Vous êtes sur la page 1sur 201

UNIVERSITY OF

CAMBRIDGE
Department of Materials
Science and Metallurgy



Thermophysical Properties of
Plasma Sprayed Thermal Barrier
Coatings

Sofia A. Tsipas


St. Johns College

June 2005






Dissertation Submitted for the Degree of Doctor of Philosophy


-i-
Preface
This dissertation is submitted for the degree of Doctor of Philosophy at the University of
Cambridge. The work described in this document was carried out between September 2001 and
June 2005 under the supervision of Prof. T.W. Clyne in the Department of Materials Science and
Metallurgy at the University of Cambridge. This dissertation is the result of my own original work
and includes nothing which is the outcome of work done in collaboration except where specifically
indicated in the text and has not been submitted, either in part or in its entity, for a degree at any
other university. This dissertation is less 60,000 words in length.






Sofia Tsipas

St Johns College
Cambridge

June 2005







-ii-
Abstract
Since advanced aerospace and power generation gas turbine engines are now designed such that nickel
superalloy components operate at temperatures very close to their melting points, current strategies for performance
improvements are centred on thermal barrier coatings (TBCs). State-of-the-art TBC systems are composed of a
zirconia ceramic top coat (300-600m thick), deposited either by air plasma spraying (APS) or electron beam
assisted physical vapour deposition (EB-PVD) over a metallic bond coat (100m thick) deposited by vacuum
plasma spraying (VPS). The top coat acts as a thermal barrier, generating a temperature drop of up to ~300C,
while the bond coat provides corrosion and oxidation protection for the substrate and creates a rough surface

which
may promote adhesion of the top coat. Failure of TBCs usually occurs by buckling and spalling of the coating.
APS top coats, which are considerably cheaper to produce, are composed of overlapping splats oriented parallel to
the plane of the coating, with poor inter-splat bonding, many fine microcracks and pores. Therefore, they have very
low stiffness. This prevents large stresses, and hence large driving forces for spallation, from being generated in
the top coat. There is current interest in pushing the top coat surface temperature up to about 1300-1400C in
aeroengines, while the aim for power generation is around 1200C. During the course of this work several aspects
of the behaviour of plasma sprayed coatings were investigated.
The effect of composition, impurity level, microstructural anisotropy and thermal cycling on the sintering
behaviour was investigated. Sintering involves healing of microcracks and strengthening of inter-splat bonding
during prolonged heat treatment. Healing of microcracks and improved inter-splat bonding due to sintering raise
the coating stiffness. This stiffening, will raise the strain energy release rate associated with a given misfit strain
and hence make spallation more likely.
TBCs operate in service under high thermal gradients. Therefore, it is crucially important to know how
such conditions will affect the top coat properties. The sintering characteristics, stress distribution and phase
constitution in the top coat were found to be strongly dependant on the presence of a through-thickness thermal
gradient.
The effect of heat treatment on the thermal conductivity of the top coat was also investigated. A key
objective for TBCs is to maximise the temperature drop across the thickness of the top coat. This requires in turn
that the thermal conductivity of the top coat should be minimised and remain low during exposure to service
conditions. Healing of microcracks and improved inter-splat bonding, due to sintering in service conditions, will
increase the thermal conductivity. A novel steady state technique was used to monitor the changes in thermal
conductivity with heat treatment. The factors that play the dominant roles in reducing the thermal conductivity of
top coat were identified. The importance of thermal conduction through the pores in the top coat was highlighted.
The use of an analytical heat flow model showed that it should be possible to predict the conductivity from
measurable microstructural parameters.
There is an increasing interest in developing new top coat materials that will be more sinter resistant and
have lower thermal conductivity. Novel material compositions have been investigated with respect to their sintering
behaviour, microstructural and phase evolution. This work will make it possible to pinpoint to the key issues for
better top coat performance and give new insight for the development of new materials for their use in TBCs.

-iii-
Acknowledgements
I gratefully acknowledge the provision of funding for the research described in this dissertation by
Sulzer Inc., Cambridge European Trust and Isaac Newton Awards. Thanks to Prof. D.J. Fray for
the provison of office, laboratory and workshop facilities at the Department of Materials Science
and Metallurgy, University of Cambridge.
There are many people who have helped me in the production of this work, but I am most indebted
to my supervisor, Prof. Bill Clyne, from whom I have learned so much. I want to express my
sincere gratitude and appreciation for his guidance and expert advice over the past three years. His
commitment and great knowledge have been an inspiration to me throughout my PhD. I would also
like to thank him for making the lab such a peasant place to work.
Thanks are also due to Dr. Jason Doesburg, Keith Harrison, Clive Britton and Andrew Nicoll from
Sulzer Inc. for providing the spraying powders, and for many useful discussions as well as their
encouragement regarding this project.
Many thanks also go to the technical staff at the Department of Materials Science and Metallurgy
for their assistance, especially Paul Stawkes and Mike Brand in the workshop, for being the people I
could rely on and for always being so happy to see me. Special thanks must go to Kevin Roberts,
for his technical and practical expertise, which has been of great value to this work, as well as for
his good company. I must also thank Terry Mosdal, for his help with spraying, Robert Stearn for
his help with experimental set ups and Mary Vickers for her help with X-ray diffraction.
It is the people that are the soul of the Gordon Laboratory, and I would like to thank them all for
making it such a pleasant place to work (and for putting up with me in general): Thank you Dr.
Igor Golosnoy for overlooking my work, for the prroper serious science, for the vodka nights, for
hanging up and worrying about my well-being in general; Athina, thank you for always being there
for me and for your invaluable friendship; Katerina, thank you for being so sweet, for your
friendship, for all your ideas that made so many days in the lab unforgettable and for the Greek
flags; Jin for being unfortunate enough to sit next to me for
almost
three years, for all your help with
technical issues and learning to dance Greek dances so well, Debu, the guru, thank you for all your
stories, for cooking proper curry and demonstrating how to take care of drunk people; Shiladitya
(whos your daddy?) for finally resigning to the fact of being a lost case and for finding the best
way to tolerate Greek noise: use ear protectors (keep up the David Hasselhoff look!); Amaia-
empanada, for being the best person to carry on and take my place and desk (keep me updated on

-iv-
the Cambridge gossip) and for sharing the same addiction with me: S&C; James C (go home). for
his unparallel ability to give me useless presents and for being so charitable (it wouldnt have
worked anyway); Andrew for reminding me that whatever I do it is meant to go wrong and for his
lack of fashion sense; James D for being the man; Martin for being the other person, other than me,
working late in the lab; Chris for the healthy food option on Fridays; TomI for being so cool and
Jamie for his Greeks, Geeks, Freaks andBrits comment. I wish you all the best of luck. Also
some past members deserve a mention, Russell-Barri for the samosas and pub-quiz expertise, Katy
for the regular lab pub visits, Dave for all the plants and Alex, yeah yeah yeah, who spent a shorter
time in the lab but became a big part of it.
A thankyou is also due to many of my friends in Cambridge, Renne, Bernard, Aldo, Milja, Ivana,
Petchya, Tony, Napoleon, Maro, George, Marianna and Kostas for their much-valued friendship,
for making my time here so enjoyable and keeping me sane. A special thank you to Carlos for his
never-failing support, encouragement, understanding and patience; I couldnt have finished this
without you.
Last, but above all, I would like to thank my loving family, Mami, Papa, and Marili, for their
constant unconditional support, which has been incredibly important to me, and for their love and
encouragement throughout my whole time in Cambridge. It is to my father, Prof. Dimitris Tsipas,
that I wish to dedicate this thesis.


-v-
TABLE OF CONTENTS
Preface .............................................................................................................................................. i
Abstract ............................................................................................................................................ ii
Acknowledgements .......................................................................................................................... iii
TABLE OF CONTENTS ................................................................................................................. v
Nomenclature ................................................................................................................................... x
1 Introduction ................................................................................................................................ 1
2 Thermal Barrier Coatings for Gas Turbine Applications ..................................................... 3
2.1 The gas turbine engine ..................................................................................................... 3
2.1.1 Materials for Gas Turbines .......................................................................................... 4
2.2 Thermal Barrier Coating Systems .................................................................................... 6
2.2.1 Nickel Based Superalloys ............................................................................................ 6
2.2.2 Bond Coat .................................................................................................................... 9
2.2.3 Top Coat ..................................................................................................................... 10
2.3 Coating Systems ............................................................................................................. 11
2.3.1 EB-PVD Coatings ...................................................................................................... 11
2.3.2 Plasma Sprayed Coatings ........................................................................................... 11
2.4 The Plasma spraying Process ......................................................................................... 13
2.4.1 The Plasma Jet ........................................................................................................... 13
2.4.2 Coatings Microstructure and Properties ..................................................................... 14
2.4.3 Parameters Affecting the Process .............................................................................. 16
3 Review of Thermomechanical Behaviour of Thermal Barrier Coatings ............................ 18
3.1 Stresses in Plasma Sprayed Coatings ............................................................................. 18
3.1.1 Mechanisms of Stress Generation .............................................................................. 18
3.1.1.1 Quenching Stress ............................................................................................................................. 18
3.1.1.2 Differential Thermal Contraction. ................................................................................................... 20
3.1.1.3 Solid State Transformations ............................................................................................................ 21
3.1.2 Measurement of Residual Stresses ............................................................................. 21
3.1.2.1 Curvature Methods .......................................................................................................................... 21

-vi-
3.1.2.2 Diffraction Methods ........................................................................................................................ 23
3.1.3 Modelling of Residual Stresses .................................................................................. 24
3.2 Mechanical Properties of TBCs ..................................................................................... 28
3.2.1 Mechanical Behaviour of PS TBCs ........................................................................... 28
3.2.2 Prediction of Mechanical Properties of PS Materials ................................................ 32
3.3 The Thermally Grown Oxide ......................................................................................... 34
3.4 Creep of Bond Coat and Thermally Grown Oxide ........................................................ 35
3.4.1 Creep in the Bond coat ............................................................................................... 36
3.4.2 Creep of the TGO ....................................................................................................... 36
3.5 Sintering of the Top Coat ............................................................................................... 37
3.5.1 Sintering mechanisms in ceramics ............................................................................. 37
3.5.1.1 Solid State Sintering ........................................................................................................................ 37
3.5.1.2 Liquid Phase Sintering .................................................................................................................... 42
3.5.2 Sintering of Zirconia based top coats ......................................................................... 43
3.6 Phase changes in Top Coat ............................................................................................ 45
4 Review of Thermal Conduction in TBCs ............................................................................... 49
4.1 Thermal Conduction in Ceramics .................................................................................. 49
4.1.1 Phonon Conduction in Ceramics................................................................................ 49
4.1.2 Radiative Heat Transfer in Ceramics ......................................................................... 51
4.2 Methods for Measuring the Thermal Conductivity of Solids ........................................ 52
4.2.1 Steady State Methods ................................................................................................. 52
4.2.2 Dynamic Methods ...................................................................................................... 54
4.2.2.1 Transient Hot Strip .......................................................................................................................... 55
4.2.2.2 Transient Plane Strip (Hot Disk) ..................................................................................................... 56
4.2.2.3 Laser Flash Method ......................................................................................................................... 58
4.3 Thermal Conductivity of TBCs ...................................................................................... 61
4.4 Modelling of Heat Flow in Plasma-Sprayed TBCs ....................................................... 65
4.4.1 Morphological representation of PS TBCs. ............................................................... 66
4.4.2 Physical Assumptions ................................................................................................ 67
4.4.3 Numerical Model ....................................................................................................... 69

-vii-
4.4.4 Two Flux Region Analytical Model .......................................................................... 70
4.4.5 Predicted Conductivity ............................................................................................... 73
5 Experimental Procedures ........................................................................................................ 74
5.1 Characterization of Starting Materials ........................................................................... 74
5.1.1 Chemical Composition ............................................................................................... 74
5.1.2 Particle Size Distribution ........................................................................................... 74
5.1.3 Powder Morphology .................................................................................................. 76
5.1.4 Phase Constitution of Starting Powders ..................................................................... 78
5.2 Coating Production ........................................................................................................ 79
5.2.1 Substrate Preparation ................................................................................................. 79
5.2.2 Plasma Spraying ......................................................................................................... 80
5.3 Sample Preparation Procedure ....................................................................................... 82
5.3.1 Detaching the Top Coat ............................................................................................. 82
5.3.2 Metallographic preparation ........................................................................................ 83
5.4 Coating Characterisation ................................................................................................ 84
5.4.1 Surface roughness ...................................................................................................... 84
5.4.2 Scanning Electron Microscopy .................................................................................. 84
5.4.3 Dilatometry ................................................................................................................ 84
5.4.4 X-Ray Diffraction ...................................................................................................... 85
5.4.5 Porosity Measurements .............................................................................................. 87
5.4.6 Stiffness Measurement ............................................................................................... 88
5.4.6.1 Cantilever Bending .......................................................................................................................... 88
5.4.6.2 Nanoindentation .............................................................................................................................. 89
5.5 Heat Treatment of Thermal Barrier Coatings ................................................................ 90
5.5.1 Isothermal Heat Treatment......................................................................................... 90
5.5.2 Heat Treatment with a High Thermal Gradient ......................................................... 90
5.6 Measurement of Thermal conductivity .......................................................................... 91
5.6.1 Steady State Rig ......................................................................................................... 91
5.6.1.1 Experimental Set Up ....................................................................................................................... 91
5.6.1.2 Data Analysis .................................................................................................................................. 94

-viii-
5.6.1.3 Validation of Technique .................................................................................................................. 95
5.6.2 Hot Disk ..................................................................................................................... 97
5.6.3 Laser Flash ................................................................................................................. 99
6 Thermal Stability of Plasma-Sprayed Top Coats ............................................................... 101
6.1 Changes in Phase Constitution of Plasma-Sprayed Top Coats due to Heat Treatment
101
6.1.1 As-sprayed ............................................................................................................... 101
6.1.1.1 Yttria stabilized zirconia top coats ................................................................................................ 101
6.1.1.2 Dysprosia stabilized zirconia top coats ......................................................................................... 104
6.1.1.3 Yttria-lanthana stabilized zirconia top coats.................................................................................. 105
6.1.1.4 Ceria stabilized zirconia top coats ................................................................................................. 106
6.1.2 Isothermal Heat Treatment....................................................................................... 108
6.1.2.1 Yttria stabilized zirconia top coats ................................................................................................ 108
6.1.2.2 Dysprosia stabilized zirconia top coats ......................................................................................... 113
6.1.2.3 Yttria-lanthana stabilized zirconia top coats.................................................................................. 114
6.1.3 Heat Treatment Under a Thermal Gradient ............................................................. 115
6.2 Microstructural development under service conditions ............................................... 117
6.2.1 Isothermal Heat Treatment....................................................................................... 117
6.2.2 Heat Treatment Under a Thermal Gradient ............................................................. 120
6.3 Effect of Heat Treatment of Detached Top Coats on sintering behaviour ................... 122
6.3.1 Effect of Composition and Morphology on Sintering Characteristics ..................... 122
6.3.2 Effect of Phase Transformation on Measured Volume Changes ............................. 129
6.3.3 Effect of Thermal Cycling on Sintering Behaviour ................................................. 131
6.3.4 Effect of Sintering on Coating Porosity ................................................................... 132
6.4 Conclusions .................................................................................................................. 135
7 Thermomechanical Behaviour of Plasma Sprayed TBCs .................................................. 137
7.1 Effect of Top Coat Sintering on Mechanical Properties .............................................. 137
7.1.1 Cantilever Bending .................................................................................................. 137
7.1.2 Nanoindentation ....................................................................................................... 138
7.1.2.1 As-sprayed and After Isothermal Heat Treatment ......................................................................... 138
7.1.2.2 Heat Treatment Under a Thermal gradient .................................................................................... 139

-ix-
7.2 Microstructure and Properties of Thick As-sprayed TBCs .......................................... 139
7.2.1 The Effect of Substrate Temperature on Microstructure and Stiffness ................... 140
7.2.2 The Effect of Substrate Temperature on Residual Stresses ..................................... 145
7.3 Stresses in TBCs after heat treatment .......................................................................... 145
7.4 Conclusions .................................................................................................................. 147
8 Thermal Conduction in Plasma-Sprayed TBCs .................................................................. 148
8.1 Thermal Conductivity of Plasma sprayed TBCs .......................................................... 148
8.1.1 Steady state Rig ........................................................................................................ 148
8.1.2 Laser Flash Measurements ....................................................................................... 150
8.1.3 Hot Disk ................................................................................................................... 152
8.2 Effect of Powder Composition on Thermal Conduction ............................................. 153
8.3 The Effect of Heat Treatment on the Thermal conductivity of TBCs ......................... 155
8.4 Effect of Pore Conductivity on Overall Thermal Conductivity ................................... 157
8.5 Experimental and Predicted Thermal Conductivity ..................................................... 158
8.6 Conclusions .................................................................................................................. 160
9 Conclusions and Future Work .............................................................................................. 162
9.1 General Conclusions .................................................................................................... 162
9.2 Future work .................................................................................................................. 164
References ....................................................................................................................................... 166



x
Nomenclature
Symbol Units Parameter
(i) Roman Symbols
A m
2
cross-sectional area
a m lattice parameter
B Pa m K
-1
a constant, with a value of 2.5 10
-5
for air
c m lattice parameter
C mol concentration
C
p
J kg
-1
K
-1

specific heat capacity at constant pressure
C
v
J kg
-1
K
-1
specific heat at constant volume
D m
2

s
-1
diffusion coefficient
d m inter-bridge (or interpore) distance.
d
v
m the thickness of the pore
E Pa Young modulus.
h W m
-2
K
-1
interfacial thermal conductance,
h m thickness (subscripts s and d refer to the thickness of the substrate and the
deposit respectively).
I m
4
second moment of area.
J kg s
-1
mass flux
k
b
J K
-1
Boltzmanns constant, 1.38 10
-23
J K
-1
k W m
-1
K
-1
thermal conductivity
k
r
W m
-1
K
-1
thermal conductivity by photon conduction
k
ph
W m
-1
K
-1
thermal conductivity from phonon conduction
k
t
W m
-1
K
-1
total thermal conductivity
k
0
W m
-1
K
-1
total conductivity of the lamellae
k
p
W m
-1
K
-1
thermal conductivity of the pore
k
g
0
W m
-1
K
-1
normal conductivity of the gas
k
res
W m
-1
K
-1
effective thermal conductivity of the contact resistance region in TFR
model
k
1,
k
2
W m
-1
K
-1
thermal conductivities of the corresponding regions in TFR model
k
eff
W m
-1
K
-1
effective thermal conductivity,
k
true
W m
-1
K
-1
actual thermal conductivity,
L m length of the specimen.
L
o
m initial sample length
L
h
m thickness of the unit cell in numerical and TFR model
L
v
m height of unit cell in numerical and TFR model
l
p
m mean free path for scattering of phonons
l
pd
m mean free path for scattering of phonons by point defects

l
r
m mean free path for photon scattering

xi
m kg mass (subscripts air, liq and laq refer to measurements made in air,
measurements made when immersed in a liquid and the sample coated with
lacquer; subscripts g and total refer to the mass of the glass container, and the
total mass of the powder and the glass container respectively; superscripts uc
and c refer to uncoated and coated sample respectively)
M N m bending moment.
n - refractive index
n - integer.
P atm pressure
P N force (describing edge loading or applied loading)
p % porosity
p
f
- porosity fraction (by volume)
q W input/output power
Q W m
-2
heat flux (subscripts upper and lower refer to the heat flux in the upper and
lower substrate respectively)
r m neck radius of sintering particles
R ohms resistance (subscript 0 refers to resistance at time t = 0)
R m radius of sintering particles
S m
2
area of the contact region in TFR model
S
br
m
2

cross section area of the bridge.
S
tot
m
2
the total cross section area of the unit cell
T K or C temperature (subscripts s and m refer to the substrate temperature and the
melting temperature respectively).
t s time (subscript max refers to total time of transient experiment)
V m
3
volume
x m distance
y m displacement, relative to the interface, through the thickness of the beam

(ii) Greek Symbols
m
2
s
-1
thermal diffusivity
o K
-1
coefficient of thermal expansion.
K
-1
temperature coefficient for the electrical resistivity
m
-1
scattering coefficient
m probing depth (in Hot Disk technique)
- linear misfit strain.
L m the length change of a material
Ak m
-1
change in curvature.
AT K temperature drop/difference (subscripts c and i refer to temperature drop
across the coating and interfaces respectively)
Ax m sample thickness/ length
c - average axial component of residual strain (subscripts s and d refer those of
the substrate and the deposit respectively).

xii
k m
-1
curvature.
m
-1
absorption coefficient
m
-1
extinction coefficient
m
4
J
-1
s
-1
grain boundary mobility
m s
-1
the mean phonon velocity
v - Poisson ratio.
kg m
-3
density (subscripts d and th refer to the experimentally determined density of
the deposit material and the corresponding theoretical (bulk) density
respectively).
o W m
-2
K
-4
Stephen-Boltzmanns constant, 5.668610
-6
o Pa stress (subscripts s and d refer to the residual stresses in the substrate and
deposit respectively).
o
0
Pa quenching stress predicted by linear elastic behaviour.
o
q
Pa quenching stress.
o
y
Pa yield stress.
s
-1
frequency of phonon lattice waves
m
3
volume per lattice site

(iii) Acronyms and Abbreviations
APS atmospheric plasma spraying
BC bond coat
CSZ ceria stabilized zirconia
CTE coefficient of thermal expansion
DC direct current
DS directional solidification
DSZ dysprosia stabilized zirconia
EB-PVD electron-beam physical vapour deposition
EDX energy dispersive x-rays
LPPS low pressure plasma spraying
PS plasma sprayed
PSZ partially stabilized zirconia
SANS small angle neutron scattering
SEM scanning electron microscope
TBC thermal barrier coating
TC top coat
TET turbine entry temperature
TGO thermally grown oxide
VPS vacuum plasma spraying
XRD x-ray diffraction
YLaSZ yttria-lanthana stabilized zirconia
YSZ yttria stabilized zirconia
Chapter 1: Introduction
1
1 Introduction
Gas turbine engines operate under severe thermal, mechanical, and chemical conditions. Turbine
technology strongly depends on development of suitable material systems that can operate in such
harsh environment. Successive modification of alloy compositions lead to the introduction of
nickel based superalloys in jet design. Further efficiency improvements of aero engines as well as
land-based gas turbines require increased operating temperatures. However, limits on the operating
temperature are imposed by the onset of melting of the superalloys. The use of thermal barrier
coatings allows increases in the operating temperature without any increase in the temperature to
which the metal is exposed, it can reduce the amount of cooling required whilst maintaining the
operating temperature of the turbine. Alternatively, improved durability and reliability will result
by reducing the component temperatures. Coatings are thus commonly applied to these components.
Thermal barrier coatings (TBCs) for use on turbine blades consist of a two-layer system. A
relatively thin bond coat is first deposited and a second layer (top coat) is then sprayed onto the
first. The bond coat is typically a MCrAlY metal alloy (e.g. Ni-22wt%Cr-10wt%Al-1wt%Y). The
purpose of the bond coat is to protect the substrate from oxidation and hot corrosion and also
improve the adhesion between the substrate and the top coat. The industry standard material used
as a top coat is zirconia, usually stabilised with yttria. During service a thin thermally grown oxide
(TGO) layer forms as a result of bond coat oxidation.
TBCs are commonly applied by plasma spraying. In plasma spraying, powder particles are injected
into a plasma flame where they are melted and accelerated towards the substrate. The molten
particles impact the substrate, spread out, cool and contract to form lenticular splats which built up
the coating. Advantages of the plasma process are that it is relatively inexpensive, offers high
deposition rates, and is capable of producing durable and reproducible coatings. The application of
thermal barrier coatings for combustor parts, augmentors, vane nozzles, transition pieces and
stationary turbine components today is widely practiced. However, the use of plasma-sprayed
TBCs as a design element in hot rotating parts of the turbine engine is still quite limited. The
application of TBCs as an integral part of such components requires reliable and predictable TBC
performance.
Failure of TBCs usually occurs by buckling and spalling of the coating. Plasma sprayed coatings
are relatively stable against spallation driven by residual or externally applied stresses. The main
reason for this is that the zirconia has a very low stiffness. This prevents large stresses, and hence
Chapter 1: Introduction
2
large driving forces for spallation, from being generated in the top coat. This low stiffness and high
strain tolerance is largely a consequence of the presence of many fine microcracks and pores in the
zirconia layer. In addition to this beneficial effect on the stiffness, these features reduce the thermal
conductivity, which is also desirable. Recent work has shown that another significant change
during prolonged heating is substantial stiffening of the top coat. This arises from sintering
processes, which can raise the stiffness.
The present research aims to investigate the thermophysical properties of plasma sprayed top coats
in TBCs. Together with the widely used zirconia-stabilized with yttria, other novel compositions,
such as dysprosia-stabilized zirconia, yttria-lanthana-stabilized zirconia and ceria-stabilized zirconia
are explored, as well the effect of impurities. The research is focused on the changes that occur in
the phase constitution, microstructure, stiffness and thermal conductivity upon exposure to high
temperatures. The differences in the sintering behaviour and thermal conductivity of differently-
stabilized zirconia top coats are investigated. In-service TBCs operate, in the presence of a
through-thickness thermal gradient, and interest is focused in particular on the behaviour of top
coats under such thermal gradients.

Chapter 2. TBCs for Gas Turbine Applications

3
2 Thermal Barrier Coatings for Gas Turbine
Applications
2.1 The gas turbine engine
A turbine is a rotary engine that converts the energy of a moving stream of fluid into mechanical
energy. The fluid can be water, steam, or gas. The basic element in a turbine is a wheel or rotor
with paddles, propellers, blades, or buckets arranged on its circumference in such a fashion that the
moving fluid exerts a tangential force that turns the wheel and imparts energy to it. This
mechanical energy is then transferred through a drive shaft to operate a machine, compressor,
electric generator, or propeller. Turbines are classified as hydraulic or water turbines, steam
turbines, or gas turbines. Most modern passenger and military aircrafts are powered by gas turbine
engines, which are also called jet engines. Todays turbine-powered generators produce most of the
world's electrical energy. There are several different types of gas turbine engines, but all have some
parts in common. The main parts of the gas turbine are the fan or inlet, the compressor, the
combustor or burner, the turbine and the exhaust or nozzle (see Figure 2.1).
All gas turbine have an inlet or fan to bring air into the engine. The incoming air enters the
compressor, which is made up from a series of aerofoil blades attached to a shaft. These control the
airflow rate and, as air goes through progressively smaller areas, the air pressure increases. The
pressure of the air can rise up to 40 atmospheres. This increases the potential energy of the air. The
high-pressure air is forced into the combustion chamber. In the combustor, air is mixed with fuel
that is sprayed into the air stream by a ring of fuel injectors. The fuel is generally kerosene,
propane or natural gas. The air/fuel mixture is ignited and burned at temperatures in a range of
1450-2000
o
C [1, 2 , 3, 4] to generate the maximum possible heat energy. The highest temperature
in a turbine engine occurs in the combustion chamber followed by the inlet temperature at the first
blades of the turbine. The combustion produces a high temperature, high pressure gas stream that
enters and expands through the turbine section. The turbine is an intricate array of alternate
stationary and rotating aerofoil-section blades. As hot combustion gas expands through the turbine,
it spins the rotating blades. The turbine is linked by a shaft to the blades in the compressor and the
inlet fan. Some of the energy from the rotating blades is used to drive compressor and the inlet fan
via the shaft. For power generation the rest of the energy is used to spin a generator to produce
electricity. For jet engines in aircrafts the combustion gases are expelled through the exhaust or
nozzle. The resultant expelled exhaust gas generates the thrust which is the force that moves a jet
Chapter 2. TBCs for Gas Turbine Applications

4
aircraft through the air. In the case of the turbine used in a power plant, exhaust gases are vented
through an exhaust pipe or are used for a heat exchanger either to extract the heat for some other
purpose or to preheat air before it enters the combustion chamber.
2.1.1 Materials for Gas Turbines
The first gas turbine operated at a temperature which was based on the limitations imposed by the
high-temperature strength of the existing materials. The need for development of new materials
with strength at high temperature arises from the fact that temperature and pressure directly
influence the efficiency of the gas turbine cycle. Both thermal efficiency and power output are
influenced by the maximum cycle temperature and pressure ratio. Broadly speaking, the maximum
temperature has the more important effect on power output whereas the pressure ratio has greater
influence on the thermal efficiency [5].
Higher pressure ratios and higher maximum temperatures are used in aeroengines than in industrial
gas turbine. Turbine gas entry temperatures today are in the range of 1100-1350
o
C for industrial
gas turbines and around 1500
o
C for advanced aircraft engines [1, 3, 4, 6-9]. Adjacent to the blades
there is layer of cooler gas that keeps the surface of the blades at lower temperatures. The
operational life requirements are different for aircrafts and industrial gas turbines. Industrial gas
turbines operate for substantially longer cycles and times between overhauls. A typical aero engine
operates for 5000-10,000 cycles between overhauls and cycle times are between 0.5-13 hours.
Industrial gas turbines operate 30,000-50,000 hrs between overhauls and cycle times are 100-
2500 hours [7, 9-11]. In aircraft gas turbine the weight must be kept low and the shorter life allows
higher stresses and temperatures to be used than would be permissible in long-life industrial plants.
From the aspect of material selection, different design parameters apply for different parts of the gas
turbine. Compressor materials must have adequate fatigue strength, in order to be able to withstand
the centrifugal and vibrational stresses. Additionally, they must have satisfactory impact resistance,
since there is some likelihood of the injection of foreign objects. The initial compressor stages
operate at relatively low pressures and temperatures, so that light alloys or composites can be used.
However, for the later stages, where temperature and pressures are higher, creep and oxidation
resistance become important and stronger alloys are necessary.
In the combustor, the fuel is ignited and the air/fuel temperature is in a range of 1450-2000
o
C [1, 2 ,
3, 4]. The combustor walls need to be continuously cooled in order to maintain the temperature at
safe levels (below1100C) [10]. Materials in the combustor experience high stress conditions.
However they need to maintain their strength at high temperatures. Furthermore oxidation
Chapter 2. TBCs for Gas Turbine Applications

5
resistance and thermal shock resistance are important. Failures in the combustor chamber occur
mainly by cracking and buckling at hot spots.

Figure 2.1 Pressure andtemperature of the gas as it passes through the various stages of a jet engine. (after
[1]).
The turbine is the most demanding environment from the materials performance point of view.
Materials operate in an environment with a combination of very high temperature and high stresses
(Figure 2.1). Turbine blades exist in a more hostile environment than compressor blades. They
must have the maximum possible resistance to creep and thermal fatigue, and high tensile strength,
since they experience high centrifugal stresses. Materials with high strength, ductility, high thermal
conductivity and low thermal expansion provide the best resistance to thermal fatigue. Also,
turbine blades must have good oxidation resistance, since they are in contact with the high
temperature combustion products with high oxygen content. In industrial turbines particularly, hot
corrosion resistance is also an important requirement. Since turbine entry temperatures are so high,
often exceeding the melting point of the metal used, blades must be actively cooled. A cooling
system is used in which cool air is continuously circulated through the blade, to keep it relatively
cool. The active cooling of blades establishes high thermal gradients between the inside and the
surface. This introduces variations in the physical properties and thermal stresses. Other required
Chapter 2. TBCs for Gas Turbine Applications

6
properties for turbine blades include good impact resistance and microstructural stability over long
periods of time. Complex materials have been developed for use in turbine blades [12, 13]. More
details on the superalloys used in turbine blades is given in 2.2.1.
Turbine disks are also rotating components in the turbine and operate under high temperatures and
centrifugal stresses. Therefore these materials must have very high tensile strength, good creep
resistance and good thermal fatigue resistance. Stator blades, and nozzle guides vanes are not
exposed to the influence of centrifugal stresses and hence the primary material requirements are
oxidation, hot corrosion and erosion resistance.
2.2 Thermal Barrier Coating Systems
Thermal barrier coatings (TBCs) in engines have been used experimentally since the 1960s and by
the 1980s they were being used in certain commercial gas turbine engines in low risk regions [14,
15]. TBCs offer benefits in terms of the performance and efficiency of gas turbines, allowing lower
metal temperatures, reduced requirements for active cooling, higher turbine entry temperatures and
longer component lifetimes. The application of thermal barrier coatings for combustor parts,
augmentors, vane nozzles, transition pieces and stationary turbine components today is widely
practiced [16]. However, the use of TBCs in hot rotating parts of the aircraft turbine engine usually
is not integrated in the design, i.e. TBCs are used to lower the superalloy temperature and thus
prolong the life-time of the blade, but failure of the TBC is not critical. If TBCs are to fulfill their
full potential, reliability is necessary.
A typical TBC system consists of (i) the top coat (TC), a porous ceramic layer that acts as the
insulator, (ii) the bond coat (BC), an oxidation-resistant metallic layer between the substrate and the
TC and (iii) the superalloy or other material substrate that carries the structural load (see Figure 2.2).
TBCs are usually produced either by electron-beam physical vapor deposition (EB-PVD) or by
plasma spraying (PS).
2.2.1 Nickel Based Superalloys
Superalloys have been described as alloys developed for elevated temperature service, usually
based on group VIIIA elements, where relatively severe mechanical stressing is encountered and
where high surface stability is frequently required [17]. Superalloys are used in aircraft, marine,
industrial and vehicular gas turbines and also in space vehicles, rocket engines, nuclear reactor,
submarines, steam power plants and petrochemical equipment. However, their most important use
is in the gas turbine aeroengines.
Chapter 2. TBCs for Gas Turbine Applications

7
High pressure turbine blades operate under severe mechanical and thermal environments. Turbine
blades are made from nickel based superalloys. Wrought turbine blade compositions are primarily
based on 80% Nickel 20% Chromium alloys. Progressively higher amounts of Aluminium,
Titanium and associated compositional modifications increased the temperature capability of the
alloys [2]. Nickel based superalloys can currently operate up to homologous temperatures of 0.8T
m

(about 1100
o
C) [8, 18].

T
e
m
p
e
r
a
t
u
r
e


~1100
o
C
~600
o
C
~1300
o
C
Substrate
Thermo-mechanical loading
Bond Coat
Oxidation protection/bonding
Top Coat
Thermal Insulation
Ni-base superalloys
ZrO
2
+(6-8)%Y
2
O
3

Hot Gases
MCrAlY or Aluminides
Coolant

Figure 2.2 Schematic of typical structure and qualitative temperature profile through a TBC.

Figure 2.3 Increase in operational temperature of turbine component made possible by alloy development,
manufacturing technology and TBCs [18].

Chapter 2. TBCs for Gas Turbine Applications

8
Nickel provides a face-centered cubic(FCC) lattice, which has a high modulus, 5 independent slip
systems and low diffusivity for substitutional solutes. The formation of coherent gamma prime
precipitates and carbides lead to strengthening. Advanced turbine blade materials have a relatively
high volume fraction of gamma prime (Ni
3
Al), around 70% [19]. Aluminium forms a stable
alumina scale during exposure to high temperature and also strengthens the alloy by forming
gamma prime. Other alloying elements are added for different reasons. Chromium may form an
adherent oxide scale which also protects against oxidation. Titanium can replace some of the
Aluminium in gamma prime as well as forming titanium carbides, which also contribute to
precipitate strengthning. Elements such as C, B, Zr and Hf provide grain boundary strengthening in
polycrystalline alloys[2, 12].
Superalloys were conventionally produced by casting methods. However, superalloys produced by
casting methods often did not exhibit consistent creep properties. This problem led to the
development of directional solidification (DS), which produces castings with grains aligned in the
direction of maximum stress and few grain boundaries normal to this direction. Directional
solidification is achieved by ensuring that the heat during solidification of the casting is removed in
a direction parallel to the desired growth direction, while a liquid/solid interface perpendicular to
the solidification direction is maintained. Several techniques have been successfully developed in
which the production cost of DS casting is no greater than that of conventional castings. DS
resulted in a significant increase in the creep strength of these superalloys, relative to
conventionally cast alloys, and led to an increase in the temperature capabilities of superalloys.
A further important development in the processing of superalloys was the production of single
crystal castings. This technology was achieved by extending the DS technology and adding a
simple modification to the casting mould, either a crystal selection device or a seed crystal.
Single crystals have improved creep properties relative to DS superalloys. The complete absence of
grain boundaries also slightly increases the solidus temperature of the alloy. This permits higher
temperature solution heat treatment, which is beneficial because it leads to reduced chemical
segregation, more uniform gamma prime distribution and consequently increased temperature
capability. Furthermore, elements such as C, B, Zr, and Hf, which are normally necessary in
polycrystalline castings to produce grain boundary strengthening, can be eliminated in single crystal
superalloys [12, 20].


Chapter 2. TBCs for Gas Turbine Applications

9
2.2.2 Bond Coat
The bond coat protects the underlying substrate from oxidation and improves adhesion between the
ceramic and the metal. Oxidation occurs due to oxygen reaching the bond coat by diffusion through
the lattice of the top coat and permeation through the pores [21]. The yield and creep characteristics
of the bond coat are thought to be significant for the performance of the TBC system.
Commonly used bond coats can be divided in two categories: MCrAlY (where M= Co or Ni or both)
and Pt-modified aluminides [14]. These coatings were developed for use as protective coatings
against oxidation and hot corrosion [22]. When exposed to an oxidizing environment, they form a
stable dense alumina layer in preference to other oxides. This alumina, often termed the thermally
grown oxide (TGO) prevents further attack of the underlying material, due to its low oxygen
diffusivity and its good adherence. MCrAlY bond coats are usually deposited by low-pressure
plasma spraying and consist of two phases (-NiAl and either -Ni solid solution or Ni
3
Al) [23].
Small amounts of Y are added in order to improve TGO adherence [24]. Yttrium additions have
been found to inhibit void formation at the TGO/BC interface. In addition, Y-rich oxide protrusions
are formed in the oxide that mechanically peg the oxide to the alloy. Furthermore, yttrium has the
effect of decreasing the grain size of the TGO and thus raising its mechanical strength [25].
Pt-modified aluminides are usually fabricated by electroplating a thin Pt layer on the superalloy and
then aluminizing by chemical vapor deposition or pack cementation. These coatings usually consist
of a single-phase- with Pt in solid solution. Platinum additions improve the spallation resistance
of conventional aluminide coatings. However, the mechanisms by which this occurs are not fully
understood.
Optimum adhesion between the bond coat and the top coat is attained differently in plasma sprayed
and EB-PVD coatings. In plasma sprayed coatings, it is achieved by mechanical interlocking of the
two interfaces, so the surface roughness of the bond coat is an important parameter [26]. In contrast,
EB-PVD coatings achieve maximum durability when applied to a smooth (preferably polished)
surface, free of absorbed gases or loose oxides. Asperities in the BC/TGO interface are thought to
serve as nucleation sites for cracks that cause coating spallation when they coalesce [23, 27].
MCrAlY bond coats creep at temperatures above 800
o
C [28-30]. At this temperature, stresses in the
BC are relieved and it is non-load bearing. The creep behaviour of the BC can have a significant
influence on the stress state of the TBC and thus on the failure mechanisms. More detailed
discussion on the influence of creep on the failure mechanism of TBCs is given in section 3.3.
Chapter 2. TBCs for Gas Turbine Applications

10
2.2.3 Top Coat
The top coat provides thermal insulation for the underlying substrate. The specifications for this
coating require a material that combines low thermal conductivity and a coefficient of thermal
expansion (CTE) that it is as similar as possible to that of the substrate, so that generation of
stresses during thermal cycling can be minimised. The preferred material for this application is
zirconia. Zirconia may exist as three solid phases, that are stable at different temperatures [31]. At
temperatures up to 1200
o
C, the monoclinic phase (m) is stable. Zirconia transforms from the
monoclinic to the tetragonal phase (t) above 1200
o
C and above 2370
o
C to the cubic phase (c).
Transformation from m to the t phase has an associated volume decrease of 4% [32]. To prevent
catastrophic cracking as a result of the volume changes accompanying the tm transformation,
which occurs at temperatures within the range of the working environment in gas turbines,
stabilizers are added to the zirconia. These stabilize zirconia into its cubic or tetragonal phases.
Early attempts used MgO to stabilize zirconia in its cubic state, by adding 25 wt% MgO [33].
However, during heat treatment the zirconia reverts to its monoclinic form and the stabilizing oxide
precipitates out from solid solution, affecting the thermal conductivity [34]. Zirconia can be fully
stabilised to its cubic phase by adding 20 wt% yttria. However, such fully stabilised zirconia
coatings perform very poorly in thermal cycling tests [35]. Typically 7-9wt% yttria is used to
partially stabilise zirconia, although other stabilizers have been used as well. Other stabilizers
include CaO, MgO, CeO
2
Sc
2
O
3
[36-38]. Rare earth dopants such as Dy
2
O
3
and Yb
2
O
3
have also
generated some interest [39, 40]. The basic criteria for the selection of a suitable stabiliser include a
suitable cation radius, similar to that of zirconium, and a cubic crystal structure. In spite of the
addition of a stabilizer in order to ensure phase stability of the top coat, phase changes in the top
coat might still be induced during service. This is described in more detail in section 3.6.
An important aspect of the performance of top coat material is its sintering behaviour. After
prolonged heating during service, sintering of the top coat can occur. This will result in healing of
the microcracks and pores that will in turn reduce the strain tolerance of the coating and increase the
likelihood for spallation. More detailed discussion of the influence of top coat sintering on the
lifetime of TBCs is given in the section 3.5.

Chapter 2. TBCs for Gas Turbine Applications

11
2.3 Coating Systems
2.3.1 EB-PVD Coatings
The EB-PVD process takes place in an evacuated chamber. A high energy electron beam is used to
heat and vaporize the coating material. The vapor formed contains the coating atoms and molecules
[7]. To ensure the correct stoichiometry of the ceramic top coat, a controlled amount of oxygen is
bled into the chamber. The preheated substrate is inserted in the vapor cloud and the vapor
condenses onto the surface of the substrate, forming a coating, usually at a deposition rate of
4-8 m min
-1
[6]. EB-PVD is used both for the deposition of bond coats and ceramic top coats.
Ceramic coatings generated by EB-PVD have a columnar microstructure, after an initial thin region
of dense ceramic (Figure 2.4). This structure exhibits high lateral strain tolerance. In addition, they
exhibit a good surface finish which is beneficial for aerofoil applications, where a low coefficient of
friction is desirable. Thermal conductivity of partially stabilized zirconia (PSZ) TBCs produced by
EB-PVD is typically in the range of 1.5-1.8 W m
-1
K
-1
[41], which is higher than values for plasma
sprayed TBCs. This is mainly due to the characteristic columnar structure of EB-PVD coatings, that
does not hinder thermal heat transfer in the through-thickness direction, since voids and interspaces
are mainly aligned parallel to the heat flux direction.
2.3.2 Plasma Sprayed Coatings
A characteristic of all thermal spray processes is a highly concentrated power source, to which the
coating material is fed in the form of powder, wire or rod. The coating material is melted and
accelerated to the substrate, forming the coating. The coating is formed of many overlapping splats,
solidifying one after another and locking one to another. Due to the high kinetic energy of the
droplets, the splats spread over the substrate, forming a pancake.
Plasma spraying is described in detail section 2.4. It is widely used for the production of TBCs.
The use of PS TBCs in the hot rotating section of aeroengines is still very limited and if they are to
fulfill their full potential, their reliability must be improved [42].
PS TBCs have the necessary strain tolerance required for most of the applications in which such
coatings are currently applied. This is largely a consequence of the presence of many fine
microcracks and pores in the microstructure, which results in low stiffness (Figure 2.4). This low
stiffness prevents large stresses from being generated in the top coat. The thermal conductivity of
plasma sprayed coatings range from 0.5-1.4 W m
-1
K
-1
[43], which is lower than corresponding
values for EB-PVD coatings. The microstructure of PS TBCs exhibits pores and grain boundaries
Chapter 2. TBCs for Gas Turbine Applications

12
aligned perpendicular to the direction of heat flux (see Figure 2.4). Grain boundaries and pores
hinder heat transfer [44]. The shape and orientation of porosity with respect to the heat flux are
more critical factors than the total amount of porosity for the thermal conductivity of PS coating
[45]. More details on the thermal conductivity of TBCs are given in section 4.3. EB-PVD coatings
offer benefits over PS coatings in terms of the erosion resistance [46-48]. In PS coatings, the
erosion occurs in the form of removal of the mechanically bonded splats by the erosive material.
Since intersplat porosity is already present, the energy required for this process is low.
The low cost associated with the PS process compared to EB-PVD makes PS TBCs the more
attractive options for many gas turbine components such as combustion chambers, nozzle guides
and abradable seals [16]. However, applications that require excellent strain tolerance, good surface
finish and erosion resistance, such as in aerofoils and aero-gas turbines, EB-PVD coatings will be
favoured [49].

Figure 2.4 Schematic illustrations of the pore morphology of (a) a plasma spray (PS) deposited coating and (b)
an electron beam physical vapour deposited (EB-PVD) coating [50].

Chapter 2. TBCs for Gas Turbine Applications

13
2.4 The Plasma spraying Process
2.4.1 The Plasma Jet
Plasma Spraying, first conducted by Reinecke in 1939, was advanced in the late 50s by several
other scientists [51]. Since then, it has become increasingly sophisticated and is nowadays widely
used in surface technology.
The plasma spraying gun consists principally of two electrodes.
Figure 2.5 shows a schematic of the plasma spray gun, with the thoriated tungsten cathode inside
the water-cooled copper anode. A gas, commonly a mixture of argon and hydrogen, is injected into
the annular space between the two. To start the process, a DC electric arc is stuck between the two
electrodes. The electric arc produces gas ionisation, i.e. gas atoms lose electrons and become
positive ions. Electrons move with high velocity to the anode, while ions move to the cathode. On
their way, electrons and atoms collide with neutral gas atoms and molecules. Hence, the electric arc
continuously converts the gas into a plasma (a mixture of ions and electron of high energy). The
plasma is on average, electrically neutral and characterized by a very high temperature [52]. The
kinetic energy of the plasma (mostly carried by free electrons) is converted into thermal energy
during collisions between ions, electrons and atoms. In this way, the plasma is capable of producing
temperatures up to approximately 10
4
K [53]. The hot gas exits the nozzle of the gun with high
velocity.
Powder material is fed into the plasma plume. The powder particles are melted and propelled by
the hot gas onto the surface of the substrate [54]. When individual molten particles hit the substrate
surface, they form splats by spreading, cooling and solidifying. These splats then incrementally
build the coating.
Plasma plumes exhibit radial temperature gradients. Whereas particles that pass through the central
core of the plasma tend to be melted, superheated or even vaporised, particles that flow near the
periphery may not melt at all. This will affect the final structure of the coating, which may contain
partially molten or unmelted particles. Voids, oxidised particles and unmelted particles can appear
in the coating (see Figure 2.6). These effects may be desirable, or they may be unwanted, depending
on the requirements of the coating.
Chapter 2. TBCs for Gas Turbine Applications

14
Water-cooled
copper anode
Tungsten
insert
Water-cooled
copper cathode
Insulating back
plate with angled
gas channels
Plasma
gases
Arc
Powder
Plasma

Figure 2.5 Schematic of Plasma spray gun [55].
2.4.2 Coatings Microstructure and Properties
Plasma sprayed coatings are built up particle by particle. Molten droplets arrive one at a time and
impact with the underlying material, which consist of previously solidified droplets (splats), and
there they solidify. Hence, the microstructure of the PS coatings is different from that of most other
materials [56]. Plasma sprayed coatings have a non-homogeneous, layered structure, consisting of
splats with a pancake-like shape. Also, partially molten particles, oxides and voids are present in a
typical PS structure (see Figure 2.6).
Porosity levels in plasma sprayed ceramic coatings are generally in the range 3-20% [57]. Porosity
might be in the form of inter splat porosity (fine gaps between one lamella and another) or larger
irregularly shaped void which result from incomplete conformation of splats to the topography of
the impact site. For thermal barrier coatings, high porosity may be desirable, since the thermal
conductivity of the coating is decreased by the presence of porosity [58]. Certain applications
require low porosity and this can be achieved by controlled spraying conditions so as to raise
droplet velocities.
Plasma sprayed coatings also have different mechanical properties from corresponding bulk
material. Strength values of the coatings are usually only a small fraction of the values of
corresponding dense materials. The Youngs modulus of metallic coatings obtained by plasma
spraying has been found to be of the order of 1/3 that of bulk material [59], though ceramic coatings
give lower modulus. For PSZ, typical values of Youngs modulus for as-sprayed coatings are
Chapter 2. TBCs for Gas Turbine Applications

15
10-40 GPa in comparison to 210 GPa for dense zirconia [60-65]. Elastic anisotropy has also been
observed in plasma sprayed deposits which is attributed to the preferred orientation of planar crack-
like defects for PS zirconia [66].
The thermal conductivity of PS coatings is usually lower than corresponding values for the bulk
material. The coatings exhibit pores and planar defects which inhibit heat transfer [67]. The
coefficient of thermal expansion, however, is not influenced by the coating morphology since pores
and voids do not contribute to expansion. This was confirmed by Kuroda and Clyne [59], who
measured the CTE of plasma-sprayed coatings and found it to be close to that of bulk material.
Schwingel et al [68] also reported an average value of ~10 10
-6
K for YSZ at room temperature,
similar to that of sintered YSZ. Ahmaniemi et al [69] observed linear thermal expansion of plasma
sprayed 8wt%-YSZ in the temperature range 50-1000C and reported a CTE of ~ 9.9 10
-6
K
-1
.
For plasma sprayed ceria stabilized zirconia the same was found to be ~ 10.8 10
-6
K
-1
. The CTE
of plasma sprayed Lanthanum hexaaluminate was found to be lower (7.7-9.3 10
-6
K
-1
) than YSZ
coatings (10-11.1 10
-6
K
-1
) in the temperature range 100-1300C[70].

Figure 2.6. Schematic diagram showing typical microstructural features in plasma sprayed coatings.

Adhesion of plasma sprayed coatings to the underlying substrate is achieved mainly by mechanical
interlocking. Therefore, the surface roughness of the substrate often plays a role in interfacial
adhesion. Solid-state diffusive bonds occurring at a molecular level is also believed to play a minor
role in the adhesion of PS coatings. These non-mechanical bonding mechanisms are not yet fully
understood [54].
During plasma spraying, stresses are generated in the coating as consequence of the rapid
solidification of the molten droplets (quenching stresses) and during subsequent cooling as a result
Chapter 2. TBCs for Gas Turbine Applications

16
of mismatch of the thermal expansion coefficients of substrate and coating. More detailed analysis
of these stresses is presented in section 3.1.
2.4.3 Parameters Affecting the Process
In plasma spraying, there are many parameters that need to be adjusted and therefore the task of
optimisation and process control is complex. In general, the properties of plasma sprayed coatings
depend on processing variables [71-73]. These include stand-off distance, plasma gas composition,
plasma power and the powder injection rate. However, these parameters are subject to certain
limitations, arising from the requirement that the powder can be melted. The effects of these
variables are now briefly considered.
The power absorbed in the plasma depends on the arc current and the voltage drop across the
electrodes. Also, the way the electric power is applied (i.e. with or without transferred arc, positive
or negatively charged) is an important factor. The power supply affects the temperature and
velocity of the plasma flame and therefore influences the melting and acceleration of the powder
particles. The plasma arc power affects the spraying efficiency. The heat available in the arc, and
the arc temperature for a set of spraying conditions, will be controlled by the arc power, which in
turn depends on the gas mixture. Low power may result in incomplete melting of the injected
powders whereas too high a power can cause vaporization of the particles. The optimum power
settings for spraying will depend on the thermophysical properties of the material sprayed.
The stand-off distance between the nozzle and the substrate influences the deposition of the powder
particles. Depending on the distance, the residence time of the powder particles in the plasma
plume, differs. This results in different velocities and temperatures of the powder particles when
impacting the substrate.
The powder variables that affect the spraying process depend mainly on the physical properties of
the material sprayed and the characteristics of the powder. The physical properties include the
melting and vaporization temperature of the bulk material, its heat transfer properties (specific heat,
thermal conductivity and latent heat of fusion) and its density. Density will affect the particle
trajectory, for a given particle size and plasma plume characteristics. Powder characteristics
include particle size and shape, which are important from the heat-transfer viewpoint. Larger
particles may be incompletely melted, whereas small particles may vaporize. Therefore, for a given
set of plasma conditions, there is an optimum particle size. In general, powders with a narrow range
of particle sizes are preferred.
Chapter 2. TBCs for Gas Turbine Applications

17
Plasma spraying offers technical advantages over other coating processes [74]. It can deposit metals,
ceramics or combinations of these, since the high plasma temperatures permit the spraying of
materials with high melting points. In addition, high particle velocities result in generating
microstructures with relatively low porosity and high bond strength.
Economically, most important is atmospheric plasma spraying (APS). It is often recommended,
provided oxidation of the coating material is not an issue, e.g. in cases of oxide ceramics or where
oxidation to a certain degree does not harm the performance of the sprayed coating. In cases where
oxidation of the coating has to be suppressed, the production can take place in an inert or protective
environment, provided a spray chamber is available. However this requires greater initial capital
investment.
Chapter 3. Review of Thermomechanical behaviour of TBCs

18

3 Review of Thermomechanical Behaviour of Thermal
Barrier Coatings
The most common failure mode of TBCs is spallation of the top coat (TC), usually by cracking
along the interface between the thermally grown oxide (TGO) and the bond coat and/or cracking in
the interface between TGO and TC. There are different factors that affect the thermo-mechanical
stability of TBCs. Significant factors are the stress state in the TC, the TGO thickness and the
interfacial adhesion between BC/TGO/TC amongst others. A brief description of the properties of
TBCs and the induced changes during service conditions will be presented here.
3.1 Stresses in Plasma Sprayed Coatings
Residual stresses in plasma sprayed coatings have been extensively studied. They can result in the
deformation of coated pieces and contribute to cracking and spallation of the coating. In this
section, an overview of the mechanisms of generation of residual stresses is presented. The
techniques used in order to measure residual stresses are assessed and numerical modelling
approaches for the prediction of residual stresses are described.
3.1.1 Mechanisms of Stress Generation
3.1.1.1 Quenching Stress
During coating formation, stresses are generated, the so called deposition or quenching stresses.
These stresses are produced when a molten droplet of the coating material impacts the cold
substrate, spreads and solidifies. As the molten splat strikes the substrate, it spreads and rapidly
loses heat through conduction to the underlying material. When it solidifies, its thermal contraction
is restricted by this material and thus a tensile stress is generated. It has been shown that the
magnitude of the quenching stress also depends on substrate temperature [59], although it is
independent of the substrate material [75]. Assuming perfect interfacial bonding and elastic
behaviour of the splat, the maximum value that the quenching stress can reach is given by [59]:
Chapter 3. Review of Thermomechanical behaviour of TBCs

19
( )
1
m s
q d d
d
T T
E

=

o o
v
(3.1)
where
d
is the coefficient of thermal expansion of the deposit, E
d
is Youngs modulus for the
deposit, v
d
is Poissons ratio for the deposit and T
m,
T
s
are the deposit melting temperature and
substrate temperature respectively. This value is commonly much larger than the yield stress; thus
plastic flow and other stress relaxation mechanisms may operate. These mechanisms are illustrated
in Figure 3.1. Stress relaxation mechanisms are different for metallic and ceramic coatings. For
metallic coatings, the yield stress is often low and the temperature of the substrate material can
often be a considerable fraction of the melting temperature of the coating material, so that creep and
yielding are common stress relaxation mechanisms. On the other hand, ceramic coatings have high
melting points and therefore do not readily relax by creep or yielding. Ceramic coatings, however,
commonly relax stress by microcracking.

Figure 3.1 Quenching stresses and relaxation mechanisms [59]..

Quenching stresses have been measured experimentally by Kuroda and Clyne [59]. The results
confirm that quenching stresses are usually much lower than the maximum theoretical value
(typically ~1 GPa). For metallic materials, the quenching stress can be quite high, especially for
materials with high yield and creep strength. For ceramics and other brittle materials, the
quenching stress is usually relatively small, as a consequence of microcracking (Table 3.1).

Chapter 3. Review of Thermomechanical behaviour of TBCs

20
Property Bond Coat
(CoNiCrAlY)
ZrO
2
-8wt% Y
2
O
3
Substrate
(Nimonic 80A)
CTE () 10
-6
K
-1

13.0 (298 K)
17.1 (773 K)
10.5 (283 K)
11.5 (811 K)
13.0 (298 K)
16.6 (773 K)
Thickness (mm) 0.1 0.3 10
Quenching Stress
(MPa)
115 5 -
Table 3.1 Quenching stress in plasma sprayed deposits [76].
3.1.1.2 Differential Thermal Contraction
The coefficient of thermal expansion is in general different for the substrate and the sprayed
material. During plasma spraying, the temperatures of the substrate and deposit are often high.
Therefore, during subsequent cooling of the sample, the mismatch in thermal expansion properties
will generate residual stresses. This mismatch in the coefficients of thermal expansion (CTE) will
also give rise to stress whenever the specimen is heated and cooled.
In addition to the magnitude of the thermal expansion mismatch, the stress state will depend upon
the thermal history, as well as the clamping arrangement during the spraying process (whether the
specimen is rigidly clamped or is allowed to bend). Assuming elastic behaviour and a thick
substrate (compared to the coating), such that the strain is accommodated by the coating alone, the
bi-axial residual stress in the coating (assuming transverse isotropy) can be estimated from the
expansivities of the substrate and the coating. Consider two bonded plates cooled from a stress-free
state through a temperature interval AT. A misfit strain Ac=AoAT will be generated and stress
generated in the coating as a result of the constraint of such strain:
) 1 (
) (
d
d s d
d
T E
v
o o
o

A
= (3.2)
where E is Youngs modulus, is the coefficient of thermal expansion, v is Poisson ratio and the
subscripts d and s refer to the deposit and substrate respectively. However, the assumption made
above in not always valid and properties of the substrate material may also be significant in the
generation of residual stresses (particularly for thin substrates).
In addition, large thermal gradients can arise during the spraying process, which may generate
stresses. These also tend to cause curvature during spraying. Thus the sample may not be in a
stress-free state when cooling commences. Evidently, the details of the generation of stress due to
differential thermal expansion can be complex and often numerical modelling is required.
Chapter 3. Review of Thermomechanical behaviour of TBCs

21
Typically, ceramics have low CTEs, compared to most metallic materials, and hence large
differential thermal contraction stresses can be produced. Residual stresses present in ceramic
coatings are therefore governed by stresses generated due to differential thermal contraction, since
quenching stresses in ceramics are in general low.
The generation of residual stresses has been thoroughly investigated for thermal barrier coatings
(TBCs) and in particular for partially stabilized zirconia (PSZ) coatings [77-80]. A gradient in the
stress state throughout the thickness of the coating has commonly been reported, with the surface
stress being more tensile and the stress at the interface more compressive [77]. The residual
stresses in the as-deposited top coat changes from tensile to compressive with increasing substrate
temperature [80, 81].
3.1.1.3 Solid State Transformations
Solid state transformations may occur during spraying and in such cases they may have an effect on
the stress field of the coating. Volumetric changes associated with solid state transformations will
induce misfit strains in the coating. A solid state transformation causing an increase in volume, will
induce compression, whereas a decrease in volume will cause tension. One such process that may
result in the generation of stresses is oxidation. This is particularly important in the case of spraying
metallic materials, such as the bond coat of TBCs, where oxidation is thought to play a significant
role in determining the coatings lifetime. To minimise oxidation effects during spraying, the
process can be carried out at low pressure (Low Pressure Plasma Spraying, LPPS). The mechanism
of bond coat oxidation and its significance for coating lifetimes is discussed in section 3.3. Solid-
state transformations may occur during annealing of zirconia top coats used in thermal barrier
coatings. This effect is discussed in section 3.6.

3.1.2 Measurement of Residual Stresses
There has been extensive research on experimental methods for measuring residual stresses in
sprayed coatings, but there is no entirely satisfactory method to achieve this. A brief overview is
given below.
3.1.2.1 Curvature Methods
Stresses produced during spraying generate a misfit strain. Applying elastic beam theory, this misfit
strain will cause an unbalanced bending moment in the sample. If the specimen is free to bend, it
Chapter 3. Review of Thermomechanical behaviour of TBCs

22
will acquire a curvature in order to minimize the strain energy. The curvature of the specimen is
hence a source of information on its stress state.
The curvature retained by the sample after the spraying process has been utilized by Hobbs and
Reater [82, 83] in order to deduce information on the stress state. However this method relies on a
single curvature measurement and there are many ambiguities associated with it. Kuroda et al [84,
85] were amongst the first to perform continuous measuring of curvature during the spraying
process and this has proved to be a powerful technique. This was realized by attaching a pair of
light contacting knife edges to a strip-shaped substrate and a contacting displacement meter at the
rear surface of the substrate. This technique was employed to calculate the quenching stress for
different materials and substrate temperatures [86, 87]. Stoneys equation [88] was used for
estimating the quenching stress:
( )
2
6 1
s s
q
s d d
E h
h h
k
o
v
c
=
c
(3.3) (Stoney equation)
where E is the Youngs modulus, h refers to thickness, is the Poisson ratio, k is the curvature and
the subscripts s and d refer to substrate and deposit respectively.
To ensure that the substrate temperature was kept constant during spraying, a pair of air jets was
provided by a pair of nozzles attached to the plasma torch. For the vacuum plasma spraying (VPS)
process a non-contacting method was employed [59, 87]. The coating material was sprayed onto
substrates of the same material. In this way, no thermal residual stress between the two layers is
generated, since there was no mismatch in the CTE for the two layers. Thus, the post spraying
curvature of the sample can be directly linked to the quenching stresses associated with the process.
The equation used for the calculation of the quenching stress is given below [89]:
1.25
6
d
s s s d
s
q
d
E
E h h h
E
h
k
o
(
| |
( +
|
(
\ .

= (3.4)
in which the variables are as described for (3.3). Further development in the curvature-monitoring
technique was achieved by Gill and Clyne [90, 91]. A non-contacting method of curvature
measurement was employed with the aid of a video camera. The video camera records deflection of
the free end of the sample during spraying and subsequent cooling. The information recorded can
be transformed into curvature data. This method has the advantage of eliminating any concern about
interference of the previously used instruments with the free bending of the sample and a wide
Chapter 3. Review of Thermomechanical behaviour of TBCs

23
range of temperatures can be examined. The technique has allowed validation of a numerical
process model described in section 3.1.3.
3.1.2.2 Diffraction Methods
Spacing between lattice planes can be determined using diffraction methods. Residual stresses will
induce strains that will alter these spacings [92]. This principle is depicted schematically in Figure
3.2. Changes in lattice parameters, and thus the strain, can be obtained from a single Braggs
measurement [93, 94]. If the elastic constants of the material are known, the corresponding stress
can be calculated. In practice, the sin
2
technique is more commonly used. This technique involves
measurement of the shift in a peak as a function of the angle of the specimen (Figure 3.2).
Diffraction methods have the advantage of being non-destructive. On the other hand, they have a
limited penetration depth and are sensitive to sample geometry and surface roughness, which can
present a problem for sprayed coatings, where the surface roughness is often large [92]. Most
diffraction techniques can only give information about the stress state near the surface. To obtain a
through-thickness profile of the residual stress, successive layers need to be removed. This
procedure will itself induce further residual stresses and affect the measurement. Furthermore, the
non-destructive nature of the technique is lost.


Figure 3.2 Schematic diagram of the XRD technique for measurement of residual stresses [95].

The most common diffraction techniques used for the purpose of measuring residual stresses are
X-ray diffraction, synchrotron X-ray diffraction and neutron diffraction [96]. Penetration depths for
X-ray diffraction are normally in the range 10-40 m [97] which is often in the order of the surface
roughness of the coating. This introduces systematic errors in the measurements. Synchrotron
o
o
normal to surface
and bisector
180
o
-2u
o
o
180
o
-2u
normal to
surface
bisector

Chapter 3. Review of Thermomechanical behaviour of TBCs

24
X-ray diffraction has higher beam intensity and lower divergence, facilitating through-thickness
profiling [97, 98]. However, penetration depths still remain in the order of 60 m. Neutron
diffraction allows much greater penetration depths, and so is more suitable for stress profiling of
thick samples [92, 99]. Scardi et al [77, 78, 100] have investigated all three aforementioned
diffraction techniques for evaluation of the residual stress state in sprayed coatings. Their work
highlights the importance and usefulness of neutron diffraction as a technique for full through-
thickness of residual stress measurement.
3.1.3 Modelling of Residual Stresses
Several researchers have used numerical models to predict stresses caused by the spraying process
[101-103]. Modelling of residual stress generation is a complex task, and many parameters need to
be considered. Understanding of the mechanisms of stress generation is critical. Accurate
knowledge of the thermal and elastic properties of the substrate and the deposit is essential. These
properties are often temperature dependent. Accurate prediction of the parameters which define the
heat and mass flux received by the specimen is also required. Furthermore, the time dependent
nature of the thermal field during spraying has to be considered and included in the model. Finally
the effect of creep should also be incorporated.
Assumptions of temperature invariant elastic properties [89, 104] or use of bulk material properties
[83] will introduce large errors in the modelling of residual stresses. Moreover, ignoring the time-
dependent thermal profiles [105] and ignoring the effects of quenching stress [83, 106] in the
generation of residual stresses can lead to errors.
A model developed by Clyne, Gill and Tsui [103, 107, 108] was used in the present work for the
prediction of residual stresses. The basic features of this model are presented in this section.
The model is based on a finite difference formulation, using a fully implicit solution scheme. The
heat flux from the plasma gun and the mass flux from the droplets determine the temperature rise in
deposit and substrate. Convective heat flux from the plasma gun and droplet mass flux were
apportioned radially symmetric Gaussian distributions, taking into account the movement of the
spraying gun. The gun movement is specified and thus the gun position in relation to the selected
modelling point can be determined at any time. The deposition is modelled by thickening the top-
most element, as splats are added, and generating a new element when the thickness exceeds the
standard element thickness (see Figure 3.3). The interface between the coating and the substrate is
placed at an element boundary.
Chapter 3. Review of Thermomechanical behaviour of TBCs

25
An equal biaxial stress-state with no through-thickness stresses, is assumed. Perfect bonding and
perfect heat conductance at the substrate-deposit interface are also assumed. Temperatures are
mapped into a mesh used for stress calculations. The temperature dependence of CTE, thermal
conductivity and specific heat capacity are taken into account. The mechanical behaviour for
metallic materials was assumed to be elastic-perfectly plastic, with temperature-dependent Youngs
modulus. The effect of creep is also simulated. For ceramic materials, the effect of microcracking
is simulated by introducing a critical stress, beyond which microcracking would inhibit further
increases in stress.
Curvature and in-plane stresses are predicted after each time step. For the calculation of stresses
and curvature, the concept of a relaxed (stress-free) element width is used (see Figure 3.3). The
complete stack of stress elements is allowed to expand or contract laterally in order to set the
summation of lateral forces to zero (force balance). For equilibrium conditions, the requirement of
a zero overall bending moment also needed to be met (moment balance). Therefore the resulting
curvature was calculated. The algorithm for the model is depicted in Figure 3.4.


Calculate gun position
Determine mass transfer and thickness increment
Is a new element required?
t = t
i

t = t
i
+t
Mass Input


Figure 3.3 (a) Deposition process as simulated by the model [109], (b) Schematic representation of the calculation
of stress using the relaxed width method. Upon cooling from temperature T
1
to T
2
, (I) substrate and coating contract
by differing amounts, due to the CTE mismatch (II). Substrate and coating elements are then constrained to have
the same width (III) (i.e. application of the force balance with no bending allowed). Finally a moment balance is
applied, such that the curvature with the least associated strain energy is adopted (IV) [76].

The numerical model also has the option of depositing additional coating layers over a coating
system which is already residually stressed, by using the files generated during the previous
deposition processes as starting files. Heat treatment of the sprayed specimen can also be simulated.
The starting point for this calculation is the generated file after the deposition process has been
(I)
(IV) (III)
(II)
Chapter 3. Review of Thermomechanical behaviour of TBCs

26
completed. Heat treatment conditions can be isothermal, i.e. the whole sample being at any
arbitrary temperature, or a thermal gradient can be simulated. A thermal gradient is simulated by
using a stationary heat flux with no mass flux, to the top of the substrate. Cooling at the rear surface
is achieved by decreasing the ambient temperature or raising the rear face heat transfer coefficient.
In this way, a steady-state heat flux is established that sets up the desired thermal gradient across the
coating. This feature of the model can be utilized for the calculation of the thermal conductivity of a
produced coating. A thermal gradient through a coating can be established and monitored
experimentally throughout the thickness of the coating. By knowing the top surface and rear
temperatures, the thermal gradient can be simulated using the model. There are no arbitrary
adjustable parameters, except the quenching stress (which can be measured separately). Sintering
of the top coat under service conditions can raise its stiffness and hence reduce its strain tolerance.
Sintering mechanisms of the top coat are discussed in more detail in section 3.5. The numerical
model has been adapted to allow prediction of residual stress development incorporating the effects
of top coat sintering during heat treatment.
Chapter 3. Review of Thermomechanical behaviour of TBCs

27


Figure 3.4 Flow chart illustrating model algorithm [76].
Chapter 3. Review of Thermomechanical behaviour of TBCs

28
3.2 Mechanical Properties of TBCs
The mechanical behaviour of dense ceramics is well known. Ceramics typically have high tensile
strength and fail in a brittle fashion. Ceramics behave differently in tension and compression and
their compressive strength is usually much higher than their tensile strength. The presence of flaws,
such as pores or inclusions, is known to affect the mechanical behaviour of ceramics. The
mechanical properties of TBCs have been studied extensively by many researchers [60, 63, 110-
112]. In this section a brief overview of the mechanical behaviour of PS TBCs and the attempts to
predict the mechanical response are presented.
3.2.1 Mechanical Behaviour of PS TBCs
Literature values show substantial scatter for the elastic modulus of APS YSZ top coats, varying
from a few GPa to values close to those of dense ceramics. An overview of the different values
reported in the literature, and the method employed for the measurement, is given in Figure 3.5.
This is because the mechanical properties of PS deposits depend very strongly on their
microstructure, especially on the porosity and the interlamellar contacts. These features depend in
turn on the processing parameters and initial powder morphology [68].
The evaluation method employed to measure the elastic properties should always be specified.
Values of the effective Youngs modulus may differ considerably for the same coating using
different evaluation methods. Measurements performed by microindentation or nanoindentation
with sharp pyramid (Berkovich) tips give values approaching that of pore-free zirconia [60]. This is
due to the fact that with such techniques the region undergoing deformation is relatively small and
should be devoid of features such as splat boundaries and microcracks, which are responsible for
the low values obtained by other techniques in which greater regions are examined. Siebert et
al[113] carried out measurements using instrumented depth-sensing micro-indentation with a
Vickers diamond pyramid indenter and reported Youngs Modulus of as-sprayed YSZ in the range
of 94-146 GPa. Due to the relatively small area of the indentations, the influence of large pores and
microcracks on the Youngs modulus is not taken in consideration using this technique.
Nevertheless, increases in the Youngs modulus due to sintering of annealed TBCs were detected.
The changes were accompanied by a changing microstructure. The Knoop indentation method has
also been used to measure the elastic modulus of YSZ plasma sprayed coatings. The indenter in
this method is an elongated diamond pyramid. Leigh et al [114] reported a modulus of 446 GPa
for an as-sprayed YSZ coating using Knoop indentation, Zhu et al [112] about 70 GPa and Vasquez
et al [115] reported values in the range of 25-56 GPa. In spherical indentation the radius of the
Chapter 3. Review of Thermomechanical behaviour of TBCs

29
indenter is usually large enough (2-3 mm) so that the indentation covers a suitable range of
deformed material with several representative microstructural features such pores, cracks and splats.
Wallace et al [111] reported values for the Youngs modulus in the range of 22-38 GPa for as-
sprayed YSZ coatings measured with spherical indentation and Eskner et al [63] a value of
384 GPa for a YSZ top coat.
0
20
40
60
80
100
120
140
160
180
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Y
o
u
n
g
'
s

M
o
d
u
l
u
s

(
G
P
a
)


c
a
n
t
i
l
e
v
e
r
[
6
0
]

s
p
h
e
r
i
c
a
l

i
n
d
e
n
t
a
t
i
o
n

[
1
1
1
]

s
p
h
e
r
i
c
a
l

i
n
d
e
n
t
a
t
i
o
n

[
6
3
]

k
n
o
o
p

i
n
d
e
n
t
a
t
i
o
n

[
1
1
5
]

k
n
o
o
p

i
n
d
e
n
t
a
t
i
o
n
[
1
1
4
]

t
e
n
s
i
l
e
[
1
1
6
]

k
n
o
o
p

i
n
d
e
n
t
a
t
i
o
n
[
1
1
2
]

m
i
c
r
o
i
n
d
e
n
t
a
t
i
o
n
[
1
1
3
]

n
a
n
o
i
n
d
e
n
t
a
t
i
o
n
[
6
0
]

m
i
c
r
o
i
n
d
e
n
t
a
t
i
o
n

[
1
1
7
]

m
i
c
r
o
i
n
d
e
n
t
a
t
i
o
n
[
1
1
7
]

u
l
t
r
a
s
o
n
i
c
s
[
1
1
5
]

m
i
c
r
o
i
n
d
e
n
t
a
t
i
o
n

[
1
1
7
]

f
o
u
r

p
o
i
n
t

b
e
n
d
i
n
g

[
6
8
]


Figure 3.5 Measured Youngs modulus data of as sprayed YSZ coatings from literature.

Methods such as cantilever bending [60] and four point bending [68], have been used to measure
macroscopically the elastic properties of the plasma sprayed coatings. The values for the elastic
modulus measured by these techniques are usually lower than those measured by indentation
techniques and arguably more representative for understanding the behaviour of these coating in
service. Thompson et al [60] measured the Youngs modulus of as-sprayed detached top coats by
Chapter 3. Review of Thermomechanical behaviour of TBCs

30
cantilever bending at 102 GPa. Schwingel [68] reported values in the range of 2-20 GPa for as-
sprayed YSZ top coats measured by four point bending.
PS coatings have different Youngs moduli along the directions parallel and perpendicular to the
surface, with higher moduli observed in the direction perpendicular to the surface [114, 117, 118].
Duan et al [117] reported a Youngs modulus of attached YSZ coatings in the direction parallel to
the substrate of 30 GPa measured by microindentation whereas in the cross section the Youngs
modulus was 61 GPa. This is attributed to the relative amount of total surface area of pores and
cracks. The surface area of the pores which are aligned parallel to the substrate is greater
compared to the area of the cracks [119].
Plasma sprayed TBCs behave differently in tension and compression (see Figure 3.6). In tension,
the microcracks will open and some splat shear is likely to occur. Under compression the
microcracks will tend to close, and at sufficiently high compressive strains, they will be fully closed.
To further deform the specimen much higher stresses will be required [62, 120, 121]. For a typical
PS PSZ top coat the microcrack width is in the order of 0.1-0.2 m, while the lateral separation of
microcracks is about 10 m. Hence, it might be deduced that strain in the order of 1-2% would be
required in order to close the microcrack [60, 122]. At higher strains, microcrack closure would
occur and an increase in stiffness is expected.

Figure 3.6 Schematic representation of the variation of coating stiffness with imposed strain (after Thompson
[60]). In tension, (a) opening of microcracks results in a low Youngs modulus. (b) at low compressive loads,
the same is true (c) at higher compressive loads an increase in stiffness is expected, due to closing of
microcracks.

Several researchers [109, 120, 121, 123, 124] have observed that PS coatings exhibit non-linear
load-strain curves under either tension or compression, resulting in an appreciable to moderate
hysteresis. The degree of non-linearity appears to be less in tension than in compression, and the
Chapter 3. Review of Thermomechanical behaviour of TBCs

31
hysterisis remained almost unchanged after a few cycles. This irreversible behaviour can be
explained in terms of the microstructure. Microcrack propagation, mechanical interlocking or
loosening and splat sliding may take place. Due to this non-linearity, different approaches have
been taken in order to rationalize the elastic properties. Choi et al [123], introduced an
instantaneous elastic modulus, whereas, Harok et al [121] used a method of partial unloading,
which gave values comparable ultrasonic tests. Hysterisis is less marked in samples that have been
heat treated [109, 123].
Heat treatment of PS YSZ top coat can result in sintering, particularly in the form of changes in the
bonding between splats and healing of microcracks. These changes strongly affect the mechanical
properties of PS YSZ TBCs [60, 111]. Sierbert [113] observed a rise in a Youngs modulus after
annealing for only 2 h at 1100
o
C from 94 GPa to 144 GPa and reaching 177 GPa after 100 h.
Similarly, Eskner et al [63] reported an increase from 38 to 68 GPa after 150 h at a temperatures as
low as 1000
o
C measured by spherical indentation. Zhu et al [112] investigated the evolution of the
Youngs modulus under high heat flux conditions, where the top coat was under a thermal gradient
with the surface at about 1100
o
C. The Youngs modulus increased after 120 h from about 70 GPa
to the final value of 125 GPa.
The effect of heat treatment on stiffness has been investigated by Thompson and Clyne [60]. The
rise in stiffness can be divided in two regimes, a rapid initial increase, followed by a more
progressive rise (see Figure 3.7a). Microstructural observations revealed the mechanisms
governing this behaviour. At short heat treatment times, there was evidence of grain growth and
bridging between splat boundaries (stage I stiffening). These changes cause an appreciable change
in stiffness. Aging for longer times, causes healing of the microcracks (stage II stiffening), which is
particularly significant at temperatures above 1100
o
C. The two stages of the stiffening process are
shown schematically in Figure 3.8. More details on the sintering of YSZ plasma sprayed top coat
and its effect on the microstructure is given is section 3.5.
Sintering and stiffening behaviour of the top coat is different when it remains attached to the
substrate. The higher CTE of the substrate results in the top coat being under tension during high
temperature heat treatment [60]. Stiffness data obtained for heat treated top coats while attached to
the substrate [60] showed that the presence of the substrate reduced top-coat stiffening (see Figure
3.7b). In fact, at lower temperatures and for short times at higher temperatures, the Youngs
modulus apparently decreased. The tensile stress state in the top coat during heat treatment opens
up microcracks and inhibits the sintering process. At higher temperatures and for longer times,
owing to diffusional processes, sintering causes the stiffness to rise.
Chapter 3. Review of Thermomechanical behaviour of TBCs

32
(a)
0
10
20
30
40
50
60
70
80
0 20 40 60 80 100 120 140
Time (hours)
1400
o
C
1300
o
C
1200
o
C
1100
o
C
1000
o
C
E

(
G
P
a
)

(b)
0
5
10
15
20
0 20 40 60 80 100
E

(
G
P
a
)
Time (hours)
1100
o
C
1200
o
C
1300
o
C
1000
o
C

Figure 3.7 Youngs modulus data obtained by applying the cantilever beam bending test to detached YSZ top coats subjected
to various prior heat treatments (a)while detached from the substrate (b) while still attached to the substrate [60].

Figure 3.8 Schematic diagram showing the microstructural defects and healing processes which result in the two
stage stiffening observed experimentally (after Thompson and Clyne [60]).

3.2.2 Prediction of Mechanical Properties of PS Materials
The mechanical properties of porous materials are different from those of dense materials and many
attempts have been made to model their mechanical response. Gibson and Ashby [125] modelled
the bulk elastic properties of a porous material, considering it to be composed of an idealised open
foam cell. Mackenzie [126] derived an expression for the Youngs modulus of a porous material,
Chapter 3. Review of Thermomechanical behaviour of TBCs

33
relating it to the Youngs modulus of the dense material and the porosity fraction, for a material
with homogenous and isotropic elastic properties and isolated, randomly distributed spherical pores.
The Eshelby model [127] addressed the problem for ellipsoidal inclusions. Based on the Eshelby
method, many researchers have modified the approach and assumed zero stiffness for the inclusions
in order to understand the mechanical response of thermal barrier coatings [128 Wang, 2003
#10277].
Leigh and co-workers [128] used a set of explicit constitutive equations for the effective elastic
constants of transversely isotropic porous materials and developed them by modelling the pores as
unidirectional aligned and two-dimensional randomly oriented ellipsoids. They obtained
quantitative microstructural data from stereological analysis, which was used to calculate the elastic
constants based on the model. Their results highlighted the strong dependence of elastic constants
on microstructural features such as void concentration, void aspect ratio and crack density.
Experimental data lay below the model predictions.
Sevostianov and Kachanov [129, 130] followed a similar approach and investigated the effect of the
crack-like shapes of the observed pores in plasma sprayed deposits. Their work indicated that crack
parameters such as crack densities and orientation play a more important role in determining the
effective elastic properties than does the overall porosity. Their work also directly related the
scatter in crack orientation to the anisotropic properties exhibit by PS coatings. However, in
practice it can be difficult to obtain reliable values for the crack density and its orientation in plasma
sprayed deposits. Kulkarni and co-workers [119, 131] used small angle neutron scattering (SANS)
in order to acquire such quantitative microstructural information of the microstructure of plasma
sprayed deposits produced from different feedstock materials. The microstructural data extracted
from SANS was incorporated in a finite element analysis and used to predict the elastic properties
of the different coatings. Their results were compared with experimental data obtained by
microindentation [119, 131]. Experimental results were lower than the predicted values and this
was attributed to the presence of splat boundaries in plasma sprayed deposits, which was not
included in the model and have an important influence on the overall modulus of the coating [132].
Generally, there is a great interest in quantitatively characterizing the microstructural features of PS
coatings and being able to relate them to their effective properties, and different techniques and
models have been investigated [133-137]. There is particular interest in the evolution of the
microstructure during sintering [138, 139]. More detail on the sintering of YSZ plasma sprayed top
coats and its effect on the microstructure is given is section 3.5.
Chapter 3. Review of Thermomechanical behaviour of TBCs

34
3.3 The Thermally Grown Oxide
Upon exposure to high temperature, an oxide layer develops at the interface between the bond coat
and the top coat, due to bond coat oxidation. The oxide layer is usually referred to as a thermally
grown oxide (TGO) and consists mainly of o-alumina. Alumina-forming alloys are preferred due
to the superior adherence (stability) of alumina and its low oxygen diffusivity. The development of
the TGO is thought to play a critical role in the durability of TBCs.
The growth rate of the TGO is essentially parabolic and depends on the oxygen activity, as well as
on time and temperature [140]. The growth is controlled mainly by the inward ingression of
oxygen anions, rather than the outward diffusion of cations. Ingression of oxygen towards the top
coat/bond coat interface can occur by two mechanisms, namely oxygen gaseous permeation through
the network or interconnected cracks and voids of the TC or oxygen diffusion thorough the lattice.
Fox [21] has shown that, over the range of operating temperatures of TBCs, oxygen transport
through the top coat by permeation is expected to dominate. However, the rate of TGO thickening is
largely controlled by the oxygen diffusion through the TGO itself.
A typical microstructure of a TGO layer formed in a MCrAlY-ZrO
2
system is depicted
schematically in Figure 3.9. The main oxide formed is -alumina. There are two zones, a columnar
zone and an equiaxed zone. In the columnar zone close to the substrate intergranular pores,
dislocation pile ups and low angle grain boundaries are present. In addition, very fine spinels and
chromium oxides are detected in this region. The thin chromia layer and spinels near the BC/TGO
interface probably formed as a result of aluminium depletion in the BC. Oxides other than alumina
and spinels are also present in the equiaxed zone near the TC. These oxides probably are formed
during the early stages of oxidation and are pushed back as the growth of alumina progresses.
Yttrium segregation to the grain boundaries is also observed. Microstructural features such as
spinels and transition oxides, segregation of species to the grain boundaries, can embrittle the
interface region. Therefore, the energy required to cause spallation of the coating decreases.
The TGO layer is typically only a few microns in thickness. However, there are thought to be large
residual stresses associated with the TGO layer. Firstly, differential thermal contraction between the
TGO and the underlying substrate can give rise to large compressive stresses in the TGO. These
stresses have been estimated to be of the order of 3-6 GPa [141]. Also, stresses arise due to the
volume changes associated with phase transformations occurring during the creation of the TGO
(~1 GPa).
Chapter 3. Review of Thermomechanical behaviour of TBCs

35

Figure 3.9 Schematic of thermally grown oxide structure [140].
According to Evans et al [23], the sequence of events that causes the failure of TBCs related to
stresses in the TGO are: (i) stresses develop in the TGO upon cooling due to differential thermal
contraction. These stresses are relieved either by buckling of the TGO or by visco-elastic
deformation of the bond coat. The relaxation processes introduce imperfections. (ii) Stresses normal
to the interface develop around imperfections and cracks are nucleated. Coalescence of these cracks
eventually leads to failure by delamination. On the other hand Tolpygo et al [142, 143], have found
that average and residual stresses in the TGO do not change significantly during thermal cycling
prior to failure. Similar results have been reported by Clarke et al [144].
Typical thicknesses of TGO are 3-10 m and failure commonly occurs when thickness of the TGO
is approximately 6-8m. Several approaches to determine the lifetime of TBCs use the time it takes
for the TGO to reach a critical thickness value as the upper limit for its durability.
3.4 Creep of Bond Coat and Thermally Grown Oxide
Creep is defined as the time-dependent deformation of a material under constant load. Creep in
polycrystalline materials occurs as a result of the motion of dislocations within grains, grain
boundary sliding and diffusion processes. Generally, creep becomes significant at temperatures
above about 0.6T
m
and at these temperatures materials will creep at stresses much lower than their
yield stress. Thus, at high temperatures creep often governs the stability of materials. Creep of the
Chapter 3. Review of Thermomechanical behaviour of TBCs

36
bond coat and the TGO is thought to play a role in the failure of TBCs. In this section the effect of
bond coat and TGO creep on the stresses in the TBC system will be discussed.
3.4.1 Creep in the Bond coat
Temperatures during typical service conditions of TBCs are above 0.6T
m
of the bond coat materials
used commonly. Hence, stresses in the bond coat will relax due to creep with time at high
temperature. It has been suggested that at temperatures above 800
o
C the bond coat becomes
essentially non-load-bearing after short times [34]. On cooling down from elevated temperatures,
stresses regenerate due to differential thermal contraction and stress levels in the bond coat might be
higher than originally, especially at high cooling rates [145]. This is because at high cooling rates,
typical of those experienced under service conditions, a limited degree of creep relaxation can take
place.
The creep strain rates of two bond coat materials, namely CoNiCrAlY and NiCrAlY have been
measured experimentally by Thompson et al [28]. The measured steady state creep rates conform
to power law expressions, with respective stress exponents of about 2.9 and 4.5. The effect of creep
of the aforementioned bond coats on their residual stress state after heat treatment of a TBC system
was predicted using a numerical model. At temperatures above 700
o
C stresses in the bond coat
relax with time due to creep. On returning to room temperature, at high cooling rates, enhanced
stress levels existing in the bond coat. The level of the stresses depends on the relative values of
expansivity, creep parameters and modulus of the bond coats. This larger stress is likely to enhance
the initiation of microcracks and debonding of the TBC leading to spallation [145].
3.4.2 Creep of the TGO
The high melting point of alumina, the main constituent of the TGO, makes creep of TGO less
significant than bond coat creep. However, creep relaxation may take place in the oxide during
prolonged heat treatment or thermal cycling at high temperatures and several researchers have
looked into TGO creep and its effect on TBC failure.
A conclusive argument on the effects of TGO creep has not been established. Rsler et al [30]
maintain that creep deformation of the TGO scale can significantly reduce growth stresses and
therefore, relaxation by creep will improve TBC life. However, it is possible that creep at high
temperatures could lead to higher stresses on subsequent cooling. On the other hand, Evans et al
[29], maintain that creep of the alumina oxide layer has no significant effect on interfacial crack
propagation rates, but that creep in the underlying NiCrAlY coating, on contrary has a significant
Chapter 3. Review of Thermomechanical behaviour of TBCs

37
effect in retarding wedge crack along the interface with the oxide. Tolpygo and Clarke [143, 146]
relate the importance of creep relaxation in the oxide to substrate thickness and maintain that TGO
creep is major relaxation process on thick substrates, whereas in thin substrates metal deformation
predominates. In all cases the effect of cooling rate on TGO creep is evidently significant, with
higher cooling rates preventing creep relaxation. Residual stresses in the oxide become smaller
with decreasing rate of cooling after oxidation, due to relaxation by creep during cooling.
3.5 Sintering of the Top Coat
The microstructure of a typical PS TBC incorporates many through-thickness microcracks,
interlamellar pores, globular voids and other defects. As a results of these, top coats have low
stiffness and hence high strain tolerance. This high strain tolerance enables the top coat to
withstand the thermal strains imposed during thermal cycling of the TBCs. However, at the
temperatures experienced under-service conditions, defects and microcracks might heal as a result
of sintering, reducing the strain tolerance and increasing the driving force for top coat spallation.
Additionally, the thermal conductivity of the top coat will tend to increase. In this section, a brief
overview of sintering effects in the top coat is presented.
3.5.1 Sintering mechanisms in ceramics
3.5.1.1 Solid State Sintering
The definition of sintering follows: sintering is a thermal treatment for bonding particles into a
coherent, predominantly solid structure via mass transport events that often occur at the atomic
scale. The bonding leads to improved strength and lower system energy [147]. The driving force
for sintering is the reduction of free surface energy and interfacial energy acting over curved
surfaces. For two adjacent particles the free surface is associated with high energy and particles
sinter by atomic level events to eliminate free surface energy. Surfaces with high curvatures have
high surface area per unit volume and mass flows towards regions of lower curvature; hence
sintering tends to eliminate surface curvature. Reduction of free surface energy occurs also by
removal or coalescence of pores, consequently densification and coarsening often occur during
sintering [148]. Similarly, grain boundaries have an energy associated with them and grain growth
during heat treatment reduces the grain boundary energy and area [149].
In response to the driving forces for sintering there is mass transport. There are several possible
mechanisms of mass transport. Possible mass transport mechanisms of sintering are summarized in
Table 3.2 and illustrated in Figure 3.10 [150]. The arrows indicate the path along which the material
Chapter 3. Review of Thermomechanical behaviour of TBCs

38
flows for each mechanism. These mechanisms are separated mainly in two classes, surface
transport and bulk transport.

Figure 3.10. Schematic representation of the possible mechanisms of sintering (after Ashby [150]).
Possible sintering mechanisms
Mechanism
Number
Transport path Source of matter Sink of Matter
1
Surface diffusion Surface Neck
2
Volume/lattice diffusion Surface Neck
3
Vapour diffusion Surface Neck
4
Grain boundary diffusion Grain boundary Neck
5
Volume/lattice diffusion Grain boundary Neck
6
Plastic flow/lattice diffusion/ Dislocation Neck
Table 3.2 Possible diffusion mechanisms during sintering [147, 150].

Surface transport processes produce neck growth without densification or shrinkage, the source and
sink of mass in these processes being the particle surface. These mechanisms are surface diffusion
(1), volume/lattice diffusion from the particle surface (2), and vapor diffusion (3) [147, 150].
Surface diffusion involves the motion of atoms from and to surface defect sites. Surface diffusion
is thermally activated and usually the activation energy is less than other mass transport
mechanisms. Therefore, surface diffusion initiates at lower temperatures in comparison with other
mass transport mechanisms and it is an initial contributor to the sintering process. Volume
diffusion from the particle surface, through the particle interior and subsequent emergence at the
particle surface is effectively a transport path form a surface source to a surface site. However this
diffusion path does not occur at very significant levels [147]. During vapor transport atoms at the
surface evaporate and condensate on a nearby surface. This process usually has a very small
contribution, however it can be significant for materials with high vapor pressures or those that
form volatile species.
Chapter 3. Review of Thermomechanical behaviour of TBCs

39
Bulk transport processes include grain boundary diffusion (4), volume/diffusion (5) and plastic flow
or lattice diffusion through dislocations (6) [147, 150]. These mass transport processes result in
densification and shrinkage and in general are active at higher temperatures. Grain boundary
diffusion is usually an important mechanism in the densification of most material systems. Grain
boundaries, due to their defective character provide a low energy path for diffusion. Mass is
removed from the grain boundary and deposited at sinter neck. The activation energy for grain
boundary diffusion is usually intermediate between that for surface and volume diffusion. When the
grain boundary has high energy, sintering may be inhibited [151, 152]. Grain boundary energy
might be lowered by impurities or additives that segregate at grain boundaries [153, 154]. If
sintering is grain boundary diffusion-controlled, most typically active species will segregate at grain
boundaries [155]. Volume diffusion involves the motion of atoms from the interparticle grain
boundary through the crystalline structure, to the neck surface, or a reverse vacancy flow, with
subsequent shrinkage and densification. The volume diffusion rate depends on temperature,
composition and curvature. Essentially, the vacancy population of the polycrystalline material,
which depends on temperature, determines the volume diffusion contribution to sintering. For ionic
materials, stoichiometry is also a controlling parameter [156]. The vacancy concentration in ionic
materials will be affected by their stoichiometry. The total contribution of volume diffusion is then
determined by the combined action of the thermally induced vacancies and vacancies induced as a
result of changes in stoichiometry. Finally, plastic flow or lattice diffusion by dislocation climb can
also contribute to sintering and densification. Dislocations essentially interact with vacancies and
improve mass transport. This transport process is most likely important during heating and their
role decreases as dislocations are annealed out.
It is worth noting that actual sintering involves a mixture and cooperative occurrence of the above
processes described. During the different stages of sintering different mechanisms shift in
dominance.
Solid state sintering can be divided into three stages. The first stage involves particles approaching
one another and the formation of a sinter bond or neck at the point of contact. In this stage, the
number and size of necks between individual particles is sufficiently small to consider that there is
no interaction between neighboring necks [157, 158]. The second or intermediate stage of sintering
is characterized by great shrinkage and densification, pore rounding and grain growth [159]. As
sintering progresses grain growth becomes increasingly active. By the end of the intermediate stage,
no open porosity remains in the structure, all remaining pores are closed. The final stage of
sintering is a slow process [159]. It involves the collapse of the pore structure, pore spheroidization,
Chapter 3. Review of Thermomechanical behaviour of TBCs

40
simultaneous coarsening and densification. Pores become isolated and move away from grain
boundaries. Once pores become spherical and isolated, further densification is difficult. Various
additives such as second phase particles and segregants can slow grain growth and assist
densification during the final stage sintering.
Numerous models have been developed in order to describe the sintering process. Most of them
describe the sintering of spherical particles of the same size. Ashby [150] developed sintering
diagrams which identify, at a given temperature, particle size and neck size, the dominant
mechanism and show the rate of sintering produced by all the mechanisms acting together. The
initial stage of sintering has been the focus of many of the sintering models [147, 150, 151, 160-
163]. For spherical particles, equal in size, the initial stages of sintering can be described in a
simplified form, by relating the neck size to particle ratio to the sintering time by the following
equation:
1
n
m
Bt r
R R
| |
=
|
\ .
(3.5)
where r is the neck radius, R is the particle radius, t is the sintering time, B
1
is a constant, and n and
m are constants that depend on the mass transport mechanism. The constant B
1
is a constant that
collects material and geometric constants and depends on the sintering mechanism. Depending on
the assumption used, different values are assigned to the exponents [164]. As mentioned earlier the
initial stage of sintering occurs mainly through surface diffusion and the diffusion coefficient
follows an Arrhenius temperature dependence embedded in the parameter B
1
. The above equation,
even though not very exact, is a good approximation for neck size ratios up to 0.3. In addition it
illustrates the relative importance of the parameters affecting the initial stage of sintering, such as
the high sensitivity to the inverse of particle size and exponential temperature dependence in
comparison with the relatively small effect of time. Kingery and Berg [160] considered the
approach of the centers of spherical particles by bulk transport mechanism and derived an equation
that related the linear shrinkage to the sintering time t, in a relationship of the following type:
2
n
o
L
B t
L
A
= (3.6)
where L is the length change of a material with initial length L
o
, and the parameter B
2
includes
physical constants and material parameters such as surface energy, volume diffusion coefficient,
atomic size, etc. The parameter B
2
is also exponentially dependant on temperature. The exponent n
takes a value of 2/5 for spherical particles. This method is an attempt to separate the mechanisms of
Chapter 3. Review of Thermomechanical behaviour of TBCs

41
surface mass transport to bulk mass transport in the initial stages of sintering and is valid only for
small shrinkages (3%). Due to the many simplifying assumptions this equation is not always
applicable to the sintering of real powders. Other efforts have also been made in literature for
quantitative descriptions of shrinkage occurring during the sintering process. A critical review of
these models is given by Exner [163, 165].
The intermediate process comprises of densification, grain and pore growth. Due to complexity of
these phenomena, there are very few attempts to model the process. Coble [159] first proposed a
microstructure model based on a highly-interconnected pore structure network. The microstructure
comprises of grains in the shape of tetrakaidecahedron with cylindrical pores occupying the grain
edges. The pore network is fully interconnected. With this model a relation between the remaining
porosity in the material and the sintering time was derived:
3
ln
o
o
t
p p B
t
= (3.7)
where p
o
and t
o
are the porosity and time at the onset of the second stage, t is the isothermal
sintering time (greater than t
o
) and p is the porosity at sintering time t. The parameter B
3
includes
physical constants of diffusional sintering and grain growth such as surface energy, diffusivity,
atomic volume, curvature and Boltzmanns constant and is also exponentially dependant on
temperature The above equation is often termed the shrinkage law. Coble further developed his
model for intermediate stage sintering and modified it to apply it to final stage sintering as well. He
assumed spherical pores located at grain corners and related the porosity to the pore diameter and
the tetrakaidecahedron grain edge length. Shi [148] proposed a microstructure pore model for
intermediate and final stage sintering based on spherical pores. He derived densification equations
relating density and densification rate to pore size and grain size ratio. The derived equations can
also be used to explain the effect of pore size distribution and the agglomeration properties. Harmer
et al [166, 167] developed sintering diagrams which highlight the conditions where pores remain
attached to grain boundaries during final-stage sintering.
Particle size has a significant effect on the sintering. During the initial stage of sintering, the
significance of particle size becomes apparent from the neck growth equations for each transport
mechanism, which are proportional to a power of the inverse particle size. Some mechanisms of
mass transport are more sensitive to particle size than others. Surface and grain boundary diffusion
are very sensitive and are favored by smaller particles sizes because of the higher surface area
content per unit volume. Volume and vapor diffusion are less sensitive to particle size changes.
Chapter 3. Review of Thermomechanical behaviour of TBCs

42
During intermediate and final stage sintering, there is a similar dependence on particle and pore size.
Small particles and small pores have a higher energy per unit volume, more available surface area,
and higher curvature, all of which contribute to higher sintering rates. Herring [168] was among the
first to report this effect and he established a scaling law, where he correlated the degree of sintering
of different sized powders to the sintering time. For smaller sized powders, less sintering time is
needed to reach a certain amount of sintering. However, is worth noting that despite enhanced
sintering of powders with smaller particle sizes, nanoscale powders tend to form agglomerates
which prove difficult to sinter [169, 170]. In general, large pores with lower curvature and higher
dihedral angle are difficult to sinter [149]. Particle size distribution also plays a role in sintering. A
wide particle size distribution results in inhomogeneities during packing, which can result in a
coarser, more porous microstructure at long sintering times [171]. Another factor that affects
sintering of compacts is the particle packing, or green compact density. Small packing densities
and packing inhomogeneities result in sintered dense regions surrounded by enlarged pores. Higher
packing densities result in more initial points of contact between particles, smaller initial pores, and
higher final densities. In general, there are many factors that affect the solid state sintering and
careful consideration of these is necessary in order to be able to control the process.
3.5.1.2 Liquid Phase Sintering
Sintering in the presence of a liquid leads to densification by quite a different process. Liquid phase
sintering occurs when a liquid is present at the sintering temperature. There are three different
stages in the densification process after the liquid is formed, illustrated schematically in Figure 3.11.
Stage 0 represents the preliminary stage, where the reactive liquid is formed. In many systems prior
to the formation of the liquid considerable densification occurs by solid state sintering. On
formation of the liquid phase, rearrangement of the liquid and solid occurs due to capillary forces
(stage I). The rearrangement of particles proceeds in the direction of reducing porosity. If the
volume of the liquid phase is sufficient, complete densification can take place in this stage. On the
other hand, at low liquid contents the solid skeleton inhibits densification, requiring the
participation of solution-reprecipitation events for further densification. At the end of the
rearrangement stage, further densification occurs by a solution-reprecipitation process (stage II),
where mass transport through the liquid controls densification. Material dissolves at the contact
points between particles and diffuses out through the liquid film between the particles. This process
causes particle centres to move closer together. The final stage of liquid-phase sintering is
controlled by grain growth and densification of the solid skeleton [172]. Other processes such as
Chapter 3. Review of Thermomechanical behaviour of TBCs

43
coalescence of grains and pores, diffusion of the liquid components into the solid and phase
transformations may also place.
The temperature of liquid-phase formation and the liquid composition can be determined from the
phase diagram for a specific system of materials. A decrease in the liquidus and solidus
temperatures with alloying indicates a tendency for solute segregation. Solute segregation is greater
for greater separation of the solidus and liquidus curves. Impurities often show up preferentially at
grain boundaries and cause the formation of an inadvertent liquid film [173].
One key factor in liquid-phase sintering is the solubility of the solid in the liquid. For high
densification, the solid solubility in the liquid is high, but the solubility of the liquid in the solid is
low to avoid formation of a transient liquid. In addition, good wetting between the solid particles
and the liquid is required [147, 174, 175].

0

I

II

III

Figure 3.11 Schematic diagram of the stages of liquid phase sintering, 0: melting; I: rearrangement; II: solution-
reprecipitation; III: pore closure (after [147]).

3.5.2 Sintering of Zirconia based top coats
The sintering behaviour of commercial yttria-stabilized zirconia powders has been investigated by
various researchers and a wide range of sintering additives that enhance sintering have been
explored. It is clear that such additives should be avoided for TBCs made of zirconia-based
materials.
The mechanisms controlling sintering kinetics in stabilized zirconia are disputed in literature. Some
authors [169, 176] reported initial stage sintering kinetics to be dominated by grain boundary and
surface diffusion whereas other authors support that volume diffusion is the dominant mechanism
[177-179]. The rate of volume diffusion is determined by the species diffusing most slowly through
Chapter 3. Review of Thermomechanical behaviour of TBCs

44
the lattice, which in zirconia is the cation [180-182]. The mechanism of cation diffusion in
stabilized zirconia is also disputed in literature [180, 183, 184].
The addition of up to 6 mol% Y
2
O
3
and up to 12 mol% CaO stabilizers in zirconia powder
compacts has been reported to enhance densification [185]. Segregation of Y
+3
, La
+3
, Ca
+2
and Ce
+4

solute dopants at the grain boundary in zirconia has been observed by various researchers [186-190].
Segregation of solid solution additives at the grain boundaries lowers grain boundary energy and
aids shrinkage of pores by grain boundary diffusion [149, 155, 191]. Larger, low valence cations
appear to be more effective in reducing the grain growth, in agreement with the solute drag
models[153, 154].
The sintering rate of stabilized zirconia is strongly dependant upon impurity content. MgO [185],
Bi
2
O
3
[192], Fe
2
O
3
[192], SiO
2
[193] CuO [194] and Al
2
O
3
[195, 196] have been shown to promote
sintering in stabilised zirconia powder compacts when added in small amounts . The most common
mechanisms involved are segregation of the impurity at the grain boundary, formation of a second
phase at the grain boundaries, or formation of a liquid phase, which migrates to the grain boundaries.
De Souza et al [173], observed the development of a silicate glassy phase during sintering of
isostatically pressed powder discs with 0.5 wt % Silica. The presence of this phase is thought to
promote liquid phase sintering. Similar results have been reported for the influence of impurities,
especially SiO
2
, in PS coatings. Eaton and Novak [197] reported a fivefold increase in shrinkage on
going from 0.2 wt % to 4 wt % silica. This was attributed to the presence of a silicate glassy phase
that promoted liquid phase sintering. An increase in shrinkage with increasing silica content (in the
range of 0.02-0.12 wt %) was also reported by Vassen et al [198]. It is important to note that ranges
of impurity content in the studies mentioned above are significantly different. Work conducted by
Klocker et al [199] indicates that, for the impurity levels encountered in baseline YPSZ powders
(0.3-0.4 wt%), there is no apparent effect on the sintering behaviour. In addition to impurity levels,
non-stoichiometry also affects sintering. Higher sintering rates have been observed by Zhu et al in
Ar + 5%H
2
atmosphere compared to air, due to enhanced cation interstitial diffusion under reducing
conditions [182].
The effect of particle size and temperature on the densification rate was studied in detail by Rhodes
[169]. He confirmed that using fine powders, free of agglomerates enhanced sintering kinetics.
Typically, PSZ compacted powders exhibit relatively rapid sintering at temperatures between
1100-1300
o
C [200], depending on the powder composition and morphology. Sintering effects in PS
TCs appear to be significant at temperatures above about 1000
o
C [197]. As a result of sintering,
there can be significant increase in the stiffness and thermal conductivity of the top coats [60, 112].
Chapter 3. Review of Thermomechanical behaviour of TBCs

45
Ilavsky et al [201] reported a significant decrease in the surface area voids (up to 33%) in PS YSZ
top coats as a result of sintering, at temperatures as low as 1000
o
C. Cracks appear to sinter first, at
lower temperatures, followed by sintering of intersplat pores at higher temperatures.

3.6 Phase changes in the Top Coat
As mentioned is section 2.2.3, zirconia exists as three solid phases that are stable at different
temperatures [31]. At temperatures up to 1200
o
C, the monoclinic phase (m) is stable. Zirconia
transforms from the monoclinic to the tetragonal phase (t) above 1200
o
C and above 2370
o
C to the
cubic phase (c). Transformation from m to the t phase has an associated volume decrease of 4%
[32]. A phase diagram for the zirconia-yttria system is presented in Figure 3.12a.
Zirconia-based TBCs are doped with trivalent or tetravalent cations, in order to stabilize the
tetragonal zirconia polymorph and prevent catastrophic cracking as a result of the volume changes
accompanying the tm transformation, which occurs at temperatures within the range of the
working environment in gas turbines. Stabilization of the tetragonal or cubic zirconia phase by the
addition of trivalent dopants is believed to be associated with the generation of oxygen vacancies
[202]. The trivalent dopant cations substitute the Zr ion in the cation network and oxygen vacancies
are created for charge compensation. The tetragonal and cubic polymorphs are stabilized when the
doping level is sufficient to create enough oxygen vacancies to reduce the co-ordination number of
the Zr ion from 8 to around 7.5 [203, 204]. The Zr ion in fully stabilized and high temperature
cubic phase has a coordination number of 7, due to the strong covalent nature of the ZrO bond and
the small size of the Zr ion.
The tetragonality c/a is an important factor in the stability of the tetragonal phase and it has been
shown to be independent of the species of the dopant, but dependant on the content of the dopant
for oversized trivalent rare earths [205, 206]. Tetragonality decreases with increasing dopant
content for oversized trivalent cations and vanishes at 11 mol% M
2
O
3
regardless of the ionic sizes,
since tetragonality is aided by the creation of oxygen vacancies only [207]. The decrease in
tetragonality is associated with increasing stability of the tetragonal phase. The lattice parameters
on the other hand of the tetragonal phases increase systematically with the ionic radii of the dopant
atoms [205, 206].
Tetravalent dopants such as ceria still stabilize the tetragonal and cubic phases of zirconia.
However, the stabilization is not the result of the generation of oxygen vacancies, as is the case with
Chapter 3. Review of Thermomechanical behaviour of TBCs

46
trivalent dopants. The Ce
4+
ion substitutes the Zr
4+
ion in the cation network, without creating any
oxygen vacancies, since both ions are tetravalent [208, 209]. In this case, the stabilization effect is
thought to be caused by the slight dilation of the cation network by the oversized Ce
4+
ion which
decreases the tetragonality and stabilizes the tetragonal phase to room temperature [208]. The
tetragonal lattice in zirconia has an inherent lattice strain caused by the smaller ionic size of the
zirconia atom. This strain can be relieved either by the presence of oxygen vacancies associated
with the Zr ions, as in the case of trivalent dopants, or in the case of tetravalent dopants by slightly
dilating the lattice by introducing a oversize ion. The tetragonality c/a of zirconia doped with
tetravalent cation decreases with increase in dopant content, as is the case with trivalent dopants.
However this effect is not so pronounced [208].
The conventional material used for the top coat in TBCs is ZrO
2
stabilized with 7-9 wt% Y
2
O
3
.
Partially stabilised zirconia with 8 wt% yttria consists of non-transformable metastable t tetragonal
phase. A metastable tetragonal phase (t) results from a diffusionless shear transformation directly
from the cubic state. During plasma spraying, the rapid cooling of molten splats to the substrate
temperature often results in the formation of t phase. This phase is thought to be stable up to
1400
o
C i.e. it does not transform directly to monoclinic on cooling to room temperature. The
resistance of the non-transformable tetragonal to transformation into the monoclinic phase is
ascribed to its smaller tetragonality. The non-transformable t tetragonal phase in PSZ with 6-
12 wt% yttria will decompose during annealing at elevated temperatures into the equilibrium high
temperature phases, cubic (c) and tetragonal (t) [38, 210, 211]. These phases will be retained at
room temperature, either as high-yttria and low-yttria t phases or as cubic and low-yttria t phase.
The rate of transformation is dependent on composition and temperature, and is slow at
temperatures below 1400
o
C. When annealing at temperatures above 1400
o
C, the tetragonal phase
can transform to monoclinic upon cooling [212, 213]. The cooling rate has a significant effect on
the tm transformation [214]. Slower cooling and heating rates will allow the tm transformation
to take place (see Figure 3.12b). Furthermore, the size of the tetragonal precipitates also affects the
stability of the t phase. Larger precipitates transform spontaneously to the m phase, whereas below
a critical grain size the metastable phase remains stable [214]. Finally, residual stresses also have
an effect on the tm transformation. It is thought that residual stresses generated during air
quenching can partially suppress the tm transformation. On the other hand, stress fields
associated with defects can promote a stress-induced tm transformation [215, 216].
The phase stability of PS top coats after prolonged heat treatment is an important issue. Different
stabilizers have been used to ensure that the tetragonal phase is retained. The basic criteria for the
Chapter 3. Review of Thermomechanical behaviour of TBCs

47
selection of a suitable stabiliser include a suitable cation radius and a cubic crystal structure. Early
attempts used MgO to stabilize zirconia in its cubic state, by adding 25 wt% MgO [33]. However,
during heat treatment the zirconia reverts to its monoclinic form and the stabilizing oxide
precipitates out from solid solution, affecting the thermal conductivity [34]. Zirconia can be fully
stabilised to its cubic phase by adding 20 wt% yttria. However, such fully stabilised zirconia
coatings perform very poorly in thermal cycling tests [35].
CeO
2
-stabilized TBCs have been investigated by various researchers [36-38, 217, 218] and the
existence of the non-transformable t tetragonal phase similar to that in the YSZ has been agreed
upon. At high CeO
2
weight percentages a second metastable cubic phase c also exists. In reducing
atmospheres the Ce
4+
ion reduces to Ce
3+
, which tends to stabilize the cubic phase [219].
Gadolinia

is in general a less effective stabilizer than Yttria, although partial substitution of Gd for
Y does enhance phase stability, if Y remains the dominant species [220, 221]. ZrO
2
-Er
2
O
3,
ZrO
2
-
Sm
2
O
3
and ZrO
2
-Nd
2
O
3
PS TBCs have been reported to consist of non-transformable t tetragonal
phase up to compositions of 6 mol% of stabilizer, and above that, coatings are entirely cubic [39,
222, 223]. Erbia stabilized coatings showed better thermal phase stability than samaria-stabilized
and neodymia-stabilized coatings, which decomposed after heat treatment to the monoclinic and
tetragonal phases. Scandia-yttria stabilized TBCs have been reported to show better phase stability
than conventional YSZ TBCs up to temperatures of 1400
o
C [224, 225]. Dy
2
O
3
and Yb
2
O
3
rare-
earth dopants have also been reported to stabilize the tetragonal phase [40, 206, 226]. Zirconia
doped with lanthana has been reported to form the cubic pyrochlore-structured phase, especially at
high mol % of stabilizer [227-229]. The cubic pyrochlore structure is stable up to its melting point
and hence it is believed that these materials might have potential as TBCs [230]. However coatings
of this material yield shorter thermal fatigue cyclic life than YSZ due to their relatively low CTE
[231].
Chapter 3. Review of Thermomechanical behaviour of TBCs

48


Figure 3.12 (a) Phase diagram for the zirconia-yttria system [232], (b) phase flow diagram showing behaviour of plasma-
sprayed 8 wt% Y
2
O
3
PSZ on cooling after aging [214].

Chapter 4. Review of Thermal Conduction in TBCs

49

4 Review of Thermal Conduction in TBCs
Trends towards ever increasing turbine inlet temperatures and the need for coatings to protect hot
sections of the turbine have increased the demand for TBC materials with low thermal conductivity.
This review is concerned with briefly describing the thermal conduction mechanisms in ceramic
oxides and reviewing the methods for determining the thermal conductivity. Also, a review of the
thermal conductivity of TBC materials and various materials with a potential use as TBC is
presented as well as an existing microstructural model for the prediction of the thermal conductivity
of TBCs.
4.1 Thermal Conduction in Ceramics
Thermal conductivity is a measure of the rate of flow of thermal energy through a material in the
presence of a temperature gradient. In oxide ceramics thermal energy transport is due to lattice
vibrations (phonons) or electromagnetic waves (photons).
4.1.1 Phonon Conduction in Ceramics
The contribution to the thermal conductivity from phonon conduction, as derived by Debye, can be
expressed by the general form [174, 233] :
1
3
ph v p
k C l v =
}
(4.1)
where C
v
is the specific heat at constant volume, the density, the mean phonon velocity
(approximately equal to the velocity of sound in the crystal) and l
p
the mean free path for scattering
of phonons. The mean free path of phonons is the distance they travel before being scattered.
Scattering of phonons in solids can occur by interaction with each other as well as by interaction
with defects, such as impurities, vacancies, grain boundaries and interfaces. Scattering of phonons
by direct interaction amongst themselves occurs by Umklapp-processes (sometimes called U-
processes), which is the name given to these interactions by Peierls [233].
At very low temperatures, phonon-phonon interactions are limited, and the mean free path is
determined by scattering from defects and grain boundaries. At very high temperatures U-processes
are very frequent and they are effectively the main source of thermal resistance. At intermediate
temperatures both U-processes and defect scattering are important.
Chapter 4. Review of Thermal Conduction in TBCs

50
The temperature dependence of phonon thermal conductivity in a non-metallic solid can be
predicted from equation 4.1. At very low temperatures, the mean free path is essentially constant
and the heat capacity C
v
varies as T
3
, so the temperature dependence of the thermal conductivity
reflects the T
3
behavior of the specific heat. At high temperatures, the specific heat is nearly
constant, and the mean free path is inversely proportional to the temperature, and hence the thermal
conductivity is proportional to 1/T. In the intermediate range of temperatures (since the incidence
of phonons rises with temperature), the thermal conductivity is determined by the intersection of the
behavior in the other temperature ranges, and is essentially represented by an exponential of 1/T
[234]. A typical conductivity curve for a non-metallic crystal is shown in Figure 4.1.

Figure 4.1 Temperature variation of phonon and photon conductivity for ceramic crystal [234].
In ceramic polycrystalline materials, the mean free path of the phonons, and hence the thermal
conductivity, may be influenced by the dimensions of the grains, the crystal structure, porosity or
impurities present. If the mean free path is of the same order of magnitude as the grain size, then
grain boundaries become active scattering centers and may reduce the mean free path of the energy
carriers and hence the thermal conductivity. For many ceramics, this is the case at low temperatures.
If the mean free path is generally smaller than the grain size, then the crystal structure is the major
influence, rather than the grain size. In this case, increased structural order results in an increase of
the mean free path and hence the thermal conductivity. The mean free path when there are a
number of independent scattering processes is given as [235]:
1 1
i
p i
l l
=

(4.2)
k
ph
- phonon conductivity
k
r
- photon conductivity
k
t
- total conductivity
k
r
~T
3
l
r

k
t
= k
ph
+ k
r

k
ph
~T
3

k
ph
~e
-1/T

k
ph
~T
-1
T
h
e
r
m
a
l

C
o
n
d
u
c
t
i
v
i
t
y

Temperature
Chapter 4. Review of Thermal Conduction in TBCs

51
where the subscript i denotes the various scattering processes operating. The above equation
suggests that the mean free path of a phonon is determined by the smallest of the mean free paths
due to the various independent scattering processes. In the case of substitutional atoms of mass
m+m in place of a normal atom of mass m, Klemens [235, 236] has shown that their contribution
to the mean free path is:
2
3 4
3
1
4
pd
m a
C
l m
A e
tv
| |
=
|
\ .
(4.3)
where l
pd
is the mean free path due to scattering of phonons by point defects, C is the concentration
of point defects per atom, a
3

is the atomic volume, is the mean phonon velocity and is the
frequency of the phonon lattice waves. Effective values of m/m

can be obtained for other defects,
including vacancies. Porosity, grain boundaries and impurities also reduce the mean free path of
phonons and hence the thermal conductivity.
4.1.2 Radiative Heat Transfer in Ceramics
In addition to phonon conduction, heat may be transferred through solids by radiative heat transfer.
Radiative heat transfer is energy transport through transmission, or absorption and reradiation of
electromagnetic waves (photons). Radiative heat transfer can occur between two surfaces and it
does not require a heat transfer medium, it can occur in vacuum. In solid materials photons can be
reflected, absorbed and emitted by the material itself. The contribution to the thermal conductivity
by photon conduction can be expressed as [234]:
r r
l T n k
3 2
3
16
o = (4.4)
where is the Stephen-Boltzmanns constant, n is the refractive index, T is the absolute temperature
and l
r
the mean free path of photons or penetration depth. The refractive index for most oxides
usually varies very little with wavelength or temperature within the region of interest. The mean
free path of a photon is the average distance a photon travels in the material between scattering
events or before being absorbed by the medium. The total mean free path is equal to the reciprocal
of the extinction coefficient. The mean free path l
r
(or penetration depth) in equation (4.4) is given
by,

l
r
1
= (4.5)
Chapter 4. Review of Thermal Conduction in TBCs

52
where = + is the extinction coefficient and is equal to the scattering () plus the absorption ()
coefficients [237]. Generally, the extinction coefficient depends on the absorption and scattering
characteristics of the material, the wavelength distribution of the incident radiation and the
temperature of the material.
The scattering coefficient will depend on the microstructural features for polycrystalline materials.
If the radiation penetration depth is much greater than the grain size, then grain boundaries become
active scattering centers, as do pores [234, 238]. The radiation heat flux in this case will be uniform
through the material and the pores, but pores will change the scattering characteristics by decreasing
the penetration depth, relative to that of fully dense material. The dependence of the scattering
coefficient on porosity is thus expected to be quite strong, and porosity at a level of few % may
result in changes of an order of magnitude in the scattering coefficient [234, 239].
The total conductivity of oxide ceramic is the summation of phonon and photon conduction:
t ph r
k k k = + (4.6)
In a perfectly transparent material or medium, the photon mean free path is infinite, and radiation
energy transfer occurs through the medium, without any interaction with the medium. In an opaque
material, the photon mean free path is extremely short, and all the radiation is absorbed and emitted
over extremely short distances. Therefore, the photon transport contribution to total energy transport
is small compared to the contribution from phonon conduction. If the photon mean free path in a
solid material is large compared to the phonon mean free path, as can be the case for
semitransparent materials, radiation conduction becomes significant. However, as can be seen from
eq.(4.4), the effect of temperature is more significant than the effect of photon mean free path on the
radiative heat transfer.

4.2 Methods for Measuring the Thermal Conductivity of Solids
4.2.1 Steady State Methods
Steady state methods for measuring the thermal conductivity generally consist of measuring a
temperature gradient and a heat flux across a sample and obtaining the thermal conductivity from:
T
q kA
x
A
A
= (4.7)
Chapter 4. Review of Thermal Conduction in TBCs

53
where is q the is the input power, A is the cross-sectional area, T is the temperature drop, x is the
distance over which the temperature drop occurs, and k (W m
-1
K
-1
) is the thermal conductivity.
The simplest steady state arrangement comprises a sample of uniform cross-sectional area A, to
which heat is supplied at one end, and removed at the other end, establishing a thermal gradient
across the length of the sample (linear flow method). An alternative is a sample which surrounds
the heat source either cylindrically or spherically, so the heat flow is radial. Further terms are then
needed in equation (4.7).
In the linear flow method, the sample is placed between a heat source and a heat sink, and so the
heat flow is linear and parallel to the sample axis and the isotherms are planar and parallel to the
cross-sectional area. The temperature difference T=T
2
-T
1
along the length x=x
2
-x
1
is measured
using thermometers or thermocouples. Since k usually varies with temperature, the derived value of
k will be that corresponding to an average temperature between T
2
and T
1
.
The steady-state linear flow method can only be applied if all the heat supplied by the heat source
travels though the sample to the colder end, with negligible heat losses from the periphery of the
sample. Heat losses are difficult to prevent, especially at higher temperatures. However, they can be
usually minimized, so that they become insignificant compared to the heat conducted through the
sample. Heat losses can be minimized by performing the measurement in vacuum and surrounding
the sample with insulation. In addition, the thermocouples should, if possible, fit tightly in the hole
and be small in cross-section, in order to minimize disruption of the heat flow. The choice of
specimen geometry, controlled mainly by the expected thermal conductivity value to be measured,
is also important. In order to ensure that the temperature difference T is sufficiently large, the
ratio of length to area must high. However, relative lateral heat loses are proportional to the surface
area, so to minimize them, a short sample with a large cross section is required. In practice, long
bars are used for good conductors and thin discs for poor conductors.
At higher temperatures, heat losses due to radiation can become very significant and it may be
necessary to place a radiation shield around the sample. Ideally the shield should be heated
independently of the sample and have a temperature distribution similar to that along the sample.
However, this is difficult and in general, the use of this method is limited to temperatures where
radiative heat loss is not significant.
A standard method for the measurement of the thermal conductivity using steady state linear flow is
the guarded hot plate method. The equipment consists of a central transverse heater and two
similar samples either side of it. The samples and heater are sandwiched in between two heat sinks
Chapter 4. Review of Thermal Conduction in TBCs

54
and the whole assembly is surrounded by a material with low conductivity. The heater is
surrounded by an annular guard, which is independently heated to the same temperature as the
heater, in order to eliminate radial heat losses. Between the heater and the guard there is a small gap
(Figure 4.2). The thermal conductivity of the sample can be calculated using equation (4.7). This
method is particularly applied when measuring material with low thermal conductivity, and
available as thin samples with low length-to-diameter ratio.

Figure 4.2 Schematic diagram of guarded hot
plate apparatus for measurement of the
thermal conductivity [240].

4.2.2 Dynamic Methods
In dynamic or non-steady state methods, the temperature distribution varies with time.
The heat balance equation for transient conditions may be written as:
dt
dT
C T k
p
= V V (4.8)
where C
p
is the specific heat at constant pressure and is the density [233, 240]. For homogenous
materials, the thermal conductivity k can be treated as a constant and equation (4.8) can be written
as
dt
dT
C T k
p
= V
2

or

dt
dT
T = V
2
(4.9)
where is the thermal diffusivity and is equal to:
p
k
C
o

= (4.10)
The thermal diffusivity is a measure of the rate at which a material will approach thermal
equilibrium with its environment.
water-cooled heat sink
water-cooled heat sink
main heater guard guard
sample
sample
insulation
Chapter 4. Review of Thermal Conduction in TBCs

55
In dynamic methods, usually both the thermal diffusivity and the specific heat need to be
determined experimentally in order to derive the thermal conductivity. Accurate determination of
the specific heat becomes essential for many materials where the specific heat may vary with
temperature. On the other hand, measurement times in dynamic methods are generally short, so the
influence of heat losses on the measurement becomes less significant. In addition, heat losses can
be taken into account and be included as boundary conditions by incorporating a heat loss
coefficient in the heat balance equation. This heat loss coefficient can be eliminated by making
additional measurements under different experimental conditions. Hence, dynamic techniques are
particularly useful for measurements at higher temperatures. There are two main categories of
dynamic methods for measurement of the thermal conductivity: periodic or transient. In periodic
methods, a periodic thermal variation is imposed to the sample and certain propagation
characteristics of the resulting temperature wave are measured, whereas in transient methods a
single application of heat is supplied and the time for subsequent temperature changes in the sample
is measured.
Transient methods employ transient, non-periodic, heating to the sample, which is initially in
equilibrium with its surrounding environment. The change in thermal flux causes a temperature
change and the thermal diffusivity is evaluated by measuring the changes in temperature of the
sample with time. There are various methods widely employed for measurement of the thermal
conductivity. In general, these methods consist of generating a thermal flux either within the
sample, i.e with a suitable heater sandwiched between two identical samples, or at the surface of the
sample. The temperature distribution throughout the sample as a function of time depends on the
thermal diffusivity of the sample. In order to obtain the theoretical relationship between the sample
temperature distribution with time and the thermal diffusivity of the sample, the diffusion equation
must be solved, subject to the specific boundary conditions imposed by the experimental set up.
The most common transient methods used for thermal conductivity measurement are discussed
briefly below.
4.2.2.1 Transient Hot Strip
The transient hot-strip (THS) [241], transient hot-wire (THW)[242], transient plane-source (TPS)
[243] (or hot-disk) and dynamic plane source (DPS) [244] methods all apply the same principle for
determining the thermal properties of a liquid or solid material, in the sense that a single element is
used simultaneously as both a heat source and a temperature sensor. The experiment generally
Chapter 4. Review of Thermal Conduction in TBCs

56
involves passing an electrical current through the source/sensor, which results in a slight
temperature increase, and recording the voltage or resistance change over the source/sensor.
The THS method employs a thin metal strip as a planar heat source, placed between two identical
samples. A constant current is passed through the strip over a short period of time and the voltage
drop across the sensor over that time is recorded. By supplying a constant current, the output power
remains constant. The voltage change is due to the temperature increase which causes an increase in
the electrical resistance of the metal strip. Assuming an infinitely long planar heat source and no
influence to the temperature increase from the outer boundaries of the sample, the temperature
distribution with time in and around the metal strip can be expressed in terms of the thermal
diffusivity of the sample, the breadth of the heat plane source and the output power per unit area
[241]. The THS method has been applied to anisotropic solids [245], to insulating solids and
liquids [246] and to electrically conducting materials [247] .
4.2.2.2 Transient Plane Strip (Hot Disk)
In the TPS technique, the resistive element consists of a thin metal pattern shaped as a disk (Hot
Disk) or as a square (Hot Square) in-between two thin layers of the disk-shaped or square-shaped
insulator respectively (see Figure 4.3). As in the THS and THW techniques, the resistive element
acts as both the heat source and the temperature sensor. The sensor is sandwiched between two
identical samples. A current pulse is passed through the sensor, sufficient to cause a slight
temperature increase in the sensor and in the sample surrounding it. The length of the pulse is
normally chosen to be short enough so that the outer boundaries of the sample do not influence the
temperature increase of the element by any measurable extent. Hence, the TPS sensor can be
considered in contact with an infinite or semi-infinite solid throughout the transient recording. In
order to ensure that this assumption holds, the distance from the resistive element to the nearest
sample boundary must be equal or larger than the probing depth which is defined as [247]:
2
1
max
) (t c = A (4.11)
where c is a constant of the order of unity, is the thermal diffusivity of the sample and t
max
is the
total time of the transient experiment. Assuming that the resistivity of the metal element in the
sensor is temperature dependant, the temperature increase in the sensor will cause a change in its
resistance and a corresponding voltage variation over the sensor. Thus, the temperature increase of
the element can be deduced from recording its resistance or voltage variation over the period of the
transient event. The resistance of the sensor can be expressed by [243]:
Chapter 4. Review of Thermal Conduction in TBCs

57
( ) {1 [ ( )]}
o i sam
R t R T T A A = + + (4.12)
where R
o
is the resistance of the TPS element before the transient recording has been initiated and
is the temperature coefficient of the electrical resistance of the sensor. T
i
is the constant
temperature difference that develops almost momentarily between the sensor and the surrounding
sample due to the insulating layer around the sensor. This temperature difference will stay constant
after a short initialization time, t
i
, due to the constant power supply (see Figure 4.4). T
sam
() is
the temperature increase of the sample, which for a disk-shaped (Hot Disk) sensor is given by [248]:
3
2
( ) ( )
sam
q
T D
rk
A t t
t
= (4.13)
where q is the total power output from the sensor, r is the radius of the sensor and k is the thermal
conductivity of the sample. The dimensionless time function D() [248] varies as a function of ,
where is defined by:
r
t

t
= = (4.14)
where

2
= is defined as the characteristic time of the measurement, r is the radius of the sensor
and is the thermal diffusivity of the sample.
For times greater than t
i
, the temperature increase T (t) of the sensor is:
1 ( )
( ) ( ) 1
i sam
o
R t
T t T T
R
A A A t
|
| |
= + =
|
\ .
(4.15)
which can estimated by the recorded change in resistance R (t) of the sensor . By substituting T()
from equation 4.13 into equation (4.15) and defining:
3
2
q
C
rk t
= (4.16)
equation (4.15) can be rewritten as:
( ) ( )
i
T t T CD A A = + (4.17)
Hence, plotting the temperature increase of the sensor T (t) versus D (), should give a straight line,
for times greater than t
i
, and if the characteristic time has its correct value. From this particular
value of the characteristic time, , the thermal diffusivity, , can be obtained ( = r
2
/ ). The slope
(C) of the final plot is proportional to k
-1
. In this way it is possible to determine the thermal
conductivity and the thermal diffusivity from one single transient event. However, in order to be
Chapter 4. Review of Thermal Conduction in TBCs

58
able to determine both the thermal conductivity and the thermal diffusivity from one single transient
event, it is necessary to perform the experiment so that the time period of the transient event is close
to the characteristic time, .
For thin samples, the probing depth as defined by equation (4.11) may be greater than the distance
from the sensor to the nearest sample boundary, and hence the assumption that the heat source is
placed in an infinite medium is not valid. Hence, for such cases, it is necessary to take the boundary
condition at one of the outer surfaces into account and consider two-dimensional radial heat flow in
an infinite plate. Such a situation has been described by Gustafsson et al [249].


Figure 4.3 (a) Schematic diagram of experimental set up with the Hot Disk sensor and (b) photo of a
Nickel/Kapton Hot Disk sensor.



Figure 4.4 (a) Schematic diagram of experiment Hot Disk set up showing the heat affected zone and (b) typical
temperature vs. time distribution in sample and sensor.

4.2.2.3 Laser Flash Method
The flash method utilizes an energy pulse (from a laser, an optical flash tube or an electron beam) to
heat the surface of a sample for a short, finite time. The energy pulse is absorbed at the surface of
the sample and it results in a temperature increase in the sample. The time-dependent temperature
on the rear of the sample is measured (see Figure 4.5). This transient response of the sample can be
Heat affected zone
Sensor
Sample
30mm
(a) (b)
(b) (a)
Chapter 4. Review of Thermal Conduction in TBCs

59
related to its thermal diffusivity. Heat losses are minimized by making the measurements in a time
short enough so that very little cooling can take place. As shown by Parker et al [250], the
temperature history T(L,t) of the rear surface of a sample of length L can be expressed in a
dimensionless form as a fraction of the maximum temperature T
max
of the rear surface by:

=
|
|
.
|

\
|
+ =
1
2
2 2
max
exp ) 1 ( 2 1
) , (
n
n
L
t n
T
t L T
(4.18)
where n is an integer and is the thermal diffusivity of the sample. The rear face temperature
history following a flash heat pulse to the front sample face is shown schematically in Figure 4.5 as
described by equation (4.18). If the heat losses are negligible, the temperature at the rear face of the
sample will reach a maximum value and remain constant. The thermal diffusivity can be deduced
from equation (4.18) in two ways. When T/T
max
= 0.5, then
2
t/L
2
=1.38 and hence:
5 . 0
2
2
38 . 1
t
L
= (4.19)
where t
0.5
is the time required for the rear surface of the sample to reach half of the maximum
temperature rise. Alternatively, the time axis intercept of the tangent to the time-temperature curve
at the point of maximum gradient occurs at 0.48, which yields the relationship:
x
t
L

2
2
48 . 0
= (4.20)
where t
x
is the time axis intercept of the time-temperature curve. Using the measured thermal
diffusivity and knowing the specific heat, C
p
, and the density of the sample, , the thermal
conductivity, k, can be calculated as the product of these three parameters ( = k/C
p
). Direct
determination of the thermal conductivity is possible, if the amount of energy absorbed, q, at the
front surface of the sample is known. Then the product of the density and the heat capacity can be
deduced using the equation:
max
p
q
C
LT
= (4.21)
The above derivation assumes that heat losses from the sample are negligible. If heat losses are not
negligible, the maximum temperature reached by the sample will be reduced, it will take a longer
time to reach the maximum temperature and it will not remain constant but will fall off with time.
According to Cape and Lehman [251], the numerical factor in equation (4.19) decreases as the heat
losses from the sample due to radiation increase. The heat losses can be reduced by reducing the
Chapter 4. Review of Thermal Conduction in TBCs

60
sample thickness. However, for the derivation of equation (4.19), the energy pulse from the flash
lamp is assumed to be of negligible duration in comparison with the time it takes for the heat pulse
to propagate through the sample length. By decreasing the sample thickness, the duration of the
energy pulse will be comparable to the time it takes for the heat pulse to propagate through the
sample length, and, in such a case, the numerical factor in equation (4.19) increases with decreasing
sample thickness. For equation (4.19) to apply, the time it takes for the heat pulse to propagate
through the sample length must be at least ten times greater than the duration of the energy pulse. If
heat losses remain important after the maximum possible reduction in thickness, corrections to
equation (4.19)must be introduced. This is particularly important for poor thermal conductors or
higher temperatures, since the parameter determining the radiation heat losses is dependent on the
temperature, the emissivity and the thermal conductivity of the sample, and it is greater for poorer
thermal conductors and higher temperatures [252].
Another consideration regarding the flash method is the uniformity of the energy density across the
face of the sample. In the case that the energy density is not uniform, large errors in the
temperature measurement of the rear surface can arise, particularly if the temperature is measured
over a small area.
Moreover, the derived solution relies on the approximation that the energy pulse is absorbed
instantaneously in a small depth and that the temperature rise in the sample is small. For poorer
thermal conductors, in order to produce a measurable temperature increase on the rear face, a larger
energy pulse in the front face will be necessary. The depth at which the energy is absorbed will also
depend upon the pulse duration and thermal diffusivity of the sample, and it will be smaller for
poorer thermal conductors. Therefore, it is possible that the temperature rise in the front face of a
poor conductor would be so large that the aforementioned assumptions will not be valid. Moreover,
for a very large temperature rise at the front face it is possible to induce vaporization of the sample
in the front face.
The flash method produces an effective value for the thermal diffusivity of the sample, however, an
effective value of corresponding temperature is not determined. Usually a time averaged mean
temperature between the front and back surface temperatures up to the time that the rear surface
temperature has reached half its maximum value is taken as the effective temperature. However,
this approximation neglects the variation of thermal diffusivity with temperature. If the temperature
rise in a single measurement is kept small, then this problem is minimized. However, in a poor
thermal conductor, where large thermal gradients can be maintained during the measurement,
assigning an effective temperature to a particular measurement may prove difficult [253].
Chapter 4. Review of Thermal Conduction in TBCs

61


Figure 4.5 (a) Rear face temperature history following flash irradiation of the front face and (b) schematic
diagram of laser flash apparatus for thermal diffusivity measurement.

4.3 Thermal Conductivity of TBCs
The thermal conductivity and diffusivity of polycrystalline or single crystal stabilized zirconia, as
well as APS or EB-PVD zirconia stabilized coatings, has been the subject of numerous
experimental investigations because of the use of these materials as TBCs. Some of the
experimental measurements available in literature for YSZ are presented in Figure 4.6. The black
markers indicate the thermal conductivity of polycrystalline tetragonal 8 wt% YSZ hot pressed
powders, measured by steady state (filled symbols) and transient techniques (open symbols). The
thermal conductivity is in the range of 2.4-2.7 W m
-1
K
-1
at room temperature and decreases slightly
with temperature, which indicates that phonon conduction predominates [254, 255]. However, in
general, the temperature sensitivity of the thermal conductivity of dense hot pressed YSZ is not
great. The thermal conductivity of cubic zirconia is 2.0-2.4 W/m K and that of monoclinic zirconia
is about 5.4 W m
-1
K
-1
at room temperature [255].
In Figure 4.6, blue symbols represent conductivity measurements in EB-PVD 8 wt% YSZ TBCs
[256-258] and red symbols measurements on PS 8 wt% YSZ TBCs, both with about 10% porosity
[67, 112, 256, 259]. Filled symbols are for measurements conducted by steady state methods and
open symbols for measurements conducted by laser flash in air or another gas at atmospheric
pressure. The values reported cover a wide range, but in the majority of the measurements, the
variation of the thermal conductivity of YSZ with temperature indicates a predominantly phonon-
1.38 0.48
0
0.2
0.4
0.6
0.8
1
0 1 2 3 4 5 6 7 8
T
'

=

T
/

T
m
a
x
t
2
ot/L
2
Temp.
recorder
amplifier recorder
trigger
laser tube
sample
furnace
(a) (b)
Chapter 4. Review of Thermal Conduction in TBCs

62
based transmission mechanism up to about 1000 K (the conductivity is inversely proportional to the
temperature). At higher temperatures the thermal conductivity increases slightly with temperature,
probably due to photon conduction (radiation).
From the data available, it is obvious that the thermal conductivity is strongly dependent on the
amount and the nature of porosity. Both EB-PVD and PS coatings have lower thermal conductivity
values than polycrystalline YSZ hot pressed powder, mainly due to the presence of about 10-15%
porosity in these coatings. EB-PVD coatings tend to have higher thermal conductivity values, since
their microstructure is comprised of columns aligned parallel to the heat flux direction. Typical
thermal conductivity values for EB-PVD coating are around 1.4-1.7 W m
-1
K
-1
[256-258]. On the
other hand, in PS TBC splat boundaries and intersplat pores lie perpendicular to the heat flux
direction and offer significant resistance to the heat flow. The thermal conductivity of PS TBCs has
usually been reported to be in the range of 0.7-1.3 W m
-1
K
-1
[67, 112, 256, 259].
The wide range in the thermal conductivity values quoted in literature, even within coatings
produced with the same production method, is indicative of the sensitivity of the thermal
conductivity to the nature of porosity and the exact microstructural features in each coating,
although it must be recognized that experimental measurement errors are sometimes quite high. For
PS coatings, different spraying conditions may result in different microstructures and, even for the
same total porosity, may have quite different thermal conductivities. Therefore, it is very important
to assess or quantify these features in order to be able to make useful comparisons between different
coatings or materials.
For EB-PVD coatings, the thermal conductivity varies through the coating thickness, since the
microstructure changes progressively in this direction. For the first few tens of micrometers of the
coating, in the nucleation zone close to the bond coat interface, the grain size is small and there are
interfaces which hinder heat transfer. Further from the interface, the grains are columnar and
parallel to the heat flux [41]. Hence quoted thermal conductivity values should include coating
thickness for PVD coatings.
Furthermore, the technique used to evaluate the thermal conductivity must be considered when
assessing the reliability of the values obtained. Features of the techniques are discussed in section
4.2. Often the thermal conductivity is indirectly measured via the thermal diffusivity, using the
(volume) specific heat. While the specific heat is not usually very sensitive to microstructure, its
value may be difficult to obtain with high accuracy. Also, measurements are often carried out in
vacuum or under reduced pressure, which could lead to an underestimate-see section 8.4.
Chapter 4. Review of Thermal Conduction in TBCs

63
0.5
1
1.5
2
2.5
3
3.5
0 500 1000 1500 2000
Dense YSZ (steady state measurements)
Dense YSZ (laser flash measurements)
EB-PVD YSZ top coat (steady state measurements)
EB-PVD YSZ top coat (laser flash measurements)
PS YSZ top coat (steady state measurements)
PS YSZ top coat (laser flash measurements)
T
h
e
r
m
a
l

C
o
n
d
u
c
t
i
v
i
t
y

(
W

m
-
1

K
-
1
)
Temperature (K)

Figure 4.6. Thermal conductivity measurements for tetragonal dense 8 wt% YSZ , 8 wt% YSZ PS coatings
and 8 wt% YSZ EB-PVD coatings measured by steady state and laser flash experiments in air. Data are
from references [112, 254, 258-263].

Heat treatment can lead to an increase in the thermal conductivity of TBCs, both for PS [112, 259,
264, 265] and EBPVD coatings [254, 257, 262]. This is due to sintering of cracks, pores and
interfaces during heat treatment. In PS coatings, poor contact between lamellae in close physical
proximity can be improved by diffusional processes at high temperatures. These effects are clearly
of potential practical significance. If representative values of the thermal conductivity need to be
used for TBC design, the thermal conductivity should be determined after a heat treatment that will
result in a structure that will remain relatively stable in service.
Microstructural defects such as pores or interfaces constitute obstacles to through-thickness heat
transfer and can be deliberately introduced in order to lower the thermal conductivity. For PS
coatings, such features are inherent and can be optimized with tailored spraying parameters. For
EB-PVD coatings, morphologies designed to reduce in the thermal conductivity can by achieved by
generating inclined columns [50, 266, 267]. Such zig-zag microstructures have increased
Chapter 4. Review of Thermal Conduction in TBCs

64
porosity between the boundaries and results in a reduction of 20-40% in the thermal conductivity.
Another effective way of reducing the thermal conductivity of EB-PVD coatings is by producing
fine micro-crystallite structures, such as in the nucleation region of EB-PVD coatings, close to the
bond coat interface. This can be achieved by periodic re-nucleation of the EB-PVD coating growth
[268, 269]. Another approach involves layering the EB-PVD coatings by introducing interfaces or
density changes parallel to the ceramic bond coat interface. This is achieved by periodically
switching between high and low direct current bias during deposition, thus altering the degree of ion
bombardment [41, 268, 270]. Thermal conductivity reductions up to 40% have been achieved.
Changing the chemical composition and structure of the TBC can lead to a reduction of the thermal
conductivity if the changes result in increased phonon scattering or photon scattering/absorption-see
eqs.(4.1) and (4.4). The mean free path due to scattering of phonons by defects is given by equation
(4.3). It can be deduced that thereduction in the thermal conductivity due to phonon scattering
increases with defect concentration. Scattering by solute or impurities increases with the difference
in mass between host ions and foreign ions. The difference in ionic radius between the foreign ion
and the host ion, and also changes in atomic bonding, can affect the strength of the reduction in
thermal conductivity [271]. Clarke et al [271], argued that low thermal conductivity in TBCs is
favoured by: (a) high molecular weight, (b) a complex crystal structure, (c) non-directional bonding
and (d) a large number of different atoms per molecule. Consequently, the thermal conductivity of
TBC top coats can be reduced by introducing atomic level defects, such as vacancies, or
substituting Zr atoms with atoms of different mass or radius. Also a reduction in the thermal
conductivity can be achieved by the addition of dopant materials that increase the opacity of TBCs,
particularly in the infrared region, and hence reduce the radiative transport (colouring) [268].
The thermal conductivity of YSZ decreases with increasing yttria percentage. The introduction of
the stabilizer with a lower valence is accompanied by the incorporation of extra vacancies for
charge compensation, which provide an effective source of scattering of phonons [270]. The
thermal conductivity of stabilized zirconia is believed to be dominated by oxygen vacancy
scattering [262, 272]. Dopant ions, regardless of their valence, also create lattice strains due to the
difference in size between host and dopant atom, which provide an additional source of phonon
scattering.
For zirconia-based TBCs, many dopants (mainly rare earths) have been investigated. Possible rare
earth dopants for zirconia-based TBCs that have been considered include Ytterbia (Yb
2
O
3
) [18],
Erbia (Er
2
O
3
) and Gadolinia (Gd
2
O
3
) [220]. Replacing Yttria with 12 mol% Dysprosia, for an EB-
PVD coating resulted in a stable cubic phase and a 40% reduction in thermal conductivity [273].
Chapter 4. Review of Thermal Conduction in TBCs

65
Replacing Yttria with 8 mol% Erbia, Gadolinia, Ytterbia and Neodymia in EB-PVD coatings all
resulted in lower thermal conductivity, with reductions up to 45%, compared to 8 mol% Yttria, with
GaSZ having the lowest conductivity followed by NdSZ and YbSZ [268]. The effect was explained
as being due to the presence of ions of differing ionic radius and mass inducing phonon scaterring
and additionally by increased radiation absorption (colouring). Lanthana additions of 1-5 mol% to
YSZ top coats have also been reported to decrease the thermal conductivity by up to 60% at room
temperature [274].
The rare earth dopants mentioned above have lower valence than the zirconium host ion, so that
oxygen vacancies are introduced, contributing to lowering the thermal conductivity by promoting
phonon scattering. The difference in mass and atomic weight between host and dopant ion is also
believed to reduce the thermal conductivity, also by promoting phonon scattering. However,
studies by Raghavan et al [272, 275], which involved introducing heavier (Ta
+5
) and lighter (Nb
+5
)
pentavalent ions of similar sizes in zirconia, suggest that phonon scattering due to mass difference
alone has little effect on thermal conductivity. The introduction of pentavalent dopants in zirconia
does not introduce any oxygen vacancies, and no reduction in the thermal conductivity was
observed by doping with Ta
+5
or Nb
+5
. This suggests that such vacancies have strong phonon
scattering effects.
Multi-component zirconia-based TBCs of Y
2
O
3
-ZrO
2
, co-doped with additional paired rare earth
oxides (Nd
2
O
3
-Yb
2
O
3
or Gd
2
O
3
-Yb
2
O
3
or Sm
2
O
3
-Yb
2
O
3
), with the paired oxides added in a ratio of
1:1 and with a total dopant concentration of about 10 mol%, showed decreases in thermal
conductivity of up to 50%, compared to YSZ with similar dopant concentration. The added paired
oxides form defective oxide clusters, which are thermodynamically stable and reduce the thermal
conductivity, as well as being sinter-resistant [276, 277].
Alternative TBC materials with lower thermal conductivities have also been investigated. Hafnia-
based TBCs with 27wt% Yttria has been reported to have lower thermal conductivity and be more
sinter-resistant [278]. Pyrochlore structure rare-earth zirconates are reported to have lower intrinsic
thermal conductivity and higher thermal expansion coefficient than zirconia [230, 279]. Coatings
of Pyrochlore/YSZ PS double layers reportedly show improved temperature capaibilities and cyclic
lifetime, compared to conventional PS YSZ coatings [280].
4.4 Modelling of Heat Flow in Plasma-Sprayed TBCs
The problem of predicting the effective thermal conductivity of materials with inhomogeneities or
inclusions is of great interest for various areas of materials science and it has been addressed by a
Chapter 4. Review of Thermal Conduction in TBCs

66
large number of authors. A few models will be mentioned briefly, focusing on models for materials
with gas-filled inhomogeneities. Materials with randomly distributed pores have been modelled by
several authors [281-283]. Shafiro [283] and Sevostianov [282] derived the effective thermal
conductivity of materials with diverse ellipsoidal inclusions, non-randomly oriented, assuming non-
interacting inclusions. This approximation is valid for materials with dilute concentration of
inclusions. For higher porosity levels, self-consistent [284, 285] and effective medium [286]
models have been developed. Bristow [287] derived the effective thermal conductivity (along with
the effective elastic moduli) in the case of crack-like pores, and in the limit of small crack density.
Other approaches include modelling incorporating a contact resistance [288-290] or modelling
periodic structures [291].
McPherson [288] developed a model specifically for plasma sprayed coatings, and other layered
systems composed of solid lamella with small contact areas between them. The model assumes two
independent heat fluxes, one through the solid contacts areas and another through the pores. A
thermal resistance is assigned to the contact areas. This model was further developed to incorporate
thermal resistance of the lamella [289] and oxidation of the contact areas [290]. Lu et al [292]
extended the original shear lag model [291] developed for composite periodic structures for thin
cracks in a uniform matrix.
An analytical and numerical model has been developed by Golosnoy et al [293] for simulating the
heat flow in plasma sprayed thermal barrier coatings. This model will be described in more detail
in the next section.
4.4.1 Morphological representation of PS TBCs
For modelling the effective thermal conductivity, the complexity of the porous space has to be
reduced to several dominant elements. The plasma sprayed TBC structure consists of pancake-
shape splats of high aspect ratio, with poor contacts between them. This structure is represented by
a periodic array of infinite lamellae, separated by pores with bridges between them. Pores in
successive layers were assumed to have aligned arrangement, as depicted in Figure 4.7. The
structure can be described by 6 characteristic parameters (4 geometrical and 2 physical). The
geometrical parameters are the height L
v
and thickness L
h
of the unit cell, the thickness of the pore
d
v
, and the inter-bridge or interpore distance d. The two physical parameters are the thermal
conductivity of the solid lamella material k
0
, and the thermal conductivity of the pore k
p
. For the
above simplified structure geometry, the normalized thermal conductivity of the material as a whole
k
eff
/k
0
, is a function of three geometrical parameters (the normalized bridge area (1-d/L
h
)
2
, the
Chapter 4. Review of Thermal Conduction in TBCs

67
normalized inter-bridge distance L
h
/L
v
, and the normalized bridge height d
v
/L
v
) and one physical
parameter (the normalized thermal conductivity of the pore k
p
/k
0
).

lamellae
interlamellar void
vertical crack
contact bridge

Figure 4.7 Modelled periodic lamellar structures, showing the pore configurations.


L
v
L
h
( L
h
- d ) d/2
d
v
k
p
k
0

Figure 4.8 Modelled unit cell for open porosity (isolated bridges) showing (a) perspective view and (b) vertical
section.
4.4.2 Physical Assumptions
Heat transfer in the polycrystalline solid occurs through phonon conduction and in the gas in pores
by molecular collision conduction. In addition, radiation heat transfer occurs through the
semitransparent material and the gas in the pores. The total heat flux through the coating can hence
be split in two parts: phonon conduction through the polycrystalline solid (or collision conductivity
in the gas) and radiative heat transfer. There are no free electrons in the constituents to consider.
Typically, TBCs consist of tetragonal Zirconia partially stabilized by 8wt% yttria. For this material,
the thermal conductivity is only weakly dependent on temperature over the range of interest, and for
dense polycrystalline solids has a value of about 2.25 W m
-1
K
-1
[255, 294].
The gas in the pores conducts heat through molecular collisions. Since pore thickness, d
v
, is
comparable to the mean free path of molecules in a typical gas at atmospheric pressure ( ~100nm),
convection in pores can be neglected. The thermal conductivity of the gas in the pores, k
g
, can be
estimated by a simple analytical expression given by Lu et al [295]:
(a) (b)
Chapter 4. Review of Thermal Conduction in TBCs

68
P d
BT
k
k
v
g
g
+
=
1
0
(4.22)
where k
g
0
is the normal conductivity of the gas at the temperature concerned, P is the pressure, T
is the temperature (in K) and B is a constant, with a value of 2.5 10
-5
Pa m K
-1
for air. Figure 4.9
shows measured values for k
g
0
for air [296], together with calculated values for k
g
, for different pore
thicknesses and pressures. It can be seen that both pore thickness and gas pressure can have
significant effects on the pore conductivity, as well as temperature.
0
0.05
0.1
0.15
400 600 800 1000 1200 1400 1600 1800 2000
k
g
(d
v
=0.1, P=1)
k
g
(d
v
=1, P=1)
k
g
(d
v
=0.01, P=1)
k
g
(d
v
=0.1, P=0.01)
k
g
(d
v
=0.1, P=40)
k
g
0
, Incropera & DeWitt, 1996
G
a
s

t
h
e
r
m
a
l

c
o
n
d
u
c
t
i
v
i
t
y
,

k
g

(
W

m
-
1

K
-
1
)
Temperature, T (K)
B = 2.5 10
-5
Pa m K
-1

Figure 4.9 Dependence on temperature of the thermal conductivity of the pores showing the effect of pore
thickness and gas pressure on the conductivity of air[296], according to equation(4.22).
The usual approach for incorporating the contribution of heat transfer by radiation in such materials
is to introduce a radiative conductivity which is additive to that of conduction [234, 237]. The
radiative conductivity is given by equation (4.4) and it is assumed that the total thermal conductivity
of the assemblage is the sum of the effective conductivity of the assemblage due to phonon
conduction and the radiative conductivity (given by equation 4.4). The mean free path l
r
(or
penetration depth) in equation (4.4) is given by,

l
r
1
= (4.23)
Chapter 4. Review of Thermal Conduction in TBCs

69
where = + is the extinction coefficient and is equal to the scattering () plus the absorption ()
coefficients [237]. Using this formulation, uniform radiation scattering is assumed and any
wavelength dependence or boundary effects are neglected. In typical plasma sprayed TBC material,
scattering is very strong and it tends to dominate, so that ~ .
For sintered zirconia, the penetration depth has been reported [297] to be about 50 m. For plasma
sprayed TBCs it is expected that the penetration depth will be lower than that for sintered zirconia.
There is some data available for the refractive indices, n, and the penetration depths for zirconia and
PS YSZ [4, 234, 298, 299]. From this data it can be inferred than appropriate values for the
penetration depth l
r
, and the refractive index n for plasma sprayed zirconia are of the order of 10 m
and 2.2 respectively. Using these figures, the radiative conductivity and the conductivity of the gas
in the pores as a function of temperature are shown in Figure 4.10, for a pore thickness of 0.1 m
and a gas pressure of one atmosphere.
0
0.05
0.1
0.15
0.2
400 600 800 1000 1200 1400 1600 1800 2000
Normal gas conductivity, k
g
o
Conductivity of gas within pores, k
g
Radiative conductivity, k
rad
T
h
e
r
m
a
l

c
o
n
d
u
c
t
i
v
i
t
y
,

k

(
W

m
-
1

K
-
1
)
Temperature, T (K)
Air in pores
P = 1 atm
d
v
= 0.1 m
penetration depth, o = 10 m
refractive index, n = 2.2

Figure 4.10 Dependence on temperature of the contributions to the thermal conductivity showing the
radiative conductivity and conductive of the gas within the pores, for air at atmospheric pressure with a pore
thickness of 0.1m, according to equation 4.22.

4.4.3 Numerical Model
The unit cell is a parallelepiped. The modelled volume incorporates the main microstructural
features. A 3-D Cartesian mesh with uniform spacing was superimposed on the cell, with
approximately 100 volume elements in each direction, giving a total of about 10
6
. The steady state
Chapter 4. Review of Thermal Conduction in TBCs

70
temperature distribution was obtained by relaxation of the transient heat transfer equation [300, 301]
(see equation (4.8)). The temperature dependence of the thermal conductivity k has been neglected
in most of the current work, except that values have been employed corresponding to low (~300 K)
and high (~1500 K) temperature regimes. The focus here is mainly on the influence of geometric
and microstructural factors. The evolving temperature distribution was obtained using an implicit
method based on the ADI technique [300, 301], with a 1% convergence criterion.
The main input parameters concerning geometry are the splat thickness (=L
v
-d
v
), inter-splat bridge
area (S
br
/S
tot
), inter-splat pore thickness, (d
v
) and the inter-splat pore volume fraction, f
p
, (which
determines the bridge spacing, L
h
). These values can be inferred from various types of
experimental measurements, including inspection of micrographs, densitometry and small angle
neutron scattering (SANS). The SANS technique [119, 136, 137, 302] gives information about the
spacings between planes which act as strong scattering centres, such as splat surfaces.
In general, the splat thickness depends only weakly on spraying conditions and the sprayed material
[303-305]: a typical value for sprayed zirconia would be about 2 m and this figure was used in the
present work. A reference value used for the inter-splat pore thickness was 0.2 m, which is
consistent with microscopy [134] and SANS [119, 136, 137, 302] evidence for sprayed zirconia.
From densitometry and SANS data, the inter-splat pore volume fraction can be estimated to be
about 3%. The relationship between the bridge area, pore volume fraction, splat thickness and
inter-splat pore thickness can be expressed
v p
br
tot v
1
L f
S
S d
= (4.24)
In practice, the bridge area and the inter-bridge spacing might vary over quite a range, particularly
as a consequence of post-deposition sintering effects, so the effects of altering these parameters was
investigated. Typical initial values [134] might be 10-30% and 5-10 m respectively.

4.4.4 Two Flux Region Analytical Model
An analytical two flux region (TFR) model was developed [293] in order to describe the heat
transfer in the porous periodic structure. The unit cell is divided into two regions. The first zone
includes the bridge, and in this region it is assumed that heat flux takes place only through the solid
material, with no conduction through the pore. The effective conductivity of this region depends on
the area S, which is the area of this contact region. The second region contains two layers of solid
material and the pore between them (see Figure 4.11). The sizes of these regions are not known a
Chapter 4. Review of Thermal Conduction in TBCs

71
priori. They are established by maximizing the conductance of the assembly. The fluxes in each
region are assumed to be independent, so that:
2 1
) ( ) ( k
S
S S
S k
S
S
S k
tot
tot
tot
eff

+ = (4.25)
where k
eff
(S) is the effective conductivity of the material, and k
1
(S) and k
2
are the thermal
conductivities of the corresponding regions and S
tot
is the total cross sectional area of the unit cell.
k
1
(S) is a function of the area S, whereas k
2
is independent and can be given by a standard series
addition expression:
1
0
2
1 1
1

|
|
.
|

\
|
+
)
`

=
p v
v
v
v
k L
d
k L
d
k (4.26)
where k
0
is the total conductivity of the lamellae, k
p
is the total conductivity of the pore, d
v
is the
pore thickness and L
v
the height of the unit cell. The required condition of maximum conductance
of the assembly can be expressed as:
) ( max
] , [
S k k
eff
S S S
eff
tot bri
e
= (4.27)
In order to find k
eff
, k
1
must be modelled as a function of the size of the region (area S), since k
2
is
independent of S. The total conductivity of contact region k
1
contains two terms: the conductivity
of the bridge itself, and the conductivity of the solid region above and below the bridge. The
conductivity of the bridge itself, k
br
, can be given by:
0
k
S
S
k
br
br
= (4.28)
where S
br
is the cross section area of the bridge. The heat flow in the region above and below the
bridge is being funneled towards (or away from) the bridge. The effective conductivity in this case
is often referred to as a contact resistance, emphasizing that it is the value for the resistivity of an
infinitesimally thin bridge. Contact resistances of this type have been investigated by several
authors [288, 290] [289, 295]. According to McPherson [288] and Li [289], for the limiting case of
small contact areas (d~L
h
), the effective thermal conductivity, referred to as contact resistance k
res
,
is given by:
1
2
0
) ( 2
1

|
|
.
|

\
|

t
+ =
v h
h
res
L d L
L
k k (4.29)
Chapter 4. Review of Thermal Conduction in TBCs

72
where L
h
is the width of the unit cell and d is the pore length between bridges. For the limiting case
of large contacts (d<<L
h
), the modified shear-lag model [295] [291]gives:
1
2
3
2
3
2
3
0
) ( 2
) 2 (
1

|
|
.
|

\
|


+ =
v h h
h
res
L L d L
d L d
k k (4.30)
A simple interpolation process between equations (4.29) and (4.30) gives the contact resistance as a
function of the region surface and bridge (contact) areas:
1
2
1
2
3
0
2 1 2
1 ) (

|
|
|
|
.
|

\
|
)
`

t +
t
+ =
br v
br
br
res
S L
S
S
S S
S
S
k S k (4.31)
The conductivity of the first (contact) region can be derived by a simple series addition expression:
1
0
1
) (
1
1 ) (

|
|
.
|

\
|
+
)
`

=
k S
S
L
d
S k L
d
S k
br v
v
res v
v
(4.32)
Consequently, equation 4.32 can be solved, since all functions are known. The appropriate value
for the area S of the first (contact) region can be determined so that the conductance of the unit cell
is maximized. The maximum will occur at some intermediate value of the fractional bridge area or
at the total bridging limit (S
br
= S
tot
). This is illustrated in Figure 4.11.


Figure 4.11. Schematic depiction of the Two Flux Regions (TFR) Model, showing (a) heat flow paths and (b)
plot of how the effective conductivity varies with the area S through which the heat flux is funneled in the
vicinity of a bridge.

(a) (b)
region
1
region
2
max
Chapter 4. Review of Thermal Conduction in TBCs

73
4.4.5 Predicted Conductivity
A comparison is made in between predictions from the numerical model, the TFR model and two
other analytical models [288] [292], using typical input data for a TBC in the absence of radiation
(radiative contributions have not been incorporated into the two models from previous work). It can
be seen that the TFR model is in good agreement with the numerical modelling predictions, while
the other two can lead to some substantial errors, depending on the regimes of bridge area and inter-
bridge distance concerned. This arises from certain simplifications and approximations
incorporated in them.
0
0.5
1
1.5
2
0 0.2 0.4 0.6 0.8 1
Current work, 3D numerical
Lu et al, modified shear lag
Current work, TFR model
McPherson, contact resistance
E
f
f
e
c
t
i
v
e

t
h
e
r
m
a
l

c
o
n
d
u
c
t
i
v
i
t
y
,

k
e
f
f

(
W

m
-
1

K
-
1
)
Normalised bridge area, S
br
/ S
tot
L
h
/L
v
= 1
d
v
/L
v
= 0.1
k
p
= 0.00008 W m
-1
K
-1
k
0
= 2.25 W m
-1
K
-1
(a)

0
0.5
1
1.5
2
2.5
0 0.5 1 1.5 2 2.5 3
Current work, 3D numerical
Current work, TFR model
Lu et al, modified shear lag
McPherson, contact resistance
E
f
f
e
c
t
i
v
e

t
h
e
r
m
a
l

c
o
n
d
u
c
t
i
v
i
t
y
,

k
e
f
f
,

(
W

m
-
1

K
-
1
)
Normalised inter-bridge distance, L
h
/ L
v
S
br
/S
tot
= 20.25%
d
v
/L
v
= 0.1
k
p
= 0.00008 W m
-1
K
-1
k
0
= 2.25 W m
-1
K
-1
(b)

Figure 4.12. Effective conductivity for different models [288] [292], with no radiative contribution, as a function
of (a) bridge area and (b) inter-bridge distance.

Chapter 5. Experimental Procedures

74

5 Experimental Procedures
5.1 Characterization of Starting Materials
Powder size and morphology, as well as composition, can have a significant effect on the behaviour
during spraying and thus influence the properties of the coatings. Therefore, characterization of the
starting materials is important in order to understand coating properties.
5.1.1 Chemical Composition
The powders utilised were supplied by SULTZER METCO Inc. The bond coat material was
CoNiCrAlY (Amdry 995C). A summary of compositions and acronyms of the top coat powders is
presented in Table 5.1. Chemical compositions of powders, as given by the supplier, are
summarised in Table 5.2. The powders were manufactured by the homogeneous spherical-powders
method (HOSP).
Product Composition Acronym
204NS-1 yttria-stabilised zirconia YSZ
204NS yttria-stabilised zirconia YSZ
205NS ceria-stabilised zirconia CSZ
SPM6-2444 dysprosia-stabilised zirconia DSZ
AE8321 yttria-lanthana stabilised zirconia YLSZ
AE8170 Al
2
O
3
-doped Yttria-stabilised zirconia Al
2
O
3
-doped YSZ
AE9018 Al
2
O
3
-doped Yttria-stabilised zirconia Al
2
O
3
-doped YSZ
Table 5.1. Acronyms and compositions for the different top coat powders.
5.1.2 Particle Size Distribution
A Malvern P580 Mastersizer E machine was used for particle size analysis. Particles were
suspended in an aqueous solution and kept in suspension by a means of a stirrer. The particle size
distribution is obtained by shining a laser beam through the suspension. The suspended particles
scatter the laser beam at different angles, depending on the size: small particles scatter at large
angles, whereas large particle scatter at lower angle. A detector gathers the scattered light and the
Chapter 5. Experimental Procedures

75
angular distribution is analysed to give the particle size distribution. The distributions are presented
in Figure 5.1. The mean particles sizes are between 50-70 m for all powders.
The microparticle population in the powders is shown in Figure 5.2. Microparticles do not
contribute much to the total mass of the powder, and do not significantly affect much the mean,
median and mode. However, they contribute to nozzle build-up during spraying, by clogging up the
nozzle and solidifying at the nozzle exit. To avoid nozzle build up, an additional carrier gas flow
rate was introduced into the nozzle, which forced particles through the nozzle and prevented them
from depositing at the nozzle exit (see Figure 5.5). All powders, expect 204NS-1, needed
additional carrier gas injection at long spraying times to avoid nozzle build-up. Powders with large
particle size distribution where more prone to nozzle built up. These two attributes are thought to
be responsible for the nozzle built up. It is concluded that in order to avoid nozzle build up, it is
important to have a narrow particle size distribution and a low population of microparticles. The
average particle size does not appear to influence build up, however, for successful melting of the
ceramic powder a relatively small average particle size is desired.



Figure 5.1 Particle size distributions for zirconia-based top
coat powders.
Figure 5.2 Distribution densities of microparticles in
zirconia-based top coats.




Chapter 5. Experimental Procedures

76
Powder Product
Chemical
Composition
(wt %)
204NS-1 204NS 205NS SPM6-2444 AE8312 AE8170 AE9018
Al
2
O
3
0.01 <0.01 <0.01 0.02 0.88 0.8
CaO 0.04 <0.01 0.02 0.01 <0.01 <0.02
CeO
2
25.15 <0.02 <0.02
Cr
2
O
3
<0.01 <0.02
Fe
2
O
3
0.01 <0.01 <0.01 0.03 0.01 0.01 0.01
HfO
2
1.7 1.51 1.2 1.5 1.54 1.59 1.59
Dy
2
O
3
10.1
La
2
O
3
6.58
MgO 0.02 <0.01 0.01 <0.01 <0.01
Na
2
O 0.01 <0.01 <0.01
P
2
O
5
0.01
SiO
2
0.3 0.07 0.0 0.10 0.04 0.05 0.05
TiO
2
0.04 0.09 0.01 0.10 0.10 0.10 0.10
U + Th <0.01 <0.01 0.01 <0.01 <0.01 <0.01
Y
2
O
3
7.8 7.6 2.4 6.50 7.51 6.5
ZrO
2
balance balance balance balance balance balance balance
Table 5.2. Chemical composition of ceramic powders supplied by SULZER METCO Inc.
5.1.3 Powder Morphology
The morphology of the top coat powders used in this study was examined using a JEOL 5800LV
scanning electron microscope. The acquisition of the images was done using a Noran Voyager
Analysis Software. To prevent charging, coatings were sputtered with gold prior to analyses in the
SEM. Figure 5.3 shows SEM micrographs of the top coat powders. All powders consist of
spherical particles with a range of different morphologies: hollow spheres, spheres with a dense
outer shell and a porous core, agglomerate spheres with smaller particles in the core and spheres
Chapter 5. Experimental Procedures

77
with a shell-structure. Representative micrographs of individual particles showing the different
particle morphologies present in all powders are shown in more detail in Figure 5.4. AE8170 and
AE8321 powders consist mainly of spheres with a dense shell and a porous core, while the YSZ
(204NS, 204NS-1) powders, mainly contain agglomerate spheres with smaller particles in the core
(Figure 5.4). However, quantitative evaluation of the morphology of the spherical particles could
not be achieved by SEM and was beyond the scope of the present study. A spherical geometry is
beneficial for good flow in the powder feed line, which in turn can lead to high feed and deposition
rates.
(a) (b) (c)
(d) (e) (f)
Figure 5.3 SEM micrographs showing the morphology of the zirconia-based top coat powders: (a) YSZ
(204NS-1), (b) YSZ (204NS), (c) CSZ (205NS), (d) DSZ (SPM6-2444), (e) YLSZ (AE8321) and (f) Al
2
O
3
-
doped YSZ (AE8170).
(a)

(b)

(c)

(d)

Figure 5.4 SEM migrographs of zirconia-based top coat powders showing the range of particle morphologies: (a)
agglomerate sphere (204NS), (b) hollow sphere (SPM6-2444), (c) shell-structured sphere (SPM6-2444) and (d) sphere
with a porous core.
Chapter 5. Experimental Procedures

78

Figure 5.5. Photograph showing nozzle build-up after spraying.

5.1.4 Phase Constitution of Starting Powders
Xray diffraction (XRD) measurements were carried out using a computer-controlled Phillips
PW1710 diffractometer, with Cu
Ko
radiation (=0.154 nm), 40 kV accelerating current and a
40 mA filament current. The incident beam optical conditions were set up with an anti-scatter slit
of 1
o
and a 15 mm horizontal mask. For the diffracted beam, a 0.2 mm receiving slit and a 1
o

divergence slit were used. The apparatus was regularly calibrated using a silicon control sample.
Quantitative phase analysis was performed using Rietveld analysis.
Phase (%) /
Powders
Monoclinic
zirconia
Cubic
zirconia
Tetragonal
zirconia
Ceria
204NS-1 (YSZ) 16 84 -
204NS (YSZ) 32 68 -
205NS (CSZ) 12 28 53 7
SPM6-2444(DSZ) 31 69
AE8321(YLaSZ) 8 28 64
AE8170 (Al
2
O
3
-dopedYSZ) 12 23 65
Table 5.3 Phase constitution of as-received top coat ceramic powders.

The X-ray diffraction spectra of the as-received zirconia-based top coat powders are shown in
Figure 5.6. The phase constitution of the as-received powders is, in general, in good agreement
with the expected compositions according to the available phase diagrams [231, 306-308]. Mostly,
the powders consist of a mixture of monoclinic and tetragonal zirconia. The x-ray data confirm that
the additives result in partial stabilization of the tetragonal phase of the zirconia polymorph. For the
Chapter 5. Experimental Procedures

79
Al
2
O
3
-doped YSZ powder (AE8170) no traces of Alumina were detected. This is because of the
very low Alumina content (0.88 wt%), which is not enough to show in x-ray scan. In the CSZ
(205NS) powder, some of the ceria is not into solution and appears as a separate peak in the scan.
0
2000
4000
6000
8000
10000
12000
14000
25 35 45 55 65 75
YSZ (204NS-1)
YSZ (204NS)
CSZ (205NS)
DSZ (SPM6-2444)
YLSZ (AE8321)
Al
2
O
3
-doped YSZ (AE8170)
C
o
u
n
t
s
2u ()

Figure 5.6 X-ray spectra of as-received top coat ceramic powders.
5.2 Coating Production
5.2.1 Substrate Preparation
The substrate materials used in the present study were mild steel and a Nickel base superalloy.
Substrates of mild steel had thicknesses of 1-1.5 mm. The Nickel base superalloy was Nimonic80A
(Special Metals Limited, Hereford, UK) and it was used in rectangular blocks of 50 mm in
thickness. A chemical analysis of the substrate for Nimonic80A, as given by the supplier, is
presented in Table 5.4.
cubic Zr
2
O
tetragonal Zr
2
O
monoclinic Zr
2
O
cubic CeO
Chapter 5. Experimental Procedures

80
Plasma spraying requires careful surface preparation. The preliminary treatment consisted of grit
blasting and cleansing. The main purpose of grit blasting is to obtain enough surface roughness to
ensure proper mechanical anchoring of the coating. Also, inorganic contaminants and oxides, that
always exist on metal substrates, can be removed by grit blasting. The blasting equipment utilised
was a Guyson Blast cleaning cabinet, with Al
2
O
3
500 m grit and a blasting pressure of 100 psi.
When grit blasting was performed on substrates in sheet form, the process introduced curving of the
substrate away from the blasted surface, as a result of residual compressive stresses. To remove this
deformation, the back surface was grit blasted as well. The cleansing process removed residual grit,
using compressed air and involved thorough cleaning with a degreasing agent. This process was
done immediately before spraying was carried out, to avoid recontamination. Over-sprayed coatings
were sprayed directly onto the sprayed coating, without any prior surface preparation.
For certain experiments, in order to facilitate top coat debonding after spraying, an aluminium
plasma sprayed coating was deposited onto the substrate, prior to the deposition top coat. The
aluminium powder used for this purpose was 1002 powder supplied by Newmet Koch. The
presence of the 100 m thick Aluminium coating facilitated detachment of the plasma sprayed top
coat, as explained in section 5.3.1.
Chemical composition of Nimonic Alloy 80A
weight % (except where stated ppm)
C Si Mn P S Al Ag (ppm) B Bi (ppm)
0.081 0.32 0.07 <0.003 <0.001 1.4 0.1 0.003 <0.1
Co Cr Cu Fe Ni Ti Pb (ppm) Ni+ Co
0.08 19.42 0.03 0.78 Bal 2.52 1.7 75.28
Table 5.4. Chemical composition of Nimonic 80A alloy, as provided by the supplier.
5.2.2 Plasma Spraying
Spraying was performed using a Plasma Technik AG VPS unit. The spraying unit is contained in a
double skinned, water cooled, cylindrical stainless steel chamber, with diameter of approximately
1.5 m and length of around 2 m. The plasma spraying was performed either at low pressure (after
evacuating the chamber-LPPS) or in air (APS).
The gun pattern and speed are programmed from a control console. The spray unit is controlled by a
dedicated computer system. A Windows based control software allows adjustment of the different
parameters of the spraying process. Table 5.5 gives a summary of the spray parameters for
deposition of coatings used in this study. For some of the ceramic powders, nozzle build up was
observed during spraying. In order to overcome this, additional gas (Ar) was injected at the nozzle
Chapter 5. Experimental Procedures

81
exit and some of the spraying parameters were modified, as specified in Table 5.5. The spraying
parameters used for depositing aluminium pre-coats were identical to those used for the
CoNiCrAlY bond coat- see Table 5.5.
To promote homogeneous coverage, gun motion in the x-y plane was programmed using a raster
pattern (see Figure 5.8). The raster pattern was repeated for a number of cycles. The number of gun
cycles, and the waiting time in-between each cycle was varied according to the requirements of the
coating being deposited. During APS, jets of compressed air were in some cases applied to the
backsurface of the substrate, for cooling purposes.

Figure 5.7 The plasma spray rig [309].

Figure 5.8 Raster pattern in the x-y plane for one cycle of gun movement.


substrate
gun
movement
one cycle
end point start point
one pass
gun
Chapter 5. Experimental Procedures

82
Table 5.5 Parameters employed in the plasma spraying process. Coatings AE8321 and AE9018 were
sprayed by Sulzer Inc.
5.3 Sample Preparation Procedure
5.3.1 Detaching the Top Coat
To carry out characterisation of the top coats, it was necessary to detach them from the substrate. In
order to achieve this, an aluminium coating of about 100 m thickness was sprayed on to substrate
prior to the deposition of the top coat. The attached coatings were placed in a hydrochloric acid (32 )
solution at 50
o
C, which dissolved the aluminium layer, allowing the top coat to detach (after
15-20 min). In the absence of the aluminium coating, the hydrochloric acid dissolved the substrate,
Material
Parameters
CoNiCrAlY YSZ
(204NS-1)
DSZ (SPM6-2444) /
CSZ (205NS) /
YSZ (204NS)/
Al
2
O
3
-doped YSZ
(AE8170)
YLaSZ (AE8321) /
Al
2
O
3
-doped YSZ
(AE9018)
Chamber pressure (mbars) 200 (Ar) Atmospheric
(air)
Atmospheric (air) Atmospheric (air)
Stand- off distance (mm) 270 105 105 120
Plasma
Arc Current (A) 500 750 750 580
Volatge (V) 50 50 50 64
Gun speed
(mm s
-1
)
100 55 55
Nozzle diameter (mm) 8 8 8 6
Plasma gas mixture
Argon flow Rate (NLPM) 50 50 50 74
Hydrogen flow rate (NLPM) 10 8 8 23
Powder Feed
Carrier gas flow rate (Ar)
(NLPM)
2 5 5 12.8
Additional gas flow rate (Ar)
(NLPM)
2
Stirrer speed (gr min
-1
) 66 80 80 80
Disc speed (%RPM) 17 20 20
Chapter 5. Experimental Procedures

83
but, the procedure was more time consuming (45-60 min) and often led to contamination of the top
coat with iron chloride. To remove any residues of iron chloride, the top coat was placed in a
hydrochloric acid solution (1 %) at 50
o
C, which was stirred continuously using a magnetic stirrer
for about 1 h, or until all residue was removed. Subsequently the top coat was thoroughly cleaned
with a degreasing agent in an ultrasonic bath and then dried.
5.3.2 Metallographic preparation
Cutting of the top coat was performed using a diamond cutting wheel in a slow speed CAPSO Q35
precision cutting machine, which allowed parallel cuts to be made, ensuring no brittle fracture of
the top coat while cutting. Fracture surfaces for examination by SEM were produced by brittle
fracture of detached top coats using three-point-bending. Mounting of TBCs for microstructural
evaluation was performed using Struers Epovac vacuum impregnation equipment. The epoxy used
was Struers Caldofix. In order to ensure low viscosity of the resin for better impregnation, the resin
was preheated to 100
o
C. Vacuum impregnation was necessary in order to ensure minimum
smearing of small pores and material removal during subsequent metallographic preparation.
Metallographic preparation of the surface and/or cross section of detached top coats was required
for some of the experiments. The procedure involved grinding with a coarse SiC grinding paper
(120 grit), until a homogenous surface was achieved, followed by finer grinding paper (800 grit and
1200 grit). Diamond polishing to 1 m finish was performed with polycrystalline diamond paste,
using polishing cloths for hard ceramics. Care was taken to minimise the polishing and grinding
times, in order to avoid unnecessary material removal and smearing of small pores during the
procedure.
For measurement of the thermal conductivity by the steady state method, samples of detached top
coats with lateral cross section of 35 30 mm were prepared. The detached TBCs were slightly
curved due to residual stresses. In preparing samples for the bi-substrate technique, special care
was taken to ensure that sample curvature was minimized during the procedure. In addition, an
attempt was made to ensure that the thickness of the sample remained uniform during preparation.
The thickness of the detached coatings after polishing was measured at 16 different points. In
general, point to point variations were below 5%. Following this procedure, it was possible to
achieve a good surface finish on detached coatings, with a roughness R
a
of less than ~1 m. The
debonded coatings had thicknesses ranging from 0.5 to 2.0 mm.
Chapter 5. Experimental Procedures

84
5.4 Coating Characterisation
5.4.1 Surface roughness
Surface roughnesses were determined using a WYKO RS-2 optical interferometric profilometer.
This apparatus characterises the surface roughness and produces a topographic map of the surface.
An incident monochromatic optical beam is scanned across the sample surface and a set of fringes
are produced inteferometrically. The fringes result from constructive or destructive interference
depending on the optical path difference between a reference beam and that reflected from the
sample. The fringe pattern is analysed and a three-dimensional topographic image of the surface is
generated. The lateral resolution is about 1 m and the depth resolution about 5 nm.
Roughness can be characterized by the values of R
a
and R
z
. The average surface roughness, R
a
, is
the average distance of departure of the profile from the surface mean, and is the most commonly
used parameter for describing surface roughness. The average maximum roughness, R
z
, is the
average height difference between the ten greatest peak-to-valley separations in the evaluation area,
and thus gives a measure of the extremes of the surface roughness and the surface texture. The
average surface roughness, R
a
, was used in the present work to characterize coatings, mainly in
order to evaluate surface roughness induced by the metallographic sample preparation.
5.4.2 Scanning Electron Microscopy
Microstructural studies were performed on fracture surfaces and polished cross sections using a
JEOL 5800LV scanning electron microscope. The acquisition of the images was done using Noran
Voyager Analysis Software. To prevent charging, coatings were sputtered with gold. For higher
magnification images carbon-sputtered samples were studied using a JEOL JSM-6340F Field
Emission Scanning Electron Microscope.
5.4.3 Dilatometry
Dilatometry experiments were performed on as-sprayed free-standing top coats in a NETSCH DIL
402C dilatometer. A schematic diagram of such a dilatometer can be seen in Figure 5.9. The
specimen under investigation is placed in the head assembly and an alumina push road applies a
constant force of 30 cN. The sample is heated in a controlled manner and changes in its dimensions
are transmitted via the alumina thrust rod and detected by an electrical transducer. The resolution is
approximately 0.1 m. The push rod and the sample holder will also undergo dimensional changes
during measurement so a calibration run is performed using a block of alumina similar in size to the
Chapter 5. Experimental Procedures

85
specimen, in order to quantify this and factor it out of the result. All tests were carried out in air.
Specimens were placed in the head assembly and heated at a heating rate of 30-50
o
C min
-1
. The
dwell times at these temperatures varied from 24 to 150 h. Cooling rates were about 30
o
C min
-1
.

Figure 5.9 Schematic diagram of the push-rod dilatometer.

Dilatometry tests were performed on free-standing top coats. As shown schematically in Figure
5.10, specimens were tested in both in-plane and through-thickness directions. Specimen
dimensions differed according to the direction being examined. Typically, to study the in-plane
characteristics, specimens approximately 10 mm in length were used. In order to examine the
sintering behaviour through the thickness of the coating, the specimen length, (coating thickness)
was about 2 mm.

Figure 5.10 Specimen configuration for dilatometry experiments.
5.4.4 X-Ray Diffraction
Xray diffraction (XRD) measurements were carried out using a computer-controlled Phillips
PW1710 diffractometer, with Cu
Ko
radiation (=0.154 nm), 40 kV accelerating current and a
40 mA filament current. The incident beam optical conditions were set up with an anti-scatter slit
in-plane
through-thickness
~10 mm
~2 mm
Chapter 5. Experimental Procedures

86
of 1
o
and a 15 mm horizontal mask. For the diffracted beam, a 0.2 mm receiving slit and a 1
o

divergence slit were used. The apparatus was regularly calibrated using a silicon control sample.
Scans over the range 2 = 25-80
o
were carried out with a step size of 0.02
o
and 4 s counting time at
each step. Two regions were of particular interest for the identification of the phases present. The
first of these was the region 2 = 27-32
o
, where ( 1 11)
m
and (111)
m
peaks, and the (111)
cub/tet
peak
are found. Slow scans were performed over this range, with a step size of 0.02
o
and a dwell time of
10 s. The second region examined by a slow scan was 2 = 72-75
o
, in order to examine the (400)
tet
,
(004)
tet
and the (400)
cub
peaks. Slow scans were performed over this range with a step size of 0.01
o

and a dwell time of 35 s.
The mole fractions of tetragonal, monoclinic and cubic phases were calculated from the peak
intensities of the slow scans, using the following equation, from Miller [232] and Toraya [310, 311]:
( )
/
111 (111)
(111)
0.82
m m
c t
m
c t
I I
M
M M I
+ | |
= |
|
+
\ .
(5.1)
( )
( ) ( )
400
004 400
0.88
c
t t
c
t
I
M
M I I
| |
| =
| +
\ .
(5.2)
1
m c t
M M M + + = (5.3)
where M
m
, M
c
and M
t
are the mole fractions of the monoclinc, cubic and tetragonal phases
respectively and I is the integrated intensity corresponding to the peak concerned. The integrated
intensity was calculated after peak deconvolution and profile fitting, performed using Phillips
PROFIT software. The software can fit Lorenzian, Gaussian, pseudo-Voigt and Pearson-Seven
profiles, all of which can be symmetrical or asymmetric, and it will assign the Ko
2
s. This
procedure allows accurate determination of peak positions, peak areas and widths.
Alternative, quantitative phase analysis was in some cases performed with the full x-ray diffraction
pattern, using the Rietveld method based on the Phillips Xpert Plus software. The Rietveld code in
the software is based on [312]. In order to perform the analysis, all phases must be identified and the
quality data for the structure of the phases present must be available. The Reitveld method permits
simultaneous refinement in each of the present phases of both structural (lattice parameters) and
microstructural (phase percentage) parameters.
The (400)
tet
, (004)
tet
and (400)
cub
peak positions were used for calculation of lattice parameters, a
for the cubic phase and a and c for the tetragonal phase. The Y
2
O
3
content of the tetragonal and
Chapter 5. Experimental Procedures

87
cubic phases were calculated from these measured lattice parameters, respectively using the
following equations [232, 306], where a and c are expressed in nanometres
1.5
1.0223
mole % YO in tetragonal phase
0.001309
c
a

= (5.4)
1.5
- 0.5104
mole % YO in cubic phase =
0.0204
a
(5.5)
5.4.5 Porosity Measurements
Porosity measurements were performed on as-sprayed and heat-treated free standing top coats,
using a liquid immersion technique. Mass measurements were performed using a Sartorius
RC120P microbalance, with a sensitivity of 10 g. Firstly, the mass of the sample was measured
and recorded (m
air
). Then, the specimen was coated with a viscous lacquer of known density
laq

and its new mass measured (m
c
). The lacquer was used in order to prevent the filling of any surface-
connected cracks or pores by capillary action of the liquid. Finally, the coated specimen was
immersed in liquid and its mass was recorded (m
liq
). For the purpose of recording the specimen
mass while immersed in liquid, an appropriate set up was utilised, which is shown schematically in
Figure 5.11. The liquid used was Flutec PP9 (per fluoro-1-methyl decalin), a dense, low surface
tension liquid with known density
liq
. The following equation was used to determine the density of
the sample:
|
|
.
|

\
|

|
|
.
|

\
|

=
laq
air c
liq
lid c
air
d
m m m m
m

(5.6)
The porosity fraction of the specimen was calculated by comparing the measured density with the
theoretical density for the bulk material
th
, following the equation below:
1
d
f
th
p

| |
=
|
\ .
(5.7)
The theoretical density of each of the coating materials was calculated in the following way. From
x-ray data, the lattice parameters of the crystal structure were deduced and the volume of the unit
cell was calculated. From the weight percentage of the dopant, and the atomic weight of the
elements present, the atomic mass of the unit cell was calculated. It was assumed that dopant
cations substitute for the zirconia ions in the structure. The presence of impurities was not taken
Chapter 5. Experimental Procedures

88
into consideration. The theoretical density was calculated from the atomic mass and the unit cell
volume. The estimated error for this technique is approximately 3-4 %. The main sources of
experimental error arise from the weight measurement when the sample is immersed in the liquid
and from uncertainties in the density of the liquid and the lacquer.

Figure 5.11 Schematic diagram of set-up for density and porosity measurements (after Murphy [313]).

5.4.6 Stiffness Measurement
5.4.6.1 Cantilever Bending
Youngs moduli of the detached top coats was determined using the Cantilever Bend Test. This is a
modified version of the cantilever beam test used by Tsui [76] and has been found suitable for
determining the Youngs modulus of materials with relatively low stiffness, such as plasma sprayed
top coats [309]. A schematic diagram of the loading configuration used is shown in Figure 5.12. A
fixed incremental load P was applied at a distance L from the clamped end. The displacement in the
y direction, at distance x from the fixed end, was measured using a non-contact laser scanning
extensometer (Lasermike). The stiffness was determined using the equation below:
(

=
6 2
3 2
x Lx
yI
P
E (5.8)
where I is the second moment of area of the sample and the other variables are as described in
Figure 5.12.
Chapter 5. Experimental Procedures

89

Figure 5.12 Loading configuration for the cantilever bend test.
5.4.6.2 Nanoindentation
Nanoindentation was carried out using a MicroMaterials NanoTest 600 indenter. Nanoindentation
testing consists of loading and unloading a diamond indenter into the sample under investigation
and monitoring how the indentation depth changes. All indentations were done with a Berkovich
diamond indenter, which is a three-sided pyramid. The indenter was pressed into the surface with a
specified loading rate. A dwell time at maximum load was included, to prevent gross errors due to
creep [314]. In this case, the maximum load was 100 mN, the loading rate was 5.1 mN s
-1
and the
dwell time at maximum load 10 s. Displacement was measured via a parallel plate capacitor. The
location of the indents and the indentation stage is accurately controlled by an actuator.
The deformation on loading/unloading of the indenter is a combination of both elastic and plastic
strain [315]. Both the hardness and modulus of the material can be determined by this method.
During, loading there is a combination of both and elastic and plastic deformation, whereas, during
unloading the material recovery is principally elastic. Youngs modulus was determined from the
unloading curve, using the Oliver and Pharr [315] technique. The reduced modulus of the material,
E
r
, is determined from the initial gradient of the unloading curve, S, using the following equation:
4
r
dP A
S E
dh t
= = (5.9)
where A is the indent contact area, which is determined from the displacement, h, according to a
pre-determined diamond area function. The Youngs modulus of the sample, E
s
, can then be
estimated using the Poissons ratio of the indenter,
i
, and the sample material,
s
, and the Youngs
modulus of diamond, E
i
, following the equation:
L
x
P
y
Chapter 5. Experimental Procedures

90
( ) ( )
2 2
1 1
1 s i
r s i
E E E
v v
= + (5.10)
The diamond indenter is commonly taken to have a Youngs modulus of 1147 GPa and Poisson
ratio of 0.07. Nanoindentation tests were performed on polished cross-sections of free-standing top
coats. The data were analysed using a program supplied by NanoTest, which uses calibration
indentations performed regularly on fused silica.

5.5 Heat Treatment of Thermal Barrier Coatings
5.5.1 Isothermal Heat Treatment
Isothermal heat treatments were performed on free-standing APS top coats with typical thickness of
0.5-2 mm, at temperatures in the range 1200-1400C. The heat treatments were performed in air in
a Lenton furnace, with a heating rate of about 30
o
C/min. Samples were removed from the furnace
once the heat treatment was completed and were air cooled at a rate of about 30C
o
/min.
5.5.2 Heat Treatment with a High Thermal Gradient
Heat treatments of TBCs under high thermal gradient were performed in a specially designed rig, in
which a high thermal gradient through the top coat could be established. The heat treatments were
carried out on Nimonic80A superalloy substrate block, about 50 mm thick, with a CoNiCrAlY bond
coat about 40 m thick and top coats 1-2.5 mm thick. A schematic of the rig is shown Figure 5.13.
The rig was constructed within a controlled atmosphere chamber. The heating source was provided
by a graphite susceptor heated by an induction coil. Typical power input into the induction coil was
3-5 kW. The base of the substrate was brazed to a water-cooled copper block, with a silver-based
braze Easyflow No.2. The flow rate of the cooling water was 4 l/min. Insulation was placed
around the specimen to minimise lateral heat losses. The graphite susceptor could be lifted away
from the induction coil by a hydraulic ram, allowing thermal cycling of the specimen. The thermal
gradient was monitored and controlled by four thermocouples embedded in the substrate. The
temperature in the top coat was monitored using a thermocouple, which was inserted into a laser
drilled hole, parallel to the bond coat/top coat. The laser drilled hole was 2 mm deep and was
drilled using a pulsed Nd-YAG laser with wavelength =1.06 m, 0.5 ms pulse duration and energy
2.5 J per pulse.
Chapter 5. Experimental Procedures

91
The holding time at temperature, for each thermal cycle, was 1 hour and samples were allowed to
cool to room temperature before being re-heated. Heating rates were ~20 K min
-1
and cooling rates
~30 K min
-1
. Typical top coat temperatures during the gradient heat treatments applied in the
present work ranged from about 1500C at the free surface to 900C at the interface. A vacuum of
2 10
-4
mbar was maintained throughout. The chamber incorporated a viewing window, which
allowed automated periodic image capture with the aid of a digital video camera.

(a)
(b)
Figure 5.13. The high thermal gradient rig: (a) schematic diagram and (b) photo taken during operation.


5.6 Measurement of Thermal conductivity
5.6.1 Steady State Rig
5.6.1.1 Experimental Set Up
Figure 5.14 shows a schematic of the setup. A sample, with lateral cross section of 35 30 mm,
was sandwiched between two Nimonic 80A alloy substrates (Special Metals Limited, Hereford,
UK). These substrates act as flux meters, each being instrumented with four K-type thermocouples
(T
1
to T
8
), inserted into holes at known distances, drilled to the centerline of each block. In
principal, only two temperature readings are required from each substrate to determine the heat flux,
but by taking additional readings, an improved average value for the flux can be obtained. Since
the (temperature-dependent) thermal conductivity of Nimonic 80A is well known over a wide
Chapter 5. Experimental Procedures

92
temperature range [316], the heat flux across each substrate can be accurately determined. Of
course, the substrate can also be made of other materials and with different cross-sections.
Temperature gradients are generated by heating the lower substrate with an electric resistance
heater, while heat is continuously removed from the upper substrate via a water-cooled copper heat
sink. To maintain consistent heat supply and removal, the heater power and cooling water flow rate
are fixed throughout the experiment. The setup is insulated with a low conductivity glass wool, to
minimise lateral heat losses and promote one-dimensional heat flow. A check can be made about
this, since significant losses would result in the heat fluxes through the two substrates being
different.
To eliminate air gaps and raise the interfacial thermal conductance (h), two types of substrate-to-
sample interfacial materials were tested:- (i) a silicone-based thermal compound (HTSP
Electrolube), with conductivity of 3.0 W m
-1
K
-1
and (ii) Sil-Pad

2000 (Bergquist), a high


performance conductive pad, with thickness of 0.38 mm and conductivity of 3.5 W m
-1
K
-1
. The
suitability of these two interfacial materials is discussed in more detail in section 8.1.1. To ensure
that the interface can be reproduced for different runs, a fixed torque of 3 N m was applied onto the
test column via the bolt. This torque generated an axial force of 750 N, corresponding to a nominal
pressure of about 0.72 MPa for the current sample cross-section. During the experiment, the
temperature readings (T
1
to T
8
) were continuously logged. A steady state was considered to have
been established when the temperature fluctuations were within 0.5C, for more than 30 minutes.
These readings were then employed in calculating the thermal conductivity of the sample using
equation (4.4) for steady state heat flux.
Chapter 5. Experimental Procedures

93

Figure 5.14. Schematic of the steady-state bi-substrate thermal conductivity setup. T
1
to T
8
designate the
eight thermocouples used for monitoring the temperature within the two substrates (flux meters).

Figure 5.15. Photo of the bi-substrate steady state thermal conductivity set up and data logging
equipment.


Chapter 5. Experimental Procedures

94
5.6.1.2 Data Analysis
The data analysis performed in order to calculate the thermal conductivity from the measured
temperature profile and sample thickness is presented here. By assuming one-dimensional heat
flow across the setup, and by taking into account changes in thermal conductivity of the blocks with
temperature, the mean heat flux, Q, can be found using eqns.(5.11) to (5.13):

Q
upper
=
1
6
T
i
T
j
x
i
x
j
k
sub
T
ave
ij
( )
j =1
3

i =2
i > j
4

,

T
ave
ij
=
T
i
+ T
j
2
(5.11)

Q
lower
=
1
6
T
i
T
j
x
i
x
j
k
sub
T
ave
ij
( )
j =5
7

i =6
i > j
8

,

T
ave
ij
=
T
i
+ T
j
2
(5.12)
where i and j designate locations of thermocouples (see Figure 5.14), while

k
sub
T
ave
ij
( )
is the
temperature-dependent thermal conductivity of the Nimonic substrate at the average temperature.
The mean heat flux can be written as:

Q =
1
2
Q
upper
+Q
lower
( )
(5.13)
The measurement is presumed to satisfy one-dimensional heat flow when:

Q
upper
Q
lower
Q
s10% (5.14)
For a typical experiment, the difference was found to be significantly less than 10%.
Figure 5.16 shows a schematic of the temperature profile established across the substrates and
sample (coating) under steady state conditions. Assuming no lateral heat losses, the same flux
flows through coating and interfaces, so that:

Q = k
eff
AT
Ax
(5.15)

Q = hAT
i
(5.16)

Q = k
true
AT
c
Ax
(5.17)
where k
eff
is the effective thermal conductivity of the coating, AT is the total temperature drop, Ax is
the coating thickness, h is the interfacial thermal conductance (assumed the same for both interfaces)
and k
true
is the actual coating thermal conductivity, while the subscripts i and c designate interface
and coating, respectively.
Chapter 5. Experimental Procedures

95
The total temperature drop, AT, consists of temperature drops across the coating and interface, that
is

AT = AT
c
+ 2AT
i
(5.18)
By substituting eqns (5.16). and (5.17) into eqn.(5.18), and rearranging, we obtain the following
linear equation:

AT
Q
=
Ax
k
true
+
2
h
(5.19)
Now, by plotting AT/Q versus Ax, the slope and intercept are given by 1/k
true
and 2/h, respectively.
Both k
true
and h can thus be determined.

Figure 5.16. Schematic depicting the different
temperature drops across the substrates and
sample (coating).
5.6.1.3 Validation of Technique
Before the technique can be used with confidence, it must be verified using a standard material of
known thermal conductivity. Fused silica (quartz glass) samples supplied by Heraeus Quarzglas
(Germany), were chosen for this purpose. In the manufacturers material datasheet, its thermal
conductivity is reported as 1.46 W m
-1
K
-1
at 100C [317]. Samples with three different thicknesses,
i.e. 0.567, 0.989 and 2.897 mm, were used. Figure 5.17 shows the temperature versus distance plot
for a silica sample with a thickness of 2.897 mm. The mean temperature of the sample was about
100C and the total temperature drop (AT) across the sample was about 48C. Similar profiles were
observed for the thinner samples, but with smaller drops. Using eqns.(5.11) to (5.13), the mean heat
flux was estimated as 26.36 kW m
-2
. Subsequently, from eqn.(5.15), the effective thermal
conductivity (k
eff
) was found to be 1.25

W m
-1
K
-1
.

Chapter 5. Experimental Procedures

96

Figure 5.17. Temperature versus distance plot showing the temperature drop across a fused silica sample.
The points (T
1
to T
8
) are thermocouple measurements corresponding to locations in the Nimocic substrate.
The line is calculated using the average heat flux, Q, and temperature-dependent thermal conductivities of
Nimonic substrates.

Figure 5.18 shows experimental data for all the fused silica samples, along with the best linear fit,
as given by eqn.(5.19). It can be seen that the data show relatively little scatter. The values of
actual thermal conductivity (k
true
) and interfacial thermal conductance (h) were found to be
1.54 W m
-1
K
-1
and 12.88 kW m
-2
K
-1
, respectively. Of course, k
true
is higher than k
eff
, since
temperature drops across the two interfaces (2AT
i
) have now been taken into account. The actual
conductivity agrees quite well with the value (1.46 W m
-1
K
-1
) quoted by the manufacturer [317]. It
may also be noted that the interfacial contact conductance measured here is of a similar order of
magnitude to values reported elsewhere (for a similar surface finish and contact pressure) [318].
Samples with lateral cross section of 35 30 mm of as-sprayed YSZ (204NS) were prepared by
debonding the top coats from the substrates, and subsequently polishing the surface to ~1 m
surface finish. The debonded coatings had thicknesses ranging from 0.5 to 2.0 mm. Care was taken
to ensure a homogenous thickness.
Chapter 5. Experimental Procedures

97

Figure 5.18. A plot of AT/Q versus Ax for fused silica samples of three different thicknesses. The points are
experimental data while the line is their best linear fit using eqn.(5.19). The values of k
true
and h can be
obtained from the slope and intercept, respectively. Conductive compound was applied at the interface, and
the mean temperature of the samples was about 100C.
5.6.2 Hot Disk
A Hot Disk Thermal Constants Analayser Version 5.0 [249] was used for thermal conductivity
measurements of free-standing top coats. In this technique a disk-shaped sensor, which acts both
the heat source and the temperature sensor, is sandwiched between two identical samples. A
current pulse is passed through the sensor, sufficient to cause a slight temperature increase in the
sensor and in the sample surrounding it. Assuming that the resistivity of the metal element in the
sensor is temperature dependant, the temperature increase in the sensor and the sample surrounding
will cause a change in its resistance and a corresponding voltage variation over the sensor. From
the change in resistivity with time, and assuming constant current and no influence to the
temperature increase from the outer boundaries of the sample, the temperature distribution with
time in and around the metal strip can be expressed in terms of the thermal diffusivity and the
thermal conductivity of the sample. Details on the underlying theory are given in section 4.2.2.2.
The Hot Disk sensor consisted of an electrically conducting double-spiral pattern of Nickel
supported in both sides by an insulating layer of Kapton. The radius of the sensor was 2.001 mm.
The output powder used for the experiments was 0.2 W and the measurement time 2.5 sec. The
Chapter 5. Experimental Procedures

98
samples used were approximately 35 30 mm and 2 mm thick. Analysis of the data was performed
using the Hot Disk Thermal Constants Analyser software.
The temperature increase in the sensor for YSZ as-sprayed top coat is presented in Figure 5.19. The
time window for analysis was chosen taking into account the sample dimensions and an initial
stabilization time. For the first few data points the temperature difference between the sample and
the sensor due to the interface resistance is not constant, and towards the end the temperature might
be influenced by the size of the sample- see Figure 5.19. The probing depth corresponding to the
time window chosen is 1.98 mm, which is just under the thickness of the samples used.
0
0.5
1
1.5
2
2.5
3
3.5
0 0.5 1 1.5 2 2.5 3
T
e
m
p
e
r
a
t
u
r
e

I
n
c
r
e
a
s
e

A
T

(
o
C
)
Time (sec)
Analysis

Figure 5.19. Temperature increase of the Hot Disk sensor as a function of time showing time-window used
for data analysis.

For the correct value of the characteristic time of the measurement

2
= , where r is the radius
of the sensor and is the thermal diffusivity of the sample, the temperature increase recorded by the
sensor can be expressed as a linear function of the dimensionless function D() (see eqns 4.13 and
4.14). Figure 5.20a shows a plot of the temperature increase as a function of D(), which is a
straight line. The slope of the line in Figure 5.20a is inversely proportional to the thermal
conductivity of the tested sample (see equations 4.16 and 4.17). The selected data points for
analysis where adjusted to a straight line by varying the characteristic time . The random scatter of
the data points around the fitted straight line is displayed in Figure 5.20b, which is a plot of the
Chapter 5. Experimental Procedures

99
difference between the measured and fitted temperature versus Sqrt(time). Since the scatter is
random, the correct value for the characteristic time has been used.
(a)
2.6
2.8
3
3.2
3.4
0.2 0.22 0.24 0.26 0.28 0.3 0.32
T
e
m
p
e
r
a
t
u
r
e

i
n
c
r
e
a
s
e

(
o
C
)
D(t)

(b)
-0.003
-0.002
-0.001
0
0.001
0.002
0.003
0.8 0.9 1 1.1 1.2 1.3 1.4 1.5
T
e
m
p
e
r
a
t
u
r
e

D
i
f
f
e
r
e
n
c
e

A
T
f
i
t
t
i
n
g

-

A
T
e
x
p


(
o
C
)
Sqrt(time)

Figure 5.20. Hot Disk measurement: (a) temperature increase as a function of the dimensionless time function D() and (b)
random scattering of experimental data around theoretical fitting as a function Sqrt(time).

5.6.3 Laser Flash
The laser flash method was used in order to calculate the thermal conductivity of free-standing top
coats. Measurements were carried out in the Thermophysical Properties Section of the NETZSCH
Application Laboratory. The thermal conductivity, k, was calculated using the following equation
(see also equation 4.10):
( ) ( ) ( )
p
k T T C T o = (5.20)
where is the thermal diffusivity, C
p
is the specific heat, is the density and T is the temperature.
The thermal diffusivity was measured using a Netzsch model 427 laser flash diffusivity apparatus.
The unit used in this work was equipped with a high-temperature, water-cooled furnace capable of
operation from 25 to 2000C. The sample chamber is isolated from the graphite heating element by
a protective tube allowing samples to be tested under vacuum or in an oxidizing, reducing or inert
atmosphere. The temperature rise on the back face of the sample is measured using an In-Sb
detector. Data acquisition and evaluation are accomplished using a comprehensive PC software
package. The thermal diffusivity measurements were conducted in a dynamic argon atmosphere at
a flow rate of ~100 ml/min between room temperature and 1400C. The presented thermal
Chapter 5. Experimental Procedures

100
diffusivity results are the average values of five individual tests. The samples were free-standing
square disk top coats with side lengths of approx. 10 mm and thicknesses of approx. 2 mm. In order
to increase both the absorption of laser energy and emission of infrared light, the samples were
coated with graphite.
The specific heat measurements were conducted using a Netzsch model DSC 404 C Pegasus


differential scanning calorimeter capable of operation from 25 to 1500C. The system is vacuum-
tight, and therefore samples can be tested under pure inert, reducing or oxidizing atmospheres, as
well as under vacuum. The specific heat is determined by running a baseline and standard over the
temperature range of interest. The sample is then run and the specific heat is calculated by the
standard ratio method. Instrument control and data acquisition are accomplished via a new 32-Bit
MS

-Windows
TM
Thermal Analysis software package. Data evaluation is carried out by a
comprehensive PC software package. The samples tested with the DSC had masses of approx. 200
mg. The measurements were carried out in a dynamic argon atmosphere (gas flow rate: 50 ml/min).
The system was equipped with a temperature-calibrated DSC-cp type S sensor. Platinum crucibles
with lids were employed for the test. The samples were heated between room temperature and
1400C at a heating rate of 20 K/min.

Chapter 6. Thermal Stability of PS top coats

101
Equation Section 6
6 Thermal Stability of Plasma-Sprayed Top Coats
6.1 Changes in Phase Constitution of Plasma-Sprayed Top Coats due
to Heat Treatment
Phase characterisation of TBCs is of great importance for both studying and designing materials for
specific applications. TBCs are often made of polymorphs whose presence is controlled by many
factors such as stabilizer content, processing parameters and heat treatment. The polymorphism of
zirconia, particularly the volume expansion associated with the tetragonal-to-monoclinic
transformation, has led to the investigation of various stabilized zirconia alloys. The phase stability
of the top coat constitutes an important factor for determining the thermomechanical properties of
TBCs.
6.1.1 As-sprayed
6.1.1.1 Yttria stabilized zirconia top coats
The x-ray diffraction patterns of the as-sprayed top coats are presented in Figure 6.1 to Figure 6.4.
The YSZ as-sprayed top coats (204NS & 204NS-1) are composed of non-transformable T
tetragonal phase. The K
o1
and K
o2
components of the T' peaks can be distinguished. This phase is
formed by a diffusionless transformation of the high-temperature cubic F phase, due to rapid
solidification and cooling during spraying. The T non-transformable tetragonal is not an
equilibrium phase and it differs from the regular tetragonal phase, notably for its smaller
tetragonality, higher yttria content and stability against further transformation into the tetragonal
phase [306, 319-322]. However, for crystallographic purposes, the T non-transformable tetragonal
phase and T tetragonal phase are fundamentally the same tetragonal polymorph in the zirconia solid
solution [207]. Stabilization of the tetragonal or cubic zirconia phase by the addition of trivalent
dopants is believed to be associated with the generation of oxygen vacancies [202]. The trivalent
dopant cations substitute the Zr ion in the cation network and oxygen vacancies are created for
charge compensation. The tetragonal and cubic polymorphs are stabilized when the doping level is
sufficient to create enough oxygen vacancies to reduce the co-ordination number of the Zr ion from
8 to around 7.5 [203, 204]. The Zr ion in fully stabilized and high temperature cubic phase has a
coordination number of 7, due to the strong covalent nature of the ZrO bond and the small size of
the Zr ion.
Chapter 6. Thermal Stability of PS top coats

102
The position of the tetragonal peaks is known to move as a function of the amount of yttria that the
tetragonal phase contains, due to the changing lattice parameter. Due to the slightly different yttia
content, the tetragonal peaks are not in exactly the same position in the two coatings. Trace
amounts of the monoclinic phase exists in the 204NS-1 YSZ powder (1.5 %), which may be due to
the presence of unmelted particles thatcould contain some monoclinic phase. This is in agreement
with previous work in this area [76, 109]. The different amount of impurities in the starting
powders 204NS and 204NS-1 of these two coatings does not have a noticeable effect of their phase
constitution.
The two YSZ top coats doped with a small amount of Al
2
O
3
(AE8170 & AE9018) also contain the
non-transformable tetragonal phase. The shoulder observed on both the tetragonal peaks are due to
the K
o1
and K
o2
components of the T' peaks. The presence of a small amount of Al
2
O
3
does not
appear to alter the phase constitution of the top coat. The solubility of Al
2
O
3
in Zr
2
O is very low
[323, 324] (<0.5 mol%), so it is possible that the Al
2
O
3
is not in solid solution. However the
amount present is too small to form a distinct peak. It can be concluded that impurities such as SiO
2

and Al
2
O
3
do not have an effect on the phase constitution of as-sprayed YSZ coatings.

27 28 29 30 31 32
I
n
t
e
n
s
i
t
y
2u ()
M (111)
T'(111) & F(111)
M (111)
(a)
M (111) M (111) M (111)

72 73 74 75
I
n
t
e
n
s
i
t
y
2u (
o
)
T'(004)
(b)
T'(400)
F(400)

Figure 6.1 XRD spectra in 2u ranges of (a) 27-32, and (b) 72-75 for as-sprayed YSZ (204NS-1) top coats.
Chapter 6. Thermal Stability of PS top coats

103
27 28 29 30 31 32
I
n
t
e
n
s
i
t
y
2u ()
M (111)
T'(111) & F(111)
M (111)
(a)

72 73 74 75
I
n
t
e
n
s
i
t
y
2u (
o
)
T'(004)
(b)
T'(400)
F(400)

Figure 6.2. XRD spectra in 2u ranges of (a) 27-32, and (b) 72-75for as-sprayed YSZ (204NS) top coats.
27 28 29 30 31 32
I
n
t
e
n
s
i
t
y
2u ()
M (111)
T'(111) & F(111)
M (111)
(a)

72 73 74 75
I
n
t
e
n
s
i
t
y
2u (
o
)
T'(004)
(b)
T'(400)
F(400)

Figure 6.3. XRD spectra in 2u ranges of (a) 27-32, and (b) 72-75for as-sprayed Al
2
O
3
-doped YSZ (AE8170)
top coats.
27 28 29 30 31 32
I
n
t
e
n
s
i
t
y
2u ()
M (111)
T'(111) & F(111)
M (111)
(a)

72 73 74 75
I
n
t
e
n
s
i
t
y
2u (
o
)
T'(004)
(b) T'(400)
F(400)

Figure 6.4. XRD spectra in 2u ranges of (a) 27-32, and (b) 72-75for as-sprayed Al
2
O
3
-doped YSZ (AE9018)
top coats.

Chapter 6. Thermal Stability of PS top coats

104
6.1.1.2 Dysprosia stabilized zirconia top coats
Dysprosia is also believed to be an effective stabilizer in zirconia, up to concentrations of about
15 mol % (~ 35 wt%) [308]. When the concentration of Dy
2
O
3
is above 15mol%, the pyrochlore-
structure Dy
2
Zr
2
O
7
phase and zirconia solid solution are supposed to co-exist. The DSZ top coat
(SPM6-2444) in the present study, which has about 3.5 mol% Dy
2
O
3
, comprises the non-
transformable tetragonal T with a trace of monoclinic phase (1.5 %). No pyrochlore phase is
observed. The K
o1
and K
o2
components of the T' peaks can be distinguished (Figure 6.5).
The tetragonality c/a is an important factor in the stability of the tetragonal phase and it has been
shown to be independent of the species of the dopant, but dependant on the content of the dopant
for oversized trivalent rare earths[205, 206]. Tetragonality decreases with increasing dopant
content for oversized trivalent cations and vanishes at 11 mol% M
2
O
3
regardless of the ionic sizes,
since tetragonality is aided by the creation of oxygen vacancies only [207]. The mol% dysprosia in
the DSZ (3.58 mol%) in the present study is slightly lower than that of yttria in the YSZ top coats
(about 4.3 mol%). Therefore, we expect the DSZ top coat will have a higher tetragonality than that
of YSZ coatings (c/a
DSZ
= 1.012, c/a
YSZ
= 1.010). The decrease in tetragonality is associated with
increasing stability of the tetragonal phase and hence the resistance of the non-transformable
tetragonal to transformation into the monoclinic phase is ascribed to its smaller tetragonality.

27 28 29 30 31 32
I
n
t
e
n
s
i
t
y
2u ()
M (111)
T'(111) & F(111)
M (111)
(a)

72 73 74 75
I
n
t
e
n
s
i
t
y
2u (
o
)
T'(004)
(b)
T'(400)
F(400)

Figure 6.5. XRD spectra in 2u ranges of (a) 27-32, and (b) 72-75for as-sprayed DSZ (SPM6-2444) top coat.



Chapter 6. Thermal Stability of PS top coats

105
6.1.1.3 Yttria-lanthana stabilized zirconia top coats
For the YLaSZ, the deconvoluted spectrum is presented in Figure 6.6b, which shows that the as-
sprayed coating probably consists of two tetragonal phases with slightly different lattice parameters.
Possibly the presence of the two tetragonal phases is a result of two different dopants with different
ionic radii. Whereas tetragonality is independent of ionic radius of the dopant and depends only on
dopant content, on the other hand lattice parameters of the tetragonal phase increase systematically
with the ionic radii of the dopant atoms [205, 206]. If the two tetragonal phase that appear to be
present are a result of the simultaneous presence of yttria and lanthana, then the tetragonal phase
with the greater lattice parameters is a result of the lanthana dopant, whereas the tetragonal phase
with slightly smaller lattice parameters is a result of the yttria dopant, since the lanthanum cation
has a bigger ionic radius (see Table 6.1).
The stabilization effect of lanthana in zirconia is disputed in the literature. As the ionic size
mismatch between the dopant and the zirconium atom becomes greater, the solubility of the
tetragonal and cubic phase decreases and the pyrochlore phase may be formed instead of the
tetragonal or cubic phase. According to the binary zirconia lanthana phase diagram [231] at above
about 2 mol% La the equilibrium phases are the monoclicnic and the pyrochlore phase. No
pyrochlore phase was detected in the present study in the as-sprayed coatings. Din et al [308]
reported that the addition of lanthana to zirconia did not stabilize the T phase. However, a
pyrochlore-type cubic phase La
2
Zr
2
O
7
is formed. Bastide et. al [227] also found that the solid
solubility of La in monoclinic Zr is 1.5 mol % and above that concentration pyrochlore-structured
La
2
Zr
2
O
7
phase is formed. However, Li et al [325] reported that 8mol% La
2
O
3
stabilized the
tetragonal phase of La
2
O
3
ZrO
2
powder compacts. In the present study the simultaneous presence
of yttria and lanthana stabilizers in zirconia appears to result in two metastable tetragonal phases in
the as-sprayed coatings.

Ion Zr
4+
Y
3+
Dy
3+
La
3+
Ce
4+
Ce
3+
Ionic radius (nm) 0.084 0.102 0.103 0.116 0.097 0.114
Table 6.1. Ionic radii for cations in PSZ top coats [326].
Chapter 6. Thermal Stability of PS top coats

106
27 28 29 30 31 32
I
n
t
e
n
s
i
t
y
2u ()
M (111)
T'(111) & F(111)
M (111)
(a)


Figure 6.6. XRD spectra in 2u ranges of (a) 27-32, and (b) 72-75 for as-sprayed YLSZ (AE8321) top coat.

6.1.1.4 Ceria stabilized zirconia top coats
The CSZ powder also contains the non-transformable tetragonal phase (see Figure 6.7). Tetravalent
dopants such as ceria stabilize the tetragonal and cubic phases of zirconia [38, 190]. However, the
stabilization is not the result of the generation of oxygen vacancies, as is the case with trivalent
dopants. The Ce
4+
ion substitutes the Zr
4+
ion in the cation network, without creating any oxygen
vacancies, since both ions are tetravalent [208, 209]. The stabilization effect is thought to be caused
by the slight dilation of the cation network by the oversized Ce
4+
ion which decreases the
tetragonality and stabilizes the tetragonal phase to room temperature [208]. In Figure 6.7 broad
XRD peaks are observed for the CSZ top coat. This has commonly been observed previously with
CSZ top coats. It is thought to relate to compositional variations, owing to the larger size of the Ce
ion within the zirconia lattice [37]. Some research [219, 327] suggests that the tetravalent Ce ion is
reduced to is trivalent state because of the reducing atmosphere in the plasma plume and the rapid
quenching thereafter of the molten droplets. In such situation, oxygen vacancies would be
introduced to balance the lower valance state of the cerium ion. This would result in stabilization of
the cubic phase and possibly the presence of both non-tranformable tetragonal T and cubic F
phases. The presence of some trivalent cerium could account for the large variation in the unit cell
responsible and thus the broad XRD peaks observed. The presence of some cubic phase is also a
possibility. However, the presence of a cubic peak cannot be confirmed in the data presented here
(see Figure 6.7), since the height of any cubic peak is expected to be of the same order as the noise,
even though the signal to noise ratio is satisfactory. Annealing at relatively moderate temperature
(700
o
C) should result in re-oxidation of the cerium into its tetravalent state and stabilization of the
Chapter 6. Thermal Stability of PS top coats

107
tetragonal phase, with no cubic phase present. The colour of CSZ plasma sprayed coatings in which
the cerium ion has been reduced to its trivalent state is brown-green [38, 218], whereas for CSZ
coatings with the majority of the substitutional Ce ions in their tetravalent state the color is pale
yellow [38]. The samples used in the present work were pale yellow in the as-sprayed state, further
suggesting that they consist mainly of tetravalent cerium and non-transformable tetragonal phase
T(see Figure 6.9). It is possible that trivalent cerium formed during spraying was re-oxidized
during cool-down, since the plasma sprayed coatings were air cooled to below 120
o
C inside the
spraying chamber. Previous work [209] suggests that the ceria ions in CSZ plasma sprayed
coatings are preferentially in their tetravalent state.
27 28 29 30 31 32
I
n
t
e
n
s
i
t
y
2u ()
M (111)
T'(111) & F(111)
M (111)
(a)

72 73 74 75
I
n
t
e
n
s
i
t
y
2u (
o
)
T'(004)
(b)
T'(400)
F(400)

Figure 6.7. XRD spectra in 2u ranges of (a) 27-32, and (b) 72-75 for as-sprayed CSZ (205NS) top coat.
27 28 29 30 31 32
I
n
t
e
n
s
i
t
y
2u ()
M (111)
T'(111) & F(111)
M (111)
(a)

72 73 74 75
I
n
t
e
n
s
i
t
y
2u (
o
)
(b)
F(400)

Figure 6.8. XRD spectra in 2u ranges of (a) 27-32, and (b) 72-75 for CSZ (205NS) top coat that has been
annealed in vacuum at 500
o
C.

The reduction of cerium atoms to its trivalent state is accompanied by a volume change. The Ce
3+

ion is larger than the Ce
4+
ion, resulting in a decrease in the solubility of CeO
2
in the tetragonal
solid solution and causing a decomposition in the T-phase. Figure 6.8 shows X-rays spectra for
Chapter 6. Thermal Stability of PS top coats

108
CSZ after annealing at 500
o
C in vacuum. The tetragonal phase has decomposed into cubic. In
addition, the colour of the as-sprayed coating changed to dark-brown/green (see Figure 6.9), which
suggests that the cerium cation has been reduced to its trivalent state. Dal Maschio et al [219]
noticed a volume diminution at 340
o
C during cool-down, believed to be due to the phase
transformation from cubic to tetragonal due to re-oxidation. The valance state of cerium in CSZ top
coats is believed to have a large influence on the stress state in the ceramic coating [219]. The
simultaneous presence of ceria and yttria makes it difficult to estimate the stabilizer content from
the peak position, using known relation between lattice parameters and stabilizer percentage.

Figure 6.9 CSZ (205NS) top coat (a) as-
sprayed and (b) after annealing at 500
o
C in
vacuum. The change in colour is due to
reduction of Ce
4+
to Ce
3+
.

6.1.2 Isothermal Heat Treatment
6.1.2.1 Yttria stabilized zirconia top coats
After holding the YSZ 204NS-1 top coat at 1300C for 100 h, the tetragonal T' phase decomposed
to a mixture of low-yttria tetragonal T'
l
,

high-yttria cubic F and T
2
with an intermediate yttria
composition (Figure 6.10). According to the phase diagram (see Figure 3.12a) YSZ top coats
composed of non-transformable T should decompose at high temperature, to a mixture of cubic
and tetragonal phases. On cooling to room temperature, the high yttria cubic phase may be retained,
or it may transform to a high yttria tetragonal phase and the low yttria equilibrium tetragonal phase
may transform to monoclinic. The estimated percentages of the phases present in the current
samples (and the yttria contents in each phase) are shown in Table 6.2. The percentage of yttria in
the tetragonal and cubic phases was deduced from the lattice parameters of each phase, according to
empirical equations (see section 5.3.4). The T
1
low-yttria tetragonal phase has a composition of
Chapter 6. Thermal Stability of PS top coats

109
1.7wt% Y
2
O
3
, the high-yttria cubic (F) has 13wt% Y
2
O
3
and the T'
2
tetragonal phase has a
composition of 6.3wt% Y
2
O
3
. The T gradually decomposes at high temperature into the
equilibrium high-yttria cubic and low yttria-tetragonal phases equilibrium, but total decomposition
is not achieved after 100 h and some intermediate T
2
is still present. Decomposition occurs by
gradual segregation of the yttria into lower and higher yttria regions, with corresponding changes in
the c/a ratio. Of course, it must be recognized that some changes in phase constitution, and possibly
in the composition of the phases, might have occurred during cooling to room temperature. The
cubic phase formed at high temperature is retained at room temperature and contains about 13 wt%
yttria, which agrees more or less with the phase diagram, which suggests that at 1300
o
C the
equilibrium concentration of the cubic phase should be about 12 wt%. This concentration exceeds
the compositional limit of the T phase and hence the cubic phase F is retained to room temperature
[217, 328].
27 28 29 30 31 32
I
n
t
e
n
s
i
t
y
2u ()
M (111)
T'(111) & F(111)
M (111)
(a)


Figure 6.10 XRD spectra in 2u ranges of (a) 27-32, and (b) 72-75 (deconvoluted) for heat treated YSZ
(204NS-1) for 100 h at 1300
o
C.

The 204NS YSZ top coat was heat treated for 50 h at 1300
o
C. The evolution of the phase content
can be seen in Figure 6.11. After 1 h at 1300
o
C, the coating is still predominantly T, but the peaks
appear to broaden. This is probably caused by compositional differences in the tetragonal phase due
to gradual diffusion of yttria, and corresponding changes in the lattice parameters. After 10 h of
heat treatment, further broadening of the tetragonal peaks is apparent and a small additional peak
appears which can be identified as the cubic peak. The trend continues for 20 h of heat treatment
and after 50 h it is possible to recognize two distinct tetragonal phases and the cubic phase. Figure
6.12 shows the deconvoluted profile the of 204NS YSZ top coat after heat treatment for 50 h at
Chapter 6. Thermal Stability of PS top coats

110
1300
o
C. The T phase has also decomposed into a mixture of low-yttria tetragonal T'
l
high-yttira
cubic F and T
2
with an intermediate yttria composition. The percentages of each phase present are
summarized in Table 6.2. As heat treatment progresses, the amount of intermediate yttria T
2

tetragonal phase appears to gradually decrease, while the low yttria T
1
and high yttria cubic phases
F appear to increase. Comparing the percentages of phases present in this top coat to those present
in the YSZ 204NS-1 top coat after heat treatment for 100 h at 1300
o
C, the amount of cubic and low
yttria phases present is less in this top coat than in the YSZ 204NS-1, since the longer heat
treatment has allowed more decomposition of the high yttria T
2
tetragonal phase. A trace of
monoclinic phase is also evident, which has probably resulted from the transformation of some of
the low yttria T
1
tetragonal phase during cooling to room temperature.
It is not clear why there is some monoclinic phase present in the YSZ 204NS top coat, while none is
present in the YSZ 204NS-1 top coat, which has the higher impurity content. The tetragonal-to-
monoclinic transformation is associated with a volume increase which can be detrimental for the
failure of TBCs, although the amount of monoclinic phase (2.9%) present in the YSZ 204NS top
coat in the current study is probably not significant enough to cause failure. It is worth noting,
however, that the presence of SiO
2
impurities probably aids the phase stability in the YSZ 204NS-1
powder.

28 28.5 29 29.5 30 30.5 31 31.5 32
as-sprayed
1 hour
2 hours
5 hours
10 hours
20 hours
50 hours
I
n
t
e
n
s
i
t
y
2u (
o
)

72 72.5 73 73.5 74 74.5 75
as-sprayed
1 hour
2 hours
5 hours
10 hours
20 hours
50 hours
I
n
t
e
n
s
i
t
y
2u (
o
)

Figure 6.11 XRD spectra in 2u ranges of (a) 27-32, and (b) 72-75 showing phase evolution for heat treated
YSZ (204NS) up to 50 h at 1300
o
C.
Chapter 6. Thermal Stability of PS top coats

111
27 28 29 30 31 32
I
n
t
e
n
s
i
t
y
2u ()
M (111)
T'(111) & F(111)
M (111)
(a)

Figure 6.12.XRD spectra in 2u ranges of (a) 27-32, and (b) 72-75 (deconvoluted) for heat treated YSZ
(204NS) for 50 h at 1300
o
C.

The Al
2
O
3
-doped-YSZ top coat (AE9018) was heat treated for 50h at 1300
o
C. This top coat has a
similar composition and impurity level to the YSZ 204NS top coat, with Al
2
O
3
being the only
significant difference. The T tetragonal phase decomposed to a mixture of low-yttria tetragonal T'
l

high-yttira cubic F and T
2
with an intermediate yttria composition (Figure 6.14). Even though the
amount of cubic phase present was slightly less than the corresponding level in the YSZ 204NS top
coat given the same heat treatment, it can not be said conclusively that the tetragonal phase appears
to be more stable. In both top coats, the percentage of intermediate yttria T
2
tetragonal phase is
comparable, indicating that approximately the same amount of decomposition into high and low
yttria containing regions took place. The gradual decomposition of the T
2
into higher and lower
yttria regions often results in compositional variations, which make the observed peaks broader, but
not necessarily identifiable as distinct phases of low and high yttria content. The addition of
alumina does not appear to make a significant difference to the stabilization of the tetragonal phases,
although it appears that yttria diffusion appears to be slightly slower in Al
2
O
3
-doped YSZ than in
YSZ top coats with similar amounts of other impurities.
The Al
2
O
3
-doped-YSZ top coat (AE8170), which has a very similar composition to the Al
2
O
3
-
doped-YSZ AE9018 top coat, was heat treated for 50 h at 1350
o
C (Figure 6.13). The percentage of
cubic phase present in this case is slightly higher than that observed in the AE9018 top coat that was
heat treated at a slightly lower temperature (Table 6.2). This highlights the influence of the heat
treatment temperature on the decomposition rate of the T phase into its equilibrium phase. The
temperature difference of 50
o
C significantly affects the percentage of cubic phase formed. As
mentioned previously, compositional variations resulting from the progressive diffusion of yttria
Chapter 6. Thermal Stability of PS top coats

112
into high- and low-yttria regions results in broadening of the observed peaks and sometimes it is
difficult to differentiate distinct phases and distinguish this from a single tetragonal phase exhibiting
a broad peak. In the current case, is possible to deconvolute the observed pattern into either a low-
yttria T
1
, high yttria cubic F and intermediate T
2
yttria phases or simply into a tetragonal and a
cubic phase. Another possibility is the existence of a mixture of phases with small compositional
variations. Thus x-ray analysis cannot in this case be reliably used to establish the phase
constitution. Other techniques, such as TEM might be useful to provide further information.

27 28 29 30 31 32
I
n
t
e
n
s
i
t
y
2u ()
M (111)
T'(111) & F(111)
M (111)
(a)


Figure 6.13 XRD spectra in 2u ranges of (a) 27-32, and (b) 72-75 (deconvoluted) for heat treated Al
2
O
3
-doped YSZ
(AE8170) for 50 h at 1350
o
C.
27 28 29 30 31 32
I
n
t
e
n
s
i
t
y
2u ()
M (111)
T'(111) & F(111)
M (111)
(a)


Figure 6.14 XRD spectra in 2u ranges of (a) 27-32, and (b) 72-75 (deconvoluted) for heat treated Al
2
O
3
-doped YSZ
(AE9018) for 50 h at 1300
o
C.

Chapter 6. Thermal Stability of PS top coats

113
6.1.2.2 Dysprosia stabilized zirconia top coats
The DSZ top coat (SPM6-2444) decomposes slowly during heat treatment and gradually a cubic
phase appears. The x-ray spectrum after 50 h at 1300
o
C is difficult to deconvolute and hence it can
not be said with certainty whether there are one or more tetragonal phases present (Figure 6.15). It
is almost certain that small compositional variations in the cubic and tetragonal intermediate phases
result in the observed broadening of the base of the two tetragonal peaks. According to the phase
diagram of this system [329], both the pyrochlore (P) phase Dy
2
Zr
2
O
7
and the zirconia solid
solution phases co-exist when the concentration of Dy
2
O
3
is above 15 mol%. No pyrochlore phase
was formed in the present work, due to the relatively low mol% of Dy
2
O
3
present. The x-ray
spectrum of DSZ after heat treatment at 1350
o
C for 100 h shows a considerably higher amount of
the cubic phase F (Figure 6.16). Since the temperature and heat treatment time were higher, it
comes as no surprise that there is more cubic phase present. A trace of monoclinic is also present.
There are no well-established equations relating the peak position to the phase composition for
dysprosia (as there are for yttria) and hence is not possible to calculate the dysprosia content in
these phases. From the present results it can be concluded that for the current amount of stabilizer
(3.58 mol%), dysprosia successfully stabilized the tetragonal phase in plasma sprayed coatings.
After heat treatment a certain amount of decomposition occurred, similar to that in conventional
YSZ top coats.

27 28 29 30 31 32
I
n
t
e
n
s
i
t
y
2u ()
M (111)
T'(111) & F(111)
M (111)
(a)

Figure 6.15 XRD spectra in 2u ranges of (a) 27-32, and (b) 72-75 (deconvoluted) for heat treated DSZ
(SMP6-2444) for 50 h at 1300
o
C.
Chapter 6. Thermal Stability of PS top coats

114
27 28 29 30 31 32
I
n
t
e
n
s
i
t
y
2u ()
M (111)
T'(111) & F(111)
M (111)
(a)

Figure 6.16 XRD spectra in 2u ranges of (a) 27-32, and (b) 72-75 (deconvoluted) for heat treated DSZ (SMP6-
2444) for 100 h at 1350
o
C.

6.1.2.3 Yttria-lanthana stabilized zirconia top coats
In the YLaSZ top coat after heat treatment, the tetragonal phases initially present in the as sprayed
coating decompose into a mixture of tetragonal cubic and the cubic pyrochlore (P) phase La
2
Zr
2
O
7
(Figure 6.17). The percentages of each phase were calculated using Rietveld analysis (see Table
6.2). The presence of the pyrochlore phase La
2
Zr
2
O
7
in LaSZ has been reported previously [227].
When both lanthana and yttria are present, it appears that decomposition of the lanthana-doped
zirconia and yttria-doped zirconia occurs simultaneously and a mixture of tetragonal, cubic and
pyrochlore phases is present. The formation of the pyrochlore phase P accompanies a volume
increase, which is thought to enhance the tetragonal-to-monoclinic phase transformation [228, 330].
The presence of both lanthana and yttria makes it difficult to estimate the stabilizer content from the
peak position. La
2
Zr
2
O
7
top coats consisting of the pyrochlore structure have been proposed as a
promising new TBC materials [230, 231, 331, 332], because of their stable structure up to the
melting points and lower thermal conductivity. Although, these top coats have lower cycling life
that YSZ top coats due to their lower CTE. However, in the present study the YLaSZ top coat was
not phase stable, probably due to the presence of both yttria and lanthana and it decomposed after
heat treatment forming a mixture of pyrochlore, cubic and tetragonal phases.

Chapter 6. Thermal Stability of PS top coats

115

27 28 29 30 31 32
I
n
t
e
n
s
i
t
y
2u ()
M (111)
T'(111) & F(111)
M(111)
(a)
La
2
Zr
2
O
7


Figure 6.17. XRD spectra in 2u ranges of (a) 27-32, and (b) 72-75 (deconvoluted) for heat treated YLaSZ
(AE8321) for 50 h at 1350
o
C.
Phases (%)
Top Coat / Heat treatment
Monoclinic M Cubic F Tetragonal T Tetragonal T
1

(low yttria)
Tetragonal T
2

(high yttria)
Pyrochlore P
204NS-1 / 100 h at 1300
o
C
Y
2
O
3
wt%
46
13
- 35
1.7
19
6.3
-
204NS / 50 h at 1300
o
C
Y
2
O
3
wt%
2.9 29.5
10
- 23.2
1.6
44.4
6.7
-
SPM6-2444 / 50 h at 1300
o
C 11 89 -
SPM6-2444 /100 h at 1350
o
C 2 30 68 -
AE8321 / 50h at 1350
o
C 39.2 37.3 12.5 11
AE8170 / 50 h at 1350
o
C
Y
2
O
3
wt%
32
14
55
1.8
13
6.6
-
AE9018 / 50 h at 1300
o
C
Y
2
O
3
wt%
17
14.6
- 35
1.7
48
6.7
-
Table 6.2. Phase constitution of heat treated detached thermal barrier coatings.

6.1.3 Heat Treatment Under a Thermal Gradient
Top coats of YSZ 204NS-1 (attached to the substrate) were heat treated in a thermal gradient for
48 h. The temperature gradient was monitored via thermocouples embedded in the substrate and by
a thermocouple embedded in a laser drilled hole in the top coat - see section 5.5.2. Typical top coat
temperatures ranged from about 1500
o
C at the free surface to 900
o
C at the interface. The XRD
spectra from material near the free surface indicates a mixture of low- and high-yttria tetragonal
phases (T'
1
and T'
2
), having compositions of 2.2wt% Y
2
O
3
and 11.8wt% Y
2
O
3
respectively, with no
residual cubic phase (Figure 6.18). No cubic phase has been retained down to room temperature,
while, as described previously, quite a significant amount is present after holding isothermally at
about 1300
o
C, even though the phase diagram indicates that more cubic material should be present
Chapter 6. Thermal Stability of PS top coats

116
at the higher temperature. Similar behaviour has been reported [217, 328] previously, particularly
for high cooling rates. The cubic phase forming at higher holding temperatures is expected to have
a lower yttria composition than if it had formed at lower temperatures, which may render it more
liable to transform during rapid cooling to the high-yttria tetragonal T
2
,

whereas when the high
temperature cubic phase has a high enough yttria content, the cubic polymorph is retained down to
room temperature. On the other hand, the XRD spectrum for the top coat material near the interface
with the bond coat indicates that this material remained predominantly T non-transformable and no
phase transformation took place (Figure 6.19). This is not surprising, since this material was being
held at about 900
o
C.

27 28 29 30 31 32
I
n
t
e
n
s
i
t
y
2u (
o
)
M (111)
T'(111) & F(111)
M (111)
(a)

72 73 74 75
I
n
t
e
n
s
i
t
y
2u (
o
)
(b)
F(400)
T'(400)
T'(004)

Figure 6.18. XRD spectra in 2u ranges of (a) 27-32, and (b) 72-75 for YSZ (204NS-1) top coats, after
gradient heat treatment, near the interface with the bond coat (which was at about 900
o
C).
27 28 29 30 31 32
I
n
t
e
n
s
i
t
y
2u (
o
)
M (111)
T'(111) & F(111)
M (111)
(a)

Figure 6.19 XRD spectra, with deconvoluted peaks and differential plots (showing the deviation between
the measured and modelled spectra) in 2u ranges of (a) 27-32
o
, and (b) 72-75
o
for YSZ (204NS-1) top coat
after thermal gradient heat treatment, near the free surface (which was at about 1500
o
C).

Chapter 6. Thermal Stability of PS top coats

117
6.2 Microstructural development under service conditions
6.2.1 Isothermal Heat Treatment
The microstructures of as-sprayed zirconia top coats, while strongly dependent on the spraying
parameters, always show certain characteristics. Figure 6.20 to Figure 6.25 show SEM micrographs
of fracture surfaces of PSZ top coats, before and after heat treatment. As-sprayed coatings exhibit
the characteristic splat structure. Through-thickness microcracks and pores are also present.
Through-thickness microcracks are formed during cooling in order to relieve the quenching stress
experienced by the splats [59]. Pores ranging in sizes can be identified. At higher magnification
(Figure 6.21(b)) adherent splats seem to be in close physical proximity to each other. The grain
structure within individual splats is columnar, due to the directional freezing of the splats after
impact with the substrate.
Compositional differences (i.e. different stabilizers) do not cause changes in observed
microstructural features (Figure 6.22 to Figure 6.25). The dimensions of the splat thickness, inter-
splat pores, large pores and through-thickness microcracks do not appear to vary significantly with
composition, when the same spraying parameters are used. However it is possible that some
systematic differences exist in the microstructure of coatings of different composition, which would
be possible to identify and quantify with other techniques, e.g. small angle neutron scattering. This
was beyond the scope of this study.
No compositional inhomogeneities could be detected in any of the structures, using EDX analysis.
This was also true for the alumina-doped YSZ, where no alumina was detected by EDX.
Isothermal heat treatment resulted in grain growth in all top coats, often bridging across interfaces
between splats in close physical proximity (see Figure 6.21). There is also evidence of the healing
of through-thickness microcracks. It is worth noting that large pores do not appear to heal even
after prolonged heat treatment. No qualitative difference in the microstructure was observed for top
coats heat treated for more that 50 h at temperatures in the range 1300-1450
o
C (Figure 6.21 to
Figure 6.23). However, quantitative characterization of the microstructure of PS YSZ top coat
using SANS [201] reported significant decreases in the surface area of the void system (up to 33%)
as a result of sintering, at temperatures as low as 1000
o
C. Cracks appear to sinter first and at lower
temperatures, followed by sintering of intersplat pores at higher temperatures. Thermal exposure
however, had no influence on large (>10m) pores [135]. This agrees well with the microstructural
features observed here.
Chapter 6. Thermal Stability of PS top coats

118

(a)

(b)

Figure 6.20. SEM micrographs of fracture surfaces of as-sprayed top coat of YSZ (204NS-1) (a) low magnification
(b) higher magnification showing some characteristic microstructural features of PS top coats.
(a)

(b)

Figure 6.21 SEM micrographs of fracture surfaces of detached top coat of YSZ (204NS-1) heat treated for 100 h at
(a)1300
o
C and (b) 1350
o
C.






Chapter 6. Thermal Stability of PS top coats

119
(a)

.(b)

Figure 6.22 SEM micrographs of detached plasma sprayed top coats heat treated at 1450
o
C for 100 h (a)
YSZ (204NS-1) and (b) DSZ (SPM6-2444).
(a)

(b)

Figure 6.23 SEM micrographs of fracture surface of DSZ (SPM6-2444) detached top coat (a) as-sprayed and
(b) heat treated at 1350
o
C for 100 h.

Chapter 6. Thermal Stability of PS top coats

120
(a)

(b)

Figure 6.24. SEM micrographs of fracture surface of YLaSZ (AE8321) detached top coat (a) as-sprayed
and (b) heat treated at 1350
o
C for 50 h (these coatings were sprayed by Sulzer).
(a)

(b)

Figure 6.25 SEM micrographs of fracture surface of Al
2
O
3
-doped YSZ (AE9018) detached top coat (a)
as-sprayed and (b) heat treated at 1350
o
C for 50 h (these coatings were sprayed by Sulzer).

6.2.2 Heat Treatment Under a Thermal Gradient
Under service conditions, TBCs operate with high thermal gradients. It is, thus, of interest to study
the microstructural changes that occur under these conditions. A YSZ top coat attached to a
substrate was heat treated under an imposed thermal gradient, for 17 h, with the top surface
temperature around 1500
o
C and the temperature near the top coat/substrate interface about 900
o
C.
Near the TC/BC interface (Figure 6.26a), there was little or no sintering. Microcracks are still
present and bonding between splats is poor. Grains in individual splats have maintained their
columnar structure. Near the free surface of the TC, on the other hand (Figure 6.26b), pronounced
Chapter 6. Thermal Stability of PS top coats

121
sintering has taken place. There has been extensive healing of microcracks and grain growth has
occurred, with many grains becoming both larger and more equiaxed in morphology.

(a)

(b)

Figure 6.26. SEM micrographs of YSZ top coat (204NS-1) after exposure to a high thermal gradient
for 7 h, (a) near the TC/BC interface and (b) near the TC outer surface.
(a)
)
(b)

Figure 6.27.Higher magnification SEM micrographs of YSZ top coat (204NS-1) after exposure to a
high thermal gradient for 17 h, (a) near the TC/BC interface and (b) near the TC outer surface.
(higher magnification).

Chapter 6. Thermal Stability of PS top coats

122
6.3 Effect of Heat Treatment of Detached Top Coats on sintering
behaviour
6.3.1 Effect of Composition and Morphology on Sintering Characteristics
The effect of heat treatment on the sintering behaviour of detached TBCs was investigated by
monitoring linear dimensional changes during isothermal heat treatment. Dilatometry experiments
were carried out on detached top coats of different compositions. The net linear contraction -
AL/L
o
(%) in the through-thickness and in-plane direction is plotted in Figure 6.28 and Figure 6.29
against time at temperature for detached top coats of different composition.
Thermal contraction (shrinkage) occurs due to sintering of the coatings. Surface transport
mechanisms do not cause shrinkage [150]. Bulk transport processes which include grain boundary
diffusion, and volume/lattice diffusion, are responsible for the shrinkage observed. For all
specimens, sintering rates change with time. Faster shrinkage rates were observed initially,
followed by a fall to relatively constant rates. The fast shrinkage observed at the early stages, even
under conditions where sintering models [151, 160-163] predict a dominance of surface diffusion,
demonstrates that volume and/or grain-boundary diffusion mechanisms are occurring as well.
Similar trends in the sintering behaviour have been observed by other researchers [182, 197, 198].
However the anisotropy between the in-plane and through-thickness contractions has not been
reported previously. Eaton and Novak [197] performed similar measurements on YSZ and other
coatings, and observed similar behaviour. Sintering of the top coat will result in enhanced intersplat
bonding, lower strain tolerance and higher stiffness of the coating. Previous work [60, 197] has
indicated that a sharp increase in stiffness occurs after a short period of exposure to high
temperature, followed by progressively slower increases. These observations are broadly in
agreement with the current dilatometry results.
Chapter 6. Thermal Stability of PS top coats

123
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
0 20 40 60 80 100 120 140 160
Time (h)
N
e
t

l
i
n
e
a
r

c
o
n
t
r
a
c
t
i
o
n

(
%
)
YSZ (204NS-1)
CSZ (205NS)
DSZ (SPM6-2444)
YSZ Al2O3-doped (AE9018)
YSZ Al2O3-doped (AE8170)
YLaSZ (AE8321)
YSZ (204NS)
Al
2
O
3
-doped
Al
2
O
3
-doped

Figure 6.28 Net linear contraction from dilatometry data for detached PS top coats heat treated at 1350
o
C in the
through-thickness direction.
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
0 20 40 60 80 100 120 140 160
Time (h)
N
e
t

l
i
n
e
a
r

c
o
n
t
r
a
c
t
i
o
n

(
%
)
YSZ Al2O3-doped (AE8170)
YSZ (204NS-1)
CSZ (205NS)
DSZ (SPM6-2444)
YLaSZ (AE8321)
YSZ (204NS)
Al
2
O
3
-doped

Figure 6.29 Net linear contraction from dilatometry data for detached PS top coats heat treated at 1350
o
C in the in
plane direction.

Chapter 6. Thermal Stability of PS top coats

124
Coatings with different compositions showed significantly different sintering rates. To analyze the
sintering behaviour of differently doped zirconia, the factors affecting bulk diffusion must be
considered. Shrinkage during sintering occurs by grain boundary diffusion and volume/lattice
diffusion. Shrinkage can also occur by dislocation diffusion, but since the contribution from this
mechanism is not likely to be significant in the present case we will not mention this mechanism
any further. At an atomic scale, sintering and shrinkage occur via mass flux of atoms, which is
driven by a gradient of chemical potential. The mass flux for sintering, which is derived from
Ficks first law is
( )
b
J D
k T
O
o = V (6.1)
where D is the diffusivity or diffusion coefficient, is the volume per lattice site, k
b
is Boltzmanns
constant, T is the temperature and V(o) the gradient of hydrostatic stress. The stress is dependent
on the surface free energy, which may be altered by impurities or dopants absorbed at the surface.
Equally, the volume per lattice site is a structural parameter, dependant on the atomic and crystal
structure. The diffusivity D is specific to each diffusion mechanism and it follows an Arrhenius-
type equation. From the above equation, it can be noted that there is a direct dependence of the
atom flux during solid state sintering on structural parameters that can be influenced by dopants and
impurities. In the case of an ionic compound, transport of anions and cations during solid state
sintering by either mechanism occurs with the help of the defects existing in the solid and with the
help of the gradient concentration of these defects existing between the surface and the bulk, or
between the bulk and the grain boundaries [147, 333]. The defects in a solid are produced either as
a result of thermal equilibrium (intrinsic) or as a result of suitable doping (extrinsic) [334]. Intrinsic
defects can be ions occupying an interstitial site (Frenkel defects) or cation and anion vacancies
with no interstitial or misplaced ions (Shottky defects). The number of defects is correctly balanced
as to maintain the stoichiometric formula and preserve electrical neutrality. In sintering of ionic
compounds, the various constituents must diffuse is the stoichiometric ratio of the compound [335].
Thus, the rate of diffusion is regulated by stoichiometric restrictions and it will be determined by
the slowest atomic species [156].
In YSZ, and equally in zirconia stabilized by other trivalent oxides, the introduction of aliovalent
components introduces defects. The majority of defects in YSZ are oxygen vacancies and yttrium
aliovalent dopants on normal cation sites. The slowest and rate controlling diffusional process in
yttria stabilized zirconia is confirmed to be transport of zirconium and yttrium cations [180, 181].
The oxygen vacancies have far lower activation energy for diffusion than the solute cations. The
Chapter 6. Thermal Stability of PS top coats

125
latter are expected to act as the controlling species for grain growth [180-182]. The defect reaction
in YSZ can be written, using the Krger-Vink notation [334], as
2
ZrO ' ..
2 3 Zr o o
YO 2Y + 3O + V

(6.2)
Other possible defects include zirconium interstitials, oxygen interstitials, yttrium interstitials and
zirconium vacancies. For stabilized zirconia with other trivalent ions, the same principle applies to
generation of defects and controlling diffusant species, since it has been confirmed the dopant
trivalent ions in zirconia occupy the cation sites [204]. Comparing the diffusion coefficients of Zr
4+

and O
2-
in ZrO
2
, the Zr
4+
is the more stable, slower diffusing species and is therefore expected to be
the rate controlling species in sintering.
Element Diffusion Coefficient (m
2
s
-1
)
Zr

10
-19

O 210
-13

Table 6.3 Reported diffusion coefficient of Zr and O
in ZrO
2

[333]

The presence of extrinsic defects will affect the transport properties in an ionic crystal. The precise
effect of these defects on diffusion depends on the defect and ion mobility [336]. Sintering rates
should increase if the concentration of slow defects is increased [191, 333]. In the YSZ the slowest
diffusing species is the Zr
4+
ion. The mechanism of cation diffusion in YSZ is disputed in the
literature. Kilo et al [183, 184]support that diffusion of cations in YSZ occurs via zirconium
vacancies formed by the creation of Schottky defects. In this case the presence of additional
oxygen vacancies should not affect the cation diffusion. Chien et al [180] support that cation
diffusion occurs by cation vacancies associated with oxygen vacancies. In this case, the addition of
oxygen vacancies might affect cation diffusion. Therefore, it is not clear how the presence of
oxygen vacancies should affect the cation diffusion rate in YSZ. The same mechanism that governs
cation diffusion in YSZ is expected to govern the diffusion of other trivalent lanthanides in zirconia
[337].
For CSZ, the stabilizing species have the same valence as the host ion and so no extrinsic defects
should be introduced. In these cases the intrinsic thermal defects contribute to the diffusion process
[335]. As with the trivalent stabilized zirconias, the cations are expected to be the rate controlling
species for sintering.
Another bulk diffusion mechanism, that also contributes to the shrinkage observed, is grain
boundary diffusion. During grain boundary diffusion, mass removed along the grain boundary is
Chapter 6. Thermal Stability of PS top coats

126
re-deposited at the sinter bond. Shrinkage is realized because of mass transport from the grain
boundary to the free surfaces [147]. As surface energy is consumed by shrinkage of pores, new
grain boundaries emerge at the sinter bonds between splats. If the grain boundary energy is high,
the replacement of free surfaces by grain boundaries might be unfavorable. Segregation of Y, La,
Ca and Ce solute dopants at the grain boundary in zirconia has been observed by various
researchers [183, 186-190]. Segregation of solid solution additives at the grain boundaries lowers
grain boundary energy and aids shrinkage of pores by grain boundary diffusion [149, 155, 191].
In the absence of a solid solution additive, grain growth occurs in order to reduce the total grain
boundary area [149]. The disappearance of grain boundaries due to grain growth will reduce
shrinkage by grain boundary diffusion. Therefore high mobility grain boundaries, which promote
grain growth, are not favorable for high shrinkage. Segregation of solid solution additives at the
grain boundaries, lowers the grain boundary mobility, and contributes to higher shrinkage [149,
155]. The size of the doping cations appears to influence their effect on grain boundary mobility
and grain growth [186]. For stabilized zirconia with different cations, including La
3+
,
Y
3+
and Ce
4+
,
larger, low valence cations are more effective in reducing the grain growth, in agreement with the
solute drag models [153, 154]. Hence top coats with, larger, low valance stabilizing cations should
exhibit higher sintering rates. The mobility of the grain boundaries, with respect to the solute drag
mechanism, is given by [154]
2
C
D
T
~ (6.3)
where is the grain boundary mobility, D is the diffusivity of the solute cation, T is the temperature
and C is the excess solute concentration at the grain boundaries. The highest sintering rates should
be exhibited by coatings in which the doping cations decrease more effectively the grain boundary
mobility. Thus, since the cations are the rate controlling species for sintering, the doping cations
should also have intrinsic low diffusivity. In addition, effective dopants for decreasing the sintering
rate should be soluble within zirconia [338], and not segregate at the grain boundaries. The
solubility limit depends both on radius and valence differences between Zr and the dopant cation.
The solid solubility limits decrease with increase in the ionic misfits (see Table 6.4).
Extensive work by Li and Chen[204, 208, 339, 340] confirmed that a common feature to all the
stabilizing cations is that, when in solution, they maintain the nature of the first neighbor bonds that
they have in the pure stabilizing oxide. Therefore, in the absence of diffusivity data, the melting
temperature of the stabilizing dopant reflects the binding energy [341] and is an indicator of
diffusion rates. Lower melting point materials have weaker atomic bonding and higher molar
Chapter 6. Thermal Stability of PS top coats

127
volume which allows faster diffusion at a given temperature. Hence, for sintering inhibition, oxides
with a high melting temperature, good solubility in zirconia and low ionic misfit would be most
effective in reducing the sintering rate (see Table 6.4).
Dopant oxide or
impurity
Ionic radius
(nm)
Solid solubility
(mol%)
Melting point

(
o
C)
Reference
ZrO
2
(host material) 0.084 - 2715
Y
2
O
3
0.102 2.9 2690 [205]
La
2
O
3
0.116 1.15 2305 [227]
CeO
2
0.097 18 2400 [307]
Dy
2
O
3
0.103 2408
Al
2
O
3
0.051 0.5 2054 [323]
Table 6.4 Ionic radii, solid solubility limits and melting points of dopant oxides or impurities.

The lowest and highest shrinkage rate are exhibited by YSZ (204NS) and YSZ (204NS-1), the top
coat with the lowest and highest amount of impurities, respectively. Impurities which are not in
solid solution, can provide a continuous low diffusivity pathway at the grain boundary which will
restrain grain boundary mobility and reduce grain growth and enhance shrinkage [153], or, provide
a high diffusivity pathway, e.g. liquid phase at the grain boundaries, which will enhance shrinkage
[172]. The effect of impurities on the sintering rate and grain growth, and particularly the role of
SiO
2
and Al
2
O
3
, has been addressed by several researchers [186, 192, 193, 196, 198, 338, 342].
Vassen et al [198] found an increase in the shrinkage rate with increased silica and alumina content,
but the effect was more pronounced for SiO
2
. This is in agreement with the results presented here,
where the sintering rate of YSZ (204NS-1) with 0.1 wt% SiO
2
is relatively high. Both the Al
2
O
3
-
doped YSZ top coats also exhibit higher sintering rates than undoped pure 204NS. Verkerk et al
[192] found that small additions of alumina reduced grain growth by the solute drag effect.
Matsumoto et al [343] reported that the addition of 1 mol% Al
2
O
3
to 4 mol% YSZ resulted in rapid
densification of EBPVD TBCs after sintering for 10 h at 1300
o
C. However they supported the
argument that the presence of a liquid phase is unlikely. Srdic et al [196] also found that a small
amount of Al
2
O
3
in zirconia suppressed grain growth and enhanced densification.
Thermal contraction is anisotropic for all materials, with greater shrinkage rates in the through
thickness direction (see Figure 6.30 to Figure 6.33). This anisotropic behaviour is attributed to the
relative amount of surface area inter-splat pores and intra-splat microcracks. PS top coats have
greater surface area of inter-splat pores, compared to intra-splat cracks [131, 136]. According to
Wang et al [131], shrinkage of the inter-splat interfaces may have greater influence on the thermal
Chapter 6. Thermal Stability of PS top coats

128
conductivity than on the elastic modulus. Their role as a barrier to heat transfer may be more
significant than as a barrier to force transfer [131].

0
0.5
1
1.5
2
0 20 40 60 80 100 120 140 160
in plane at 1350
o
C
through thickness at 1350
o
C
in plane at 1450
o
C
through thickness at 1450
o
C
N
e
t

L
i
n
e
a
r

C
o
n
t
r
a
c
t
i
o
n

(
%
)
Time (hours)

0
0.1
0.2
0.3
0.4
0.5
0 20 40 60 80 100
in plane at 1300
o
C
through thickness at 1300
o
C
in plane at 1350
o
C
through thickness at 1350
o
C
N
e
t

L
i
n
e
a
r

C
o
n
t
r
a
c
t
i
o
n

(
%
)
Time (hours)

Figure 6.30 Net linear contraction for detached top
coats of YSZ (204NS-1) heat treated at 1350
o
C and
1450
o
C, showing anisotropy of contraction.
Figure 6.31 Net linear contraction for detached top
coats of YSZ (204NS) heat treated at 1300
o
C and
1350
o
C, showing anisotropy of contraction.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0 20 40 60 80 100 120
in plane at 1350
o
C
through thickness at 1350
o
C
in plane at 1450
o
C
through thickness at 1450
o
C
N
e
t

L
i
n
e
a
r

C
o
n
t
r
a
c
t
i
o
n

(
%
)
Time (hours)

0
0.2
0.4
0.6
0.8
1
1.2
1.4
0 20 40 60 80 100
in plane at 1350
o
C
through thickness at 1350
o
C
in plane at 1450
o
C
through thickness at 1450
o
C
N
e
t

L
i
n
e
a
r

C
o
n
t
r
a
c
t
i
o
n

(
%
)
Time (hours)

Figure 6.32Net linear contraction for detached top coats
of DSZ (SPM6-2444) heat treated at 1350
o
C and
1450
o
C, showing anisotropy of contraction.
Figure 6.33 Net linear contraction for detached top
coats of CSZ (205NS) heat treated at 1350
o
C and
1450
o
C, showing anisotropy of contraction.

Chapter 6. Thermal Stability of PS top coats

129
6.3.2 Effect of Phase Transformation on Measured Volume Changes
An estimate can be made of the magnitude of volume change associated with the relevant phase
changes, or at least an upper bound can be placed on them. There is a particular interest in the
tetragonal-to-cubic transformation, since this is likely to occur during extended heating at high
temperature. This can be seen from the phase diagram (Figure 3.12a) [306]. The volume change
associated with the tetragonal-monoclinic transformation is well established [212] in literature, to
be about 4%. For the tetragonal to cubic transformation, since the number of atoms per unit cell is
the same for (all of the possible) tetragonal and cubic phases, to estimate the volume change it is
only necessary to establish the volume of the unit cell in each case. The volume of the unit cell in
each case can be obtained using lattice parameter values inferred from measured X-ray peak
positions

. The composition of the cubic and tetragonal phases can also be deduced from these
measured lattice parameters using equations 5.3 and 5.4. Measured peak positions, lattice
parameters and unit cell volumes for YSZ (204NS-1) and YSZ (204NS) after various heat
treatments are given in Table 6.5 and Table 6.6.
The volume contraction associated with the phase changes in YSZ (204NS-1) after heat treatment at
1300
o
C for 100 h is about 0.5 %. The actual shrinkage observed after cooling to room temperature,
derived from the linear contractions, amounts to about 1.9 %. Hence microstructural effects due to
sintering, and not phase changes, must be primarily responsible for the contraction observed.
For the YSZ 204NS-1 top coat, no transformation of the monoclinic phase was observed and, since
this transformation has the most significant volume change, the volume changes due to phase
transformations associated with the YSZ (204NS) top coat, which contained 2.9 (%) monoclinic
phase after 50 h at 1300
o
C, were investigated. In this case, the volume change associated with
phase transformation after heat treatment is V/V
o
= -0.35%. The actual shrinkage observed after
cooling to room temperature is L/L
o
= -0.26% in the in-plane direction and L/L
o
= -0.73% in the
through-thickness direction, which amounts to a total volume change of V/V
o
= -1.2%. Hence,
the shrinkage observed can be attributed primarily to microstructural changes associated with
sintering effects.
Phase /
Treatment
tetragonal T' /
as-sprayed
tetragonal T'
1
/
100 h @ 1300C
tetragonal T'
2
/
100 h @ 1300C
cubic F /
100 h @ 1300C

Of course, the X-ray data were obtained at room temperature, whereas interest centres on volume at high
temperature. However, the difference in thermal expansivities between cubic and tetragonal phases is apparently [366]
less than 1 10
-6
K
-1
, so the possible error from this source must be less than about 0.1% in linear dimensions.
Chapter 6. Thermal Stability of PS top coats

130
2u for (004) K
o1
73.139 73.030 73.236 -
2u for (400) K
o1
73.981 74.447 74.217 73.738
c (nm) 0.51737 0.51803 0.51678 -
a (nm) 0.51231 0.50956 0.51091 0.51375
wt% Y
2
O
3
7.8 1.7 6.3 13
vol. of unit cell (nm
3
) 0.13579 0.13451 0.13489 0.13561
Composition (%)
100 35 19 46
AV/V (%) 0 -0.94 -0.66 -0.13
Table 6.5. Measured X-ray peak positions, deduced lattice parameters and associated unit cell volumes for
phases within YSZ (204NS-1) coating, with and without a prior heat treatment. Also shown are percentages
of phases present and phase compositions from the lattice parameters, using eqns.5.3 and 5.4.
An upper bound on the expected volume change can be obtained by assuming that thermodynamic
equilibrium is attained at the heat treatment temperature. For example, at ~1300C, up to ~50% of
the tetragonal phase could be transformed to the cubic phase, for material with an overall
composition of about ZrO
2
-7wt%Y
2
O
3
. It can be seen that, a volume contraction of up to about
0.50.6% might be expected when tetragonal material containing about 7wt% Y
2
O
3
transforms to
50% of F phase containing 13wt% Y
2
O
3
and 50% of T'
1
phase containing 2wt% Y
2
O
3
. The
corresponding linear contraction is thus about 0.2%. It can be concluded that phase changes are
expected to make only relatively small contributions to dilatometry data, indicating linear
contractions of the order of 1% or more. This calculation thus also confirms that the contraction
can be attributed primarily to microstructural changes associated with sintering effects. If anything,
the effect of taking phase transformation-induced linear contraction into account would be to
increase the deduced anisotropy of the sintering contraction.
Another interesting observation on the measured linear shrinkage is the difference observed
between the total linear shrinkage observed at temperature and the total linear shrinkage observed
after cooling at room temperature. For the YSZ (204NS) top coat heat treated at 1300
o
C, in the in-
plane direction, there is a difference of about 0.15 (%) and in the through-thickness direction, of
about 0.47 (%) (see Figure 6.35). In order to confirm that these differences were genuine,
calibration with a blank alumina sample was performed, which, as expected, showed no difference
between the shrinkage observed at high temperature and after cooling (the shrinkage in both cases
was zero) (see Figure 6.34). This difference is attributed to phase changes occurring during cooling
down, particularly the tetragonal to monoclinic transformation, which is known to have a 4 %
volume change.
Chapter 6. Thermal Stability of PS top coats

131
Phase /
Treatment
tetragonal T' /
as-sprayed
tetragonal T'
1
/
50 h @ 1300C
tetragonal T'
2
/
50 h @ 1300C
cubic F /
50 h @ 1300C
Monoclinic/
50 h @ 1300C
2u for (004) K
o1
73.1060 73.3219 - -
2u for (400) K
o1
74.5336 74.2704 73.8420 -
c (nm) 0.51696 0.51757 0.51626 - -
a (nm) 0.51156 0.50905 0.51060 0.51313 -
wt% Y
2
O
3
7.3 1.6 6.7 10 -
vol. of unit cell
(nm
3
)
0.13528 0.13412 0.13459 0.13511 -
Composition (%)
100 23.2 44.4 29.5 2.9
AV/V (%) 0 -0.86 -0.51 -0.13 4
Table 6.6 Measured X-ray peak positions, deduced lattice parameters and associated unit cell volumes for phases
within YSZ (204NS) coating, with and without a prior heat treatment. Also shown are percentages of phases present
and phase compositions from the lattice parameters, using eqns.5.3 and 5.4.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0
200
400
600
800
1000
1200
1400
0 1 2 3 4 5
AL/L
o
for T= 600
o
C
AL/L
o
for T= 1300
o
C
T = 600
o
C
T = 1300
o
C
A
L
/
L
o

(
%
)
T
e
m
p
e
r
a
t
u
r
e

(
o

C
)
Time (h)

-1
-0.5
0
0.5
1
1.5
2
2.5
3
0
200
400
600
800
1000
1200
1400
0 10 20 30 40 50 60
through-thickness
in-plane
Temperature
A
L
/
L
o

(
%
)
T
e
m
p
e
r
a
t
u
r
e

(
o
C
)
Time (hours)
YSZ (204NS)

Figure 6.34. Dilatometry calibration plots for alumina,
showing no difference in the shrinkage observed at high
temperature and after cooling to room temperature.
Figure 6.35. Dilatometry plots for YSZ (204NS) in-plane
and through-thickness, showing the difference in the total
linear shrinkage observed at high temperature and after
cooling to room temperature.

6.3.3 Effect of Thermal Cycling on Sintering Behaviour
Thermal barrier coatings under service conditions, particularly in the aerospace industry are
thermally cycled routinely. The effect of thermal cycling on the sintering behaviour was examined
Chapter 6. Thermal Stability of PS top coats

132
using the dilatometry technique. Detached top coats of YSZ (204NS) were thermally cycled and
the shrinkage was measured in-situ in the in-plane direction. The thermal cycle performed and the
shrinkage observed is shown in Figure 6.36. The individual high temperature segments are plotted
cumulatively in Figure 6.37 and compared with the shrinkage observed during continuous heat
treatment for the same time. During thermal cycling, faster shrinkage is observed during the initial
cycles, followed by slower shrinkage rates at later cycles, as is the case for continnous heat
treatment. Thus the tendency for high initial sintering rates is not altered by thermal cycling.
However, surprisingly, the shrinkage observed during thermal cycling is less than that observed
when heat treatment happens continuously, for the same amount of time at high temperature.
Cooling down and re-heating appears to slow down the shrinkage, possible due to phase changes
occurring during repeated cool down-heat up cycles which might affect the diffusion processes at
high temperature, which do not resume at the same rate.

-0.5
0
0.5
1
1.5
2
0 10 20 30 40 50 60 70 80
Time (h)
A
L
/
L
o

(
%
)
0
200
400
600
800
1000
1200
1400
T
e
m
p
e
r
a
t
u
r
e

(
o
C
)
dL/L
o
(%)
Temperature (
o
C)

0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0 10 20 30 40 50 60
Time (hours )
-
A
L
/
L
o

(
%
)
50h continuous
0-1 hours
1-2 hours
2-5 hours
5-10 hours
10-20 hours
20-50 hours
1300
o
C
YSZ (204NS)

Figure 6.36. Linear shrinkage observed during thermal
cycling of detached YSZ (204NS) top coat in the in-plane
direction.
Figure 6.37. Comparison between linear shrinkage observed
during heat treatment of detached YSZ (204NS) top coat at
1300
o
C, carried out continuously and during thermal cycling.

6.3.4 Effect of Sintering on Coating Porosity
Porosity measurements on as-sprayed and isothermally heat treated PSZ samples were conducted to
complement the dilatometry results, since it is well known that relative shrinkage during sintering is
related to the initial porosity of the material, and the higher the initial porosity, the higher the
shrinkage observed [147]. The theoretical density was calculated from the unit cell volume and the
Chapter 6. Thermal Stability of PS top coats

133
atomic weight of the elements present, as described in section 5.3.5. The theoretical density was
calculated for the as-sprayed coatings. Where more than one phase was present in the as-sprayed
coating (as in the YLaSZ coating), the unit cell volume of each of these phases and the relative
percentage of each phase, was calculated in order to derive the overall density. The theoretical
densities after heat treatment, taking into account the phase changes, were not calculated. Since the
error involved in the measurement of porosity is of the order of the change in porosity due to heat
treatment, no additional information could be obtained by more detailed evaluation of the
theoretical density. The porosity levels of as-sprayed top coats of all materials are presented in
Figure 6.38. The porosity of all coatings is similar with a mean between 10-12 % and a standard
deviation of about 3-4 %. Since the porosity of plasma sprayed coatings does not differ much with
different materials, it can be concluded that the trends observed during dilatometry are not due to
difference in initial porosity levels of the samples, but probably represent inherent difference in
their sintering characteristics.
Top Coat Material YSZ
(204NS-1)
YSZ
(204NS)
DSZ
(SMP6-2444)
CSZ
(205NS)
YLaSZ
(AE8321)
doped-YSZ
(AE8170)
doped-YSZ
(AE9018)
Theoretical density
(gr cm
-3
)
6.0429 6.0406 6.2953 6.3658 6.1136 6.04980 6.0632
Table 6.7. Theoretical density of as-sprayed top coats deduced from X-ray data and top coat composition.
Figure 6.39 is a plot of the porosity as a function of heat treatment time at 1350
o
C and at 1450
o
C,
for YSZ(204NS-1) and DSZ detached top coats. Porosity decreases slightly with heat treatment, and
the change is the order of 3-4%. This observation is more or less consistent with the linear changes
observed during dilatometry for these coatings after the same heat treatment time at 1350
o
C and
1450
o
C, where linear contraction in the through-thickness direction is of the order of 1.2-1.7% and
in the-plane direction, of the order 0.3-0.5%, which corresponds to a volume contraction of about
2-3%. It is not expected that it would be possible to differentiate between the changes in porosity
for different top coats (YSZ and DSZ), since the errors involved in this measurement are larger than
the difference in the shrinkage rate between the top coats. For the same reason, no significant
difference between the changes in porosity after heat treatment at 1350
o
C and 1450
o
C is apparent.
The small changes in porosity are consistent with the microstructural features observed after heat
treatment (see section 6.3). Most of the porosity is in the form of large pores that do not appear to
sinter with heat treatment. Changes in porosity are, therefore, due to sintering of finer pores.
Chapter 6. Thermal Stability of PS top coats

134
5
7
9
11
13
15
17
19
P
o
r
o
s
i
t
y

(
%
)
YSZ (204NS-1)
YSZ(204NS)
CSZ (205NS)
DSZ (SPM6-2444)
YLaSZ (AE321)
doped-YSZ (AE8170)
doped YSZ (AE9018)
Average

Figure 6.38. Porosity measurements for as-sprayed top coats.
4
5
6
7
8
9
10
11
12
13
14
0 20 40 60 80 100 120 140 160
Heat treatment time (h)
P
o
r
o
s
i
t
y

(
%
)
YSZ heat treated at 1350C
DSZ heat treated at 1350C
YSZ heat treated at 1450C
DSZ heat treated at 1450C

Figure 6.39 Porosity measurement of YSZ and DSZ top coats after heat treatment at 1350
o
C and 1450
o
C.

Chapter 6. Thermal Stability of PS top coats

135
6.4 Conclusions
The following conclusions can be drawn from this work.
1 Exposure to high temperatures induced phase changes in all top coats studied, which may affect
the microstructure and perhaps the stress state in TBCs. The YSZ as-sprayed top coats are
entirely the so-called non-transformable tetragonal T' phase. At temperatures above 1300
o
C
they decompose to their equilibrium tetragonal (T) and cubic (F) phases. On cooling to room
temperature the coatings consist of mixtures of cubic (F) phases, and high-yttria tetragonal
phases (T'
1
and T'
2
). The presence of SiO
2
and Al
2
O
3
impurities appears to decrease the rate of
decomposition.
2 Dysprosia, in the amount present in the current study (3.58 mol%), successfully stabilized the
tetragonal phase in plasma sprayed zirconia coatings. After heat treatment, a certain amount of
decomposition occurred, similar to that in conventional YSZ top coats.
3 YLaSZ as-sprayed top coats consisted of two tetragonal phases. After heat treatment, the
YLaSZ top coat was not phase-stable, probably due to the presence of both yttria and lanthana
and it decomposed to a mixture of cubic (F), high-yttria tetragonal phases (T'
1
and T'
2
), and the
pyrochlore (P) phase.
4 The Ce
4+
ion in the CSZ top coat was reduced to Ce
3+
during annealing at 500
o
C in vacuum. As
a result, the non-transformable tetragonal T' phase decomposed into the cubic (F) phase. The
cerium valence state has a great influence on the phase composition, and it is believed to affect
the stress state in the coating.
5 The high thermal gradients commonly present in TBCs under service conditions lead to widely
varying conditions within the top coat. Sintering is likely to occur much more rapidly near the
free surface, not only because the temperatures are higher there, but because the residual
stresses tend to be more compressive. After heat treatment in a high thermal gradient, material
near the bond coat remained entirely T', whereas material near the free surface was composed of
a mixture of low- and high-yttria tetragonal phases (T'
1
and T'
2
), which formed on cooling from
the high temperature tetragonal (T) and cubic (F) phases, respectively.
6 Detached top coats exhibit substantial reductions in linear dimensions following exposure to
elevated temperatures. These linear contractions are greater in the through-thickness direction
than the in-plane directions. This is thought to be at least primarily associated with
microstructural changes induced by sintering processes and the anisotropy is a consequence of
Chapter 6. Thermal Stability of PS top coats

136
the splat and pore geometry in these coatings. Any contributions arising from phase changes
occurring at high temperature are expected to be relatively small.
7 Stabilizers affect the shrinkage rate of top coats. It is thought that this is primarily due to
changes in the diffusion kinetics, since the microstructural characteristics of all the top coats
studied where qualitatively similar. Stabilizers may affect the volume diffusion coefficient,
grain boundary diffusion coefficients, grain boundary mobility or alter the ratio of grain
boundary to surface energy. Any of these could lead to differences in the shrinkage observed.
8 The impurity level, specifically that of SiO
2
, significantly increases the shrinkage observed. It
is thought that impurities which are not in solid solution provide a continuous low diffusivity
pathway at the grain boundary, which will restrain grain boundary mobility, reduce grain
growth and enhance shrinkage. Alternatively, impurities may also do so by providing a high
diffusivity pathway, e.g. liquid phase at the grain boundaries, which will enhance shrinkage.
9 Thermal cycling, commonly experienced by TBCs under service conditions, affects their
shrinkage. The shrinkage observed during thermal cycling is less than that observed under
continuous heat treatment. Cooling down and re-heating appears to slow down the shrinkage,
possibly due to phase changes occurring during repeated cool down-heat up cycles which might
affect the diffusion processes at high temperature, which do not resume at the same rate.

Chapter 7: Thermomechanical Behaviour of PS TBCs

137

7 Thermomechanical Behaviour of Plasma Sprayed
TBCs
7.1 Effect of Top Coat Sintering on Mechanical Properties
7.1.1 Cantilever Bending
This technique, which measures the global in-plane stiffness of the coating, was applied to detached,
as-sprayed top coats. A typical loading/unloading plot is presented in Figure 7.1. The hysteresis
observed could be attributed to the opening of cracks during initial loading, that fail to close when
the load is removed. Diminishing hysteresis was observed after loading and unloading several
times. The average value for the gradient of a load/unload/load/unload sequence was calculated.
This value was used to determine the stiffness, by applying equation 5.8. The stiffness of as-
sprayed zirconia-based top coats was in the range of 10-15 GPa (see Figure 7.2). It is worth noting
that elastic moduli of PS top coats depend on microstructure, which in turn depend on the
processing parameters, so values reported in the literature vary over a wide range corresponding to
the different parameters used (see section 3.2.1). However, the global stiffness of PS top coats is
expected to be considerably lower than that of the corresponding dense ceramic, due to the
microcracks, interlamella pores and the poor intersplat bonding of as-sprayed coatings. After
isothermal heat treatment for 88 h at 1350
o
C in air, the elastic modulus increases significantly due
to sintering effects (see Figure 7.2). Similar effects of heat treatment on stiffening have been
previously reported [60, 112].
It is known that the apparent stiffness of these materials varies with applied stress level, tending to
be higher under compressive loading, which closes the microcracks. During cantilever bending, the
detached top coat is placed both under tension and compression. In the results presented here, this
effect is neglected, but it may be noted that the resultant error in stiffness is probably [344] of the
order of 5 GPa. This is a significant uncertainty, but the observed trend of increasing stiffness
after heat treatment is nevertheless quite clear. No significant difference is observed for coatings of
different compositions. However, due to the errors involved, the differences in the sintering and
stiffening behaviour between top coats of different compositions can not be precisely evaluated
with this technique.
Chapter 7: Thermomechanical Behaviour of PS TBCs

138
7.1.2 Nanoindentation
7.1.2.1 As-sprayed and After Isothermal Heat Treatment
Nanoindentation measurements were performed on as-sprayed detached top coats of YSZ (204NS-1)
in the through thickness direction. The indents were located in positions remote from visible cracks
in the polished surface. The results obtained give an average stiffness value of 130 20 GPa. It is
expected that the value of Youngs Modulus given by indentation would approach the value of the
monolithic material [61] since the region undergoing deformation is small and thus expected to be
free of significant defects. The slightly lower values measured are possibly due to some
accommodation of the indent by intersplat porosity or cracks. In other techniques, such as cantilever
bending (see section 7.1.1), the volume of material tested is more representative of the material as a
whole. Similar values for the Youngs modulus by nanoindentation of PS YSZ top coats have been
found in-plane (144 GPa) [109].
After heat treatment at 1300
o
C for 100 h, the stiffness was found to be 205 10 GPa, which is close
to the value for bulk zirconia. Similar values have also been found in the in-plane direction [109].
This suggests that sintering has caused healing of microcracks and interfaces between splats and
thus, the effect of these features in the measurement is reduced.

0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0.2
0 0.2 0.4 0.6
Displacement (mm)
F
o
r
c
e

(
N
)
Load
Unload

0
10
20
30
40
50
60
70
80
alumina-doped
YSZ (AE8170)
DSZ
(SPM6-2444)
YSZ
(204NS)
YSZ
(204NS-1)

Y
o
u
n
g
'
s

M
o
d
u
l
u
s

E

(
G
P
a
)
As sprayed Heat treat for 88 h at 1350C

Figure 7.1. Typical load/unload plot for as as-sprayed
YPSZ top coat during cantilever bend testing.
Figure 7.2. Youngs modulus of zirconia-based detached top
coats as sprayed and after 88 h at 1350
o
C.

Chapter 7: Thermomechanical Behaviour of PS TBCs

139
7.1.2.2 Heat Treatment Under a Thermal gradient
Measurements were made on a top coat that had been heat treated for 17 h with a high through-
thickness thermal gradient. Data are shown in Figure 7.3. A through-thickness gradient in the
stiffness is apparent. Near the TC/BC interface, values were similar to those obtained with the as-
sprayed top coat, but near the outer surface the average value increases, approaching that of
monolithic dense YSZ. There is, of course, quite a lot of scatter in the data, since the value
obtained is sensitive to the presence or absence of neighbouring fine scale flaws. These results are
consistent with the microstructural observations, indicating pronounced sintering near the outer
surface. The absolute values of Youngs modulus obtained using this technique are, as expected,
much higher than those given by methods that measure the global stiffness, since gross flaws affect
the latter, but not the former. Also, the indentation method senses the compressive stiffness only,
which is known to be larger than that under tension in these materials, particularly at relatively high
strains.
0
50
100
150
200
250
0 0.5 1 1.5 2 2.5 3
S
t
i
f
f
n
e
s
s

(
G
P
a
)
Distance from TC/BC interface (mm)

Figure 7.3. Stiffness data obtained by nanoindentation on a polished transverse section of a top coat
after heat treatment in thermal gradient (~900C at the interface, ~1500C at the free surface) for 17 h.

7.2 Microstructure and Properties of Thick As-sprayed TBCs
Thick TBCs provide greater temperature drops across the coating, which offers advantages with
respect to reduction of the cooling air required by hot components or allows increases in gas turbine
inlet temperature, thus increasing the efficiency and performance of the engine. However thick
Chapter 7: Thermomechanical Behaviour of PS TBCs

140
TBCs show low thermal shock resistance [345]. With increased thickness, the elastic strain energy
stored in the coating increases and hence the driving force for spallation increases. Microstructural
modifications that induce microcracks and porosity increase the thermal shock resistance [68, 346].
7.2.1 The Effect of Substrate Temperature on Microstructure and Stiffness
Thick YSZ top coats (about 1.5 mm) were produced by PS. Coatings sprayed at four different
average substrate temperatures were examined. The substrate temperature was controlled by
cooling the back surface of the substrate during spraying with a jet of compressed air and/or by
waiting 35 s in-between each gun pass. The average backsurface temperature during spraying for
the different spraying conditions is shown in Figure 7.8. The temperature of the substrate changes
continuously during spraying. The error bars in Figure 7.8 indicate the maximum and minimum
temperatures of the backsurface of the substrate during spraying. The four samples produced have
been labelled as A, B, C and D with decreasing average substrate temperature. The sprayed
specimens were vacuum impregnated with epoxy, ground and finely polished. Cross sections of the
samples were examined by optical microscopy, and fracture cross sections were also prepared for
microstructure characterization. The polished cross sections are presented in Figure 7.4 to Figure
7.7. The microstructural features observed can be defined as follows: (i) segmentation cracks
(cracks running perpendicular to the coating surface and penetrating at least half the coating
thickness), (ii) branching cracks (cracks parallel to the coating plane, starting from segmentation
cracks), (iii) microcracks (other cracks in the coating not fulfilling the criteria of segmentation or
branching cracks) (iv) porosity (spherical porosity in the cross section of the coating). The
segmentation crack density was calculated by dividing the number of segmentation cracks present
in each cross section with the cross section length.
Coating A, sprayed with the highest average surface temperature, had the highest levels of
segmentation cracks (1.1 mm
-1
), and significant branching cracks originating from them. Coating D,
with the lowest average back surface temperature, produced by cooling the substrate and waiting in-
between gun cycles, had a segmentation crack density of 0.7 mm
-1
. Also, its worth noting that in
this coating no branching cracks were present. Coatings B and C had a similar average backsurface
temperature. However, the difference between the minimum and maximum temperatures in coating
B was about 100
o
C more than in coating C. In coating B, the crack segmentation density is
relatively low (0.72 mm
-1
) and branching cracks are present. Coating C has a segmentation crack
density of 0.91 mm
-1
and very few branching cracks are observed
Chapter 7: Thermomechanical Behaviour of PS TBCs

141


Figure 7.4. Polished cross section of as-sprayed YSZ
top coat sprayed without cooling and without waiting in-
between each gun cycle. [average backsurface
substrate temperature : 507
o
C].
Figure 7.5. Polished cross section of as-sprayed YSZ
top coat sprayed with 35 sec wait in-between each gun
cycle. [average backsurface substrate temperature :
445
o
C].


Figure 7.6 Polished cross section of as-sprayed YSZ
top coat sprayed with air cooling at the back substrate
surface. [average backsurface substrate temperature :
429
o
C].
Figure 7.7. Polished cross section of as-sprayed YSZ
top coat sprayed with air cooling at the substrate back
surface and with 35 sec. wait in-between each gun
cycle. [average backsurface substrate temperature :
212
o
C].
These observations suggest that higher average substrate temperatures and lower differences
between the maximum and minimum substrate temperature give rise to coatings with higher density
of segmentation cracks. Segmentation cracks are generated when the sample is cooled to room
temperature, after the deposition process is complete, as a result of relaxation of bi-axial stresses
[347]. The segmentation crack initiates at the free surface and propagates through the thickness of
the coating. When it reaches the weak interfaces that exist in-between layers due to poor splat
bonding, the stresses at the crack tips are relieved by opening up these cracks. At high average
Chapter 7: Thermomechanical Behaviour of PS TBCs

142
substrate temperature and low differences between the maximum and minimum substrate
temperature, there will be improved contact between splats [348-350], and stresses in the top coat
cannot relax by generation of cracks in-between splats or microcracking. Thus relaxation occurs by
the generation of segmentation cracks. When the difference between the minimum and maximum
spraying temperature is high, even for a high average substrate temperature, bonding between splats
is poor and the segmentation crack density is relatively low. Similar results have been observed by
other researchers [68, 346]. Coatings with higher densities of segmentation cracks density perform
better under thermal cycling [68, 346].
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
0 1 2 3 4 5
A
v
e
r
a
g
e

B
a
c
k
s
u
r
f
a
c
e

T
e
m
p
e
r
a
t
u
r
e

(
o
C
)
continuous
spraying
A
waiting in-
between gun
cycles
B
continuous
spraying with
substrate
cooling
C
waiting in-
between cycles
with substrate
cooling
D

Figure 7.8 Average substrate back-surface temperature during spraying for different
cooling conditions. The error bars indicate the temperature range of the substrate during
spraying.
The numerical model described in section 3.1.3 and developed by Clyne and co-workers [95, 99,
351-353] was used to calculate the residual stresses in the top coat and confirm the segmentation
and branching crack initiation process. The measured thermal history was compared with that
obtained using the numerical model. Comparisons of experimental and modelled backsurface
thermal histories for each of the different spraying conditions are shown in Figure 7.9 to Figure 7.12.
The agreement for all cases is good. The residual stress distribution through the coating was then
obtained from the numerical model. Figure 7.14 shows the residual stress distribution through the
thickness of the coating produced by waiting in-between each gun cycle. There is a stress gradient
through the thickness of the coating, with compressive stresses at the TC/BC interface and tensile
stresses in the outer surface. The results from the model confirm that near the free surface of the
Chapter 7: Thermomechanical Behaviour of PS TBCs

143
top coat there are high tensile stresses that probably lead to initiation of the segmentation cracks.
Branching cracks initiate from the segmentation cracks at places of weak splat bonding and where
stresses are tensile (see Figure 7.5).
0
100
200
300
400
500
600
700
800
0 50 100 150 200 250
Time (sec)
T
e
m
p
e
r
a
t
u
r
e

(
o
C
)
experiment
model

0
100
200
300
400
500
600
700
800
0 100 200 300 400 500
Time (sec)
T
e
m
p
e
r
a
t
u
r
e


(
o
C
)
experiment
model

Figure 7.9 Experimental and modelled back surface
thermal histories for YPSZ-TC sprayed without cooling
and without waiting in-between each gun cycle.
[average backsurface substrate temperature : 507
o
C]
Figure 7.10 Experimental and modelled back surface
thermal histories for YPSZ-TC sprayed with 35 sec wait
in-between each gun cycle. [average backsurface
substrate temperature : 445
o
C]
0
100
200
300
400
500
600
700
0 50 100 150 200
Time (sec)
T
e
m
p
e
r
a
t
u
r
e

( o
C
)
experiment
model

0
100
200
300
400
500
600
700
800
0 100 200 300 400
Time (sec)
T
e
m
p
e
r
a
t
u
r
e

(
o
C
)
experiment
model

Figure 7.11 Experimental and modelled back surface
thermal histories for YPSZ-TC sprayed with air cooling
at the back substrate surface. [average backsurface
substrate temperature : 429
o
C].
Figure 7.12 Experimental and modelled back surface
thermal histories for YPSZ-TC sprayed with air cooling
at the substrate back surface and with 35 sec. wait in-
between each gun cycle. [average backsurface
substrate temperature : 212
o
C].

Chapter 7: Thermomechanical Behaviour of PS TBCs

144
Segmentation and branching cracks are characteristic of thick TBCs and are not present in coatings
less than about 1 mm thick. The average stress state in the top coat after spraying coatings of
different thickness was estimated using the numerical model by matching the measured and
modelled thermal histories. The results show that the average residual stress in the top coat changes
from compressive to tensile with increasing coating thickness (see Figure 7.13). Stresses in the
outer surface of thicker coatings are higher (more tensile) than thinner coatings. Even when the
average stress in the top coat is compressive, the stresses near the free surface of thick coatings are
tensile. For thinner coatings the stresses near the free surface are possibly not high enough for
initiation of the segmentation cracks.
-16
-14
-12
-10
-8
-6
-4
-2
0
2
4
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
Coating thickness (mm)
A
v
e
r
a
g
e

S
t
r
e
s
s

i
n

C
o
a
t
i
n
g

(
M
P
a
)
)


Figure 7.13 Averages residual stress in APS-YPSZ top
coats of different thickness.
Figure 7.14 Residual stress through thickness of as-
sprayed YSZ top coat sprayed by waiting in-between
each gun cycle.

The Youngs modulus of coating A, B, C, and D was determined by cantilever bending, with the
free surface of the coating in tension. The results are presented in Figure 7.15 as a function of
segmentation crack density. The Youngs moduli are in the range of 4.5-10 GPa. Segmentation
cracks are expected to have a strong influence on the Youngs Modulus measured by this technique.
Coating A, with the highest segmentation crack density, has the lowest Youngs modulus. Due to
the fact that segmentation cracks do not always penetrate entirely the thickness of the coating, lower
values for the stiffness of the coating are expected when the free surface is loaded under tension.
-60
-40
-20
0
20
40
60
80
1 1.2 1.4 1.6 1.8 2 2.2
Distance (mm)
S
t
r
e
s
s

(
M
P
a
)
maximum stress
(segmentation
crack initiation)
TC/BC interface
initiation of
branching cracks
Chapter 7: Thermomechanical Behaviour of PS TBCs

145
A
C
D
B
0
2
4
6
8
10
12
14
0.6 0.7 0.8 0.9 1 1.1 1.2
Segmentation crack density (mm
-1
)
Y
o
u
n
g
'
s


m
o
d
u
l
u
s

(
G
P
a
)

-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
3
100 200 300 400 500 600
Average Substrate Temperature (
o
C)
A
v
e
r
a
g
e

S
t
r
e
s
s

i
n

C
o
a
t
i
n
g

(
M
P
a
)

Figure 7.15. Influence of segmentation crack density on
Youngs modulus of thick PS TBCs. Error bars indicate the
standard deviation and the letters A, B, C, D refer to the
corresponding coatings sprayed with conditions specified in
the text.
Figure 7.16 Average residual stress in APS YPSZ top
coat against average temperature at the substrate back
surface during spraying.
7.2.2 The Effect of Substrate Temperature on Residual Stresses
Figure 7.16 shows that the average TC residual stress changes from tensile to compressive with
increasing average substrate temperature. There is a stress gradient through the thickness of the TC.
It can be seen in Figure 7.14 that stresses at the outer surface are tensile, with higher stresses
(reaching 68 MPa) for the lower substrate temperature. At the TC/BC interface, stresses are
compressive. Coatings deposited at higher temperatures showed the highest compressive stresses at
the interface. With increasing substrate temperature, stresses become more compressive. The
reason for this is that, as the substrate temperature is increased, the stresses due to differential
thermal contraction generated on cooling increase, producing greater compressive residual stresses
in the top coat. Similar results on the effect of substrate temperature on the residual stress state of
the top coat have been reported [77, 348, 354, 355].
7.3 Stresses in TBCs after heat treatment
The model described in section 3.1.3 and developed by Clyne and co-workers [95, 99, 351-353] has
been used in the present work for prediction of residual stresses after spraying and subsequent heat
treatments. The predicted effect of imposing a high through-thickness thermal gradient is shown
Figure 7.17b. It can be seen that the stress level in the coating is moved towards more compressive
Chapter 7: Thermomechanical Behaviour of PS TBCs

146
values when the thermal gradient is imposed, particularly near the free surface, where also the
temperature is the highest. This is expected to accelerate the rates of sintering and stiffening.
Figure 7.17a shows the predicted effect on the residual stresses of heat treatment isothermally at
1100
o
C. In this case, the stress in the top coat is approximately uniform and tensile and it is
expected that sintering processes will be retarded by the presence of tensile stresses which hold
open the microcracks and pores [60]. The difference in the stress state when heat treated under a
thermal gradient and heat treated isothermally is quite significant. This must be considered if
realistic conclusions about the thermomechanical behaviour of TBCs are to be drawn.
-10
-5
0
5
10
15
20
25
30
0
200
400
600
800
1000
1200
1400
1600
0 10 20 30 40 50
Distance from base (mm)
stress
temperature
S
t
r
e
s
s

(
M
P
a
)
T
e
m
p
e
r
a
t
u
r
e

(
o

C
)
SubstrateCoating
(a)

-200
-100
0
100
200
300
0
200
400
600
800
1000
1200
1400
1600
0 10 20 30 40 50
stress
temperature
S
t
r
e
s
s

(
M
P
a
)
T
e
m
p
e
r
a
t
u
r
e

(
o
C
)
Distance from base (mm)
Substrate Coating
(b)

Figure 7.17. Predicted through-thickness distributions of stress and temperature in a
TBC system (a) when isothermal at 1100C and (b) when subjected to a high thermal
gradient, similar to that employed in the experimental work.

Chapter 7: Thermomechanical Behaviour of PS TBCs

147
7.4 Conclusions
The following conclusions can be drawn from this work.
1. The Youngs modulus of as-sprayed zirconia-based top coats, measured by cantilever bending,
is in the range of 10-15 GPa. After heat treatment the elastic modulus increases substantially
due to sintering. Due to the errors involved, the differences in the sintering and stiffening
behaviour between top coats of different compositions cannot be precisely evaluated with this
technique. The observed trend of increasing stiffness after heat treatment is nevertheless quite
clear. The Youngs moduli of YSZ top coats measured by indentation approach the value of the
monolithic material and are not representative of the global stiffness of the coating due to the
small volume of material undergoing deformation.
2. A through thickness gradient in the stiffness of the top coat is obtained when the coating is heat
treated in a thermal gradient. This is because sintering progresses more rapidly near the free
surface where the temperature is highest and the stresses are more compressive.
3. Thick thermal barrier coatings exhibit microstructural features, such as segmentation and
branching cracks, which affect the elastic properties and which are not present in thinner TBCs.
These cracks are generated as a result of relaxation of bi-axial tensile residual stresses near the
free surface of thick TBCs.
4. Crack segmentation density in thick thermal barrier coatings is strongly dependant on the
substrate temperature during deposition. Higher average substrate deposition temperatures,
combined with low variations between the minimum and maximum substrate temperature, give
higher crack segmentation densities. Coatings with higher crack segmentation densities have a
lower Youngs modulus.
5. The average top coat residual stress changes from tensile to compressive with increasing
average substrate temperature.
6. The high thermal gradients commonly present in TBCs under service conditions lead to widely
varying conditions within the top coat. Sintering is likely to occur much more rapidly near the
free surface, not only because the temperatures are higher there, but because the residual
stresses tend to be more compressive. It is known that tensile stresses retard the progression of
sintering.

Chapter 8: Thermal Conduction in PS TBCs

148

8 Thermal Conduction in Plasma-Sprayed TBCs
8.1 Thermal Conductivity of Plasma sprayed TBCs
8.1.1 Steady state Rig
Figure 8.1 and Figure 8.2 show experimental data, and corresponding best linear fits from eqn 5.19,
for samples tested using both silicone-based thermal compound and conductive pads. By
employing a thermal compound, k
true
and h were found to be 1.38 0.27 W m
1
K
1
and
16.7 kW m
2
K
1
, respectively (Figure 8.1). Using conductive pads, on the other hand, measured
k
true
and h values were relatively lower, i.e. 1.1 0.10 W m
1
K
1
and 3.77 kW m
2
K
1
, respectively
(Figure 8.2). The silicone-based compound appears to be more effective than conductive pads in
raising the interfacial thermal conductance (h), by filling more gaps between the matching surfaces.
However, the use of such compound also raises the true conductivity, by as much as ~30%. With
porous materials, there is a serious danger of liquids or soft solids entering the pores during
specimen preparation or during measurement, which will tend to raise the measured conductivity.
Therefore, the use of thermal compounds should be avoided for coatings containing relatively large
surface-connected pores, such as those typically found in sprayed TBCs.
It is worth pointing out that some thermophysical properties of PS coatings are often sensitive to
powder composition and spraying parameters. Nevertheless, the measured value of thermal
conductivity of detached YSZ coatings (1.1 0.10 W m
1
K
1
) is in relatively good agreement with
literature data for PS coatings of similar composition and porosity (see Figure 4.6).
It can be seen that both cases exhibited larger scatter than the fused silica samples that were used to
validate the technique - see figure 5.18. The detached YPSZ coatings were found to be slightly
curved, as a result of residual stresses, making preparation of samples with flat surfaces difficult.
Also, the inherent variability in the microstructure of such material is much greater than that of
fused silica. Thin conductive pads are recommended for measuring the conductivity of YPSZ. Its
application does not present the danger of soft solids (e.g. silicone compound) infiltrating the larger
pores and altering the conductivity. Also, conductive pads are sufficiently compliant to
accommodate the slightly curved surface of detached coatings, hence raising the interfacial thermal
conductance.

Chapter 8: Thermal Conduction in PS TBCs

149
0.0
0.5
1.0
1.5
2.0
0.0 0.5 1.0 1.5 2.0 2.5
YPSZ coating
(with silicone thermal compound)
A
T
/
Q

(
K

m
-
2

k
W
-
1
)
Ax (mm)
k
true
= 1.380.27 W m
-1
K
-1
h = 16.7 kW m
-2
K
-1
1/k
true
= 0.7235
2/h = 0.12

Figure 8.1. Plot of ATQ against specimen thickness for detached as-sprayed YSZ top coats, obtained using
the conductive compound. Deduced values of k
true
and h are shown.

0.0
0.5
1.0
1.5
2.0
2.5
0.0 0.5 1.0 1.5 2.0 2.5
YPSZ coating
(with conductive pads)
A
T
/
Q

(
K

m
-
2

k
W
-
1
)
Ax (mm)
k
true
= 1.10.1 W m
-1
K
-1
h = 3.77 kW m
-2
K
-1
1/k
true
= 0.9083
2/h = 0.53

Figure 8.2. Plot of AT/Q against specimen thickness for detached as-sprayed YSZ top coats, obtained using
the conductive pads. Deduced values of k
true
and h are shown.

Chapter 8: Thermal Conduction in PS TBCs

150
8.1.2 Laser Flash Measurements
The thermal conductivity of detached as-sprayed YSZ (204NS) and DSZ top coats was measured
using laser flash technique. Details on the experimental procedure are given in section 5.6.3. The
theoretical background of this technique is given in 4.2.2.3. Measured thermal diffusivity data for
detached YSZ (204NS) top coats are presented in Figure 8.3 together with measured specific heat
data and deduced thermal conductivity. The density of the YSZ detached top coat was measured by
the Archimedes method (see section 5.3.5) to be about 5.35 g cm
-3
which corresponds to about 11%
porosity, (typical of as-sprayed coatings -see section 6.4.5).

Figure 8.3. Thermal conductivity, thermal diffusivity and specific heat measurements from 25-1400
o
C of
detached YSZ as-sprayed top coat (204NS) The thermal diffusivity was measured by the laser flash
technique.

The measured thermal conductivity decreases slightly with increasing temperature, up to about
1100
o
C, indicating a phonon conduction mechanism (see section 4.1.1) and then increases slowly
above about 1100
o
C. The increase at higher temperature is probably due to radiative transport (see
section 4.1.2). At 100
o
C the thermal conductivity is 1.13 W m
-1
K
-1
which is close to the value
measured using the steady state bi-substrate technique (1.1 W m
-1
K
-1
). The data for DSZ follow a
similar trend (Figure 8.4). The porosity of the DSZ as-sprayed sample was also about 11 %
(density ~ 5.8 g cm
-3
). The measured thermal conductivity at 100
o
C is about 1.38 W m
-1
K
-1
. The
Chapter 8: Thermal Conduction in PS TBCs

151
error expected in the values measured here is about 10%. The thermal conductivity of as sprayed
DSZ is slightly greater than YSZ (204NS). This will be discussed in more detail in section 8.2.

Figure 8.4. Laser flash data for the thermal diffusivity of detached DSZ as-sprayed top coats (SPM6-2444)
together with measured specific heat data and deduced thermal conductivity.

Laser flash has been extensively used for the measurement of the thermal diffusivity of the thermal
barrier coatings because of its rapid measurement and high temperature capabilities- see sections
4.2.2.3 and 4.3. It is important to consider several factors that could be significant in the
determination of the thermal diffusivity of TBCs. One of these factors is the gas content and gas
pressure of the pores in the TBC during the measurement, since this will affect the measured value
of the thermal diffusivity. In addition, TBCs are translucent in the infrared region and hence the
energy pulse is not absorbed at the surface of the coating. To overcome this problem the surface of
translucent specimens is routinely treated for laser flash measurements by applying an opaque
coating of graphite, Au or Pt. However, high temperatures can cause partial evaporation of these
layers and possible reaction with the sample is another cause of concern [356]. Also, since the laser
flash method measures the thermal diffusivity of the sample, calculation of conductivity requires
prior knowledge of the volume specific heat at different temperatures. Since TBCs show large
variations in porosity and crystallite size [14, 357], it can be difficult to measure the volume specific
heat accurately. Another problem lies in determining suitable laser pulse power and test duration
Chapter 8: Thermal Conduction in PS TBCs

152
values [358]. In general, these are inherent drawbacks with all transient methods. The reliability of
laser flash thermal diffusivity measurements for thermal barrier coatings was addressed by Wang at
el [358]. Their work suggests that due to the inherent low thermal conductivity and low sample
thickness of TBC, large temperatures can develop at the front and rear surface of the sample. This
could introduce errors in the temperature measurement at the rear surface due to non-linear
temperature-voltage response of IR detectors, and yield lower values for the thermal diffusivity of
the TBCs.
8.1.3 Hot Disk
The thermal conductivity of as-sprayed detached YSZ (204NS) top coats was measured using the
hot disk method. For the theoretical background of the technique see section 4.2.2.2. and for the
experimental procedure see section 5.5.2. The thermal conductivity was estimated to be
1.2 0.1 W m
-1
K
-1
and the thermal diffusivity 0.48 mm s
-1
, which is in reasonable agreement with
the values measured by the bi-substrate steady state technique and laser flash. Figure 8.5 is a
comparative plot of the thermal conductivity values of as-sprayed YSZ top coats measured with the
techniques mentioned above. The error bars indicate the errors associated with the technique (for
laser flash and hot disk techniques) or the experimental scattering (for steady state technique).
Most of the techniques currently employed for coatings are either steady-state [258, 263, 359] or
laser flash [358, 360] (transient) methods- see section 4.3. However, measured values obtained
using different techniques often show significant variations. This is largely because the
assumptions employed are not strictly valid [361]. Laser flash is often considered to be the most
convenient and accurate method over a wide range of temperatures. It does not require large
samples and can be used for remote measurement, both under vacuum and in gas atmospheres.
However, there are several issues surrounding the application of this method to thermal barrier
coatings (TBCs) [358, 360]. The Hot Disk method has not been extensively used for the
measurement of thermal conductivity of TBC coatings. This technique is very sensitive to sample
dimensions and two identical samples of the same dimensions are needed for a single measurement.
In addition, there are problems in determining suitable power and test duration values as well as a
time window for analysis -see section 5.5.2 and 4.2.2.2. On the other hand, in the steady-state
method presented here, only the temperature drop across the sample and the heat flux are required
to calculate the thermal conductivity. For relatively low temperatures (from room temperature to
about 500C), the steady state method presented is a simple set up which can reliably be used to
Chapter 8: Thermal Conduction in PS TBCs

153
measure the thermal conductivity of TBCs and ceramic coatings of different thicknesses and
porosity levels.

Figure 8.5. Thermal conductivity of YSZ (204NS) top coats
measured by different techniques.
8.2 Effect of Powder Composition on Thermal Conduction
The thermal conductivity of DSZ top coats has been found to be higher than that of YSZ, while that
of YLaSZ is lower than YSZ (see Figure 8.6 and Figure 8.7). The thermal conductivity of
stabilized zirconia is believed to be dominated by oxygen vacancy scattering [262, 272]-see section
4.3. Scattering by point defects, such as vacancies and foreign ions, increases with defect
concentration and difference in the mass between the host ion and the defect (see equation 4.3).
Larger misfits between the ionic radii of the host and dopant ions are also expected to introduce
lattice strain which will promote phonon scattering [18].
In a DSZ top coat the dopant ion (Dy
3+
)

has a high atomic weight (162.5) and the ionic radius is
0.103 nm, which is larger than the ionic radius of the host ion Zr
4+
(0.084 nm) and about the same
as Y
3+
(0.102 nm) (see table Table 8.1). However, the DSZ top coat contains 3.58 mol% of dopant,
whereas YSZ contains 4.3 mol% of dopant, and YLaSZ has a total of 6.25 mol% (2.6 mol% La
2
O
3

and 3.65 mol% Y
2
O
3
). The thermal conductivity of YSZ, is slightly lower than the thermal
conductivity of DSZ which has a dopant ion with a large atomic weight and slightly less stabilizer.
In the YLaSZ while the dopants have lower atomic mass than Dy
3+
(atomic weight of La
3+
is 138.9
and Y
3+
88.9) the total amount of stabilizer is greater and hence the level of oxygen vacancies is
higher, which promotes scattering of phonons. Additionally, the ionic radii misfit between the Zr
4+

host ion (0.084 nm), the La
3+
(0.116 nm) and Y
3+
(0.102 nm) ions create lattice strains, again
promoting phonon scattering. This effect is expected to be more pronounced for the La
3+
ion, since
Chapter 8: Thermal Conduction in PS TBCs

154
it has the largest misfit. The lower thermal conductivity of YLaSZ and YSZ top coats confirms that
oxygen vacancies play a dominant role in phonon scattering and hence in reducing the thermal
conductivity. Similar effects of La
2
O
3
additions have been found by other researchers [274].
It must be pointed out that, even though the coatings have similar porosity (about 10%), there might
be differences in the microstructure, such as many interfaces with poor bonding between splats,
with little effect on the total porosity level, but a strong effect on thermal conductivity.
Material Ion/
Stabilzing
Ion
Ionic
radius
(nm)
Ionic
mass
Valence
difference
from Zr
4+
Atomic
concentration
in powder
(mol%)
Measured k
at 100
o
C
(W m
-1
K
-1
)
ZrO
2
Zr
4+
0.084 91.2 0 -
YSZ Y
3+
0.102 88.9 -1 4.3 1.1
DSZ Dy
3+
0.103 162.5 -1 3.58 1.29
YLaSZ La
3+
/
Y
3+

0.116 /
0.102
138.9 /
88.9
-1 /
-1
2.6 /
3.65
0.68
Table 8.1 Ionic radius, ionic mass, valance and atomic concentration of stabilizing ion in YSZ, DSZ
and YLaSZ top coats together with measured values of thermal conductivity at 100
o
C by the bi-
substrate steady state technique.

(a)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
0 2 4 6 8 10 12
YSZ
DSZ
YLaSZ
T
h
e
r
m
a
l

c
o
n
d
u
c
t
i
v
i
t
y

(
W

m
-
1

K

-
1
)
Time (hours)
1400
o
C

(b)
0.0
0.4
0.8
1.2
1.6
2.0
2.4
0 2 4 6 8 10 12
YSZ
DSZ
T
h
e
r
m
a
l

c
o
n
d
u
c
t
i
v
i
t
y

(
W

m
-
1

K

-
1
)
Time (hours)
1200
o
C

Figure 8.6. Thermal conductivity of detached YSZ, DSZ and YLaSZ top coats at 100
o
C as a function of heat treatment time
(a) at 1400
o
C. and (b) at 1200
o
C. Measurements where performed using the steady state technique.
Chapter 8: Thermal Conduction in PS TBCs

155
(a)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
0 2 4 6 8 10 12
YSZ
DSZ
T
h
e
r
m
a
l

c
o
n
d
u
c
t
i
v
i
t
y

(
W

m
-
1

K

-
1
)
Time (hours)
1400
o
C

(b)
0.0
0.4
0.8
1.2
1.6
2.0
2.4
0 2 4 6 8 10 12
YSZ
DSZ
T
h
e
r
m
a
l

c
o
n
d
u
c
t
i
v
i
t
y

(
W

m
-
1

K

-
1
)
Time (hours)
1200
o
C

Figure 8.7. Thermal conductivity of detached YSZ and DSZ top coats at 100
o
C as a function of heat treatment time. (a) at
1400
o
C and (b) at 1200
o
C. Measurements were performed using the laser flash technique.

8.3 The Effect of Heat Treatment on the Thermal conductivity of TBCs
The thermal conductivity of detached plasma sprayed top coats after isothermal heat treatment at
1200
o
C and 1400
o
C was measured using the steady state and the laser-flash techniques. Figure 8.6
shows the evolution of the thermal conductivity for YSZ, DSZ and YLaSZ top coats after heat
treatment at 1200
o
C and 1400
o
C measured by the steady state technique. For each time and
temperature, at least two samples of different thickness were used for the measurement. Conductive
pads between the sample and substrate were used in order to minimize the contact resistance. For
calculating the thermal conductivity, the interface conductance h was assumed to be 3.77 kW m
-2
K
-
1
, as previously calculated for as-sprayed YSZ top coats of different thicknesses (see section 8.1.1).
The surface finish in all samples was the same and the same pressure was applied, so the
assumption that the interfacial conductance was similar should be valid.
The thermal conductivity increases after heat treatment, which is more pronounced for the sample
heat treated at 1400
o
C. Very little increase has occurred after heat treatment at 1200
o
C. This
increase is due to sintering of the top coat which causes the disappearance of small pores and
interphases between splats as well as grain growth. Small pores, splat interphases and grain
boundaries inhibit heat transfer and cause scattering of phonon and photons. Therefore the increase
of the thermal conductivity observed is due to sintering. At 1200
o
C not much sintering takes place
Chapter 8: Thermal Conduction in PS TBCs

156
and hence a smaller increase in the thermal conductivity is measured. Similar trends have been
observed by other researchers [112, 254, 257, 259, 262, 264, 265](see section 4.3). Increase in the
thermal conductivity is due to sintering of cracks, pores and interfaces during heat treatment. In PS
coatings, poor contact between lamellae in close physical proximity bond together assisted by
diffusion processes at high temperatures. It should be pointed out that even though no significant
increase in the thermal conductivity is observed after 10

h at 1200
o
C, there are significant increases
in the Youngs modulus at this temperature and heat treatment times [362].
The thermal conductivity of the detached YSZ (204NS) and DSZ (SPM6-2444) coatings after
isothermal heat treatment was also measured by the laser flash technique. Results are presented in
Figure 8.7. Both techniques confirm similar trends: there is an increase in the thermal conductivity
after heat treatment for both coatings and the increase is more pronounced for the coating that has
been heat-treated at 1400
o
C. The thermal conductivity of the heat treated coatings follows the same
trend as the as-sprayed coatings with temperature; firstly decreases and then at above 1100
o
C it
slightly increases due to the effect of radiation. Both the bi-substrate and laser flash techniques give
comparable results and trends.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
0 2 4 6 8 10 12
Heat Treatment Time (hours)
N
o
r
m
i
l
i
s
e
d

T
h
e
r
m
a
l

C
o
n
d
u
c
t
i
v
i
t
y

k
/
k
o
YSZ
DSZ
YLaSZ

Figure 8.8. Relative change in the thermal conductivity of detached top coats after heat treatment at 1400
o
C
measured by the steady state technique. k/k
o
is defined as the thermal conductivity of coating after heat
treatment (k) over the thermal conductivity of the as-sprayed coating (k
o
).

The relative increase in the thermal conductivity after heat treatment at 1400
o
C is presented in
Figure 8.8. It is apparent than even though the initial thermal conductivity of YLaSZ is
considerably lower than those of YSZ and DSZ top coats, it quickly increases with heat treatment
due to sintering. This implies that the microstructural features, such as interfaces and poor bonding
Chapter 8: Thermal Conduction in PS TBCs

157
between splats that cause a reduction in thermal conductivity, are not stable and hence some of the
benefits of such features quickly disappear after a few hours of exposure to high temperature.
Nevertheless, the absolute value of the thermal conductivity of YLaSZ after 10 h at 1400
o
C is still
lower than that of YSZ and DSZ top coats. The relative increase of DSZ and YSZ top coats is very
similar, however, the initial thermal conductivity of DSZ is higher than YSZ.
8.4 Effect of Pore Conductivity on Overall Thermal Conductivity
The effect of pore conduction on the overall thermal conductivity of PS YSZ was explored by
experimental measurements in air and in vacuum. Measurements were performed at atmospheric
pressure in air with the steady state technique and compared to corresponding measurements in
vacuum. The thermal conductivity of YSZ plasma sprayed TBCs in vacuum was found to be ~15%
lower than that in air (Figure 8.9a). Similar results have been found by other researchers [363].
The analytical TFR model described in section 4.4.4 was used to interpret the results (Figure 8.9b).
For comparative purposes, some measured data available in literature on the thermal conductivity of
7-8 wt% YSZ with about 10% porosity are also plotted on the same graph. Open symbols indicate
measurements performed in vacuum and filled symbols measurements performed in air or another
gas at atmospheric pressure. For the model, the values used for the splat thickness (=L
v
-d
v
) was
1.9 m, the inter-splat pore thickness, (d
v
), 0.1 m and inter-bridge spacing, L
v
, 6 m (see section
4.4.3 for more detailed explanation for the values used). These values correspond well to the
microstructural feature observed in our PS top coat (see Figure 8.10). For this input data there is
good agreement between the measured and predicted effective thermal conductivity. It can be
concluded that thermal conduction through the pores affects the overall thermal conductivity of
TBCs and thus cannot be neglected.
In noting the effect of gas pressure, it may be recalled that in the hot end of a typical gas turbine
engine, the pressure varies between about 40 atmospheres and one atmosphere. Furthermore the gas
permeability of TBC top coats is known to be very high [21]. Therefore, such pressures will
quickly become established throughout most of the pores within such coatings. At atmospheric and
reduced pressure, convection in pores can be neglected, since the mean free path of air molecules is
comparable with the pore dimensions. The pore conductivity is low and temperature independent.
Higher gas pressures can significantly raise the pore conductivity, particularly at high temperature
(see Figure 8.9b). However experimental measurements of the thermal conductivity are normally
carried out either at one atmosphere or under reduced pressure, in which case the pore conductivity
Chapter 8: Thermal Conduction in PS TBCs

158
will be appreciably lower. Furthermore, measurements are often made at room temperature, which
will also have the effect of reducing the pore conductivity.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
air vacuum
T
h
e
r
m
a
l

C
o
n
d
u
c
t
i
v
i
t
y

(
W

m
-
1

K
-
1
)

0.5
1
1.5
2
2.5
500 1000 1500 2000
model (current work) in air 40atm
model (current work) in air 1atm
model (current work) in vacuum
measurement (from literature) in air
measurement (from literature) in vacuum
measurement (current work) in air
measurement (current work) in vacuum
T
h
e
r
m
a
l

C
o
n
d
u
c
t
i
v
i
t
y

(
W

m
-
1

K
-
1
)
Temperature (K)

Figure 8.9. (a) Thermal conductivity of YSZ top coat measured by the steady state technique in air and in vacuum
(b) Comparison of measured values of the thermal conductivity of 7-8wt%YSZ with prediction from the model
described in section 4.4. The literature values were taken from references [67, 112, 259, 264, 364, 365].

8.5 Experimental and Predicted Thermal Conductivity
Microstructural features from YSZ top coats and measured thermal conductivity by the steady state
method were compared with the predictions of the microstructure-based model described in section
4.4.4. There are difficulties in making reliable measurements of microstructural features such as
inter-splat bridge areas and spacings. These are often more clearly seen on fracture surfaces, such
as those shown in Figure 8.10, rather than on polished sections, but it is possible that the
microstructure could be affected by the fracturing process. The input parameters of the model were
inferred from SEM micrographs (Figure 8.10) of the YSZ top coats and densitometry.
For the particular input data employed, the TFR model gives predictions for the thermal
conductivity which are consistent with these measurements. While this procedure is clearly an
approximate one, it shows the potential of using mictrostructural parameters which can be identified
from various types of experimental measurement including inspection of micrographs, densitometry
and small angle neutron scattering (SANS), to predict the thermal conductivity of PS top coats.
Furthermore, the model can give an idea of how the sintering mechanisms generate microstructural
changes, which affect the coating conductivity.

Chapter 8: Thermal Conduction in PS TBCs

159



Figure 8.10. SEM micrographs of fracture surfaces of as-sprayed YSZ top coats showing typical microstructural
features.

In addition, the model predictions were also compared with the experimental data from literature as
well as other microstructure-based models for the thermal conductivity [288, 291] - see Figure 8.9b).
Unfortunately, while there are many reported thermal conductivity measurements for plasma
sprayed TBCs, there are certain problems in making comparisons with predictions from
microstructure-based models. In particular, the properties of plasma sprayed TBCs are often very
sensitive to powder composition and spraying conditions [137, 302], so it is important that both
thermal properties and microstructural features should be measured on the same specimens.
However, there have been a few studies in which systematic measurements have been made of both
conductivity and microstructural parameters. Data from one such study [119] are presented in
Figure 8.11. It can be seen that substitution of appropriate input data into the TFR model gives
predictions for the thermal conductivity which are consistent with these measurements. For these
particular input data, the predictions of the McPherson [288] model look rather unreliable, while
those of Lu et al [291] are closer to experiment and to the TFR model.

Chapter 8: Thermal Conduction in PS TBCs

160
0
0.4
0.8
1.2
1.6
2
2.4
0 0.2 0.4 0.6 0.8 1
TFR model (current work)
measured (Kulkarni et al)
measured (current work)
modified shear lag model (Lu et al)
contact resistance model (McPherson)
E
f
f
e
c
t
i
v
e

T
h
e
r
m
a
l

C
o
n
d
u
c
t
i
v
i
t
y
,

k
e
f
f

(
W

m
-
1

K
-
1
)
Normalised bridge area, S
br
/ S
tot
Splat thickness L
v
= 1.9 m
Bridge thickness d
v
= 0.1 m
Bridge spacing, L
h
= 6 m

0
0.5
1
1.5
2
0 0.2 0.4 0.6 0.8 1
TFR model (current work)
measured (current work)
modified shear lag model (Lu et al)
contact resistance model (McPhearson)
E
f
f
e
c
t
i
v
e

t
h
e
r
m
a
l

c
o
n
d
u
c
t
i
v
i
t
y
,

k
e
f
f

(
W

m
-
1

K
-
1
)
Normalised bridge area, S
br
/ S
tot
Splat thickness L
v
= 1.9 m
Bridge thickness d
v
= 0.1 m
Bridge spacing, L
h
= 6 m

Figure 8.11. Comparison between predicted conductivities obtained from TFR (see section 4.4.4.), contact resistance [288]
and the modified shear lag [291] models and experimental measurements [119] (a) in vacuum and (b) in air (1 atm).

8.6 Conclusions
The following conclusions can be drawn from this work.
1. A novel thermal conductivity measurement technique is described. The setup is simple and can
be applied to ceramic coatings of different thickness and porosity levels. The setup is designed
to generate steady one-dimensional heat flow through the coating. The heat flux is calculated as
the average through upper and lower substrates, taking into account their thermal conductivities,
including any temperature dependence. The method involves testing samples of different
thickness, so that a plot can be constructed of temperature drop over heat flux against specimen
thickness. This allows the true thermal conductivity of the sample, k
true
, and the interfacial
thermal conductance, h, to be evaluated. The technique has been validated using fused silica
samples of known thermal conductivity-see section 5.6.1.3, employing conductive paste at the
interface. The measured values were consistent with this figure, within the expected
experimental error.
2. The thermal conductivity of as-sprayed YSZ TBCs was measured using the bi-substrate
technique and was found to be 1.1 0.10 W m
-1
K
-1
, using conductive pads and specimens
about 0.5-2 mm in thickness. However, a higher value of 1.38 0.27 W m
-1
K
-1
was obtained
when the conductive paste was employed. This is attributed to entry of the paste into the
extensive surface-connected porosity which is known to be present in these TBCs. The use of
Chapter 8: Thermal Conduction in PS TBCs

161
such conductive pastes is not recommended when relatively large surface-connected pores are
known to be present. Laser flash and Hot disk techniques were also used to measure the thermal
conductivity of as-sprayed TBCs, with reasonably good agreement, within experimental error.
The drawbacks of each of the techniques were briefly discussed.
3. The thermal conductivity of YSZ and DSZ as-sprayed and heat treated top coats was measured
using laser flash technique up to 1400
o
C. The thermal conductivity decreases slightly with
increasing temperature, up to about 1100
o
C, indicating a phonon conduction mechanism. It then
increases slowly above about 1100
o
C. The increase at higher temperature is probably due to
radiative transport.
4. The thermal conductivity of DSZ top coats has been found to be higher than that of YSZ, while
that of YLaSZ is lower than YSZ. The results suggest that oxygen vacancies play a dominant
role in phonon scattering and hence in reducing the thermal conductivity, whereas phonon
scattering due to the mass difference between dopant and host ion has a weaker effect.
5. Heat treatment results in an increase of the thermal conductivity of YSZ, DSZ and YLaSZ
coatings due to sintering. The increase is more pronounced at 1400
o
C, whereas after 10 h
at1200
o
C hardly any increase is observed. The relative increase in the thermal conductivity is
higher for the YLaSZ top coats. This suggests that some microstructural features in YLaSZ
responsible for a reduction in thermal conductivity are not stable, and hence some of the
benefits of such features quickly disappear after a few hours at high temperature. Nevertheless,
the absolute value of the thermal conductivity of YLaSZ top coats after 10 h at 1400
o
C is still
lower than YSZ and DSZ top coats.
6. The effect of thermal conduction through the gas in the pores on the overall thermal
conductivity of TBCs was explored. The thermal conductivity of YSZ plasma sprayed TBCs in
vacuum was found to be ~15% lower than in air. Predictions from the TFR model show that
higher gas pressures in the pores, up to the range typical of gas turbine service conditions, can
significantly raise the pore conductivity, particularly at relatively high temperatures.
Conversely, the pore conductivity is predicted to be low under reduced pressure, which is
sometimes used for experimental measurements. Depending on temperature and microstructure,
there may be a significant effect on the overall conductivity.
7. Some limited comparisons have been made between experimentally measured conductivities
and predictions from the TFR model. These suggest that it should be possible to predict the
conductivity from measurable microstructural parameters.
Chapter 9: Conclusions and Future Work

162
9 Conclusions and Future Work
9.1 General Conclusions
In the present work plasma sprayed zirconia-based top coats were characterized with respect to their
sintering behaviour, microstructural features, phase constitution, thermal conductivity and residual
stress state.
Exposure to high temperatures induced phase changes in all the zirconia-based top coats studied,
and this may affect the microstructure and perhaps the stress state in TBCs. The presence of SiO
2

and Al
2
O
3
impurities appears to decrease the rate of decomposition. Dysprosia stabilizer, in the
amount present in the current study (3.58 mol%), successfully stabilized the tetragonal phase in as-
sprayed zirconia coatings. The oxidation state of cerium in ceria-stabilized-zirconia has a great
influence on the phase composition, and it is believed to affect the stress state in the coating.
The addition of La
2
O
3
significantly lowered the thermal conductivity. However La
2
O
3
coatings
decomposed after 10 h and the pyrochlore phase was formed. The formation of the pyrochlore
phase accompanies volume increase, which is thought to enhance the tetragonal-to-monoclinic
phase transformation above 1 mol% La
2
O
3.
Hence the amount of La
2
O
3
added must be tailored in
order to avoid degradation of the thermal barrier coating.
Due to sintering processes, detached top coats exhibit substantial reductions in linear dimensions
during exposure to at elevated temperatures. These linear contractions are greater in the through-
thickness direction than in the in-plane directions. Any contributions to the observed shrinkage
arising from phase changes are expected to be relatively small. Top coat compositon affects the
shrinkage rate of top coats, due to changes in the diffusion kinetics. The presence of SiO
2
and
Al
2
O
3
impurities enhances the sintering kinetics by providing a continuous low diffusivity pathway
at the grain boundary which will restrain grain boundary mobility, reduce grain growth and enhance
shrinkage. It may also do so by providing a high diffusivity pathway, e.g. liquid phase at the grain
boundaries, which will enhance shrinkage. Thermal cycling, commonly experienced by TBCs
under service conditions, affects their shrinkage. The shrinkage observed during thermal cycling is
less than that observed when heat treatment is continuous.
The high thermal gradients, commonly present in TBCs in-service, lead to widely varying
conditions within the top coat. Sintering is likely to occur much more rapidly near the free surface,
not only because the temperatures are higher there, but because the residual stresses tend to be more
compressive. The phase constitution near the bond coat interface is different from that near the free
Chapter 9: Conclusions and Future Work

163
surface. In addition, a through thickness gradient in the stiffness of the top coat is obtained when
the coating is heat treated in a thermal gradient.
The Youngs modulus of as-sprayed zirconia-based top coats, measured by cantilever bending, is in
the range of 10-15 GPa. After heat treatment, the elastic modulus increases substantially due to
sintering. Measurements by the nanoindentation technique gave Youngs Modulus values
approaching that of the monolithic material and are not representative of the global stiffness of the
coating due to the small volume of material undergoing deformation.
Thick thermal barrier coatings exhibit microstructural features such as segmentation and branching
cracks, which affect the elastic properties, and are not present in thinner TBCs. Crack segmentation
density is strongly dependent on the substrate temperature during deposition. Coatings with higher
crack segmentation densities have lower Youngs modulus. The presence of branching-cracks
could be significant when considering spallation of the top coat.
The thermal conductivity of as-sprayed YSZ TBCs was measured using a novel steady-state
thermal conductivity measurement technique and was found to be 1.1 0.10 W m
-1
K
-1
. Laser flash
and hot disk techniques were also used to measure the thermal conductivity of as-sprayed TBCs,
with reasonably good agreement.
The thermal conductivity of DSZ top coats has been found to be higher than that of YSZ, while that
of YLaSZ is lower than YSZ. The results suggest that oxygen vacancies play a dominant role in
phonon scattering and hence in reducing the thermal conductivity, whereas phonon scattering due to
mass difference between dopant and host ion has a weaker effect. Heat treatment results in an
increase of the thermal conductivity due to sintering. This increase is more pronounced at 1400
o
C,
whereas at 1200
o
C hardly any increase is observed. The relative increase in the thermal
conductivity is higher for the YLaSZ top coats.
Thermal conduction through the pores affects the overall thermal conductivity of TBCs. Thermal
conductivity of YSZ plasma sprayed TBCs in vacuum was found to be ~15% lower than in air.
Predictions from the TFR model show that higher gas pressures in the pores, up to the range typical
of gas turbine service conditions, can significantly raise the pore conductivity, particularly at
relatively high temperatures.
Comparisons between experimentally measured conductivities, and predictions from the TFR
model, suggest that it should be possible to predict the conductivity from measurable
microstructural parameters.

Chapter 9: Conclusions and Future Work

164
9.2 Future work
In light of the work carried out in this study, future work should focus on the optimization of the
performance of the top coat. Better understanding of the sintering mechanisms is required. For
example, the relative contributions of the different sintering mechanisms to the sintering behaviour
of the top coats and the factors that influence them must be better understood. To achieve this, the
diffusion kinetics of the top coat materials must be studied in more detail and independently of the
microstructural features of PS top coats. It is suggested that simpler powder geometries, e.g.
spheres, should be used to study the diffusion kinetics of top coat materials. Also, it has been
shown that the role of impurities in sintering kinetics is of considerable importance, and studies
need to be undertaken to explore this further, for new coatings with differing impurity levels.
It has been shown that sintering of the top coat causes healing of small defects within the
microstructure and leads to an increase in stiffness and thermal conductivity. In parallel with
understanding the diffusion kinetics independently of microstructure, the microstructural features of
PS top coats must be investigated quantitatively and their role in sintering must be explored. It is
suggested that techniques such as small angle neutron scattering, measurement of surface area,
mercury intrusion porosimetry, and helium pycnometry, can be used to characterize the
microstructure of PS top coats. In addition the effect, if any, of top coat composition and impurity
levels on these microstructural features as well as the evolution of the microstructure after heat
treatment should be explored. In parallel, characterization of the microstructural features, and
understanding of their role in sintering, can be used in order to optimize deposition processes, so
that coatings may be deposited with larger flaw sizes, such that locking up of the structure by
sintering of such defects cannot occur.
The mechanisms of generation of segregation and branching cracks in thick TBCs, and their
significance on the thermomechanical stability of the top coat must be investigated. The presence
of branching cracks could be significant with respect to the spallation of part of top coat as these
interfaces are weak. Study of their evolution during isothermal heat treatment and heat treatment
under a thermal gradient is of interest.
An immediate aim must be to continue the investigation of the effect of heat treatment of TBC
under a strong thermal gradient on the stiffness of the top coat. The presence of the observed
layered microstructure, with a sintered outer surface and a porous structure near the BC/TC
interface, affects the strain tolerance of the top coat. Different coating elements will experience
different thermal histories and thus attain different stiffnesses. A more detailed analysis of the
Chapter 9: Conclusions and Future Work

165
effects of top coat stress at high temperatures on the rate of sintering is required. Measurement of
corresponding changes in porosity is necessary.
The potential of using different top coat materials to lower the thermal conductivity has been
demonstrated. In order to achieve the maximum reduction in thermal conductivity, studies should
be performed to investigate further the role of microstructure as well as the role of intrinsic thermal
conductivity of the top coat materials. Guidelines for the selection of top coat materials as well as
optimum microstructures to reduce the thermal conductivity need to be established.
This work could lead to significantly more sinter-resistant, low thermal conductivity plasma
sprayed top coats, which could provide a real alternative to EB-PVD coating in some applications.

References

166
References
1. Rolls-Royce, The Jet Engine. 1996, Birmingham: Renault Printing Co Ltd.
2. Sims, C.T., Stoloff N.S., Hagel W.C., Superalloys II. 1987: John Wiley & Sons.
3. Becker, B. Potential Benefits of Advanced Materials for Performance and Life Cycle Cost of
Large Industrial Gas Turbines. in 5th International Charles Parson Turbine Conference:
Advanced Materials for 21st Century Turbines and Power Plant. 2000, edited by Strang, A.,
Banks, W.M., Conroy R.D., McColvin, G.M., Neal, J.C., and Simpson, S., Cambridge, UK,
p. 658-670.
4. Siegel, R. and Spuckler, C. M., Analysis of Thermal Radiation Effects on Temperatures in
Turbine Engine Thermal Barrier Coatings, Materials Science & Engineering A, 1998,
245(2), p. 150-159.
5. Endres, W. Design Principles of Gas Turbines. in Proceedings of the symposium on High
Temperature Materials in Gas Turbines. 1973, edited by Sahm PR, S.M., Baden,
Switzerland: Elsevier Scientific Publishing Company.
6. Kaysser W.A., Bartsch M., Krell T., Fritscher K., Leyens C., Schulz U., and M., Peters,
Ceramic Thermal Barriers for Demanding Turbine Applications, CFI-Ceramic Forum
International, 2000, 77(6), p. 32-35.
7. Beele, W., Marijnissen, G., and van Lieshout, A., The Evolution of Thermal Barrier
Coatings - Status and Upcoming Solutions for Today's Key Issues, Surface & Coatings
Technology, 1999, 121, p. 61-67.
8. Stover, D. and Funke, C., Directions of the Development of Thermal Barrier Coatings in
Energy Applications, Journal of Materials Processing Technology, 1999, 93, p. 195-202.
9. Winstone, M.R., Partridge, A., and Brooks, J.W. The Contribution of Advanced Materials to
Future Aeroengines. in 5th International Charles Parson Turbine Conference: Advanced
Materials for 21st Century Turbines and Power Plant. 2000, edited by Strang, A., Banks,
W.M., Conroy R.D., McColvin, G.M., Neal, J.C., and Simpson, S., Cambridge, UK, p. 780-
797.
10. Kimmel, J., Mutasim, Z., and Brentnall, W., Effects of Alloy Composition on the
Performance of Yttria Stabilized Zirconia-Thermal Barrier Coatings, Journal of
Engineering for Gas Turbines and Power - Transactions of the ASME, 2000, 122, p. 393-
400.
11. Price, F.C., Stiles, W.R., and Singh, C. Materials for Marine, Light Industrial and
Aeroderivative Gas Turbines. in 5th International Charles Parson Turbine Conference:
Advanced Materials for 21st Century Turbines and Power Plant. 2000, Cambridge, UK.
12. McLean, M., Nickel-base superalloys: current status and potential, Philosophical
Transactions of the Royal Society of London, 1995, A(351), p. 419-433.
13. Backmann, D.G. and Williams, J.C., Advanced Materials for Aircraft Engine Applications,
Science, 1992, 255(February), p. 1082-1087.
14. Miller, R. A., Thermal Barrier Coatings for Aircraft Engines: History and Directions,
Journal of Thermal Spray Technology, 1997, 6(1), p. 35-42.
15. Taylor, R., Brandon, J.R., and Morell, P., Microstructure, Composition and Property
Relationships of Plasma Sprayed Thermal Barrier Coatings, Surface & Coatings
Technology, 1992, 50, p. 141-149.
References

167
16. DeMasi-Marcin, J. T. and Gupta, D. K., Protective coatings in the gas turbine engine,
Surface & Coatings Technology, 1994, 68/69, p. 1-9.
17. Sims, C.T. and Hagel, W. C., The Superalloys. 1972: John Wiley & Sons.
18. Schulz, U., Leyens, C., Fritscher, K., Peters, M., Saruhan-Brings, B., Lavigne, O., Dorvaux,
J-M., Poulain, M., Mevrel, R., and Caliez, M., Some Recent Trends in Research and
Technology of Advanced Thermal Barrier Coatings, Aerospace Science and Technology,
2003, 7, p. 73-80.
19. Meetham, GW, Superalloys in gas turbine engines, in The Metallurgist and Materials
Technologist, 1982, p. 387-392.
20. Caron, P. and Khan, T., Evolution of Ni-based superalloys for single crystal gas turbine
blade applications, Aerospace Science and Technology, 1999, 3, p. 513-523.
21. Fox, A.C. and Clyne, T.W., Oxygen Transport through the Zirconia Layer in Plasma
Sprayed Thermal Barrier Coatings, Surface & Coatings Technology, 2004, 184, p. 311-321.
22. Wood J.H., Goldman E.H., Protective Coatings, in Superalloys II, edited by Sims, C.T.,
Stoloff N.S., Hagel W.C., 1987, John Wiley & Sons. p. 359-384.
23. Evans, A.G, Mumm, D.R., Hutchinson, J.W., Meier, G.H., and Pettit, F.S., Mechanisms
controlling the Durability of Thermal Barrier Coatings, Progress in Materials Science, 2001,
46, p. 505-553.
24. Tawancy, H. M., Abbas, N. M., and Bennett, A., Role of Y during High Temperature
Oxidation of an M-Cr-Al-Y Coating on an Ni-base Superalloy, Surface & Coatings
Technology, 1994, 68/69, p. 10-16.
25. Choquet, P. and Mevrel, R., Microstructure of Alumina Scales Formed on NiCoCrAl Alloys
with and without Yttrium, Materials Science & Engineering, 1989, A120, p. 153-159.
26. Taylor, M. P. and Evans, H. E., The Influence of Bond Coat Surface Roughness and
Structure on the Oxidation of a Thermal Barrier Coating System, Materials Science Forum,
2001, 369-3, p. 711-717.
27. Busso, E.P., Lin, J., Sakurai, S., and 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, p. 1515-1528.
28. Thompson, J.A, Tsui, Y.C., Reed, R.C., Rickerby, D.S., and Clyne, T.W., Creep of Plasma
Sprayed CoNiCrAlY and NiCrAlY Bond Coats and its Effects on Residual Stresses During
Thermal Cycling of Thermal Barrier Coating Systems, in High Temperature Surface
Engineering, edited by Nicholls, J. and Rickerby, D.S., 2000, IOM: Edinburgh, UK. p. 199-
212.
29. Evans, H. E. , Strawbridge, A. , Carolan, R. A. , and Ponton, C. B., Creep Effects on the
Spallation of an Alumina Layer from a NiCrAlY Coating, Materials Science & Engineering,
1997, A225, p. 1-8.
30. Rosler, J., Baker, M., and Volgmann, M., Stress state and failure mechanisms of thermal
barrier coatings: Role of creep in thermally grown oxide, Acta Materialia, 2001, 49(18), p.
3659-3670.
31. Herman, H. and Shankar, N. R., Survivability of Thermal Barrier Coatings, Materials
Science & Engineering, 1987, 88, p. 69-74.
References

168
32. Langhahr, P.A, Oberacker, R., and Hoffmann, M.J., Long-Term Behaviour and Application
Limits of Plasma-Sprayed Zirconia Thermal Barrier Coatings, Journal of the American
Ceramic Society, 2001, 84(6), p. 1301-1208.
33. Porter, D.L. and Heuer, A.H., Microstructural Development in MgO-Partially Stabilized
Zirconia (Mg-PSZ), Journal of the American Ceramic Society, 1979, 62, p. 298-305.
34. Bennett, A., Properties of Thermal Barrier Coatings, Materials Science and Technology,
1986, 2, p. 257-261.
35. Maier, R.D., Scheuermann, C.M., and Andrews, C.W., Degradation of Two Layer Thermal
Barrier Coating under Thermal Cycling, American Ceramic Society Bulletin, 1981, 60(5), p.
555-560.
36. Wilden J., Wank A., Application study on Ceria Based Thermal Barrier Coatings, Mat.-
wiss. u. Werkstofftech, 2001, 32, p. 654-659.
37. Lee, E.Y., Sohn, Y., Jha, S.K., Holmes, J.W., and Sisson, R.D. Jr., Phase transformations of
plasma sprayed zirconia-ceria thermal barrier coatings, Journal of the American Ceramic
Society, 2002, 85(8), p. 2065-2071.
38. Brandon, J. R. and Taylor, R., Phase Stability of Zirconia-Based Thermal Barrier Coatings
Part II. Zirconia-Ceria Alloys, Surface & Coatings Technology, 1991, 46, p. 91-101.
39. Khor, K. A. and Yang, J., Transformability of t-ZrO
2
and lattice parameters in plasma
sprayed rare-earth oxides stabilized zirconia coatings, Scripta Materialia, 1997, 37(9), p.
1279-1286.
40. Hamacha, R, Fauchais, P., and Nardou, F., Influence of Dopant on the Thermal Properties
of Two Plasma-sprayed Zirconia Coatings 1. Relationship Between Powder Characteristics
and Coating Properties, Journal of Thermal Spray Technology, 1996, 5(4), p. 431-438.
41. Nicholls, J.R., Lawson K.L, Rickerby D.S., and Morrell, P., Advanced Processing of TBC's
for Reduced Thermal Conductivity, in AGARD-R-823, 1997, NASA: Aalborg, Denmark, p.
6.1-6.9.
42. Grunling, H. W. and Mannsmann, W., Plasma-Sprayed Thermal Barrier Coatings For
Industrial Gas- Turbines - Morphology, Processing and Properties, Journal de Physique IV,
1993, 3(C7), p. 903-912.
43. Morrell, P. and Taylor, R., eds. Thermal Diffusivity and Phase Structure of Zirconia-Based
Thermal Barrier Coatings. Advances in Ceramics. Vol. Volume 24: Science and
Technology of Zirconia III. 1988, The American Ceramics Society.
44. Klemens, P.G. and Gell, M., Thermal Conductivity of Thermal Barrier Coatings, Materials
Science & Engineering, 1998, A245, p. 143-149.
45. Scardi, P., Leoni, M., Cernuschi, F., and Figari, A., Microstructure and heat transfer
phenomena in ceramic thermal barrier coatings, Journal of the American Ceramic Society,
2001, 84(4), p. 827-835.
46. Nicholls, J.R., Jaslier, Y., and Rickerby, D.S., Erosion and Foreign Object Damage of
Thermal Barrier Coatings, Materials Science Forum, 1997, 251, p. 935-948.
47. Nicholls, J.R., Deakin, M.J., and Rickerby, D.S., A Comparison between the Erosion
Behaviour of Thermal Spray and Electron Beam Physical Vapour Deposition Thermal
Barrier Coatings, Wear, 1999, 235, p. 352-361.
References

169
48. Strangman, T.E., Thermal Barrier Coatings for Turbine Airfoils, Thin Solid Films, 1985,
127, p. 93-105.
49. Morrell, P. and Rickerby, D.S., Advantages/Disadvantages of Various TBC Systems as
perceived by the Engine Manufacturer, in AGARD-R-823, 1998, NASA: Denmark, p. 20-1 -
20-9.
50. Hass, D.D., Slifka, A. J., and Wadley, H. N. G., Low Thermal conductivity Vapor Deposited
Zirconia Microstructures, Acta Materialia, 2001, 49(6), p. 973-983.
51. Fisher, I.A., Variables Influencing the Characteristics of Plasma-Sprayed Coatings,
International Metals Reviews, 1972, 17, p. 117-129.
52. Fauchais, P. and Vardelle, M., Plasma Spraying : Present and Future, Pure & Applied
Chemistry, 1994, 66(6), p. 1247-1258.
53. Apelian, D., Paliwal, M., Smith, R.W., and W.F., Schilling, Melting and Solidification in
Plasma Deposition - Phenomenological Review, International Metals Reviews, 1983, 28(5),
p. 271-294.
54. Heimann, R. B., Plasma-Spray Coating Principles and applications. 1996, Weinhein,
Germany: VCH Verlagsgesellschaft.
55. Gill, S. C., Residual Stresses in Plasma Sprayed Deposits, PhD Thesis, in Department of
Materials Science and Metallurgy, 1991, University of Cambridge: Cambridge.
56. Sampath, S and Herman, H. Microstructural Development of Plasma Sprayed Coatings. in
12th International Conference on Thermal Spray. 1989, edited by Berndt, C.C., The
Welding Institute, Cambridge, U.K., p. 761-766.
57. Matejka, D. and Benko, B., Plasma Spraying of Metallic and Ceramic Materials. 1989:
John Wiley & Sons. 280.
58. Schlichting, K.W., Padture, N.P., and Klemens, P.G., Thermal Conductivity of Dense and
Porous Yttria-stabilized Zirconia, Journal of Materials Science, 2001, 36, p. 3003-3010.
59. Kuroda, S. and Clyne, T.W., The Quenching Stress in Thermally Sprayed Coatings, Thin
Solid Films, 1991, 200, p. 49-66.
60. Thompson, J.A. and Clyne, T.W., The Effect of Heat Treatment on the Stiffness of Zirconia
Top Coats in Plasma-Sprayed TBCs, Acta Materialia, 2001, 49(9), p. 1565-1575.
61. Green, D.J., An Introduction to the Mechanical Properties of Ceramics. 1 ed. Cambridge
Solid State Science Series, ed. Clarke, D.R., Suresh, S. and Ward, I. M. 1998: Cambridge
University Press.
62. Malzbender, J. and Steinbrech, R. W., Determination of the Stress-Dependent Stiffness of
Plasma-Sprayed Thermal Barrier Coatings Using Depth-Sensitive Indentation, Journal of
Materials Research, 2003, 18(8), p. 1975-1984.
63. Eskner, M. and Sandstrom, R, Measurement of the Elastic Modulus of a Plasma-Sprayed
Thermal Barrier Coating using Spherical Indentation, Surface & Coatings Technology,
2004, 177, p. 165-171.
64. Kulkarni, A, Vaidya, A. , Goland, A., Sampath, S., and Herman, H., Processing Effects on
Porosity-Property Correlations in Plasma Sprayed Yttria-Stabilized Zirconia Coatings,
Materials Science and Engineering A, 2003, 359, p. 100-111.
References

170
65. Siebert, B., Funke, C., Vassen, R., and Stover, D., Changes in porosity and Young's
Modulus due to sintering of plasma sprayed thermal barrier coatings, Journal of Materials
Processing Technology, 1999, 93, p. 217-223.
66. Guo, H. B., Xu, H. B., Gong, S. K., and Liu, F. S., Effect of surface strengthening on high
temperature oxidation and bonding strength of EB-PVD thermal barrier coating, Rare
Metal Materials Engineering, 2001, 30(4), p. 314-317.
67. Pawlowski, L. and Fauchais, P., Thermal Transport Properties of Thermally Sprayed
Coatings, International Materials Reviews, 1992, 37(6), p. 271-289.
68. Schwingel, D., Taylor, R., Haubold, T., Wirgen, J., and Gaulco, C., Mechanical and
Thermophysical properties of thick PYSZ Thermal Barrier Coatings:Correlation with
Microstructure and Spraying Parameters, Surface & Coatings Technology, 1998, 108-109,
p. 99-106.
69. Ahmaniemi, S., Vuoristo, P., and Mantyla, T., Mechanical and elastic properties of
modified thick thermal barrier coatings, Materials Science and Engineering A, 2004, 366, p.
175-182.
70. Gadow, R. and Lischka, M., Lanthanum Hexa-aluminate - Novel Thermal Barrier Coatings
for Gas Turbine Applications - Materials and Process Development, Surface & Coatings
Technology, 2002, 151, p. 392-399.
71. Pfender, E., Fundamental Studies Associated with the Plasma Spray, Surface & Coatings
Technology, 1988, 34, p. 1-16.
72. Vaidya A., Bancke G., Sampath S., and H., Herman. Influence of Process Variables on the
Plasma Sprayed Coatings: An Integrated Study. in Thermal Spray 2001 :New Surfaces for a
New Millenium. 2001, edited by Bernd C.C., K.K.A., Lugscheider E.F.,, Materials Park,
Ohio, USA: ASM International.
73. Fauchais P., Vardelle M., Vardelle A., Bianchi L., Leger A.C.,, Parameters Controlling the
Generation and Properties of Plasmas Sprayed Zirconia Coatings, Plasma Chemistry and
Plasma Processing, 1996, 16(1), p. 99S-125S.
74. Smith, R. W., Wei, D., and Apelian, D., Thermal Plasma Materials Processing -
Applications and Opportunities, Plasma Chemistry and Plasma Processing, 1989, 9, p.
135S-165S.
75. Kuroda, S., Fukushima, T., and Kitahara, S., Generation Mechanisms of Residual-Stresses
in Plasma-Sprayed Coatings, Vacuum, 1990, 41(224), p. 1297-1299.
76. Tsui, Y. C., Adhesion of Plasma Sprayed Coatings, PhD thesis, in Department of Material
Science and Metallurgy, 1996, University of Cambridge: Cambridge, UK.
77. Scardi, P., Leoni, M., Bertini, L., and Bertamini, L., Residual Stress in Partially-Stabilised
Zirconia TBCs: Experimental Measurement and Modelling, Surface & Coatings Technology,
1997, 94-5, p. 82-88.
78. Scardi, P., Leoni, M., and Bertamini, L., Residual Stresses in Plasma Sprayed Partially
Stabilised Zirconia TBCs - Influence of the Deposition Temperature, Thin Solid Films, 1996,
278, p. 96-103.
79. Watson, J.W. and Levine, S.R., Deposition Stress Effects on the Life of Thermal Barrier
Coatings on Burner Rigs, Thin Solid Films, 1984, 119, p. 185-193.
References

171
80. Teixeira, V., Andritschky, M., Fischer, W., Buchkremer, H.P., and Stover, D., Analysis of
Residual Stresses in Thermal Barrier Coatings, Journal of Materials Processing Technology,
1999, 93, p. 209-216.
81. Teixeira, V., Andritschy, M., Fischer, W., Buchkremer, H.P., and Stover, D., Effects of
Deposition Temperature and Thermal Cycling on Residual Stress State in Zirconia-Based
Thermal Barrier Coatings, Surface & Coatings Technology, 1999, 121, p. 103-111.
82. Hobbs, M. K. and Reiter, H., Residual Stresses in ZrO
2
-8%Y
2
O
3
Plasma-Sprayed Thermal
Barrier Coatings, Surface & Coatings Technology, 1988, 34, p. 33-42.
83. Kingswell, R., Scott, K. T., and Sorensen, B., Measurement of Residual Stress in Plasma
Sprayed Ceramic Coatings, in 2nd Plasma-Technik-Symposium, edited by Blum-Sandmeier,
S., Eschnauer, H., Huber, P., and Nicoll, A.R., 1991, Plasma Technik. p. 377-388.
84. Kuroda, S., Kukushima, T., and Kitahara, S., In situ Measurement of Coating Thickness
during Thermal Spraying using an Optical Displacement Transducer, Journal of Vacuum
Science & Technology, 1987, A5, p. 82-87.
85. Kuroda, S., Kukushima, T., and Kitahara, S., Simultaneous Measurement of Coating
Thickness and Deposition Stress During Thermal Spraying, Thin Solid Films, 1988, 164, p.
157-163.
86. Kuroda, S., Fukushima, T., and Kitahara, S. Significance of the Quenching Stress in the
Cohesion and Adhesion of Thermally Sprayed Coatings. in 13th Int. Conf. Thermal Spraying.
1992, edited by Berndt, C.C., Orlando, Florida, USA: ASM, Metals Park, Ohio, p. 903-909.
87. Kuroda, S. and Clyne, T.W. The Origin and Quantification of the Quenching Stress
Associated with Splat Cooling during Spray Deposition. in 2nd Plasma Technik Symposium.
1991, edited by Eschenauer, H., Huber, P., Nicoll, A.R., and Sandmeier, S., Lucerne,
Switzerland: Plasma Technik, p. 273-284.
88. Stoney, G.G., The Tension of Metallic Films Deposited by Electrolysis, Proceedings of the
Royal Society of London, 1909, A82, p. 172-175.
89. Brenner, A. and Senderoff, S., Calculation of Stress in Electrodeposits from the Curvature
of a Plated Strip, Journal of Research of the National Bureau of Standards, 1949, 42, p. 105-
123.
90. Gill, S. C. and Clyne, T. W., Residual Stress Modelling and Characterisation of Thermally
Sprayed Ceramic Coatings, in High Performance Ceramic Films and Coatings, edited by
Vincenzini, P., 1991, Elsevier. p. 339-352.
91. Gill, S. C. and Clyne, T. W., Investigation of Residual Stress Generation during Thermal
spraying by Continuous Curvature Measurement, Thin Solid Films, 1994, 250, p. 172-180.
92. Matejicek, J., Sampath, C., Brand, P.C., and Prask, H.J., Quenching, Thermal and Residual
Stress in Plasma Sprayed Deposits: NiCrAlY and YSZ Coatings, Acta Materialia, 1999,
47(2), p. 607-617.
93. Levit, M., Grimberg, I., and Weiss, B.-Z., Residual Stresses in Ceramic Plasma-Sprayed
Thermal Barrier Coatings: Measurement and Calculation., Materials Science and
Engineering, 1995, A206, p. 30-38.
94. Zhuang, H. and Zhang, T. A study of the Residual Stresses in a ZrO
2
+ 8%Y
2
O
3
Coatings. in
2nd Plasma-Technik-Symposium. 1991, edited by Blum-Sandmeier, S., Eschnauer, H.,
Huber, P., and Nicoll, A.R., Lucerne, Switzerland: Hafliger Druck AG, Wettingen, p. 311-
318.
References

172
95. Clyne, T.W. and Gill, S.C., Residual Stresses in Thermally Sprayed Coatings and their
Effect on Interfacial Adhesion - A Review of Recent Work, Journal of Thermal Spray
Technology, 1996, 5, p. 1-18.
96. Kesler, O., Matejicek, J., Sampath, S., Suresh, S., Gnaeupel-Herold, T., Brand, P.C., and
Prask, H.J., Measurement of Residual Stress in Plasma-Sprayed Metallic, Ceramic and
Composite Coatings, Materials Science and Engineering, 1998.
97. Lifshin, E., X-ray Characterization of Materials. 1999: Wiley-VCH.
98. Reimers, W., Broda, M., Dantz, G., Liss, K.-D., Pyzala, A., Schmakers, T., and Tschenscher,
T., Evaluation of Residual Stresses in Bulk Materials by High Energy Synchrotron
Radiation, Journal of Non-destructive Evaluation, 1998, 17(3), p. 129-140.
99. Thompson, J.A., Matejicek, J., and Clyne, T.W. Modelling and Neutron Diffraction
Measurement of Stresses in Sprayed TBCs. in Superalloys 2000. 2000, edited by Pollock,
T.M., Kissinger, R.D., Bowman, R.R., Green, K.A., McLean, M., Olsen, S.L., and Schirra,
J.J., Seven Springs, USA: TMS, p. 639-647.
100. Scardi, P., Leoni, M., Bertini, L., Bertamini, L., and Cernuschi, F., Strain Gradients in
Plasma - Sprayed Zirconia Thermal Barrier Coatings, Surface and Coatings Technology,
1998, 108-9, p. 93-98.
101. Rickerby, D. S., Scott, K. T., and Eckold, G. Analysis of the Residual Stresses in Plasma
Sprayed Coatings. in First Plasma-Technik Symposium. 1988, edited by Eschenauer, H.,
Huber, P., Nicoll, A.R., and Sandmeier, S., Whlen, Switzerland: Plasma Technik, p. 267-
276.
102. Bengtsson, P. and Persson, C., Modelled and Measured Residual Stresses in Plasma
Sprayed Thermal Barrier Coatings, Surf. & Coatings Tech., 1997, 92, p. 78-86.
103. Gill, S. C. and Clyne, T. W., Modeling of the generation of residual stress during thermal
spraying of ceramic coatings, High Performance Ceramic Films and Coatings, 1991.
104. Rickerby, D. S., Eckold, G., Scott, K. T., and Buckley-Golder, I. M., The Interrelationship
between Internal Stress, Processing Parameters and Microstructure of Physically Vapour
Deposited and Thermally Sprayed Coatings, Thin Solid Films, 1987, 154, p. 125-141.
105. Stover, D., Jager, D. A., and Schutz, H. G. Residual Stresses in Low Pressure Plasma
Sprayed Chromia Coatings. in Proceedings of The Fourth National Thermal Spray
Conference. 1991, edited by Bernecki, T.F., Pittsburgh, PA, USA: The Materials
Information Society, p. 215-219.
106. Elsing, R., Knotek, O., and Balting, U., Calculation of Residual Thermal Stress in Plasma-
Sprayed Coatings, Surface & Coatings Technology, 1990, 43/44, p. 416-425.
107. Tsui, Y.C. and Clyne, T.W. Adhesion of Thermal Barrier Coating Systems and
Incorporation of an Oxidation Barrier Layer. in Thermal Spray: Practical Solutions for
Engineering Problems. Proceedings of the 9th Nat. Thermal Spray Conf. 1996, edited by
Berndt, C.C., Cincinatti: ASM, p. 275-284.
108. Tsui, Y.C. and Clyne, T.W., An Analytical Model for Predicting Residual Stresses in
Progressively Deposited Coatings. Part 1: Planar Geometry, Thin Solid Films, 1997, 306,
p. 23-33.
109. Thompson, J.A., Thermomechanical Behaviour of Plasma-Sprayed Thermal Barrier
Coatings, PhD Thesis, in Department of Materials Science and Metallurgy, 2000,
University of Cambridge: Cambridge, UK.
References

173
110. Lugscheider, E., Bobzin, K., Barwulf, S., and Etzkorn, A., Mechanical properties of EB-
PVD-thermal barrier coatings by nanoindentation, Surface & Coatings Technology, 2001,
138(1), p. 9-13.
111. Wallace, J.S. and Ilavsky, J., Elastic Modulus Measurements in Plasma Sprayed Deposits,
Journal of Thermal Spray Technology, 1998, 7(4), p. 521-526.
112. Zhu, D.M. and Miller, R.A., Thermal Conductivity and Elastic Modulus Evolution of
Thermal Barrier Coatings under High Heat Flux Conditions, Journal of Thermal Spray
Technology, 2000, 9, p. 175-180.
113. Siebert, S., Funke, C., Vassen, R., and Stover, D., Changes in Porosity and Young's
Modulus due to Sintering of Plasma-Sprayed Thermal Barrier Coatings, Journal of
Materials Processing Technology, 1999, 93(217-223).
114. Leigh, S.H, Lin, C.K, and Berndt, C. C., Elastic response of thermal spray deposits under
indentation tests, Journal of the American Ceramic Society, 1997, 80(8), p. 2093-2099.
115. Vasquez, D.L., Kucuk, A., Lima, R. S., Senturk, U., and Berndt, C. C. Elastic modulus
measurements of air plasma sprayed yttria partially stabilized zirconia coatings using laser
ultrasonics and indentation techniques. in Thermal Spray 2001:New Surfaces for a New
Millenium. 2001, edited by Berndt, C.C., Khor, K.A., and Lugscheider, E.F., Ohio, USA:
ASM International, p. 1045-1050.
116. Steffens, H.D. and Fischer, U. Correlation between microstructure and physical properties
of plasma sprayed zirconia properties. in Thermal Spray: Advances in Coatings Technology.
1998, edited by Houck, D.L., Ohio, USA: ASM International, p. 167-173.
117. Duan, K. and Steinbrech, R. W., Influence of sample deformation and porosity on
mechanical properties by instrumented microindentation technique, Journal of the European
Ceramic Society, 1998, 18(2), p. 87-93.
118. Pawlowski, L., The Science and Engineering of Thermal Spray Coatings. 1995, Chichester,
England: J. Wiley & Sons Ltd.
119. Kulkarni, A., Wang, Z., Nakamura, T., Sampath, S., Goland, A., Herman, H., Allen, J.,
Ilavsky, J., Long, G., Frahm, J., and Steinbrech, R. W., Comprehensive microstructural
characterization and predictive property modeling of plasma-sprayed zirconia coatings,
Acta Materialia, 2003, 51(9), p. 2457-2475.
120. Kroupa, F., and Plesek, J., Nonlinear Elastic Behaviour in Compression of Thermally
Sprayed Materials, Materials Science and Engineering, 2002, 328(1-2), p. 1-7.
121. Harok, V. and Neufuss, K., Elastic and inelastic effects in compression in plasma-sprayed
ceramic coatings, Journal of Thermal Spray Technology, 2001, 10(1), p. 126-132.
122. Kroupa, F., and Dubsky, J., Pressure Dependence of Young's Moduli of Thermal Sprayed
Materials, Scripta Materialia, 1999, 40(11), p. 1249-1254.
123. Choi, S. R., Zhu, D. , and Miller, R. A., Deformation and Tensile Cyclic Fatigue of Plasma-
Sprayed ZrO
2
-8wt % Y
2
O
3
Thermal barrier Coatings, Ceramic Engineering and Science
Proceedings, 2001, 22(4), p. 427-434.
124. Rejda, E. F., Socie, D. F., and Beardsley, B., Fatigue behavior of a plasma-sprayed
8%Y2O3-ZrO2 thermal barrier coating, Fatigue &Fracture of Engineering Materials and
Structures, 1997, 20(7), p. 1043-1050.
125. Gibson, L.J. and Ashby, M.F., The Mechanics of Three-Dimensional Cellular Materials,
Proceedings of the Royal Society of London, 1982, A382, p. 43-59.
References

174
126. MacKenzie, J.K., The Elastic Constants of a Solid Containing Spherical Holes, Proceedings
of the Physical Society, 1950, B63, p. 2-11.
127. Eshelby, J.D., The Determination of the Elastic Field of an Ellipsoidal Inclusion and
Related Problems, Proceedings of the Royal Society of London, 1957, 241, p. 376-396.
128. Leigh, S.H. and Berndt, C.C., Modelling of Elastic Constants of Plasma Spray Deposits with
Ellipsoid-Shaped Voids, Acta Materialia, 1999, 47(5), p. 1575-1586.
129. Sevostianov, I. and Kachanov, M., Modeling of the anisotropic elastic properties of plasma-
sprayed coatings in relation to their microsturcture, Acta Materialia, 2000, 48, p. 1361-
1370.
130. Kachanov, M., Sevostianov, I., and Shafiro, B., Explicit cross-property correlations for
porous materials with anisotropic microstructures, Journal of Mechanics and Physics of
Solids, 2001, 49(1), p. 1-25.
131. Wang, Z, Kulkarni, A, Deshpande, S, Nakamura, T, and Herman, H, Effects of pores and
interfaces on effective properties of plasma sprayed zirconia coatings, Acta Materialia, 2003,
51, p. 5319-5334.
132. McPherson, R. and Shafer, V., Interlamellar Contact within Plasma-Sprayed Coatings, Thin
Solid Films, 1982, 97, p. 201-204.
133. Dobbins, T.A., Allen, A.J., Ilavsky, J., Long, G.G., P.R.Jemian, Kulkarni, A., and Herman,
H., Recent Development in the Characterization of Anizotropic Void Populations in Thermal
Barrier Coatings Using Ultra-small Angle X-ray Scattering, Ceramic Engineering and
Science Proceedings, 2003, 24(3), p. 517-524.
134. McPherson, R., A Review of Microstructure and Properties of Plasma Sprayed Ceramics
Coatings, Surface & Coatings Technology, 1989, 39/40, p. 173-181.
135. Strunz, P., Schumacher, G., Vassen, R., and Wiedenmann, A., In Situ SANS Study of Pore
Microstructure in YSZ Thermal Barrier Coatings, Acta Materialia, 2004, 52, p. 3305 -3312.
136. Allen, A. J., Ilavsky, J., Long, G. G., Wallace, J. S., Berndt, C. C., and Herman, H.,
Microstructural Characterisation of Yttria-stabilised Zirconia Plasma-sprayed Deposits
using Multiple Small-Angle Neutron Scattering, Acta Materialia, 2001, 49, p. 1661-1675.
137. Boukari, H., Allen, A. J., Long, G. G., Ilavsky, J., Wallace, J. S., Berndt, C. C., and Herman,
H., Small-angle neutron scattering study of the role of feedstock particle size on the
microstructural behavior of plasma-sprayed yttria-stabilized zirconia deposits, Journal of
Materials Research, 2003, 18(3), p. 624-634.
138. Ilavsky, J., Long, G.G., Allen, A.J., and Berndt, C.C., Evolution of the Void Structure in
Plasma-Sprayed YSZ Deposits During Heating, Materials Science and Engineering, 1999,
272(1), p. 215-221.
139. Lughi, V., Tolpygo, V.K., and Clarke, D. R., Microstructural aspects of the sintering of
thermal barrier coatings, Material Science and Engineering, 2004, A272, p. 215-221.
140. Lelait, L., Alperine, S., and Mevrel, R., Alumina Scale Growth at Zirconia-MCrAlY
Interface: A Microstructural Study, Journal of Materials Science, 1992, 27, p. 5-12.
141. Lipkin, D.N. and Clarke, D. R., Measurement of the stress in oxide scales formed by
oxidaton of alumina-forming alloys, Oxidation of Metals, 1996, 45(3/4), p. 267-280.
References

175
142. Tolpygo, V.K., Dryden, J.R., and Clarke, D.R., Determination of the Growth Stress and
Strain in o-Alumina Scales during the Oxidation of Fe-22Cr-4.8Al-0.3Y Alloy, Acta
Materialia, 1998, 46, p. 927-937.
143. Tolpygo, V. K. and Clarke, D. R., Spalling failure of oAlumina Films grown by Oxidation.
I: Dependence on Cooling Rate and Metal Thickness, Materials Science and Engineering,
2000, A278, p. 142-150.
144. Clarke, D.R., Christenson, R.J., and Tolpygo, V., The Evolution of Oxidation Stresses in
Zirconia Thermal Barrier Coated Superalloy Leading to Spalling Failure, Surface &
Coatings Technology, 1997, 94-5, p. 89-93.
145. Ali, M.Y., Nusier, S.Q., and Newaz, G.M., Creep effects on early damage initiation in a
TBC system, Journal of Materials Science, 2004, 39(10), p. 3383-3390.
146. Tolpygo, V. K. and Clarke, D. R., Spalling failure of o-Alumina Films grown by Oxidation:
II. Decohesion Nucleation and Growth, Materials Science and Engineering, 2000, A278, p.
151-161.
147. German, R.N., Sintering Theory and Practice. 1 ed. 1996, New York: John Wiley & Sons.
148. Shi, J.L., Thermodynamics and densification kinetics in solid state sintering of ceramics,
Journal of Materials Research, 1999, 14(4), p. 1398-1408.
149. Lay, K.W. Grain growth during sintering. in 3rd International Conference on Sintering and
Related Phenomena. 1972, edited by Kuczynski, G.C., Indiana, USA: Plenum Press, p. 65--
80.
150. Ashby, M.F., A First Report on Sintering Diagrams, Acta Metallurgica, 1974, 22, p. 275-
289.
151. Pask, J.A. and Hoge, C.E. Thermodynamic aspects of solid state sintering. in Sintering and
Related Phenomena. 1972, edited by Kuczynski, G.C., Notre Dame, Indiana, USA: Plenum
Press, New York, p. 229-238.
152. Burke, J.E., Role of Grain Boundaries in Sintering, Journal of the American Ceramic
Society, 1957, 40(3), p. 80-85.
153. Mendelev, M.I. and D.J., Srolovitz, Impurity effects on grain boundary migration,
Modelling and Simulation in Materials Science and Engineering, 2002, 10, p. R79-R-109.
154. Cahn, J.W., The impurity-drag effect in grain boundary motion, Acta Metallurgica, 1962, 10,
p. 769-798.
155. Brook, R.J., Howlett, S.P., and Wu, S.X. The use of solid solution additives in sintering. in
Sintering- Theory and Practice. 1981, edited by Kolar, d., Pejovnik, S., and Ristic, M.M.,
Yugoslavia: Elsevier Scientific Publishing Company, Amsterdam, p. 135-144.
156. Kroger, F.A. Point defects in compounds and their role in diffusion. in Sintering and
Related Phenomena. 1967, edited by Kuczynski, G.C., Hooton, N.A., and Gibbon, C.F.,
New York: Gordon and Breach.
157. Coblenz, W.S., Dynys, J.M., Cannon, R.M., and Coble, R.L. Initial Stage Solid State
Sintering Models. A Critical Analysis and Assessment. in Sintering and Related Phenomena.
1980, edited by Kuczynski, G.C.: Plenum Press, p. 36-43.
158. Kingery, W.D. and Berg, M., Study of the Initial Stages of Sintering Solids by Viscous Flow,
Evaporation-Condensation and Self-Diffusion, Journal of Applied Physics, 1955, 26(10), p.
1205-1212.
References

176
159. Coble, R.L., Sintering crystalline solids: I. Intermediate and Final Stage Diffusion models,
Journal of Applied Physics, 1961, 32(5), p. 787-792.
160. Kingery, W.D. and Berg, M., Study of the Initial Stages of Sintering Solids by Viscous Flow,
Evaporation-Condensation, and Self Diffusion, Journal of Applied Physics, 1955, 26(10), p.
1205-1212.
161. Kuczynski, G.C., Self-diffusion in sintering of metallic particles, Trans A.I.M.E., 1949,
185(February), p. 169-177.
162. Maximenko A.L. and Olevsky, E.A., Effective diffusion coefficients in solid-state sintering,
Acta Materialia, 2004, 52, p. 2953-2963.
163. Exner, H.E. and Petzow, G. A critical evaluation of shrinkage equations. in Sintering
Processes. 1979, edited by Kuczynski, G.C., Indiana, USA: Plenum Press, p. 107-120.
164. Coblenz. Initial stage solid state sintering models. A critical analysis and assessment. in
Sintering Processes. 1979, edited by Kuczynski, G.C., Indiana, USA: Plenum Press, p. 141-
157.
165. Exner, H.E., Solid-state sintering: critical assessment of theoretical concepts and
experimental methods, Powder Metallurgy, 1980, 4, p. 203-209.
166. Harmer, M.P. Use of Solid-Solution Additives in Ceramic Processing. in Advances in
Ceramics: Structure and Properties of MgO and Al
2
O
3
ceramics. 1983, edited by Kingery,
W.D., Cambridge Massachusetts, USA: The American Ceramic Society.
167. Zhao, J. and Harmer, M.P., Effect of pore distribution on microstructure development: II
First- and second-generation pores, Journal of the American Ceramic Society, 1988, 71(7),
p. 530-539.
168. Herring, C., Effect of scale on sintering phenomena, Journal of Applied Physics, 1950, 21, p.
301-303.
169. Rhodes, W.H., Agglomerate and particle size effects on sintering of yttria stabilzed zirconia,
Journal of the American Ceramic Society, 1981, 64(1), p. 19-22.
170. Shi, J.L., Gao, J.H., Lin, Z.X., and Yen, T.S., Sintering Behavior of Fully Agglomerated
Zirconia Compacts, Journal of the American Ceramic Society, 1991, 74(5), p. 994-997.
171. Chappell, J.S., Ring, T.A., and Birchall, J.D., Particle size distribution effects on sintering
rates, Journal of Applied Physics, 1986, 6(1), p. 383-319.
172. German, R. M., Liquid Phase Sintering. 1985, New York: Plenum Press.
173. De Souza, D.P.F. and De Souza, M.F., Liquid Phase Sintering of RE
2
O
3
: YSZ Ceramics -
Part I - Grain Growth and Expelling of the Grain Boundary Glass Phase, Journal of
Materials Science, 1999, 34, p. 4023-4030.
174. Kingery, W.D., Introduction to Ceramics. 1960: Jonh Wiley and Sons.
175. Richerson, D.W., Modern Ceramic Engineering. 2nd ed, ed. Peter, H. 1992: Marcel Dekker
Inc.
176. del Olmo, L., Duran, P., and Moure, C. Sintering behaviour of an yttria stabilized zirconia.
in Sintering-theory and Practice. 1981, edited by Kolar, D., Pejovnik, S., and Ristic, M.M.,
Portooz, Yugoslavia: Elsevier Scientific, Amsterdam, p. 401-408.
References

177
177. Nasar, R. S., Cerqueira, M., Longo, E., and Varela, J. A., Sintering mechanisms of ZrO
2

center dot MgO with addition of TiO
2
and CuO, Ceramics International, 2004, 30(4), p. 571-
577.
178. Jorgensen, P.J. Diffusion controlled sintering in oxides. in Sintering and Related Phenomena.
1967, edited by Kuczynski, G.C., Hooton, N.A., and Gibbon, C.F.: Gordon and Breach,
New York, p. 401-422.
179. Brook, R.J. Preparation and Electrical Behaviour of Zirconia Ceramics. in Science and
Technology of Zirconia. 1981, edited by Heuer, A.H. and Hobbs, L.W., Cleveland, Ohio,
USA: The American Ceramic Society Inc., p. 272-285.
180. Chien, F. R. and Heuer, A. H., Lattice diffusion kinetics in Y
2
O
3
-stabilized cubic ZrO
2
single
crystals: a dislocation loop annealing study, Philosophical Magazine A, 1996, 73, p. 681-
697.
181. Solmon, H., Chaumont, J., Dolin, C., and Monty, C. Zr, Y and O self diffusion in Zr
(1-x)
Y
x
O
2-
x/2
. in Ceramic Transactions- Point Defects and Related Properties of Ceramics. 1991,
edited by Mason, T.O. and Routbort, J.L., p. 175-184.
182. Zhu, D.M. and Miller, R.A., Sintering and Creep Behaviour of Plasma-Sprayed Zirconia-
and Hafnia-Based Thermal Barrier Coatings, Surface & Coatings Technology, 1998, 109(1-
3), p. 114-120.
183. Kilo, M., Taylor, B., Argirusis, Ch., Borchardt, G., Lesage, B., Weber, S., Scherrer, S.,
Scherrer, H., Schroeder, M., and Martin, C., Cation self-diffusion of Ca, Y, and Zr in single
crystalline calcia- and yttria-doped zirconia, Journal of Applied Physics, 2003, 94(12), p.
7547-7552.
184. Kilo, M., Taylor, M.A., Argirusis, Ch., Borchardt, G., Weber, S., Scherrer, H., and Jackson,
R.A., Lanthanide transport in stabilized zirconias: Interaction between ionic radius and
diffusion coefficient, Journal of Chemical Physics, 2004, 12(11), p. 5482-5487.
185. Wu, S.X. and Brook, R.J., Sintering Aids for Zirconia Ceramics, Transactions and journal of
the British Ceramic Society, 1983, 82(6), p. 200-205.
186. Alleman, J.A., Michel, B., Marki, H.B., Gauckler, L.J., and Moser, E.M., Grain growth of
differently doped Zirconia, Journal of the European Ceramic Society, 1995, 15, p. 951-958.
187. Hwang, S-L. and Chen, I-W., Grain size control of tetragonal zirconia polycrystals using
the space change concept, Journal of the American Ceramic Society, 1990, 73(11), p. 3269-
3277.
188. Theunissen, G.S.A.M., Winnubst, A.J.A., and Burggraaf, A.J., Surface and grain boundary
analysis of doped zirconia ceramics studies by AES and XPS, Journal of Materials Science,
1992, 27, p. 5057-5066.
189. Stemmer, S., Vleugels, J., and Van Der Biest, O., Grain boundary segregation in high-
purity yttria-stabilized tetragonal zirconia polygrystals (T-TZP), Journal of the European
Ceramic Society, 1998, 18, p. 1565-1570.
190. Boutz, M.M.R., Winnubst, A.J.A., and Burggraaf, A.J., Yttria-Ceria Stabilized Tetragonal
Zirconia Polycrystals - Sintering, Grain Growth and Grain Boundary Segregation, Journal
of the European Ceramic Society, 1994, 13(2), p. 89-102.
191. Lynn Johnson, D. Impurity Effects in the Initial Sintering of Oxides. in Sintering and related
phenomena. 1967, edited by Kuczynski, G.C., Hooton, N.A., and Gibbon, C.F.: Gorden and
Breach, New York.
References

178
192. Verkerk, M., Winnubst, A., and Burggraaf, A., Effect of Impurities on Sintering and
Conductivity of Yttria-Stabilised Zirconia, Journal of Materials Science, 1982, 17, p. 3113-
3122.
193. Tekeli, S., Erdogan, M., and Aktas, B., Microstructural evolution in 8 mol% Y
2
O
3
-stabilized
cubic zirconia (8YSCZ) with SiO
2
addition, Materials Science and Engineering a-Structural
Materials Properties Microstructure and Processing, 2004, 386(1-2), p. 1-9.
194. Shi, J.L., Yen, T.S., and Schubert, H., Effect of Small Amounts of Additives on the Sintering
of High-Purity Y-TZP, Journal of Materials Science, 1997, 32, p. 1341-1346.
195. Tsubakino, H., Fujiwara, T., Satani, K., and Ioku, S., Composition of grain boundary phase
formed in zirconia-3 mol% yttria containing alumina, Journal of Materials Science Letters,
1997, 16(17), p. 1472-1475.
196. Srdic, V.V., Winterer, M., and Hahn, H., Sintering Behaviour of Nanocrystalline Zirconia
Doped with Alumina Prepared by Chemical Vapour Synthesis, Journal of the American
Ceramic Society, 2000, 83(8), p. 1853-1860.
197. Eaton, H.E. and Novak, R.C., Sintering Studies of Plasma Sprayed Zirconia, Surface &
Coatings Technology, 1987, 32, p. 227-236.
198. Vassen, R., Czech, N., Mallener, W., Stamm, W., and Stoever, D., Influence of Impurity
Conetent and Porosity of Plasma-Sprayed Yttria-Stabilized Zirconia Layers on the Sintering
Behaviour, Surface & Coatings Technology, 2001, 141, p. 135-140.
199. Klocker, T. and Clyne, T.W., The Effects of Bond Coat Roughness and Top Coat Purity on
the Adhesion of Thermal Barrier Coatings, Surface & Coatings Technology, 2002, p.
submitted.
200. Gibson, I.R., Dransfield, G.P., and Irvine, J.T.S., Sinterability of Commercial 8 mol%
Yttria-Stabilized Zirconia Powders and the Effect of Sintered Density on the Ionic
Conductivity, Journal of Materials Science, 1998, 33(17), p. 4297-4305.
201. Ilavsky, J., Long, G.G., and Allen, A.J. Evolution of the Microstructure of Plasma Sprayed
Deposits During Heating. in 5th International Thermal Spray Conference. 1998, edited by
Coddet, C., Nice, France: ASM International, p. 1641-1644.
202. Fabris, S., Paxton, A.T., and Finnis, M.W., A stabilization mechanism of zirconia based on
oxygen vacancies only, Acta Materialia, 2002, 50, p. 5171-5178.
203. Ho, S-M., On the structural chemistry of zirconium oxide, Materials Science and
Engineering, 1982, 54, p. 23-29.
204. Li, P. and Chen, I-W., Effect of dopants on zirconia stabilization - an-x-ray absorption study:
I. trivalent dopants, Journal of the American Ceramic Society, 1994, 77(1), p. 118-1128.
205. Yoshimura, M., Phase Stability of Zirconia, Ceramic Bulletin, 1988, 67(12), p. 1950-1955.
206. Yoshimura, M., Yashima, M., Noma, T., and Somiya, S., Formation of diffusionlessly
transformed tetragonal phases by rapid quenching of melts in ZrO
2
-RO
1.5
systems (R=rare
earths), Journal of Materials Science, 1990, 25, p. 2011-2016.
207. Sheu, T-S., Tien, T-Y., and Chen, I-W., Cubic-to-tetragonal (t') transformation in zirconia-
containing systems, Journal of the American Ceramic Society, 1992, 75(5), p. 1108-1116.
208. Li, P. and Chen, I-W., Effect of dopants on zirconia stabilization - an-x-ray absorption study:
II. tetravalent dopants, Journal of the American Ceramic Society, 1994, 77(1), p. 1281-1288.
References

179
209. Mastelaro, V.R., Briois, V., de Souza D.P.F., and Silva, C.L., Structural studies of a ZrO
2
-
CeO
2
doped system, Journal of the European Ceramic Society, 2003, 23(3), p. 273-282.
210. Brandon, J.R. and Taylor, R., Thermal Properties of Ceria and Yttria Partially Stabilized
Zirconia Thermal Barrier Coatings, Surface & Coatings Technology, 1989, 39/40, p. 143-
151.
211. Lee, C.H., Kim, H.K., Choi, H.S., and Ahn, H.S., Phase Transformation and Bond coat
Oxidation Behaviour of plasma-Sprayed Zirconia Thermal Barrier Coating, Surface &
Coatings Technology, 2000, 124, p. 1-12.
212. Harmsworth, P. D. and Stevens, R., Phase Composition and Properties of Plasma-Sprayed
Zirconia Thermal Barrier Coatings, Journal of Materials Science, 1992, 27, p. 611-615.
213. Ilasky J., Stalick J.K., Wallace J., Thermal Spray yttria-Stabilized Zirconia Phase Changes
during Annealing, Journal of thermal Spray Technology, 2001, 10(3), p. 491-501.
214. VanValzah, J.R. and Eaton, H.E., Cooling Rate Effects on the Tetragonal to Monoclinic
Phase Transformation in Aged Plasma-Sprayed Yttria Partially Stabilised Zirconia, Surface
& Coatings Technology, 1991, 46, p. 289-300.
215. Lee, J.D., Ra, Y.H., Hong, K.T., and S.K., Hur, The Effect of Residual Stress on Martensitic
Transformation in Plasma-Sprayed ZrO
2
(8 wt % Y
2
O
3
) Coatings, Surface & Coatings
Technology, 1992, 54/55, p. 64-69.
216. Gupta, T. K., Lange, F. F., and Bechtold, J. H., Effect of Stress-Induced Phase
Transformation on the Properties of Polycrystalline Zirconia containing Metastable
Tetragonal Phase, Journal of Materials Science, 1978, 13, p. 1464-1470.
217. Moon, J, Choi, H, Kim, H, and Lee, C, The effects of heat treatment on the phase
transformation behaviour of plasma-sprayed stabilized ZrO
2
coatings, Surface & Coatings
Technology, 2002, 155(1), p. 1-10.
218. Choi, H., Kim, H., and Lee, C., Phase evolution of plasma sprayed ceria and yttria
stabilized zirconia thermal barrier coating, Journal of Materials Science Letters, 2002, 21, p.
1359-1361.
219. Dal Maschio, R., Scardi, P., and Lutteroti, L., Influence of Ce
3+
/Ce
4+
ration on the phase
stability and residual stress field in ceria-yttria stabilized plasma-sprayed coatings, Journal
of Materials Science, 1992, 27, p. 5591-5596.
220. Portinha, A., Teixeira, V., Carneiro, J., Costa, M.F., Barrabas, N.P., and Sequeira, A.D.,
Stabilization of ZrO
2
PVD coatings with Gd
2
O
3
, Surface & Coatings Technology, 2004,
188-189, p. 107-1115.
221. Rebollo, R.N., Fabrichnaya, O., and Levi, C.G., Phase stability of Y+Gd co-doped zirconia,
Z. Metallkd., 2003, 94(3), p. 163-170.
222. Khor, K. A. and Yang, J., Lattice parameters, tetragonality (c/a) and transformability of
tetragonal zirconia phase in plasma sprayed ZrO
2
-Er
2
O
3
coatings, Materials Letters, 1997,
31, p. 23-27.
223. Khor, K. A. and Yang, J., Plasma spraying of samaria-stabilized zirconia powders and
coatings, Materials Letters, 1997, 31, p. 165-171.
224. Jones, R. L. and Mess, D., Improved tetragonal phase stability at 1400
o
C with scandia,
yttria-stabilized zirconia, Surface & Coatings Technology, 1996, 87-8(1-3), p. 94-101.
References

180
225. Leoni, M., Jones, R.L., and Scardi, P., Phase Stability of Scandia - Yttria Stabilized Zirconia
TBCs, Surface and Coatings Technology, 1998, 108-9, p. 107-113.
226. Stubican, V.S. and Hellman, J.R. Phase Equilibria in some Zirconia Systems. in First
International Conference on the Science and Technology of Zirconia. 1980, edited by Heuer,
A.H. and Hobbs, L.W., Cleveland, Ohio: American Ceramic Society, p. 25-36.
227. Bastide, B., Odier, P., and Coutures, J.P., Phase Equilibrium and Martensitic
Transformation in Lanthana-Doped Zirconia, Journal of the American Ceramic Society,
1988, 71(6), p. 449-453.
228. Andrievskaya E.R. and L.M., Lopato, Influence of composition of the t->m transformation
in the systems ZrO
2
-Ln
2
O
3
(Ln= La, Nd, Sm, Eu), Journal of Materials Science, 2001, 30(10),
p. 2592-2596.
229. Thangadurai, P., Sabarinathan, V., Chandra Bose, A., and Ramasamy, S., Conductivity
behaviour of a cubic/tetragonal phase stabilized nanocrystalline La
2
O
3
-ZrO
2
, Journal of
Physics and Chemistry of Solids, 2004, 65, p. 1905-1912.
230. Vassen, R., Cao, X. Q., Tietz, F., Basu, D., and Stover, D., Zirconates as new materials for
thermal barrier coatings, Journal of the American Ceramic Society, 2000, 83(8), p. 2023-
2028.
231. Cao, X. Q., Vassen, R. and Stoever, D., Ceramic Materials for Thermal Barrier Coatings,
Journal of the European Ceramic Society, 2004, 24, p. 1-10.
232. Miller, R.A., Smialek, J.L., and Garlick, R.G. Phase Stability in Plasma-Sprayed, Partially
Stabilized Zirconia-Yttria. in First International Conference on the Science and Technology
of Zirconia. 1980, edited by Heuer, A.H. and Hobbs, L.W., Cleveland, Ohio: American
Ceramic Society, p. 241-253.
233. Berman, R., Thermal Conduction in Solids. 1976, Oxford: Clarendon.
234. Lee, D.W. and Kingery, W.D., Radiation Energy Transfer and Thermal Conductivity of
Ceramic Oxides, Journal of the American Ceramic Society, 1960, 43(11), p. 594-605.
235. Klemens, P. G., Theory of the Thermal Conductivity of Solids, in Thermal Conductivity,
edited by Tye, R.P., 1969, Academic Press: Cambridge, Massachusetts, U.S.A. p. 1-65.
236. Klemens, P. G., Thermal Resitance due to Point Defects at High Temperatures, Physical
Review A, 1960, 119(21), p. 507-509.
237. Siegel, R and Howell, J R, Thermal Radiation Heat Transfer. 1st ed. 1972, New York:
McGraw-Hill. 814.
238. Peelen, J.G. and Metselaar, R., Light scattering by pores in polycrystalline materials:
Transmission properties of alumina, Journal of Applied Physics, 1974, 45(1), p. 216-220.
239. Manara, J., Caps, R., Raether, F., and Fricke, J., Characterization of the pore structure of
alumina ceramics by diffuse radiation propagation in the near infrared, Opt. Commun.,
1999, 168(1-4), p. 237-250.
240. Parrott, J. E. and Stuckes, A. D., Thermal Conductivity of Solids. 1975, London: Pion
Limited. 157.
241. Gustafsson, S.E., Karawacki, E, and Khan, M.Nazim, Transient hot-strip method for
simultaneously measuring thermal conductivity and thermal diffusivity of solids and fluids,
Journal of Physics D, 1979, 12, p. 1411-1421.
References

181
242. Pettersson, S., Theoretical considerations on the hot-wire method applied to solids under
pressure, Review of Scientific Instruments, 1990, 61(4), p. 1308-1313.
243. Log, T. and Gustafsson, S.E., Transient Plane Source (TPS) Technique for Measuring
Thermal Transport Properties of Building Materials, Fire and Materials, 1995, 19, p. 43-49.
244. Karawacki, E and Suleiman, M., Dynamic plane source technique for simultaneous
determination of specific heat, thermal conductivity and thermal diffusivty of metallic
samples, Measurement Science & Technology, 1991, 2, p. 744-750.
245. Gustafsson, S.E. and Khan, M.N., Determination of the thermal-conductivity tensor and the
heat capacity of insulating solids with the transient hot-strip method, Journal of Applied
Physics, 1981, 52(4), p. 2596-2600.
246. Gustafsson, S.E. and Karawacki, E, Transient hot-strip probe for measuring thermal
properties of insulating solids and liquids, Review of Scientific Instruments, 1983, 54(6), p.
744-747.
247. Gustafsson, S.E., Karawacki, E., and Chohan, M.A., Thermal transport studies of
electrically conducting materials using the transient hot-strip technique, Journal of Physics
D, 1986, 19, p. 727-735.
248. Gustafsson, S.E., Transient Plane Source Techniques for Thermal Conductivity and
Thermal Diffusivity Measurements of Solid Materials, Review of Scientific Instruments,
1990, 62(3), p. 797-804.
249. Gustafsson, M., Karawacki, E., and Gustafsson, S.E., Thermal conductivity, thermal
diffusivity, and specific heat of t in samples from transient measurements with hot disk
sensors, Review of Scientific Instruments, 1994, 65, p. 3856-3859.
250. Parker, W.J., Jenkins, R.J., Butler, C.P., and Abbott, G.L., Flash Method of Determining
Thermal Diffusivity, Heat Capacity and Thermal Conductivity, Journal of Applied Physics,
1961, 32, p. 1679-1684.
251. Cape, J.A. and Lehman, G.W., Tempeature and Finite Pulse-Time Effects in the Flash
Method for Measuring Thermal Diffusivity, Journal of Applied Physics, 1963, 34(7), p.
1909-1913.
252. Clark, L.M. and Taylor, R.E., Finite Pulse Time Effects in the Flash Diffusivity Method,
Journal of Applied Physics, 1975, 46, p. 714-719.
253. Mardolcar, U.V. and Nieto de Castro, C.A., The Measurement of Thermal Conductivity at
High Temperatures, High Temperature - High Pressures, 1992, 24, p. 551-580.
254. Azzopardi, A., Mevrel, R., Saint-Ramond, B., Olson, E., and Stiller, K., Influence of aging
on structure and thermal conductivity of Y- PSZ and Y-FSZ EB-PVD coatings, Surface &
Coatings Technology, 2004, 177, p. 131-139.
255. Raghavan, S. , Wang, H. , Dinwiddie, R. B. , Porter, W. D. , and Mayo, M. J., The Effect of
Grain Size, Porosity and Yttria Content on the Thermal Conductivity of Nanocrystalline
Zirconia, Scripta Materialia, 1998, 39(8), p. 1119-1125.
256. Alperine, S., Derrien, M., Jaslier, Y., and Mevrel, R., thermal Barrier Coatings: the
Thermal Conductivity challenge, in AGARD-R-823, 1997, NASA: Aalborg, Denmark, p. 1.
257. Zhu, D. M., Miller, R. A., Nagaraj, B. A., and Bruce, R. W., Thermal conductivity of EB-
PVD thermal barrier coatings evaluated by a steady-state laser heat flow technique, Surface
& Coatings Technology, 2001, 138(1), p. 1-8.
References

182
258. Slifka, A. J. and Filla, B. J., Thermal Conductivity Measurement of an Electron-Beam
Physical-Vapor-Deposition Coating, Journal of Research of the National Institute of
Standards and Technology, 2003, 108(2), p. 147-150.
259. Trice, R W, Su, Y J, Mawdsley, J R, Faber, K T, De Arellano-Lopez, A R, Wang, H, and
Porter, W D, Effect of Heat Treatment on Phase Stability, Microstructure, and Thermal
Conductivity of Plasma-sprayed YSZ, Journal of Materials Science, 2002, 37(11), p. 2359-
2365.
260. Zhu, D. and Miller, R.A., Thermal-Barrier Coatings for Advanced Gas-Turbine Engines,
MRS Bulletin, 2000, 25(7), p. 43-47.
261. Zhu, D., Miller, R.A., Nagaraj, B.A., and Bruce, R.W., Thermal Conductivity of EB-PVD
Thermal Barrier Coatings Evaluated by a Steady-State Laser Heat Flux Technique, in NASA
TM-2000-210238, 2000, NASA Lewis Research Center: Cleveland, Ohio.
262. Dinwiddie, R.B., Beecher, S.C., Porter, W.D., and Nagaraj, B.A. The Effect of Thermal
Aging on the Thermal Conductivity of Plasma Sprayed and EB-PVD Thermal Barrier
Coatings. in International gas turbine and aeroengine congress & exhibition. 1996,
Birmingham, UK: ASME, p. 96-GT-282.
263. Slifka, A.J., Filla, B.J., Phelps, J.M., Bancke, G., and Berndt, C.C., Thermal Conductivity of
a Zirconia Thermal Barrier Coating, Journal of Thermal Spray Technology, 1998, 7(1), p.
43-46.
264. Dutton, R., Wheeler, R., Ravichandran, K.S., and An, K., Effect of Heat Treatment on the
Thermal Conductivity of Plasma-Sprayed Thermal Barrier Coatings, Journal of Thermal
Spray Technology, 2000, 9(2), p. 204-209.
265. Taylor, R.E., Thermal Conductivity Determinations of Thermal Barrier Coatings, Materials
Science and Engineering, 1998, 245, p. 160-167.
266. Lawson, K.L, Nicholls, J.R., and Rickerby, D.S., Thermal conductivity and Ceramic
Microstructure, in High Temperature Surface Engineering, edited by Nicholls, J.R. and
Rickerby, D.S., 1997, Edinburgh:The Institute of Materials: London, UK. p. 8.
267. Gu, S., Lu, T.J., Hass, D.D., and Wadley, H.N.G., Thermal Conductivity of Zirconia
Coatings with Zig-Zag Pore Microstructures, Acta Materialia, 2001, 49, p. 2539-2547.
268. Nicholls, J. R., Lawson, K. J., Johnstone, A., and Rickerby, D. S., Methods to Reduce the
Thermal Conductivity of EB-PVD TBCs, Surface & Coatings Technology, 2002, 151, p.
383-391.
269. Singh, J., Wolfe, D.E., and Singh, J., Architecture of TBCs Produced by Electron Beam-
Physical Vapor Deposition (EB-PVD), Journal of Materials Science, 2002, 37(15), p. 3261-
3268.
270. Nicholls, J. R., Lawson, K. J., Johnstone, A., and Rickerby, D. S., Low Thermal
Conductivity EB-PVD Thermal Barrier Coatings, Materials Science Forum, 2001, 369-372,
p. 595-606.
271. Clarke, D.R., Materials Selection Guidelines for Low Thermal Conductivity Barrier
Coatings, Surface & Coatings Technology, 2003, 163-164, p. 67-74.
272. Raghavan, S., Wang, H., Dinwiddie, R. B., Porter, W. D., Vassen, R., Stover, D., and Mayo,
M. J., Ta
2
O
5
/Nb
2
O
5
and Y
2
O
3
co-doped zirconias for thermal barrier coatings, Journal of
the American Ceramic Society, 2004, 87(3), p. 431-437.
References

183
273. Schulz, U., Saruhan, B., Fritscher, K., and Leyens, C., Review on Advanced EB-PVD
Ceramic Topcoats for TBC Applications, International Journal of Applied Ceramic
Technology, 2004, 1(4), p. 302-315.
274. Matsumoto, M., Aoyama, K., Matsubara, H., Takayama, K, Banno, T., Kagiya Y., and
Sugita, Y., Thermal conductivity and oahse stability of plasma sprayed ZrO
2
-Y
2
O
3
-La
2
O
3
coatings, Surface & Coatings Technology, 2005, 194, p. 31-35.
275. Raghavan, S., Wang, H., Porter, W. D., Dinwiddie, R. B., and Mayo, M. J., Thermal
properties of zirconia co-doped with trivalent and pentavalent oxides, Acta Materialia, 2001,
49(1), p. 169-179.
276. Zhu, D. and Miller, D A, Development of Advanced Low Thermal Conductivity Thermal
Barrier Coatings, in NASA/TM-2004-212961, 2004, Glenn Research Center: Cleveland,
Ohio.
277. Zhu, D. and Miller, R.A., Thermal Conductivity of Advanced Ceramic Thermal Barrier
Coatings Determined by a Steady-State Laser Heat-Flux Approach, in NASA/TM-2004-
213040, 2004, Glenn Research Center: Cleveland Ohio.
278. Singh, J., Wolfe, D.E., Miller, D.A., Eldridge, J.I., and Zhu, D., Tailored microstructure of
zirconia and hafnia-based thermal barrier coatings with low thermal conductivity and high
hemispherical reflectance by EB-PVD, Journal of Material Science, 2004, 36(6), p. 1975-
1985.
279. Lehmann, H., Pitzer, D., Pracht, G., Vassen, R., and Stover, D., Thermal conductivity and
thermal expansion coefficients of the lanthanum rare-earth-element zirconate system,
Journal of the American Ceramic Society, 2003, 86(8), p. 1338-1344.
280. Vassen, R., Traeger, F., and Stover, D., New Thermal Barrier Coatings Based on
Pyrochlore/ YSZ Double-Layer Systems, International Journal of Applied Ceramic
Technology, 2004, 1(4).
281. Clyne, T. W. and Withers, P. J., An Introduction to Metal Matrix Composites. Cambridge
Solid State Science Series, ed. Davis, E. and Ward, I. 1993, Cambridge: Cambridge
University Press.
282. Sevostianov, I. and Kachanov, M., Anisotropic thermal conductivities of plasma-sprayed
thermal barrier coatings in relation to the microstructure, Journal of Thermal Spray
Technology, 2000, 9(4), p. 478-482.
283. Shafiro, B. and Kachanov, M., Anisotropic effective conductivity of materials with
nonrandomly oriented inclusions of diverse ellipsoidal shapes, Journal of Applied Physics,
2000, 87(12), p. 8561-8569.
284. Tzou, D. Y., The Effect of Internal Heat-Transfer in Cavities on the Overall Thermal-
Conductivity, International Journal of Heat and Mass Transfer, 1991, 34(7), p. 1839-1846.
285. Hashin, Z. and Shtrikman, S., A Variational Approach to the Theory of the Elastic
Behaviour of Multiphase Materials, Journal of Mechanics and Physics of Solids, 1963, 11, p.
127-140.
286. Bauer, T. H., A General Analytical Approach Toward the Thermal Conductivity of Porous
Media, International Journal of Heat and Mass Transfer, 1993, 36(17), p. 4181-4191.
287. Bristow, J.R., Microcracks, and the static and dynamic elastic constants of annealed and
heavily cold-worked metals, British Journal of Applied Physics, 1960, 11(2), p. 81-85.
References

184
288. McPherson, R., A Model for the Thermal Conductivity of Plasma-Sprayed Ceramic
Coatings, Thin Solid Films, 1984, 112, p. 89-95.
289. Li, C. J. and Ohmori, A., Relationships between the microstructure and properties of
thermally sprayed deposits, Journal of Thermal Spray Technology, 2002, 11(3), p. 365-374.
290. Boirelavigne, S., Moreau, C., and Saintjacques, R. G., The Relationship between the
Microstructure and Thermal- Diffusivity of Plasma-Sprayed Tungsten Coatings, Journal of
Thermal Spray Technology, 1995, 4(3), p. 261-267.
291. Lu, T. J. and Hutchinson, J. W., Thermal-Conductivity and Expansion of Cross-Ply
Composites with Matrix Cracks, Journal of the Mechanics and Physics of Solids, 1995,
43(8), p. 1175-1198.
292. Lu, T. J., Levi, C. G., Wadley, H. N. G., and Evans, A. G., Distributed Porosity as a Control
Parameter for Oxide Thermal Barriers made by Physical Vapor Deposition, Journal of the
American Ceramic Society, 2001, 84, p. 2937-2946.
293. Golosnoy, I.O., Tsipas, S.A., and Clyne, T.W., An Analytical Model for Simulation of Heat
Flow in Plasma Sprayed Thermal Barrier Coatings, Journal of Thermal Spray Technology,
2005, 14(2), p. in press.
294. Ravichandran, K.S., An, K., Dutton, R.E., and Semiatin, S.L., Thermal Conductivity of
Plasma-Sprayed Monolithic and Multilayer Coatings of Alumina and Yttria-Stabilized
Zirconia, Journal of the American Ceramic Society, 1999, 83(3), p. 673-682.
295. Lu, T. J., Levi, C. G., Wadley, H. N. G., and Evans, A. G., Distributed porosity as a control
parameter for oxide thermal barriers made by physical vapor deposition, Journal of the
American Ceramic Society, 2001, 84(12), p. 2937-2946.
296. Incropera, F.P. and Dewitt, D.P., Introduction to Heat Transfer. 3rd ed. 1996, NY, USA:
John Wiley & Sons, Inc.
297. Makino, T., Kunitomo, T., Sakai, I., and Kinoshita, H., Thermal Radiation Properties of
Ceramic Materials, Heat Transfer. Japanese Research., 1984, 13(4), p. 33-50.
298. Ferriere, A., Lestrade, L., and Robert, J. F., Optical properties of plasma-sprayed ZrO
2
-Y
2
O
3

at high temperature for solar applications, Journal of Solar Energy Engineering-
Transactions of the ASME, 2000, 122(1), p. 9-13.
299. Eldridge, J.I., Spuckler, C.M., Street, K.W., and Markham, J.R. Infrared Radiative
Properties of Yttria-Stabilized Zirconia Thermal Barrier Coatings. in Ceramic and
Engineering Science Procedings. 2002, Cocoa Beach, Florida, USA: The American Ceramic
Society, p. 417-430.
300. Samarskii, A.A. and Vabishevich, P.N., Computational heat transfer. V.1. Mathematical
Modelling. 1995, Chichester: Wiley. 406.
301. Tannehill, J.C., Anderson, D.A., and Pletcher, R.H., Computational Fluid Mechanics and
Heat Transfer. 2nd ed. 1997, Washington, London: Taylor & Francis. 792.
302. Ilavsky, J., Allen, A.J., Long, G.G., Krueger, S., Berndt, C.C., and Herman, H., Influence of
Spray Angle on the Pore and Crack Microstructure of Plasma-Sprayed Deposits, Journal of
the American Ceramic Society, 1997, 80(3), p. 733-742.
303. Trapaga, G., Matthys, E.F., Valencia, J.J., and Szekely, J., Fluid Flow, Heat Transfer and
Solidification of Molten Metal Droplets Impinging on Substrates; Comparison of Numerical
and Experimental Results, Metallurgical Transactions, 1992, 23B, p. 701-718.
References

185
304. Zhang, H., Theoretical analysis of Spreading and Solidification of Molten Droplet during
Thermal Spray Deposition, int. J. Heat Mass Transfer, 1999, 42, p. 2499-2508.
305. Bennett, T. and Poulikakos, D., Splat-quench Solidification: Estimating the Maximum
Spreading of a Droplet impacting a Solid Surface, Journal of Materials Science, 1993, 28, p.
963-970.
306. Scott, H.G., Phase Relationships in the ZrO
2
-Y
2
O
3
System, Journal of Materials Science,
1975, 10, p. 1527-1535.
307. Tani, E., Yoshimure, M., and Somiya, S., Revised phase-diagram of the system ZrO
2
-CeO
2

below 1400
o
C, Journal of the American Ceramic Society, 1983, 66(7), p. 506-510.
308. Din, S. U. and Kaleem, A., Vickers hardness study of zirconia partially stabilized with
lanthanide group oxides, Materials Chemistry and Physics, 1998, 53(1), p. 48-54.
309. Thompson, J.A., Thermomechanical Behaviour of Plasma-Sprayed Thermal Barrier
Coatings, in Department of Materials Science and Metallurgy, 2000, University of
Cambridge: Cambridge, UK.
310. Toraya, H., Yoshimura, M., and Somiya, S., Calibration Curve for Quantitative Analysis of
the Monoclinic-Tetragonal ZrO
2
System by X-ray Diffraction, Journal of the American
Ceramic Society, 1984, p. C119-121.
311. Toraya, H., Yoshimura, M., and Somiya, S., Quantitative Analysis of Monoclinic -
Stabilized Cubic ZrO
2
Systems by X-ray Diffraction, Journal of the American Ceramic
Society, 1984, 67, p. C183-184.
312. Wiles, D.B. and Young, R.A., A new computer program for Rietveld analysis of X-ray
powder diffraction Patterns, Journal of Applied Crystallography, 1981, 14, p. 149-151.
313. Murphy, A.M., Clustering in Particulate Metal Matrix Composites, PhD thesis, in Materials
Science and Metallurgy, 1997, University of Cambridge: Cambridge, UK.
314. Chudoba, T. and Richter, F., Investigation of Creep Behaviour Under Load During
Indentation Experiments and its Influence on Hardness and Modulus Results, Surface &
Coatings Technology, 2001, 148, p. 191-198.
315. Oliver, W C and Pharr, G M, An Improved Technique for Determining Hardness and Elastic
Modulus Using Load and Displacement Sensing Indentation Experiments, Journal of
Materials Research, 1992, 7(6), p. 1564-1583.
316. Nimonic Alloy 80A Material Datasheet: Physical and Mechanical Properties, Special
Metals Corporation: Hereford, UK.
317. Heraeus Quarzglas - Quartz Glass for Optics: Data and Properties, 2004, Heraeus:
Germany.
318. Incropera, F.P. and Dewitz, D.P., Fundamentals of Heat and Mass Transfer. Fifth ed. 2002:
John Wiley & Sons.
319. Miller, R., Garlick, R., and Smialek, J., Phase Distributions in Plasma-Sprayed Zirconia-
Yttria, Ceramic Bulletin, 1983, 62, p. 1355-1358.
320. Heuer, A H. and Ruhle, M. Phase transformations in ZrO2-containing ceramics: I. the
instability of c-ZrO2 and the resulting diffusion-controlled reactions. in Advances in
ceramics: Science and technology of zirconia II. 1983, edited by Claussen N., Ruhle, M.,
and Heuer, A.H., Stuttgard, Germany: American Ceramic Society, Inc., p. 1-13.
References

186
321. Heuer, A.H., Chaim, R., and Lanteri, V., The Displacive Cubic Tetragonal Transformation
in ZrO
2
Alloys, Acta Metallurgica, 1987, 35, p. 661-666.
322. Lanteri, V., Heuer, A.H., and Mitchell, T.E. Tetragonal Phase in the System ZrO
2
-Y
2
O
3
. in
Second International Conference on the Science and Technology of Zirconia. 1983, edited
by Claussen, N., Ruhle, M., and Heuer, A.H., Stuttgart: American Ceramic Society, p. 118-
131.
323. Lakiza, S.M. and Lopato, L.M., Stable and metastable phase relations the system alumina-
zirconia-yttria, Journal of the American Ceramic Society, 1997, 80(4), p. 893-902.
324. Stough, M.A., Solid solubility and precipitation in a single-crystal alumina-zirconia system,
Journal of the American Ceramic Society, 2002, 85(12), p. 2895-2902.
325. Li, M.J., Feng, Z.C., Ying, P.L., Xin, Q., and Li, C., Phase transformation in the surface
region of zirconia and doped zirconia detected by UV Raman spectroscopy, Physical
Chemistry Chemical Physics, 2003, 5(23), p. 5326-5332.
326. Shannon, R.D., Revised effective ionic radii and systematic studies of interatomic distances
in halides and chalcogenides, Acta Crystallographica, 1976, A32, p. 751-767.
327. Langjahr, P.A., Oberacker, R., and Hoffmann, M.J., Long-term behaviour and application
limits of plasma sprayed zirconia thermal barrier coatings, Journal of the American
Ceramic Society, 2001, 84(6), p. 1301-1308.
328. Brandon, J. and R.Taylor, Phase Stability of Zirconia-Based Thermal Barrier Coatings Part
I. Zirconia-Yttria alloys, Surface & Coatings Technology, 1991, 46, p. 75-90.
329. Michel, D., Perezyjo.M, and Collongu.R, Study of Order-Disorder Transformation of
Fluorite Structure with Pyrochlore Type Phases of Formula (1-X) ZrO
2
-XLn
2
O
3
, Materials
Research Bulletin, 1974, 9(11), p. 1457-1468.
330. Jang J.W., Kim, D.J., and Lee, D.Y., Size effect of trivalent oxides on low temperature phase
stability of 2Y-TZP, Journal of Materials Science, 2001, 36(22), p. 5391-5395.
331. Vassen, R., Dietrich, M., Lehmann, H., Cao, X., Pracht, G., Tietz, F., Pitzer, D., and Stover,
D., Development of oxide ceramics for an application as TBC, Materialwissenschaft Und
Werkstofftechnik, 2001, 32(8), p. 673-677.
332. Cao, X. Q., Vassen, R., Jungen, W., Schwartz, S., Tietz, F., and Stover, D., Thermal
stability of lanthanum zirconate plasma-sprayed coating, Journal of the American Ceramic
Society, 2001, 84(9), p. 2086-2090.
333. Anthony, A.M. Structure of point defects in ionic materials. in Sintering and Related
Phenomena. 1972, edited by Kuczynski, G.C., Indiana: Plenum press, New York, London.
334. Kroger F.A., The Chemistry of Imperfect Crystals. 1964, Amsterdam, New York: North
Holland Publ. Co. and Wiley.
335. Tilley, R.J.D., Defect Crystal Chemistry and its Applications. 1987: Blackie & Son Ltd.
336. Borg, R.J. and Dienes, G.J., An Introduction to Solid State Diffusion. 1988: Academic Press,
Inc.
337. Weber, S., Scherrer, S., Scherrer, H., Kilo, M., Taylor, M.A., and Borchardt, G., SIMS
analysis of multi-diffusion profiles of lanthanides in stabilzed zirconias, Applied Surface
Science, 2003, 203-204, p. 656-659.
References

187
338. Chaim, R., Basat, G., and Kats-Demyanets, A., Effect of oxide additives on grain growth
during sintering of nanocrystalline zirconia alloys, Materials Letters, 1998, 35(3-4), p. 245-
250.
339. Li, P., Chen, I-W., and Penner-Hahn, J.E., X-ray-absorption studies of zirconia polymorphs.
II. Effect of Y
2
O
3
dopant on ZrO
2
structure, Physical Review B, 1993, 48(14), p. 10074-
10081.
340. Li, P., Chen, I-W., and Penner-Hahn, J.E., X-ray-absorption studies of zirconia polymorphs.
I. Characteristic local structures, Physical Review B, 1993, 48(14), p. 10063-10073.
341. Kamigaito, O., Idea strength of brittle materials as a function of their melting or
sublimation points, Journal of Materials Science Letters, 1991, 10, p. 567-568.
342. Tekeli, S., Erdogan, M., and Aktas, B., Influence of alpha-Al
2
O
3
addition on sintering and
grain growth behaviour Of 8 mol% Y
2
O
3
-stabilised cubic zirconia (c-ZrO
2
), Ceramics
International, 2004, 30(8), p. 2203-2209.
343. Matsumoto, M., Yamaguchi, N., and Matsubara, H., Low thermal conductivity and high
temperature stability of ZrO
2
-Y
2
O
3
-La
2
O
3
coatings produced by electron beam PVD, Scripta
Materialia, 2004, 50(6), p. 867-871.
344. Malzbender, J. and Steinbrech, R.W., Mechanical methods to determine layer compliances
within multilayered composites, Journal of Materials Research, 2003, 18(6), p. 1374-1382.
345. Bengtsson, P., Ericsson, T., and Wigren, J., Thermal Shock Testing of Burner Cans Coated
with a Thick Thermal Barrier Coating, Journal of Thermal Spray Technology, 1998, 7(3), p.
340-348.
346. Guo, H.B., Vassen, R., and Stover, D., Thermophysical properties and thermal cycling
behaviour of plasma sprayed thick thermal barrier coatings, Surface & Coatings
Technology, 2005, 192, p. 48-56.
347. Hutchinson, J.W. and Suo, Z., Mixed Mode Cracking in Layered Materials, Advanced
Applied Mechanics., 1991, 29, p. 63-191.
348. Bengtsson, P. and Johannesson, T., Characterization of Microstructural Defects in Plasma-
Sprayed Thermal Barrier Coatings, Journal of Thermal Spray Technology, 1995, 4(3), p.
245-251.
349. Sampath, S., Jiang, X.Y., Matejicek, J., Leger, A.C., and Vardelle, A., Substrate
temparature effects on splat formation, microsture devlopment and properties of plasma
sprayed coatings Part I: Case study for partially stabilized zirconia, Materials Science and
Engineering, 1999, A272, p. 181-188.
350. Sampath, S. and Jiang, X.Y., Spalta formation and microstructure developemtn during
plasma spraying: deposition temperature effects, Materials Science and Engineering, 2001,
A304-306, p. 144-150.
351. Gill, S.C. and Clyne, T.W., Stress Distributions and Material Response in Thermal Spraying
of Metallic and Ceramic Deposits, Met. Trans., 1990, 21B, p. 377-385.
352. Clyne, T.W. and Gill, S.C., Heat Flow and Thermal Contraction during Plasma Spray
Deposition, in Heat Transfer in Manufacturing and Processing of New Materials, edited by
Tanasawa, I., 1991, Hemisphere: New York. p. 33-48.
353. Tsui, Y.C., Thompson, J.A., and Clyne, T.W. The Effect of Bond Coat Creep on Residual
Stresses and Debonding in Plasma Sprayed Thermal Barrier Systems. in Thermal Spray:
References

188
Meeting the Challenges of the 21st Century. Proceedings of the 15th Int. Thermal Spray
Conf. 1998, edited by Coddet, C., Nice, France: ASM International, p. 1565-1570.
354. Teixeira, V., Andritschky, M., Fischer, W., Buchkremer, H. P., and Stover, D., Effects of
deposition temperature and thermal cycling on residual stress state in zirconia-based
thermal barrier coatings, Surface & Coatings Technology, 1999, 121, p. 103-111.
355. Takeuchi, S., Ito, M., and Takeda, K., Modelling of Residual Stress in Plasma-Sprayed
Coatings: Effect of Substrate Temperature, Surface & Coatings Technology, 1990, 43/44, p.
426-435.
356. Stover, D., Pracht, G., Lehmann, H., Dietrich, M., Doring, J. E., and Vassen, R., New
Material Concepts for the Next Generation of Plasma Sprayed Thermal Barrier Coatings,
Journal of Thermal Spray Technology, 2004, 13(1), p. 76-83.
357. Miller, R.A., Current Status of Thermal Barrier Coatings - An Overview, Surface &
Coatings Technology, 1987, 30, p. 1-11.
358. Wang, H. and Dinwiddie, R.B., Reliability of the Laser Flash Thermal Diffusivity
Measurements of the Thermal Barrier Coatings, Journal of Thermal Spray Technology,
2000, 9(2), p. 210-214.
359. Slifka, A. J., Thermal Conductivity Apparatus for Steady-State, Comparative Measurement
of Ceramic Coatings, Journal of Research of the National Institute of Standards and
Technology, 2000, 105(4), p. 591-605.
360. Taylor, R., Construction of Apparatus for Heat Pulse Thermal Diffusivity Measurements
from 300-3000K, Journal of Physics E: Scienctific Instruments, 1980, 13, p. 1193-1196.
361. Mirmira, S.R. and Fletcher, L. S., Review of the Thermal Conductivity of Thin Films,
Journal of Thermophysics and Heat Transfer, 1998, 12(2), p. 121-131.
362. Thompson, J.A., Ji, W., and Clyne, T.W., The Effect of Top Coat Sintering on Ceramic
Spallation in Plasma-Sprayed Thermal Barrier Coatings, in 13th Int. Conf. on Surface
Modific. Techn. (SMT XIII), edited by Sudarshan, T.S., Khor, K.A., and Jeandin, M., 1999,
ASM. p. 177-182.
363. Mogro-Campero, A., Johnson, C. A., Bednarczyk, P. J., Dinwiddie, R. B., and Wang, H.,
Effect of Gas Pressure on Thermal Conductivity of Zirconia Thermal Barrier Coatings,
Surface & Coatings Technology, 1997, 94-5(1-3), p. 102-105.
364. Zhu, D., Bansal, N.P., Lee, K.N., and Miller, R.A., Thermal Conductivity of Ceramic
Thermal Barrier and Environmental Barrier Coating Materials, in NASA TM-2001-211122,
2001, NASA Lewis Research Center: Cleveland, Ohio.
365. Pawlowski, L., Lombard, D., and Fauchais, P., Structure-Thermal Properties - Relationship
in Plasma Sprayed Zirconia Coatings, Journal of Vacuum Science & Technology a-Vacuum
Surfaces and Films, 1985, 3(6), p. 2494-2500.
366 Cardarelli, F., Materials Handbook: a concise desktop reference. 2000, London: Springer-
Verlag.