evaluation and characterisation of thermal barrier coatings
Post on 31-May-2022
10 Views
Preview:
TRANSCRIPT
Evaluation and characterisation of
thermal barrier coatings
A thesis submitted to The University of Manchester for the degree of
Doctor of Philosophy
in the Faculty of Engineering and Physical Sciences
2013
Yang Zhao
Materials Science Centre
School of Materials
LIST OF CONTENTS
PAGE 2
List of Contents
List of Contents ··················································································· 2
List of Figures ····················································································· 6
List of Tables ······················································································13
Abstract ······························································································14
Declaration ·························································································15
Copyright Statement ···········································································16
Acknowledgement··············································································17
Chapter 1 Introduction ·······································································18
1.1 Demand for advanced gas turbine engine ··············································· 18
1.2 Introduction of thermal barrier coatings ·················································· 19
1.3 Goal of the dissertation ············································································· 21
Chapter 2 Literature review ································································24
2.1 Thermal barrier coating system ······························································· 24
2.2 Ceramic topcoat ························································································ 26
2.2.1 yttria-stabilised zirconia ·············································································· 26
2.2.2 New TBC materials ···················································································· 28
2.3 Bond coat and its oxidation ······································································ 30
2.3.1 Diffusion and overlay coatings ··································································· 32
2.3.2 Bond coat properties ··················································································· 33
2.3.3 Bond coat oxidation ···················································································· 37
2.4 Processing of thermal barrier coating ······················································ 39
2.4.1 Electron beam physical vapour deposition of TBC ······································ 42
2.4.2 Atmospheric plasma spray deposition of TBC ············································· 44
2.4.3 Alternative processing technologies ···························································· 47
2.5 Failure phenomena···················································································· 50
2.5.1 General principles in TBCs failure ······························································ 51
2.5.2 Specific mechanisms of imperfections governing ······································· 53
LIST OF CONTENTS
PAGE 3
2.5.3 Foreign attack and environmental degradation ············································ 58
2.6 Summary ··································································································· 60
Chapter 3 Investigation of interfacial properties of atmospheric plasma
sprayed thermal barrier coatings with four-point bending and
computed tomography technique ························································61
3.1 Introduction ······························································································· 61
3.2 Experiments······························································································· 63
3.2.1 Materials ···································································································· 63
3.2.2 Preparation of the four-point bending samples ············································ 63
3.2.3 Experimental procedures ············································································ 64
3.3 Results ······································································································· 66
3.3.1 TBC phase fractions ··················································································· 66
3.3.2 Four-point bending test ··············································································· 67
3.3.3 Microstructure observation ········································································· 68
3.3.4 Determination of the mechanical properties of TBCs ·································· 70
3.4 Discussion ································································································· 71
3.4.1 Analytical considerations ············································································ 71
3.4.2 Estimation of the interfacial toughness························································ 72
3.4.3 Further discussion ······················································································· 77
3.5 Summary ··································································································· 79
3.6 Appendix ··································································································· 79
Chapter 4 Local stress around spherically symmetrical portions of
thermally grown oxide layer formed on a metal substrate ··················81
4.1. Introduction ······························································································ 81
4.2 Experiments······························································································· 83
4.3 Results ······································································································· 83
4.3.1 Morphology characterisation and microstructure observation······················ 83
4.3.2 Stress measurements ··················································································· 88
4.4 Discussion ································································································· 89
LIST OF CONTENTS
PAGE 4
4.4.1 Analytical solutions for the local stress around spherically symmetrical
portions of TGO ·································································································· 89
4.4.2 Effect of curvature radius, depth and TGO thickness on local stress ············ 90
4.4.3 Effect of oxidation time and substrate thickness on local stress ··················· 92
4.4.4 Micro crack patterns caused by TGO stress················································· 97
4.5 Summary ··································································································· 98
Chapter 5 Microstructure evolution and interface morphology in
thermal barrier coatings studied by X-ray microtomography··············99
5.1 Introduction ······························································································· 99
5.2 Experiments····························································································· 101
5.2.1 X-ray computed tomography ···································································· 101
5.2.2 Experimental procedures ·········································································· 103
5.3 Results and discussions ·········································································· 106
5.3.1 3D visualisation of TBCs ·········································································· 106
5.3.2 Microstructure evolution and damage accumulation ································· 110
5.3.3 Characterisation of interface morphology ················································· 114
6. Summary ··································································································· 121
Chapter 6 Structure, oxidation resistance and mechanical properties of
simple and Pt-modified aluminide coatings on superalloy ················ 122
6.1 Introduction ····························································································· 122
6.2 Experiments····························································································· 123
6.2.1 Experimental procedures ·········································································· 123
6.2.2 Micro instrumented indentation ································································ 124
6.3 Results and discussions ·········································································· 128
6.3.1 Microstructure characterisation ································································· 128
6.3.2 Oxidation resistance ················································································· 132
6.3.3 Mechanical properties ··············································································· 133
6.4 Summary ································································································· 137
Chapter 7 Temperature dependence of Raman scattering of
yttria-stabilised zirconia ··································································· 138
LIST OF CONTENTS
PAGE 5
7.1 Introduction ····························································································· 138
7.2 Experiments····························································································· 138
7.3 Results ····································································································· 139
7.4 Discussions ······························································································ 141
7.5 Summary ································································································· 146
Chapter 8 Conclusions and future work············································ 147
8.1 Discussion and Conclusions ··································································· 147
8.2 Future work ····························································································· 150
References ························································································ 152
Words count: 38,248
LIST OF FIGURES
PAGE 6
List of Figures
Figure 1.1 Increase of operational temperature of turbine components made possible by
alloy development, manufacturing technology and thermal barrier coatings. [6] .......... 19
Figure 1.2 photography of a turbine blade (~10 cm long) with thermal barrier coating
(TBC) from the high-pressure hot section of an Engine Alliance GP7200 aircraft engine,
and a scanning electron microscope (SEM) image of a cross-section of an electron beam
physical vapor deposited 7 wt% yttria-stabilised zirconia TBC. [3, 22] ....................... 21
Figure 2.1 Cross-section scanning electron micrograph (SEM) of (a) atomspheric
plasma sprayed (APS) after annealing at 1150 °C for 5 hours and (b) electron beam
physical vapour deposited (EBPVD) TBCs, showing the constituents in TBCs. .......... 25
Figure 2.2 Schematic of the multi-layer structure in thermal barrier coatings, with
properties or functions for different layers indicated. (Redrawn from reference [3]) .... 26
Figure 2.3 Phase diagram of the ZrO2- Y2O3 binary system (zirconia rich corner). [32]
................................................................................................................................... 28
Figure 2.4 Ternary Ni-Al-Cr phase diagram predicted by the Calphad method, and
approximate composition of three different bond coat classes shown in scanning
electron microscope images. Shifts in the amounts of Al, Cr, and Ni in these coatings
permit changes in the predominant phase. [44] ............................................................ 31
Figure 2.5 Coefficients of thermal expansion (CTEs) for a standard diffusion aluminide
bond coat and commercial third generation Rene N5 Ni-based superalloy. [44] .......... 35
Figure 2.6 The tensile strength of a NiCoCrAlY overlay bond coat, a platinum modified
diffusion aluminide bond coat, and a development ruthenium aluminide alloy as a
function of temperature. [44] ....................................................................................... 36
Figure 2.7 Topographic profilometer (optical) images (top view) of the identical area of
a diffusion aluminide bond coat (without topcoat TBC) after polishing flat and then
thermal cycled (1 hour cycle) between room temperature and 1150 °C for the cycles
indicated. As is evident from the sequence of images, the magnitude of rumpling
surface instability increases with cycling but the microstructure does not. The colour
scale at the right indicates the rumpling height variation. [3, 90] ................................. 39
Figure 2.8 Microstructures and defects in electron beam physical vapour deposition
(EBPVD) and atmospheric plasma sprayed (APS) thermal barrier coatings (TBCs). The
APS TBC was annealed at 1150 °C for 5 hours; (a, c) scanning electron micrograph
LIST OF FIGURES
PAGE 7
(SEM) of polished coating cross sections; (b) SEM of fractured EBPVD TBC cross
section revealing feathery features and inter-columnar gaps. (d) SEM of fractured APS
TBC (top view) showing the “splats” (impacted and solidified droplets). (image a, b are
adapted from reference [14]) ....................................................................................... 41
Figure 2.9 Schematic of electron beam physical vapour deposition (EBPVD) processing
is shown, where the orange and green represent two vapour clouds of different
chemistry that can be mixed on the airfoil [92], (a-c) scanning electron micrographs of
three as-deposited EBPVD TBC morphologies with “intermediate”, “fine” and “coarse”
columns produced by different processing conditions, (d-f) are top view images [94]. 44
Figure 2.10 Schematic of atmospheric plasma spray deposition process for thermal
barrier coatings in which a ceramic feedstock is carried to the substrate by a plasma
spray. Scanning electron micrographs show (a) the feedstock powder, (b) an individual
component of the coating assembly (splat), and (c) a polished cross section of an
aggregated coating. [92] .............................................................................................. 47
Figure 2.11 Images of the cross sections of (a) traditional (non-segmented) and (b)
highly segmented thermal barrier coatings produced at low and high substrate
temperature, respectively, with enlarged scanning electron micrographs obtained from
fractured cross sections of the coatings. [92] ............................................................... 49
Figure 2.12 (a) Incipient buckling of a TBC coating viewed under reflected light. (b)
The surface revealed by spallation of the TBC consists of a mixture of local failure
between the TGO and the bond coat (appearing dark) and in the TBC itself (light
regions). [14] .............................................................................................................. 52
Figure 2.13 Schematic illustration of the buckling of a compressed film with a
pre-existing flaw of diameter dc................................................................................... 53
Figure 2.14 (a) A schematic of two major categories of TGO imperfection that govern
the TBC failure sequence; (b) a thickness imperfection in a TGO grown on a
NiCoCrAlY bond coat; (c) an undulation imperfection that develops in a Pt-aluminide
system upon thermal cycling. [9]................................................................................. 55
Figure 2.15 Microstructure of an initially flat aluminide bond coat after 50×1h cycles at
1200 °C: (a) surface rumpling; (b) cross section showing a rather uniform oxide layer
and strong surface undulations (γ‟ phase is revealed by etching); (c, d) optical
micrographs showing etched cross section before and after cyclic oxidation. Dark areas
on the optical images correspond to β phase while γ‟ phase appears white. [119] ........ 57
LIST OF FIGURES
PAGE 8
Figure 2.16 (a) Cross section scanning electron micrograph of 7YSZ EBPVD TBC fully
penetrated by a model calcium-magnesium-alumino-silicate (CMAS) melt in a
laboratory experiment. Crystalline phases with different compositions from the parent
7YSZ material (lighter gray) are noted (a) at the interface between the coating and the
melt. (b) The corresponding Si elemental map showing the extensive CMAS penetration.
[3] ............................................................................................................................... 59
Figure 2.17 Extended menu of failure mechanisms typical of current thermal barrier
coatings (TBCs) [11]. Three general modes of CMAS damage (lower right),
characteristic of higher temperature operation, have been identified so far. One involves
delamination cracks propagating through the TBC, another leads to chemical attack of
the thermally grown oxide (TGO) with concomitant loss of adherence, and a third result
from creep cavitation of the bond coat below a heavily penetrated TBC. [123]............ 60
Figure 3.1 A schematic of the four-point bend test (a) [143] in a typical experiment (b)
[144]. .......................................................................................................................... 64
Figure 3.2 A schematic of the Xradia X-ray computed tomography arrangement. ....... 66
Figure 3.3 (a) XRD patterns of the TBC exposed at 1150 ºC for 0, 10, 50, 100 and 200h
in the 20-90º 2 range. (b) XRD patterns in the 27-33º 2 range and the volume fraction
of the monoclinic phase as a function of thermal exposure time (inset)........................ 67
Figure 3.4 A typical load-displacement curve of the four-point bend test..................... 68
Figure 3.5 SEM images of cross-sections at the interface between the TBC and bond
coat after four-point bend test in the (a) as-sprayed condition and after (b) 10h, (c) 100h
and (d) 200h of thermal exposure at 1150 °C, showing that the cracks propagate above
the TGO within the top coat. [144] .............................................................................. 69
Figure 3.6 Equivalent micro-tomography slices of approximately the same region taken
from 3D images of the microstructure of a APS TBCs sample exposed at 1150 ºC for 0
(a), 20 (b), 120 hours (c), indicating various kinds of imperfections developed near the
YSZ/BC interface. ...................................................................................................... 69
Figure 3.7 (a) Young‟s modulus and (b) hardness of the components in the TBCs as a
function of thermal exposure time. .............................................................................. 70
Figure 3.8 A schematic of the interface cracking model. [144] .................................... 72
Figure 3.9 The energy release rate as a function of the thermal exposure time. ............ 73
LIST OF FIGURES
PAGE 9
Figure 3.10 Stress intensity factor as a function of the thermal exposure time. The circle
and diamond marks indicate the stress intensity factor of mode I and mode II,
respectively. Dashed lines connect the average values. ................................................ 74
Figure 3.11 Reconstructed images of (a) the TBC and bond coat and (b) the morphology
of the interface on the TBC side extracted from (a). The interfacial information can then
be analysed such as the interfacial roughness, the amplitude and wavelength of the
surface etc. It is found that the interfacial roughness did not increase obviously after
oxidation. .................................................................................................................... 78
Figure 4.1 The profiles of a typical indent created by a spherical indenter with 200 µm
radius using 30 N loading prior to (a) and after (b) oxidation at 1200 ºC for 25 hrs. The
inset is the line scan of the location indicated by the dotted line. ................................. 85
Figure 4.2 (a) SEM image of a typical indent created by a spherical indenter with 200
µm radius using 30 N loading after oxidation at 1200 ºC for 1 hrs; (b) higher
magnification of (a); (c) the microstructure of TGO inside the indent; (d) ZrC
precipitates formed on surface of Fecralloy after oxidation. The ZrC comes from the
diffusion from the substrate. ........................................................................................ 86
Figure 4.3 (a) SEM image of an indent created by a spherical indenter with 200 µm
radius using 30 N loading after oxidation at 1200 ºC for 9 hrs; (b) the grain
microstructure of TGO formed on the substrate. .......................................................... 87
Figure 4.4 Sketch of an undulating TGO. The predominant growth mechanisms are
schematically depicted particularly for TGO convex and concave portions. [166] ....... 88
Figure 4.5 A typical profile of the peak shift of the characteristic R-line of α-Al2O3 scale
around the indent created by a 200 µm-radius indenter with 30 N loading after oxidation
at 1200 ºC for 25 hours. .............................................................................................. 89
Figure 4.6 (a) Peak shift of the characteristic R-line of α-Al2O3 scale formed inside the
indents on Fecralloy after oxidation at 1200 ºC for 25 hours as a function of indent
radius and indent depth. (b) Normalised σzz and σxx (σyy) as a function of the indent
radius and oxide thickness ratio, R/H. ......................................................................... 91
Figure 4.7 (a) Peak shift of the characteristic R-line of α-Al2O3 scale at the indents with
20 and 200 µm radius as a function of (a) oxidation time at 1200 ºC (given substrates
are 2 mm thick) and (b) substrate thickness (given the oxidation time is 25 hours). With
increasing oxidation time the TGO thickness increases and the TGO stress in the indents
decreases. Meanwhile the substrate thickness has no effect on the TGO stress. ........... 93
LIST OF FIGURES
PAGE 10
Figure 4.8 Calculated growth stress of the oxide scale formed on a flat surface and the
indents with 20 and 200 µm radius as a function of oxidation time at 1200 ºC. ............ 95
Figure 4.9 (a) Micro cracks initiate at the ridges of the indenter where tensile hoop
stress arises at this convex portion; (b) higher magnification image of the circled area in
(a), showing the cracks propagate along grain boundary. ............................................. 96
Figure 4.10 (a) 70 degree tilted SEM image of a typical undulating morphology of TGO
formed on FeCoCrAlY substrate after oxidation at 1200C for 4 hrs; (b) one spallation of
TGO on the surface by buckling. ................................................................................. 97
Figure 5.1 Schematic of X-ray computed tomography. .............................................. 102
Figure 5.2 (a) EBPVD TBCs sample prepared, ready for acquisition of X-ray
radiographs. The sample is glued by epoxy onto a plastic tube fastened on a nail and
then put on the sample stage between X-ray source and detector in Versa X-ray µCT
machine (b). .............................................................................................................. 103
Figure 5.3 Procedures of X-ray tomography study, (a) acquisition of radiographs of an
EBPVD TBCs, (b) reconstruction of 2D slices from radiographs, (c) stacking the 2D
slices to build the 3D data, ready for image analysis such as visualisation, segmentation
and quantification. .................................................................................................... 105
Figure 5.4 (A) Reconstructed X-ray tomography images of the microstructure in an
EBPVD TBC deposited on a platinum-modified aluminde bond coat coated on CMSX-4
superalloy after thermal cycling (50×1h at 1150 °C), (B, C) slices of microstructure in x
and y directions, (D-F) microstructure at different slices in the through coating thickness
direction (z), showing the interdiffusion zone and the inter-columnar porosity that
increases from the bottom to the top of TBC. ............................................................ 110
Figure 5.5 The evolution of the microstructure in the EBPVD TBCs as a function of
thermal cycling number, showing (A) inter-columnar spacing and cracking in TBC
caused by sintering of the TBC, (B) TGO thickening, and (C) severe damage at the edge
of the bond coat. ....................................................................................................... 112
Figure 5.6 The reconstructed volume rendering images of the EBPVD TBCs as a
function of thermal cycling number, (A) the outer surface of the sample, revealing the
damage accumulation at the outer surface, and (B) the view of the inside structure by
slicing the data, both showing the damages at the bond coat edge and voids increase
with thermal cycling. ................................................................................................ 114
Figure 5.7 Reconstruction for each constituent layer in the as-deposited and as-heated
TBCs (100×1h at 1150 °C) after segmentation, revealing each interface between layers,
LIST OF FIGURES
PAGE 11
(A) substrate superalloy, (B) interdiffusion zone in bond coat, (C) outer layer in bond
coat, (D) TGO (the TGO in the as-deposited sample is too thin to be identified), (E)
TBC topcoat (it is set translucent to allow underneath layers seen). This segmentation
technique makes each interface can be extracted for further study. ............................ 117
Figure 5.8 The evolution of the TGO/bond coat interface with thermal cycling
(TBC/bond coat interface for as-deposited TBCs,). The interface is extracted from
segmented data. ........................................................................................................ 118
Figure 5.9 Quantification of interface morphologies in matlab by input of segmented
data. .......................................................................................................................... 118
Figure 5.10 Cross-sectional electron scanning microscopy (SEM) micrographs near the
thermally grown oxide (TGO) interface for as-deposited and as-thermally-cycled
EBPVD TBCs. Such 10 images are combined to quantify the interfacial roughness. . 120
Figure 5.11 the TGO/bond coat interface morphologies for as-deposited and thermally
cycled TBCs obtained from cross-sectional SEM images. The calculated interfacial
roughness is indicated along with the plotted lines. ................................................... 121
Figure 6.1 Schematic of indentation load–displacement data showing important
measured parameters. [197] ...................................................................................... 125
Figure 6.2 Schematic representation of the indenter-sample contact. [197] ................ 125
Figure 6.3 (a) Schematic of Vickers-produced indentation-fracture system, showing
peak load P and characteristic dimensions c and a of cracks, (b) Scanning electron
micrographs of radial crack system in a brittle material, sapphire, with P=10 N load
[198]. ........................................................................................................................ 127
Figure 6.4 Surface microstructure of simple aluminide coating. ................................ 128
Figure 6.5 (a) Cross-section SEM micrographs of simple aluminide coating, (b) the
concentration profile of elements along the line marked in (a). .................................. 129
Figure 6.6 Surface microstructure of Pt-modified aluminide coating. ........................ 130
Figure 6.7 X-ray diffraction (XRD) patterns from the Pt-modified aluminide coating
surface. ..................................................................................................................... 130
Figure 6.8 (a) Cross-section SEM micrographs of Pt-aluminide coating, (b, c) the
concentration profile of elements along the line marked in (a). .................................. 131
Figure 6.9 Thermogravimetric analysis (TGA) results of the simple and Pt-modified
aluminide coatings tested at 1150 °C for (a) 9 hours and (b) 20 hours. ...................... 133
Figure 6.10 Berkovich indentations of 2 μm depth on the simple and Pt-modified
aluminide coatings, labelled are indents. ................................................................... 133
LIST OF FIGURES
PAGE 12
Figure 6.11 Indentation load-displacement curves of several tests on both simple and
Pt-modified aluminide coatings, indicating more ductility for the later one. .............. 134
Figure6.12 Load-displacement curves of indentation made with 1 and 2 N loads on
simple aluminide coating. Inserted are the optical images of the corresponding Vickers
indentations, revealing large cracks for the 2 N case, but not for 1 N. ........................ 136
Figure6.13 Load-displacement curves of indentation made with 2 and 4 N loads on
Pt-modified aluminide coating. Inserted are the optical images of the corresponding
Vickers indentations.................................................................................................. 136
Figure 7.1 Raman spectra of tetragonal 8YSZ at various temperatures after subtracting
baselines. .................................................................................................................. 140
Figure 7.2 Peak positions of the Raman spectra for tetragonal 8YSZ as a function of
temperature (solid dots). The black solid lines are the best linear fits to the experimental
data and the red dash lines are the results predicted by theoretical calculations using the
methods from references [208, 210]. ......................................................................... 141
Figure 7.3 FWHM of each Raman band of tetragonal 8YSZ as a function of temperature.
................................................................................................................................. 144
Figure 7.4 Raman spectra of cubic YSZ single crystals with 9 wt% (a) and 30 wt% (b)
yttria content at various temperatures after subtracting baselines. .............................. 145
Figure 7.5 Temperature dependence of the Raman band F2g for cubic 9YSZ and 30YSZ.
................................................................................................................................. 146
LIST OF TABLES
PAGE 13
List of Tables
Table 3.1 Thermal treatment conditions and x-ray micro CT settings for the samples. . 65
Table 3.2 The experimental data used to calculate the energy release rate and the values
obtained by the two methods. ...................................................................................... 75
Table 4.1 The depth, span and calculated radius of a typical indent shown in Figure 4.1.
................................................................................................................................... 84
Table 5.1 X-ray microtomography experiment settings for the TBC sample scanning 104
Table 5.2 3D interfacial roughness of the TGO/bond coat interface in this EBPVD
TBCs sample with thermal cycling. ........................................................................... 119
Table 6.2 Chemical compositions in area 1, 2 and 3 in Figure 6.8a. ........................... 132
Table 6.3 Hardness and elastic modulus of simple and Pt-modified aluminide coatings
measured by instrumented indentation. ..................................................................... 134
Table 7.1 Fitting parameters of the temperature dependence of Raman shift for
tetragonal 8YSZ both free standing and deposited on superalloy. .............................. 141
ABSTRACT
PAGE 14
Abstract
Evaluation and characterisation of thermal barrier coatings
Yang Zhao
The University of Manchester for the degree of Doctor of Philosophy in the
Faculty of Engineering and Physical Sciences
2013
Evaluation and characterisation of thermal barrier coatings (TBCs) have been conducted
correlating microstructure with physical and mechanical properties, to further
understand TBC failure mechanisms and performances in this thesis.
A modified four-point bending test was employed to investigate the interfacial
toughness of atmospheric plasma sprayed TBCs. The delamination of the TBCs
occurred mainly within the topcoat. The energy release rate increased from ~50 J/m-2
for as-sprayed conditions to ~120 J/m-2
after annealing at 1150 ºC for 200 hours with a
loading phase angle about 42º. Micro X-ray tomography revealed how various types of
imperfections developed near the interface and the 3D interface was characterised.
Stress measurements by photoluminescence piezospectroscopy (PLPS) and analytical
solutions were combined to investigate the local stress around spherically symmetrical
portions of a TGO layer formed on Fecralloy. Spherical indenters were used to create
curvature with different curvature radii and depths on alloys. The effect of curvature
radius on stress was found to be more significant than the depth of local curved area.
TGO stress as a function of oxidation time at the curved areas was also discussed.
Electron beam physical vapour deposited (EBPVD) TBCs with a β-(Ni,Pt)Al bond coat
on CMSX4 substrate were investigated by micro X-ray computed tomography (XCT).
The 3D microstructures evolution and damage accumulation were studied. 3D
interfacial roughness was calculated and compared to scanning electron microscope
image analysis. The calculated interfacial roughness did not change much even after
200 thermal cycles, indicating there was not obvious rumpling in this TBCs sample.
Commercial simple and Pt-modified aluminide coatings were studied and compared.
Both coatings consisted mainly of β-NiAl phase. Thermogravimetric analysis (TGA)
tests indicated that the Pt-modified aluminide coating was much more resistive for
oxidation than simple aluminide coating. Instrumented indentation was used to measure
the mechanical properties, showing the coatings had similar young‟s modulus around
130 GPa while Pt-modified aluminide coating was more ductile and had a higher
fracture toughness than simple aluminide coating.
The Raman spectra of yttria-stabilised zirconia (YSZ) in the temperature range of
25-1100 ºC were investigated. The peak shift and broadening were carefully analysed.
The thermal mismatch stress was found to have a negligible effect on the Raman shift.
The dependence can be used to monitor the temperature in YSZ without contact.
DECLARATION
PAGE 15
Declaration
No portion of the work referred to in this thesis has been submitted in support of an
application for another degree or qualification of this or any other university or other
institution of learning.
COPYRIGHT STATEMENT
PAGE 16
Copyright Statement
i. The author of this thesis (including any appendices and/or schedules to this
thesis) owns certain copyright or related rights in it (the “Copyright”) and s/he
has given The University of Manchester certain rights to use such Copyright,
including for administrative purposes.
ii. Copies of this thesis, either in full or in extracts and whether in hard or
electronic copy, may be made only in accordance with the Copyright, Designs
and Patents Act 1988 (as amended) and regulations issued under it or, where
appropriate, in accordance with licensing agreements which the University has
from time to time. This page must form part of any such copies made.
iii. The ownership of certain Copyright, patents, designs, trade marks and other
intellectual property (the “Intellectual Property”) and any reproductions of
copyright works in the thesis, for example graphs and tables (“Reproductions”),
which may be described in this thesis, may not be owned by the author and may
be owned by third parties. Such Intellectual Property and Reproductions cannot
and must not be made available for use without the prior written permission of
the owner(s) of the relevant Intellectual Property and/or Reproductions.
iv. Further information on the conditions under which disclosure, publication and
commercialisation of this thesis, the Copyright and any Intellectual Property
and/or Reproductions described in it may take place is available in the
University IP Policy (see
http://www.campus.manchester.ac.uk/medialibrary/policies/intellectual-property
.pdf), in any relevant Thesis restriction declarations deposited in the University
Library, The University Library‟s regulations (see
http://www.manchester.ac.uk/library/aboutus/regulations) and in The
University‟s policy on presentation of Theses.
ACKNOWLEDGEMENT
PAGE 17
Acknowledgement
I would like to express my sincere appreciation and gratitude to my supervisor, Prof.
Ping Xiao for giving me the chance to join his research group and his supervision and
support throughout my PhD study. I would also thank very much my senior colleague
Dr. Xiaofeng Zhao for his great help and guidance on my experiments and advice on my
project.
I wish to acknowledge China Scholarship Council for provision of financial support to
my study. I am very lucky to have the experience of studying in University of
Manchester in UK which not only helps my academic career, but also broadens my
horizons. I am more knowledgable and mature now and it will benefit me in future.
Thanks to all my dear colleagues, Fan Yang, Akio Shinmi, Ian Shapiro, Huixing Zhang,
Fangwei Guo, Fanfei Wang, Eddie Honorato, Chao Zhu, Ying Long, Yong Zhang, Yu
Dang, Nadia Rohbeck, Ying Chen, Mingwen Bai, Xin Gen and Xiaoxiao Lu from our
ceramic coating group. It has been a great pleasure to work and play with you. I have
had a lot of fun and do not feel lonely with your companies in a foreign country. I am
also very grateful to the friendly and helpful technician staffs in Materials Science
Centre, including Mr. Kenneth Gyves, Mr. Andrew Forrest, Mr. Andy Wallwork, Mr.
Michael Faulkner, Dr. Christopher Wilkins, Mr. Andrij Zadoroshnyj and Mr. Gary
Harrison etc. In addition, my thanks go to Prof. Phillip Withers and Dr. Robert Bradley
in the Henry Moseley X-ray Imaging Facility for their valuable help and provision of
access to X-ray tomography facility. Thanks to all friends I have made in UK. I will
cherish the memory of life in UK forever.
Last but not least, I would like to thank my dear parents, Yubao Zhao and Baohua
Wang who always believe I am the best, for their everlasting love and encouragement
throughout all my life.
CHAPTER 1 INTRODUCTION
PAGE 18
Chapter 1
Introduction
1.1 Demand for advanced gas turbine engine
In 21st century, there is a huge conflict between ever growing energy demand of human
society and environmental issues such as climate change, which results from carbon
dioxide released by burning fossil fuels for energy, until large scale energy generation
from renewable sources become more viable economically. The gas-turbine engines are
the most efficient engines humans have ever made, which are widely used to propel
airplanes and generate electricity all over the world. Even one minor improvement in
the efficiency of gas-turbine engines will have a very positive impact on the world‟s
energy situation.
The market of gas-turbine engines was worth about $ 42 billion worldwide in 2010,
with commercial airplane engine production accounting for $ 21 billion and the
reminder land-based engines for electricity generation [1]. With expected growth of
electricity demand and the recent technology of exacting natural gas from vast shale
resources originating in US, the number of gas-turbine engines will inevitably increase
in the coming decades [1]. At the same time, air travel is predicted to more than double
in the next 20 years [2]. Together, these demands are the driving forces for technology
innovations in advanced gas-turbine engines with improved efficiency [3].
Gas-turbine engines convert the energy from burning fuel into useable work via three
main elements, i.e., a compressor, combustor and turbine. The engine efficiency can be
increased by higher pressure ratio generated in the compressor or more efficient
CHAPTER 1 INTRODUCTION
PAGE 19
secondary systems (cooling, sealing), or from the material perspectives, higher turbine
entry temperature (TET), which is restricted by the high temperature capability of
material components applied in the hottest part of gas-turbine engine, namely, the
combustor and turbine [4]. In these arrangements, the nickel-based superalloys are used
almost exclusively. The last five decades have seen continual increase of TET provided
by the enhancement of temperature capability of superalloys and improvements of the
materials design and processing [5]. Figure 1.1 shows superalloys have matured over
the years from wrought to cast, then to the directionally solidified alloys, whereas in the
latest generation of turbines for the most demanding applications, such as the high
pressure turbine blade, single crystal material is employed [6].
Figure 1.1 Increase of operational temperature of turbine components made possible by
alloy development, manufacturing technology and thermal barrier coatings. [6]
1.2 Introduction of thermal barrier coatings
The superalloys seem to have reached its limit in temperature capability since in today‟s
engines the hot gas temperature exceeds the melting point of nickel-based superlloys by
250 ºC [7]. The only way the airfoil can survive in such an environment is by extensive
internal and external cooling, which, however, would reduce its engine efficiency [7].
Nevertheless, further increases in thrust-to-weight ratio of next generation aero engine
and electricity output will require even higher gas temperatures. To meet the
requirements, advanced coating systems, particularly thermal barrier coatings (TBCs)
have been developed and applied in the hottest part of gas-turbine engine, enabling
CHAPTER 1 INTRODUCTION
PAGE 20
gas-turbine engine to run at significantly higher temperature than its predecessor, thus
leading to a major step in increasing engine efficiency [8-11].
TBCs, typically made of 7 wt% yttria-stabilised zirconia (YSZ) provide a thermal
protection for the underlying metallic engine components as shown in Figure 1.2. The
application of TBCs combined with advanced air cooling technology has been a great
achievement in materials science and engineering, resulting in a dramatic increase in
gas temperature, much bigger than any earlier materials development including
application of single crystal superalloy [12, 13]. However, there are still some problems
and new challenges to the exiting TBCs with further demands for higher gas
temperature in pursuit for better engine efficiency. First, the TBCs were not „„prime
reliant‟‟ [9-11], which means TBCs were not considered in the design of temperature
capability of superalloys. Actually, in today‟s engines, only about half of the
temperature capabilities of TBCs are taken into account because of the lack of
confidence in processing reproducibility and reliability [3]. Even a small variation of
processing parameters may lead to huge changes to microstructure and thus to a bigger
scatter in mechanical properties and lifetimes [12, 14]. Second, because of the sheer
complexity of the multi-layer coating system in which inter-diffusions and mechanical
interactions occur between each layer at high temperature, also evolving with service
time, it is important to consider TBCs as a complex and integral material systems
[13-15]. There is a variety of failure modes of TBCs depending on deposition methods
and engine operating conditions etc. Accurate testing and evaluation of TBCs is
challenging. But it is essential to have a more comprehensive understanding of the
evolution of TBCs and failure behavior with a better characterisation of the material
properties especially at high temperature in order to take full use of the potential
afforded by TBCs. The third problem comes from issues with increasing gas
temperature, such as radiation heat transportation at higher temperature rather than the
phonon scattering mechanism predominant in the current temperature range, and also
the degradation of topcoat by molten deposits, which is due to the ingestion of
particulates like sand or volcanic ash. It is commonly referred to CMAS attack
(calcium-magnesium-alumino-silicate) [16-21]. These challenges require new designs
and innovations in TBCs development.
CHAPTER 1 INTRODUCTION
PAGE 21
Figure 1.2 photography of a turbine blade (~10 cm long) with thermal barrier coating
(TBC) from the high-pressure hot section of an Engine Alliance GP7200 aircraft engine,
and a scanning electron microscope (SEM) image of a cross-section of an electron beam
physical vapor deposited 7 wt% yttria-stabilised zirconia TBC. [3, 22]
1.3 Goal of the dissertation
As mentioned in the introduction part about the challenges to the existing TBCs, it is
essential to obtain a better description of TBCs properties by new testing and evaluation
methods and relate the properties to the microstructure evolution to further understand
the failure behavior of TBCs.
CHAPTER 1 INTRODUCTION
PAGE 22
Various samples have been investigated thanks to the supply of air plasma spray (APS)
TBCs samples mainly from Volvo Aero and Cranfield University and electron beam
physical vapor deposition (EBPVD) TBCs samples from Rolls-Royce plc. The material
properties have been tested and evaluated with various methods and carefully analysed.
Hopefully this work can contribute something new to the understanding of the complex
coating system.
The structure of the thesis is as following. In chapter 2, the fundamentals and
understandings of TBCs achieved so far by previous research are reviewed, which
provides basic understandings and are helpful in the guidance of experiments and
analysis of results.
In chapter 3, interfacial toughness between the topcoat and bond coat in APS TBCs is
measured by a modified four-point bending test and micro X-ray tomography is also
used to observe the 3D interface evolution of one sample non-destructively. The
purpose of the work is that interfacial toughness is supposed to be the failure criteria of
the coating system even though it is not easy to get reliable and reproducible results and
also in the complex TBC system there are other factors besides the interfacial toughness
which can affect the failure behavior significantly. The direct observation of 3D
microstructure evolution is always interesting.
In chapter 4, local stress in thermally grown oxide (TGO) formed on a spherical portion
of Fecralloy substrate made by spherical indentation is measured by photoluminescence
piezospectroscopy (PLPS) and the TGO growth stress in the curved area is derived by
analytical solutions and then discussed. Although the TGO stress measurement by PLPS
has been well established by renowned researchers since 1994, most work reported are
on the stress in flat TGO surfaces (in-plane stress) and only some numeric calculations
have been done on curved areas because explicit solutions are impossible for complex
geometry. In this work, spherical indentations make analytical solutions possible.
Combined with experimental measurements, TGO growth stresses at curved area are
obtained, which are rarely studied in literatures. It is important to understand the local
stress since it is in the local area where crack initiates, grows and coalesces before final
spallation or delamination.
CHAPTER 1 INTRODUCTION
PAGE 23
In chapter 5, besides to the tomography study of APS TBCs in Chapter 3, EBPVD
TBCs with Pt modified aluminide bond coat are investigated by micro X-ray
tomography with the purpose to study the damage accumulation with thermal cycling.
The microstructure evolution is clearly characterised and 3D interfacial roughness is
calculated.
In chapter 6, structure, oxidation resistance and mechanical properties of Pt modified
and simple aluminide bond coat are studied, with emphasis on the comparison of the
mechanical properties measured by instrumented indentation.
In chapter 7, temperature dependence of Raman scattering of yttria-stabilised zirconia
(YSZ) is studied. The series of Raman spectrum of YSZ attached with and without
substrate are measured from at room temperature up to 1100 ºC. The temperature
dependence is then analysed and discussed, with indication that the Raman spectrum
can be used to measure the temperature of YSZ, such as in TBCs without contact.
In chapter 8, the main results and conclusions are summarised, along with outlook for
future work.
CHAPTER 2 LITERATURE REVIEW
PAGE 24
Chapter 2
Literature review
2.1 Thermal barrier coating system
Thermal barrier coatings (TBCs) are refractory-ceramic coating applied to the surface of
metallic components in the hottest part of gas-turbine engine, widely used in the
propulsion and power generation industry [3, 9, 10, 13-15]. They comprise thermally
insulating materials with sufficient thickness and durability to sustain thermal gradients
between underlying metals and hot coating surface. The primary function of TBCs is to
provide a thermal barrier to protect the load bearing alloys. They also bring additional
benefits such as resistance to oxidation, corrosion and thermal shock. The application of
TBCs combining with adequate internal cooling makes it possible to reduce the
temperature of metallic parts thereby increasing the lifetime of the components or to
allow higher turbine entry temperature (TET) leading to higher engine efficiency.
The development of TBCs started from 1950s with the manufacture of first enamel
coatings for military engine components [23]. Then in the 1960s the first flame sprayed
ceramic layers with NiAl bond coats were used in commercial aero engine [24]. Since
then continual improvements in both TBC materials development and processing
technology have taken place in the subsequent decades. It was in 1980s when the TBCs
were significantly improved [12]. During this decade, yttria-stabilised zirconia (YSZ)
was identified as an exceptional TBC material, and has been established as a standard
material for TBC since then [25, 26].
CHAPTER 2 LITERATURE REVIEW
PAGE 25
Figure 2.1 Cross-section scanning electron micrograph (SEM) of (a) atomspheric
plasma sprayed (APS) after annealing at 1150 °C for 5 hours and (b) electron beam
physical vapour deposited (EBPVD) TBCs, showing the constituents in TBCs.
Originally, the thermal barrier coating refers to the ceramic topcoat itself, however,
because of the coupled diffusion and interactions between the ceramic coating and
underlying substrate at high temperature, it is essential to consider it as a complex,
interrelated and evolving materials system [3]. They comprise not only the ceramic
CHAPTER 2 LITERATURE REVIEW
PAGE 26
topcoat but also underlying alloy and two layers between them. These include an
intermetallic or metallic bond coat which provides the oxidation protection to the
superalloy and a thin scale, thermally grown oxide (TGO) which forms between the
topcoat and bond coat due to the oxidation of bond coat in service. The four primary
constituents (e.g., TBC, TGO, bond coat and substrate) in typical atmospheric plasma
sprayed (APS) and electron beam vapour deposited (EBPVD) TBCs are shown in
Figure 2.1. The microstructure of the TBC in EBPVD TBCs shown in Figure 2.1b is not
perfect columnar structure because during the manufacturing process the substrate is not
completely vertical to the deposition direction. Each of the elements in TBCs is
dynamic and all interact to control TBCs performance and durability [9]. Figure 2.2
illustrates the multi-layer structure in a typical TBC system with properties or functions
of each layer.
Figure 2.2 Schematic of the multi-layer structure in thermal barrier coatings, with
properties or functions for different layers indicated. (Redrawn from reference [3])
2.2 Ceramic topcoat
2.2.1 yttria-stabilised zirconia
The ceramic topcoat is a thermal insulator with its prime function to reduce the heat
transfer to metallic substrate. Today, the majority of topcoat are made of yttria
CHAPTER 2 LITERATURE REVIEW
PAGE 27
partially-stabilised zirconia (YSZ) with a composition of zirconia with ~7 wt% yttria.
Originally, the ceramic material was chosen empirically based on its low thermal
conductivity that is more than one order of magnitude below that of the superlloy and
one of the lowest at elevated temperature of all ceramics (~2.3 W·m-1
·K-1
at 1000 ºC for
a fully dense materials [27]) because of its high concentration of point defects (oxygen
vacancies and substitutional solute atoms), which scatter heat-conducting phonons
(lattice waves) [28]. YSZ also has a large thermal expansion (~11×10-6
ºC-1
) which
comes close to that of metal substrate (~14×10-6
ºC-1
) [7]. This helps alleviate the
stresses as a result of thermal expansion mismatch between the ceramic topcoat and
underlying substrate. In addition, Porosity are deliberately incorporated into the ceramic
coating to further mitigate the stress, making it “strain tolerant” and highly compliant. It
also has a high melting point (~2700 ºC) and resistance to erosion and corrosion [10].
Finally, YSZ has a demonstrated manufacturing capability for depositing it with
constant composition [3, 7, 29].
The high temperature phases of ZrO2 can be stabilised to room temperature by the
addition of other oxides such as Y2O3, MgO, CaO, CeO2, Sc2O3 and In2O3 etc., among
which Y2O3 is the most widely used [30]. Depending on the concentration and the type
of the stabilizer, zirconia ceramics can be classified into three major types according to
crystal structure: full stabilised zirconia (FSZ), partially stabilised zirconia (PSZ) and
tetragonal zirconia polycrystals (TZP) [31]. In FSZ, zirconia is in its cubic phase and is
widely used in oxygen sensors, fuel cell electrolytes and fake diamonds. The PSZ
consists of nanosized tetragonal or monoclinic particles that have precipitated out in a
cubic matrix. TZP is monoliths of tetragonal phase, which may contain a secondary
cubic phase [30]. Figure 2.3 shows the phase diagram of the ZrO2- Y2O3 binary solid
solution (zirconia rich corner) [32].
Different from all the phases mentioned above, the 7YSZ, currently used in most TBCs
are metastable tetragonal phase (t‟). It is desirable for TBC application because unlike
its tetragonal counterpart, t‟ phase does not undergo a martensitic phase transformation
during which tetragonal phase transforms to the monoclinic one, leading to large
volume change. Therefore, the thermodynamic metastable tetragonal phase is more
“stable” in this application. Besides, 7YSZ has been shown to have exceptionally high
fracture toughness especially at high temperature due to ferroelastic toughening [33, 34].
CHAPTER 2 LITERATURE REVIEW
PAGE 28
Different from other transformation-toughened zirconia-based ceramics used in bearing,
cutting tools and knives, the toughened toughness of 7YSZ does not come from a
martenistic phase transformation which is an irreversible and diffusionless collective
movement of atoms, but rather from reversible ferroelastic domain switching from one
tetragonal variant to another when stressed [3, 33, 34]. Also martenistic transformation
toughening can only occur at low temperature while ferroelastic toughening can operate
at high temperature, typical of those engine temperatures, which may be why 7YSZ has
the exceptional thermal cycling lifetime.
Figure 2.3 Phase diagram of the ZrO2- Y2O3 binary system (zirconia rich corner). [32]
2.2.2 New TBC materials
Despite of all the advantages of YSZ for TBC material application mentioned above,
there are certain limitations for the standard material due to sintering and phase
transformation especially at even higher temperature in pursuit of higher engine
CHAPTER 2 LITERATURE REVIEW
PAGE 29
efficiency. On prolonged exposure at elevated temperature, the metastable tetragonal (t‟)
will decompose into higher and lower yttria phase [12]. The later will transform to
monoclinic phase on cooling associated with a large volume change which would cause
TBC failure. Also sintering of topcoat will make TBC stiffer and reduce its compliance
thereby leading to early failure. In addition, the radiation heat transfer through TBC will
be much bigger at elevated temperature. Therefore, currently the accepted limit
temperature for use of YSZ is 1200 ºC [35-37]. As a result, research have been done on
searching for new ceramics better than YSZ, for instance, TBC materials with
pyrochlore structure A2B2O7 offer very attractive properties for application at service
temperature above 1300 ºC [12], specifically, the lower thermal conductivity of several
zirconate pyrochlore makes this kind of materials promising [38, 39]. In addition, the
thermal stability of the pyrochores is excellent which is probably related to the fixed
positions of cations in the crystal. Among the widely investigated pyrochlores are the
rare-earth zirconates (Ln2Zr2O7), where Ln is any or combination of La, Gd, Sm, Nd, Eu
and Yb [40-43] . And among the pyrochlores, La2Zr2O7 (LZ) seems to be one of the
most promising for TBC application due to the its outstanding bulk properties compared
to standard YSZ with a high thermal stability up to 2000 ºC, a low thermal conductivity
of 1.56 W·m-1
·K-1
and a low sintering tendency [12, 39]. However, LZ has a relatively
low thermal expansion (9×10-6
ºC-1
) compared to YSZ (11×10-6
ºC-1
) which would leads
to higher thermal mismatch stress upon cooling and possibly shorter lifetime, and
besides, LZ is thermodynamically incompatible with alumina as a TGO. Therefore, to
combine the advantages of both YSZ and pyrochores, a so-called double-layer system,
with YSZ as the first layer attached to bond coat and a top layer made of pyrochlore
materials, has been proposed and lifetime is significantly improved when tested in
thermal gradient cyclic rigs [42]. These double-layer systems based on
pyrocholore/YSZ revealed excellent high-temperature capability significantly better
than that of YSZ and they are expected to improve the performance of gas-turbine
engines during application [42]. These new types of TBC materials are very promising
with excellent thermal and mechanical properties and demonstrated processing
capability. Further development might reveal certain advantages of other ceramic
materials with respect to thermal cyclic performance and thermal stability [12].
CHAPTER 2 LITERATURE REVIEW
PAGE 30
2.3 Bond coat and its oxidation
The bond coat is an oxidation-resistant intermetallic (or mixed metallic and
intermetallic) layer, with primary function to provide a reservoir from which Al can
diffuse to form a protective α-Al2O3 thermally grown oxide during service. The bond
coat is arguable the most crucial component of TBCs as the coating system performance
is usually linked to the bond coat creep and yield properties governed by its
composition and microstructure, thereby, dictating the failure behaviours of TBCs [3, 9,
10, 44].
In early applications of propulsion and land-base turbines even before thermal barrier
coatings were developed, intermetallic coatings was used as environmental coatings to
serve a single function [44]. Aluminide coatings based on NiAl and NiCoCrAl coatings
became the standards for oxidation protection, while diffusion chromides and overlay
CoNiCrAl coatings were applied to protect against hot corrosion [5, 45]. As the turbine
entry temperature has increased and substrate alloys have evolved to nickel-based single
crystal superalloys, the environmental coatings have become multifunctional. They have
to provide hot corrosion resistance in cooler sections and intermediate-temperature
regions (blade shanks and below the blade platform), oxidation resistance in hottest
sections (turbine blade tips, platforms and airfoils), maintain adhesion to the ceramic
topcoat without reacting with it and minimise interdiffusion of deleterious refractory
elements at the interface from underlying substrate which may degrade the bond coat
properties and the interface adhesion with TGO [44, 46-48].
The intermediate layer between the ceramic topcoat and metallic substrate is called the
bond coat as the bonding to the deposited topcoat and underlying alloy was a major
concern, particularly for plasma-sprayed coatings, in the early days of TBCs
development [14]. The thickness of bond coat varies between 30-150 µm, depending on
the processing methods and service time [10, 44]. The bond coat is chemically complex
because of the need to optimise a broad set of thermomechanical and thermochemical
properties to serve multiple functions. Because the bond coat is deposited on
nickel-based superalloys, and either chromia or alumina forms due to oxidation during
service, the Ni-Al-Cr ternary diagram at 1100 °C is relevant (Figure 2.4) [44].
According to it, phases that exit include the β-NiAl phase, the fcc γ-Ni and γ-Al phases,
CHAPTER 2 LITERATURE REVIEW
PAGE 31
the Ll2 γ‟-Ni3Al phase and the α-Cr phase. Bond coat can be broadly classified by these
major phase constituents [5, 11, 49]. Coatings consisting primarily of the β-NiAl phase
are typically referred to as nickel aluminide coatings. If platinum is added into the
coatings, they are classified as platinum-modified aluminides. A combination of β and γ‟
phases form the basis of MCrAlY coatings (M=Ni, Co+Ni, or Fe). Although these
coatings are very different in terms of composition and microstructure, the challenges
are similar: minimise the deformation of bond coat at intermediate and operating
temperatures, mitigate the interdiffusion with substrate to prevent the formation of
brittle intermetallics, and deliver critical elements in addition to Al, such as Hf and Y, to
the growing TGO to retard its inelastic plastic deformation under thermal cycling [3].
As coating compositions have evolved to keep up with the functional demands of
engineering components, so have the processing methods for deposition. The challenge
of delivering ever-increasing coating functionality while maintaining robust and
cost-effective manufacturing procedures has resulted in a multiplicity of bond coat
processing, broadly classified into two categories: diffusion and overlay coatings.
Figure 2.4 Ternary Ni-Al-Cr phase diagram predicted by the Calphad method, and
approximate composition of three different bond coat classes shown in scanning
electron microscope images. Shifts in the amounts of Al, Cr, and Ni in these coatings
permit changes in the predominant phase. [44]
CHAPTER 2 LITERATURE REVIEW
PAGE 32
2.3.1 Diffusion and overlay coatings
Diffusion coatings, particularly aluminide coatings, are the product of interdiffusion
between the metallic substrate and an aluminium source (donor) [44]. The simplest and
earliest industrialised process is pack cementation. Components to be aluminised are
embedded into a mixture of the aluminium source, an inert matrix (typically alumina
sand) and a halide salt activator (for instance, AlF3 or NH4F). The aluminium sources
can be aluminium or aluminium alloy such as CrAl, CoAl, or NiAl. Alloy donors are
used to increase the donor melting temperature and define the chemical activity of
aluminium, allowing manipulation of coating microstructures for improved performance.
The pack is located in a retort and heated to 650-1200 ºC in a non-oxidising atmosphere,
such as argon or hydrogen. During the treatment, the halide activator transports
aluminium from the donor to the surface of target component. It then decomposes,
releasing the aluminum and allowing it to diffuse into the substrate and cycling the
halide back to the donor. Then the whole process repeats. At thermal treatment
temperature above ~1050 ºC and with low-aluminium activity donors, NiAl coatings
form via predominantly outward diffusion of cations. These coatings are typically single
β phase and have Al:Ni ratios less than unity. While for aluminisation at temperature
below~1000 ºC, and especially with high-activity donors (e.g., aluminium-containing
alloys), NiAl coatings grow mainly by inward diffusion of anions.
Other processes such as vapour phase aluminisation (VPA) and chemical vapour
deposition (CVD) have been developed to enable long range vapour phase
transportation of the aluminium source without the need to place the components
directly in the pack. In both cases, the components to be coated are placed in a high
temperature retort, and the aluminium-bearing vapours are transported to them by an
inert gas [50]. These processes have the benefit of allowing both external and internal
surfaces to be coated.
Elements including Cr, Si, Hf, Zr and Y have been incorporated into the diffusion
aluminide coatings to impart additional performance benefits such as enhanced
resistance against corrosion, cyclic oxidation resistance and improvement in interface
adhesion [51, 52]. One of the most widely adopted implementation is the addition of
platinum into simple aluminide to form platinum-modified aluminide [53-55]. Typically,
CHAPTER 2 LITERATURE REVIEW
PAGE 33
the platinum-modified aluminide coating (PtNiAl) are formed by first electroplating
5-10 µm platinum onto the substrate and then an inert atmosphere interdiffusion heat
treatment, followed by aluminising process (either pack or vapour phase) during which
aluminium diffuses into the surface of the substrate while nickel diffuses out and reacts
with the aluminium and platinum to form the PtNiAl bond coat. A variant of the
platinum diffusion coatings (Pt-γ/γ‟) which directly platinises the substrate without a
subsequent aluminising step has also been developed and demonstrated to improve the
bond coat performances [56-59].
While the diffusion aluminide coatings are widely used in turbine components, the
degree to which their composition, microstructure and thickness can be tuned is limited
by the constraints inherent to diffusion process [14]. Although the diffusion aluminides
are still the standards for all internal coatings, the external coatings are also
manufactured using overlay processes which provide the flexibility of tailoring the
complex multi-component systems [44].
MCrAlY overlay coatings can be deposited by a number of processes, producing
varying degrees of coatings density and process-induced oxidation [60]. Electron-beam
physical vapour deposition (EBPVD) can deposit clean, dense high quality MCrAlY
coating with highly reactive element additions (e.g., Hf, Y, Zr and Si) [61], however it
involves costly equipment and maintenance. Consequently various plasma spraying
methods are commonly used to deposit overlay coatings due to its lower cost compared
to physical vapour deposition processes and its ability to coat large components. Among
plasma spraying processes, low-pressure plasma spray (LPPS) is broadly used [62] but
still relatively costly due to the accompanying vacuum chamber [44]. Therefore, the
emergence of atmospheric plasma spray, including inert-gas shrouded plasma spray [63,
64] and high-velocity oxyfuel (HVOF) [65] has made the high volume deposition of
complex MCrAl-family coatings with reactive elements routine.
2.3.2 Bond coat properties
The bond coat properties and performance are crucial not only because TBCs durability
is governed through the structure and morphology of the TGO created when it oxidises,
but also the physical and mechanical properties of bond coat itself are essential to
CHAPTER 2 LITERATURE REVIEW
PAGE 34
influence TBCs failure modes, for example, the bond coat creep/yield strength (or
tensile strength), coefficient of thermal expansion (CTE) which is associated the thermal
misfit stress with superalloy substrate, phase stability, modulus and hardness, and
porosity change etc. A major challenge is to have reliable measurements and predictions
of the properties especially at elevated temperatures. The conventional experiments are
difficult to conduct mainly because of its reduced dimensionality. Although some of the
relevant properties of the most common coating systems have been measured,
unfortunately, property measurement is rarely incorporated as a tool for developing new
generations of coatings [44]. It is due to the lack of confidence and bond coat design is
mainly based on composition and structure consideration.
Although the strain energy in the TGO resulting from thermal mismatch and growth
strain is the driving force for TBCs failure as the coating system usually spalls or
delaminates at TGO/bond coat or TGO/TBC interface [9], the stress in the bond coat is
found to be essential as it influences the bond coat deformation especially at high
temperature and thus morphology stability. At temperature above 1000 ºC, CTE
difference between bond coat and superalloy substrate of 1×10-6
ºC-1
will generate
thermal stresses on the order of hundreds of MPa which is big enough to promote
plastic deformation of the bond coat at elevated temperatures and interfacial
delamination upon cooling [9]. The thin coatings prevent conventional dilatometer
measurements of CTE. Non-contact digital image correlation is employed to measure
the CTE for a standard diffusion aluminide bond coat and commercial single crystal
Rene N5 superalloy (Figure 2.5) [44]. Difference in CTE of about 2×10-6
ºC-1
are
present at room temperature but varies as a function of temperature and even reverses at
high temperature. Comparison of the CTE for commercial MCrAlY bond coat and
superalloy does not show such a phenomenon, with CTE difference of 2 to 3×10-6
ºC-1
at room temperature and increasing at high temperature. To more closely match the
thermal expansion of superalloy substrate, coatings with the same phase constituents
(γ/γ‟ phases), more optimal compositions for oxidation have been investigated [66, 67].
In addition to the CTE matching, these coatings are desirable because of the lower
driving forces for interdiffusion. Nevertheless, a drawback is the lower amount of Al
available in the coating reservoir for TGO formation.
CHAPTER 2 LITERATURE REVIEW
PAGE 35
Figure 2.5 Coefficients of thermal expansion (CTEs) for a standard diffusion aluminide
bond coat and commercial third generation Rene N5 Ni-based superalloy. [44]
The mechanical properties of bond coat have a crucial influence on the coating
durability in service. Diffusion aluminides have been shown to have a ductile-brittle
transition temperature (DBTT) of approximately 600 ºC [44]. Below the DBTT, the
bond coat is linearly elastic. But above the temperature, the ultimate tensile strength and
creep response of the diffusion aluminide bond coat drop dramatically and are very
temperature dependent (Figure 2.6). Strengths of 400 MPa have been measured at
intermediate temperatures, but above 1000 ºC, the strength of commercial diffusion
aluminide bond coat is below 50 MPa. Attempts to improve the elevated temperature
strength have been only minimally successful, with the greatest high temperature
strength being achieved by the development of a ruthenium aluminide bond coat
[68-70].
CHAPTER 2 LITERATURE REVIEW
PAGE 36
Figure 2.6 The tensile strength of a NiCoCrAlY overlay bond coat, a platinum modified
diffusion aluminide bond coat, and a development ruthenium aluminide alloy as a
function of temperature. [44]
Due to the variation of composition and microstructure of overlay coatings, their
properties can vary much more widely than those of diffusion coatings. It is possible to
characterise the properties better since very thick overlay coatings can be deposited such
that conventionally-sized test specimens can be machined though it may lose some
similarity of the microstructure. Overall, the physical properties of overlay coatings are
very similar to those of superalloys. However, the thermal expansion coefficient,
particularly of the coatings containing Co, tends to be greater than that of superalloys,
resulting in thermal misfit stress in the coatings. MCrAlY coatings have been reported
to be very strong at room temperature with an ultimate tensile strength of 1.4 GPa and
significant (>2%) ductility, but their elevated temperature strength is dramatically
reduced [71].
During thermal exposure, formation of the TGO and interdiffusion between the bond
coat and underlying substrate significantly deplete the aluminium content in the bond
coat which can lead to the martensite phase transformation of β-NiAl to γ‟-Ni3Al, and
eventually γ-Ni. The remaining β-NiAl phase regions often have the characteristic lath
structure of a martensite, so sometimes termed as martensite β‟-NiAl. The martensite
phase transformation is diffusionless and reversible. It is a displacive transformation.
CHAPTER 2 LITERATURE REVIEW
PAGE 37
The formation of the metastable L10 γ‟-Ni3Al can have a dramatic effect on rumpling if
the diffusionless transformation accompanied with volume change occurs at
intermediate temperatures, when the bond coat is easily plastically deformed [44]. The
martensitic start temperature (Ms) is extremely sensitive to composition. Ms of pure
β-NiAl phase is known to be around from room temperature to 300 ºC [72], the
additional Pt, Co, and Cr present in the PtNiAl bond coat increase the Ms to ~600 ºC
[73]. The addition of Pt raises the transformation temperature which promotes plastic
deformation of the bond coat [74, 75].
2.3.3 Bond coat oxidation
Upon service, a thermally grown oxide (TGO) forms due to the oxidation of bond coat
since the TBC microstructure is highly defective with micro-cracks and porosity and
YSZ is transparent to oxygen. The compositions of bond coat are selected in order to
preferentially form α-Al2O3 TGO because it is thermodynamically compatible with YSZ
[76] and furthermore, α-Al2O3 is usually considered to be the slowest growing oxide at
high temperature on account of its smallest oxygen diffusivity [77].
The TGO exerts a central role in controlling TBC system durability because of the
strain energy built and accumulated in the TGO which provides the motivation for
cracking in coatings. Actually some manufacturers are believed to use the critical
thickness of TGO as criterion to predict average life [78]. The essential mechanics of
this form of failure are similar to the origin of a critical thickness for the loss of
coherence of epitaxial thin films, namely when the release of stored elastic strain energy
in the growing film exceeds the fracture toughness [3].
There are two contributions to the stress in the TGO. One is associated with thermal
expansion misfit between the TGO and underlying substrate upon cooling, and the other
results from TGO growth which consists of a simple thickening component and another
lateral expanding one. During service, the simple thickening of TGO can be
accommodated by rigid displacement, but the lateral elongation in turn drives
out-of-plane instability as well as other mechanical responses [79]. Both these stresses
may be alleviated by TGO creep [80, 81] and redistributed in the vicinity of
imperfections [82, 83]. The origin of the lateral growth strain is poorly understood but is
CHAPTER 2 LITERATURE REVIEW
PAGE 38
generally attributed to the counter-diffusion of inward diffusing O2-
and outward
diffusing Al3+
resulting in formation of new Al2O3 in the TGO grain boundaries [84].
Measurements by X-ray diffraction [85] and photoluminescence piezospectroscopy
(PLPS) [86, 87] indicate that the thermal expansion mismatch stress is in compression,
on average, ranging from 3-6 GPa at room temperature. There have been a limited
number of direct measurements of the TGO growth stress, without the presence of TBC,
using X-ray synchrotron sources [85, 88, 89] showing the growth stress is also
compressive (0-1 GPa) and much smaller than thermal misfit stress, but not nearly
enough to follow the stress evolution during oxidation or thermal cycling. More
revealing have been the non-contact measurements by PLPS of the strain measured
through the TBC [22]. In this technique, a laser beam is used to penetrate through the
ceramic topcoat and excite the R-line luminescence from trace Cr3+
ions which are
invariably present in the TGO. The local mean stress in the TGO is proportional to the
frequency shift of the R-lines. This has enabled correlations to be mapped between
luminescence shifts and the development of local damage as the bond coat and TGO
rumple, as shown in Figure 2.7 [3, 90].
There remain several important unresolved questions about the lateral growth strain
including how minor elements, at the ppm level and above, affect the growth and
mechanical behaviour of the TGO [3]. Of particular interest are the elements Y, Zr, and
Hf that segregate, on account of their large ionic radii, to the grain boundaries of the
TGO. Among the key questions being raised are whether these elements alter the
counter-diffusion along the TGO grain boundaries that create the lateral growth strain
and how they affect the high-temperature creep and plasticity of the TGO. It is known
that rare-earth ions dramatically increase the creep resistance of alumina ceramics [91].
CHAPTER 2 LITERATURE REVIEW
PAGE 39
Figure 2.7 Topographic profilometer (optical) images (top view) of the identical area of
a diffusion aluminide bond coat (without topcoat TBC) after polishing flat and then
thermal cycled (1 hour cycle) between room temperature and 1150 °C for the cycles
indicated. As is evident from the sequence of images, the magnitude of rumpling surface
instability increases with cycling but the microstructure does not. The colour scale at the
right indicates the rumpling height variation. [3, 90]
2.4 Processing of thermal barrier coating
The widespread application of thermal barrier coatings in both propulsion and power
generation industries has, to a large extent, been enabled by the development of
advanced processing technologies [92]. The deposition processes for the bond coat have
been mentioned above in section 2.3.1. This section will review the manufacturing
methods for the TBC ceramics.
Due to the refractory nature of TBC materials such as yttria-stabilised zirconia (YSZ)
which has a melting point in excess of 3000 K, ultrahigh temperature processing
technologies are required. Therefore, the thermal plasma spray and electron beam
physical vapour deposition (EBPVD) have become the primary and preferred choices.
The former involves melt fabrication of powder ceramics while the latter is based on
evaporation and vapour deposition from ceramic ingots. Nowadays, the implementation
of such advanced processing is remarkable. Some 1-1.5 million kilograms of YSZ was
atmospheric plasmas sprayed (APS) onto engine components in 2011 [92]. Parts in
aero-engines deposited by APS include combustors, vanes, and turbine shrouds, and
also TBCs are plasma sprayed onto both rotating and stationary parts of large
land-based power generators, while the hottest section in turbine blades of aircraft
CHAPTER 2 LITERATURE REVIEW
PAGE 40
engines contain TBCs deposited by EBPVD. These applications of TBCs are expected
to grow in the coming decades due to the predicted increasing air travel and continuous
demand for fuel efficiency amid energy crisis.
Though the microstructures of APS and EBPVD TBC are quite different, they are both
designed to incorporate defective architectures to impart strain compliance which is
desirable for the coatings to endure thermal misfit between the ceramics and underlying
alloys, and also to help reduce the thermal conductivity. Indeed, both APS and EBPVD
coatings contain some 10-30 % porosity which results in a drop in thermal conductivity
of YSZ by 35-55% of bulk materials for EBPVD coatings and as high as 80% for APS
ones [93, 94]. In EBPVD TBC, the lateral strain compliance results from the columnar
structure and inter-columnar gaps produced by rotation of the component during
deposition (Figure 2.8a, b). The individual columns also contain microscopic porosity
that can reduce the thermal conductivity as well [14]. While in APS coatings, the lateral
strain compliance and low thermal conductivity are conferred by the incorporation of
porosity between “splats” of successively deposited materials (Figure 2.8c, d). The
EBPVD TBC is more compliant due to the columnar structure vertical to the TBC/bond
coat interface while APS TBC is more effective in thermal insulation because of its
lamellar structure and plate-like porosity parallel to the interface.
CHAPTER 2 LITERATURE REVIEW
PAGE 41
Figure 2.8 Microstructures and defects in electron beam physical vapour deposition
(EBPVD) and atmospheric plasma sprayed (APS) thermal barrier coatings (TBCs). The
APS TBC was annealed at 1150 °C for 5 hours; (a, c) scanning electron micrograph
(SEM) of polished coating cross sections; (b) SEM of fractured EBPVD TBC cross
section revealing feathery features and inter-columnar gaps. (d) SEM of fractured APS
TBC (top view) showing the “splats” (impacted and solidified droplets). (image a, b are
adapted from reference [14])
Although the two methods dominate current deposition processing for TBCs, there has
been significant interest and progress both in industry and academia to develop new
methods that combine the benefits of EBPVD and APS. Of particular interest is the
development of dense vertically cracked (DVC) or segmented crack microstructures
synthesised via advanced APS processing which also display vertical separations
similar to those of EBPVD [95, 96]. The introduction of suspension and solution
precursor plasma spraying further hybridise the benefits of feathery EBPVD coatings
and vertical cracking of the DVC structures [97]. The underlying principles of both
contemporary manufacturing methods (EBPVD and APS) and modification to
traditional APS processes (segmented TBC and suspension/solution sprays) and some
emerging hybrid technologies are described below.
CHAPTER 2 LITERATURE REVIEW
PAGE 42
2.4.1 Electron beam physical vapour deposition of TBC
Electron beam physical vapour deposition processing relies on evaporation of a material
from a melt, utilising a high vapour pressure over an overheated molten pool [98]. A
highly energetic electron beam is scanned over YSZ and evaporates it within a vacuum
chamber. Preheated substrates are positioned in the vapour cloud and the vapour is
deposited onto the substrates at deposition rates of several µm/minute. To achieve a
defined stoichiometry of the zirconia, oxygen is bled into the deposition chamber to
compensate for the deficit caused by dissociation. Due to the formation of the coating
from the vapour phase and combined actions of surface diffusion, shadowing and
crystallographic growth selection, a columnar microstructure of the TBC can be
achieved, providing a high level of strain tolerance (Figure 2.8a, b). To ensure
continuous growth of the ceramic coating, cylindrical ingots of the ceramic are
bottom-fed into the crucibles [92].
Formation of the microstructure of EBPVD TBC is closely connected to the processing
condition used [99]. Columns and inter-columnar gaps originate from vapour phase
condensation and macroscopic shadowing caused by the curved column tips, triggered
by rotation of the parts during deposition. Since shadowing occurs primarily along the
plane of vapour incidence, columns are significantly wider in the direction parallel to
the rotation axis than perpendicular to it, leading to an anisotropy of the in-plane
compliance with notable consequences to the strain tolerance of the TBC. Globular and
elongated spheroid pores are a consequence of rotation and the feathery features (Figure
2.8b) are due to the shadowing by growth steps on the column tips. Recent investigation
using ultra-small angle x-ray diffraction and small-angle neutron diffraction indicate
that the distribution of most elongated and feathery pores is also highly anisotropic
[100]. Intra-columnar pores are a combination of globular and elongated spheroids and
range between 18 and 25 nm in size. Image analyses indicate opening dimensions of
200-250 nm at feathery features with a typical aspect ratio of 1 to 10. To lower thermal
conductivity, EBPVD TBC relies mainly on elongated and feathery pores
(intra-columnar porosity), while inter-columnar porosity primarily provides compliance
[94].
CHAPTER 2 LITERATURE REVIEW
PAGE 43
The microstructure factors regarding intra-columnar porosity are size, distribution,
concentration, and morphology, all of which can be manipulated by processing
parameters including deposition temperature, rotational speed, chamber pressure,
pattern of vapour incidence, condensation rate and partial shadowing. Hence,
microstructure tailoring is viable, within limits, which might be set by durability,
processing cost issues and physical restrictions due to shadowing [101]. Schematic of an
EBPVD facility and examples of different microstructure produced by different
processing conditions are shown in Figure 2.9 [92, 94].
In summary, EBPVD processing produces TBC with the columnar structure which
provides the desirable strain tolerance. In comparison to plasma sprayed TBC, a higher
erosion resistance, a smoother surface finish that offers aerodynamic advantages, and
the fact that cooling holes stay open through the processing stages are key benefits. On
the other hand, high cost, higher thermal conductivity, and limits in chemical variability
due to vapour pressure issues are the drawbacks along with a low utilisation of the raw
materials.
CHAPTER 2 LITERATURE REVIEW
PAGE 44
Figure 2.9 Schematic of electron beam physical vapour deposition (EBPVD) processing
is shown, where the orange and green represent two vapour clouds of different
chemistry that can be mixed on the airfoil [92], (a-c) scanning electron micrographs of
three as-deposited EBPVD TBC morphologies with “intermediate”, “fine” and “coarse”
columns produced by different processing conditions, (d-f) are top view images [94].
2.4.2 Atmospheric plasma spray deposition of TBC
Thermal plasma spray is a molten droplet deposition technology in which tens of
micron-sized particles of metals and ceramics are introduced in powder form into an
arc-plasma jet and projected onto a prepared substrate [92]. The particles acquire
thermal energy and momentum from the thermal plasma and undergo melting, followed
CHAPTER 2 LITERATURE REVIEW
PAGE 45
by impact and rapid solidification. Typically, for oxides and even most metallic alloys,
plasma spraying is conducted under ambient condition, hence it is termed atmospheric
plasma spraying (oxidation sensitive metals can be sprayed in a low pressure
environment, LPPS), The resultant sprayed materials comprise an assembly of “splats”
which are rapidly solidified materials assemble to form a plate-like structure. Typical
powder particle sizes range from 10-100 microns with resultant splats of the order of a
few microns in thickness and 100-150 microns in diameter. Under typical APS
conditions, most particles solidify independently, resulting in a chaotic assemblage of
the deposited microstructure that consists of splat gaps and porosity between unfilled
regions. Figure 2.10 provides an illustration of APS process, and also microstructure
images of source material (feedstock powder), top view of a single splat, and polished
cross section of an assembly of many splats [92].
As expected, the characteristics of the deposited microstructure are strongly dependent
on processing [102]. Parameters of critical importance include the characteristics of the
spray stream (particle trajectory and thermal and kinetic state); the location and state of
the substrate, including substrate roughness, temperature, position, geometry and
relative movement; and speed of torch and part. Both attributes of the feedstock powder
as well as the spray device are of significance. All of these processing conditions govern
the microstructural nature of the deposit build-up. With respect to TBC system based on
YSZ polymorph, the critical microstructure elements include splat interface, intra-splat
cracking due to the relief of very large quenching stresses upon impact and
solidification, and finally incompletely filled layers which result in highly varied
porosity [103]. Much progress has been achieved in the ability to characterise the
interplay between processing and microstructure. Particularly there has been specific
progress in macroscopic quantification of the structure-property relationship (e.g.
porosity-thermal conductivity relations) and underlying fundamentals in terms of the
generation of defect types and their characteristics [104]. This is significant for
monitoring and controlling the coating compliance, which in recent years has been
attributed to the unique nonlinear/anelastic response of the porous ceramic coating on
metallic substrate [9, 105]. These developments have not only enhanced the
applicability of YSZ TBC systems, but have also paved the way for expanded
opportunities for other oxide systems.
CHAPTER 2 LITERATURE REVIEW
PAGE 46
In summary, APS processing is a mature and cost-effective method to deposit TBC
ceramic coatings which have the characteristics of lamellar and splat structure imparting
the superior thermal insulation. However, APS TBCs are fabricated by small to large
factories all over the world to specifications defined by the engine manufacturer. As
these spray factories are entitled to use any available (approved) spray device and
powder, the specification tend to be rather wide to accommodate microstructure
variation. This has, to some extent, resulted in design engineers reducing the reliance on
the TBC coating for reliable performance and focusing more on life extension of the
underlying superalloy substrates [92]. Namely, the coatings are not “prime reliant” yet.
However, as the increase in turbine entry temperature is demanded for higher engine
efficiency, the need for reliable and controlled coatings is becoming increasing critical.
As such, advancements in processing science have been sought to develop
methodologies not only for process control, but also effective material properties for
microstructure control.
CHAPTER 2 LITERATURE REVIEW
PAGE 47
Figure 2.10 Schematic of atmospheric plasma spray deposition process for thermal
barrier coatings in which a ceramic feedstock is carried to the substrate by a plasma
spray. Scanning electron micrographs show (a) the feedstock powder, (b) an individual
component of the coating assembly (splat), and (c) a polished cross section of an
aggregated coating. [92]
2.4.3 Alternative processing technologies
Both APS and EBPVD coatings, which are now used in main stream TBC
manufacturing, offer benefits in terms of both compliance and thermal conductivity due
to their defective microstructure. Typical APS coatings offer lower conductivity while
EBPVD coating provide better strain compliance and erosion resistance due to their
different microstructure characteristics. As outline earlier, there is significant interest to
combine the benefits of both APS and EBPVD TBC, which has led to both industrial
and academic research on alternative processing technologies such as segmented or
vertically cracked ceramic coatings and their successful implementation in gas turbine
engines [95, 96, 106, 107].
Plasma sprayed zirconia splats undergo extremely rapid quenching (108 K·s
-1) as they
impact a substrate or the surface of already deposited layers, spread into sheets of a few
CHAPTER 2 LITERATURE REVIEW
PAGE 48
micron thickness, and are cooled down predominantly by thermal conduction to the
underlying solid body [92]. As a result, an intricate net work of micro-cracks can be
seen in a single splat formed on a flat metal substrate (Figure 2.10b). The patterns and
principles behind the formation of these cracks are similar to mud cracking or crazing in
porcelains, where the shrinkage of the surface layer is restricted by the underlying body,
resulting in large lateral tensile stress upon cooling. Since the coefficient of thermal
expansion for YSZ is approximately 10-5
K-1
, and the underlying substrate almost does
not go through temperature change while the temperature drop ΔT that plasma sprayed
splats undergo is more than 2000 K, the tensile strain generated is about 2×10-2
, which
is an order of magnitude larger than the failure strain of most ceramics including
zirconia. As a result, APS coatings usually contain a large number of micro-cracks.
When the substrate/deposition temperature is raised significantly, the temperature
change for the deposited coatings decreases and hence less micro-cracks in splats.
However, large macroscopic cracking (i.e., so-called segmented cracks or dense vertical
cracks) tend to form. Figure 2.11 shows the microstructures of a conventional
non-segmented APS coating and a segmented dense vertically crack (DVC) coating [92].
It has been shown that the density of segmented cracks increases at the expense of
micro-pores as the substrate/deposition temperature increases [95]. Significant
improvement in terms of thermal cycle life has been reported for such vertically cracked
TBCs, and this technology has been used in advanced engines for more than 15 years
[106, 107].
In summary, the segmented or dense vertically crack processing provides the plasma
sprayed TBC coatings with the benefits of vertical crack similar to the columnar
structure in EBPVD coatings. However, several fundamental issues remain, such as
what optimal crack density is needed to achieve optimum performance in TBC.
Furthermore, thermal conductivity is to some extent sacrificed, as there are more
channels for heat transport through the vertical cracks compared to conventional
lamellar layered APS TBC. To compensate this, DVC coatings are typically sprayed to
a much greater thickness to impart similar heat resistance.
CHAPTER 2 LITERATURE REVIEW
PAGE 49
Figure 2.11 Images of the cross sections of (a) traditional (non-segmented) and (b)
highly segmented thermal barrier coatings produced at low and high substrate
temperature, respectively, with enlarged scanning electron micrographs obtained from
fractured cross sections of the coatings. [92]
Besides the segmented cracked TBC via advanced APS processing, there are other
emerging technologies such as suspension and solution plasma spraying (SPS) which
combines the advantages of traditional APS, segmented crack structures, and EBPVD
attributes, and overcomes some limitations of the segmented crack TBCs [108, 109].
Plasma spray physical vapour deposition (PSPVD) is a novel extension to vacuum
plasma spray technology allowing for vapour generation within the thermal plasma
followed by deposition onto a hot substrate in a way similar to PVD but at higher rates
[98, 110]. PSPVD has the advantages of coating complex shaped objects at a relatively
higher rate and is being investigated by industry for potential commercialisation.
To summarise, the widespread use of thermal barrier coating in propulsion and power
generation industry has advanced ceramic deposition technologies, such as plasma
spraying, EBPVD and emerging processing methods. As advanced TBCs are now
CHAPTER 2 LITERATURE REVIEW
PAGE 50
expected to be prime reliant (i.e., they are guaranteed not to fail), future development of
TBCs will involve interplay among processing science and technologies, microstructure
control and its association with coating performance. Research and innovations will
continue in all areas driven by the need for more reliable TBC system used in gas
turbine engines.
2.5 Failure phenomena
Thermal barrier coatings (TBCs) are very complex multi-functional coating systems not
only because TBCs can be made of various compositions (more than 15 elements) and
microstructures (several phases), but also TBCs deposited by different processes such as
APS and EBPVD are quite different, furthermore, TBCs evolves with thermal exposure
during service, also depending on thermal history (isothermal/thermal cycling). In
addition, the engine environment can strongly influence TBCs durability and
performance as well.
As a result, one of the chronic problems is that the life of present TBC coatings
invariably shows a wide distribution, with a high proportion clustered about a median
value but with a significant proportion failing at much earlier times [14]. As the turbine
entry temperature is required to increase and TBCs are moving towards prime reliant, it
is essential to understand the failure phenomena and use it as guidance for future
improvement and new coating designs.
Although there are various failure modes observed in TBCs and different mechanisms
proposed, generally, the failure phenomena can be classified into two categories, one is
caused by thermally-activated issues including oxidation-induced TGO growth, stress
redistribution in the vicinity of imperfection and bond coat deformation, namely,
intrinsic mechanisms, and the other results from foreign object damage and
environmental degradation, i.e., extrinsic mechanisms. The section will mainly review
the former with the latter one mentioned briefly.
CHAPTER 2 LITERATURE REVIEW
PAGE 51
2.5.1 General principles in TBCs failure
There is substantial circumstantial evidence to suggest that many of the TBC failures
are associated with bond coat oxidation [14, 111]. Indeed, one oxidation criterion for
predicting average TBC life is using a critical thickness of TGO as TGO growth is
driven by oxidation. Another, embodied in the Coatlife software, is an aluminium
depletion criterion based on the combined time and temperature for the concentration of
aluminium at the bond coat surface to fall below a critical value. In the case of MCrAlY
bond coats, the rationale for this is that when the Al concentration falls below a certain
amount, alumina is no longer the thermodynamic preferred phase and other oxides such
as spinels may form [112]. These oxides do not form such a protective scale, and
consequently the alloy oxidises faster.
Although related to the oxidation behaviour of the bond coat, neither the concept of a
critical thickness nor aluminium depletion can account for the wide distribution in
coating lives, especially under thermal cycling conditions. Indeed, in the majority of
materials examined after failure above about 1000 °C, the aluminium concentration,
albeit depleted somewhat, has not fallen to the critical value [14]. Meanwhile, the
short-lived coatings have failed before the TGO thickness has reached the thickness of
their counterparts that have shown the longest lives. Together these findings indicate
that there must be other factors that govern the failure behaviours.
The prevailing mode of failure is that a part of the coating buckles and spalls away from
the underlying substrate, typically on cooling down to room temperature [9, 113]. A
typical buckling failure, in this case nucleated from the edge of a test coupon, is shown
in Figure 2.12 [14]. Such buckling and subsequent spallation is a common mode of
failure for all films and coatings under compression that consists of thermal expansion
misfit stress between the ceramic coating and underlying substrate upon cooling and
TGO growth stress generated at high temperature. The mechanics of the failure by
buckling of a thin, elastically isotropic film under compression from a flat surface is
well understood [114], provided an unbounded region of a critical size, dc, exists at the
interface (Figure 2.13). For a fixed film thickness and residual stress, the stress at which
buckling will occur is given by the relationship:
CHAPTER 2 LITERATURE REVIEW
PAGE 52
σ/E = 4.8(h
dc)2 (1)
Figure 2.12 (a) Incipient buckling of a TBC coating viewed under reflected light. (b)
The surface revealed by spallation of the TBC consists of a mixture of local failure
between the TGO and the bond coat (appearing dark) and in the TBC itself (light
regions). [14]
The striking feature of this relation is that the flaw size depends linearly on the
thickness of the film. Since the TBC is over 100 µm thick, the critical size to which an
interface flaw must grow before buckling occurs should be several millimetres [14]. As
interface separations of this large size are not usually present, one of the major disputed
questions is how these interface separation nucleate and then grow to such a large size.
Such progressive failure consisting of nucleation of local interface separation and their
subsequent growth has indeed been observed [115]. Mechanics calculations have shown
that interface perturbations from flatness can decrease the critical size at which buckling
can initiate, then grow, to form a spallation [116]. Nevertheless, localised flaws must
first initiate and then grow for failure to occur. Therefore, understanding the nucleation,
growth and subsequent linkage of these flaws is essential before realistic failure models
can be developed.
CHAPTER 2 LITERATURE REVIEW
PAGE 53
Figure 2.13 Schematic illustration of the buckling of a compressed film with a
pre-existing flaw of diameter dc.
In summary, the strain energy in the TGO provides the driving force for TBCs failure,
and cracks (flaws) are needed to nucleate, propagate and coalesce before they become
large enough to trigger spallation. Evans etc. [9, 83] have proposed the specific
mechanisms about how the strain energy and imperfections in the vicinity of TGO
govern the TBCs durability which will be described below.
2.5.2 Specific mechanisms of imperfections governing
The specific ways in which the cracks nucleate and grow relate to the increase in the
severity of the imperfections as the system is exposed and cycled [9]. While this occurs
in many ways, all are ultimately linked to the magnitude and scale of tensile σzz stresses
that amplify as either the TGO thickens or the imperfection increase in size, or both. In
turn, the stresses translate into stress intensity factors acting on cracks that nucleate and
propagate around the imperfections [83]. The formation of tensile σzz results from stress
redistribution. The subsequent questions come, what are these imperfections and how
do they form? From microstructure observation, there are mainly two types of
imperfections including interface undulations and thickness heterogeneities (Figure
CHAPTER 2 LITERATURE REVIEW
PAGE 54
2.14). Direct 3D observations of these imperfections by novel X-ray computed
tomography technique are obtained and shown in chapter 3.
CHAPTER 2 LITERATURE REVIEW
PAGE 55
Figure 2.14 (a) A schematic of two major categories of TGO imperfection that govern
the TBC failure sequence; (b) a thickness imperfection in a TGO grown on a
NiCoCrAlY bond coat; (c) an undulation imperfection that develops in a Pt-aluminide
system upon thermal cycling. [9]
CHAPTER 2 LITERATURE REVIEW
PAGE 56
TGO thickness imperfections form and enlarge in regions where the O2-
diffusivity
through the TGO is exceptionally large. This happens at locations where the TGO
contains oxides other than α-Al2O3 having intrinsically lower resistance to O2-
diffusion.
Examples comprise TGOs that entrain Y from the bond coat to form yttrium aluminates
and become locally thick (Figure 2.14b) [117]. Above a critical size, the tensile stresses
around these imperfection are predicted to nucleate interfacial separations [116].
In terms of the interface undulation (or so-called rumpling/roughening), it is more
disputable. However, at least two mechanisms have been identified that can lead to such
rumpling [14]. The rumpling has been attributed to a “ratcheting” phenomenon
motivated by the lateral compressive stress in the growing TGO and facilitated by
thermal cycling [9]. TGO thickens during oxidation, accompanied with compressive
stresses. As TGO is attached to the bond coat, the only way in which it can decrease its
elastic strain energy is by undulation. In this way, its length increases and it remains
attached to the alloy. This undulation requires the alloy to deform to accommodate the
undulation, and the oxide must also deform concurrently. According to the ratcheting
mechanism, this accommodation is by plastic deformation of both the TGO and bond
coat during thermal cycling. As the lateral growth of the thickening oxide continues
during the high-temperature portion of the thermal cycles, it continues to generate
compressive stress that is relaxed by ratcheting during the thermal cycle so the process
is ongoing. Many of the essential features of the mechanism have been substantiated by
finite element computations [118] and are consistent with observations of the increase
in TGO length as the surfaces roughen.
Another mechanism shown to cause roughening is the surface displacement associated
with volumetric changes in the bond coat as aluminium depletion occurs [14]. This
roughening is illustrated in Figure 2.15, together with etched cross sections revealing
the presence of both γ‟ and β phases in the bond coat [119]. After aluminising and YSZ
deposition, the platinum-modified aluminide (PtNiAl) bond coat is chemically
homogeneous and has the β-NiAl crystal structure. After high temperature exposure, the
initially flat bond coat roughens and etching reveals that the bond coat has partially
transformed to γ‟-Ni3Al. In addition, the remaining β-NiAl phase regions often have the
martensitic characteristic structure. These observations are the results of aluminium
depletion in the bond coat and concurrent enrichment of nickel diffused from the
CHAPTER 2 LITERATURE REVIEW
PAGE 57
underlying substrate. As aluminium is being consumed, the average composition of the
bond coat becomes increasingly enriched in nickel until formation of γ‟-Ni3Al. At even
later times, the composition can extend to γ-Ni phase.
Figure 2.15 Microstructure of an initially flat aluminide bond coat after 50×1h cycles at
1200 °C: (a) surface rumpling; (b) cross section showing a rather uniform oxide layer
and strong surface undulations (γ‟ phase is revealed by etching); (c, d) optical
micrographs showing etched cross section before and after cyclic oxidation. Dark areas
on the optical images correspond to β phase while γ‟ phase appears white. [119]
When these imperfections mentioned above exist in the coatings, the stresses in the
vicinity of them can deviate from average values and also be redistributed by creep or
yielding of bond coat. The local stresses, especially the tensile σzz stresses normal to the
TBC surface initiate cracks along directions having lowest toughness. Local stresses in
undulated TGO are measured and discussed in chapter 4. In summary, the basic
principles that govern TBCs failure are as follows [9]:
1. The TGO experiences large in-plane compressions, especially upon cooling. It
attempts to alleviate the stress (associated with strain energy) by lengthening itself,
through out-of-plane displacements. This can occur by buckling as well as by
visco-plastic deformation of the bond coat. These displacements induce tensile σzz
stresses normal to the interface that motivate delamination mechanisms.
CHAPTER 2 LITERATURE REVIEW
PAGE 58
2. When imperfections exist (or are developed) around the TGO, tensions are induced
normal to the TGO/bond coat interface, as well as in the TBC, that nucleate and
grow cracks in this vicinity. The propagation of these cracks leads to the coalescence
of the cracks, and eventually resulting in failure.
3. The TBC, despite its compliance, has sufficient stiffness to suppress small scale
buckling of the TGO. Accordingly, eventual failure often occurs by large scale
buckling which requires a sufficiently large separation developing near the interface,
typically several millimetres in diameter. The durability of the TBC is governed by
the time/cycles needed to develop such separation: through a nucleation,
propagation and coalescence sequence, involving the energy density in the TGO, as
well as the size and spacing of the prominent imperfections.
Another issue worth to be mentioned is that TBCs made of nominally the same
superalloy, with the same bond coat and the same YSZ coating, all produced by the
same manufacturer in the same process manner can have various failure behaviours,
which suggests that even small, but as yet unidentified, concentrations of dopants can
have a large effect on life [120].
2.5.3 Foreign attack and environmental degradation
Higher engine temperatures are also creating new materials issues in ceramic topcoats,
namely the degradation of YSZ TBCs due to the molten silicate deposits [16-19],
formed by the ingestion of fine particulates from the environments (sand [18], volcanic
ash [121-123]). Because of the major components in the silicate glass formed, this
phenomenon is commonly referred to as CMAS (calcium-magnesium-alumino- silicate)
attack. This primarily affects high performance jet engines on account of their higher
maximum temperatures and electricity generation engines, but it is likely to affect more
engines as operation temperature are increased in pursuit of greater engine efficiency [3].
In the case of land-based electricity-generation engines, it is not always practical to
filter out the finest particles that can be carried along with the input air and from
alternative fuels such as synthesis gas [124, 125]. It appears that the wetting of TBCs by
the molten CMAS glass, and dissolution/reprecipitation of YSZ grains in that glass,
contribute to the CMAS attack of TBCs [19, 20]. This manifests itself as continued
penetration of the CMAS glass into the TBC and affects both APS and EBPVD TBCs
CHAPTER 2 LITERATURE REVIEW
PAGE 59
alike. For example, Figure 2.16 shows complete penetration of EBPVD YSZ TBC by
molten CMAS in a laboratory test [3]. Therefore, being able to mitigate CMAS attacks
becomes an additional critical requirement for future TBCs.
Figure 2.16 (a) Cross section scanning electron micrograph of 7YSZ EBPVD TBC fully
penetrated by a model calcium-magnesium-alumino-silicate (CMAS) melt in a
laboratory experiment. Crystalline phases with different compositions from the parent
7YSZ material (lighter gray) are noted (a) at the interface between the coating and the
melt. (b) The corresponding Si elemental map showing the extensive CMAS penetration.
[3]
In summary, TBCs failure mechanisms include the intrinsic ones resulting from
oxidation-induced TGO growth, imperfection development, stress redistribution and
bond coat deformation etc, and extrinsic ones from foreign object damage and
environmental degradation by molten deposits. Together all these are illustrated in
Figure 2.17.
CHAPTER 2 LITERATURE REVIEW
PAGE 60
Figure 2.17 Extended menu of failure mechanisms typical of current thermal barrier
coatings (TBCs) [11]. Three general modes of CMAS damage (lower right),
characteristic of higher temperature operation, have been identified so far. One involves
delamination cracks propagating through the TBC, another leads to chemical attack of
the thermally grown oxide (TGO) with concomitant loss of adherence, and a third result
from creep cavitation of the bond coat below a heavily penetrated TBC. [123]
2.6 Summary
Thermal barrier coatings (TBCs) are widely used in gas turbine engines in propulsion
and power generation industries for improved engine efficiencies. The development
history and constituents of the complex multi-layer evolving coating system have been
reviewed in the chapter, along with the processing sciences and technologies. Failure
mechanisms identified to date are also briefly described. The promise of even higher
efficiencies and other benefits is driving TBCs research worldwide, and the continuous
development of these fascinating systems provides rich opportunities for materials
research community.
CHAPTER 3 INTERFACIAL PROPERTIES OF APS TBCs
PAGE 61
Chapter 3
Investigation of interfacial properties of atmospheric plasma
sprayed thermal barrier coatings with four-point bending and
computed tomography technique
3.1 Introduction
Air plasma sprayed thermal barrier coatings (APS TBCs) have been developed for
advanced gas turbine and diesel engine applications to improve engine reliability and
efficiency [9, 14, 126]. For aeroengines, these materials include a metallic bond coat to
improve bonding and oxidation resistance applied on a nickel superalloy substrate, on
top of which is a ceramic thermal barrier coating providing the necessary thermal
insulation. The state of the art TBC typically comprises 6-8 wt.% yttria stabilised
zirconia (YSZ) and the bond coat is made of MCrAlY where M stands for Ni, Co or
both for APS TBCs. During high temperature exposure, a thermally grown oxide (TGO)
forms between the TBC and bond coat. Upon cooling, large compressive stresses
develop in the ceramic layers due to the thermal mismatch between them and the metal
substrate. APS TBCs usually fail within the TBC near the TBC/bond coat interface,
with local segments entering the TGO as well as the interface [127-131]. In contrast,
TBCs produced by electron beam physical vapour deposition (EBPVD) typically fail at
the TGO/bond coat interface [9, 13, 132]. Generally, it is believed that the failure of
TBCs is driven by the strain energy in the ceramic layers and is resisted by the
interfacial toughness. Therefore, to take full advantage of the potential of TBCs and
further understand the failure mechanism, it is important to evaluate the interfacial
toughness of the TBCs.
CHAPTER 3 INTERFACIAL PROPERTIES OF APS TBCs
PAGE 62
There are a number of methods available to determine the adhesion in coating/substrate
systems, such as compressive tests [132, 133] , tensile tests [127, 134], indentation tests
[117, 127, 135] and bending tests [136-141]. Although Zhao et al. [132] have recently
employed a compressive test to determine the interfacial toughness of EBPVD TBCs,
this method requires buckling of the coating on the substrate. Our experiment showed
that a compressive test did not generate buckling for APS TBCs before large scale
delamination of the coating from the substrate. In the case of tensile tests, it is difficult
to detect the critical point when debonding of the interface occurs from the
load-displacement curve. Indentation tests usually produce cracks which deflect into the
coating and a wide scatter in the data exists, especially for APS TBCs which usually
have a rough TBC/bond coat interface. Hofinger et al. [138] modified the four-point
bend test proposed by Charalambides et al. [142] to evaluate the interfacial fracture
energy of the plasma sprayed ZrO2 coatings on flame sprayed high alloyed steel
substrates. This involved bonding a stiffener on top of the TBC in order to suppress the
cracks in the coatings normal to the coating/substrate interface thereby increasing the
stored energy in the coating which is the driving force for delamination. This method
has the advantage of creating a stable crack growth under loading condition, which
enables the interfacial toughness to be evaluated without precise measurement of crack
length. Therefore the method was adopted in this work.
X-ray computed tomography (CT) allows the study of materials microstructures
non-destructively [18]. In principle, it can also provide quantitative information such as
coating density, pore size distribution, surface roughness as well as enabling the
structural monitoring of in-situ tests. In this study micro X-ray CT was deployed to
observe the evolution of the microstructure in TBCs during thermal treatments.
The aim of the work is to evaluate the interfacial toughness of an APS TBCs system
using a modified four-point bend test. This has been combined with micro CT, along
with other techniques such as SEM, XRD and indentation, to study the relationship
between the microstructure and mechanical properties in the TBCs as a function of
aging by thermal exposure.
CHAPTER 3 INTERFACIAL PROPERTIES OF APS TBCs
PAGE 63
3.2 Experiments
3.2.1 Materials
The atmospheric plasma sprayed (APS) TBCs investigated in this study were produced
using agglomerated and sintered (A&S) yttria stabilised zirconia powder (Amperit
827.873, HC Starck). The thickness of the ceramic coating is ~200 µm. A NiCoCrAlY
(Amdry 365-2, Sulzer Metco, USA) bond coat with a thickness of ~150 µm was applied
by APS on a Ni-based superalloy (IN718) substrate of 5 mm in thickness, 30 mm in
width and 50 mm in length. Both the top coat and the bond coat were sprayed using a
Sulzer Metco F4 plasma gun and Ar/H2 mixture as plasma gas; the gun operating
powers were 45 kW and 42 kW, respectively. Prior to spraying, the bond coat and the
substrate surfaces were degreased and grit blasted with grit 60 alumina using an air
pressure of 5 bars.
3.2.2 Preparation of the four-point bending samples
The samples were first exposed at 1150 °C for various times using heating and cooling
rates of 3 K/min. A temperature of 1150 °C was used to accelerate the experiment and
to simulate the flame test temperature. After thermal exposure, samples were sliced into
the required geometry as shown in Figure 3.1. A stiffener, which is identical to the
substrate, was bonded on top of the TBC using adhesive (Araldite, precision 2011) after
the thermal treatment of the TBC samples. The thickness of the adhesive is around 100
µm and the infiltration of the adhesive to the TBC is low so that its effect on the
measurement of interfacial toughness is negligible [140]. A notch of about 0.4 mm in
width was made by hand using a diamond cutting blade at the centre of the bending
sample. The notch was observed under an optical microscope until it reached the
TBC/bond coat interface to ensure the TBC was not damaged. A four-point bending
device was used at room temperature on an Instron 5569 mechanical testing machine
with a constant crosshead speed of 0.2 mm/minute. Pre-cracks were introduced by
loading during which the load-displacement curve was monitored. As soon as the slope
deviated from linearity, the machine was stopped and pre-cracks were created unstably.
CHAPTER 3 INTERFACIAL PROPERTIES OF APS TBCs
PAGE 64
Figure 3.1 A schematic of the four-point bend test (a) [143] in a typical experiment (b)
[144].
3.2.3 Experimental procedures
The phases of the TBC were identified using X-ray diffraction (XRD) with Cu-K
radiation at 40 mA and 50 kV (Philips, PW1830). Step scans of 0.05° over the 5° to 85°
2 range were measured. From the integrated intensities of the monoclinic and
tetragonal phase of X-ray diffraction peaks, Im and It, the volume fraction of the
monoclinic phase, fm, can be inferred [145]:
m
mm
X
Xf
3 1 1.01
3 1 1.1
(1)
where Xm is the integrated intensity ratio expressed by :
CHAPTER 3 INTERFACIAL PROPERTIES OF APS TBCs
PAGE 65
)101()111()111(
)111()111(
tmm
mm
mIII
IIX
(2)
The Young‟s modulus and hardness of the TBC, bond coat and substrate were
determined by micro-indentation. The cross-sections of the samples had been ground
and polished following a standard metallurgical procedure finishing with 1m diamond
paste. For each indentation test, it took 30 seconds to reach the maximum load (3N)
before pausing for 10 seconds and then unloading over another 30 seconds. The
modulus was evaluated from the unloading curve. At least 20 indentations were
undertaken for each condition.
The microstructures of the TBCs were examined by scanning electron microscopy
(SEM) using a Philips XL30. Two APS TBCs specimens were prepared and then
scanned by an Xradia micro CT (in Manchester X-Ray Imaging Facility, UK) prior to
and after thermal exposure. One sample was machined to the size of ~0.8×1×6 mm and
the other was carefully ground manually to a tip shape. The thermal treatment
conditions and tomography settings are shown in Table 3.1. During a scan, Xradia low
energy filter 2 was applied, over a thousand 2D radiographs were taken over a rotation
of 183 degrees. These were reconstructed to form a 3D virtual volume using filtered
back projection and then analysed using commercial software (Avizo).
Table 3.1 Thermal treatment conditions and x-ray micro CT settings for the samples.
CHAPTER 3 INTERFACIAL PROPERTIES OF APS TBCs
PAGE 66
Figure 3.2 A schematic of the Xradia X-ray computed tomography arrangement.
3.3 Results
3.3.1 TBC phase fractions
The XRD pattern in Figure 3.3a shows that after thermal exposure the TBC mainly
consists of tetragonal (t) phase and a small amount of monoclinic (m) phase. The )101(t ,
)111(m and )111(m peaks in a range of 2 between 27o and 33
o were used to obtain the
volume fraction of the m phase using Equation 1 and 2 after deconvolution. The inset in
Figure 3.3b gives the relationship between the volume fraction of the m phase and the
exposure time. The content of m phase increases as a function of thermal exposure time
but remains below 4% even after exposure for 200 hours.
CHAPTER 3 INTERFACIAL PROPERTIES OF APS TBCs
PAGE 67
Figure 3.3 (a) XRD patterns of the TBC exposed at 1150 ºC for 0, 10, 50, 100 and 200h
in the 20-90º 2 range. (b) XRD patterns in the 27-33º 2 range and the volume fraction
of the monoclinic phase as a function of thermal exposure time (inset).
3.3.2 Four-point bending test
Four-point bending tests were conducted to measure the adhesion between the TBC and
bond coat in the TBCs system. With increasing bending moment the strain energy in the
TBC increases. When the external applied energy exceeds the fracture toughness stable
CHAPTER 3 INTERFACIAL PROPERTIES OF APS TBCs
PAGE 68
crack starts to grow under the loading conditions. Figure 3.4 shows a typical
load-displacement curve of the sample in this experiment. The slope of the straight line
at the beginning is related to the stiffness of the sample. The load continues to increase
until reaching the peak and then drops to PC and starts to increase again gradually. The
peak load indicates the point where some minor cracks occur inside the sample and the
critical value PC at the given pre-crack length corresponds to the point where stable
crack propagation happens. This method takes the advantage of the constant bending
momentum in the four-point bending setting to determine the energy release rate for
stable crack propagation at the TBCs interface. Both the critical load and geometrical
dimensions will be used in the calculation of the fracture toughness in Section 3.4.2.
Figure 3.4 A typical load-displacement curve of the four-point bend test.
3.3.3 Microstructure observation
Figure 3.5 shows SEM images of fractured TBCs after the four-point bending test. In all
cases, the crack has propagated mainly in the TBC just above the interface between the
TBC and bond coat, even after thermal exposure for 200 hours. Figure 3.6 displays the
microstructure and TGO growth in the TBCs as a function of thermal exposure time
studied by X-ray micro CT, which shows that imperfections develop at the interface
including undulation (location A); regions in the TBC that exhibit local TGO growth
(location B) and domains of TGO formation in the bond coat beneath the interface
(location C) without the introduction of the damage from cutting and grinding. It is
CHAPTER 3 INTERFACIAL PROPERTIES OF APS TBCs
PAGE 69
expected that local stresses will redistribute around these imperfections and help to
nucleate and propagate cracks during cooling which will be discussed in Section 3.4.3.
Figure 3.5 SEM images of cross-sections at the interface between the TBC and bond
coat after four-point bend test in the (a) as-sprayed condition and after (b) 10h, (c) 100h
and (d) 200h of thermal exposure at 1150 °C, showing that the cracks propagate above
the TGO within the top coat. [144]
Figure 3.6 Equivalent micro-tomography slices of approximately the same region taken
from 3D images of the microstructure of a APS TBCs sample exposed at 1150 ºC for 0
(a), 20 (b), 120 hours (c), indicating various kinds of imperfections developed near the
YSZ/BC interface.
CHAPTER 3 INTERFACIAL PROPERTIES OF APS TBCs
PAGE 70
3.3.4 Determination of the mechanical properties of TBCs
Figure 3.7 gives the Young‟s modulus and hardness of the components that constitute
the present TBC system as recorded by micro-indentation on the TBC cross-sections.
Due to the sintering, the Young‟s modulus of the TBC coating increases with the
thermal exposure time. It should be noted that the Young‟s modulus of the TBC may be
overestimated by the indentation technique, because it only measures a single splat
rather than a sufficient amount of splats which usually include pores and micro-cracks.
In addition, the elastic modulus can increase dramatically in the presence of
compressive stress [146]. An accurate determination of the aggregated mechanical
properties of the TBC is often difficult. However, one advantage of the modified
four-point bending method is that there is no need to acquire an accurate value of the
modulus of the coating as the stiffener is much thicker than the coating. This means that
the elastic modulus of the coating has a negligible effect on the strain energy in the
model calculation. On the other hand, the Young‟s moduli of the bond coat and
substrate are almost constant ~200 GPa regardless of the thermal exposure time. The
hardness of the TBC increases while that of the bond coat decreases to become closer to
the value of the substrate with increasing thermal exposure time. This is attributed to the
inter-diffusion between the bond coat and substrate during which Al diffuses towards
the substrate and concurrently Ni diffuses into the bond coat.
Figure 3.7 (a) Young‟s modulus and (b) hardness of the components in the TBCs as a
function of thermal exposure time.
CHAPTER 3 INTERFACIAL PROPERTIES OF APS TBCs
PAGE 71
3.4 Discussion
3.4.1 Analytical considerations
The specimen is loaded in the four-point bending as shown in Figure 3.1. In this case,
the total energy release rate, G which comprises the contributions from residual stress in
the coating and those generated by the application of an external mechanical load, can
be given by [147],
rprP GGGG (3)
The first term in Equation 3, GP, is the energy release rate applied by the external load
derived from linear elastic fracture mechanics. Gr is the energy release rate for the
relaxation of the residual stress. Gpr stands for the interaction between the external load
and the residual stress distribution.
A schematic of the model is shown in Figure 3.8 where the top coat and stiffener are
taken as material 1, and the substrate and bond coat as material 2. The energy release
rate GP can be derived by two methods.
Method 1: The interfacial fracture energy can be obtained according to Suo-Hutchinson
analysis on interface crack between two elastic layers [148]. The energy release rate is
given in Appendix A [148]:
sin2
16 23
22
1
hAI
PM
Ih
M
Ah
PcG (4)
The complex stress intensity factor K is written as:
Gcc
K21
cosh4
(5)
Therefore, the energy release rate and stress intensity factor for the bend testing can be
obtained from Equation 4 and Equation 5.
Method 2: The energy release rate GP can be calculated according to Hofinger et al.
[138],
cIIE
MG
11
2
)1(
22
2
2
2
0 (6)
CHAPTER 3 INTERFACIAL PROPERTIES OF APS TBCs
PAGE 72
with the constant bending moment
b
LLPM
4
)'(00
(7)
and the second moments of inertia
12
3
22
hI
)(4
)2(
333 12
2
1
22
1
2
22
1
2
1
33
1
3
2
d
dd
dd
d
chhh
hhhhhhhhh
hhhI
(8)
with
)1(
)1(2
2
2
2
d
d
E
E
)1(
)1(2
12
2
21
E
E (9)
where the parameter h represents the layer thickness. E and ν denote the Young‟s modulus
and Poisson‟s ratio. The subscripts 1, 2, and d refer to the ceramic coating, substrate and
stiffener, respectively. Based on the Equations 6-9 the energy release rate can be
calculated. The two methods above will be adopted to estimate the interfacial toughness
of the APS TBCs.
Figure 3.8 A schematic of the interface cracking model. [144]
3.4.2 Estimation of the interfacial toughness
In the above sections, the solutions and parameters required for the calculation of the
interfacial toughness have been outlined. This part will be devoted to an estimation of
the interfacial toughness as well as the effect of thermal exposure time on it. Method 1
CHAPTER 3 INTERFACIAL PROPERTIES OF APS TBCs
PAGE 73
refers to a single-layer system. However, in this work, a stiffener is attached to the TBC
top coat. The bilayer can be approximated as a single layer according to Vasinonta and
Beuth [149] with the effective Young‟s modulus given by:
d
dd
hh
hEhEE
1
11 (10)
where subscripts E is the effective Young‟s modulus, E1, Ed denote the modulus of TBC
and the stiffener and h1, hd refer to the thickness of TBC and the stiffener, respectively.
A four-point bending device with inner span of L=20mm and outer span of L’=40mm is
used. Dimensions of the samples, h2, hd, and b are measured from experiments. The
thickness of TBC h1 is about 0.2 mm. The bond coat has similar elastic properties to the
substrate, so it is regarded as a part of the substrate. The Poisson‟s ratio of the TBC and
substrate are taken as 0.2 and 0.3, respectively and the elastic modulus of the TBC is
obtained from indentation. Young‟s moduli of the substrate and stiffener E2, Ed are the
same ~200 GPa. The critical load is determined from the load-displacement curve by
the four-point bend test. Substituting these parameters (Table 3.2) into the Equations
4-5 and Equations 6-9 produces the mixed mode energy release rate and stress intensity
factor summarised in Figure 3.9 and 3.10.
Figure 3.9 The energy release rate as a function of the thermal exposure time.
CHAPTER 3 INTERFACIAL PROPERTIES OF APS TBCs
PAGE 74
Figure 3.10 Stress intensity factor as a function of the thermal exposure time. The circle
and diamond marks indicate the stress intensity factor of mode I and mode II,
respectively. Dashed lines connect the average values.
The two methods give almost the same results. Figure 3.9 displays that the energy
release rate shows an increasing trend with the thermal exposure time initially and tends
to be stable after 200 hours exposed at 1150 ºC. The interfacial toughness is indeed
sensitive to the loading phase angle Ψ, i.e. the ratio between mode II (Ψ=90º) and mode
I (Ψ=0º) fracture [147]. Since the edge and buckle delaminations are controlled by the
mode II toughness, it is essential to have the knowledge of the phase angel associated
with toughness [150]. The calculated phase angles (Appendix A) depending on the
loading geometry and elastic properties of the materials are about 42º for this
experiment which implies heavily mixed mode loading is applied at the crack tip.
CHAPTER 3 INTERFACIAL PROPERTIES OF APS TBCs
PAGE 75
Table 3.2 The experimental data used to calculate the energy release rate and the values
obtained by the two methods.
In this work the delamination mainly occurs within the YSZ layer, it seems that the
toughness of YSZ is weaker than the interfacial toughness between YSZ and the bond
coat up to this point. It is known that clean metal/oxide interfaces devoid of reaction
products are inherently tough and ductile: with toughness exceeding 200 J/m-2
[151,
152]. However, a broad range of toughness has been cited in reports [139-141, 153,
154]. This may be due to the sintering and phase transformation in the TBC,
composition and structure evolution in the bond coat, and contamination or segregations
at interfaces. Besides, different testing methods often lead to various measured values of
the interfacial toughness of YSZ TBCs which is also dependent on deposition method,
thermal treatment history, fracture mode, etc. For example, Kim et al. [153]
implemented a push-out method for EB-PVD TBC on a NiCoCrAlY bond coat and
CHAPTER 3 INTERFACIAL PROPERTIES OF APS TBCs
PAGE 76
reported that the interfacial toughness reduced from 115 J/m-2
to 15-20 J/m-2
when
oxidized at 1150 ºC from 10 to 100 h, while Thery et al. [139] adopted the modified
four-point bending test to study the EB-PVD TBC on a β-NiAl bond coat finding the
interfacial toughness to be 110 J/m-2
. However, for the APS TBC system, Yamazaki et
al. [140] used the modified four-point bending test and evaluated 130 J/m-2
for the
as-deposited specimen increasing to 250 J/m-2
after aging at 1000 ºC for 2000 h which
is generally in agreement with the current work. Zhou et al. [141] also investigated
as-deposited APS TBC with a NiCrAlY bond coat on stainless steel using both tensile
testing and four-point bend testing and found that the stress intensity factor was
0.67-0.94 MPa m1/2
for tensile test with mode II fracture and 1.0-1.27 MPa m1/2
for
four-point bending with mixed mode. The stress intensity factor increased to 4.26-7.21
MPa m1/2
just by increasing the gun operating power from 32.5 kW to 38.5 kW which
indicates the effect of deposition parameters on the adhesion of the coating/ substrate
system. Xu et al. [154] analysed a flame-sprayed TBC on a NiCrAl bond coat by shear
testing and an inverse finite element modelling which showed the mode II toughness
increased from 260 J/m-2
for as-deposited sample to 290 J/m-2
for sample annealed at
500 ºC. In general, the magnitude and variation of the interfacial toughness in this study
are in broad agreement with the reported data for APS TBCs.
It should be noted that the strain energy caused by the large residual stress in the TGO
was not included in the calculation above as the crack did not propagate through the
TGO. It can be expected that with further exposure the crack propagation will move to
the TBC/BC interface and penetrate the TGO [130]. The external applied energy release
rate GP will decrease as the strain energy in the TGO will increase with the increasing
TGO thickness when the energy in the TGO must contribute to the overall fracture
energy for motivating crack propagation. The added energy release rate can be
expressed as [13]:
TGO
TGO
TGO
TGO
h
EG
)1(2
(11)
where 1 , TGO refers to the residual stress in TGO. The introduction of is
because there is no mechanism capable of transmitting all of the strain energy in the
TGO into the delamination.
CHAPTER 3 INTERFACIAL PROPERTIES OF APS TBCs
PAGE 77
3.4.3 Further discussion
The energy release rate increases with the thermal exposure time because the crack
mainly propagates in the TBC itself just above the TBC/BC interface and the fracture
toughness of the top coat increases due to the sintering. The fracture toughness KIC of
the free-standing APS TBC film increases from 1 MPa m1/2
to around 2.5 MPa m1/2
after exposure at 1316 ºC for 500 h according to the work by Choi SR et al. [155]. The
tetragonal to monoclinic phase transformation detected by XRD may affect the property
measurements because of transformation-toughening in ceramics [156].
Since the fracture toughness of the TBC gradually increases with increase in thermal
exposure time, another mechanism must be responsible for the ultimate failure
behaviour after extended exposure. X-ray micro CT was employed to follow the
microstructural evolution at the buried TBC/bond coat interface at a specific location as
a function of thermal exposure time non-destructively. The images demonstrate all three
kinds of imperfections local to the interface reported by Rabiei and Evans [130]. Further,
the cracking of the TBC near the interface appears to be related to the TGO growth in
the vicinity of the imperfections [130]. Busso et al. [157] reported by finite element
analysis that considerable out-of-plane tensile stresses can develop adjacent to the
imperfections at the interface. These would help to nucleate and propagate cracks
during cooling to ambient temperature. A recent study by Limarga and Clarke [158]
shows that the tensile stress measured by Raman spectroscopy can reach 50 MPa in the
coating cooled from both front and back sides for PS TBCs after thermal cycling. Here
we have reconstructed the 3D interface between the TBC from the micro CT images as
shown in Figure 3.11. The 3D interfacial roughness was calculated to be 17.2 and 17.5
µm for the 1st sample prior to, and after, thermal exposure for 50 hrs at 1150 ºC,
respectively suggesting that the roughness of the interface did not change significantly.
This demonstrates the potential of micro-tomography images to show and, in future,
quantify the imperfections that develop sub-surface at the interface as well as the
interfacial morphology of TBC/bond coat systems during thermal treatments.
CHAPTER 3 INTERFACIAL PROPERTIES OF APS TBCs
PAGE 78
Figure 3.11 Reconstructed images of (a) the TBC and bond coat and (b) the morphology
of the interface on the TBC side extracted from (a). The interfacial information can then
be analysed such as the interfacial roughness, the amplitude and wavelength of the
surface etc. It is found that the interfacial roughness did not increase obviously after
oxidation.
CHAPTER 3 INTERFACIAL PROPERTIES OF APS TBCs
PAGE 79
3.5 Summary
1. A modified four-point bend test was successfully employed to quantify the interfacial
toughness of APS TBCs with a NiCoCrAlY bond coat systems after thermal exposure at
1150 ºC. Delamination was found to occur mainly within the TBC just above the
interface between the TBC and bond coat.
2. The calculated energy release rate was found to increase from ~50 J/m-2
after
deposition to ~120 J/m-2
after thermal exposure for 200 hours with a loading phase
angle about 42º. This is attributed to the sintering of the TBC. A small amount (below
4%) of tetragonal to monoclinic phase transformation may also play a role.
3. X-ray micro CT was used to observe the evolution of the microstructure at a specific
sub-surface location non-destructively. The 3D interface was reconstructed suggesting
that the interfacial roughness varied little (from 17.2 to 17.5 µm before and after
thermal exposure for 50 hrs at 1150 ºC). The tomography images show that different
kinds of imperfections developed sub-surface near the interface during the thermal
treatments.
3.6 Appendix
The energy release rate for an interfacial crack between two elastic layers can be
expressed according to Suo and Hutchinson as:
sin2
16 23
22
1
hAI
PM
Ih
M
Ah
PcG (A1)
where ic is the compliance parameter, i
i
ic
1 with ii 43 for plane strain.
is Poisson‟s ratio and the shear modulus with)1(2 i
i
i
E
and i=1,2, refer to the
coating and substrate.
Other non-dimensional parameters A , I , angle , , , and are given by:
)364(1
132
A )1(12
13
I
CHAPTER 3 INTERFACIAL PROPERTIES OF APS TBCs
PAGE 80
AI)1(6sin 2
1
1
H
h (A2)
)1()1(
)1()1(
12
12
)1()1(
)1()1(
12
12
where 2
1
, h and H are the thicknesses of the coating and substrate, whereas
subscripts 1 and 2 refer to the coating and substrate, respectively. The load per unit
thickness P and the moment per unit thickness M are given by:
bh
LLPCP
4
)'(02
b
LLPCM
4
)'(03
(A3)
with
0
1A
C
)2
11(
0
2
I
C 0
312I
C
10A (A4)
3
2
03
11
3
111
I
)1(2
21 2
where 0P is the critical load and b the width of the specimen. 'L and L are the length of
the outer and inner spans.
The phase angle Ψ can also be obtained as:
I
II
K
K11 tan)sin(cos
)cos(sintan
Where
M
Ph
A
I 31.52 (A5)
Gcc
K21
cosh4
cosKK I s i nKK II
CHAPTER 4 TGO STRESS AT UNDULATED SURFACE
PAGE 81
Chapter 4
Local stress around spherically symmetrical portions of
thermally grown oxide layer formed on a metal substrate
4.1. Introduction
During oxidation of high temperature alloys, e.g., Ni based superalloys and FeCrAl
alloys, the stress in the alumina scale formed on substrates plays an important role in
spallation of the oxide scale. The thermally grown oxide (TGO), mostly alumina, forms
in thermal barrier coatings (TBCs) when the TBCs are exposed to a high temperature
environment. The spallation of TGO leads to failure of TBCs [9, 14, 159-161]. It is
generally accepted that the stress in TGO varies with the undulated morphology of the
oxide scale [162]. Rumpling of the TGO during thermal cycling of TBCs is believed to
contribute to the failure mechanism of TBCs [9, 163]. The presence and growth of the
TGO not only can introduce significant stress arising from either growth deformation or
thermal expansion mismatch but also can roughen the original interface and even cause
morphological instability during extensive thermal cycling [164, 165]. Therefore, it is
very essential to study the local stresses in TGO and understand factors which have an
effect on them.
The TGO stresses are made up of two parts, thermal expansion mismatch and TGO
growth stress after oxidation. The former can generate very large stress at room
temperature due to drastic temperature drop in a coating system subjected to thermal
cycling oxidation. The later plays an important role for the case where high
temperatures must be maintained for a long period of time such as for power generators
[166]. In comparison with thermal mismatch stress, the growth stress is much less
CHAPTER 4 TGO STRESS AT UNDULATED SURFACE
PAGE 82
studied because of its intrinsic complexities. The stress in the TGO is strongly affected
by the TGO morphology. Stress redistribution will be imposed at the locations where
complex morphology appears and the local stress field will deviate from the mean stress
state. Some enhancement in stress will lead to potential micro cracks. The local TGO
stress is believed to be responsible for micro crack initiation. The understanding of the
TGO stress in a general sense is yet incomplete, let alone the failure mechanisms in a
much more complex coating systems like TBCs.
Extensive experimental studies have been reported on the measurements of TGO stress
in TBCs or Fecralloy using various methods. The average stress in an oxide scale is
significant and can reach the order of several gigapascal, normally varying with
oxidation time [88, 162, 167, 168]. Using piezospectroscopy and a plate model,
Tolpygo et al.[169] successfully determined the lateral growth strain and the growth
stress in the TGO formed on a FeCrAlY alloy substrate. Their results show that the
lateral growth strain increases with oxidation time, which essentially differs from the
constant strain due to molar volume expansion. Clarke [170] proposed a theoretical
model to account for the experimental observation and quantify the lateral growth-strain
evolution. Using Clarke‟s model, the mean growth stress in a flat oxide/metal system
can be predicted. But in many cases, especially when fracture is concerned, local stress
will be of greatest importance. Then the corners or curved portions of an oxide scale
must be considered. Hsueh and Evans [171] developed a series of cylinder models
imposed by plane strain condition to calculate the elastic and viscoelastic stresses for a
curved surface, in which the dependence of the stress state on the oxidation site and
growth strain were discussed. They also evaluated the effect of growth stress on
cracking and spalling.
However, as mentioned above, most of the previous studies were concerned with the
stresses in a TGO formed on a flat oxide/metal system. Even when a curved
morphology was considered, only numerical calculation and finite element analysis
were conducted. Few studies with both carefully designed experiments and analytical
models have been reported possibly because of experimental difficulties. In this work,
we used different sizes of spherical indenters with instrumented micro indentation
machine to produce perfectly spherical indents of different radii and depths on Fecralloy
substrate before formation of the TGO by oxidation. Fecralloy rather than a superalloy
CHAPTER 4 TGO STRESS AT UNDULATED SURFACE
PAGE 83
was used because it forms a dense and adherent α-alumina TGO which excludes other
complex factors, and then the stresses around the indents were measured by
piezospectroscopy method. The stress analysis is mainly based on a spherical model
which permits closed form solutions, preferable for systematic study, and is capable of
giving good approximation of the local stresses around indents as verified by the
experimental results.
4.2 Experiments
Fecralloy (Fe72.8Cr22Al5Y0.1Zr0.1 in wt. %, Goodfellow, UK) was used for the study
as it forms a uniform and adherent α-Al2O3 scale. Button-like substrates, of diameter
25.4 mm, were cut and then mechanically polished to a 0.25 µm finish on both sides,
and cleaned in acetone. Three spherical indenters with radius of 20, 200 and 1000 µm
were employed with various loadings to create roughness on the polished samples with
different radii and depths. 20 and 200 µm radius indents were made by instrumented
micro-indentation while 1000 µm one was created by a Rockwell tester. One group of
the samples with the same thickness of ~2 mm was oxidised in ambient air at 1200 ºC
for 1, 4, 9 and 16 hours. The other consisting of the samples with thickness of ~1, 2, 3
and 4 mm was oxidised at 1200 ºC for 25 hours. After oxidation, all samples were taken
from the furnace immediately and cooled by air blasting to prevent plastic relaxation of
the substrate. Several specimens were additionally cooled in liquid nitrogen. The stress
was measured and compared to the values for specimens cooled by air blasting. The
results showed the same stress value and, therefore, it is reasonable to assume that no
obvious plastic relaxation of the substrate occurred during cooling.
4.3 Results
4.3.1 Morphology characterisation and microstructure observation
The profiles of indents were measured by interferometer (MicroXAM, KLA-Tencor)
prior to and after oxidation shown in Figure 4.1. Significant spikes are present at the
edge of the indent after oxidation. The depth, h and span, l of a typical indent (created
by 200 μm radius indenter) are listed in Table 4.1. The radius of the indent can be
CHAPTER 4 TGO STRESS AT UNDULATED SURFACE
PAGE 84
calculated according to h
lhR
2
2/22
. The radii obtained from indents before
oxidation agree very well with the geometry of spherical indenters but deviation occurs
after oxidation because the depths of indents increase after oxidation (see Table 4.1),
which indicates that the oxide scale formed is generally thicker at the flat surface than at
indents. This also agrees with a previous report by Tolpygo and Clarke [162].
Table 4.1 The depth, span and calculated radius of a typical indent shown in Figure 4.1.
CHAPTER 4 TGO STRESS AT UNDULATED SURFACE
PAGE 85
Figure 4.1 The profiles of a typical indent created by a spherical indenter with 200 µm
radius using 30 N loading prior to (a) and after (b) oxidation at 1200 ºC for 25 hrs. The
inset is the line scan of the location indicated by the dotted line.
CHAPTER 4 TGO STRESS AT UNDULATED SURFACE
PAGE 86
Figure 4.2 (a) SEM image of a typical indent created by a spherical indenter with 200
µm radius using 30 N loading after oxidation at 1200 ºC for 1 hrs; (b) higher
magnification of (a); (c) the microstructure of TGO inside the indent; (d) ZrC
precipitates formed on surface of Fecralloy after oxidation. The ZrC comes from the
diffusion from the substrate.
CHAPTER 4 TGO STRESS AT UNDULATED SURFACE
PAGE 87
Figure 4.3 (a) SEM image of an indent created by a spherical indenter with 200 µm
radius using 30 N loading after oxidation at 1200 ºC for 9 hrs; (b) the grain
microstructure of TGO formed on the substrate.
SEM images show a typical perfect spherical indent in substrate as seen in Figure 4.2a,
b. After oxidation at 1200 ºC for 1 hour, a thin layer of TGO formed on the substrate
and there are also ZrC precipitates dotted around the surface which come from the
diffusion from the substrate (Figure 4.2d). The TGO shows a consistent microstructure
both inside and outside the indent (Figure 4.2c). Figure 4.3 displays that after a longer
period of oxidation, a ring of ridge oxide appears on the edge of indents because of the
oxidation of the pile-ups created by indentation, which also corresponds to Figure 4.1b.
CHAPTER 4 TGO STRESS AT UNDULATED SURFACE
PAGE 88
This ridge morphology can be analogous to the convex portion of TGO undulation
(Figure 4.4). The TGO grain grows to about 500 nm after 9 hours oxidation at 1200C
(Figure 4.3b).
Figure 4.4 Sketch of an undulating TGO. The predominant growth mechanisms are
schematically depicted particularly for TGO convex and concave portions. [166]
4.3.2 Stress measurements
The residual stress in the alumina scale was measured at room temperature using
photoluminescence piezospectroscopy (PLPS) with an optimal probe size of about 3 µm,
where the peak shift of spectra is used to determine the residual stress. Figure 4.5 shows
a typical profile of the peak shift around an indent. The peak shift is nearly constant at
the flat surface and decreases almost to zero at the edge of indent then increases to a
constant value across most part of the indent. Significant decohesion of TGO from the
substrate should have occurred at the edge of indents corresponding to Figure 4.1b and
4.3a, therefore, leading to zero stress in the TGO. The following study will focus on the
peak shift of TGO inside the indents.
CHAPTER 4 TGO STRESS AT UNDULATED SURFACE
PAGE 89
Figure 4.5 A typical profile of the peak shift of the characteristic R-line of α-Al2O3 scale
around the indent created by a 200 µm-radius indenter with 30 N loading after oxidation
at 1200 ºC for 25 hours.
4.4 Discussion
4.4.1 Analytical solutions for the local stress around spherically symmetrical
portions of TGO
The stress can be derived from the peak shift ( ) of the characteristic R-lines of the
Cr3+
luminescence relative to the stress-free alumina given by [86, 172]
jjii 3
1 (1)
Where jj are the hydrostatic components of the stress tensor and ii are the
components of the piezospectroscopic tensor ( ii =7.60 cm-1
/GPa for α-Al2O3) [172].
For a flat scale, the stress can be assumed to be biaxial, i.e. σxx = σyy = σ and σzz =0.
Therefore, the biaxial stress can be determined by 06.5 . However, for a curved
alumina scale, the component σzz normal to the specimen surface is no longer zero and
the in-plane components σxx and σyy should vary from place to place depending on the
geometry of oxide. For a spherically symmetric case like the spherical indent in this
study, an analytical solution for the thermal mismatch stress is obtained as [173]
CHAPTER 4 TGO STRESS AT UNDULATED SURFACE
PAGE 90
33
3
12
1
2
121
1)(
R
HR
E
E
R
HR
R
HRTE
sub
sub
oxox
ox
oxsubox
zz
(2)
zzyyxx
R
HR
R
HR
3
3
1
2
11
(3)
Where the oxide scale and the substrate are distinguished by subscript „sub‟ and „ox‟, H
is the oxide thickness, R is the radius of the indent, α is the thermal expansion
coefficient, ν is Poisson ratio, and ΔT is the temperature change. From Equations (1)-(3),
the peak shift induced by thermal mismatch stress can be calculated.
4.4.2 Effect of curvature radius, depth and TGO thickness on local stress
PLPS measurements were made on the alumina scale formed at the central part of
indents with various radii and depths. Figure 4.6a summarises the peak shift as a
function of indent radius and indent depth after oxidation at 1200 ºC for 25 hours. As
the radius of indent (or the reciprocal of curvature) increases, the peak shift, i.e., stress
increases. For the case of 1000 µm radius which has a very small curvature, the value is
almost the same as that of a flat surface. According to Equation (2) and (3), normalised
tangential stress σxx, σyy and normal stress σzz as a function of the ratio between the
indent radius and oxide thickness are shown in Figure 4.6b, where σ0 is the stress at a
flat surface, ox
ox TE
10 . σzz decreases but σxx increases with increasing radius.
Since )(3
1zzyyxxii , the peak shift follows the same trend as the
tangential stress, hence, it goes up with an increase in the radius. According to Figure
4.6a, indent depth has no obvious effect on the peak shift which is also consistent with
Equations (2) and (3) (in the equations there is no indent depth, h, involved).
CHAPTER 4 TGO STRESS AT UNDULATED SURFACE
PAGE 91
Figure 4.6 (a) Peak shift of the characteristic R-line of α-Al2O3 scale formed inside the
indents on Fecralloy after oxidation at 1200 ºC for 25 hours as a function of indent
radius and indent depth. (b) Normalised σzz and σxx (σyy) as a function of the indent
radius and oxide thickness ratio, R/H.
CHAPTER 4 TGO STRESS AT UNDULATED SURFACE
PAGE 92
4.4.3 Effect of oxidation time and substrate thickness on local stress
Figure 4.7 reveals the peak shift from measurements at indents as a function of
oxidation time and substrate thickness. The peak shift decreases as the oxidation time
increases. During oxidation the oxide scale thickens and according to Figure 4.6b, the
total stress decreases with increasing oxide thickness given the same radius of indent,
hence, the peak shift decreases with the oxidation time. In addition, the substrate
thickness has no obvious effect on the peak shift because the substrate is much thicker
than the oxide scale in the study and it does not exhibit any plastic deformation during
cooling, and also according to Equations (1)-(3) the substrate thickness is not included
in the calculations given the substrate is significantly thicker than the alumina scale,
therefore the peak shifts on the substrates of different thickness give similar values.
CHAPTER 4 TGO STRESS AT UNDULATED SURFACE
PAGE 93
Figure 4.7 (a) Peak shift of the characteristic R-line of α-Al2O3 scale at the indents with
20 and 200 µm radius as a function of (a) oxidation time at 1200 ºC (given substrates
are 2 mm thick) and (b) substrate thickness (given the oxidation time is 25 hours). With
increasing oxidation time the TGO thickness increases and the TGO stress in the indents
decreases. Meanwhile the substrate thickness has no effect on the TGO stress.
It is noted that the peak shift becomes stable gradually after an initial decrease with
extended oxidation time, where the change in stresses cannot be explained purely based
on thermal mismatch stress according to the Equations (1)-(3) where only thermal
CHAPTER 4 TGO STRESS AT UNDULATED SURFACE
PAGE 94
mismatch stress has been considered. The residual stress in the oxide scale at room
temperature comprises two parts, the thermal mismatch stress and the growth stress.
Generally the mismatch stress is larger than the growth stress, so the changing trend of
the total stress based on the peak shift as a function of indent depth and radius can be
explained well with Equations (1)-(3). However, in order to understand the effect of
oxidation time on the stresses in alumina scale and also conduct quantitative analysis,
the growth stress must be included which will be discussed below.
The growth stress, σgrowth, in the scale at the oxidation temperature can be calculated
from the room temperature stress data using the following equation:
growthT
ox
ox
zzyyxxE
E 260.7
3
1 (4)
Where the thermal expansion coefficients of the substrate and oxide are, αsub=14.0 ×
10-6
/ºC and αox=8.2 × 10-6
/ºC, respectively [174], the elastic modulus of the substrate
and oxide are, Eox=400 GPa, Esub=200 GPa, and T
oxE =330 GPa is the oxide Young‟s
modulus at 1200 ºC [174], ΔT is the temperature drop from oxidation temperature
(ΔT=1175 ºC). Poisson ratio is taken to be νox=0.25, νsub=0.3. The oxide thickness H is
measured by SEM images and taken as 0.7, 1.4, 2.2, 2.9 and 3.6 µm for samples
oxidised at 1200 ºC for 1, 4, 9, 16 and 25 hours, respectively. Using the peak shift data
in Figure 4.7a and Equations (2)-(4), the growth stress can be obtained.
CHAPTER 4 TGO STRESS AT UNDULATED SURFACE
PAGE 95
Figure 4.8 Calculated growth stress of the oxide scale formed on a flat surface and the
indents with 20 and 200 µm radius as a function of oxidation time at 1200 ºC.
The growth stress of the oxide scale is presented in Figure 4.8. The difference between
the oxide growth stresses at the indents of different radii is obvious. For the indent of 20
µm radius, the growth stress is tensile from the beginning of oxidation. To the contrary,
the growth stress is compressive for the indent of 200 µm radius and it decreases as a
function of oxidation time. The growth stress formed on a flat surface is almost constant
through the oxidation period carried out in the study. The possible reason for the tensile
stress within the indent of small radius (big curvature) is that partial stress relaxation
may occur in this local region during cooling. The analysis above is based on the
assumption that there is no plastic deformation of the substrate during temperature
change. It is reasonable for a flat surface or locations with small curvatures because the
stress in the substrate is small and relatively uniform. But the presence of an indent with
a very big curvature should lead to a non-uniform stress redistribution in the subscale
region of the substrate with relatively high stresses in specific areas [162, 173]. In
addition, the assumption of the spherical configuration of the indent is not exactly valid
for samples after oxidation, as shown in Table 4.1. These can result in some errors in
the data. However, the experimental results unambiguously indicate that the curvature
has a more important effect on the growth stress in the oxide scale than the depth of
CHAPTER 4 TGO STRESS AT UNDULATED SURFACE
PAGE 96
local morphology. More careful work needs to be done on the TGO growth stress
including lateral growth strain and through-thickness strain at undulations in future.
Figure 4.9 (a) Micro cracks initiate at the ridges of the indenter where tensile hoop
stress arises at this convex portion; (b) higher magnification image of the circled area in
(a), showing the cracks propagate along grain boundary.
CHAPTER 4 TGO STRESS AT UNDULATED SURFACE
PAGE 97
4.4.4 Micro crack patterns caused by TGO stress
Figure 4.9 illustrates the micro crack patterns around the undulating TGO (ridge parts of
indents) where stress acts as the driving force for fracture. It is a tensile stress at this
convex portion. Normally, the TGO has a columnar grain structure and its most
common failure modes are delamination and spallation. Figure 4.10 shows a typical
spallation of TGO on the undulated surface of Fecralloy. It occurs by buckling when the
initial crack reaches a critical size the large stress in the TGO buckles the oxide scale off
the substrate. These patterns of micro cracks illustrated in Figure 4.9 are one of the
failure modes experimentally observed in coating system such as TBCs because in
reality the failure process can be controlled by multiple mechanisms, and the nucleation
and propagation of micro cracks can be driven by different physical and mechanical
forces, e.g. the presence of top coat and thermal stress under temperature
variation/gradient. The interaction of different failure mechanisms is more complex,
however, this study of the simple experimental model still give an insight into TGO
local stress and how to reduce the fracture driving force. For instance, ensuring the
uniformity of TGO is helpful for suppressing TGO hoop tension. The curvature radius
at a curved area has a more significant role than its depth. A smooth (large radius R) or
thin TGO layer (small h) is beneficial since tensile stress decreases with R/h.
Figure 4.10 (a) 70 degree tilted SEM image of a typical undulating morphology of TGO
formed on FeCoCrAlY substrate after oxidation at 1200C for 4 hrs; (b) one spallation of
TGO on the surface by buckling.
CHAPTER 4 TGO STRESS AT UNDULATED SURFACE
PAGE 98
4.5 Summary
A sphere-based analytical model is used and combined with experiments by
indentations to investigate the local stress around the spherically symmetrical portions
of an undulating TGO layer on Fecralloy, which incorporates normal and tangential
stresses. Stress measurement by photoluminescence piezospectroscopy is performed
which confirms the validity of the sphere model to predict the local stress. The TGO
concave portions inside the indents are considered mainly. It is seen that the normal and
tangential stress compete in the determination of the stress state around TGO and one
can overwhelm the other, depending on the ratio of undulation curvature and oxide
thickness (R/H). The radius R can be determined by measuring the surface morphology
with interferometer. In the case of this work where the curvature radius is much bigger
than the TGO thickness, the effect of radius is more significant than the depth of local
curved area. When R is small and comparable to the oxide thickness H, the situation
will be more complex and the radius and depth of curved area will both affect the TGO
stress. More importantly, tension can arise at the undulating TGO/substrate interface.
More careful work needs to be done on TGO growth stress at curved area, considering
the nonuniform and anisotropic nature of the growth stress including the lateral and
through-thickness one.
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 99
Chapter 5
Microstructure evolution and interface morphology in
thermal barrier coatings studied by X-ray microtomography
5.1 Introduction
Thermal barrier coatings (TBCs) have been extensively applied to the hot sections in
gas turbine engines in propulsion and power generation industries to provide protection
against thermal shock, high temperature oxidation and hot corrosion [23, 175, 176]. The
TBC system is a very complex evolving coating system, consisting of a refractory-oxide
ceramic coating, a load bearing substrate alloy and an oxidation-protective bond coat
between the substrate and topcoat, and also a thermally grown oxide (TGO) as an
oxidation product of the bond coat. Each layer is dynamic and evolving and all interact
to control the performance and durability of TBCs.
Even though the coating system is widely used nowadays, testing and evaluation of
TBCs are still challenging. First, the conditions under which they operate are often
extremely harsh, fast temperature transients, high pressures and additional mechanical
loading, as well as oxidative and corrosive environments, which are difficult to
reproduce in laboratory [177]. The coating systems also changes with thermal exposure
or cycling as interdiffusion occurs, microstructure evolves, and the properties of the
constituent layers change. For instance, the ceramic topcoat sinters during service,
leading to increase in elastic modulus and thermal conductivity due to the decrease of
porosity. The properties that need to be evaluated are rarely those of the constituent bulk
materials themselves. For example, while the intrinsic fracture toughness of the TBC,
typically made of 7 wt% yttria-stabilised zirconia (7YSZ) is important, it is the
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 100
toughness that a crack experiences as it extends in or near the interface with TGO that
directly influences the TBC lifetime. As TBCs are expected to become prime reliant
which requires that they can be implemented into the design of turbine engine with
reliable performance, it is essential to develop sensors and non-destructive evaluation
methods to monitor TBC properties and performance, ideally in-situ. There are several
approaches that have been explored, including infrared imaging, Raman spectroscopy,
thermography, impedance spectroscopy, acoustic emission and photoluminescence
piezospectroscopy (PLPS) [22, 86, 87]. Infrared image is used to detect the damaged
region but with rather poor spatial resolution as shown at NASA [178]. Raman
spectroscopy can be applied for phase analysis and determination of residual stress in
TBC by Raman peak shift. More revealing is the PLPS approach which takes use of
luminescence from trace Cr3+
invariably present in alumina TGO and relates the
frequency shift to the local mean stress in the TGO. Using this method, the stress/strain
tomography (mapping) of the TGO, which is the driving force for coating failure, has
been obtained for an electron beam physical vapour deposited TBC [179], however, it
does not give direct visualisation of the microstructure. With continuous development in
X-ray computed tomography (XCT), especially high resolution X-ray microtomography
(µCT) [180], it provides a promising non-destructive method to study the microstructure
and damage evolution in TBCs including TBC sintering, porosity change, flaw
development and interface morphology. When combing µCT with other approaches,
such as conventional scanning electron microscopy (SEM) imaging that has better
resolution but usually involves cutting and polishing, and the PLPS method mentioned
above that gives stress tomography, the microstructure and properties can be related,
thus giving a better understanding of TBCs performance and durability.
Furthermore, the failure mechanism is not well defined and understood. Generally, it is
believed that the coating system fails by a sequence of crack nucleation, growth
(propagation), link-up (coalescence) until a large-scale spallation or delamination occurs
after extended thermal cycling or isothermal exposure upon cooling [9, 83]. Evans etc.
[9] propose that the cracks nucleate due to the tensile stress normal to the interface
induced in the vicinity of imperfections including thickness heterogeneity and interface
undulation, the so-called rumpling or roughening. However, this is still debatable. The
mechanisms for this phenomenon are not sure either. At least two mechanisms are
indentified, including the “ratcheting” mechanism, which briefly means that the
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 101
rumpling is motivated by the TGO lateral compressive stress and accommodated by
plastic deformation of both TGO and bond coat during thermal cycling [9]. Another
mechanism is due to the surface displacement associated with volumetric changes in
bond coat as aluminium depletes [14, 119]. One of the reasons for the uncertainty of
failure phenomenon study is that the conventional ways of sample preparations and
SEM cross-sectioning imaging inevitably involve cutting and polishing which may
introduce artificial cracks and flaws, or enlarge separations. What‟s more, the
destructive cross sectioning of samples exclude the study of the microstructure in an
identical location with thermal exposure. The X-ray tomography provides not only
non-destructive evaluation of the microstructure and damage evolution in an identical
place in one sample, but also gives 3D quantitative information, such as porosity size
and shape, 3D interface morphology and roughness etc. Previous work on atmospheric
plasma sprayed TBCs described in chapter 3 reveals that various kinds of imperfections
nucleate and develop in or near the TGO after isothermal exposure.
The aim of the work is to have an elementary investigation of the microstructure
evolution of the TBC/bond coat interface morphology in an electron beam physical
vapour deposited (EBPVD) TBC on a platinum-modified aluminde bond coat coated on
superalloy by X-ray microtomography combining with SEM analysis. Further
improvements in the acquisition equipment and imaging analysis software, as well as
in-situ testing rigs such as mechanical tester and furnace, are currently being developed,
and can be expected to yield faster imaging of the tested samples and make it possible
to follow the damage accumulation in much greater detail and have better and more
comprehensive quantitative data about the microstructure in TBCs.
5.2 Experiments
5.2.1 X-ray computed tomography
X-ray computed tomography (XCT) provides a non-destructive way of making
three-dimensional (3D) measurements of structure and, by extension through repeated
image acquisition, of following structural evolution over time [181]. Consequently, it is
becoming an important emerging tool for the study of degradation and damage
accumulation processes in a wide variety of environments.
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 102
It is based on taking a series of high-spatial-resolution digital radiographs (projections),
with the sample rotated by a small increment between each image. The image can be
created using fan beam, cone beam, or parallel beam illumination. Typically, the
projections are captured over a rotation angle of 180°. Image contrast is generated with
samples by variations in attenuation or, for phase contrast imaging, in refractive index.
The acquisition process can be illustrated by a schematic (Figure 5.1). A 3D image of
the contrast variation is then reconstructed mathematically. This procedure consists of
two stages. This first stage is a filtering step, preprocessing the radiograph and reducing
the effect of noise. The filtering step can be adapted to a specific application and usually
involves a Radon transform. The second part of the reconstruction process is to
back-project the filtered projections onto a grid and to sum up the contributions from
each of the radiographs to obtain a 3D representation of the object. Originally,
two-dimensional (2D) slices (tomograms; tomos means slice or section in Greek) were
reconstructed using fan beams and line detectors, but 2D detectors mean it is now more
usual to reconstruct many 2D slices to form 3D data sets [181].
Figure 5.1 Schematic of X-ray computed tomography.
X-ray microtomography (µCT) is now approximately 30 years old [182, 183]. The
potential of X-ray tomography was immediately acknowledged, first in medicine [184]
and soon after in materials science [185]. Today, laboratory and synchrotron X-ray µCT
systems can provide resolutions in the micrometer range or even better. X-ray
tomography has been used not only as a non-destructive tool for structural
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 103
characterisation in 3D, but more importantly, for understanding fatigue and damage
accumulation in materials.
5.2.2 Experimental procedures
The TBC system investigated was cut from a turbine blade provided by Rolls Royce
plc., UK. It comprises ~200 µm thick EBPVD TBC made of 7YSZ on a
platinum-modified aluminde bond coat (~40 µm) on CMSX-4 superalloy. One sample
was cut and ground into the required size (diameter should be less than 0.5 mm) in order
to fit into the field of view with highest resolution (Figure 5.2a). The as-deposited
sample was scanned by X-ray µCT machine (Versa, shown in Figure 5.2b) in the Henry
Moseley Imaging Facility, University of Manchester. Then it underwent thermal cycling
(1150 °C, 1 hour duration, fan cooling) for 10, 50 and 100 times. After each thermal
treatment, an identical location of the sample was scanned.
Figure 5.2 (a) EBPVD TBCs sample prepared, ready for acquisition of X-ray
radiographs. The sample is glued by epoxy onto a plastic tube fastened on a nail and
then put on the sample stage between X-ray source and detector in Versa X-ray µCT
machine (b).
The experimental settings are summarised in table 5.1. 20× lens was chosen to have a
proper field of view for the sample. The sample stage was moved to the source as close
as possible for a more energetic X-ray beam, thus saving scanning time. The minimum
distance was 10 mm with avoidance of collision. The distance of the detector was
adjusted for phase contrast, hence the resolution. Two voltages are available, 90 or
140KV. To get the X-ray transmittance rate in the range of 20-25% for best quality
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 104
images, the voltage was set 90KV. The low energy filter No.6 (LE6) was used, which
filtered out the lower energy part of the X-ray beam. As YSZ and the nickel-based
superalloy are quite absorbing for X-ray, exposure time for one radiograph was set to 7
seconds, in order to have a sufficient amount of signals to be received by the detector.
As many as 3201 radiographs were generated during 180° rotation to have a high
signal-to-noise ratio. Binning 2 was used to averages 2×2 pixels into one. This reduces
both the data size and noise significantly, which is desirable for the later analysis, but at
the price of resolution. A comparison of images reconstructed by binning 1 and 2 was
made. It was found that the difference in image quality was negligible, therefore,
binning 2 was adopted in this work. Pixel size (or called voxel size in 3D) was claimed
as 0.85 µm by the machine. However, this pixel size does not represent the minimum
feature size that can be resolved. Many commercial suppliers of laboratory equipment
quote the pixel size in a rather loose manner. Such approaches are not helpful, and more
formal methods to quantify the resolution of a 2D projection as a function of contrast
difference should be used. Realistically, the minimum feature in the images of this work
that can be discerned is about 2 pixels, i.e. 1.7 µm provided there is little noise. So to
improve the tomography image quality, one should either increase the resolution or
reduce the data noise.
Table 5.1 X-ray microtomography experiment settings for the TBC sample scanning.
Lens Distance of
detector(mm)
Distance of
source (mm)
Voltage
(KV)
Power (W)
20 6 10 90 8
Exposure
time (s)
Filter Number of
images
Binning Pixel size
(um)
7 LE6 3201 2 0.85
After acquisition of the radiographs, they were reconstructed using a filtered
back-projection algorithm and 2D slices were then stacked to build 3D images of the
sample. Subsequent image analysis including segmentation of different phases and
quantification of porosity and interface morphology can be implemented in image
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 105
software Azivo. Figure 5.3 shows the three steps of X-ray tomography study described
above.
Figure 5.3 Procedures of X-ray tomography study, (a) acquisition of radiographs of an
EBPVD TBCs, (b) reconstruction of 2D slices from radiographs, (c) stacking the 2D
slices to build the 3D data, ready for image analysis such as visualisation, segmentation
and quantification.
Except for the study of one sample by X-ray microtomography, another specimen cut
from the same turbine blade were prepared and went through the same thermal cycling
treatments, then the sample was cross-sectioned and investigated by SEM (Philips
XL30), following metallurgical preparation, i.e., mounting in a resin, grinding and
polishing to 1 µm diamond finish. 10 SEM images were taken at the TGO/bond coat
interface and then the interfaces were extracted by Image J software based on the phase
contrast. The 10 images of interface morphology were put together and interfacial
roughness was calculated to compare with the tomography results. After the SEM
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 106
experiment, the sample was taken out of the resin by heating the resin in a hot stage.
Subsequently, the sample was put back into furnace for further thermal treatment. The
process repeats.
5.3 Results and discussions
5.3.1 3D visualisation of TBCs
3D microstructures of the EBPVD TBCs sample after thermal cycling (50×1h at
1150 °C) are shown in Figure 5.4. The reconstructed images give promising results with
reasonable details of the microstructure in the TBCs despite the high attenuation of
zirconia and nickel-based superalloy, and also the noise effect caused by the porosity.
As expected, the illustrated microstructure contains YSZ TBC with a columnar structure,
a thermally grown oxide (TGO), which appears black grey below the TBC, a
platinum-modified aluminde bond coat (BC) with an interdiffusion zone (IDZ)
appearing quite bright, and the underlying superalloy substrate (Figure 5.4B). The gray
scale of one phase depends on its X-ray absorption (or attenuation). Generally, the
heavier the atoms are, the more they absorb X-ray and the brighter they appear in the
reconstructed images. Indeed, the alumina TGO is the darkest and the IDZ containing
some heavy elements due to the diffusion from substrate during thermal exposure shows
high brightness.
The slices of the microstructure in the TBCs in different directions (x, y and z) are
shown in Figure 5.4A-D. When going through the slices, microstructure features such as
cracks, interface fluctuation, pore size and shape can be observed. Figure 5.4D-F
demonstrate the microstructure at different slices in the coating thickness direction. The
band of pores in Figure 5.4B is due to the interrupted deposition processing. The TBCs
samples were prepared and taken out of the retort to do quality control measurement and
then put back into the retort to continue the deposition process, thus resulting in the
band of pores near the top of the coating. These reveal the change in the inter-columnar
spacing (porosity) through the coating thickness. A fine porosity is shown at the bottom
of TBC near the interface with TGO (Figure 5.4E), where columns of all
crystallographic orientations nucleate and then the dominant orientation favoured by the
deposition conditions grow further. The columns are seen along with inter-columnar
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 107
pores to extend through the thickness direction. The pores become larger gradually to
the top of the TBC because the columns near the interface are smaller and denser,
meanwhile the columns at the top of the TBC become bigger but scarcer after growth,
leading to larger inter-columnar pores as seen in Figure 5.4D. The quantitative analysis
for porosity information involves the following procedure, (a) generation of a histogram
of the linear attenuation coefficient from the gray scale images. These show bimodal
peaks, one due to the pores and the other due to solid material, (b) selection of
region-of-interest in individual images, and (c) conversion of the gray scale density
maps to “black and white” images by a process of segmentation involving assignment
(materials and pores) for each voxel. Finally, the porosity measurements are done by
voxel counting of black and white regions in the segmented images. This will be
conducted in future work with improved resolution of the pore/grain boundary.
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 108
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 109
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 110
Figure 5.4 (A) Reconstructed X-ray tomography images of the microstructure in an
EBPVD TBC deposited on a platinum-modified aluminde bond coat coated on CMSX-4
superalloy after thermal cycling (50×1h at 1150 °C), (B, C) slices of microstructure in x
and y directions, (D-F) microstructure at different slices in the through coating thickness
direction (z), showing the interdiffusion zone and the inter-columnar porosity that
increases from the bottom to the top of TBC.
5.3.2 Microstructure evolution and damage accumulation
One of the reasons for the complexity of the TBC system is due to its evolving
microstructure and properties with thermal exposure or cycling. All of the constituent
layers in the TBCs interact with each other and evolve with time. Hence it is essential to
characterise the microstructure evolution and tract the damage including pores, cracks,
and interface instability etc. which govern the TBCs durability. Figure 5.5 shows the
reconstructed slices of the microstructure at an identical location of the TBC sample as a
function of thermal cycling numbers (strictly, the slices do not represent an exactly
identical place as during scanning the sample was tilted, even for a very small angle,
thus it is unable to show the exactly same place in a 2D slice). The images show several
changes discernable in the microstructure: (A) inter-columnar pores and cracking in
TBC. During exposure ceramic TBC sinters, leading to reduction of micro pores and
compromise of strain compliance, but maybe resulting in bigger inter-columnar pores.
Upon cooling large thermal expansion misfit stress may cause cracking in TBC and
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 111
early failure. It is worth to mention that in the current work the image resolution is not
enough to support a detailed study of pores and cracks evolution. Further work needs to
be done to investigate the sintering of TBC; (B) Thickening of the TGO as expected.
This does not give new insight into the TBC research. However, the very clear contrast
between TGO and other layers allows 3D interface to be readily and reliably segmented
and quantified which will be discussed in the next section; (C) Severe damage at the
edge of bond coat. Although the edge of bond coat is not covered by TBC and exposed
directly to the hot oxygen during thermal cycling in atmosphere, these damages which
appear black grey in Figure 5.5 are not oxides. Actually they are voids shown clearly in
both the outer surface and inside zone in the reconstructed volume rendering images
(Figure 5.6). The figure reveals that the voids grow with thermal cycling. The formation
of the voids is due to the aluminium depletion in the bond coat on account of its
oxidation and inter-diffusion with the substrate. This leads to the phase transformation
from β to the higher-density γ‟-Ni3Al phase, accompanied with a magnitude of volume
decrease in the range from 8% to 38%, depending on which of the two depletion
processes prevail [186]. It can be predicted that this TBC coating will fail by
delamination from the edge with prolonged thermal cycling. It has been proposed and
studied that the phase transformation in the bond coat associate with volume change can
cause the TGO interface undulation (rumpling) [186, 187], which is the interest of the
next section.
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 112
Figure 5.5 The evolution of the microstructure in the EBPVD TBCs as a function of
thermal cycling number, showing (A) inter-columnar spacing and cracking in TBC
caused by sintering of the TBC, (B) TGO thickening, and (C) severe damage at the edge
of the bond coat.
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 113
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 114
Figure 5.6 The reconstructed volume rendering images of the EBPVD TBCs as a
function of thermal cycling number, (A) the outer surface of the sample, revealing the
damage accumulation at the outer surface, and (B) the view of the inside structure by
slicing the data, both showing the damages at the bond coat edge and voids increase
with thermal cycling.
5.3.3 Characterisation of interface morphology
As mentioned above, the phases in the reconstructed data can be segmented based on
gray scale intensity by sophisticated algorithms. The segmentation allows quantification
of the microstructure possible, such as the volume fraction of each phase and interfacial
roughness. Each of the multiple layers in the TBC system is separated and shown with
its associated interface morphology for as-deposited and as-heated TBCs (100×1h at
1150 °C) (Figure 5.7). It appears that the interface morphologies evolve with changes in
wavelength and amplitude in Figure 5.7A-C. Quantitative analyses are needed for the
study.
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 115
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 116
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 117
Figure 5.7 Reconstruction for each constituent layer in the as-deposited and as-heated
TBCs (100×1h at 1150 °C) after segmentation, revealing each interface between layers,
(A) substrate superalloy, (B) interdiffusion zone in bond coat, (C) outer layer in bond
coat, (D) TGO (the TGO in the as-deposited sample is too thin to be identified), (E)
TBC topcoat (it is set translucent to allow underneath layers seen). This segmentation
technique makes each interface can be extracted for further study.
The TGO/bond coat interface is of the most interest as EBPVD TBCs usually fail at or
near the interface through crack nucleation, growth, and linkage until large scale
spallation by a rumpling phenomenon [9]. Especially the platinum-modified aluminide
bond coat is found to tend to rumple [119]. The roughening in platinum-modified
aluminide bond coat is due to the surface displacement associated with volumetric
changes in the bond coat as aluminium depletion occurs. In this work, the TGO/bond
coat interface is extracted from the reconstructed data (Figure 5.8) and transferred to
grey scale images in matlab. The coordinates and heights of the images are then used to
calculate the 3D interfacial roughness (Figure 5.9).
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 118
Figure 5.8 The evolution of the TGO/bond coat interface with thermal cycling
(TBC/bond coat interface for as-deposited TBCs,). The interface is extracted from
segmented data.
Figure 5.9 Quantification of interface morphologies in matlab by input of segmented
data.
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 119
Table 5.2 3D interfacial roughness of the TGO/bond coat interface in this EBPVD
TBCs sample with thermal cycling calculated by matlab.
*The edges of the interface with large voids are cropped before roughness calculation.
The calculated 3D interfacial roughness is shown in Table 5.2. The results show that the
interfacial roughness is of the magnitude of ~1 micron for all thermal treatment
conditions, and there is no obvious change in the interfacial roughness. The roughness
for as-deposited condition is even bigger than that for as-heated ones. It may be because
of the smoothing of the interface at initial stage of thermal exposure, also indicated in
Figure 5.7C where the thermally cycled sample shows a smoother interface with larger
wavelength than the as-deposited one. The interfacial roughness is expected to increase
with extended thermal cycling as already shown in the data of 10, 50 and 100 cycles.
The absence of the obvious rumpling in this study may be due to the formation of
extensive voids at the bond coat edge, resulting from the volumetric change in the bond
coat by phase transformation. These voids relax the stress and hence reduce the strain
energy significantly in the TGO that is the motivation for TGO undulation, but the
extent of relaxation is unknown. To verify the accuracy of the data obtained by X-ray
tomography results, the cross-sectional SEM micrographs of the TBCs have been also
analysed (Figure 5.10). The interfacial morphologies are plotted in the Figure 5.11,
along with the calculated interfacial roughness indicated. The comparison shows that
the values obtained from the two methods are similar in general, although the values for
as-deposited condition show some deviation. Therefore, the X-ray tomography provides
a promising method to study the TBCs with a good accuracy (at least in interface
quantification and the technique is still improving with higher resolution and in-situ
testing development), as well as with the benefits of 3D information and non-destructive
evaluation. It is also worth to note, the standard deviation of interface height, defined as
Roughness (standard deviation of
interface height) (µm)
As_deposited 1.24
1150_10x1h 0.91
1150_50x1h 0.92
1150_100x1h 0.98*
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 120
interfacial roughness, is not sufficient to characterise the interfacial morphology as it
has been described in Chapter 4 that the curvature influences the local stress more than
the depth in an undulated area. More careful work needs to be done in the analysis of
the interface wavelength and amplitude. Finally, it is very interesting to combine the
X-ray µCT and PLPS methods, to generate a comprehensive tomography
(microstructure and stress) of the TGO interface in TBCs.
Figure 5.10 Cross-sectional electron scanning microscopy (SEM) micrographs near the
thermally grown oxide (TGO) interface for as-deposited and as-thermally-cycled
EBPVD TBCs. Such 10 images are combined to quantify the interfacial roughness.
CHAPTER 5 X-RAY TOMOGRAPHY STUDY OF EBPVD TBCs
PAGE 121
Figure 5.11 the TGO/bond coat interface morphologies for as-deposited and thermally
cycled TBCs obtained from cross-sectional SEM images. The calculated interfacial
roughness is indicated along with the plotted lines.
6. Summary
X-ray microtomography has been employed to study the microstructure evolution, and
interface morphology in an EBPVD TBC deposited on a platinum-modified aluminide
bond coat with thermal cycling. The 3D microstructure and its evolution with thermal
cycling have been characterised clearly, including the inter-columnar porosity and
sintering of TBC, TGO growth and formation of voids in the bond coat due to phase
transformation associated with volume change. In addition, the TGO/bond coat interface
has been investigated with quantification of the 3D interfacial roughness. Tomography
results there is no obvious TGO/bond coat interface rumpling indicated by the
calculated standard deviation of interface height, which is consistent with results from
SEM image analysis. Further work needs to be done in the characterisation of pore size
and shape during TBC sintering, and more detailed analysis on interface morphologies
including wavelength and amplitude.
CHAPTER 6 SIMPLE AND PLATINUM MODIFIED ALUMINIDE COATINGS
PAGE 122
Chapter 6
Structure, oxidation resistance and mechanical properties of
simple and Pt-modified aluminide coatings on superalloy
6.1 Introduction
The intermetallic bond coat serves as an environment barrier in thermal barrier coatings
(TBCs) to provide protection for the underlying superalloy against oxidation, hot
corrosion and thermal fatigue [5, 44, 49]. The bond coat is arguably the most crucial
component in the TBC system. Its chemistry and microstructure influence durability
through the structure and morphology of the TGO created as it oxidizes and, moreover,
the system performance is linked to its creep and yield characteristics [9].
Bond coats are generally in two categories, NiCoCrAlY overlay coating and diffusion
coating. The diffusion coatings, particularly aluminide coatings that are the product of
interdiffusion between the superalloy component and an aluminium source, together
with electron beam physical vapour deposited (EBPVD) yttria stabilised zirconia (YSZ)
topcoat are widely used in gas turbine engine in aeroplanes. Additional performance
benefits can be gained to the aluminide coating by incorporating rare earth elements,
such as Zr, Hf and Y, which impart the bond coat enhanced resistance against corrosion,
cyclic oxidation or improvement in TBC retention [51, 52, 188, 189]. One of the most
widely adopted durability enhancements to the simple aluminide is platinum modified
nickel aluminide, fabricated by electroplating a thin layer of Pt onto the superalloy and
then aluminising by either pack cementation or chemical vapour deposition. These
coatings are typically made of single β phase, with Pt in solid solution [190, 191]. The
mechanism of the Pt effect in an aluminide bond coat is yet to be understood completely.
CHAPTER 6 SIMPLE AND PLATINUM MODIFIED ALUMINIDE COATINGS
PAGE 123
Some believe that the doping of Pt is to improve the adhesion of the alumina to the bond
coat, perhaps by mitigating the effects of sulphur in the alloy and coating [192, 193].
There have been intensive researches on the microstructure and performance of the
aluminide coatings. Swadzba [194] investigated the structure and cyclic oxidation
resistance of Pt, Pt/Pd-modified and simple aluminide coatings on CMSX-4 alloys and
found that the Pt/Pd-modified aluminide exhibited the best performance. Haynes et al.
[195] compared platinum modified aluminide (β-NiPtAl) and γ-γ‟ coatings on N5Y and
CMSX-4 alloys. Both classes of coating showed protective scale formation up to 1000
cycles at 1050 °C. The thermally grown oxide (TGO) growth rate on the γ-γ‟ coatings
was faster than that on CVD β-NiPtAl. However, the platinum modified aluminide
coatings are prone to degradation by rumpling. Tolpygo and Clarke [196] concluded
that a relatively small temperature decrease in each oxidation cycle, of the order of
100 °C or perhaps even smaller is sufficient to cause rumpling of the aluminide coating
on a second generation superalloy substrate. Observations indicated that neither the
martensite nor the β-γ‟ phase transformation occurring during heating and cooling is
fully responsible for rumpling.
Although many studies on the oxidation resistance of the aluminide coatings have been
reported, few are focused on mechanical properties of the bond coat, such as its fracture
toughness. This goal of the work is to study the mechanical properties of simple and
Pt-modified aluminide coatings, along with characterisations of their microstructure and
oxidation resistance.
6.2 Experiments
6.2.1 Experimental procedures
The investigation concerned two types of commercial aluminide coatings cut from high
pressure turbine blade provided by Rolls Royce, plc.. The simple aluminide coatings
deposited on a directionally solidified MARM-002 superalloy were prepared by the
means of pack cementation at 880 ± 10 °C for 21 hours ± 15 min while Pt-modified
coatings on CMSX-4 superalloy were manufactured by first electroplating with Pt to 7.5
µm thick, which took around 4 hours depending on the cathode efficiency, and then
CHAPTER 6 SIMPLE AND PLATINUM MODIFIED ALUMINIDE COATINGS
PAGE 124
treatment in vacuum at 1100 °C for 1 hour, followed by the aluminisation as mentioned
above. Some of the samples were prepared in a metallurgical way, grinding and
polishing to 1 µm diamond paste finish for cross-section examination by Philips XL30
electron scanning microscope (SEM) equipped with energy dispersive spectroscopy
(EDS) and mechanical testing by micro instrumented indentation. The other samples
were tested by thermogravimetric analysis (TGA) for oxidation resistance
measurements. Phase compositions were identified using X-ray diffraction (XRD) with
Cu-Kα radiation at 40 mA and 50 kV (Philips, PW1830).
6.2.2 Micro instrumented indentation
These years have seen an increasing interest in using instrumented indentation
techniques including micro and nano indentations to investigate thin film or coating
materials due to the advantages of indentations in term of accuracy, convenience and
ability to measure small scales. But it has the limitations that the results obtained from
indentation only represent local rather than overall properties.
The method Oliver and Pharr [197] introduced in 1992 for measuring hardness and
elastic modulus by instrumented indentation techniques has widely been adopted and
used in the characterisation of mechanical behaviours of materials at small scales. Its
attractiveness stems largely from the fact that mechanical properties can be determined
directly from indentation load and displacement measurements without the need to
image the hardness impression. During the past two decades, several important changes
that both improve its accuracy and extend its realm of application have been developed
through experience in testing a large number of materials and by improvements to
testing equipment and techniques. For example, the measurement of contact stiffness by
dynamic techniques allows for continuous measurement of properties as a function of
depth and also facilitates more accurate identification.
The method was developed to measure the hardness and elastic modulus of a material
from indentation load–displacement data obtained during one cycle of loading and
unloading. A schematic representation of a typical data set obtained with a Berkovich
indenter is presented in Figure 6.1, where the parameter P designates the load and h the
displacement relative to the initial surface.
CHAPTER 6 SIMPLE AND PLATINUM MODIFIED ALUMINIDE COATINGS
PAGE 125
Figure 6.1 Schematic of indentation load–displacement data showing important
measured parameters. [197]
There are three important quantities that must be measured from the P–h curves: the
maximum load, Pmax, the maximum displacement, hmax, and the elastic unloading
stiffness, S=dP/dh, defined as the slope of the upper portion of the unloading curve
during the initial stages of unloading (also called the contact stiffness). The accuracy of
hardness and modulus measurement depends inherently on how well these parameters
can be measured experimentally. Another important quantity is the final depth, hf, the
permanent depth of penetration after the indenter is fully unloaded. Figure 6.2 shows a
schematic illustration of the unloading process, showing parameters characterising the
contact geometry.
Figure 6.2 Schematic representation of the indenter-sample contact. [197]
CHAPTER 6 SIMPLE AND PLATINUM MODIFIED ALUMINIDE COATINGS
PAGE 126
According to the Oliver & Pharr procedure, the unloading curve needs to be fit by a
power law relationship:
m
phhBP )( (1)
Where P is the indentation load, h is the displacement, B and m are empirically
determined fitting parameters, and hp is the final displacement after complete unloading.
The unloading stiffness S is then established by differentiating equation (1) at the
maximum depth of penetration, h=hmax:
1
maxmax )()( m
phhmBhhdh
dPS
(2)
The contact depth is also estimated from the load-displacement data:
S
Phhc
max
max (3)
Where Pmax is the peak load and ε is the constant which depends on the indenter
geometry
From the basic measurements contained in the load-displacement data, the projected
area A of the hardness impression is estimated via evaluating the empirically
determined indenter shape function at the contact depth, hc:
)( chfA
(4)
Once the contact area is determined from the load-displacement data, the hardness H
and effective elastic modulus Eeff could be expressed as follow:
A
PH max
(5)
and
A
SEeff
2
1
(6)
The effective elastic modulus, which accounts for the fact that elastic deformation
occurs in both the specimen and the indenter, is given by:
i
i
eff EEE
22 111
(7)
In this study, the instrumented indentation method detailed above is used to measure the
hardness and elastic modulus of simple and Pt-modified aluminide coatings, and also
CHAPTER 6 SIMPLE AND PLATINUM MODIFIED ALUMINIDE COATINGS
PAGE 127
characterise the ductility and fracture toughness of the coatings. Fracture toughness
determination by indentation was originally designed for brittle bulk materials. The
equation is [198]
2/1
2/3))((0154.0
H
E
c
PKc (8)
Where P is indentation load, E is elastic modulus, H is hardness and c is the length of
generated crack. A schematic illustration of the cracking system induced by indentation
is shown in Figure 6.3, along with a SEM image of cracking in sapphire. A modification
has been proposed to address the case for interfacial fracture toughness where i, C and S
represent interface, coating and substrate, respectively [199].
2/1
2/3))((0154.0 i
v
cH
E
c
PK (9)
2/1
2/1
2/1
2/1
2/1
)(1
)(
)(1
)(
)(
S
C
C
C
S
S
i
H
HH
E
H
HH
E
H
E
(10)
Figure 6.3 (a) Schematic of Vickers-produced indentation-fracture system, showing
peak load P and characteristic dimensions c and a of cracks, (b) Scanning electron
micrographs of radial crack system in a brittle material, sapphire, with P=10 N load
[198].
CHAPTER 6 SIMPLE AND PLATINUM MODIFIED ALUMINIDE COATINGS
PAGE 128
6.3 Results and discussions
6.3.1 Microstructure characterisation
Figure 6.4 presents the surface microstructure of simple aluminide coating on
superalloy. The chemical composition of the coating surface corresponds to β-NiAl
phase and there are precipitates of elements such as Cr and Co present. They can form
ZrCr and ZrCo.
Figure 6.4 Surface microstructure of simple aluminide coating.
The cross-section micrograph is shown in Figure 6.5a, along with the micro-areas
examined by EDS summarised in Table 6.1. There should not be any platinum
concentration in the simple aluminide coating. However, the platinum content in the
Table 6.1 is not exactly zero maybe because of equipment error. The amount is very
small and can be neglected. The green coating is about 40 μm thick and at the bottom of
the coating there is a thin diffusion layer appearing bright. The diffusion zone will grow
with further thermal treatment during which Al diffuses towards the substrate and Ni
diffuses out from substrate to the coating. For pack aluminising at temperature below
~1000 °C, such as in this case (880 °C), and especially when the packs have
high-activity donors (e.g., metallic aluminium), coatings grow by predominantly inward
diffusion of aluminium. Precipitates of Co and Cr are dotted in the β-NiAl phase. The
concentration profile of elements along the depth of the coating indicated in Figure 6.5a
is examined using EDS and shown in Figure 6.5b. It reveals that the Al level decreases
CHAPTER 6 SIMPLE AND PLATINUM MODIFIED ALUMINIDE COATINGS
PAGE 129
and Ni increases towards the substrate as expected. There are also differences in Cr and
Co contents in the coating and substrate.
Figure 6.5 (a) Cross-section SEM micrographs of simple aluminide coating, (b) the
concentration profile of elements along the line marked in (a).
Table 6.1 Chemical compositions in area 1, 2 and 3 in Figure 6.5a.
CHAPTER 6 SIMPLE AND PLATINUM MODIFIED ALUMINIDE COATINGS
PAGE 130
Figure 6.6 shows the surface microstructure of Pt-modified aluminde coating. It appears
similar to that of simple aluminide coating (Figure 6.4). The composition is β-NiAl
phase with Pt in solid solution, confirmed by XRD pattern (Figure 6.7).
Figure 6.6 Surface microstructure of Pt-modified aluminide coating.
Figure 6.7 X-ray diffraction (XRD) patterns from the Pt-modified aluminide coating
surface.
The cross-section image (Figure 6.8a) reveals its zonal structure consisting of an outer
layer (~50 μm thick) which is β-NiAl phase and diffusion layer (~15 μm thick) which is
located at the bottom of the coating above the substrate and is rich in bright precipitates
comprising refractory elements, such as Ta, W and Ti etc.. The EDS results in Table 6.2
obtained from the marked area in Figure 6.8a indicate that the outer layer is rich in Pt.
The concentration profiles along the depth of the coating shown in Figure 6.8a are
CHAPTER 6 SIMPLE AND PLATINUM MODIFIED ALUMINIDE COATINGS
PAGE 131
depicted in Figure 6.8b and c. The concentration of Ni is increasing gradually through
the coating except within the diffusion zone to the level in the substrate alloy. The
decreasing rate of Al content is relatively low in the outer zone. After the diffusion zone
the Al content drops more rapidly and reaches the level in the substrate. Pt
concentration decreases faster in the outer zone and then becomes more stable. At a
depth of about 50 μm there is a significant increase of elements such as W, Ta, Cr and
Ti which indicates the presence of the diffusion zone (Figure 6.8c).
Figure 6.8 (a) Cross-section SEM micrographs of Pt-aluminide coating, (b, c) the
concentration profile of elements along the line marked in (a).
CHAPTER 6 SIMPLE AND PLATINUM MODIFIED ALUMINIDE COATINGS
PAGE 132
Table 6.2 Chemical compositions in area 1, 2 and 3 in Figure 6.8a.
6.3.2 Oxidation resistance
Thermogravimetric analysis (TGA) of the simple and Pt-modified aluminide coatings
cut from the commercial turbine blades was conducted in order to compare their
oxidation resistance. One batch of samples was tested at 1150 °C for 9 hours and
another for 20 hours. As shown in Figure 6.9, the Pt-modified aluminide coating
displays much better oxidation resistance than its counterpart in this isothermal
condition. In the initial stage, the simple aluminide coating oxidises and gains weight
rapidly and even after 9 and 20 hours the oxidation rate of simple aluminide coating is
still bigger than that of Pt-modified aluminide coating revealed by the slopes of the
weight gain curves. The weight gain of simple aluminide coating after 9 hours is around
3 mg/cm2 and 3.5 mg/cm
2 after 20 hours while Pt-modified aluminide coating gains
weight less than 1 mg/cm2 after 20 hours. The different slopes of the later stable TGO
growth curves indicate that the oxidation mechanisms may be different because at this
stage all TGO formed is dominantly α-Al2O3 and the growth rate is controlled by
diffusion. It indicates that the Pt addition might mitigate the Al diffusion towards the
coating surface to form alumina. The experiments of oxidation resistance comparison
between simple aluminide and Pt-modified aluminide coating are quite reproducible,
showing a large difference in oxidation rate as seen in Figure 6.9. Further work needs to
be done on the examinations of the cross-section of the oxidised coatings.
CHAPTER 6 SIMPLE AND PLATINUM MODIFIED ALUMINIDE COATINGS
PAGE 133
Figure 6.9 Thermogravimetric analysis (TGA) results of the simple and Pt-modified
aluminide coatings tested at 1150 °C for (a) 9 hours and (b) 20 hours.
6.3.3 Mechanical properties
Instrumented indentation was employed to characterise the mechanical properties of the
simple and Pt-modified aluminide coatings. A berkovich indenter was used for the
determination of hardness and elastic modulus of the coatings (Figure 6.10). The
indentation depth was fixed as 2 μm. The load-displacement curves of several tests on
both coatings are shown in Figure 6.11. The maximum loads for the simple and
Pt-modified aluminide coatings reach about 0.6 and 0.4 N respectively. At the same
indentation depth, the Pt-modified aluminide coating shows a smaller maximum load P
and larger final depth hr (the residual indent depth after the removal of indenter)
displayed in Figure 6.11. This indicates that the Pt-modified aluminide coating exhibits
more ductility than the simple aluminide coating, which imparts additional benefit to the
bond coat besides the oxidation resistance enhancement.
Figure 6.10 Berkovich indentations of 2 μm depth on the simple and Pt-modified
aluminide coatings, labelled are indents.
CHAPTER 6 SIMPLE AND PLATINUM MODIFIED ALUMINIDE COATINGS
PAGE 134
Figure 6.11 Indentation load-displacement curves of several tests on both simple and
Pt-modified aluminide coatings, indicating more ductility for the later one.
The hardness and elastic modulus of the two coatings are summarised in Table 6.3. The
elastic modulus of the two coatings are very similar, around 130 GPa while the simple
aluminide coating is much harder than the Pt-modified one (9.1 and 5.3 GPa
respectively), which may account for the ductility difference. The reason may be that
the simple aluminide coating has a higher aluminium content than the Pt-modified
coating, which leads to more brittleness and higher value in hardness. More
investigation needs to be done on this.
Table 6.3 Elastic modulus and hardness of the simple and Pt-modified aluminide
coatings measured by instrumented indentation.
CHAPTER 6 SIMPLE AND PLATINUM MODIFIED ALUMINIDE COATINGS
PAGE 135
One way to evaluate the fracture toughness of a material is to make a Vickers
indentation on it and increase the load until it is cracking. From the load and crack
length combined with material‟s hardness and elastic modulus, fracture toughness can
be estimated. As shown in Figure 6.12 an indentation made by 2 N load generates large
cracks in the simple aluminide coating, but 1 N does not. The large cracks are parallel to
the coating/substrate interface and no cracks penetrating the interface occur. This
indicates that the fracture toughness of the coating is lower than the interfacial
toughness of the coating and substrate so the cracks are obstructed to propagate through
the interface. The load-displacement curve shows a plateau on the loading curve
between 1 and 2 N which implies the critical load to crack the coating. Then the load
can be used to deduce fracture toughness if crack lengths are known. Figure 6.13
displays the results for Pt-modified aluminide coating. No cracks can be generated for
loads up to 4 N. Sharp and symmetric Vickers indentation are seen in the coating. An
accurate calculation of the fracture toughness needs symmetrical radial cracks shown in
Figure 6.3b. Although the crack pattern as seen in simple aluminide coating (Figure
6.13a) excludes a quantitative determination of the fracture toughness, a qualitative
comparison of the fracture toughness between simple aluminide and Pt-modified
aluminide coating can be done. The Pt-modified coating exhibits higher fracture
toughness to accommodate the impression by indentation. Generally, higher aluminium
content leads to more brittleness and hence lower fracture toughness although it can
increase high temperature oxidation resistance. The Pt-modified aluminide coating has
lower aluminium content so it should have higher fracture toughness than simple
aluminide coating. Meanwhile, it also shows good oxidation resistance despite its lower
aluminium level. This may be due to the mitigation of aluminium diffusion by platinum
during oxidation. The indentation method provides a convenient and straight forward
tool to characterise coatings fracture toughness. The addition of Pt not only enhances
the oxidation resistance for the aluminide coating, but also increases its fracture
toughness to resist cracking.
CHAPTER 6 SIMPLE AND PLATINUM MODIFIED ALUMINIDE COATINGS
PAGE 136
Figure 6.12 Load-displacement curves of indentation made with 1 and 2 N loads on
simple aluminide coating. Inserted are the optical images of the corresponding Vickers
indentations, revealing large cracks for the 2 N case, but not for 1 N.
Figure 6.13 Load-displacement curves of indentation made with 2 and 4 N loads on
Pt-modified aluminide coating. Inserted are the optical images of the corresponding
Vickers indentations.
CHAPTER 6 SIMPLE AND PLATINUM MODIFIED ALUMINIDE COATINGS
PAGE 137
6.4 Summary
Commercial simple and Pt-modified aluminide coatings cut from airfoil of high
pressure turbine blade are investigated in the work including microstructure, oxidation
resistance and mechanical properties. Both coatings are made of β-NiAl phase, with Pt
in solid solution for Pt-modified coating. Thermogravimetric analysis shows that the
simple aluminide coating gains 3.5 mg/cm2 weight after 20 hours exposure at 1150 °C
compared to less than 1 mg/cm2 for Pt-modified one. The different oxidation rates
indicate that the Pt addition may mitigate the Al diffusion. It is also worth to mention
that during the thermogravimetric analysis the samples were cut from the turbine blades
directly, so there were two sides of the samples that were exposed to the atmosphere and
not protected by the aluminide coating. Therefore, they gained more weight during TG
tests than the samples completely covered by the simple and Pt-modified coatings. The
absolute values of the oxidation rate of the two coatings will be different from the
current data, but the trend will still be the same. Pt-modified coating shows a much
better oxidation resistance. Elastic modulus are similar for the coatings, around 130 GPa
while simple aluminide coating is harder than the Pt-modified one with 9.1 and 5.3 GPa,
respectively. Moreover, indentation method provides a convenient and straight forward
way to evaluate the fracture toughness of coatings. It shows that Pt-modified aluminide
coating exhibits more ductility and higher fracture toughness than simple aluminide one
because of its lower aluminium content. Therefore, the addition of Pt imparts not only
better oxidation resistance, but also enhanced fracture toughness for the aluminide
coatings.
CHAPTER 7 TEMPERATURE DEPENDENCE OF RAMAN SPECTRUM OF YSZ
PAGE 138
Chapter 7
Temperature dependence of Raman scattering of
yttria-stabilised zirconia
7.1 Introduction
Yttria-stabilised zirconia (YSZ) is one of the most important ceramic materials for
modern technological applications. It has been widely used in thermal barrier coatings
(TBCs) in jet turbine engines because of its excellent mechanical properties and
structural stability [9, 14]. At ambient pressure, pure zirconia exists as three polymorphs
between room temperature and its melting point at 3100 K, i.e., monoclinic (m-ZrO2),
tetragonal (t-ZrO2) and cubic (c-ZrO2) [200, 201]. The cubic and tetragonal structure
can be stabilised by doping with rare-earth oxide like yttria. Raman spectroscopy
provides a viable method to study its structure, transformation characteristic and
structural stability against temperature, pressure and stabilisers [202-205].
In this work we present the temperature dependence of Raman scattering measurements
of tetragonal YSZ coatings and cubic YSZ single crystals with different yttria contents.
Experimental data are analysed and compared to those predicted using simple models in
the literatures.
7.2 Experiments
The materials used in this study are free standing 8 wt%, tetragonal prime, YSZ about
100 um thick and ones deposited by electron-beam physical vapor deposition (EBPVD)
on a superalloy, and also cubic YSZ single crystals with two yttria contents (9 and 30
CHAPTER 7 TEMPERATURE DEPENDENCE OF RAMAN SPECTRUM OF YSZ
PAGE 139
wt %). Raman spectra were acquired with samples mounted in a heating stage
(TS1500EV-6, Linkam Scientific Instruments, Ltd., UK) under a Raman microscope
(RenishawTM
2000, UK). During the measurements, a laser beam (Ar ion, 514.5 nm)
was focused to a minimum observable spot size about 3 um on the YSZ surface. Raman
spectra were collected at room temperature and intervals of 50 ºC between 50 and 400
ºC and intervals of 100 ºC between 500 and 1100 ºC. The collection time was about 20
seconds.
7.3 Results
Raman spectra obtained under various temperatures are deconvoluted after subtracting
the baseline, assuming a lorentzian distribution using a commercial fitting software
(Grams AI, Galactic). A mixture of lorentzian and Gaussian distribution was first
employed but the results showed that the Gaussian was fitted to the limit, indicating that
the spectra only contain a lorentzian distribution. Figure 7.1 shows that the Raman
spectra of tetragonal zirconia exhibit six characteristic bands (3Eg+2B1g+1A1g) and there
is no evidence of a phase transition, implying that the t‟-YSZ is stable throughout the
temperature range in this study. The mode assignment adopted in the work follows
those recently proposed by Quintard et al. [206] and Milman et al. [207] although there
are still disputes about the mode at ~260 cm-1
[200, 204]. All the bands shift to lower
frequency and broaden with increasing temperature.
CHAPTER 7 TEMPERATURE DEPENDENCE OF RAMAN SPECTRUM OF YSZ
PAGE 140
Figure 7.1 Raman spectra of tetragonal 8YSZ at various temperatures after subtracting
baselines.
As one B1g band appears on the shoulder of a more intense Eg band and the uncertainty
of the fitting for the specific band becomes bigger with increasing temperature (Figure
7.1), only five centre peak positions are fitted and shown in Figure 7.2. All bands shift
to lower wavenumber as the temperature increases, as observed in some other zirconia
materials by the previous reports [202, 208-210]. The data can be fitted very well with a
linear function and the least-squares fitting parameters are summarised in the Table 7.1.
The data can also be fitted with a polynomial, such as a parabola. However, since the
linear fitting is very good, the higher order terms in the polynomial will be very small
and negligible. The poor correlation coefficient (R-square) for the mode Eg (~140 cm-1
)
results from the uncertainty of deconvolution of the peak when background removal is
conducted. The fitting parameters in Table 7.1 show that there is little difference
between the measurements from free-standing YSZ and intact YSZ coatings on a
superalloy which suggests that the effect of thermal mismatch stress is negligible. All
the analysis will be detailed in the following section where simple models are used to
explain the effect of temperature on the Raman shift.
CHAPTER 7 TEMPERATURE DEPENDENCE OF RAMAN SPECTRUM OF YSZ
PAGE 141
Figure 7.2 Peak positions of the Raman spectra for tetragonal 8YSZ as a function of
temperature (solid dots). The black solid lines are the best linear fits to the experimental
data and the red dash lines are the results predicted by theoretical calculations using the
methods from references [208, 210].
Table 7.1 Fitting parameters of the temperature dependence of Raman shift for
tetragonal 8YSZ both free standing and deposited on superalloy.
7.4 Discussions
There are two contributions to the thermally induced Raman shift. The first is the
volume (or implicit) effect which is a consequence of the change in the equilibrium
CHAPTER 7 TEMPERATURE DEPENDENCE OF RAMAN SPECTRUM OF YSZ
PAGE 142
interatomic spacings with temperature. The second is phonon-excitation (or explicit)
effect which is due to the changes in vibrational amplitudes of atoms, i.e., the
occupation number of phonon states [208, 210, 211]. The thermally induced Raman
shift at constant pressure can be expressed as
TVP PTT
(1)
where ω is the Raman shift, T is the temperature, P is the pressure, α is the thermal
expansion coefficient (CTE) and κ is the compressibility,. The first and second parts
represent the explicit and implicit effect, respectively.
In addition, if the sample is constrained to a substrate there is also a contribution from
the thermal mismatch stress due to the difference of CTE between the coating and
substrate. The magnitude of this contribution has been estimated by Lughi and Clarke
[208] as follows.
TEPmis 2 (2)
where Πp is the in-plane piezospectroscopic coefficient, E is the biaxial elastic modulus
of YSZ coating, Δα is the CTE difference and ΔT is the temperature change considered
in the work. For the estimation of all the parameters, the thermal mismatch is taken as ~
6×10-6
K-1
[174], the elastic modulus of columnar structural YSZ is generally low, about
~1-10 GPa, ΔT is 1000 K. The Πp is taken approximately as the value of the uniaxial
piezospectroscopic coefficient reported by Cai et al [212]. The maximum estimation of
the thermal mismatch contribution to the Raman shift can be then evaluated, in the
magnitude of 10-5
which is two orders lower than the slopes of the fitting parameters in
Table 7.1. This suggests that the thermal mismatch has a negligible effect on the
thermally induced Raman shift. Although the claim is not entirely new, it is the first
time that experimental data is provided and it agrees with the analysis very well.
The shift caused by the volume thermal expansion effect alone can be given by [211]
'
0'
)( dTP
T
T
T T
vol
(3)
where both the thermal expansion coefficient α(T) and compressibility κ(T) are
functions of temperature and will be estimated with the available data from literatures.
CHAPTER 7 TEMPERATURE DEPENDENCE OF RAMAN SPECTRUM OF YSZ
PAGE 143
The piezosepectroscopic coefficient TP
is assumed as independent of temperature
following Cai et al.[210].
Thermal expansion coefficient α(T) is estimated from functional temperature
dependence of lattice parameters as follows
dT
Tdc
TcdT
Tda
TaT
)(
)(
1)(
)(
12)( (4)
where the a(T) and c(T) are lattice parameters of tetragonal structure, respectively.
The compressibility κ can be determined from standard elasticity relations,
ca 2 (5)
2
13121133
1333
2)( cccc
cca
(6)
2
13121133
131211
2)(
2
cccc
cccc
(7)
Using the Equation (3)-(7) and parameters summarised in Lughi and Clarke‟s work
[208], the Raman shifts due to volume effect contribution have been evaluated for each
Raman peak. The red dash lines in Figure 7.2 display the results, with the calculated
contribution added to the Raman peak measured at room temperature. In general, the
theoretical predictions agree with the experimental data very well, similar to the results
of previous work [208] which suggests that the volume effect accounts for more
temperature-induced Raman shift than the phonon-excitation effect does, implying that
the bonding in YSZ has ionic character [210]. The only exception is the peculiar mode
A1g (~260 cm-1
) which has been confirmed by many researches about pressure
dependence of Raman scattering of zirconia materials as a softening mode and it has a
negative piezospectroscopic coefficient [204, 207, 212-214], hence, leading to the
positive contribution of the calculated volume effect. The phenomenon of the
temperature dependence of this mode has been observed in previous works [202, 208,
209, 214], but no explanation was provided. It needs further study for the origin of the
question.
The full width at half maximum (HWHM) of each Raman band as a function of
temperature is shown in the Figure 7.3. It is well established that Raman bands broaden
CHAPTER 7 TEMPERATURE DEPENDENCE OF RAMAN SPECTRUM OF YSZ
PAGE 144
as the temperature increases due to the reduced phonon lifetime. A simple model that
quantitatively describes the broadening has been proposed by Hart et al [215],
12/exp
21)(
0
0kTh
T
(8)
where Γ0 is the intrinsic linewidth at the absolute zero, ν0 is the center band position of
the selected mode, h is the Plank constant, and k is the Boltzmann constant. The
Equation (8) describes the temperature dependence of the Raman band width of the
YSZ used in this work reasonably well.
Figure 7.3 FWHM of each Raman band of tetragonal 8YSZ as a function of
temperature.
Raman scattering measurement from cubic YSZ single crystals with two yttria content
(9 and 30 wt %) in the temperature range of 25-1000 ºC are also presented in Figure 7.4.
The main feature of these spectra is the asymmetric band peaking at about 600 cm-1
which appears strongly in the cross-polarization configuration (F2g symmetry) [210].
This band shifts to lower frequency continuously up to the highest temperature of the
measurement (Figure 7.5) without obviously abrupt change of slope, indicating that the
material is very stable throughout the temperature range. The same analysis method
used for tetragonal YSZ coating can be applied for the cubic YSZ crystals and it shows
that the volume effect also accounts for most of the thermally induced Raman shift.
Different yttria contents in the cubic YSZ lead to various lattice parameters, a(T) and
CHAPTER 7 TEMPERATURE DEPENDENCE OF RAMAN SPECTRUM OF YSZ
PAGE 145
c(T), so the compressibility k(T) and thermal expansion α(T) are changed, resulting in
different temperature dependence of Raman spectra. New models need to be developed
to analyse these spectra.
Figure 7.4 Raman spectra of cubic YSZ single crystals with 9 wt% (a) and 30 wt% (b)
yttria content at various temperatures after subtracting baselines.
CHAPTER 7 TEMPERATURE DEPENDENCE OF RAMAN SPECTRUM OF YSZ
PAGE 146
Figure 7.5 Temperature dependence of the Raman band F2g for cubic 9YSZ and 30YSZ.
7.5 Summary
The temperature dependence of Raman scattering of tetragonal YSZ coatings and cubic
YSZ single crystals with two yttria contents has been studied under the temperature
range from ambient temperature to 1100 ºC. The materials are stable over the
temperature range and show no evidence of phase transformation. The comparison of
the measurements between the free-standing tetragonal YSZ and intact YSZ coatings on
substrates shows that the thermal mismatch expansion has a negligible effect on the
Raman shift. The Raman bands shift to lower frequency and broaden with increasing
temperature. The shift can be mostly attributed to the volume expansion effect due to
the piezospectroscopic shift. One exception is the A1g mode which has the opposite
trend. Further work is needed to study this phenomenon.
CHAPTER 8 CONCLUSIONS AND FUTURE WORK
PAGE 147
Chapter 8
Conclusions and future work
8.1 Discussion and Conclusions
(1) A modified four-point bending test was employed to investigate the interfacial
toughness of atmospheric plasma sprayed (APS) yttria-stabilised zirconia (YSZ)
thermal barrier coatings (TBCs) after isothermal heat treatments at 1150 ºC. The
delamination of the TBCs occurred mainly within the topcoat, several to tens of microns
above the interface between the topcoat and bond coat. X-ray diffraction analysis
revealed that the topcoat was mainly tetragonal in structure with a small amount of the
monoclinic phase. The calculated energy release rate increased from ~50 J/m-2
for
as-sprayed conditions to ~120 J/m-2
after annealing at 1150 ºC for 200 hours with a
loading phase angle about 42º. X-ray micro-tomography was used to track in 3D the
evolution of the topcoat microstructure non-destructively at a single location as a
function of thermal exposure time. This revealed how various types of imperfections
developed near the interface after exposure which could be responsible for the initiation
of cracks. The 3D interface was reconstructed and showed no significant change in the
interfacial roughness after thermal exposure.
Interfacial toughness is critical for understanding the TBCs performances and failure
behaviours. In this work, the interfacial toughness of the APS TBCs has been
successfully determined by a modified four-point bending test. However, this method
requires complicated sample preparations such as carefully designed geometry and the
attachment of stiffeners. Besides, there is a sample size constraint and the experimental
results are very sensitive to the notch as the crack usually starts to grow from the place
CHAPTER 8 CONCLUSIONS AND FUTURE WORK
PAGE 148
near the notch. In this work, the notch was made manually to the TGO/bond coat
interface, which could not be very accurate due to man made errors. A better method
with reliability and convenience to measure the interfacial toughness of the TBCs needs
to be developed. The indentation method is easy to conduct, but the results are not very
reproducible due to the uncertainty of the crack pattern generated, especially near the
interface. Conventional mechanical tests such as tensile, compressive or buckling tests
require large amount of samples with fixed geometry and complex sample preparations.
A miniature compression test by indentation will have the potential to measure the
interfacial toughness with both robustness and ease.
(2) The growth of thermally grown oxide (TGO) can induce significant local stress
which can potentially lead to interfacial delamination and failure in a coating system.
Direct measurements of the local stress, particularly around the undulating TGO is very
important to understand the failure mechanisms of a coating system as micro cracks
usually initiate from these parts. The object of this study is to combine stress
measurements by photoluminescence piezospectroscopy (PLPS) and analytical
solutions to investigate the local stress around spherically symmetrical portions of a
TGO layer formed on Fecralloy. Spherical indenters are used to create curvature with
different curvature radii and depths on alloys before oxidation. From theoretical analysis
the normal and tangential stresses compete at the curved areas and one can overwhelm
the other, depending on the ratio between the curvature radius and TGO thickness (R/H).
In the case of this work where the curvature radius is much bigger than the TGO
thickness, the tangential stress overwhelms the normal stress. The effect of curvature
radius on stress is found more significant than the depth of local curved area. The total
stress at the curved areas as a function of oxidation time is obtained and discussed. The
TGO growth stress is also derived.
As discussed above, understanding the TGO stress especially growth stress at the
curved area is crucial to develop any realistic TBC failure model. Despite the
development of understanding TBCs performances in last several decades, a complete
TGO growth model is still yet to be developed partially because the TGO growth is
asymmetrical in nature, i.e., in the lateral and through-thickness directions. And also the
TGO stress at the curved area is very complex, involving stress redistribution and
relaxation. A comprehensive experiment, preferably with analytical solutions, along
CHAPTER 8 CONCLUSIONS AND FUTURE WORK
PAGE 149
with numerical calculation which is not constrained by the geometry will be helpful to
study the growth stress at the curved area.
(3) Electron beam physical vapour deposited (EBPVD) TBCs with a β-(Ni,Pt)Al bond
coat on CMSX4 substrate from Rolls Royce pls. are investigated by micro X-ray
computed tomography (XCT). A single location of the sample is scanned by XCT after
thermal cycling at 1150 °C non-destructively. The 3D microstructures evolution and
damage accumulation are characterised with a reasonably good resolution. TGO layer
thickens and severe oxidation damage occurs at the edge of bond coat. The 3D
interfacial morphology between topcoat and bond coat is successfully extracted by
reconstruction. The 3D interfacial roughness is calculated by matlab and compared to
the value obtained by conventional 2D scanning electron microscope (SEM) image
analysis, showing they are similar hence proving the accuracy of tomography method.
The calculated interfacial roughness does not change much even after 200 thermal
cycles, indicating there is not obvious rumpling in this TBCs sample.
Micro-tomography exhibits a great potential to study the properties and failure
mechanisms of TBCs with an improving resolution.
The micro-tomography has the advantage to characterise materials non-destructively in
3D, but has a lower resolution compared to the conventional SEM. The tomography can
measure the microstructure evolution while the photoluminescence piezospectroscopy
technique can measure the stress evolution. Therefore, experiments should be done to
combine the tomography and SEM with photoluminescence piezospectroscopy to obtain
a comprehensive characterisation of TBCs.
(4) The commercial simple aluminide coatings from Rolls Royce pls. are prepared by
the means of pack cementation at 880 °C for 21 hours while Pt-modified coatings are
electroplated with Pt prior to the aluminisation. Microstructural and phase analysis
reveal that both the coatings consist mainly of β-NiAl phase. The cross-section
microstructure of Pt-modified coating is zonal and composed of β-NiAl phase zone,
diffusion zone and secondary reaction zone. The concentration profiles show that both
Pt and Al contents decrease gradually towards the substrate. Thermogravimetric
analysis (TGA) tests at 1150 °C for 20 hours indicate that the Pt-modified aluminide
coating is much more resistive for oxidation than simple aluminide coating. The
CHAPTER 8 CONCLUSIONS AND FUTURE WORK
PAGE 150
mechanical tests made by instrumented micro-indentation demonstrate a convenient and
feasible method to measure the mechanical properties and test the ductility of coatings.
It shows both coatings have similar young‟s modulus around 130 GPa while
Pt-modified aluminide coating is more ductile and has a higher fracture toughness than
simple aluminide coating.
(5) The Raman spectra of 8 wt% tetragonal yittria-stabilised zirconia (YSZ) coatings
and cubic YSZ single crystals with two yttria concentrations (9 and 30 wt %) in the
range of 25-1100 ºC are investigated. The materials are stable over the temperature
range and show no evidence of phase transformation. The comparison of the
measurements between the free-standing tetragonal YSZ and intact YSZ coatings on
superalloy illustrates that the thermal mismatch expansion has a negligible effect on the
Raman shift. All Raman bands except the A1g band shift to lower frequency and
broaden with increasing temperature. The shift can be mostly attributed to the volume
expansion effect due to the piezospectroscopic shift. This method indicates that the
Raman spectrum can be used to monitor the temperature in YSZ without contact, in
applications, such as thermal barrier coatings.
8.2 Future work
(1) Although the four-point bending test provides a reliable way to measure the
interfacial toughness of TBC coatings, it has limitations such as sample size
requirement and complex preparation procedures and also need of a stiffener (usually
the same substrate material) adhered to the tested samples, which is troublesome to
prepare. These factors limit its application as a fast, convenient and reliable method to
measure the coating interfacial toughness. As described in Chapter 6, instrumented
indentation provides a fast and straight forward way to study the coating toughness
qualitatively and compare different coatings although quantitative data can be difficult
to obtain because random cracks can be generated by indentation in coatings with
complex microstructure. Future work should be carried out to develop a new method to
characterise interfacial toughness of coatings in a fast and convenient way and it should
also allow the observation of crack propagation.
CHAPTER 8 CONCLUSIONS AND FUTURE WORK
PAGE 151
(2) The TGO growth mechanism should be understood further as it is essential to
develop a realistic model to study the TGO stress. The nonuniform and anisotropic
TGO growth stress at the undulated region is of great interest as it is the local
redistributed stress that initiates and accelerates crack propagation. Systematic study of
the TGO growth stress at curved areas with analytical solutions or numerical
calculations, combined with careful experiments as proof will be done in future work.
(3) Micro X-ray computed tomography has been demonstrated in Chapter 3 and 5 as a
very promising method to evaluate and investigate the TBCs properties and failure
mechanisms. It not only allows non-destructive observations of microstructure
evolution at an identical location, but also study of damage accumulation and 3D data
quantification. In-situ experiments such as samples scanned by X-ray in a furnace rig or
under a mechanical loading are currently being developed. This will provide much more
information about TBCs properties and performance and help understand TBCs failure
modes. Much more work will be done on both APS and EBPVD TBCs by micro X-ray
tomography, including TBC sintering, TGO interface stability and combination of
microstructure tomography with stress tomography measured by photoluminescence
piezospectroscopy which correlates the TBC microstructure with its performance.
REFERENCES
PAGE 152
References
[1] Langston LS. Powering Ahead. Mechanical Engineering 2011; 133:30-33.
[2] www.faa.gov/news/press_releases/news_story.cfm?newsId=12439.
[3] Clarke DR, Oechsner M, Padture NP. Thermal-barrier coatings for more efficient
gas-turbine engines. Mrs Bulletin 2012; 37:891-902.
[4] www.rolls-royce.com/about/technology/gas_turbine_tech/introduction.jsp.
[5] Reed RC. The superalloys Fundamentals and Applications. Cambridge University
Press 2006.
[6] Morrell P, Rickerby DS. Advantages/Disadvantages of Various TBC Systems as
perceived by the Engine Manufacturer. AGARD Report 823, Neuilly-sur-Seine, France
1998:20-1-9.
[7] Schulz U, Leyens C, Fritscher K et al. Some recent trends in research and
technology of advanced thermal barrier coatings. Aerospace Science and Technology
2003; 7:73-80.
[8] Levi CG. Emerging materials and processes for thermal barrier systems. Current
Opinion in Solid State & Materials Science 2004; 8:77-91.
[9] Evans AG, Mumm DR, Hutchinson JW et al. Mechanisms controlling the durability
of thermal barrier coatings. Progress in Materials Science 2001; 46:505-553.
[10] Padture NP, Gell M, Jordan EH. Materials science - Thermal barrier coatings for
gas-turbine engine applications. Science 2002; 296:280-284.
[11] Evans AG, Clarke DR, Levi CG. The influence of oxides on the performance of
advanced gas turbines. Journal of the European Ceramic Society 2008; 28:1405-1419.
[12] Vassen R, Jarligo MO, Steinke T et al. Overview on advanced thermal barrier
coatings. Surface & Coatings Technology 2010; 205:938-942.
[13] Wright PK, Evans AG. Mechanisms governing the performance of thermal barrier
coatings. Current Opinion in Solid State & Materials Science 1999; 4:255-265.
[14] Clarke DR, Levi CG. Materials design for the next generation thermal barrier
coatings. Annual Review of Materials Research 2003; 33:383-417.
[15] Meier SM, Gupta DK. The Evolution of Thermal Barrier Coatings in Gas-Turbine
Engine Applications. Journal of Engineering for Gas Turbines and Power-Transactions
of the Asme 1994; 116:250-257.
[16] Smialek JL, Archer FA, Garlick RG. Turbine Airfoil Degradation in the
Persian-Gulf-War. Jom-Journal of the Minerals Metals & Materials Society 1994;
46:39-41.
[17] Stott FH, Dewet DJ, Taylor R. Degradation of Thermal-Barrier Coatings at Very
High-Temperatures. Mrs Bulletin 1994; 19:46-49.
[18] Borom MP, Johnson CA, Peluso LA. Role of environmental deposits and operating
surface temperature in spallation of air plasma sprayed thermal barrier coatings. Surface
& Coatings Technology 1996; 86:116-126.
[19] Kramer S, Yang J, Levi CG, Johnson CA. Thermochemical interaction of thermal
barrier coatings with molten CaO-MgO-Al2O3-SiO2 (CMAS) deposits. Journal of the
American Ceramic Society 2006; 89:3167-3175.
[20] Aygun A, Vasiliev AL, Padture NP, Ma XQ. Novel thermal barrier coatings that
are resistant to high-temperature attack by glassy deposits. Acta Materialia 2007;
55:6734-6745.
[21] Evans AG, Hutchinson JW. The mechanics of coating delamination in thermal
gradients. Surface & Coatings Technology 2007; 201:7905-7916.
REFERENCES
PAGE 153
[22] Christensen RJ, Lipkin DM, Clarke DR, Murphy K. Nondestructive evaluation of
the oxidation stresses through thermal barrier coatings using Cr3+ piezospectroscopy.
Applied Physics Letters 1996; 69:3754-3756.
[23] Miller RA. Thermal barrier coatings for aircraft engines: History and directions.
Journal of Thermal Spray Technology 1997; 6:35-42.
[24] Bose S, DeMasiMarcin J. Thermal barrier coating experience in gas turbine
engines at Pratt & Whitney. Journal of Thermal Spray Technology 1997; 6:99-104.
[25] Stecura S. Advanced Thermal Barrier System Bond Coatings for Use on
Nickel-Base, Cobalt-Base and Iron-Base Alloy Substrates. Thin Solid Films 1986;
136:241-256.
[26] Wing R. White hot technology - a turbine revolution. Materials World 2000;
8:10-12.
[27] Hasselman DPH, Johnson LF, Bentsen LD et al. Thermal-Diffusivity and
Conductivity of Dense Polycrystalline Zro2 Ceramics - a Survey. American Ceramic
Society Bulletin 1987; 66:799-806.
[28] Klemens PG, Gell M. Thermal conductivity of thermal barrier coatings. Materials
Science and Engineering a-Structural Materials Properties Microstructure and
Processing 1998; 245:143-149.
[29] Stecura S. 2-Layer Thermal Barrier Coating for High-Temperature Components.
American Ceramic Society Bulletin 1977; 56:1082-&.
[30] Yang F. PhD Thesis, Electrical and thermal properties of yttria-stabilised zirconia
(YSZ)-based ceramic materials. University of Manchester 2011.
[31] Chevalier J, Gremillard L, Virkar AV, Clarke DR. The Tetragonal-Monoclinic
Transformation in Zirconia: Lessons Learned and Future Trends. Journal of the
American Ceramic Society 2009; 92:1901-1920.
[32] http://www.doitpoms.ac.uk/tlplib/fuel-cells/sofc_electrolyte.php. University of
Cambridge.
[33] Virkar AV. Role of ferroelasticity in toughening of zirconia ceramics. Key
Engineering Materials 1998; 153-154:183-210.
[34] Mercer C, Williams JR, Clarke DR, Evans AG. On a ferroelastic mechanism
governing the toughness of metastable tetragonal-prime (t ') yttria-stabilized zirconia.
Proceedings of the Royal Society a-Mathematical Physical and Engineering Sciences
2007; 463:1393-1408.
[35] Brandon JR, Taylor R. Phase-Stability of Zirconia-Based Thermal Barrier
Coatings .1. Zirconia Yttria Alloys. Surface & Coatings Technology 1991; 46:75-90.
[36] Vanvalzah JR, Eaton HE. Cooling Rate Effects on the Tetragonal to Monoclinic
Phase-Transformation in Aged Plasma-Sprayed Yttria Partially-Stabilized Zirconia.
Surface & Coatings Technology 1991; 46:289-300.
[37] Taylor R, Brandon JR, Morrell P. Microstructure, Composition and Property
Relationships of Plasma-Sprayed Thermal Barrier Coatings. Surface & Coatings
Technology 1992; 50:141-149.
[38] Clarke DR, Phillpot SR. Thermal barrier coating materials. materials today 2005;
8:22-29.
[39] Wang YF, Yang F, Xiao P. Glass-like thermal conductivities in
(La1-x1Yx1)(2)(Zr1-x2Yx2)(2)O7-x2 (x = x(1) + x(2), 0 <= x <= 1.0) solid solutions.
Acta Materialia 2012; 60:7024-7033.
[40] Bansal NP, Zhu DM. Effects of doping on thermal conductivity of pyrochlore
oxides for advanced thermal barrier coatings. Materials Science and Engineering
a-Structural Materials Properties Microstructure and Processing 2007; 459:192-195.
REFERENCES
PAGE 154
[41] Saruhan B, Francois P, Fritscher K, Schulz U. EB-PVD processing of
pyrochlore-structured La2Zr2O7-based TBCs. Surface & Coatings Technology 2004;
182:175-183.
[42] Vassen R, Traeger E, Stover D. New thermal barrier coatings based on
pyrochlore/YSZ double-layer systems. International Journal of Applied Ceramic
Technology 2004; 1:351-361.
[43] Wu J, Wei XZ, Padture NP et al. Low-thermal-conductivity rare-earth zirconates
for potential thermal-barrier-coating applications. Journal of the American Ceramic
Society 2002; 85:3031-3035.
[44] Pollock TM, Lipkin DM, Hemker KJ. Multifunctional coating interlayers for
thermal-barrier systems. Mrs Bulletin 2012; 37:923-931.
[45] Nicholls J. Advances in coating design for high-performance gas turbines. Mrs
Bulletin 2003; 28:659-670.
[46] Wu RT, Kawagishi K, Harada H, Reed RC. The retention of thermal barrier
coating systems on single-crystal superalloys: Effects of substrate composition. Acta
Materialia 2008; 56:3622-3629.
[47] Smeggil JG, Funkenbusch AW, Bornstein NS. A Relationship between Indigenous
Impurity Elements and Protective Oxide Scale Adherence Characteristics. Metallurgical
Transactions a-Physical Metallurgy and Materials Science 1986; 17:923-932.
[48] Hou PY. Beyond the sulfur effect. Oxidation of Metals 1999; 52:337-351.
[49] Stiger MJ, Yanar NM, Topping MG et al. Thermal barrier coatings for the 21st
century. Zeitschrift Fur Metallkunde 1999; 90:1069-1078.
[50] Gauje GMCA. US Patent 3,486,927. 1969.
[51] Goward GW, Cannon LW. Pack Cementation Coatings for Superalloys - a Review
of History, Theory, and Practice. Journal of Engineering for Gas Turbines and
Power-Transactions of the Asme 1988; 110:150-154.
[52] Bianco R, Rapp RA. Pack Cementation Aluminide Coatings on Superalloys -
Codeposition of Cr and Reactive Elements. Journal of the Electrochemical Society 1993;
140:1181-1190.
[53] Niu Y, Wu WT, Boone DH et al. Oxidation Behavior of Simple and Pt-Modified
Aluminide Coatings on In738 at 1100-Degrees-C. Journal De Physique Iv 1993;
3:511-519.
[54] Patnaik PC, Elder J, Thamburaj R, Immarigeon JP. Protective Coatings for Aero
Engine Gas-Turbine Hot Section Components. Cim Bulletin 1987; 80:105-105.
[55] Strangman TE. US Patent 5,514,482. 1996.
[56] McGill IR, Selman GL. US Patent 4,399,199. 1983.
[57] Tatlock GJ, Hurd TJ. Platinum and the Oxidation Behavior of a Nickel Based
Superalloy. Oxidation of Metals 1984; 22:201-226.
[58] Nagaraj BA, Connor WB, Jendrix RW et al. US Patent 5,427,866. 1995.
[59] Rickerby DS, Bell SR, Wing RG. US Patent 5,667,663. 1997.
[60] Sampath S, Jiang XY, Matejicek J et al. Role of thermal spray processing method
on the microstructure, residual stress and properties of coatings: an integrated study for
Ni-5 wt.%Al bond coats. Materials Science and Engineering a-Structural Materials
Properties Microstructure and Processing 2004; 364:216-231.
[61] Boone DH, Shen S, Mckoon R. Electron-Beam Evaporation of Low
Vapor-Pressure Elements in Mcral Coating Compositions. Thin Solid Films 1979;
64:299-304.
[62] Shankar S, Koenig DE, Dardi LE. Vacuum Plasma Sprayed Metallic Coatings.
Journal of Metals 1981; 33:13-20.
[63] Jackson JE. US Patent No. 3,470,347. 1969.
REFERENCES
PAGE 155
[64] Taylor TA, Overs MP, Gill BJ, Tucker RC. Experience with Mcral and Thermal
Barrier Coatings Produced Via Inert-Gas Shrouded Plasma Deposition. Journal of
Vacuum Science & Technology a-Vacuum Surfaces and Films 1985; 3:2526-2531.
[65] Browning JA. US Patent 5,120,582. 1992.
[66] Sato A, Harada H, Kawagishi K. Development of new coatings for advanced
Ni-base superalloys. Journal of the Japan Institute of Metals 2006; 70:192-195.
[67] Kawagishi K, Sato A, Harada H. A concept for the EQ coating systell for
nickel-based superalloys. Jom 2008; 60:31-35.
[68] Tryon B, Cao F, Murphy KS et al. Rutheniumm-containing bond coats for thermal
barrier coating systems. Jom 2006; 58:53-59.
[69] Tryon B, Feng Q, Wellman RG et al. Multilayered ruthenium-modified bond coats
for thermal barrier coatings. Metallurgical and Materials Transactions a-Physical
Metallurgy and Materials Science 2006; 37A:3347-3358.
[70] Cao F, Pollock TM. Creep deformation mechanisms in Ru-Ni-Al ternary B2 alloys.
Metallurgical and Materials Transactions a-Physical Metallurgy and Materials Science
2008; 39A:39-49.
[71] Hemker KJ, Mendis BG, Eberl C. Characterizing the microstructure and
mechanical behavior of a two-phase NiCoCrAlY bond coat for thermal barrier systems.
Materials Science and Engineering a-Structural Materials Properties Microstructure and
Processing 2008; 483-84:727-730.
[72] Smialek JL, Hehemann RF. Transformation Temperatures of Martensite in
Beta-Phase Nickel Aluminide. Metallurgical Transactions 1973; 4:1571-1575.
[73] Chen MW, Ott R, Hufnagel TC et al. Microstructural evolution of platinum
modified nickel aluminide bond coat during thermal cycling. Surface & Coatings
Technology 2003; 163:25-30.
[74] Karlsson AM, Hutchinson JW, Evans AG. The displacement of the thermally
grown oxide in thermal barrier systems upon temperature cycling. Materials Science
and Engineering a-Structural Materials Properties Microstructure and Processing 2003;
351:244-257.
[75] Mumm DR, Evans AG, Spitsberg IT. Characterization of a cyclic displacement
instability for a thermally grown oxide in a thermal barrier system. Acta Materialia
2001; 49:2329-2340.
[76] Meier GH, Pettit FS. The Oxidation Behavior of Intermetallic Compounds.
Materials Science and Engineering a-Structural Materials Properties Microstructure and
Processing 1992; 153:548-560.
[77] Kofstad P. High Temperature Corrosion. Amsterdam: Elsevier 1988.
[78] Sims CT, Stoloff NS, Hagel WC. Superalloys II. 1987.
[79] Balint DS, Hutchinson JW. An analytical model of rumpling in thermal barrier
coatings. Journal of the Mechanics and Physics of Solids 2005; 53:949-973.
[80] Cannon RM, Rhodes WH, Heuer AH. Plastic-Deformation of Fine-Grained
Alumina (Al2o3) .1. Interface-Controlled Diffusional Creep. Journal of the American
Ceramic Society 1980; 63:46-53.
[81] Cho JH, Harmer MP, Chan HM et al. Effect of yttrium and lanthanum on the
tensile creep behavior of aluminum oxide. Journal of the American Ceramic Society
1997; 80:1013-1017.
[82] Evans AG, Hutchinson JW, He MY. Micromechanics model for the detachment of
residually compressed brittle films and coatings. Acta Materialia 1999; 47:1513-1522.
[83] Evans AG, He MY, Hutchinson JW. Mechanics-based scaling laws for the
durability of thermal barrier coatings. Progress in Materials Science 2001; 46:249-271.
REFERENCES
PAGE 156
[84] Rhines FN, Wolf JS. Role of Oxide Microstructure and Growth Stresses in
High-Temperature Scaling of Nickel. Metallurgical Transactions 1970; 1:1701-&.
[85] Sarioglu S, Blachere JR, Pettit FS et al. The effects of reactive element additions,
sulfur removal, and specimen thickness on the oxidation behaviour of alumina-forming
Ni- and Fe-base alloys. High Temperature Corrosion and Protection of Materials 4, Pts
1 and 2 1997; 251-2:405-412.
[86] Lipkin DM, Clarke DR. Measurement of the stress in oxide scales formed by
oxidation of alumina-forming alloys. Oxidation of Metals 1996; 45:267-280.
[87] Ma Q, Clarke DR. Piezospectroscopic Determination of Residual-Stresses in
Polycrystalline Alumina. Journal of the American Ceramic Society 1994; 77:298-302.
[88] Hou PY, Paulikas AP, Veal BW. Growth strains in thermally grown Al2O3 scales
studied using synchrotron radiation. Jom 2009; 61:51-55.
[89] Schumann E, Sarioglu C, Blachere JR et al. High-temperature stress measurements
during the oxidation of NiAl. Oxidation of Metals 2000; 53:259-272.
[90] Dryepondt S, Porter JR, Clarke DR. On the initiation of cyclic oxidation-induced
rumpling of platinum-modified nickel aluminide coatings. Acta Materialia 2009;
57:1717-1723.
[91] Cho J, Wang CM, Chan HM et al. Role of segregating dopants on the improved
creep resistance of aluminum oxide. Acta Materialia 1999; 47:4197-4207.
[92] Sampath S, Schulz U, Jarligo MO, Kuroda S. Processing science of advanced
thermal-barrier systems. Mrs Bulletin 2012; 37:903-910.
[93] Kulkarni A, Wang Z, Nakamura T et al. Comprehensive microstructural
characterization and predictive property modeling of plasma-sprayed zirconia coatings.
Acta Materialia 2003; 51:2457-2475.
[94] Renteria AF, Saruhan B, Schulz U et al. Effect of morphology on thermal
conductivity of EB-PVD PYSZ TBCs. Surface & Coatings Technology 2006;
201:2611-2620.
[95] Guo HB, Vassen R, Stover D. Atmospheric plasma sprayed thick thermal barrier
coatings with high segmentation crack density. Surface & Coatings Technology 2004;
186:353-363.
[96] Guo HB, Murakami H, Kuroda S. Thermal cycling behavior of plasma sprayed
segmented thermal barrier coatings. Materials Transactions 2006; 47:306-309.
[97] Fauchais P, Montavon G. Latest Developments in Suspension and Liquid Precursor
Thermal Spraying. Journal of Thermal Spray Technology 2010; 19:226-239.
[98] von Niessen K, Gindrat M. Plasma Spray-PVD: A New Thermal Spray Process to
Deposit Out of the Vapor Phase. Journal of Thermal Spray Technology 2011;
20:736-743.
[99] Schulz U, Saruhan B, Fritscher K, Leyens C. Review on advanced EB-PVD
ceramic topcoats for TBC applications. International Journal of Applied Ceramic
Technology 2004; 1:302-315.
[100] Renteria AF, Saruhan B, Ilavsky J, Allen AJ. Application of USAXS analysis and
non-interacting approximation to determine the influence of process parameters and
ageing on the thermal conductivity of electron-beam physical vapor deposited thermal
barrier coatings. Surface & Coatings Technology 2007; 201:4781-4788.
[101] Schulz U, Terry SG, Levi CG. Microstructure and texture of EB-PVD TBCs
grown under different rotation modes. Materials Science and Engineering a-Structural
Materials Properties Microstructure and Processing 2003; 360:319-329.
[102] Vaidya A, Sirinivasan V, Streibl T et al. Process maps for plasma spraying of
yttria-stabilized zirconia: An integrated approach to design, optimization and reliability.
REFERENCES
PAGE 157
Materials Science and Engineering a-Structural Materials Properties Microstructure and
Processing 2008; 497:239-253.
[103] Kuroda S, Clyne TW. The Quenching Stress in Thermally Sprayed Coatings. Thin
Solid Films 1991; 200:49-66.
[104] Nakamura T, Liu YJ. Determination of nonlinear properties of thermal sprayed
ceramic coatings via inverse analysis. International Journal of Solids and Structures
2007; 44:1990-2009.
[105] Chi WG, Sampath S, Wang H. Microstructure-thermal conductivity relationships
for plasma-sprayed yttria-stabilized zirconia coatings. Journal of the American Ceramic
Society 2008; 91:2636-2645.
[106] Taylor TA. US Patent 5,073,433. 1989.
[107] Gray DM, Lau YC, Johnson CA et al. US Patent 5,830,586. 1996.
[108] Rampon R, Filiatre C, Bertrand G. Suspension plasma spraying of YPSZ coatings:
Suspension atomization and injection. Journal of Thermal Spray Technology 2008;
17:105-114.
[109] Delbos C, Fazilleau J, Rat V et al. Phenomena involved in suspension plasma
spraying part 2: Zirconia particle treatment and coating formation. Plasma Chemistry
and Plasma Processing 2006; 26:393-414.
[110] Hospach A, Mauer G, Vassen R, Stover D. Columnar-Structured Thermal Barrier
Coatings (TBCs) by Thin Film Low-Pressure Plasma Spraying (LPPS-TF). Journal of
Thermal Spray Technology 2011; 20:116-120.
[111] Miller RA. Oxidation-Based Model for Thermal Barrier Coating Life. Journal of
the American Ceramic Society 1984; 67:517-521.
[112] Hindam H, Whittle DP. Microstructure, Adhesion and Growth-Kinetics of
Protective Scales on Metals and Alloys. Oxidation of Metals 1982; 18:245-284.
[113] Sergo V, Clarke DR. Observation of subcritical spall propagation of a thermal
barrier coating. Journal of the American Ceramic Society 1998; 81:3237-3242.
[114] Hutchinson JW, Suo Z. Mixed-Mode Cracking in Layered Materials. Advances in
Applied Mechanics, Vol 29 1992; 29:63-191.
[115] Peng X, Clarke DR. Piezospectroscopic analysis of interface debonding in
thermal barrier coatings. Journal of the American Ceramic Society 2000; 83:1165-1170.
[116] Hutchinson JW, He MY, Evans AG. The influence of imperfections on the
nucleation and propagation of buckling driven delaminations. Journal of the Mechanics
and Physics of Solids 2000; 48:709-734.
[117] Mumm DR, Evans AG. On the role of imperfections in the failure of a thermal
barrier coating made by electron beam deposition. Acta Materialia 2000; 48:1815-1827.
[118] Karlsson AM, Hutchinson JW, Evans AG. A fundamental model of cyclic
instabilities in thermal barrier systems. Journal of the Mechanics and Physics of Solids
2002; 50:1565-1589.
[119] Tolpygo VK, Clarke DR. Surface rumpling of a (Ni, Pt)Al bond coat induced by
cyclic oxidation. Acta Materialia 2000; 48:3283-3293.
[120] Tolpygo VK, Clarke DR, Murphy KS. The effect of grit blasting on the oxidation
behavior of a platinum-modified nickel-aluminide coating. Metallurgical and Materials
Transactions a-Physical Metallurgy and Materials Science 2001; 32:1467-1478.
[121] Kim J, Dunn MG, Baran AJ et al. Deposition of Volcanic Materials in the Hot
Sections of 2 Gas-Turbine Engines. Journal of Engineering for Gas Turbines and
Power-Transactions of the Asme 1993; 115:641-651.
[122] Drexler JM, Gledhill AD, Shinoda K et al. Jet Engine Coatings for Resisting
Volcanic Ash Damage. Advanced Materials 2011; 23:2419-+.
REFERENCES
PAGE 158
[123] Levi CG, Hutchinson JW, Vidal-Setif MH, Johnson CA. Environmental
degradation of thermal-barrier coatings by molten deposits. Mrs Bulletin 2012;
37:932-941.
[124] Bons JP, Crosby J, Wammack JE et al. High-pressure turbine deposition in
land-based gas turbines from various synfuels. Journal of Engineering for Gas Turbines
and Power-Transactions of the Asme 2007; 129:135-143.
[125] Gledhill AD, Reddy KM, Drexler JM et al. Mitigation of damage from molten fly
ash to air-plasma-sprayed thermal barrier coatings. Materials Science and Engineering
a-Structural Materials Properties Microstructure and Processing 2011; 528:7214-7221.
[126] Zhu DM, Miller RA. Sintering and creep behavior of plasma-sprayed zirconia-
and hafnia-based thermal barrier coatings. Surface & Coatings Technology 1998;
108:114-120.
[127] Khan AN, Lu J, Liao H. Effect of residual stresses on air plasma sprayed thermal
barrier coatings. Surface & Coatings Technology 2003; 168:291-299.
[128] Lima CRC, Guilemany JM. Adhesion improvements of thermal barrier coatings
with HVOF thermally sprayed bond coats. Surface & Coatings Technology 2007;
201:4694-4701.
[129] Sadeghi-Fadaki SA, Zangeneh-Madar K, Valefi Z. The adhesion strength and
indentation toughness of plasma-sprayed yttria stabilized zirconia coatings. Surface &
Coatings Technology 2010; 204:2136-2141.
[130] Rabiei A, Evans AG. Failure mechanisms associated with the thermally grown
oxide in plasma-sprayed thermal barrier coatings. Acta Materialia 2000; 48:3963-3976.
[131] Markocsan N, Nylen P, Wigren J et al. Effect of Thermal Aging on
Microstructure and Functional Properties of Zirconia-Base Thermal Barrier Coatings.
Journal of Thermal Spray Technology 2009; 18:201-208.
[132] Zhao X, Liu J, Rickerby DS et al. Evolution of interfacial toughness of a thermal
barrier system with a Pt-diffused gamma/gamma ' bond coat. Acta Materialia 2011;
59:6401-6411.
[133] Wang JS, Evans AG. Measurement and analysis of buckling and buckle
propagation in compressed oxide layers on superalloy substrates. Acta Materialia 1998;
46:4993-5005.
[134] Qian G, Nakamura T, Berndt CC, Leigh SH. Tensile toughness test and high
temperature fracture analysis of thermal barrier coatings. Acta Materialia 1997;
45:1767-1784.
[135] Begley MR, Mumm DR, Evans AG, Hutchinson JW. Analysis of a wedge
impression test for measuring the interface toughness between films/coatings and
ductile substrates. Acta Materialia 2000; 48:3211-3220.
[136] Bahr HA, Balke H, Fett T et al. Cracks in functionally graded materials. Materials
Science and Engineering a-Structural Materials Properties Microstructure and
Processing 2003; 362:2-16.
[137] Hirakata H, Kitazawa M, Kitamura T. Fatigue crack growth along interface
between metal and ceramics submicron-thick films in inert environment. Acta
Materialia 2006; 54:89-97.
[138] Hofinger I, Oechsner M, Bahr HA, Swain MV. Modified four-point bending
specimen for determining the interface fracture energy for thin, brittle layers.
International Journal of Fracture 1998; 92:213-220.
[139] Thery PY, Poulain M, Dupeux M, Braccini M. Spallation of two thermal barrier
coating systems: experimental study of adhesion and energetic approach to lifetime
during cyclic oxidation. Journal of Materials Science 2009; 44:1726-1733.
REFERENCES
PAGE 159
[140] Yamazaki Y, Schmidt A, Scholz A. The determination of the delamination
resistance in thennal barrier coating system by four-point bending tests. Surface &
Coatings Technology 2006; 201:744-754.
[141] Zhou YC, Tonomori T, Yoshida A et al. Fracture characteristics of thermal barrier
coatings after tensile and bending tests. Surface & Coatings Technology 2002;
157:118-127.
[142] Charalambides PG, Lund J, Evans AG, McMeeking RM. A test specimen for
determining the fracture resistarim of bimaterial interfaces. Journal of Applied
Mechanics-Transactions of the Asme 1989; 56:77-82.
[143] Shinmi A. PhD thesis, Characterisation of the materials used in air plasma
sprayed thermal barrier coatings. University of Manchester 2011.
[144] Zhao Y, Shinmi A, Zhao X et al. Investigation of interfacial properties of
atmospheric plasma sprayed thermal barrier coatings with four-point bending and
computed tomography technique. Surface & Coatings Technology 2012;
206:4922-4929.
[145] Toraya H, Yoshimura M, Somiya S. Calibration Curve for Quantitative-Analysis
of the Monoclinic-Tetragonal Zro2 System by X-Ray-Diffraction. Journal of the
American Ceramic Society 1984; 67:C119-C121.
[146] Johnson CA, Ruud JA, Bruce R, Wortman D. Relationships between residual
stress, microstructure and mechanical properties of electron beam physical vapor
deposition thermal barrier coatings. Surface & Coatings Technology 1998; 108:80-85.
[147] Howard SJ, Tsui YC, Clyne TW. The Effect of Residual-Stresses on the
Debonding of Coatings .1. A Model for Delamination at a Bimaterial Interface. Acta
Metallurgica Et Materialia 1994; 42:2823-2836.
[148] Suo ZG, Hutchinson JW. Interface Crack between 2 Elastic Layers. International
Journal of Fracture 1990; 43:1-18.
[149] Vasinonta A, Beuth JL. Measurement of interfacial toughness in thermal barrier
coating systems by indentation. Engineering Fracture Mechanics 2001; 68:843-860.
[150] Choi SR, Hutchinson JW, Evans AG. Delamination of multilayer thermal barrier
coatings. Mechanics of Materials 1999; 31:431-447.
[151] Evans AG, Hutchinson JW, Wei Y. Interface adhesion: Effects of plasticity and
segregation. Acta Materialia 1999; 47:4093-4113.
[152] Gaudette F, Suresh S, Evans AG et al. The influence of chromium addition on the
toughness of gamma-Ni/alpha-Al2O3 interfaces. Acta Materialia 1997; 45:3503-3513.
[153] Kim SS, Liu YF, Kagawa Y. Evaluation of interfacial mechanical properties
under shear loading in EB-PVD TBCs by the pushout method. Acta Materialia 2007;
55:3771-3781.
[154] Xu ZH, Yang YC, Huang P, Li XD. Determination of interfacial properties of
thermal barrier coatings by shear test and inverse finite element method. Acta Materialia
2010; 58:5972-5979.
[155] Choi SR, Zhu DM, Miller RA. Effect of sintering on mechanical properties of
plasma-sprayed zirconia-based thermal barrier coatings. Journal of the American
Ceramic Society 2005; 88:2859-2867.
[156] Porter DL, Evans AG, Heuer AH. Transformation-Toughening in
Partially-Stabilized Zirconia (Psz). Acta Metallurgica 1979; 27:1649-1654.
[157] Busso EP, Lin J, Sakurai S, Nakayama M. A mechanistic study of
oxidation-induced degradation in a plasma-sprayed thermal barrier coating system. Part
I: Model formulation. Acta Materialia 2001; 49:1515-1528.
REFERENCES
PAGE 160
[158] Limarga AM, Vassen R, Clarke DR. Stress Distributions in Plasma-Sprayed
Thermal Barrier Coatings Under Thermal Cycling in a Temperature Gradient. Journal of
Applied Mechanics-Transactions of the Asme 2011; 78.
[159] Clarke DR, Christensen RJ, Tolpygo V. The evolution of oxidation stresses in
zirconia thermal barrier coated superalloy leading to spalling failure. Surface &
Coatings Technology 1997; 94-5:89-93.
[160] V T, Andritschky M, Fischer W et al. Effects of deposition temperature and
thermal cycling on residual stress state in zirconia-based thermal barrier coatings.
Surface & Coatings Technology 1999; 120:103-111.
[161] Sohn YH, Kim JH, Jordan EH, Gell M. Thermal cycling of EB-PVD/MCrAlY
thermal barrier coatings: 1. Microstructural development and spallation mechanisms.
Surface & Coatings Technology 2001; 146:70-78.
[162] Tolpygo VK, Clarke DR. Wrinkling of alpha-alumina films grown by thermal
oxidation - I. Quantitative studies on single crystals of Fe-Cr-Al alloy. Acta Materialia
1998; 46:5153-5166.
[163] He MY, Evans AG, Hutchinson JW. The ratcheting of compressed thermally
grown thin films on ductile substrates. Acta Materialia 2000; 48:2593-2601.
[164] Karlsson AM, Evans G. A numerical model for the cyclic instability of thermally
grown oxides in thermal barrier systems. Acta Materialia 2001; 49:1793-1804.
[165] Karlsson AM, Levi CG, Evans AG. A model study of displacement instabilities
during cyclic oxidation. Acta Materialia 2002; 50:1263-1273.
[166] Sun YL, Zhang WX, Li JG, Wang TJ. Local stress around cap-like portions of
anisotropically and nonuniformly grown oxide layer in thermal barrier coating system.
Journal of Materials Science 2013; 48:5962-5982.
[167] Tolpygo VK, Clarke DR. Competition between stress generation and relaxation
during oxidation of an Fe-Cr-Al-Y alloy. Oxidation of Metals 1998; 49:187-212.
[168] Veal BW, Paulikas AP. Growth strains and creep in thermally grown alumina:
Oxide growth mechanisms. Journal of Applied Physics 2008; 104.
[169] Tolpygo VK, Clarke DR, Murphy KS. Evaluation of interface degradation during
cyclic oxidation of EB-PVD thermal barrier coatings and correlation with TGO
luminescence. Surface & Coatings Technology 2004; 188:62-70.
[170] Clarke DR. The lateral growth strain accompanying the formation of a thermally
grown oxide. Acta Materialia 2003; 51:1393-1407.
[171] Hsueh CH, Evans AG. Oxidation Induced Stresses and Some Effects on the
Behavior of Oxide-Films. Journal of Applied Physics 1983; 54:6672-6686.
[172] He J, Clarke DR. Determination of the Piezospectroscopic Coefficients for
Chromium-Doped Sapphire. Journal of the American Ceramic Society 1995;
78:1347-1353.
[173] Gong XY, Clarke DR. On the measurement of strain in coatings formed on a
wrinkled elastic substrate. Oxidation of Metals 1998; 50:355-376.
[174] Munro RG. Evaluated material properties for a sintered alpha-alumina. Journal of
the American Ceramic Society 1997; 80:1919-1928.
[175] Demasimarcin JT, Gupta DK. Protective Coatings in the Gas-Turbine Engine.
Surface & Coatings Technology 1994; 68:1-9.
[176] Brindley WJ, Miller RA. Tbcs for Better Engine Efficiency. Advanced Materials
& Processes 1989; 136:29-33.
[177] Vassen R, Kagawa Y, Subramanian R et al. Testing and evaluation of
thermal-barrier coatings. Mrs Bulletin 2012; 37:911-916.
REFERENCES
PAGE 161
[178] Eldridge JI, Spuckler CM, Martin RE. Monitoring delamination progression in
thermal barrier coatings by mid-infrared reflectance imaging. International Journal of
Applied Ceramic Technology 2006; 3:94-104.
[179] Heeg B, Tolpygo VK, Clarke DR. Damage Evolution in Thermal Barrier Coatings
with Thermal Cycling. Journal of the American Ceramic Society 2011; 94:S112-S119.
[180] Stock SR. Recent advances in X-ray microtomography applied to materials.
International Materials Reviews 2008; 53:129-181.
[181] Withers PJ, Preuss M. Fatigue and Damage in Structural Materials Studied by
X-Ray Tomography. Annual Review of Materials Research, Vol 42 2012; 42:81-103.
[182] Elliott JC, Dover SD. X-Ray Microtomography. Journal of Microscopy-Oxford
1982; 126:211-213.
[183] Sato T, Ikeda O, Yamakoshi Y, Tsubouchi M. X-Ray Tomography for
Microstructural Objects. Applied Optics 1981; 20:3880-3883.
[184] Klems H. X-Ray-Diagnosis of a Fracture of the Tibial Plateau - Tomography.
Chirurg 1983; 54:417-419.
[185] Kress JW, Feldkamp LA. X-Ray Tomography Applied to Nde of Ceramics.
Mechanical Engineering 1983; 105:91-91.
[186] Tolpygo VK, Clarke DR. On the rumpling mechanism in nickel-aluminide
coatings - Part I: an experimental assessment. Acta Materialia 2004; 52:5115-5127.
[187] Tolpygo VK, Clarke DR. On the rumpling mechanism in nickel-aluminide
coatings - Part II: characterization of surface undulations and bond coat swelling. Acta
Materialia 2004; 52:5129-5141.
[188] Guo HB, Cui YJ, Peng H, Gong SK. Improved cyclic oxidation resistance of
electron beam physical vapor deposited nano-oxide dispersed beta-NiAl coatings for
Hf-containing superalloy. Corrosion Science 2010; 52:1440-1446.
[189] Hong SJ, Hwang GH, Han WK et al. Effect of zirconium addition on cyclic
oxidation behavior of platinum-modified aluminide coating on nickel-based superalloy.
Intermetallics 2010; 18:864-870.
[190] Tawancy HM, Sridhar N, Abbas NM, Rickerby DS. Comparative performance of
selected bond coats in advanced thermal barrier coating systems. Journal of Materials
Science 2000; 35:3615-3629.
[191] Wang YQ, Sayre G. Factors affecting the microstructure of platinum-modified
aluminide coatings during a vapor phase aluminizing process. Surface & Coatings
Technology 2009; 203:1264-1272.
[192] Zhang Y, Lee WY, Haynes JA et al. Synthesis and cyclic oxidation behavior of a
(Ni, Pt) Al coating on a desulfurized Ni-base superalloy. Metallurgical and Materials
Transactions a-Physical Metallurgy and Materials Science 1999; 30:2679-2687.
[193] Birks N, Meier GH, Pettit FS. Introduction to the High-Temperature Oxidation of
Metals. Cambridge University Press 2006.
[194] Swadzba R, Hetmanczyk M, Sozanska M et al. Structure and cyclic oxidation
resistance of Pt, Pt/Pd-modified and simple aluminide coatings on CMSX-4 superalloy.
Surface & Coatings Technology 2011; 206:1538-1544.
[195] Haynes JA, Pint BA, Zhang Y, Wright IG. Comparison of the oxidation behavior
of beta and gamma-gamma ' NiPtAl coatings. Surface & Coatings Technology 2009;
204:816-819.
[196] Tolpygo VK, Clarke DR. Rumpling of CVD (Ni,Pt)Al diffusion coatings under
intermediate temperature cycling. Surface & Coatings Technology 2009;
203:3278-3285.
REFERENCES
PAGE 162
[197] Oliver WC, Pharr GM. An Improved Technique for Determining Hardness and
Elastic-Modulus Using Load and Displacement Sensing Indentation Experiments.
Journal of Materials Research 1992; 7:1564-1583.
[198] Anstis GR, Chantikul P, Lawn BR, Marshall DB. A Critical-Evaluation of
Indentation Techniques for Measuring Fracture-Toughness .1. Direct Crack
Measurements. Journal of the American Ceramic Society 1981; 64:533-538.
[199] Lesage J, Chicot D. Role of residual stresses on interface toughness of thermally
sprayed coatings. Thin Solid Films 2002; 415:143-150.
[200] Feinberg A, Perry CH. Structural Disorder and Phase-Transitions in Zro2-Y2o3
System. Journal of Physics and Chemistry of Solids 1981; 42:513-518.
[201] Clarke DR, Adar F. Measurement of the Crystallographically Transformed Zone
Produced by Fracture in Ceramics Containing Tetragonal Zirconia. Journal of the
American Ceramic Society 1982; 65:284-288.
[202] Ishigame M, Sakurai T. Temperature-Dependence of Raman-Spectra of Zro2.
Journal of the American Ceramic Society 1977; 60:367-369.
[203] Perry CH, Liu DW, Ingel RP. Phase Characterization of Partially Stabilized
Zirconia by Raman-Spectroscopy. Journal of the American Ceramic Society 1985;
68:C184-C187.
[204] Bouvier P, Lucazeau G. Raman spectra and vibrational analysis of nanometric
tetragonal zirconia under high pressure. Journal of Physics and Chemistry of Solids
2000; 61:569-578.
[205] Bouvier P, Djurado E, Lucazeau G, Le Bihan T. High-pressure structural
evolution of undoped tetragonal nanocrystalline zirconia. Physical Review B 2000;
62:8731-8737.
[206] Quintard PE, Barberis P, Mirgorodsky AP, Merle-Mejean T. Comparative
lattice-dynamical study of the Raman spectra of monoclinic and tetragonal phases of
zirconia and hafnia. Journal of the American Ceramic Society 2002; 85:1745-1749.
[207] Milman V, Perlov A, Refson K et al. Structural, electronic and vibrational
properties of tetragonal zirconia under pressure: a density functional theory study.
Journal of Physics-Condensed Matter 2009; 21.
[208] Lughi V, Clarke DR. Temperature dependence of the yttria-stabilized zirconia
Raman spectrum. Journal of Applied Physics 2007; 101.
[209] You JL, Jiang GC, Yang SH et al. Temperature dependence of the Raman spectra
and phase transition of zirconia. Chinese Physics Letters 2001; 18:991-993.
[210] Cai JG, Raptis C, Raptis YS, Anastassakis E. Temperature-Dependence of
Raman-Scattering in Stabilized Cubic Zirconia. Physical Review B 1995; 51:201-209.
[211] Zallen R, Slade ML. Influence of Pressure and Temperature on Phonons in
Molecular Chalcogenides - Crystalline As4s4 and S4n4. Physical Review B 1978;
18:5775-5798.
[212] Cai JG, Raptis YS, Anastassakis E. Stabilized Cubic Zirconia - a Raman-Study
under Uniaxial-Stress. Applied Physics Letters 1993; 62:2781-2783.
[213] Limarga AM, Clarke DR. Piezo-spectroscopic coefficients of tetragonal-prime
yttria-stabilized zirconia. Journal of the American Ceramic Society 2007;
90:1272-1275.
[214] Kourouklis GA, Liarokapis E. Pressure and Temperature-Dependence of the
Raman-Spectra of Zirconia and Hafnia. Journal of the American Ceramic Society 1991;
74:520-523.
[215] Hart TR, Aggarwal RL, Lax B. Temperature Dependence of Raman Scattering in
Silicon. Physical Review B-Solid State 1970; 1:638-&.
REFERENCES
PAGE 163
top related