-
A THICK MULTILAYER THERMAL BARRIER COATING:
DESIGN, DEPOSITION, AND INTERNAL STRESSES
Hamed Samadi
A thesis submitted in Conformity with
the requirement for the degree
Doctor of Philosophy
Department of Materials Science and Engineering
University of Toronto
© Copyright by Hamed Samadi 2009
-
II
A Thick Multilayer Thermal Barrier Coating:
Design, Deposition, and Internal Stresses
Hamed Samadi
Doctor of Philosophy
Department of Materials Science and Engineering
University of Toronto
2009
Abstract
Yttria Partially Stabilized Zirconia (Y-PSZ) plasma-sprayed
coatings are widely used in turbine
engines as thermal barrier coatings. However, in diesel engines
Y-PSZ TBCs have not met with
wide success. To reach the desirable temperature of 850-900˚C in
the combustion chamber from
the current temperature of 400-600˚C, a coating with a thickness
of approximately 1mm is
required. This introduces different considerations than in the
case of turbine blade coatings,
which are on the order of 100µm thick. Of the many factors
affecting the durability and failure
mechanism of TBCs, in service and residual stresses play an
especially important role as the
thickness of the coating increases. For decreasing the residual
stress in the system, a multi-layer
coating is helpful. The design of a multilayer coating employing
relatively low cost materials
with complementary thermal properties is described. Numerical
models were used to describe
the residual stress after deposition and under operating
conditions for a multilayer coating that
exhibited the desired temperature gradient. Results showed that
the multilayer coating had a
lower maximum stress under service conditions than a
conventional Y-PSZ coating. Model
validation with experiments showed a good match between the
two.
-
III
Acknowledgments
In His name…,
A journey is easier when you travel together. Interdependence is
certainly more valuable than
independence. This thesis is the result of five years of work
whereby I have been accompanied
and supported by many people. It is a pleasant aspect that I
have now the opportunity to express
my gratitude for all of them.
Firstly, I would like to thank my Supervisor, Prof. Tom Coyle. I
could not have imagined
having a better advisor and mentor for my PhD, and without his
common sense, knowledge and
perceptiveness, I would never have finished. Thank-you to my
dissertation committee, Prof.
Utigard, Prof. Hibbard, Prof. Kesler, Prof. Mostaghimi and Dr.
Marple for managing to read the
whole thing so thoroughly, and for a surprisingly enjoyable
viva. I would also like to thank all
the rest of the academic and support staff of the Centre for
Advanced Coating Technologies,
particularly those who have put up with my drifting a long way
away from my original title.
Much respect to my officemates, and hopefully still friends,
Hanif, Fardad, Reza, Hamid,
Mehdi, Ala, Amir, Bob, Ben, Ken, Arash, Tommy, Rajeev, Babak,
Sanaz and Nikoo for putting
up with me for more than five years. Thanks to Prof. Mostaghimi,
Dr. Larry Pershin, Dr.
Eugenio García, Dr. Salimijazi and everyone else for all that
serious discussion and helps in
running the equipment. In addition, thanks to Sal Boccia, who
trained and helped me during use
of SEM facilities in the Department of Materials Science and
Engineering. Also thanks to those
people and groups who provided me with so much whether they know
it or not.
On a different note, I would like to thank the coffee club of
CACT for keeping me thinking (you
have to ask why?), medical researches for making me feel okay
about drinking coffee and
traditional Persian alongside classical music for keeping me
sane.
Finally, I have to say 'thank-you' to: all my friends and
family, wherever they are, particularly
my Mom and Dad and my in-laws; and, most importantly of all, to
Narges and Amir Mahdi, my
beloved family, for everything. Without their encouragement and
understanding, it would have
been impossible for me to finish this work.
Is that everyone?
Hamed, July 1st 2009.
-
IV
To my wife, Narges, who made me a
very happy man…
Her love has made me wealthy beyond
my dreams.
-
V
Table of Contents
Abstract..........................................................................................................................................II
Acknowledgments
.......................................................................................................................
III
List of Tables
.................................................................................................................................X
List of
Figures..............................................................................................................................
XI
1. Chapter I: Introduction
..........................................................................................................
1
1.1. Coating
techniques..........................................................................................................
1
1.2. Thermal
spraying............................................................................................................
1
1.3. Application of coatings in diesel engines
.......................................................................
2
1.4. Diesel engines and turbine engines: similarities and
differences ................................... 2
1.5. Thesis objective
..............................................................................................................
3
1.6. Structure of the thesis document
....................................................................................
4
2. Chapter II: Thermal Spray
Process........................................................................................
5
2.1. Introduction
....................................................................................................................
5
2.2. Plasma
spraying..............................................................................................................
6
2.3. Process
parameters..........................................................................................................
7
2.3.1. Plasma current
................................................................................................................
8
2.3.2. Primary gas flow rate:
argon...........................................................................................
8
2.3.3. Auxiliary gas flow rate: hydrogen
..................................................................................
9
2.3.4. Carrier gas flow rate
.....................................................................................................
10
2.4. Feedstock injection
.......................................................................................................
13
2.5. Particle-particle interactions
.........................................................................................
14
2.6. Particle
trajectory..........................................................................................................
15
2.7. Particle- substrate
interactions......................................................................................
16
-
VI
2.8. Coating
formation.........................................................................................................
20
2.9.
Summary.......................................................................................................................
22
3. Chapter III: Coating
Materials.............................................................................................
23
3.1. Introduction
..................................................................................................................
23
3.2. Failure mechanisms in
TBCs........................................................................................
24
3.3.
Alternatives...................................................................................................................
27
3.3.1. Zirconates
.....................................................................................................................
28
3.3.2.
Garnets..........................................................................................................................
28
3.3.3. Al2O3.SiO2.MgO
system...............................................................................................
29
3.3.3.1. Cordierite
...................................................................................................................
30
3.3.3.2. Forsterite
....................................................................................................................
31
3.3.3.3. Spinel
.........................................................................................................................
32
3.3.3.4. Mullite
.......................................................................................................................
34
3.3.4. Multilayer system
.........................................................................................................
35
3.3.5. Phase stability
...............................................................................................................
37
3.4.
Summary.......................................................................................................................
40
4. Chapter IV: Coating Deposition
..........................................................................................
41
4.1. Introduction
..................................................................................................................
41
4.2. Feedstock powder
.........................................................................................................
41
4.2.1. Morphology
..................................................................................................................
41
4.2.2. Size
distribution............................................................................................................
43
4.2.3. X-ray diffraction and phase analysis
............................................................................
45
4.3. Optimization of deposition
parameters.........................................................................
48
4.3.1. Measurement of in-flight particle temperature and
velocity ........................................ 50
4.3.2. Statistical design of experiments
..................................................................................
52
-
VII
4.4. Analysis of the results of the Taguchi design of
experiments ...................................... 55
4.5. Coating
deposition........................................................................................................
60
4.6. Characterization of deposits
.........................................................................................
63
4.6.1. Sample preparation
.......................................................................................................
63
4.6.2. Coating microstructure
.................................................................................................
64
4.6.3. Porosity
measurement...................................................................................................
67
4.6.4. Crystallinity index
........................................................................................................
70
4.6.5. Thermal diffusivity and conductivity
...........................................................................
72
4.6.6. Elastic modulus measurement
......................................................................................
72
4.7. Single splat
collection...................................................................................................
74
4.8. Results and discussion
..................................................................................................
77
4.8.1. Deposition
efficiency....................................................................................................
77
4.8.2. Porosity
.........................................................................................................................
78
4.8.3. Crystallinity index
........................................................................................................
80
4.8.4. Single splat
collection...................................................................................................
81
4.8.5. Optimization of plasma parameters
..............................................................................
84
4.8.6. Thermal diffusivity and conductivity
...........................................................................
85
4.8.7. Modulus of elasticity
....................................................................................................
87
4.9.
Summary.......................................................................................................................
88
5. Chapter V: Residual Stress
..................................................................................................
89
5.1. Introduction
..................................................................................................................
89
5.2. Quenching
stress...........................................................................................................
89
5.3. Thermal
stresses............................................................................................................
92
5.4. Total residual stress
......................................................................................................
93
5.5. Stress
relaxation............................................................................................................
94
-
VIII
5.6. Residual stress determination
.......................................................................................
96
5.6.1. Hole
drilling..................................................................................................................
96
5.6.2. X-ray diffraction
...........................................................................................................
99
5.6.3. Synchrotron
................................................................................................................
101
5.6.4. Neutron diffraction
.....................................................................................................
101
5.6.5. Curvature and layer removal
......................................................................................
102
5.6.6. Magnetic method
........................................................................................................
104
5.6.7. Ultrasonic methods
.....................................................................................................
104
5.6.8. Piezo-spectroscopic (Raman)
.....................................................................................
105
5.6.9. Comparison between techniques
................................................................................
105
5.7. Curvature measurement
..............................................................................................
108
5.7.1. Setup
...........................................................................................................................
108
5.7.2. Calculating curvature from the displacement
.............................................................
110
5.7.3. Results and discussion
................................................................................................
111
5.8.
Summary.....................................................................................................................
115
6. Chapter VI: Numerical Modeling of Curvature and In-plane
Stresses.............................. 116
6.1. Introduction
................................................................................................................
116
6.2. Previous
models..........................................................................................................
116
6.3. The numerical model
..................................................................................................
119
6.3.1. Model description
.......................................................................................................
120
6.3.2. Material
properties......................................................................................................
121
6.3.3. Heat transfer boundary
conditions..............................................................................
122
6.3.4. Mechanical boundary
conditions................................................................................
123
6.3.5. Meshing
......................................................................................................................
124
6.4. Results and discussion
................................................................................................
124
-
IX
6.4.1. Thermal
behaviour......................................................................................................
124
6.4.2. In-situ curvature
calculations......................................................................................
127
6.5. Comparison with existing models
..............................................................................
129
6.6. Internal stress under service
conditions......................................................................
133
6.7.
Summary.....................................................................................................................
137
7. Chapter VII: Conclusion and Suggestions for Further Work
............................................ 138
7.1.
Summary.........................................................................................................................
138
7.2. Original
contributions.....................................................................................................
138
7.3. Suggestions for future
work............................................................................................
139
8. Appendices
........................................................................................................................
140
8.1. Appendix A: Cost
calculation.....................................................................................
140
8.2. Appendix B: Effect of in-flight particle properties on
deposition of air plasma sprayed
forsterite.................................................................................................................................
141
8.2.1. Introduction
................................................................................................................
141
8.2.2. Experimental
procedure..............................................................................................
142
8.2.2.1. Powder
.....................................................................................................................
142
8.2.2.2. Coating
deposition...................................................................................................
143
8.2.2.3. In-flight particle properties
......................................................................................
145
8.2.3. Results and discussion
................................................................................................
145
8.2.4. Conclusion
..................................................................................................................
150
References..................................................................................................................................
151
-
X
List of Tables
Table 2-1. Comparison of different spraying methods [2].
............................................................. 6
Table 4-1. The L9 DOE matrix for spinel
....................................................................................
54
Table 4-2. The L9 DOE matrix for mullite and forsterite
............................................................ 54
Table 4-3. ANOVA of in-flight particle velocity for forsterite
................................................... 56
Table 4-4. ANOVA of in-flight particle temperature for
forsterite............................................. 56
Table 4-5. ANOVA of in-flight particle velocity for
spinel........................................................
56
Table 4-6. ANOVA of in-flight particle temperature for
spinel.................................................. 57
Table 4-7. ANOVA of in-flight particle velocity for mullite
...................................................... 57
Table 4-8. ANOVA of in-flight particle temperature for
mullite................................................ 57
Table 4-9. Grinding preparation steps.
........................................................................................
63
Table 4-10. Physical Properties of the
Coatings..........................................................................
80
Table 4-11. Chosen parameters for multilayer material
..............................................................
85
Table 4-12. Modulus of
elasticity................................................................................................
88
Table 5-1. Practical issues with different technique materials
[119]. .......................................... 106
Table 5-2. Materials issues with different techniques [119].
....................................................... 107
Table 5-3. Physical Characteristics
[119].....................................................................................
107
Table 6-1. Material properties used in the model
......................................................................
122
Table 6-2. Values used in heat transfer physics.
.......................................................................
127
Table 8-1. Cost calculations for duplex using SG-100 and
multilayer coatings using SG-100 (for
spinel and mullite) and CACT gun (for forsterite).
...................................................................
140
Table 8-2. Spray parameters
......................................................................................................
144
-
XI
List of Figures
Figure 2.1. Effect of plasma current on: (a) plasma gas
temperature, (b) plasma gas velocity [21].
.......................................................................................................................................................
8
Figure 2.2. Effect of Ar flow rate on: (a) plasma gas
temperature and (b) plasma gas velocity [21].
.......................................................................................................................................................
9
Figure 2.3. Effect of hydrogen flow rate on: (a) plasma gas
temperature , (b) degree of Ar and H
ionization and (c) plasma gas velocity [21].
..................................................................................
10
Figure 2.4. Particle trajectories expected for low and high
powder carrier gas flow rates [23]. ... 11
Figure 2.5. Effect of carrier gas flow rate on (a) density and
(b) deposition efficiency [23]. ....... 12
Figure 2.6. Schematic picture of particle trajectory for
different particle sizes [13]..................... 15
Figure 2.7. Dispersed trajectories of alumina particles of
different sizes in a turbulent free
plasma argon jet
[2].......................................................................................................................
16
Figure 2.8. Schematic deformation of a splat at impact [2].
......................................................... 17
Figure 2.9. Microstructure of fractured as-sprayed partially
stabilized zirconia showing
columnar grains [28].
.....................................................................................................................
17
Figure 2.10. Morphology of different ceramic splats on AISI304
substrate [29].......................... 18
Figure 2.11. Dependence of fraction of disk shape splats on
substrate temperature (substrate:
AISI 304, particle Al2O3) [29].
......................................................................................................
19
Figure 2.12. Dependence of Al2O3 splat morphology on substrate
(AISI 304) temperature [29]. 19
Figure 2.13. Schematic rendering of the chaotic structure of
plasma sprayed coating: 1. Thin
molten shell; 2. Unmelted core; 3. Liquid splash; 4. ‘Pancake’
splat; 5. Interlocked splat; 6.
Oxidized particle; 7. Unmelted particle; 8. Pore; 9. Void; 10.
Roughened substrate; 11.
Substrate[2].
..................................................................................................................................
20
Figure 2.14. Dependence of deposition efficiency on various
plasma spray parameters [2]........ 21
Figure 3.1. Air Plasma-Sprayed (APS) coating [42].
....................................................................
23
Figure 3.2. SEM image of plasma sprayed mullite coatings
....................................................... 25
-
XII
Figure 3.3. Burner rig showing four-specimen rotating carousel
[23]. ......................................... 27
Figure 3.4. Surface temperature of burner rig specimens during
rig cycle [23]. ........................... 27
Figure 3.5. Experimentally measured thermal conductivity of
different garnets [55]................... 29
Figure 3.6. Al2O3.SiO2.MgO phase diagram
[56]..........................................................................
30
Figure 3.7. SEM micrograph of plasma sprayed forsterite on
80Ni-20Cr bond coat and mild
steel substrate after 719 thermal
cycles[57]...................................................................................
31
Figure 3.8. As-sprayed forsterite
[57]............................................................................................
32
Figure 3.9. Crystal structure of spinel [64].
...................................................................................
33
Figure 3.10. Phase diagram of MgO-Al2O3 [56].
..........................................................................
33
Figure 3.11. Spinel coating after 719 thermal cycles [57].
............................................................ 34
Figure 3.12. Binary phase diagrams of Al2O3-SiO2 [56].
..............................................................
34
Figure 3.13. DTA of plasma sprayed mullite; the
recrystallization temperature is shown as an
exothermic peak [78].
....................................................................................................................
35
Figure 3.14. Calculated surface stresses of zirconia and mullite
during cooling following the
establishment of steady state temperature and stress
distributions under exposure to a heat flux
of 270 kW.m-2
[15].........................................................................................................................
36
Figure 3.15. XRD of mullite - forsterite mixture (JCPDS card
numbers 85-1358, 79-1454 and
02-0646).......................................................................................................................................
38
Figure 3.16. Mullite - forsterite sample after
heating..................................................................
38
Figure 3.17. XRD pattern for forsterite-spinel heated sample
(JCPDS card numbers 85-1358,
73-1959 and
87-0653)..................................................................................................................
39
Figure 3.18. XRD pattern for forsterite-spinel heated sample
(JCPDS card numbers 79-1454,
73-1959 and
11-0607)..................................................................................................................
39
Figure 3.19. Spinel-forsterite (left) and spinel-mullite (right)
samples after heating.................. 40
Figure 4.1. SEM micrographs of the starting powders: forsterite
(top left), spinel (top right),
mullite (bottom).
..........................................................................................................................
42
-
XIII
Figure 4.2. SEM micrographs of cross sections of starting
powders: forsterite (top left), spinel
(top right) and mullite
(bottom)...................................................................................................
43
Figure 4.3. Particle size analysis of mullite sample. X axis is
logarithmic. ................................ 44
Figure 4.4. Particle size analysis of powders. X axis is
logarithmical. ....................................... 44
Figure 4.5. Binary phase diagrams of Al2O3-SiO2 (top left),
MgO-SiO2 (top right) and MgO-
Al2O3 (bottom) [56].
......................................................................................................................
46
Figure 4.6. XRD result for forsterite powder (JCPDS card number
85-1358)............................ 47
Figure 4.7. XRD result for spinel powder (JCPDS card number
73-1959)................................. 47
Figure 4.8. XRD result for mullite powder (JCPDS card number
79-1454)............................... 48
Figure 4.9. Argon concentration isocontours (current = 450
A)[80]............................................. 49
Figure 4.10. Alternatives in tailoring the
properties...................................................................
49
Figure 4.11. (a) DPV-2000 set up and (b) DPV-2000 particle
detector mechanism................. 51
Figure 4.12. Typical DPV-2000
results.......................................................................................
52
Figure 4.13. Factors with the highest influence on forsterite
in-flight particles: (a) velocity, (b)
temperature.
.................................................................................................................................
58
Figure 4.14. Results of ANOVA analysis of the most influential
process parameters on the in-
flight characteristics of spinel particles: (a) velocity, (b)
temperature. ....................................... 59
Figure 4.15. Factors with the highest influence on mullite
in-flight particles: (a) velocity, (b)
temperature.
.................................................................................................................................
60
Figure 4.16. Spraying setup.
........................................................................................................
61
Figure 4.17. Spraying pattern.
.....................................................................................................
62
Figure 4.18. Substrate temperature measured at the back surface
of the substrate vs. time
(mullite sample). The lower curve is for a non-preheated
substrate, while the upper curve was
collected for a substrate preheated to ~300°C before deposition.
............................................... 62
Figure 4.19. SEM micrograph of a cross section of a forsterite
coating. .................................... 64
Figure 4.20. SEM micrograph of a cross section of a spinel
coating. ......................................... 65
-
XIV
Figure 4.21. SEM micrograph of a cross section of a mullite
coating. ....................................... 65
Figure 4.22. XRD pattern for spinel. Powder (bottom), coating
(top) (JCPDS card numbers 73-
1959 and
88-0826).......................................................................................................................
66
Figure 4.23. XRD pattern for mullite. Powder (bottom), coating
(top) (JCPDS card numbers 79-
1454 and
88-0826).......................................................................................................................
66
Figure 4.24. Effect of magnification and number of fields of
view on (a) porosity and (b)
measurement variability of plasma sprayed Al2O3 [97].
...............................................................
68
Figure 4.25. A typical SEM micrograph
.....................................................................................
69
Figure 4.26. A typical micrograph used for image analysis.
....................................................... 69
Figure 4.27. Grey threshold
.........................................................................................................
70
Figure 4.28. Main peak of a mullite sample.
...............................................................................
71
Figure 4.29. XRD patterns of the starting powder (left) and a
typical coated sample (right)
(JCPDS card
79-1454).................................................................................................................
71
Figure 4.30. Four-point bending test assembly for modulus of
elasticity measurement. Strain
gauge is located on the bottom of the
sample..............................................................................
73
Figure 4.31. Shutter system for collecting single splats within
the deposition footprint of the
plasma plume.
..............................................................................................................................
75
Figure 4.32. Upper half of a mullite footprint produced in front
of the stationary gun as
described in the text.
....................................................................................................................
76
Figure 4.33. In-flight temperature and velocity maps for mullite
sample including single splat
collection. Point c is out of the left side
map...............................................................................
77
Figure 4.34. The thickness of coatings deposited with different
deposition parameters: (a)
forsterite, (b) spinel, (c)
mullite...................................................................................................
78
Figure 4.35. Porosity of different layers: (a) forsterite, (b)
spinel, (c) mullite. ........................... 79
Figure 4.36. Crystallinity Index for mullite coatings.
.................................................................
80
Figure 4.37. Tp-Vp map for
forsterite...........................................................................................
81
Figure 4.38. Forsterite single splats for samples 9 (left) and 2
(right) ........................................ 82
-
XV
Figure 4.39. Tp-Vp map for spinel.
.............................................................................................
82
Figure 4.40. Spinel single splats for sample 3 (left) and sample
4 (right)................................... 83
Figure 4.41. Tp-Vp map for mullite.
...........................................................................................
83
Figure 4.42. Mullite single splat for 8 (left) and 5 (right).
.......................................................... 84
Figure 4.43. Thermal diffusivity and thermal conductivity of the
coating materials versus
temperature.
.................................................................................................................................
86
Figure 4.44. XRD patterns of the layers, before and after
thermal conductivity test. In each
image, top pattern is the sample after heating to 1000°C and the
bottom is the one as-sprayed. 87
Figure 5.1. Schematic depiction of impact, spreading and cooling
of a single splat [46]. ............ 90
Figure 5.2. Specimen setup for quenching stress measurement
[46]............................................. 91
Figure 5.3. Quenching stress dependency on substrate temperature
[46]...................................... 91
Figure 5.4. Quenching stress of various powders on different
substrates [115]. ........................... 92
Figure 5.5. Schematic diagram of the variation of the final
residual stress in the coating with
substrate temperature Ts during spraying. σr (T0) is the final
residual stress after the sprayed
deposit and the substrate cool down to T0 [46].
.............................................................................
93
Figure 5.6. Surface stress of mullite and zirconia
[117].................................................................
94
Figure 5.7. Experimental averaged interface crack length (ai)
versus coating thickness (lc)[118]. 95
Figure 5.8. Residual stress measurement distribution according
to techniques [119]. The numbers
represent the industries which use the specific measurement
method in the above mentioned
study.............................................................................................................................................
96
Figure 5.9. Hole-drilling apparatus and residual stress strain
gauge rosette design [119]. ............ 97
Figure 5.10. Dataset collected for stress in shot peened Ni
based alloy using hole drilling
method alongside the suitable arrangement of the strain gauges
[124].......................................... 98
Figure 5.11. Practical example of a diffractometer, showing the
X-ray source, sample stage,
detector and goniometer
[119]........................................................................................................
99
Figure 5.12. Stress distribution in an 8 mm wide tungsten inert
gas weld in 3 mm thick plate [124].
............................................................................................................................................
100
-
XVI
Figure 5.13. A stress map collected for welded plate using
neutron diffraction method [124] ... 102
Figure 5.14. Curvature resulting from coating
deposition.........................................................
103
Figure 5.15. Stress distribution in a tungsten carbide coating
on Ti substrate using layer
removal[124].
...............................................................................................................................
103
Figure 5.16. The Almen strip system
[139]..................................................................................
104
Figure 5.17. Schematic illustration of the continuous curvature
measurement set-up.............. 108
Figure 5.18. Schematic diagram showing spray pattern of the
gun........................................... 109
Figure 5.19. Schematic diagram showing geometry of the curved
sample ............................... 110
Figure 5.20. Scale drawing of the fixture and mask for curvature
measurement (values in mm).
...................................................................................................................................................
112
Figure 5.21. Curvature for forsterite single layer coating.
........................................................ 113
Figure 5.22. Curvature of multilayer coating after deposition of
each layer: forsterite (top left),
spinel (top right) and mullite (bottom).
.....................................................................................
114
Figure 6.1. Substrate
dimensions...............................................................................................
121
Figure 6.2. Mechanical boundary condition for (a): side view of
in-process sample and (b) in-
service sample (From left to right: multilayer and duplex).
...................................................... 123
Figure 6.3. Zoom-in of the meshed
area....................................................................................
124
Figure 6.4. Temperature of the back of the substrate during
spraying: (a) forsterite, (b) spinel,
and (c)
mullite............................................................................................................................
126
Figure 6.5. Curvature calculation versus measurement for the
multilayer coating: (a) forsterite ,
(b) spinel, and (c) mullite.
.........................................................................................................
128
Figure 6.6. Quenching stress for forsterite
................................................................................
129
Figure 6.7. Comparison between numerical and analytical
solutions with experiment in: (a)
forsterite and (b) mullite.
...........................................................................................................
131
Figure 6.8. Comparison between numerical and analytical
solutions in multilayer coatings. .. 132
Figure 6.9. Comparison between this study and the work done by
Zhang et al.[158]. ................ 133
-
XVII
Figure 6.10. In-plane stress distribution of coatings after
deposition and cooling to room
temperature: (a) duplex coating (b) multilayer coating.
............................................................
134
Figure 6.11. Geometry and boundary condition of multilayer and
duplex systems.................. 135
Figure 6.12. The steady state temperature distribution in the
two systems............................... 136
Figure 6.13. Stress distribution in coatings at service
temperature: (a) duplex coating, (b)
multilayer coating.
.....................................................................................................................
136
Figure 8.1. Cross sectional SEM micrographs of as-received
(left) and calcined (right) powders.
...................................................................................................................................................
142
Figure 8.2. Particle size distribution of as-received, calcined
and calcined-sieved powders. ... 143
Figure 8.3. Spraying pattern.
.....................................................................................................
144
Figure 8.4. Temperature measurement of the back of the substrate
during deposition............. 145
Figure 8.5. In-flight particle properties of SG-100. From top
left: temperature, velocity, particle
size and particle number distributions. Particles are injected 5
mm inside the gun. The oval is the
highest particle density
trajectory..............................................................................................
146
Figure 8.6. In-flight particle properties of CACT gun. From top
left: temperature, velocity,
particle size and particle number distributions. Particles are
injected 5 mm inside the gun. The
oval is the highest particle density trajectory.
...........................................................................
146
Figure 8.7. In-flight particle characteristics for CACT gun with
sieved powder injected 5 mm
inside the gun. From top left: temperature, velocity, particle
size and particle number
distributions. The oval is the highest particle density
trajectory. .............................................. 147
Figure 8.8. Cross sections of (a) calcined powder deposited with
SG-100 (b) calcined powder
deposited with CACT (c) calcined and sieved powder deposited
with CACT. All injected 5 mm
inside the
gun.............................................................................................................................
149
Figure 8.9. XRD result for forsterite powder (JCPDS card number
85-1358).......................... 150
-
XVIII
Nomenclature
c
•
ε Strain rate
1/R Curvature
a Surface area of a particle
A Constant
a Substrate displacement
b Width of four-point bend sample
b Width of coating
CI Crystalinity index
cp Specific heat capacity
dc Thickness of coating
ds Thickness of substrate
Ē1 Average modulus of elasticity
Ec Modulus of elasticity of coating
Ed Modulus of elasticity of the coating material
Es Modulus of elasticity of the substrate
Es(Ts) Modulus of elasticity of the substrate at Ts
F Initiated force
GF Gage factor
h Thickness of four-point bend sample
H Thickness of substrate
h Heat transfer coefficient
hd Thickness of the coating
-
XIX
hs Convective heat transfer coefficient
hs Thickness of the substrate
I Moment of inertia
Ia Intensity of the background
Ic Intensity of the main peak
l Length of substrate
M Moment
m. Plasma flow rate
n Constants
P Applied load
qc Convective heat flux
qr Radiative heat flux
Rf Lead wire resistance
Rg Gage resistance
S1 Outer span of four-point bend setup
S2 Inner span of four-point bend setup
T∞ Free-stream plasma temperature
Ta Ambient temperature
Tmd Melting point of the coating material
TP Particle surface temperature
TS Substrate initial temperature
Ts* Transition temperature
Tsurf Surface temperature
-
XX
V Voltage
w Thickness of coating
y Distance from the neutral axis
z Layer number
zNI Neutral axis
α Thermal diffusivity
αc Linear coefficient of thermal expansion of coating
αd Coefficient of thermal expansion of deposit
αs Linear coefficient of thermal expansion of substrate
β Moduli ratio
δ Distance from the bending axis to the substrate/coating
interface
ΔH Activation energy
Δh Thickness of coating
ΔT Temperature gradient
Δκ Curvature difference
ε Particle emissivity
εi In-plane strain of layer i
θ Curved angle
κn Curvature of layer n
ν Poisson’s ratio
ρ Density
ρ% Percent influence
σ Von Mises equivalent stress
-
XXI
σc In-plane stress of coating
σdj In-plane stress at the midpoint of jth layer
σi In-plane stress of layer i
σq Quenching stress
σsb In-plane stress at the bottom of substrate
σst In-plane stress at the top of substrate
σT Stefan-Boltzmann constant
υc Poisson’s ratio of coating
υs Poisson’s ratio of substrate
-
1
1. Chapter I: Introduction
1.1. Coating techniques
Coatings are applied to surfaces of materials for various
reasons. Corrosion, oxidation and wear
resistance are only a few examples. There are different
techniques for applying a coating onto a
substrate, such as chemical and physical vapour deposition,
sol-gel, electro-deposition, dip
coating, and thermal spray. For different applications, a
specific method should be chosen
according to deposition rate, desired microstructure, cost,
final surface finish and so on. Among
these methods, thermal spray process sales are increasing
regularly. Thermal spraying
techniques may continue to grow as they are more environmentally
friendly than some
competitive coating techniques and full of undiscovered
potentials [1].
In this work we used the air plasma spray technique (APS), which
is the most commonly used
thermal spray process, and studied the generation of residual
stresses in a multilayer coating.
1.2. Thermal spraying
Thermal spray is the name of a category of coating deposition
techniques which use thermal
energy to melt and/or pyrolyze the precursor (powder, solution
or wire) and accelerate it
towards the substrate. At the substrate, the individual molten
particles flatten and solidify.
One of the first techniques invented in this category was flame
spaying, which uses a flame
(usually oxy-propane) to melt the powder and gravity or
compressed air to accelerate the
droplets towards the substrate. The temperature generated (on
the order of 3000°C) may be
sufficient to melt metals, but in most cases it is insufficient
to melt ceramics [2].
Another type of thermal spray is air plasma spray (APS), in
which a DC arc is formed between
an anode and cathode, which generates a plasma [3]. The
precursors are injected into the high
temperature plasma and accelerated towards the substrate as a
result of the high velocity gas
flow. This process can also be performed in an inert atmosphere
or vacuum, in which case the
process is named vacuum plasma spray (VPS). Due to the high
temperatures attained, plasma
spray is a good application method for ceramic coatings, as it
can ensure that the precursors
have been melted before hitting the substrate.
-
2
Plasma spray process parameters govern the resulting coating
characteristics and when changing
these parameters (gas flow rate, current, standoff distance,
feeding rate, etc.), coating
characteristics change. Thus, we need to optimize the plasma
parameters to have a coating with
the desired characteristics.
1.3. Application of coatings in diesel engines
Diesel engines are commonly used in buses, trucks and in
passenger cars. In a diesel engine, the
fuel is compressed to a very high pressure and it automatically
ignites and burns. The idea of
thermally insulating the engine is as old as the internal
combustion engine itself and is based on
the knowledge that only 30-40% of the entering fuel energy is
converted to useful work on the
output shaft [4]. In the 1980s there was an effort to use
thermal barrier coatings (TBCs) in diesel
engines in pursuit of advantages, including higher power
density; fuel efficiency; reducing
specific fuel consumption; emissions and noise; improving the
engine life and cold start
reliability; and multi-fuel capacity due to higher combustion
chamber temperatures (900°C vs.
650 °C) [5-8]. The goal was to have an engine with 48%
efficiency rather than the ordinary 33% [9]. Seker and Kamo
developed an adiabatic engine for a passenger car which showed
an
increase of 12% in performance [10]. Prasad and Samria coated
the piston crown with partially
stabilized zirconia (PSZ) and reported a 19% reduction in heat
loss through the piston [11].
Studies done by Hejwowski and Weronski showed that using TBCs in
a diesel engine could
increase engine power by 8%, decrease the specific fuel
consumption by 15-20%, and increase
the exhaust gas temperature by 200K [12]. A significant problem
was still unsolved: durability.
1.4. Diesel engines and turbine engines: similarities and
differences
Thermal barrier coatings (TBCs) have been in use for three
decades in the hot sections of
turbines [13]. High performance turbine blades consist of single
crystal super alloys, which
cannot tolerate the high temperature (more than 1000°C) and
corrosive environment
experienced in current turbine engines. A thin layer of yttria
stabilized zirconia (YSZ), with a
thickness of approximately 200 µm, decreases the alloy surface
temperature by 100-300°C,
which largely fulfills current requirements [14].
The service environment of the coating in the turbine is
markedly different from that in the
diesel engine. In the former, the service temperature is high
(1000-1100°C). The superalloy
substrate’s maximum service temperature is about 800°C [14]. The
thickness of coating is a few
-
3
hundred microns and the coating is applied to protect it against
oxidation, hot corrosion,
thermomechanical fatigue, and creep. Due to the high substrate
temperature, oxidation of the
bond coat and creep play major roles in coating failure.
On the other hand, in the diesel engine the gas temperature,
currently less than 650°C, would
ideally approach 900°C[5, 6]. The substrate temperature is
limited to approximately 200°C, and
therefore a thick coating (at least 1mm) is required, which
leads to a large thermal gradient. In a
thick thermal barrier coating (TTBC), the bond coat temperature
is too low for severe oxidation
and creep [15]. Another major difference between the two systems
is that due to the on/off nature
of the diesel engine operation, the thermal barrier coating in a
diesel engine experiences more
transient mode conditions than in a turbine engine. These
differences in service conditions and
coating thickness result in different failure mechanisms,
primarily thermocyclic fatigue and
thermal shock in the TTBC.
Numerical modeling done by Kokini et al. [15] suggested that in
a thick thermal barrier coating,
when the surface of the yttria-stabilized zirconia coating is
heated, a large compressive stress is
developed in the surface, which may be relaxed after two hours
of steady state heating. Upon
cooling, the stress may become tensile and initiate cracks which
grow with each thermal cycle.
Modelling also indicated that during cyclic operation, rapid
cooling can result in the
development of a transient tensile stress at the surface of the
coating, leading to crack initiation
and thermal shock failure [15]. Due to the mismatch in
thermomechanical properties of the top
coat and bond coat, this interface is a source for cracking and
delamination [16].
1.5. Thesis objective
The objective of the thesis research is to design and fabricate
a relatively low cost multilayer
ceramic thick thermal barrier system, which is physically and
chemically stable under the
operating conditions typical of a diesel engine and which
minimizes the internal stress
experienced when subjected to a thermal gradient typical of that
experienced in service. In this
regard, traditional refractory ceramic material compositions
have been chosen instead of
expensive synthetic oxides. Two paths have been pursued
simultaneously: optimization of the
coating process for each material and development of a model of
the stress in the coating during
processing and service. The model was validated using an in-situ
curvature measurement
-
4
technique, and then used to predict the stress distribution that
would be experienced under
service conditions.
1.6. Structure of the thesis document
The structure of the thesis is as follows. The thermal spray
process and the most significant
process parameters are described in Chapter 2. The third chapter
discusses the selection of the
coating materials for the specified application. In Chapter 4
the experimental work involved in
establishing the deposition process parameters for each material
is presented. This includes
characterization of the feedstock powder and deposited coatings.
Porosity, crystallinity, thermal
conductivity and modulus of elasticity are measured for coatings
of each material. Chapter 5
reviews the nature of residual stress in such coatings, and
direct and indirect methods for its
measurement, before concentrating on the technique of curvature
measurement. The
experimental implementation of this technique is described, and
in-situ measurements of
curvature during coating deposition presented. Numerical
modeling of the build-up of internal
stresses during deposition, the prediction of the resulting
curvature, and the associated residual
stresses are included in Chapter 6. To validate the model, the
curvature predictions are
compared with the experimentally measured behaviour. In Chapter
7 the results of the study are
discussed, conclusions drawn, and suggestions for further
studies presented.
-
5
2. Chapter II: Thermal Spray Process
2.1. Introduction
Thermal spraying is a process in which a stream of molten and/or
semi-molten particles is
sprayed onto a substrate. A high temperature flame or plasma jet
is used to melt the feedstock
powder and form a coating on the substrate [2]. The molten
particles impact, flatten, and form
splats on the substrate.
Coating properties depend on five subsystems on which the
operator can have some control [1]:
1. Flame or jet formation: linked to the torch design, gas
composition and flow rate, power,
and other process parameters.
2. Powder and its injection: depends on powder chemical
composition, particle size
distribution, structure and morphology, injector design and
positioning.
3. Deposition atmosphere: spraying can be applied in air, inert
gas or vacuum.
4. Substrate material and preparation: roughness, oxidation
state, preheating time and
temperature during and after spraying.
5. Relative motion of the torch and substrate: traverse speed
controls coating thickness per
pass and partially the heat transferred to the coating and
substrate.
There are different ways of categorizing spraying methods, based
on principal energy source,
maximum achievable temperature [2], particle impact velocity and
temperature [17], and even
popularity in the industry [18].
Based on achievable temperature, we may use either plasma
spraying or combustion based
spraying techniques. In combustion based spraying techniques,
the maximum flame temperature
is governed by the enthalpy of the chemical reaction between
combustion gases, which is no
greater than 3300K [2]. On the other hand, when using plasma,
the only limitation is the amount
of electrical energy supplied, which in turn is a function of
the cross section of the power leads [2]. Plasma temperatures of
5000 K to 25000 K are reported for different types of plasma
spraying methods (Table 2-1) [2, 17, 18].
-
6
Since in thermal spraying the particles must be in a molten or
semi-molten state [19] before
deposition, according to Table 2-1, plasma spraying tends to be
a suitable choice for ceramic
materials, which generally have higher melting points than
metals.
Table 2-1. Comparison of different spraying methods [2].
Flame D-Gun HVOF APS IPS VPS RFS
Gas Temperature (°C)
p=1 atm 2700 3200 3000 14000 14000 --- ---
p=0.25 atm --- --- --- --- --- 10000 8000
Particle velocity (ms-1)
Al2O3- 30 μm
p= 1 atm 70 500 350 230 250 --- ---
p= 0.25 atm --- --- --- --- --- 380 30
Plume length (cm)
p= 1 atm < 10 --- < 20 < 7 < 10 --- ---
p= 0.25 atm --- --- --- --- --- < 15 < 15
Particle injection Axial Axial Axial Orthogonal or Axial
Axial
APS: air plasma spray, IPS: inductive plasma spray, VPS: vacuum
plasma spray, RFS: radio frequency spraying.
2.2. Plasma spraying
As discussed earlier, plasma spraying is a suitable method for
applying ceramic coatings. Due to
the high temperature, most ceramic materials can be sprayed
using this method. However, to
have satisfactory deposition efficiency, the melting temperature
should be at least 300 K lower
than the vaporization or decomposition temperature [2].
There are some unique features in plasma spraying such as
[18]:
-
7
1. A wide range of materials can be deposited, from polymers to
metals to ceramics and their
combinations.
2. Combinations of ceramics and alloys having different vapour
pressure can be deposited
without significant changes in final composition.
3. Homogenous coatings can be produced with no significant
difference in composition through
the thickness parallel to spray direction.
4. Fine, equiaxed grains without columnar defects can be
produced.
5. Functionally graded coatings can be generated.
6. High deposition rates of the order of mm.s-1 are
achievable.
7. Near-net shape free-standing thick coatings can be
sprayed.
8. The process can be carried out in different environments such
as air, reduced pressure,
controlled atmosphere, and underwater.
2.3. Process parameters
The microstructure and mechanical properties of plasma sprayed
coatings are determined by the
size, temperature and velocities of the droplets in the jet at
deposition [20] and the substrate
temperature. The temperature and velocity of the particles are
determined by the plasma jet
characteristics [21]. A common way to achieve a satisfactory
deposit is to apply a design of
experiment (DOE) to find a combination of process parameters
that produce acceptable coating
properties[22]..Understanding the relationships between the
process parameters, plasma
characteristics, and the in-flight particle characteristics is
vital to assure reproducible, high-
quality coatings.
In plasma spraying, the relationship between the processing
variables and the resulting plasma
and particle characteristics is complex because of the large
number of processing parameters,
including plasma current, plasma gas composition and flow rate,
powder size, powder feed rate,
carrier gas flow rate, standoff distance, and plasma gun
traverse speed [21]. For example, Zhao et
al. tried to interpret plasma characteristics in an A2000 vacuum
plasma spray system (Sulzer
Metco AG, Winterthur , Switzerland) by modeling the influence of
different plasma parameters
on the energy balance of the jet [21]. Some of these results are
shown below.
-
8
2.3.1. Plasma current
Figure 2.1 shows variations of plasma gas temperature and
velocity, with plasma current in the
range of 400-1000 A, at a fixed Ar flow rate of 35 l.min-1 and
H2 flow rate of 8 l.min-1 [21]. The
increases shown are due to increasing the plasma power, plasma
energy and plasma gas
ionization [21].
Figure 2.1. Effect of plasma current on: (a) plasma gas
temperature, (b) plasma gas velocity [21].
2.3.2. Primary gas flow rate: argon
Figure 2.2 illustrates the effect of Ar flow rate on plasma gas
temperature and velocity found by
Zhao et al.[21]. Ar flow rate was in the range of 15-100 l min-1
at a constant current of 700 A, H2 flow rate of 8 l min-1[21]. The
decrease in temperature, in this range, is because the amount of
gas
to be heated increases faster than the energy input to the
plasma. In contrast, velocity increases
almost linearly with increasing the Ar flow rate, which is due
to more moles Ar in the jet.
-
9
Figure 2.2. Effect of Ar flow rate on: (a) plasma gas
temperature and (b) plasma gas velocity [21].
2.3.3. Auxiliary gas flow rate: hydrogen
Zhao et al.[21] found that H2 flow rate has a dual effect on
plasma gas characteristics (Figure
2.3). The H2 flow rate was in the range of 0-20 l min-1, at a
fixed plasma current of 700 A and
Ar flow rate of 35 l min-1. H2 is often used as the secondary
gas in APS and VPS, because its
dissociation and ionization contribute strongly to the overall
plasma enthalpy [2]. As shown in
Figure 2.3, increasing H2 up to 2 l min-1 increases both the
plasma gas temperature and the
degree of Ar and H ionization (Figure 2.3 (a), (b)). However,
when gas flow rate exceeds 2 l
min-1, the volume of gas to be heated per unit time increases at
a faster rate than the plasma
energy input rate [21].
-
10
Figure 2.3. Effect of hydrogen flow rate on: (a) plasma gas
temperature , (b) degree of Ar and H
ionization and (c) plasma gas velocity [21].
2.3.4. Carrier gas flow rate
Miller et al.[23] showed that the carrier gas flow rate affects
coating properties via changing the
particle dwell time and trajectories for radially injected
particles. Figure 2.4 illustrates the effect
of radially injected carrier gas flow rate on particle
trajectory [23]. With a low carrier gas flow
rate, fine powders bounce off the plasma jet, and most are not
melted. Heavier (coarser)
particles, though, can penetrate to the centreline of the jet,
heat up and accelerate towards the
substrate. This situation results in a porous coating containing
unmelted particles, which
deposits above the centreline of the torch.
Using a high carrier gas flow rate in radial injection, finer
particles may travel along near the
centreline, while coarser particles pass the centreline,
spending more of their residence time in
-
11
the hottest section of the jet. The coating forms mainly beneath
the centreline with this condition [23].
Figure 2.4. Particle trajectories expected for low and high
powder carrier gas flow rates [23].
-
12
(a)
(b)
Figure 2.5. Effect of carrier gas flow rate on (a) density and
(b) deposition efficiency [23].
-
13
Increasing the carrier gas flow rate increases the density and
to some extent increases the
deposition efficiency by melting more particles (Figure 2.5)
[23]. If the carrier gas flow rate is too
high, the plasma temperature and velocity can be significantly
affected and many particles may
pass completely through the plasma.
2.4. Feedstock injection
The ideal situation for a high density plasma-sprayed coating
would be when all particles
injected reach the substrate with a temperature well above their
melting point (and below the
decomposition or vaporization point) with velocities as high as
possible but compatible with
achieving a fully melted state. High velocities decrease the
particle residence time and thus their
heating [19]. This “ideal situation” is impossible to achieve
because of the following reasons [1].
(a) Particles resulting from milling and atomization processes
have a relatively wide size
distribution: typically between 22 and 45 μm in diameter. For
example if the diameter
distribution has a maximum ratio of 2, the mass distribution has
a ratio of 8, which leads
to particles with very different thermal history.
(b) Particles, at least in radial injection in DC arc plasma
spraying, should be injected with a
momentum similar to that of the plasma jet. This is achieved by
using a carrier gas and
an injector. At the exit of the injector, particles are not all
parallel to the injector axis.
Thus, some of the injected particles do not penetrate the plasma
jet. Furthermore,
depending on their trajectories and masses, not all particles
which penetrate the plasma
will melt. The best situation is to optimize the carrier gas
flow rate for the mean of the
size distribution.
(c) Particle melting depends on plasma gas velocity (which is
related to the residence time)
and the plasma gas composition. The plasma gas composition
determines the plasma
enthalpy and controls the heat transfer to the particle. For
materials with a low thermal
conductivity (e.g, thermal barrier coating materials such as
zirconia and mullite), high
plasma temperatures and velocities may result in larger
particles not being uniformly
heated and melted.
(d) For low melting point materials, the particle size
distribution is still a problem, as the
question of whether the radially injected particles penetrate
into the jet or not remains.
With smaller particles, the carrier gas velocity has to be
increased drastically
-
14
(proportional to the negative third power of the particle
diameter) and, for example,
below 5-10 μm the carrier gas flow rate disturbs the plasma
jet.
2.5. Particle-particle interactions
Typically powder, 20-100 μm in diameter, is injected radially
into the plasma jet and while
accelerating toward the substrate, melts rapidly. The velocity
of particles ranges from 100 to 400
ms-1 and the thermal conditions are such that even materials
with a high melting point melt in
the residence time [24].
Powder particles in the plasma jet are heated through convection
by the plasma gas and
conduction inside the powder particle and at the same time
release energy through radiation to
the ambient [2]. The amount of heat gained by the powder from
the plasma is:
)( Sc TThaq −= ∞ ( 2.1)
and the amount of heat lost by radiation to the surroundings
is:
)( 44 aSTr TTaq −= εσ ( 2.2)
where:
h: plasma/particle convective heat transfer coefficient
a: surface area of the particle
T∞: free-stream plasma temperature
TS: particle surface temperature
Ta: temperature of the ambient
σT: Stefan-Boltzmann constant
ε: particle emissivity
For a particle to gain energy in a plasma jet, the net energy
change must be positive:
0>>−= rcn qqq ( 2.3)
Since qr is generally small compared to qc because of the small
value of the Stefan-Boltzmann
constant, which is in the range of 5.67×10-8 Wm-2K-4, the net
energy is much greater than zero
and the particles heat up in the jet [2].
-
15
Solid particles tend to bounce off the substrate or previously
formed coating. Therefore, the first
essential task in successful plasma spraying is to melt the
majority of powder fed into the jet. In
practice, due to large temperature and velocity gradients inside
a plume, particles with different
thermal and velocity histories exist [24]. In the case of porous
or low thermal conductivity
materials, there is always a thermal gradient between surface
and the core of the particles, which
might exceed 1000 K [2]. These particles are the ones that can
have the most significant effect on
the coating structure and properties.
2.6. Particle trajectory
Radially injected particles in a plasma jet have different
thermal histories even for a constant
carrier gas flow rate. Particle density is one factor affecting
the trajectory. Low density particles
having lower momentum cannot easily penetrate into the plasma
jet and may bounce off it [13].
These particles spend little or no time in the hot section of
the jet and might not melt. On the
other hand, larger size particles may penetrate and travel
through the plasma jet, and may not
make it to the substrate. Larger particles need more dwell time
in the jet to melt or they may end
up as non- or semi-melted particles (Figure 2.6).
For a given set of plasma parameters and particle structure,
there is only one particle size which
will be correctly heated, accelerated and deposited on the
substrate. Therefore, a narrow particle
size distribution should be used [13].
Figure 2.6. Schematic picture of particle trajectory for
different particle sizes [13].
-
16
Figure 2.7 illustrates the dispersed particle trajectories of
alumina particles with three different
mean diameters in a turbulent-free argon plasma jet [2]. Smaller
particles disperse more and
dispersion becomes more prominent when moving downstream from
the injection point due to
accumulated random walk influence [2].
Figure 2.7. Dispersed trajectories of alumina particles of
different sizes in a turbulent free
plasma argon jet [2].
2.7. Particle- substrate interactions
At impact onto the substrate, the molten or semi-molten/plastic
state particles flatten, solidify
and form splats (Figure 2.8). Splats usually have columnar
structures with grain size between 50
and 200 nm (Figure 2.9Error! Reference source not found.).
Different coating properties are
highly linked to the quality of contacts between splats
[19].
Computer simulations show that the most important reason for a
change in splat shape is a
change in the thermal contact resistance between the splat and
the substrate[25, 26]. The growth of
an oxide layer on metal substrates is one way in which the
contact resistance can change
significantly. The thickness of the oxide film on a surface
depends on the heating temperature
and time [27].
-
17
Figure 2.8. Schematic deformation of a splat at impact [2].
Figure 2.9. Microstructure of fractured as-sprayed partially
stabilized zirconia showing
columnar grains [28].
Commonly it is believed that splat morphology should have an
important effect on adhesion and
coating quality. Tanaka and Fukumoto found that splat shape is
highly dependent on substrate
temperature. The transition temperature (Tc) has been introduced
as the substrate temperature at
which half of the splats were disk type [29]. The transition
temperature is found to depend more
on the coating material than on the substrate material [30]. The
most widely accepted explanation
-
18
for the transition temperature is that Tc is the temperature at
which adsorbates and condensates
at the substrate surface are eliminated [31] (Figure 2.10 to
Figure 2.12). As generally known,
clean surfaces energetically attract foreign species, resulting
in adsorption and condensation of
these molecules. At a given temperature these condensed species
begin to vaporize and the
evaporation rate increases with temperature. By heating the
substrate the condensed molecules
would be removed from the substrate and the splats deposit on a
clear surface. Increasing the
substrate too much and metallic substrate may oxidize which
forms an entirely new layer on the
substate.
On the other hand, Pasandideh-Fard et al. found that the rate of
solidification is much more
sensitive to values of thermal contact resistance than to
substrate temperature [27]. Thermal
contact resistance can be affected by gases or solid impurities
trapped at the particle-substrate
interface [30].
Figure 2.10. Morphology of different ceramic splats on AISI304
substrate [29].
-
19
Figure 2.11. Dependence of fraction of disk shape splats on
substrate temperature (substrate:
AISI 304, particle Al2O3) [29].
Figure 2.12. Dependence of Al2O3 splat morphology on substrate
(AISI 304) temperature [29].
-
20
Studies by Pershin et al. showed that heating the substrate
while applying a plasma spray
coating can significantly increase adhesion strength and reduce
porosity [30]. Similar results were
reported by Salimijazi et al., showing a 40% decrease in
porosity after substrate preheating [32].
2.8. Coating formation
Coatings are built up by successive splats. A thermally sprayed
coating has a complex
microstructure containing splats with different thermal
histories, pores, voids, and microcracks
(Figure 2.13). In this regard, optimizing the coating structure
for a specific application is of
great importance.
There has been a great deal of work done to find correlations
between coating properties and
spray parameters, but attempts to link at a more fundamental
level coating properties to in-flight
particle characteristics, particle properties at impact, and
substrate properties are less than a
decade old [19].
Figure 2.13. Schematic rendering of the chaotic structure of
plasma sprayed coating: 1. Thin
molten shell; 2. Unmelted core; 3. Liquid splash; 4. ‘Pancake’
splat; 5. Interlocked splat; 6.
Oxidized particle; 7. Unmelted particle; 8. Pore; 9. Void; 10.
Roughened substrate; 11.
Substrate[2].
-
21
The coating structure depends on many parameters in addition to
the degree of melting and
velocity of the particles, such as s [1, 18, 19]:
(a) Surface preparation: cleaning, grit blasting or surface
roughening has an important role
in increasing the adhesion of the coating. The main adhesion
mechanism is splat
shrinkage while cooling around peaks of the roughened substrate.
Chemical adhesion
may exist at some points in metal coating/metal substrate
interfaces.
(b) Splat layering: this depends on the particle parameters at
impact, the shape and topology
of already deposited layers, the ability of the flattening
particle to accommodate the
existing pores, asperities, etc, and, finally, of their
temperature at the moment of impact.
(c) The substrate and coating temperature: this controls the
inter-lamella contacts and the
residual stress distribution.
To address the coating growth rate, deposition efficiency is
used. It is defined as the ratio of the
weight of coating formed on the substrate to the weight of
powder consumed. The effect of
different plasma parameters on deposition efficiency is shown in
Figure 2.14 [2].
Figure 2.14. Dependence of deposition efficiency on various
plasma spray parameters [2].
The absolute value of deposition efficiency will depend on the
substrate size, shape, and the
spray pattern. For comparison purposes, with the same feeding
rate, number of gun passes,
traverse speed, and coating densities, the thickness of the
deposited layers will reflect the
-
22
deposition efficiency. If coating densities differ by a small
amount, coating thickness could still
be a useful parameter for qualitatively comparing the deposition
efficiencies. For example, if
two coatings had the same thickness but a 5 vol% difference in
porosity, the actual deposition
efficiency would be different by only 6%.
2.9. Summary
The plasma spray process is governed by a large number of
process parameters. To reliably
obtain an acceptable coating, the effect of the spraying
parameters on the coating properties
must be understood. These parameters can then be optimized
within the technical limitations of
the system in use.
-
23
3. Chapter III: Coating Materials
3.1. Introduction
Thermal Barrier Coatings (TBCs) in diesel engines lead to
advantages including higher power
density, fuel efficiency, and multifuel capacity due to the
possibility of higher combustion
chamber temperature (900°C vs. 650°C) [5, 6]. The increased
operating temperature possible
when using TBCs can increase engine power by 8%, decrease the
fuel consumption by 15-20%
and increase the exhaust gas temperature 200K [12]. Although
several systems have been used as
TBCs for different purposes, yttria-stabilized zirconia with 7-8
wt% yttria has received the most
attention [33, 34]. Several factors play important roles in
overall TBC performance including
thermal conductivity, microstructural and chemical stability at
the service temperature, creep
resistance, and the coefficient of thermal expansion (CTE) [5,
6, 16, 34-40].
Plasma spraying is the most common method of depositing TBCs for
diesel applications. It
creates a splat structure with 10-20 % volume fraction of voids
and cracks [41] (Figure 3.1). The
high porosity of this structure makes it an ideal choice for
TBCs. Although the potential
advantages of utilizing thermal sprayed TBCs in diesel engines
were recognized over 30 years
ago, widespread application has been limited by insufficient
lifetimes and unacceptable cost.
Premature failure by spallation has been a frequently observed
failure mode for YSZ thermal
barrier coatings. This issue has again become very important in
the context of future engines
designed for improved efficiency, durability and reliability
[37].
Figure 3.1. Air Plasma-Sprayed (APS) coating [42].
-
24
In contrast to the case of TBCs for diesel engines, TBCs have
become almost universally
employed within the hot sections of gas turbine engines over the
course of the past three
decades. Partially stabilized zirconia (PSZ) coatings on turbine
blades enhance the oxidation and
corrosion resistance of the Ni- based alloy blades and allow
operation of the turbine engines at
higher gas temperatures. The coating increases the lifetime of
the blades and limits the
maintenance costs [43].
There are two major differences in the two cases. The cost of
the coating represents a much
smaller portion of the total cost of a turbine engine than it
does of the cost of a diesel engine.
From a technical perspective, the service environment of the
coating in the turbine is markedly
different than in the diesel engine. In the former, the service
temperature is high (1000-1100°C).
The superalloy substrate’s maximum service temperature is about
800°C. The thickness of
coating is a few hundred microns and is applied to protect
against oxidation, hot corrosion,
thermo-mechanical fatigue, and creep. Due to the high substrate
temperature, oxidation of the
bond coat plays a major role in coating failure. In contrast, in
the diesel engine the gas
temperature, currently less than 650°C, would ideally approach
900°C. The substrate
temperature is limited to approximately 200°C, and therefore a
thick coating (at least 1mm) is
required, which leads to a large thermal gradient. In a thick
thermal barrier coating (TTBC) the
bond coat temperature is too low for significant oxidation and
creep [15]. As a consequence the
dominant failure mechanisms are different in the two cases.
3.2. Failure mechanisms in TBCs
Mechanical behaviour of a material is highly dependent on its
microstructure, which in turn is
dependent on the processing technique. Plasma sprayed coatings
have complicated
microstructures (Figure 3.2). Plasma sprayed TBC coatings have a
high density of microcracks,
isolated large pores, and weak interfaces between splats. These
defects can be sources of
mechanical failure in thermal barrier coatings. Another
characteristic of plasma spray coatings
is anisotropy in microstructure and mechanical properties.
Cracks can propagate more easily in
the plane parallel to the coating-substrate interface than in
the perpendicular plane [44].
Horizontal (parallel to the coating-substrate interface) and
vertical (perpendicular to the coating-
substrate interface) cracks in the TBC perform different roles.
Horizontal cracks, mostly located
at splat boundaries, may considered non-detrimental to the
coating, and helpful to reduce the out
-
25
of plane heat transfer in it, making the TBC more effective.
However, these cracks can grow
during thermal cycling, link together, and cause coating
spallation [45]. Vertical cracks, which
may propagate through the coating thickness (then often referred
to as segmentation cracks) can
increase the coating compliance and extend its lifetime
[45].
Figure 3.2. SEM image of plasma sprayed mullite coatings
Coating failures in diesel engines are known to occur due to
either loss of cohesion in the
ceramic layer or loss of adhesion at the coating/bond coat or
the bond coat/substrate interface.
Loss of adhesion may occur at high service temperatures due to
the growth of an oxide layer
between the bond coat and top coat, known as a thermally grown
oxide (TGO) layer. This
mechanism is not significant in water-cooled diesel engines, as
the maximum service
temperature does not exceed 1000°C. At the lower service
temperatures, thermo-mechanical
fatigue and residual stresses play a more important role in
coating failure [5].
Thermal stresses generated by the difference in the coefficient
of thermal expansion between the
substrate and coating are one of the major factors contributing
to failure in plasma sprayed
coatings [46]. The residual stresses which are induced in the
fabrication process of the TBCs are
-
26
associated with many mechanical failures of the coating. For
example, delamination may occur
along the interface of the pre-tensioned coatings [47] while
compressive residual stress may cause
spalling inside the coating [48]. The mechanism whereby residual
stresses are generated within
thermal barrier coatings will be discussed in chapter 5.
The failure of plasma-sprayed TBCs under thermal cycling is a
highly complex process
involving an interplay between several general phenomena listed
below: (i) thermal expansion
mismatch stress; (ii) growth of thermally grown oxide (TGO) at
the interface; (iii) cyclic creep
of the bond coat; (iv) depletion of Al in the bond coat leading
to the formation of brittle oxides
other than α- Al2O3; (v) sintering of the porous TBC and the
attendant deterioration of strain
tolerance and thermal resistivity; (vi) degradation of the metal
ceramic interface toughness; (vii)
delamination and cracking; (viii) crack coalesces. The TBC
failure mechanisms are highly
system and application specific, where one or more of the above
phenomena dominate [37]. For
example, in thick thermal barrier coatings (for diesel engine
applications), service temperature is
not high enough for TGO formation, top coat sintering or cyclic
creep of the bond coat. Thermal
stresses are the most important factors in this application
[20].
In thick coatings, additional thickness and low thermal
conductivity result in a higher thermal
resistance. These coatings are prone to cracking and
delamination near the interface due to the
mismatch in thermomechanical properties between the top coat and
the substrate [16]. Kokini et
al showed that the stress relaxation process occurring in thick
TBC systems at high temperature
is a significant cause of crack initiation and propagation [15].
Hence, developing a multilayer
thick thermal barrier coating may be a solution that reduces the
driving force for TBC-substrate
interface fracture [16].
Plasma sprayed coatings start out with microcrac