Evaluation and characterisation of thermal barrier coatings

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.

http://www.materials.manchester.ac.uk/people/staff/profile/?ea=andrew.zadoroshnyj

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


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


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.


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.


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.


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].


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


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


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].


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


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].


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,


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]


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,


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


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.


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].


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.


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


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].


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


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.


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.


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].


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.


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


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.


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.


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


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.


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


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.


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:


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.


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


PAGE 54

2.14). Direct 3D observations of these imperfections by novel X-ray computed

tomography technique are obtained and shown in chapter 3.


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]


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


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.


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


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.


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.


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.


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.


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 :


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.


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.


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


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


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.


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.


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)


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


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.


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.


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


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.


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.


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.


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


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


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


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


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.


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.


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.


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.


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.


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]


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).


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.


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.


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


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.


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


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.


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.


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


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


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.


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


PAGE 103

characterisation in 3D, but more importantly, for understanding fatigue and damage

accumulation in materials.


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


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


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


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


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.


PAGE 108


PAGE 109


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


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.


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.


PAGE 113


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.


PAGE 115


PAGE 116


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).


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.


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*


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.


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.


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


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


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.


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]


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


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].


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


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.


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


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).


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.


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.


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.


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.


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.


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


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.


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.


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


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.


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


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


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.


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


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


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


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.


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.

http://www.faa.gov/news/press_releases/news_story.cfm?newsId=12439

http://www.rolls-royce.com/about/technology/gas_turbine_tech/introduction.jsp

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


[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.

http://www.doitpoms.ac.uk/tlplib/fuel-cells/sofc_electrolyte.php

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.


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


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


[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.


[161] Sohn YH, Kim JH, Jordan EH, Gell M. Thermal cycling of EB-PVD/MCrAlY

thermal barrier coatings: 1. Microstructural development and spallation mechanisms.


[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.


[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


[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.


[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


[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

Documents

Evaluation and characterisation of thermal barrier coatings