 plasma spray parameters! ZrO2 – MgO
50 mm
AM100
Fabricated at Thermal Spray Laboratory, SUNY, Stony Brook (Jan Ilavsky)
Young’s Modulus versus Porosity 220
ZrO 2 - 8 w/o Y 2O3 (sintered) 2.381 mm f WC Indenter 4.8 N Load (Evaluated at 4 N)
200
Literature Value
Modulus [GPa]
180 160
Sintered Samples
140 120 annealed, 1 h
100
T
80
cross section
60 40
Amdry 142: 115 mm spray distance
20
intralamellar microcracks
plan section
0 80
85
90
95
100
Density [% rth] J. S. Wallace and J. Ilavsky, J. of Thermal Spray Tech., 7, [4], 521-526 (1998).
Calculating Average Elastic Properties of a Representative Region
85 mm
Effective Elastic Young’s Modulus Calculated From Microstructural Finite Elements Porosity Section Moduli (%) (GPa) Ex Calc.:
12 ± 1
Expt.:
11
Porosity (%)
Ey
39 ± 10 13 ± 7
6
28
11
Plan Moduli (GPa) Ex
Ey
83
90 19
ZrO2 – 8 wt% Y2O3: E = 214 GPa n = 0.310
How to Generate Effective Meshes?
Original Image
Low Res. Mesh Ex = 136 GPa Ey = 100 GPa Nodes = 10,894
High Res. Mesh Ex = 83 GPa Ey = 40 GPa Nodes = 43,887
Modulus versus Mesh Resolution (
160
)
AMLH (Ex) AMLH (Ey) AMTTA (Ex) AMTTA (Ey) SXLD (Ex) SXLD (Ey) SXTTC (Ex) SXTTC (Ey)
140
Modulus (GPa)
120
100
80
60
40
20
0 0
2 10
4
4 10
4
6 10
4
8 10
4
Number of Nodes
1 10
5
1.2 10
5
Modulus versus Mesh Resolution (
160
140
If the structure is truly random, then errors in simulation might vary with the spacing between nodes:
Eµ
1 N
Modulus (GPa)
How accurate should mesh be?
120
100
)
AMLH (Ex) AMLH (Ey) AMTTA (Ex) AMTTA (Ey) SXLD (Ex) SXLD (Ey) SXTTC (Ex) SXTTC (Ey)
80
60
40
????
20
0 0
0.002
0.004
0.006 1/Root(N)
0.008
0.01
0.012
Critical Issues § Does the mesh resolution capture the
essential features that affect behavior? E.g., Does the mesh capture the essence of the fine cracks?
§ If the mesh resolution is changed, does the simulated behavior also change?
§ This there asymptotic behavior? § Can these techniques be validated in a general manner?
Modeling Mechanical Behavior of TBC’s SUMMARY: § Microstructure-based, finite-element simulations
§ § §
provide a new paradigm for property measurements of complex materials, such as, TBC’s. Sample preparation & image analysis are critical for obtaining accurate, quantitative measures of behavior. Mesh resolution can have significant influences on determined properties. Finite-element simulations help to elucidate the influences of stochastic microstructural features (e.g., porosity) on the elastic behavior of complex TBC microstructures.
