Modified Thick Thermal Barrier Coatings
Modified Thick Thermal Barrier Coatings
Modified Thick Thermal Barrier Coatings
Publication 473
Tampere 2004
Thesis for the degree of Doctor of Technology to be presented with due permission for
public examination and criticism in Konetalo Building, Auditorium K1702, at Tampere
University of Technology, on the 28th of May 2004, at 12 noon.
PREFACE
The work for this thesis was mainly carried out during the years 1999-2003 in the Tampere
University of Technology, Institute of Materials Science (TUT/IMS). The supervisor of the
thesis was professor Tapio Mntyl. I want to thank professor Mntyl for all his guidance
and for giving me the opportunity to prepare the thesis at TUT/IMS. I am grateful to
professor Petri Vuoristo with whom I worked closely for years in TUT/IMS, too. During
those years he always deepened my knowledge of coatings and coating technologies.
Many thanks also to co-authors from TUT/IMS (Dr. Minnamari Vippola, M. Sc. Jari
Tuominen) as well as to the technical and assistant personnel (Mikko Kylmlahti, Ulla
Mnnikk, Sari Iltanen, Mari Honkanen, Katri Kosme).
I completed part of the work at the University of Trento, Italy (08/2001-07/2002). I am
grateful especially to Dr. Luca Lutterotti, Dr. Rosa Di Maggio and professor Roberto Dal
Maschio who gave me the opportunity to work at the University of Trento. They all
supported me scientifically, but also helped me with the language and the Italian way of
living. I thank all the organisations (IVO Sti, Henry Fordin sti, Ehnrothin sti,
Tampereen kaupunki and Kaupallisten ja teknillisten tieteiden sti) that awarded me the
funding for this exchange period in Italy.
The work with thick thermal barrier coatings in diesel engines started in the project
Development of the wall construction of the combustion chamber in a Diesel engine. The
project was funded by National Technology Agency (Tekes), Wrtsil Technologies Oy
and Patria Finavitec Oy. The project lasted for three years (12/1999-08/2002) and was
coordinated by the Internal Combustion Engine Laboratory, Helsinki University of
Technology (HUT/ICELAB). Co-operation with HUT/ICELAB continued in the Extreme
Value Engine (EVE) project. The EVE project (06/2000-12/2003) was funded by the
Academy of Finland. I thank all the financial supporters related to these projects. I would
like also to thank the personnel of the HUT/ICELAB for their fruitful, interdisciplinary cooperation in the field of diesel engines and materials science.
In 1999-2003 TUT/IMS took part in the COST 522 Program (Ultra Efficient, Low Emission
Power Plant/Gas Turbine Group) in which the TTBCs were considered more from the
standpoint of gas turbines. Here I thank Federico Cernuschi (CESI, Italy), Carlo Gualco
(Ansaldo Richerche, Italy) and Robert Vassen (Forschungszentrum Jlich GmbH,
Germany) for their contributions to our joint studies.
I thank also the personnel of the Institute of Materials Science and especially the people in
the Surface Engineering Laboratory where the atmosphere is both scientific and relaxing.
Last but not least I thank my wife Riikka for her positive attitude towards my work. Finally I
am grateful to my daughter Ella who keeps my feet on the ground by saying once in a
while "Daddy, you just an average engineer.
Muurame, 16 of February, 2004
Samppa Ahmaniemi
ABSTRACT
This thesis studies the microstructures of modified zirconia based thick thermal barrier
coatings as well as their properties. Plasma sprayed yttria stabilised zirconia (8Y2O3-ZrO2)
was the basic reference coating, but magnesia (MgO) and ceria (CeO2) stabilised zirconia
coatings were also studied. Coating microstructures were mainly modified by post
treatments, such as phosphate based sealing treatments and laser glazing. These
procedures were carried out in order to improve particular coating properties such as
erosion resistance, thermal cycling resistance and hot corrosion resistance. The work
concentrated mainly on optimising the coating modification procedures, performing
detailed coating characterisation, determining the coating mechanical and thermal
properties and testing their high temperature properties in hot corrosion and thermal
cycling experiments.
The modification procedures changed coating microstructures near the surface.
Phosphate sealants penetrated approximately 300-400 m into the coating microcracks
and pores reducing the open porosity by 24-48 % depending on the coating material. It
was found that the sealant improved the cohesion of the splat boundaries by adhesive
binding and chemical bonding mechanisms. In laser glazing it was possible to control the
melting of the ceramic coating surface. Optimal thickness of the melted layer was 50-150
m leading to a dense surface layer with specific vertical macrocrack structure.
Modification processes strongly affected on the coating mechanical and wear properties.
Microhardness of the phosphate sealed coatings was increased by 15-55 % and as much
as 70-100 % in the case on laser-glazed coatings. The strengthening effect of the
phosphate sealing was clearly seen in the four-point bending tests, where the modulus of
rupture in bending (RB) of the 8Y2O3-ZrO2 coating was increased by more than 200 %. At
the same time, the bending modulus (EB) of the phosphate sealed coating was almost
eight times higher than the as-sprayed reference coating. In the laser-glazed 8Y2O3-ZrO2
coating the modulus of rupture in bending was one fourth and the bending modulus only
one fifth that of the as-sprayed coating. Erosion resistance of the 22MgO-ZrO2 and 8Y2O3ZrO2 coatings was improved by 65-70 % due to the phosphate based sealing treatment.
The average improvement in the laser-glazed 8Y2O3-ZrO2 coating was 35 %.
Thermal conductivity (k(T)) of all studied zirconia based coatings at a temperature range of
RT-1250oC was more than doubled by the phosphate sealing. Sealing also weakened the
high temperature phase stability of the 8Y2O3-ZrO2 coating at temperatures over 1000oC.
Laser glazing had only a minor effect on the thermal properties of the coating. Depending
on the macrocrack structure and its orientation, laser glazing either slightly raised or
slightly lowered thermal conductivity.
Modification processes had no clear beneficial effect on coating hot corrosion resistance,
when exposed in air to a NaSO4-V2O5 based deposit at 650, 750, and 850oC for 48-1000
hours. The penetration of melt deposit into the phosphate sealed coatings was lowered in
some degree if compared to the as-sprayed coatings. However, the phosphate sealed
coatings failed in hot corrosion tests mainly because of the strong compressive stresses
generated during the test. The compressive stresses were mainly induced when tetragonal
and cubic zirconia phases transformed to monoclinic zirconia. The microstructure of the
laser-glazed coatings was not optimal considering the hot corrosion test method (melt
deposit exposure). The melt deposit penetrated through the vertical cracks in the laserglazed top layer and affected the coating structure much as it did in the case of as-sprayed
coatings. The laser-glazed zone itself at the top of the coating was rather unaffected.
Thermal cycling resistance of the 8Y2O3-ZrO2 coating was lowered by the phosphate
sealing treatment. The reasons for the deterioration of the strain tolerance of the
phosphate sealed coating were the increased elastic modulus due to better cohesion of
splats and compressive internal stresses. Thermal cycling behaviour of the laser-glazed
8Y2O3-ZrO2 coatings was superior compared to the reference coating. Reduced elastic
modulus due to the macrocracks made the laser-glazed coating much more strain tolerant.
.
TABLE OF CONTENTS
PREFACE............................................................................................................................1
ABSTRACT .........................................................................................................................3
TABLE OF CONTENTS ......................................................................................................5
LIST OF INCLUDED PUBLICATIONS................................................................................9
LIST OF SYMBOLS AND ABBREVIATIONS ...................................................................11
1.
INTRODUCTION ........................................................................................................13
1.1
1.1.1
Gas turbine...................................................................................................13
1.1.2
Diesel engine................................................................................................14
1.2
1.3
1.4
TBC materials......................................................................................................16
1.4.1
1.4.2
1.5
1.5.1
1.5.2
1.5.3
1.6
2.
2.1.1
2.1.2
2.1.3
2.1.4
2.2
2.2.1
Microscopy ...................................................................................................27
2.2.2
2.2.3
2.3
2.3.1
Microhardness..............................................................................................29
2.3.2
2.3.3
Erosion resistance........................................................................................30
2.3.4
Abrasion resistance......................................................................................30
2.4
3.
2.4.1
Thermal expansion.......................................................................................30
2.4.2
2.4.3
Specific heat.................................................................................................30
2.4.4
2.5
2.6
Microstructural characterisation...........................................................................34
3.1.1
3.1.2
3.1.3
3.1.4
3.2
3.2.1
Microhardness..............................................................................................43
3.2.2
Elastic properties..........................................................................................43
3.2.3
3.2.4
Wear properties............................................................................................46
3.3
3.3.1
Thermal expansion.......................................................................................47
3.3.2
3.3.3
3.4
3.4.1
3.4.2
3.4.3
3.4.4
3.5
4.
3.5.1
3.5.2
3.5.3
3.5.4
REFERENCES ..................................................................................................................72
Thermal diffusivity
Bulk density
Poissons ratio
Diffraction angle
4PB
Four-point bending
a, w, h
AP
APS
ATCS
CP(T)
CTE
CVD
Youngs modulus
EB
Bending modulus
EB-DVD
EB-PVD
EDS
ESEM
FGM
HIP
HVOF
IA
Image analysis
k(T)
Thermal conductivity
LASER
Laser-glazed coating
LPPS
MP
Mercury porosimetry
OM
Optical microscopy
OPA
RB
SAED
SEG
SEM
SOLGEL
m-ZrO2
Monoclinic zirconia
c-ZrO2
Cubic zirconia
t-ZrO2
t-ZrO2
Tetragonal zirconia
Youngs modulus
wt%
Weight percent
DGUN
TBC
TEM
TGO
TTBC
VPS
vol%
Volume percent
XRD
X-ray diffraction
1. INTRODUCTION
Thermal barrier coatings (TBCs) have been used since the 60s in thermal protection of
gas turbine hot section components [1,2]. From the early 1980s, many investigators have
applied TBCs to the combustion chambers of diesel engines as well to lower heat losses.
[3-6]. As a TBC material, most investigators have used zirconia (ZrO2), partially stabilised
by magnesia (MgO), calcia (CaO) or yttria (Y2O3), because of its low thermal conductivity,
high temperature stability and relatively high coefficient of thermal expansion (CTE)
compared to other ceramic materials. Traditional TBCs have been manufactured by
atmospheric plasma spraying (APS) using the partially stabilised zirconia in powder form
as the raw material for coating.
Surface temperature of metallic components working at high temperatures can be reduced
by 100-300oC by using TBCs [7,8]. This temperature drop is significant considering the
mechanical properties of the structural materials, such as cobalt or nickel based
superalloys. In practice TBCs can extend the maintenance interval and component
lifetime. On the other hand TBCs make it possible to improve the process efficiency by
increasing the combustion temperature. Continually increasing process temperatures set
high requirements for TBC development too. Fig. 1 illustrates the effect of TBC on the
temperature gradient of a diesel engine piston head.
base
material
bond
coat
TBC
combustion
chamber
Temperature
cooling
system
Examples of TBC coated gas turbine components are presented in Fig. 2. On the firststage vanes of a gas turbine the coating thickness is normally in the range of 250-500 m
and in the combustion chamber component even 500 -1000 m. Weight and aerodynamic
considerations limit the coating thickness on rotating parts, such as blades, to 125-380 m
[16]. The large-scale industrial use in gas turbines of thick TBCs (> 1.0 mm) is still rather
limited.
a)
b)
c)
Fig. 2. TBC coated gas turbine components: a) first-stage vane, b) burner can and c) heat
shield of a combustor.
1.1.2 Diesel engine
Mean component surface temperatures in diesel engines are much lower than in gas
turbines. However, in a diesel engine almost 30 % of the fuel energy is wasted due to heat
losses through combustion chamber components [4]. For that reason, lots of research
activity has focused on applying TBCs to diesel engines. Fig. 3 a illustrates a crosssectional view of the diesel engine combustion chamber and points out the components
that might be effectively coated with TBC. Fig. 3 b presents a TBC coated piston head of
a test engine.
a)
b)
TBCs the coefficient of thermal efficiency of the diesel process can be increased or fuel
consumption lowered [19,20]. Some published studies also trace the effect of TBCs on
reducing diesel engines emissions [21,22]. In any case, the diesel process had to be
adjusted correctly to realise the benefits of TBCs.
1.2
a)
b)
Fig. 4. Plasma spray process: a) principle of plasma spraying [25] and b) robotized plasma
spraying of a gas turbine transition duct.
In modern plasma spray systems all the spray parameters can be computer controlled.
Component handling as well as spray gun movement can be fully automated. Many
plasma gun designs are available for different types of component geometries. These
factors have made plasma spraying, which is a very sensitive process with various
parameters, easier and more reliably consistent. In the last decades numerous studies
have focused on plasma spray parameters and their effect on the microstructure of TBCs
[26-28]. During the last ten years on-line diagnostics has brought a new dimension to the
understanding of the relationship between plasma spray parameters and microstructure
[29-31]. By using on-line diagnostics researchers can collect information about in-flight
spray particles in a plasma plume and monitor possible changes in the spraying process.
TBCs are also increasingly manufactured by electron-beam physical vapour deposition
(EB-PVD) [32-36]. EB-PVD TBCs are mostly used in first-stage blades of gas turbines.
The strain tolerant microstructure and aerodynamically beneficial smooth surface of EBPVD coating suit them very well for that type of component. However, manufacturing costs
of EB-PVD coatings are higher than those of APS coatings [16] and the coating process is
not flexible enough to use when coating large components, for instance, or components
with complex shapes or inside diameter surfaces [13]. Some other coating processes such
as chemical vapour deposition (CVD) techniques [37-38] and electron beam directed
vapour deposition (EB-DVD) [39] have been lately studied as alternatives to conventional
EB-PVD.
15
1.3
The thermal barrier coating system consists of a thermal insulation layer and bond coating.
The typical thickness of a TBC layer is 150-500 m and 150-250 m for bond coating.
Schematic illustrations of plasma sprayed and EB-PVD TBC structures are presented in
Fig. 5. The properties required from a TBC layer are low thermal conductivity, high stain
tolerance, long-term stability at high temperatures, good erosion and hot corrosion
resistance. The lamellar and porous microstructure of plasma sprayed coating is
advantageous if considering low thermal conductivity and strain tolerance, but erosion and
hot corrosion properties can be moderate. APS TBC is mechanically bonded to the bond
coat, whereas chemical bonding is formed in EB-PVD coating (due to thermally grown
oxide (TGO)). The columnar microstructure of EB-PVD coating is extremely strain tolerant
[9,32,40], but its thermal conductivity is higher than that of plasma sprayed coating
[9,32,41,42].
bond coat
superalloy
superalloy
a)
b)
TBC materials
As stated earlier, the TBC material should have low thermal conductivity k(T) and a CTE
close to those of the metallic bond coatings and substrates. It should also have long-term
phase stability at whole service temperature range and adequate corrosion resistance
against impurities present in the process (such as Na, S, V). TBC material should have a
low sintering tendency to maintain the strain tolerant microstructure. Sufficient mechanical
properties are also needed (bond strength, erosion resistance). Various materials, mostly
oxide ceramics, have been studied as TBC candidates. Partially stabilised zirconia is the
most used TBC material, and 8Y2O3-ZrO2 has been the industrial standard composition for
years. The following two chapters present the partially stabilised zirconia structures as well
as the other TBC material alternatives.
16
powders are agglomerated and sintered (manufactured by spray drying) and part of them
further plasma densified. Spray drying gives excellent possibilities to vary spray powder
composition and particle size distribution and even grain size of primary particles.
1.4.2 Other TBC materials
The limited maximum service temperature of partially stabilised zirconia coatings has
prompted researchers to seek totally new material alternatives for very high temperatures
[79,80]. Promising results have been reported for high temperature stability of lanthanum
zirconate (La2Zr2O7) [81,82] and lanthanum hexaluminates [83]. Glass-matrix structures
[84,85] and NZP (NaZr2P3O12) [86] have also been studied lately as TBC materials. Mullite
(3Al2O32SiO2) has been studied for its good hot corrosion and high thermal stability. Due
to its relatively low CTE it may prove useful for coating diesel engine piston heads where
the local temperature variation might be very high. In diesel engine tests, reported by
Yonushonis [87], mullite based multilayer coating performed better than zirconia coatings.
Various oxides, silicates and titanates have been proposed for TBC materials [88-92].
However, even if most of these other TBC materials offer some improved features, they
still have not surpassed the good overall properties of yttria stabilised zirconia or they are
not yet commercially available.
1.5
No exact definition exists for the thickness of thick thermal barrier coating, but generally
the term has been used with TBCs thicker than 0.5 mm. The following chapters explain the
motivation to develop TTBCs as well as their potential use in gas turbines and diesel
engines. In addition, the chapters discuss the drawbacks and risks of thick coatings and
present the state-of-the-art TTBCs with modified microstructures as well as other potential
modification procedures.
1.5.1 Demand for thicker coatings
More efficient thermal insulation of the hot path components of state-of-the-art gas
turbines is needed because of the increasing demands of higher process temperatures
and the limited service temperatures of present superalloys. Higher combustion
temperatures improve process efficiency and fuel economy. In gas turbines the
temperatures of the hot path component are mainly controlled by various cooling
techniques like film cooling and serpentine cooling as well as by thermal barrier coatings.
Although component air-cooling is essential, it is not sufficient for controlling component
surface temperatures. For that reason lower thermal conductance (thermal conductivity of
the coating/coating thickness) TBCs are extensively developed. The lowering of thermal
conductance of TBCs can be achieved in three ways: 1) lowering the thermal conductivity
of the coating material, 2) lowering the thermal conductivity by increasing the porosity of
the coating and 3) increasing the thickness of the coating. When tailoring new thermal
barrier coatings, all these ways should be considered. Calculations have shown that a
traditional 500 m thick TBC effects a temperature drop in the range of 150oC, but a 1.8
mm thick TBC produces a drop of 320oC (if the coating surface temperature is 1250oC)
[93]. TTBCs could be used in the static components of gas turbines like heat shields in
combustion chambers, combustor cans, transition ducts and afterburners (aeroengines).
There are some studies in literature where TTBCs in gas turbines have been reported
[15,94-96] containing service or laboratory testing of real gas turbine components.
TTBCs have been studied for diesel engines since the advent of the idea of the adiabatic
diesel engine [3,87,97] or the low heat rejection engine [98,99]. Most of the TTBCs studies
18
for diesel engines have been focused on small and medium sized diesel engines, used in
vehicles and ships. TTBCs could potentially be utilized in high-powered diesel engines
(even up to 80 MW) designed mainly for marine and power station use. The basic goal of
applying TTBCs in diesel engines has been to minimize the heat losses through the
combustion chamber components. Since 30 % of the heat losses of combustion chamber
wall structures flow through the piston [4] it has been the component most often targeted
for applying TTBCs. Piston head coatings of up to 3.5 mm thick have been studied in order
to minimize the heat losses and to reach the targeted temperature drop through the
coating [100]. If the heat losses of a diesel engine were lowered, the extra heat, available
for the exhaust gases, could be converted in a flue gas boiler to heat or electricity or in a
turbocharger to mechanical energy. In such ways the total process efficiency could be
improved. Several studies have been published documenting the testing of TTBC coated
diesel engine components [87,101-104].
1.5.2 Drawbacks of TTBCs
Several studies [96,105,106] have shown that, as the thickness of plasma sprayed TBCs
increases, their reliability deteriorates, especially when exposed to thermal cycling. So only
increasing the coating thickness, without modifying the coating microstructure, will not
produce strain tolerant thick thermal barrier coatings. With thicker coatings the problems
with residual stresses, originating in the coating manufacturing, are emphasized. When the
coating thickness is increased by introducing more spray passes, the substrate and
coating temperature rises step by step unless adequate cooling is used. This temperature
increase reduces the cooling rate of individual splats and leads to better contact of
lamellae and decreased number of vertical microcracks in lamellae. These are the
mechanisms through which the tensile (quenching) stresses impact the coating. After the
spraying, when the component cools down, compressive (thermal) stress is induced to
TBC (CTETBC < CTESUBSTRATE). The final residual stress state of the coating is a sum of all
the stress components, in this case mainly the quenching and thermal stresses. The
formation of residual stresses (or strains) in plasma sprayed coatings and TBCs has been
widely studied [107-111]. It has been reported that residual stresses in plasma sprayed
TBCs can be tensile or compressive and can be affected by controlling the substrate
temperature during spraying [112-114]. In the same studies it was also reported that the
stress state change in high temperature exposure is towards compression. Considering
the combined effect of residual stresses and the stresses caused by thermal cycling loads
on TBCs, the residual stresses, as low as possible, should be beneficial. The bond
strength or the intrinsic cohesion of the coating is also lowered in thicker coatings [106].
The following chapter will discuss how the stresses can be affected in plasma sprayed
TTBCs.
All these drawbacks of traditionally prepared TTBCs, residual stresses, low bond strength
and low strain tolerance, combine to lower the reliability of the coating. With increased
coating thickness the temperature drop through the coating increases at service
temperatures and at the same time the dimensional mismatch of the coating surface and
bond coat interface becomes higher, due to low strain tolerance. This dynamic induces
more stresses into the structure and increases total strain energy available for crack
initiation. Typically with TTBCs the crack is initiated near the bond coat interface leading to
macroscopic coating delamination. In practice the coating failure mechanism is not so
simple: varying thermal loads due to thermal cycling, thermal shocks and local hot spots
make the situation even more difficult.
19
Several other risks have to be taken into account when considering the use of TTBCs in
modern gas turbines, where the turbine inlet temperatures are extremely high (13501500oC). The use of thicker coatings generally leads to higher coating surface
temperatures that can be detrimental in many ways, if certain limit are exceeded: 1) The
phase structure of yttria stabilised zirconia 8Y2O3-ZrO2 is not stable above the 1250oC and
can destabilise quite rapidly at 1400oC [59,63], 2) sintering of the plasma sprayed zirconia
can take place already at 1200oC [115,116], that increase the coating stiffness and reduce
the strain tolerance of the coating, 3) the creep rate of the coating increases with higher
temperatures, which still can weaken the strain tolerance of the coating [116,117].
The literature also contains accounts of some diesel engine experiments with TTBC
coated piston heads in which the coating lifetime has been poor [87,101-103,118]. In the
piston head surface, the local stresses on the coating can be very high in hot spots where
the fuel is injected. Even if the mean surface temperatures of the TBC in diesel engine
remains at lower level than in gas turbines, the surface temperature swing during the one
engine cycle can be 240-350oC higher [19,119]. Pressure variations in the combustion
chamber and the high velocity of the piston exacerbate the severe high cycle fatigue
loading on the piston head surface.
1.5.3 Microstructural modifications of TBCs
Modification of the microstructure of plasma sprayed TTBCs as well as traditional thin TBC
has been widely studied as a means of improving a variety of coating properties such as
strain tolerance, thermal conductivity, hot corrosion and erosion resistance. In this work
the modification processes have been divided into three classes. Class A includes
processes where the coating structure is influenced during manufacturing, through
processes such as spray parameter controlling and special cooling techniques. Class B
contains modifications in which the coating structure is not the typical double layer, but a
graded or multilayered structure. Class C includes different types of post treatments such
as sealing, densification and surface remelting processes. Each of these classes is
presented in more detail in the following paragraphs.
A) In the case of TTBCs the structural modifications have been mainly concentrated on
lowering the Young's modulus (E) and residual strains/stresses of the coating for obtaining
better strain tolerance [93,94,106,120]. This modification has been approached by
introducing segmentation cracks [95,120] or a special microcrack network into the coating
structure [121-123] or by increasing the coating porosity [93].
Vertical segmentation cracks can be obtained by using rather thick spray passes, short
spray distance and particular substrate preheating [120]. A. S. Grot et al. [124] as early as
1981 studied the segmented 6Y2O3-ZrO2 structures where the vertical macrocracks went
through the whole coating thickness. In burner rig type hot corrosion tests with 30.5 l/h
SO2 gas, 20 ppm sea salt at 704oC and 899oC, they showed that some corrosives
penetrated into the segmentation cracks. The overall performance of the segmented
coatings in burner rig tests was good. D. Schwingel et al. [120] and P. Bengtsson [95]
found in their studies that the lifetime of the segmentation cracked TTBC was significantly
better compared to normal TTBC structure. At the same time the Youngs modulus of the
coating was much reduced. Several studies [121-123] of atmosphere and temperature
controlled spraying (ATCS) have been reported. In the ATCS technique cryogenic surface
cooling is used during spraying in order to intensify the formation of microcracks in the
lamellae. Microcracks are formed due to the increased cooling rate of the splats. By ATCS
it was possible to improve coating strain tolerance and thermal cycling lifetime as well as
20
to reduce coating residual stresses [122,123]. H.-D. Steffens et al. [106] presented results
for TTBCs of reduced residual stresses and improved thermal shock resistance when
using various cooling techniques in plasma spraying. In plasma spraying it is possible to
affect the TBC porosity to some degree. However, the normal porosity of TBCs is already
at a rather high level (10-15 %) and further porosity increase by changing spray
parameters could be difficult. Extremely high porosity values, up to 25 vol%, of TBCs have
been obtained by spraying polymers together with zirconia [93]. Increasing the coating
porosity decreases thermal conductivity and Youngs modulus is expected to decrease
too.
Some further drawbacks should be taken into account as well. Due to the increased
number of cracks and pores the mechanical properties like adhesion and cohesion,
erosion resistance and hot corrosion resistance of the modified TTBCs, presented in
previous paragraph, might be slightly weakened.
B) Many studies have focused on functionally graded materials (FGMs) in order to improve
the properties of TTBCs. The gradient has often been constructed by mixing the starting
material powders TBC and MCrAlY (bond coat) in various fractions. In many cases the
focus has been on lowering the critical stresses in the structure caused by differences in
the CTEs of the coating and substrate material [104,125-130]. But also other properties
such as enhanced erosion resistance [130] and bond strength [128] as well as lowered
oxygen transport in TBC [126] have been reported. However it should be remembered that
the metal phase in graded structures have very large surface areas and for that reason are
susceptible to oxidation at high temperatures.
C) Lots of work has been done in modifying the properties of the TBCs by various post
treatment processes. Post treatments, such as different sealing treatments and surface
remelting and densification procedures, have been used mainly for improving the hot
corrosion and erosion resistance of the coatings by closing the open pores on the coating
surface. Most of these studies have focused on thin TBCs (< 1 mm).
A. Ohmori et al. [131,132] studied sealing of TBCs by liquid manganese and manganese
alloys (Mn-Cu, Mn-Sn, Mn-In). With the liquid metal impregnation it was possible to
increase elastic modulus, microhardness and fracture toughness of the coatings.
I. Zaplatynsky [133] studied the effect of laser glazing (CO2 laser) on the microstructure
and properties of 8Y2O3-ZrO2 coatings. The lifetime of the laser-glazed coatings was
extended four times in burner rig type hot corrosion tests, where 100 ppm of NaCl + 0.05
wt% S in fuel was used at Tmax=843oC. The result was explained by the reduced
permeability of the coating surface. Laser glazing did not affect the coating behaviour in
cyclic oxidation tests, even if there were vertical cracks in the coating. R. Sivakumar et al.
[134] performed a comprehensive study of the CO2 laser melting of the plasma sprayed
CaO, MgO and Y2O3 stabilised zirconia coatings. In the hot corrosion exposure to molten
salt of 95Na2SO4-5NaCl at 950oC for 100 h, the laser-glazed zirconia coatings performed
worse than the as-sprayed ones. The melt deposit penetrated into the vertical cracks,
induced by laser glazing, and caused severe oxidation of the bond coat. H. L. Tsai et al.
[135,136] studied sealing of 6-20 wt% yttria stabilised zirconia TBC coatings with CO2
laser. Coatings were exposed to thermal cycling/oxidation tests in which the coatings were
kept at 11005oC for 1 hour and then cooled to ambient temperature in 10 minutes by
pressurized air. They did not find any effect of laser glazing on the bond coat oxidation, but
the lifetime in thermal cycling tests was increased by 2 -6 times, depending on the coating
21
composition. A. Petitbon et al. [137] studied surface melting and over-cladding of the Y2O3
and Y2O3/HfO2 stabilised zirconia coatings by CO2 laser. The cladding was made using
Al2O3 powder. Laser treatments improved thermal cycling, Tmax 1200oC, dwell 5 min, Tmin
100oC, dwell 5 min, properties as well as friction and erosion resistance. Finally the Al2O3
cladded TBC coatings were proved to be superior in an in-service test, where adjacent
flaps of the FALCON F16 fighter turbine were tested for 150 hours. K. A. Khor et al. [138]
performed sealing experiments with Nd-YAG laser for 5CaO-ZrO2 coatings.
Microhardnesses of properly melted surface areas were doubled if compared to assprayed coating. A. Zhou et al. [139,140] studied the hybrid spray process, combined
plasma spraying and Nd-YAG laser, in manufacturing 8Y2O3-ZrO2 coatings. It was found
that coating microhardness and wear resistance were increased.
H. Kuribayashi et al. [141] studied densification of TBC coatings by the hot isostatic
pressing (HIP) process. They found that mechanical properties of the coatings increased
significantly, hardness from 5 GPa to 13,3 GPa, tensile strength from 5 MPa to 60 MPa. K.
A. Khor et al. [142,143] studied HIPing of the 8Y2O3-ZrO2 and 5CaO-ZrO2 coatings.
Coating porosity was reduced whereas thermal diffusivity and microhardness was
increased.
K. Moriya et al. [144,145] studied sealing of plasma sprayed coatings by the sol-gel
process, where Al2O3 and SiO2 based precursors were impregnated into Al2O3 and 8Y2O3ZrO2 coatings. Metal alkoxides, Al(OC3H7)3 and Si(OC2H5)4, together with water and HCl,
were used as starting materials for Al2O3 and SiO2 based precursors. Adhesive strength of
the coatings increased significantly due to the sealing process. Porosity of the coatings
was also reduced. G. John et al. [146] made sealing experiments for 8Y2O3-ZrO2 coatings
with alumina and silica based sol-gels. Potentiodynamic polarization tests in aqueous 3
wt% NaCl solution and gas permeability tests showed the reduction of coating open
porosity as a function of impregnation time. Coating adhesion was also improved. I.
Berezin et al. [147] used a silica based precursor (pre-hydrolyzed ethyl silicate, Si(OC2H5))
in sealing 8Y2O3-ZrO2 coatings. Microhardness of the sealed coatings was increased even
if it was estimated that only 1/10 of the open porosity could be sealed with one infiltration
cycle. J. Kathikeyan at al. [148] made sealing experiments for free standing 8Y2O3-ZrO2
coatings with aqueous based aluminium hydroxide precursor. Mercury porosimetry
showed the porosity reduction and the change of the pore size distribution. T. Troczynski
et al. [149,150] studied physico-chemical sealing treatments for yttria stabilised ZrO2
coatings with sol-gel impregnation and laser glazing (CO2 laser). They also performed
laser glazing for sol-gel sealed specimens. In thermal shock tests at Tmax=1270oC and air
cooling, the sol-gel sealed coatings lasted longer than as-sprayed coatings, but the laserglazed as well as the sol-gel sealed + laser-glazed coatings performed best.
Borisova et al. [151] sealed flame sprayed zirconia coatings by phosphate based sealants.
In sealing experiments they used aluminium-chromium phosphate and orthophosphoric
acid (H3PO4). It was found that the sealing treatment strengthened the coating structure.
22
1.6
The aim of the study was to improve the properties of thick thermal barrier coatings by
modifying their microstructures by several post treatments, mainly concentrating on
phosphate sealing and laser glazing. Phosphate sealing was mainly performed in order to
densify the surface layer of the porous plasma sprayed TTBC. The purpose of the surface
densification processes was to increase erosion and hot corrosion resistance of TTBCs
without deteriorating the other important coating properties such as thermal conductivity
and strain tolerance. The coating microstructures were modified also by laser glazing to
densify the surface of the coating and to introduce a special crack structure into the
coating. In laser-glazed coatings, in addition to erosion and hot corrosion resistance, also
the strain tolerance was expected to improve if beneficial vertical macrocrack networks
could be created.
The study started with the optimisation of each modification procedure and continued with
coating microstructural characterisation. Then the mechanical, wear and thermal
properties of the coatings were determined, and finally their high temperature behaviour
was tested in hot corrosion and thermal cycling experiments. At a rather early stage of the
study the phosphate sealing and laser glazing seemed to be the most promising ways to
affect coating microstructures. For that reason this thesis mostly focuses on the results of
these two modification processes and only briefly discusses the other processes, such as
sol-gel sealing and dense overlay coatings prepared by detonation gun spraying.
23
2. EXPERIMENTAL PROCEDURES
This chapter introduces the materials, coatings and coating modification procedures
studied in this work and describes the characterisation and testing methods used.
2.1
Coatings were produced by thermal spraying techniques. Ceramic TTBCs and their bond
coatings were prepared mainly by APS. In some special cases HVOF and detonation gun
spray processes were applied. All the coatings were sprayed using commercial feedstock
powders. Most of the coating modification procedures were post-treatments which were
made for as-sprayed coatings.
2.1.1 Reference coatings and substrate materials
Zirconia based TTBCs (8Y2O3-ZrO2 and 25CeO2-2.5Y2O3-ZrO2 and 22.5MgO-ZrO2) were
air plasma sprayed with plasma spray equipment (Plasma-Technik A3000S, Sulzer Metco
AG, Wohlen, Switzerland) using a F4 plasma gun. Bond coatings were sprayed using
either APS or HVOF systems. The HVOF spraying was done by Diamond Jet Hybrid 2600
HVOF gun (Sulzer Metco AG, Wohlen, Switzerland). Before applying zirconia the HVOF
bond coat was diffusion heat-treated for 2 h at 1120oC and for 24 h at 845oC. Substrates
were cleaned and grit blasted before applying the bond coat. Surface roughness, Ra, after
the grit blasting with corundum of 40 grit, was at the range of 6-7 m. Coating temperature
was measured with a handheld infrared thermometer during the spraying and it was kept
below 200oC by pressurized air-cooling. The targeted coating thickness of TTBCs was 1.0
mm and 200 m for bond coats. The data of coating compositions and used powders with
main spray parameters are presented in Table 1.
Table 1. Nominal compositions of coatings, powder data and spray parameters.
Coating
abbreviation
Nominal
composition
Powder
tradename
Spray
process
8Y
8Y
25C
22M
A962
A995
8Y2O3-ZrO2
8Y2O3-ZrO2
25CeO2-2.5Y2O3-ZrO2
22MgO-ZrO2
Ni22Cr10Al1Y
Co32Ni2Cr8Al0.5Y
Metco 204NS*
ZRO-113/114**
Metco 205NS*
ZRO-103**
Amdry 962*
Amdry 995C*
APS
APS
APS
APS
APS
APS
SICOAT 2453
Ni10Co23Cr12Al0.6Y3Re
SICOAT 2453***
HVOF
Powder suppliers: * Sulzer Metco, Wohlen, Switzerland, ** Praxair, Indianapolis, IN, USA, ***H. C. Starck GmbH,
Laufenburg, Germany.
Several substrate materials were used in preparing the coating specimens for different
tests. Mild steel Fe37 (AISI 1023) was used with coatings in erosion, abrasion and fourpoint bending tests. Tempered steel 42CrMo4 (AISI4142) was used for samples prepared
for characterisation purposes and microhardness measurements. Alloy 600 and Nimonic
80A were substrate materials in hot corrosion tests and IN738 in thermal cycling tests.
Nominal compositions of substrate materials are presented in Table 2 on page 25.
In some cases the specimens had to be tested as freestanding coatings. Freestanding
coating specimens were etched from the substrates using 50HCl/50H2O solution. If
freestanding specimens were needed, the phosphate sealing procedure was made after
etching in order to avoid the reaction between the sealant and etchant.
24
Si
Mn
Cr
Mo
<
0.18
0.380.45
-
0.150-50
0.150.40
0.5
<
1.00
0.600.90
1.0
<
0.045
<
0.035
-
<
0.045
<
0.035
-
<
0.25
0.901.20
16.0
<
0.10
0.150.25
1.75
<
0.10
< 1.0
< 1.0
0.015
18.021.0
0.17
16.0
Ni
<
0.30
Fe
Co
Cu
Zr
Ta
Al
Ti
<
0.30
bal
bal
bal
8.0
bal
<
3.0
0.5
<
2.0
<
0.008
<
0.2
<
0.15
1.01.8
1.82.7
2.6
bal
8.5
1.75
3.4
3.4
a)
sealant infiltration
b)
c)
d)
25
beam
Laser source
a)
b)
Fig. 7. Schematic illustration of the laser glazing process and the surface of the laserglazed 8Y2O3-ZrO2 coating.
Laser glazing parameters were optimised by comparing coating microstructures with
different specific laser energy densities using continuous and pulsed laser beams. In the
optimisation stage the predetermined melting depth of the coating surface was reached,
without causing coating spallation. Also formation of too long vertical cracks, which pass
through the thickness of the coating, was avoided. The optimised laser glazing parameters
for studied coatings are presented in Table 3. Abbreviation LASER is used here for all
laser-glazed coatings.
Table 3. Laser glazing parameters for studied TTBCs.
Laser power [kW]
Surface speed [mm/min]
Surface distance from the mirror [mm]
Laser beam specific energy density [J/mm2]
8Y2O3-ZrO2
3.5-4.0
3500-4500
80
4.7-6.9
25CeO2-2.5Y2O3-ZrO2
3.0
4000
80
4.5
22.5MgO-ZrO2
3.5
4500
80
4.7
velocities, but still sufficient heat, obtained by the D-gun system. Spray parameters and
powder information are listed in Table 4. Abbreviation DGUN is used for detonation gun
sprayed coatings.
Table 4. Powder data and spray parameters for detonation gun sprayed coatings.
Nominal
composition
Powder
tradename
Powder
supplyer
8Y2O3-ZrO2
65ZrO2-35SiO2
Cr2O3
Amperit 727.054
Amperit 840.1
Amperit 706.072
H.C Starck
H.C Starck
H.C Starck
Acetylene flow
rate [l/min]
12
12
12
O2 flow rate
[l/min]
21
21
25
Several methods were used to characterise the relationship between the effect of coating
modification processes and the structure/behaviour of the materials. The microstructures
were studied by microscopy and phase structures by x-ray diffraction (XRD). The influence
of the sealing treatments on coating densification was studied by porosity measurements.
2.2.1 Microscopy
Optical microscopy (OM) with magnification range of 10x-100x was used in the
examination of the coating overall microstructure. Three systems were used, namely Leitz
27
(Wetzlar, Germany), Versamet 3 (Union Co., Japan) and Carl Zeiss Axiophot (Germany).
Scanning electron microscopy (SEM/ESEM, model Model XL-30, Philips, Eindhoven,
Netherlands) was used with higher magnifications (100x-10 000x). Energy dispersive
spectrometry (EDS, Model DX-4, EDAX International, New Jersey, USA) was used in
elemental analysis in SEM studies. Transmission electron microscopy (TEM, Model JEM
2010, Jeol, Tokyo, Japan) was used at magnifications higher than 10 000x. In TEM studies
selected area electron diffraction (SAED) was used to study the crystal structures.
Cross-sectional samples for microscopy were cut by a precision cut-off machine and cold
mounted in vacuum. Specimens were grinded by diamond grinding discs or by SiC papers.
The final polishing was carried out by polishing cloths using diamond spray or diamond
paste. In SEM investigations, where electrical conductivity of the sample is required, a thin
layer of gold or carbon was sputtered on the specimens.
2.2.2 X-ray diffraction
X-ray diffraction was used in phase identification, quantitative phase analysis, texture
determination and residual stress studies.
The phase compositions of the coatings were identified with X-ray diffractometer (XRD,
Siemens D500, Karlsruhe, Germany) using CuK radiation with scan step of 0.02o and
step time of 1.2 s. For more detailed quantitative phase analysis image plate X-ray
diffractometer (XRD, Italstructures, Riva del Garda, Italy) was used. The image plate XRD
system worked with CuK radiation operating at 40kV and 30mA. The used exposure time
was two hours and the analysed spectra were taken from 2 range of 20-120o. The
constant incident angle () between the x-ray source and the specimen surface was 15o.
The image plate (x-ray sensitive film) diffraction pattern was scanned into a computer and
the data was analysed using MAUD software (Material Analysis Using Diffraction, version
1.87 (Luca Lutterotti, University of Trento, Italy). In MAUD software the quantitative and
texture analyses were carried out by the Rietveld method [152,153].
Residual stresses were measured using a XStress3000 stress analyser (Stresstech Oy,
Vaajakoski, Finland). CrK -radiation was used with 30 kV, 5.0 mA and 30 s exposure
time. The traditional sin2 -method was carried out using specimen tilts of =0o, 21.8o,
31.7o and 40o. In there the least squares method was used in fitting the measured
points to a line (d(sind2) graph). Error of each measurement, presented as error bars in
results, is an average error that expresses the goodness of fit of points to a line. The peak
shifts of zirconia coatings were studied on (3 1 3) crystalline plane of tZrO2 at 2 position
of 153o. Bulk material constants E = 205 GPa and = 0.23 for zirconia were used in stress
calculations. Through thickness stress profiles were determined by repeating the
measurements and layer removal steps. Layers were removed with careful grinding to
avoid producing additional stresses or cracks.
2.2.3 Porosity and bulk density determination
Total porosity was evaluated from the coating cross-section by image analysis (IA) using
optical microscope (Carl Zeiss Axiophot, Germany) and image acquisition and analysis
software (QWin, Leica Microsystems, Switzerland). The results are presented as a mean
value with standard deviation of five separate analyses from each type of coating. Open
porosity was measured with mercury porosimetry (MP, models Pascal 140 and
Porosimeter 2000, CE-instruments, Milan, Italy) over the pressure range of 0.1 kPa 200
MPa. Bulk density of the coating was determined by the method of Archimedes [154].
28
2.3
P
2
bending specimen
P
2
P
2
RB =
3Pmax a
wh 2
a (3L2 4a 2 ) P
EB =
wh 3
(1)
(2)
Thermal expansion studies were used to study the high temperature stability of the
modified coating structures. As low thermal conductivity is one of the most important
features of TBCs, thermal diffusivity and specific heat measurements were carried out.
2.4.1 Thermal expansion
Thermal expansion studies and the determination of CTE were carried out by dilatometer
(Adamel Lhomargy SAS, model DI-24, France) in air at a temperature range of 50-1300oC.
The temperature ramping rate varied from 5oC/min to 10oC/min and dwell times at
maximum temperature from 5 minutes to 5 hours.
2.4.2 Thermal diffusivity
Thermal diffusivity, (T), measurements were carried out with laser flash apparatus Theta
(Theta Industries Inc., Port Washington, NY, USA) in vacuum (< 0,01 Pa). Measurements
were performed at 7 different temperatures in the temperature range of 100-1300oC.
Measurements were repeated five times at each temperature for statistical reasons. Prior
to evaluating the thermal diffusivity, in order to make the sample surfaces opaque, thin
layers of colloidal graphite were painted on both the front and the rear faces. The
measurement cycle was repeated 3 times for each coating in order to find out the effect of
high temperature exposure of the previous measurement on (T).
2.4.3 Specific heat
Specific heat, CP(T), measurements were performed by a Differential Scanning
Calorimeter DSC 404 C (Netzsch-Gertebau GmbH, Selb, Germany). The scanning rate
was 15C/min at the temperature range of 100C up to 1250C. Measurements were
carried out in air and in argon atmospheres using either alumina or platinum crucibles.
Weight of the free-standing coating specimen was approximately 80 mg. For each sample
three consequent measurements cycles were performed in order to lower the statistical
error of the measurement.
30
k (T ) = (T ) C P (T ) B
(3)
, where (T) is thermal diffusivity, CP(T) specific heat at constant pressure and B bulk
density,
2.5
Hot corrosion resistance of the modified TTBCs was tested at 600-850 C for 48-1000
hours. Coatings were exposed to mixtures of vanadium pentoxide (V2O5) and sodium
sulphate (Na2SO4) in order to simulate the deposits and temperatures present in a diesel
engine combustion chamber. The first test series were made with a mixture of 65Na2SO4 35V2O5 (mol-%) for 48 and 200 hours. The mixed starting materials were melted in an
alumina crucible at 800oC for 2 hours, and the solidified deposit was crushed manually
after the heat treatment. This solid deposit in powder form was placed on each specimen
before the test (5-10 mg). The second test series was made with 18Na2SO4 - 82V2O5 (mol%) mixture for 100-1000 hours. Starting materials were mixed with ethanol in order to get
better mixing and easier application. The deposit (approximately 0.1 g) was spread on the
coating surface every 100 hours. The test furnace was cooled down to RT at these 100 h
intervals. A phase diagram of the Na2SO4 - V2O5 system is presented in Fig. 9. The
mixture compositions are marked into the phase diagram by numbers 1. and 2. according
to test series.
Fig. 9. Phase diagram of the Na2SO4 - V2O5 system and the used compositions [156].
31
2.6
Thermal cycling resistance of the modified TTBCs was studied in a thermal cycling facility,
illustrated in Fig. 10. Two separate specimen holders were used simultaneously with 8
specimens mounted on each holder. Fig. 10 b illustrates the positions of the heating and
cooling stations of each rotation of one holder. At the heating station the specimen was
heated up by an oxyacetylene burner. A maximum temperature of the coatings, 10001300oC, was fixed with the burner distance (L) from the coating surface and with the
burner gas flow rates, see Fig. 10 c. At the heating station the temperature gradient
through the coating was emphasised by backside pressurised air-cooling. At the primary
cooling station the specimens were cooled from the front and back by pressurised air. At
the top position of the holder there was an optional front side cooling station, which was
used only in part of the tests.
Additional
cooling station
Heating
station
(+ backside
cooling)
Front- and
backside
cooling
Automatic stepwise
rotation (1/8 r)
a)
b)
Specimen casing
and mounting
Oxyacethylene burner
c)
Pressurized air
backside cooling
Fig. 10. Illustration of the thermal cycling device: a) photo of the upper specimen holder
during the test, b) positions of heating and cooling stations and c) side view of the
specimen mounting.
The test parameters of each series are presented in Table 5. Test series 2 was divided in
two parts a and b, since the additional cooling was used only after the first 500 cycles.
32
Test 1
70
100050
1,2
30025
20
500
Test 2 a
60
115050
1,2
30025
20
500
Test 2 b
60
115050
1,2,3
30025
20
500
Test 3
55
130050
1,2,3
30025
20
230
* 1 = backside cooling at the heating station, 2 = front and backside cooling at primary cooling station, 3
= additional front side cooling.
One or two samples of each material were used in all test series. Coating was rated to be
a failure by visual inspection when 10 % of the coating was peeled off. At the beginning of
each test series the coating surface temperatures were roughly calibrated. In the
calibration phase the coating surface temperatures were measured by pyrometer based
thermal camera (ThermaCam PM 595 FLIR Systems, Portland, OR, USA) working at a
spectral length of 7.5-13 m. Since the heating cycle was the same for all specimens the
surface temperature of the phosphate sealed coatings remained 50-100oC lower than with
other coatings. This fact explains the maximum temperature window ( 50 oC), presented
in Table 5.
33
Microstructural characterisation
8Y AP
8Y LASER
Fig. 11. Optical micrographs of the as-sprayed and modified 8Y based TTBCs.
34
25C
25C AP
25C LASER
Fig. 12. Optical micrographs of the as-sprayed and modified 25C based TTBCs.
35
22M
22M OPA
22M LASER
Fig. 13. Optical micrographs of the as-sprayed and modified 22M based TTBCs.
Results of the porosity and bulk density measurements are presented in Table 6. The
results of heat treated specimens appear in the grey columns and are discussed later in
chapter 3.3.2. The modified coatings were only analysed in the region of the sealed top
layers. Total porosity was determined by image analysis, and open porosity by mercury
porosimetry and the method of Archimedes [Publications III,IV].
Phosphate sealing reduced the total porosity of the coatings by 30-39 % [Publication III].
Total porosity values of the as-sprayed and phosphate sealed coatings were likely affected
by the pull-outs that were introduced in specimen preparation. This result is typical for
plasma sprayed oxide coatings. However, it could be assumed that the amount of pull-outs
was lower in phosphate sealed coatings in which the cohesion of lamellae was increased
due to sealing. So the measured total porosity reduction in phosphate sealed coatings was
partly due to reduced porosity, but also due to reduced amount of pull-outs. Porosity
reduction in the phosphate sealed coatings was seen also in open porosity measurement
where the reduction ranged from 24-48 % [Publication III]. It should also be noted that the
mercury porosimetry result represents the mean open porosity of the entire coating, so the
real open porosity of the sealed coating surface might be even lower than reported here.
36
Table 6. Porosity and bulk density of the reference and modified TTBCs.
Coating
Total Porosity,
Image analysis
[ Vol.-%]
Original
8Y
8YL
8Y AP
25C
25CL
25C AP
22M
22ML
22M OPA
20.7 1.8
2.8 2.6*
12.6 1.9
18.4 3.3
4.9 2.1*
12.9 2.4
12.1 2.2
3.3 1.6*
7.5 1.6
Heat
treated
9.4 0.6
2.2 0.9*
6.9 1.1
9.8 0.6
1.4 0.7*
5.4 0.6
8.2 1.1
3.4 1.3*
12.0 2.0
Open Porosity,
Mercury porosimetry
[Vol.-%, 1%]
Original
9.3
n.a
5.3
10.4
n.a
5.4
9.5
n.a
7.2
Heat
treated
10.0
n.a
9.0
8.5
n.a
5.4
13.7
n.a
9.0
Open Porosity,
Archimedes
[Vol.-%, 1%]
Original
9.0
n.a
3.9
7.5
n.a
5.2
13.4
n.a
3.9
Heat
treated
9.1
n.a
5.9
8.0
n.a
5.4
11.9
n.a
5.3
Bulk density,
Archimedes
[g/cm3, 0.1 g/cm3]
Original
5.3
n.a
5.4
5.6
n.a
5.7
4.2
n.a
4.4
Heat
treated
5.4
n.a
5.4
5.7
n.a
5.7
4.5
n.a
4.8
* reliable porosity measurement for the laser-glazed coatings was possible to perform only by the image analysis.
The analyses were taken from the melted top layer.
Laser-glazed coatings were highly densified within the melted surface layer, with the
exception of some closed pores and vertical macrocracks formed in the laser glazing
process. Total porosity of the laser-glazed layers was lowered by 73-86 % exclusive of the
vertical macrocracks [Publication III]. Most of the pores were spherical and located at the
lower region of the melted layer. The rest of the porosity took the form of vertical
microcracks. These techniques could not produce a reliable determination of open porosity
of the laser-glazed coatings.
3.1.2 Characteristic microstructure of the phosphate sealed coatings
The microstructure of the phosphate sealed coatings was investigated in more detail by
SEM and TEM [Publications I-III]. This investigation clarified the process of sealant
penetration into the coating as well as yielded a better understanding of the bonding and
strengthening mechanism related to the phosphate sealing.
The penetration of the aluminium phosphate sealant into the 8Y coating near the surface is
illustrated by SEM micrographs in Figs. 14 a and b. EDS analyses showed that aluminium
rich areas were mostly located in the coating cracks, which is clearly shown in the
aluminium EDS map, Fig. 14 c, over the region shown in Fig. 14 b [157].
a)
b)
c)
Fig. 14. Penetration of the aluminium phosphate sealant into the 8Y coating: a) SEM BSE
micrograph of the coating, b) higher magnification of the sealant filled crack and c)
aluminium EDS map from the same region [157].
37
In our earlier studies [158,159] we found that, depending on the coating material (plasma
sprayed oxides), the strengthening in phosphate sealing resulted from two different
mechanisms; chemical bonding or/and adhesive binding. In the first case a chemical
reaction bonds the coating material and the sealant. In the latter case the strengthening
depends on the formation of condensed phosphates in the structural defects of the
coating. Phosphate sealant, penetrated into the interlamellar crack in 8Y AP coating, is
presented in TEM micrograph in Fig. 15. The high magnification TEM images showed no
visible reaction layer in the coating/sealant interface, so it can be assumed that here the
bonding is based mainly on the latter mechanism [Publication III]. SAED ring patterns
verified the amorphous structure of the sealant, if compared to the SAED pattern taken
from the coating lamella, see Fig. 15. In our earlier study [160] we showed that, in the case
of orthophosphoric acid sealed 22MgO-ZrO2 coating, the sealant also takes amorphous
form. But in that case the bonding and strengthening is mostly based on the chemical
reaction between the coating material and the sealant. The strengthening mechanism of
the 25C AP coating was not possible to study by TEM, since the number of specimens to
be characterised by TEM was limited.
Fig. 15. TEM micrograph of the 8Y AP coating and SAED patterns for t-ZrO2 and
amorphous sealant phases [Publication III].
3.1.3 Characteristic microstructure of the laser-glazed coatings
Top-view SEM microstructures of the laser-glazed TTBCs are presented in Fig. 16. The 8Y
LASER coatings had a rather smooth and even surface. The colour of the yttria stabilised
zirconia coating changed from light grey to transparent yellowish/white due to the laser
glazing procedure. Optically it could be described as transparent and glassy [Publications
I-III]. The surface topography of the 25C LASER coating was also rather smooth, but some
craters, 200-500 m in diameter, were opened to the surface. The craters were likely
generated when entrapped gas, from the coating porosity, escaped from the melt pool
during the glazing process. The light greenish/yellow colour of the coating changed to
black in laser glazing [Publications II,III]. In contrast, the surface of the 22M LASER
coating was quite coarse with lots of craters, but the white colour of the 22M did not
change in the laser treatment [Publications I-III]. The coating colour shade variations in
laser glazing most likely resulted from the changes in stoichiometry during rapid heating
and cooling processes. The reversibility of the colour change was demonstrated in a
simple heat treatment in air at 1250oC for 5 hours, after which the colour of 8Y LASER and
25C LASER coatings nearly matched the original colour of the feedstock powder.
38
8Y LASER
25C LASER
22M LASER
39
8Y LASER
25C LASER
22M LASER
40
8Y LASER
25C LASER
22M LASER
The quantitative XRD analysis of the phosphate sealed coatings was made after grinding
off the 50 m thick coating surface layer. The sealant phases could not be identified by
XRD, mainly because of their amorphous microstructure [Publication III]. And even if the
sealant had reacted with the coating material, in the case of a chemical bonding
mechanism, the total volume of the reaction layer at the interfaces of the sealant and
coating material was too small to detect in XRD analyses. However, when the XRD
analyses for the phosphate sealed coatings were performed with removing only the extra
sealant from the surface, the results were different. These analyses identified clear
zirconium phosphate (ZrP2O7) peaks in the 22M OPA [Publication II] and 25C AP coatings,
but not in the 8Y AP coating. The same reactions could be expected to take place also in
cracks and pores, but in smaller volumes. These results suggest that the bonding and
strengthening mechanism in the 22M OPA and 25C AP could be based on chemical
bonding; they also indicate an adhesive binding mechanism in the 8Y AP coating.
Table 7. Quantitative XRD phase analysis results for the feedstock powders and assprayed and modified coatings before and after the heat treatment at 1250oC for 5 h in air.
Powder/
Coating
204NS
8Y
8Y LASER
m-ZrO2
[vol%, 3%]
t-ZrO2
[vol%, 3%]
c-ZrO2
[vol%, 3%]
Other phases
[vol%, 3%]
Original
Heat
treated
Original
Heat
treated
Original
Heat
treated
Original
20
3
-
n.a
3
-
80
92
100
n.a
92
100
5
-
n.a
5
-
8Y AP
50
92
48
205NS
25C
25C LASER
29
-
n.a
-
36
72
96
n.a
89
99
31
25
-
n.a
9
-
25C AP
60
54
39
38
ZRO-103
22M
n.a
65
19
n.a
8
65
55
n.a
1
22M LASER
54
16
22M OPA
85
19
55
CeO2 = 4
CeO2 = 3
CeO2 = 4
CeO2 = 1, traces
of ZrP2O7*
MgO = 35
MgO = 26
Mg2Zr5O12 = 66,
MgO = 18
MgO = 26, traces
of ZrP2O7*
Heat treated
Traces of
AlPO4*
CeO2 = 2
CeO2 = 1
CeO2 = 3
MgO = 26
Mg2Zr5O12 = 29,
MgO = 17
MgO = 12
* ZrP2O7 and AlPO4 were found only at the coating surface. These phases were not detected if the surface layer of 50
m was grinded off before the XRD analysis
In laser glazing the zirconia phase structure in the 8Y and 25C coatings was totally
stabilised to t-ZrO2 [Publication I,III]. In both coatings the structure was identified as pure
t-ZrO2 with no cubic phase present. The pure t-ZrO2 structure indicated the complete
melting of the surface layer in the laser glazing process and very rapid solidification and
cooling of the crystals at the surface. The discrete lines and spots in the image plate
spectrum of the 8Y LASER coating indicated large grain size at the coating surface. In 25C
LASER coating there still was some free CeO2 as in the spray powder and in the assprayed coating. The 22M LASER coating consisted mostly of rhombohedral Mg2Zr5O12
phase after the laser glazing, but c-ZrO2 and c-MgO were also present [Publications I-III].
EDS analyses showed approximately 8 wt % of the MgO within the dendrite structure, and
respectively 17 wt % between the dendrites. This indicated that the dendrites were
composed of Mg2Zr5O12 crystals and the rest of the structure of c-ZrO2 and c-MgO.
3.2
The mechanical and wear properties of modified TTBCs are presented in this chapter.
Coating residual stresses are also considered. Mechanical properties were characterised
42
elongation of the 8Y LASER coating were caused by the vertical macrocracks that opened
at the coating surface under the tensile bending load.
0.060
8Y AP
0.050
Coating
8Y AP HT (1250C, 5h)
Load [kN]
0.040
8Y HT (1250C, 5h)
0.030
0.020
8Y
0.010
0.000
0.00
0.50
1.00
RB [MPa]
EB [GPa]
8Y
39.7 2.7
9.9 0.6
8Y HT
91.3 3.9
40.9 4.1
8Y AP
130.7 5.7
83.2 3.0
8Y AP HT
134.4 11.4
71.1 5.1
8Y LASER
10.8 2.1
13*
8Y LASER HT
19.1 1.4
58*
* No exact values could be determined
1.50
8Y LASER
2.00
Extension [mm]
Fig. 20. 4PB test results in a form of load-displacement curves (median curves of six
tested specimens) and calculated bending modulus (EB) and modulus of rupture in
bending (RB).
Heat treatment clearly increased RB and EB of the 8Y and 8Y LASER coatings [Publication
V]. In the aluminium phosphate sealed coating the effect was very low. The increase of RB
and EB in the 8Y and 8Y LASER coatings results from the crack healing and the improved
bonding of the lamellae, induced by sintering during the high temperature exposure. This
behaviour has also been reported in other studies for yttria stabilised zirconia coatings
[161-163]. However, the bending modulus of the heat-treated 8Y LASER coating still
remained very low, even lower than the value of the as-sprayed 8Y coating. Even if the
determination of EB for 8Y LASER coating was inaccurate, its stress-strain behaviour
makes it interesting when considering the coatings resistance against thermal cycling
loads. In 8Y AP coating the effect of the heat treatment was almost negligible. In this case,
the lamellae had probably already bonded well due to the sealant, and heat treatment
caused no further improvement. In other words, the sealant in the coating structure had
effectively hindered the sintering.
3.2.3 Residual stresses
Residual stresses of TBCs have been widely studied by different techniques [95,107-110].
A lot of effort has been put into understanding the mechanisms through which the stresses
are generated to the coatings in coating manufacturing [107,108] and how they develop at
high temperatures [112-114,164]. Here the residual stress analyses were carrier out using
XRD based sin2 -method. As the penetration depth of CuK is approximately 5-10 m
into the zirconia, the stress profiles were generated by repeating the measurements after
slightly grinding the coating layer by layer. Results are presented in Fig. 21. Residual
stress analyses were performed only for 8Y based coating, since it has good peak to
background ratio in XRD diffraction pattern at high 2 -angles if compared to 22M and 25C
coatings.
44
200
Stress [MPa]
100
0
-100
-200
8Y REF
-300
8Y LASER
-400
8Y AP
-500
0
200
400
600
800
45
tensile shrinkage stresses, 3) not all tensile stresses were relaxed, since some residual
stresses remained in each dense segment between the vertical cracks.
3.2.4 Wear properties
Wear tests were performed only for 8Y and 22M based coatings. Dry erosion tests showed
the typical erosion wear behaviour of brittle material considering the particle impact angle.
In all coatings the wear volume was highest at an angle of 90o and lowest at 30o. Wear
rates as a function of impact angle of erosive are presented in Fig. 22.
Weight loss [mg]
60
50
90
40
60
30
30
20
120
100
60
30
80
60
40
10
20
0
a)
140
8Y
8Y AP
22M
8Y LASER
22M OPA
22M LASER
Fig. 22. Dry erosion test results of the 8Y and 22M based modified TTBCs.
500 m
500 m
Erosion resistance of both the phosphate sealed coatings was increased significantly
[Publication V]. At low impact angles the weight losses were only one fifth that of the
reference coatings. In all impact angles the average improvement in erosion resistance
was 65-70%. In phosphate sealed coatings the better integrity of the structure probably
hindered the erosion-induced crack growth at splat boundaries and prevented the microchipping of lamellae. Fig. 23 shows the cross-sectional optical micrographs of the wear
traces of the as-sprayed and phosphate sealed coatings at different erosive impact angles.
Fig. 23. Cross-sectional optical micrographs of the a) 8Y and 8Y AP coatings and b) 22M
and 22M OPA coatings after the erosion tests with different impact angles.
Laser glazing improved the erosion wear resistance of the yttria stabilised coating on the
average by 35%, but the effect was negative on magnesia stabilised coating [Publication
V]. The erosion wear traces showed that the wear mechanism of the laser-glazed coatings
differed from that of the as-sprayed coatings. The wear surface of the as-sprayed coating
indicated that erosion particle impact on the coating surface caused material removal in
the scale of one lamella (~1-5 m). By contrast, in the laser-glazed coating the final
fracture, leading to material removal, seemed to occur at the scale of the melted layer. Fig.
46
24 shows the top view of the erosion wear traces of the 8Y LASER and 22M LASER
coatings.
a)
b)
Fig. 24. SEM micrographs of the erosion wear traces in a) 8Y LASER and b) 22M LASER
coatings.
At the edges of the removed material the columns are clearly visible, and severe wear
seems to concentrate in the cavities. In Fig. 24 the centre line of the wear trace is marked
with the white line. The thicker and more uniform layer of the melted zone in the 8Y
LASER coating explains its better erosion resistance, compared to 22M LASER coating.
Due to the branched and coalesced cracks below the melted layer of 22M LASER coating,
it eroded away very rapidly at beginning of the test. This explains the high total wear
volume of the coating.
The results of the abrasion tests were well in line with the erosion test results and could be
interpreted analogously. Even though the test procedure was different, the order of the
wear resistance of the coatings was almost the same. Both phosphate sealed coatings
again showed excellent wear resistance, 70-80 % lower weight losses than in reference
coatings. However, in this test the 8Y LASER coating was even more abrasion resistant
than the phosphate sealed coating. In this case, only a minor part of the melted top-layer
was worn away.
3.3
Thermophysical properties
temperature range of 1000-1300oC, see Figs. 25 a and d. D. Zhu et al. [116] have
presented the same type of results of sintering shrinkage for various plasma sprayed
TBCs. The second type of irreversible change was seen as strong shrinkage in the case of
magnesia stabilised coating when the MgO was precipitated from the zirconia matrix, see
Figs. 25 e and f [Publication IV]. Reversible volume changes (phase transformations) were
clearly seen in totally destabilised coatings, in heat treated 22M and 22M OPA coatings,
for instance.
Coefficient of thermal expansion (CTE) of the 8Y coating was approximately 9.910-6 K-1 at
a temperature range of 50-1000oC, see Fig. 25 a. The major shrinkage occurred very
quickly at 1000-1300oC, and there was only a slight difference in total shrinkage if the
dwell time at the maximum temperature was extended from 5 minutes to 5 hours. In the
heat treated coating the total shrinkage [dl/lo] of the measurement cycle was very limited
and it was only ~10% of the shrinkage of the as-sprayed coating (0.02 % vs. 0.27 %). No
indication of phase changes was observed in 8Y coatings. Instead thermal expansion of
the 8Y AP coating was not as linear as that of the as-sprayed coating, see Fig. 25 b. When
the coating was heated up to 980oC, no indication of shrinkage or phase changes were
observed. But if heated up to 1300oC some irreversible behaviour could be observed. For
some reason the t-ZrO2 phase structure was partially destabilised at high temperature,
which could be seen as a phase transformation in the return curve. If some chemical
reaction took place between the sealant and stabilising oxide (Y2O3), that reaction could
not be shown by XRD. The phase changes of zirconia were even more clearly seen in the
case of the heat treated 8Y AP coating. This phase change was also detected in XRD
studies, presented in Table 7 on page 42. The phase change regions (t-ZrO2 to m-ZrO2,
m-ZrO2 to t-ZrO2) are marked on the curves as textured areas in Fig. 25.
In 25C and 25C AP coatings the thermal expansion was almost identical, see Figs 25 c
and d. The sintering shrinkage could be seen at the temperature range of 1000-1300oC
with no impression of phase changes. The heat treated coatings showed no phase
changes, but some minor shrinkage was detectable. CTE for 25C and 25C AP coatings in
temperature range of 50-1000oC was approximately 10.810-6 K-1.
Magnesia stabilised coatings 22M and 22M OPA started to destabilise at temperatures of
900-950oC, see Figs. 25 e and f. Strong shrinkage was seen in both coatings (shrinkage
was about 2,57 %, so it was 10 times higher than the maximum shrinkage of the 8Y or
25C coatings). The zirconia phase structure of both coatings was almost totally changed to
m-ZrO2 in the heat treatment, see Table 7 on page 42. After the heat treatment the further
shrinkage of the 22M HT and 22M OPA HT coatings was very limited. The phase changes
(m-ZrO2 t-ZrO2 and t-ZrO2 m-ZrO2) of zirconia could be clearly detected, see the
textured areas in the Figs. 21 e and f. For some reason these phase changes occurred at
higher temperature in the 22M OPA HT than in the 22M HT coating. CTE of the assprayed 22M coating at a temperature range of 50-700oC was approximately 8.810-6 K-1.
48
49
Fig. 25. Thermal expansion of a) 8Y coatings, b) 8YAP coatings, c) 25C coatings, d) 25CAP coatings, e) 22M coatings and f)
22MOPA coatings. Marked areas in the figures refer to the structural changes of zirconia.
Fig. 26. SEM micrographs of heat-treated 8Y coating illustrating the sintering of the
structure. String of the fine pores and closed cracks at splat boundaries are marked with
black arrows.
Open porosity of the 8Y and 25C coatings, determined by mercury porosimetry or
Archimedes method, remained fairly constant before and after the heat treatment
[Publication IV]. Sintering shrinkage seemingly did not affect the open porosity, and
mercury porosimetry results showed no clear evidence of the reduction of the very fine
pores or microcracks. However, the heat treatment increased the open porosity of the 8Y
AP and 25C AP coatings to some degree [Publication IV]. This increase was probably due
to the shrinkage of the sealant at high temperatures, caused by the crystallisation of the
amorphous structure. Porosity of the heat treated 22M based coatings was difficult to
measure and interpret, because the high amount of m-ZrO2 made the structure very brittle.
Moreover, the MgO precipitates, which were seen as dark spots in optical micrographs,
complicated the image analysis.
Quantitative XRD phase analysis results strongly supported the thermal expansion data
[Publication IV], see the grey columns in Table 7 on page 42. The t-ZrO2 phase structure
of the 8Y and 8Y LASER coatings did not change in the heat treatment although in 8Y AP
coating the tetragonal structure was partially destabilised to m-ZrO2 (50 vol%). After the
heat treatment a small amount of AlPO4 was identified at the coating surface, and it could
be assumed also that the amorphous sealant in the coating cracks was crystallised.
However, the amount of the sealant penetrated into the coating structure was probably too
low to detect by XRD. Unfortunately the TEM studies, in which this inference could be
verified, were impossible to carry out. Heat treatment had only a slight effect on the phase
structure of the 25C and 25C LASER coatings. 25C AP coating was more stable in heattreatment than 8Y AP and only 5 vol% of m-ZrO2 was detected. The ZrP2O7 phase that
50
was identified at the coating surface in 25C AP coating was not present after the heat
treatment. Phase structure of the magnesia stabilised coatings were strongly affected by
the heat treatment. The c- ZrO2 and t'-ZrO2 structures were almost totally destabilised and
the major part of the coatings was transformed to m-ZrO2. Precipitates of MgO were
possible to observe in SEM studies. XRD peaks of the ZrP2O7 phase at 22M OPA coating
surface were not identified after the heat treatment, as was the case with 25C AP coating.
After the heat treatment in 22M LASER coating the amount of m-ZrO2 was lower
compared to 22M and 22M OPA coatings. It seemed that the Mg2Zr5O12 phase in 22M
LASER coating was slightly more stable than t'-ZrO2/c-ZrO2 at high temperatures.
3.3.3 Thermal conductivity
8.00
7.00
6.00
5.00
4.00
3.00
8Y-2
8Y LASER-2
8Y AP-2
3.00
2.50
2.00
1.50
1.00
1.00
0.50
0.00
0
200
400
a)
600
800
1000
1200
1400
Temperature [C o ]
2.50
2.00
1.50
1.00
0.50
0.00
0
25C-1
25C-2
25C LASER-1
25C LASER-2
25C AP-1
25C AP-2
200
400
600
800
Temperature [C o]
200
400
b)
3.00
c)
8Y-1
8Y LASER-1
8Y AP-1
3.50
2.00
0.00
4.00
600
800
1000
1200
1400
Temperature [C ]
6.00
5.00
22M-1
22M-2
22M LASER-1
22M LASER-2
22M OPA-1
22M OPA-2
4.00
3.00
2.00
1.00
1000
1200
1400
0.00
d)
200
400
600
800
1000
1200
1400
Temperature [C o ]
The data found from the literature [41,166] for 8Y2O3-ZrO2 was compared to the results of
8Y coating, see Fig. 27 a. The modification processes had clear effects on thermal
conductivity of TTBCs. Phosphate sealing significantly increased thermal conductivity due
to sealant filling the cracks and pores [Publication IV]. In the case on 8Y AP and 22M OPA
coatings, the sealant induced or accelerated destabilisation of zirconia structure which
further increased thermal conductivity. The effect of laser glazing on thermal conductivity
varied little between each coating material [Publication IV]. In the 8Y LASER coating, in
which the macrocracks were straight and vertical, the dense laser-glazed top layer slightly
increased thermal conductivity. But in the case of 25C LASER and 22M LASER coatings
the effect was the opposite. This difference can be explained by the fact that the
macrocracks in those coatings were not perfectly vertical and some cracks were even
laterally branched.
In all coatings k(T) was obviously higher in the second measurement cycle [Publication IV].
In 8Y and 25C based coatings this was mainly due to the better integrity of the lamellar
structure induced by the sintering based phenomena, which were discussed in the
previous chapter. D. Zhu et al. [162] demonstrated by isothermal k(T) measurements at
990, 1100 and 1320oC that the major increase in k(T) takes place during the first 5-10
hours. Repeating the measurements three times shows that the major increase of the k(T)
occurs really quickly, mainly during the first measurement cycle. In 22M based coatings
the increase of k(T) was caused by another mechanism, mainly by the precipitation of
MgO, leading to destabilisation of c-ZrO2/t-ZrO2 zirconia and formation of m-ZrO2.
3.4
TTBCs were exposed to mixtures of vanadium pentoxide (V2O5) and sodium sulphate
(Na2SO4) at various temperatures (600-850C). Test duration varied between 48 and 1000
hours. Since not all the coatings were available for all test series, the test parameters are
presented here case by case. Following chapters consider the hot corrosion resistance of
modified TTBCs based on SEM+EDS investigations, XRD studies and residual stress
analyses.
3.4.1 Melt deposit penetration into the coatings
In Figs. 28-30 the cross sectional SEM micrographs of the hot corrosion exposed (750850oC, 18Na2SO4 - 82V2O5 (mol-%), 400 h, in air) coatings are presented illustrating the
corrosion reaction layers and melt deposit penetration into the coating.
52
8Y
8Y AP
8Y LASER
Fig. 28. Cross-sectional SEM micrographs of the hot corrosion tested 8Y based coatings.
(18Na2SO4 - 82V2O5 (mol-%) deposit in air at 750oC for 400h).
53
25C
25C AP
25C LASER
Fig. 29. Cross-sectional SEM micrographs of the hot corrosion tested 25C based coatings.
(18Na2SO4 - 82V2O5 (mol-%) deposit in air at 750oC for 400h).
54
22M
22M OPA
22M LASER
Fig. 30. Cross-sectional SEM micrographs of the hot corrosion tested 22M based
coatings. (18Na2SO4 - 82V2O5 (mol-%) deposit in air at 850oC for 400h).
SEM studies and EDS analysis found that coating surface areas were depleted from the
stabilising oxides. The molten deposit penetration into the coating structure can be seen
as dark vanadium rich phases at splat boundaries. Observation of the coating surface
region reveals that the reaction layer in phosphate sealed coatings is slightly thinner than
in as-sprayed coatings. Also the amount of penetrated corrosion deposit below the
reaction layer was found to be lower in phosphate sealed coatings. The thickness of the
reaction layer at the surface of the laser-glazed coatings was even lower than that in
phosphate sealed coatings. The reaction layer thickness was closely related to the
density/porosity of the surface. In other words, the specific surface area for corrosion
reaction gets smaller as the surface porosity decreases. Although the laser-glazed layer
itself was dense and more corrosion resistant than the other structures, the molten deposit
was able to penetrate into the coating structure via the vertical macrocracks. There the
deposit was spread out as in as-sprayed coatings and even reached the areas below the
dense laser-glazed layer.
The molten deposit penetration into the 8Y and 8Y AP coatings is demonstrated in Fig. 31
by optical micrographs and vanadium EDS analyses, taken from the different depths from
the coating surface.
55
a)
1.8
8Y
At % of Vanadium
1.6
8Y AP1
1.4
1.2
8Y AP2
8Y AP3
0.8
0.6
0.4
0.2
0
b)
200
400
600
800
1000
1200
Fig. 31. Melt deposit penetration into the 8Y AP coatings: a) cross-sectional and top-view
optical images and b) EDS vanadium area analyses at different depth from the coating
surface. (65Na2SO4 - 35V2O5 (mol-%) deposit at 600oC for 48 hours in air).
Each analysed area was approximately 150 m high and 1500 m wide. The analysed
coatings were exposed at 600oC to 65Na2SO4 - 35V2O5 (mol-%) for 48 hours in air. Crosssectional optical micrographs, Fig. 31 a, showed that in 8Y coating the melt deposit has
spread throughout the coating. By contrast, in phosphate sealed coatings the penetration
was smaller but spread uniformly or evenly in a vertical direction. The EDS analyses
showed lowered V concentrations at the surface of the 8Y AP coatings, but deeper in the
coating values deviated.
The melt deposit penetrated into the laser-glazed coatings via the vertical macrocracks as
stated earlier. The SEM micrographs and related elemental maps of the exposed laserglazed coatings are presented in Figs. 32-34. Here the coatings were exposed to
18Na2SO4 - 82V2O5 (mol-%) deposit at 750oC for 100 h in air. It was clearly seen that the
melt deposit penetrated down to the vertical crack tip and even further in a horizontal
direction.
56
Fig. 32. Cross-sectional SEM micrographs and elemental map of vanadium of the hot
corrosion tested 8Y LASER coating demonstrating the melt penetration into the
macrocracks. (18Na2SO4 - 82V2O5 (mol-%) deposit at 750oC for 100 h in air).
57
Fig. 33. Cross-sectional SEM micrographs and elemental map of cerium of the hot
corrosion tested 25C LASER coating demonstrating the melt penetration into the
macrocracks. (18Na2SO4 - 82V2O5 (mol-%) deposit at 750oC for 100 h in air).
EDS elemental mapping showed that in 25C LASER and 22M LASER coatings depletion
of stabilising oxide took place at crack edges. The same has probably happened in 8Y
LASER coating, but it was impossible to detect due to overlapping peaks of Zr and Y in
EDS spectrum.
58
Fig. 34. Cross-sectional SEM micrograph and elemental maps of vanadium and
magnesium of the hot corrosion tested 22M LASER coating demonstrating the melt
penetration into the macrocracks. (18Na2SO4 - 82V2O5 (mol-%) deposit at 750oC for 100 h
in air).
3.4.2 Zirconia destabilization and corrosion reactions
XRD diagrams of the 8Y and 22M based coatings, exposed at 650oC to 65Na2SO4 35V2O5 (mol-%) for 200 hours in air, are presented in Figs. 35 and 36 [Publication I].
These showed strong destabilisation of the t-ZrO2 and c-ZrO2 zirconia phases. Some
reaction products and remains of the Na2SO4 - V2O5 deposit can also be identified.
However, the most common phenomenon with all coatings was the increase of the
proportion of the m-ZrO2.
In the case of all 8Y based coatings the stabilising oxide Y2O3 had reacted with vanadium
and formed YVO4, see Fig. 35. This reaction, see equation 4, has been reported in several
other studies [64,65,167] and it is known to be a problem of yttria stabilised zirconia in
vanadium containing environments at temperature ranges of 600-900oC.
Y2O3 (in t-ZrO2) + V2O5 2YVO4 with formation of m-ZrO2
(4)
59
After the exposure the major phase in the 8Y and 8Y AP coating was m-ZrO2, whereas in
8Y LASER the t-ZrO2 phase had the highest XRD intensity peaks. Two possible reasons
account for this: 1) either the transformed t-ZrO2 structure in laser-glazed coating has
been more resistant to the reaction with the deposit or 2) the surface area for the corrosion
reaction has been much lower, since the analysis was made from the dense glazed layer.
The latter explanation is well supported by the micrographs presented in Figs. 28 and 32.
a)
b)
Fig. 35. XRD diagrams of the 8Y based coatings: a) hot corrosion exposed coatings and
b) original coatings. Coatings were exposed at 650oC to 65Na2SO4 - 35V2O5 (mol-%) for
200 hours in air. Phase markings: m = m-ZrO2, t = t-ZrO2, y = YVO4, n = Na2SO4.
Correspondingly, destabilisation of the c-ZrO2 took place in all 22M based coatings, see
Fig. 36. In the laser-glazed coating the rhombohedral Mg2Zr5O12 phase appeared to be
slightly more stable in the test environment compared to c-ZrO2. Some unidentified
diffraction peaks were present in the XRD diagram in the case of 22M OPA and 22M
LASER coatings. These peaks did not exactly fit any reaction products expected, but some
correlation was found with MgV2O6. Other presumable reaction products according to
phase diagram of MgO and V2O5 [168] were Mg2V2O7, MgV6O17, Mg3V2O8, but these
phases were not found from exposed 22M based coatings. MgO has been reported to
form MgSO4 in the presence of Na2SO4(l) and SO3(g) [169]. However, magnesium
sulphate was not found in XRD studies. In all the original 22M based coatings there were
some free c-MgO phase, which completely disappeared during the exposure according to
XRD data. Lack of the free c-MgO and destabilisation of the c-ZrO2 in exposed coatings
mean that MgO has reacted to some extent with the Na2SO4 - V2O5 deposit.
Ceria stabilised coatings were not analysed by XRD, but SEM studies and EDS analysis of
the 25C LASER coating found that the ceria content at the interfaces of the coating and
corrosion deposit was significantly lowered. This was demonstrated in Fig. 33 on page 57.
60
R. L. Jones [64] has reported that ceria stabilised zirconia can react with V2O5 leading to
destabilised zirconia (m-ZrO2) and CeVO4.
a)
b)
Fig. 36. XRD diagrams of the 22M based coatings: a) hot corrosion exposed coatings and
b) original coatings. Coatings were exposed at 650oC to 65Na2SO4 - 35V2O5 (mol-%) for
200 hours in air. Phase markings: m = m-ZrO2, c = c-ZrO2, z = Mg2Zr5O12 and x =
unidentified peak.
3.4.3 Stress generation in the hot corrosion exposed coatings
Residual stresses of the hot corrosion exposed specimens (65Na2SO4 - 35V2O5 (mol-%),
48 h, 600oC, air) were analysed in order to better understand the failure mechanism of the
8Y AP coatings. Stresses were measured from the exposed and non-exposed area (see
the specimens in Fig. 31 a on page 55).
The results are presented in Fig. 37. It can be noted that the stresses after sealing are
lower here compared to results presented in Fig. 21 on page 45. After the test, a 50-100
m thick layer had to be ground off the coating surface of the samples used here to
remove the extra corrosion products. This was also carried out for the original sealed
coatings, which means the top region was ground off, where the highest compression
appeared. It can be clearly seen that compressive stresses have been induced in all
coatings in the exposed areas and that the stresses were extremely high in phosphate
sealed coatings. A slight compressive stress component was also induced in the nonexposed areas in all coatings. Compressive stress generation was probably related to the
volume expansion that is linked to the phase change t-ZrO2 to m-ZrO2. Due to the different
test parameters the increase of m-ZrO2 was not as dramatic here as it was presented in
Fig. 35. The melt deposit penetration into the coating may also have affected the stresses.
If the deposit remained in the coating microcracks and open pores, it would have induced
compressive stresses after the test when solidifying in cooling down to TR.
61
com pression
tension
8Y AP3 exposed area
8Y AP3 non-exposed area
8Y AP3 after sealing
8Y AP2 exposed area
8Y AP2 non-exposed area
8Y AP2 after sealing
8Y AP1 exposed area
8Y AP1 non-exposed area
8Y AP1 after sealing
8Y exposed area
8Y non-exposed area
8Y as-sprayed
-700
-600
-500
-400
-300
-200
-100
100
Stress [MPa]
Fig. 37. Effect of the hot corrosion exposure on stress states in the 8Y and 8Y AP
coatings. Coatings were exposed to 65Na2SO4 - 35V2O5 (mol-%) for 48 hours at 600oC in
air.
3.4.4 Conclusions of the hot corrosion experiments
Tests showed that the melt deposit (Na2SO4 - V2O5) exposure was a very severe test for
zirconia based TTBCs. In testing times over 200 h, all the coatings were peeled off due to
the phase transformations induced by zirconia destabilisation. Melt deposit was found to
be more aggressive when containing higher fractions of V2O5. Some general findings can
be derived from all the tested specimens: 1) Visual inspection of the exposed specimens
showed that the 22M and 25C coatings were in better condition than the 8Y coatings when
the temperatures were 600-750oC. At 850oC the opposite was true. 2) Phosphate sealed
coatings lowered the melt deposit penetration into the coating. However, original
compressive stress state of the phosphate sealed coating seemed to increase significantly
in hot corrosion tests due to zirconia phase changes. In some cases this was seen as a
violent cracking, and coating fragments were bounced off from the substrate when the
specimens were cooled down after the test. 3) Laser glazing did not effectively prevent
melt deposit penetration into the coating. The deposit was able to enter into the vertical
macrocracks and from there spread even under the dense laser-glazed layer. Finally there
was no noteworthy difference in general hot corrosion resistance of reference and laserglazed coatings Nevertheless, the dense melted zone itself at the coating surface was
rather corrosion resistant.
3.5
Thermal cycling properties of the 8Y and 25C based modified TTBCs were studied in three
test series in which the maximum coating temperature was fixed at 1000, 1150 and
1300oC [Publication VI]. In addition to as-sprayed, phosphate sealed and laser-glazed 8Y
coatings, some segmentation cracked 8Y based coatings were studied, since they
represent the existing state-of-the-art strain tolerant TTBC structure [95,120]. It should be
noted that here the laser glazing was performed for normal as-sprayed coating, but
aluminium phosphate sealing for segmented coating. In this context the bond coat
deposition technique is marked in brackets after the coating abbreviation, since the
segmentation cracked coatings were sprayed on the bond coat prepared either by HVOF
or APS. Ceria stabilised coatings were tested in as-sprayed (25C), aluminium phosphates
sealed (25C AP) and laser-glazed (25CL) state only with APS bond coat.
62
In the test type used here the coating failure resulted from the stresses generated by the
high temperature gradients in heating and cooling steps. On that basis it was easier to
compare the pure thermal cycling resistance and strain tolerance properties of each
modified TTBC structure. In service the delamination of TBCs may occur due to several
different mechanisms, but when the coating is exposed to high temperatures for long
periods the stresses at the interface of the bond coat and zirconia have significant effect
on coating lifetime, as reported in the literature [170,171]. This type of stress is
emphasised when the layer thickness of thermally grown oxide (TGO) at the TBC/bond
coat interface reaches a certain level. In this experiment the total dwell time at maximum
temperature was too low for allowing the growth of TGO (this was also verified by
SEM/EDS analysis). The other factor affecting the stresses in long term, zirconia
destabilisation, was also negligible here (verified by XRD).
3.5.1 Test series 1
In test series 1 the Tmax was fixed up to 1000oC. One of the 25C AP coatings was
damaged after 277 cycles, but the other coatings showed no visible delamination or
cracking after 500 cycles. The failure mode of 25C AP coating was the same as in test
series 2 and 3, described later. Optical micrographs of the undamaged coatings after 500
cycles are presented in Fig. 38 on pages 63-64.
8Y (HVOF bc)
63
25C
25C AP
25C LASER
Fig. 38. Optical micrographs of the modified coatings after the 500 cycles in test series 1
(Tmax = 100050oC).
Some microstructural changes took place in the coatings during the test series 1. In
reference to 8Y (HVOF bc) coating, a horizontal crack slightly above the bond coat
indicated that the coating delamination process had already started. The segmentation
cracked coatings 8Y SEG (APS bc) and 8Y SEG (HVOF bc) and the phosphate sealed 8Y
SEG + AP (APS bc) coating did not show any changes. In laser-glazed coating a
horizontal crack appeared within the melted top layer, but it probably had formed during
laser glazing or in the specimen cutting process as in the thermal cycling test. Rather
pronounced microstructural changes were seen in the 25C based coatings. Some vertical
macrocracks, comparable to segmentation cracks in 8Y SEG coatings, had formed in the
25C coating. In the 25C AP coating these cracks, with plenty of branched horizontal
elements, appeared even more clearly. This type of crack structure also formed in the 25C
LASER coating in laser glazing process, but the length of macrocracks and number and
length of branching cracks increased in the test. A horizontal crack, propagated from the
edge of the specimen, was also seen near the bond coat of 25C LASER coating.
64
test 2b
F
1000
25C LASER
809
F
9
9
25C AP
522
25C
864
1000
1000
1000
915
35
128
8Y (HVOF bc)
200
400
600
800
1000
1200
Cycles to failure
Fig. 39. Number of thermal cycles leading to coating failure and propagation of the coating
delamination in combined test series 2a and 2b (Tmax = 115050oC).
Reference 8Y (HVOF bc) coating was peeled off only after a couple of cycles.
Delamination occurred at the interface of zirconia and bond coat, and the coatings
detached in one piece. Indication of that type of failure was already obtained in optical
micrographs after the test series 1, Fig. 38. The 25C coating resisted more than 500
cycles, but the failure mode was the same as in the reference 8Y (HVOF bc) coating. The
crack structure of the 25C coating, if developed during the experiment as in test series 1,
might explain its advantage to 8Y (HVOF bc) coating.
Segmentation cracked coatings showed excellent performance in the test series 2. In the
one 8Y SEG (APS bc) coating the outer rim (~20 % of the total area) of the coating peeled
off near the bond coat after 915 cycles. Aluminium phosphate sealing decreased
significantly thermal cycling resistance of the 8Y SEG (APS bc) and 25C coatings. In all
phosphate sealed coatings the failure occurred in the same mode and nearly half of the
coating area was delaminated from the edge of the specimen in the form of a sickle. In
these cases the coating was fractured within the ceramic layer in such way that the
thickness of the delaminated layer was higher in the specimen edges. So it is possible that
the delamination of the phosphate sealed coating occurred at the depth corresponding to
sealant penetration. 8Y LASER (APS bc) and 25C LASER coatings performed very well. In
one 8Y LASER (APS bc) coating the area outside from the laser-glazed track peeled off
gradually, see Fig. 39. One 25C LASER coating detached from the substrate in one piece
at the interface on the bond coat after 809 cycles.
65
209
9
12
25C AP
25C
F
F
15
F
216
230
171
F
173
173
30
11
F
2
3
8Y (HVOF bc)
50
100
150
200
250
Cycles to failure
Fig. 40. Number of thermal cycles leading to coating failure in test series 3 (Tmax =
130050oC).
Here the test was continued until all the specimens were damaged. Reference 8Y (HVOF
bc) and 25C coatings as well as all phosphate sealed coatings failed at early stage of the
test. When the test was finished (230 cycles) more that 80 % of coating area remained in
three specimens; 8Y LASER (APS bc), one 8Y SEG (HVOF bc) and one 8Y SEG (APS
bc). Optical micrographs of these coatings, taken from the undamaged area, are
presented in Fig. 41.
66
8Y LASER
Fig. 41. Optical micrographs of the modified TTBCs after 230 thermal cycles (Tmax =
130050oC).
3.5.4 Discussion of the test results and failure modes
The delamination of the coating started in all cases from the specimen edge region; see
the propagation of the delamination of each type of coating in Figs. 39 and 40. The failure
mode of each coating type was almost equivalent in all test series. The edge regions were
most sensitive to coating failure, because there the heating and cooling occurred most
rapidly.
Thermal cycling resistance of the reference coatings 8Y (HVOF bc) and 25C was poor.
Coatings peeled off in one piece at the bond coat interface indicating good coating
integrity but low strain tolerance. The opposite behaviour was found in segmentation
cracked and laser-glazed coatings.
Both the 8Y SEG (HVOF bc) and 8Y SEG (APS bc) coatings were found to have excellent
thermal cycling resistance. The slightly better thermal cycling resistance of the 8Y SEG
(HVOF bc) coating could be explained by the original segmentation crack structure of
these two coatings. The average segmentation crack length in 8Y SEG (HVOF bc) coating
67
was at least twice as high as in 8Y SEG (APS bc) coating (300-800 m vs. 100-300 m).
The cracks were also more open in the original 8Y SEG (HVOF bc) coating. The effect of
HVOF and APS bond coat was difficult to interpret here, since there were also differences
in zirconia coatings that were prepared separately. However, our earlier study [93] showed
that there is no difference in thermal cycling resistance whether the zirconia was deposited
onto VPS, HVOF of APS sprayed bond coat.
The laser-glazed coatings also showed excellent thermal cycling resistance. In the 8Y
LASER (APS bc) coating only the segments, outside the laser-glazed track, were peeled
off and the laser-glazed area remained unaffected. 25C LASER coatings also managed
well although their failure mode differed from the 8Y LASER (APS bc) coating. The
delamination also started from the specimen edge, but at some point the whole coating
was peeled off near the bond coat.
It can be assumed that the segmentation cracked and laser-glazed coating structures
tolerate better tensile stresses as compressive stresses. Under tensile load the coating
cracks can be opened, but in compression that is not possible. However, the stress
situation is not so simple in practice, because there should always be a temperature
gradient in TBC. This was also the case in these tests. Due to temperature gradient and
differences in CTE of the coating and substrate, a stress gradient is induced into the
coating. The stress gradient creates bending stresses that still increase the effect of crack
initiation and growth at weak points of the coating such as the edge regions of the coating
and the TBC/bond coat interface. The advantage of a segmentation cracked and laserglazed coating is that tensile and bending stresses do not directly accumulate into a
coating in macro scale as they do in the case of normally structured APS TBC.
Thermal cycling resistance was dramatically deteriorated due to the aluminium phosphate
sealing. The failure mode of the aluminium phosphate sealed coatings, 8Y SEG + AP
(APS bc) and 25C AP, differed from the other coatings. These coatings were not
delaminated regularly at the bond coat interface, but were fractured within the ceramic
layer in the form of a sickle. These fractured coating pieces were thicker at the specimen
edge side referring to irregular local sealant penetration and coating densification. As
presented earlier in chapters 3.1.2 and 3.1.3, aluminium phosphate sealing significantly
increased the elastic modulus and compressive stresses of the 8Y coating. These findings
can be associated with the lowered thermal cycling resistance of aluminium phosphate
sealed coatings and support the failure mode observed in these tests.
68
4. CONCLUDING REMARKS
The use of combined cycle power and heat generation is increasing due to continuously
expanding energy consumption. These increases will be inevitable especially in those
countries that are shutting down their nuclear power plants. High power diesel engines will
account for a share of the increase too, perhaps in smaller units such as reserve power
plants or the power stations of industrial plants. Energy producers are searching for ways
to reduce their costs in highly competitive markets. The aviation industry is also
experiencing a tremendous need for lower costs. The development of highly efficient
power plants and aviation engines as well as advanced maintenance/overhaul services
comes in response to these and other background pressures. In this development thermal
barrier coatings play a small but essential role. TBCs indirectly affect engine efficiency,
fuel economy and maintenance costs. It is not wrong to say that there is steady demand
for better TBCs.
Thermal barrier coatings have been used and studied for decades so the published data
available on the topic is voluminous. During the last twenty years, much work has focused
on studying the properties of APS and EB-PVD 8YSZ TBCs (present industrial standards),
but recently an increasing number of publications investigate interestingly novel topics
such as new TBC materials, low thermal conductance zirconia based TBCs and new
coating techniques in producing strain tolerant TBCs. More of this type of work will be
needed to find solutions for the next generation of heat engines.
This work was undertaken for the purpose of improving the properties of thick thermal
barrier coatings by modifying their microstructure by laser glazing and phosphate based
sealing treatments. The research shows how the microstructures can be affected by each
modification procedure, and their basic mechanical and thermal properties were
determined and compared to normal structured TTBCs. The work also included high
temperature testing of the coatings in hot corrosion and thermal cycling experiments to
better understand and estimate their behaviour in real service conditions. The main results
of the work are summarized in here:
Microstructures
Microstructures of zirconia based TTBC were modified by phosphate sealing and laser
glazing. By phosphate sealing it was possible to reduce the open porosity of the coatings
in 300-400 m thick surface layer. Sealant filled the cracks and open pores and
strengthened the coating structure by adhesive binding or chemical bonding mechanisms.
In laser glazing the 50-150 m thick layer was melted resulting in a dense surface with
special vertical macrocrack structure.
Mechanical and elastic properties
Phosphate sealing and laser glazing significantly affected the coating mechanical and
elastic properties. In both cases the microhardness was greatly increased. The
strengthening effect of the phosphate sealing was also seen in a sharp increase in
modulus of rupture in bending (RB) and bending modulus (EB). It was also found that
strong compressive stresses were generated in coatings in phosphate sealing. In laserglazed coatings both RB and EB were reduced due to the macrocracks. Phosphate sealing
significantly improved the erosion resistance of TTBCs. Laser glazing positively affected
the erosion resistance of the 8Y2O3-ZrO2 coatings, but negatively affected the 22MgOZrO2 coatings. This difference was based on dissimilar macrocrack structure in 8Y LASER
and 22M LASER coatings.
69
Thermophysical properties
High temperature phase stability of the 8Y2O3-ZrO2 coating was deteriorated by phosphate
sealing. This effect was found to be much slighter in the case of 25CeO2-2.5Y2O3-ZrO2
coating. Phosphate sealing also increased thermal conductivity of all studied coatings.
Laser glazing had negligible effect on coating high temperature phase stability. The effect
on coating thermal conductivity was also small and mainly influenced by the macrocrack
structure and its orientation.
Hot corrosion resistance
Results showed that phosphate sealing or laser glazing can not be use to improve the hot
corrosion resistance of TTBCs when they are exposed to molten NaSO4-V2O5 based
corrosion deposit at 600-850oC. Phosphate sealing slightly decreased the molten deposit
penetration into the coatings, but problems arose due to very high compressive stresses
induced by the corrosion exposure. The compressive stresses those were already present
after the sealing mainly grew out of the destabilisation of zirconia (t-ZrO2 m-ZrO2). In
laser-glazed coatings the molten deposit found its way through the macrocracks, so the
coating structure was not protected by the top layer. For that reason there was no notable
difference in general hot corrosion resistance of reference and laser-glazed coatings.
However the dense laser-glazed zone proved to be rather corrosion resistant.
Thermal cycling resistance
Phosphate sealing lowered the thermal cycling resistance of TTBCs. Obviously the
increased elastic modulus (better cohesion of splats) and compressive internal stresses
decrease the coating strain tolerance. Correspondingly, laser glazing significantly
improved the thermal cycling resistance of the TTBCs. Laser-glazed coatings were
superior to the reference coatings and comparable to the segmentation cracked coatings.
The favourable strain tolerant structure of the laser-glazed coatings was caused by its low
elastic modulus due to vertical macrocracks.
Final conclusions
Based on these results, it can be concluded that the phosphate sealed coatings are not
suitable for use in gas turbine hot section components where the TBC surface
temperatures typically approach 1000oC or even higher. They can be neither
recommended to use in such combustion chamber components of diesel engine where the
maximum temperatures affect coating phase stability or where the hot corrosion conditions
are severe. It can be stated that the excellent erosion resistance of the phosphate sealed
TTBC coating would be possible to exploit only in low temperature diesel processes. In
that case the coating behaviour and durability should be extensively tested in service,
because also in such conditions there still might be a risk of coating failure due to lowered
strain tolerance.
The laser-glazed TTBCs, especially laser-glazed 8Y2O3-ZrO2 coatings, might work well in
static gas turbine components. Even if their hot corrosion resistance against molten
NaSO4-V2O5 at 600-850oC is not better than normal TTBC, their excellent strain tolerance
and erosion resistance make them very promising compared to current TTBCs. They could
be first exploited in parts where the laser glazing would not present too much difficulty
(complex geometries). Rotation symmetrical inside diameter surfaces or plane surfaces
should be possible to process with existing techniques (robot controlled Nd-YAG lasers
with special optics). These types of parts include, for example, combustion cans or heat
shields of the combustion chamber. Even if the data presented here supports the
70
71
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
72
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
73
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
76
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.
100.
101.
102.
103.
77
104.
105.
106.
107.
108.
109.
110.
111.
112.
113.
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
130.
131.
132.
133.
134.
79
135.
136.
137.
138.
139.
140.
141.
142.
143.
144.
145.
146.
147.
148.
149.
150.
H.L. Tsai, P.C. Tsai, Performance of Laser-glazed Plasma-sprayed (ZrO212wt.%Y2O3)/(Ni-22wt.%Cr-10wt.%Al-1wt.%Y) Thermal Barrier Coatings in Cyclic
Oxidation Tests, Surface and Coatings Technology 71 (1995) 53-59.
H.L. Tsai, P.C. Tsai, Microstructures and Properties of Laser-Glazed PlasmaSprayed ZrO2-YO1.5/Ni-22Cr-10Al-1Y Thermal Barrier Coatings, Journal of Thermal
Spray Technology 4[6] (1995) 689-696.
A. Petitbon, R. Queriaud, Strengthened Thermal Barrier Coatings for Use in Diesel
and Gas Turbine Engines, in T. S. Sudarshan and M. Jeandin (Eds.), Surface
Modification Technologies VIII, The Institute of Materials, London, UK, 1995, p. 772777.
K. A. Khor, S. Tana, Pulsed Laser Processing of Plasma Sprayed Thermal Barrier
Coatings, Journal of Materials Processing Technology 66 (1997) 4-8.
Z. Zhou, N. Eguch, A. Ohmori, Microstructure Control of Zirconia Thermal Barrier
Coatings by Using YAG Laser Combined Plasma Spraying Technique, in C. C.
Berndt (Ed.), Thermal Spray: A United Forum for Scientific and Technological
Advances, ASM International, 1997, p. 315-321.
Z. Zhou, N. Eguchi, H. Shirasawa, A. Ohmori, Microstructure and Characterization
of Zirconia-Yttria Coatings Formed in Laser and Hybrid Spray Process, Journal of
Thermal Spray Technology 8[3] (1999) 405-413.
H. Kuribayashi, K. Suganuma, Y. Miyamoto, M. Koizumi, Effect of HIP Treatment on
Plasma-Sprayed Ceramic Coating onto Stainless Steel, American Ceramic Society
Bulletin 65[9] (1986) 1306-1310.
K. A. Khor, N. L. Loh, Hot Isostatic Pressing of Plasma Sprayed Thermal Barrier
Coating Systems, Materials and Manufacturing Processes 10[6] (1995) 1241-1256.
K.A. Khor, Y. W. Gu, Hot Isostatic Pressing of Plasma Sprayed Yttria-Stabilized
Zirconia, Materials Letters 34 (1998) 263268.
K. Moriya, H. Tomino, Y. Kandaka, T. Hara, A. Ohmori, Sealing of Plasma-Sprayed
Ceramic Coatings by Sol-Gel Process, in C. C. Berndt and S. Sampath (Eds.),
Thermal Spray Industrial Applications, ASM International, USA, 1994, p. 549-553.
K. Moriya, W. Zhao, A. Ohmori, Improvement of Plasma-Sprayed Ceramic Coatings
Treated by Sol-Gel Process, A. Ohmori (Ed.), Thermal Spraying: Current Status
and Future Trends, High Temperature Society of Japan, Japan, 1995, p. 10171021.
G. John, T. Troczynski, Surface Modification of Thermal Sprayed Coatings, in C. C.
Berndt (Ed.), Thermal Spray: Practical Solutions for Engineering Problems, ASM
International, USA, 1996, p. 483 488.
I. Berezin, T. Troczynski, Surface Modification of Zirconia Thermal Barrier Coatings,
Journal of Materials Science Letters 15 (1996) 214-218.
J. Kathikeyan, C. C. Berndt, A. Ristorucci, H. Herman, Ceramic Impregnation of
Plasma Sprayed Thermal Barrier Coatings, in C. C. Berndt (Ed.), Thermal Spray:
Practical Solutions for Engineering Problems, ASM International, USA, 1996, p.
477-482.
T. Troczynski, L. Pawlowski, N. Third, L. Covelli, I. Smurov, Physico-Chemical
Treatment of Zirconia Coatings for Thermal Barriers, in C. Coddet (Ed.), Thermal
Spray: Meeting the Challenges of the 21st Century, ASM International, USA, 1998,
p. 1337-1342.
T. Troczynski, Q. Yang, G. John, Post-Deposition Treatment of Zirconia Thermal
Barrier Coatings Using Sol-Gel Alumina, Journal of Thermal Spray Technology 8[2]
1999 229-234.
80
151.
152.
153.
154.
155.
156.
157.
158.
159.
160.
161.
162.
163.
164.
165.
166.
167.
168.
169.
170.
171.
82