Materials and Engineering A359 (2003) 100
/111

www.elsevier.com/locate/msea

Processing effects on porosity-property correlations in plasma

sprayed yttria-stabilized zirconia coatings
Anand Kulkarni *, A. Vaidya, A. Goland, S. Sampath, H. Herman
Department of Materials Science and Engineering, Center for Thermal Spray Research, State University of New York, Old Engineering, Room 314,
Stony Brook, NY 11794-2275, USA

Received 11 November 2002; received in revised form 4 April 2003

Abstract

For plasma sprayed thermal barrier coatings (TBCs), control of thermal conductivity is critical since low thermal conductivity
depends not only on the intrinsic property of the yttria-stabilized zirconia (YSZ) TBC, but also on the morphology of pores and
cracks introduced during spray process. They are closely linked to process methodology as well as to chemistry, structure and
morphology of the ceramic feed materials. This paper addresses the influence of feedstock characteristics on particle state in the
plasma and the resultant coating properties. In addition, substrate temperature, angle-of-impact and thermal cycling effects on
porosity (quantity and morphology) and its resultant influence on thermal conductivity and elastic modulus of plasma sprayed YSZ
TBCs. The results show increased porosity with particle size, due to an increase in the degree of particle fragmentation and unmelted
particles, leading to lower thermal conductivity and modulus. Furthermore, higher substrate temperatures and low particle velocity
lead to lower porosity and improved inter-splat contact and, thus, enhanced coating properties. Sintering during thermal cycling
reduces porosity and increases thermal conductivity and modulus.
# 2003 Elsevier B.V. All rights reserved.

Keywords: Thermal barrier coating; Plasma; Yttria-stabilized zirconia; Porosity; Thermal conductivity; Elastic modulus

1. Introduction plasma spray and electron beam-physical vapor deposi-

tion (EB-PVD). In the EB-PVD process, vapors are
There is an extensive effort to incorporate prime- produced by heating the source material with an
reliant ceramic thermal barrier coatings (TBCs) into electron beam, and the evaporated atoms condense on
advanced gas turbine and diesel engine components. the substrate. Crystal nuclei form on favored sites,
This is because of the increased thermodynamic effi- growing laterally and in thickness to form individual
ciency due to the higher operating temperatures and columns, which provide in-plane compliance [5]. For the
reduced cooling air requirements allowed by these case of plasma spraying, feedstock powder is melted and
coatings as well as prolonged substrate lifetimes due to accelerated to high velocities, impinging upon the
lower metal surface temperatures. TBCs based on 6
/8 substrate, and rapidly solidifying to form a ‘‘splat’’ (a
mol% yttria-stabilized zirconia (YSZ) have been widely flattened particle). The deposit develops by successive
used, owing to their low thermal conductivity, chemical impingement and inter-bonding among the splats. The
and thermal stability, relatively high coefficient of deposit microstructure is strongly dependent on proces-
thermal expansion (relative to the metallic substrate) sing conditions, spray parameters and feedstock materi-
and high resistance to spallation under thermal fatigue als. The splats are separated by interlamellar pores
(toughening) [1
/4]. The development and acceptance of resulting from rapid solidification of the lamellae, very
TBCs are also closely linked to processing technology: fine voids formed by incomplete inter-splat contact or
around unmelted particles, and cracks due to thermal
* Corresponding author. Tel.: /1-631-632-4511; fax: /1-631-632-
stresses and tensile quenching relaxation stress [6
/8].
8440. The pores and cracks interfere with the direct flow of
E-mail address: [email protected] (A. Kulkarni). heat (thermal barrier) resulting in lowered thermal
0921-5093/03/$ - see front matter # 2003 Elsevier B.V. All rights reserved.
doi:10.1016/S0921-5093(03)00342-3
 A. Kulkarni et al. / Materials and Engineering A359 (2003) 100
/111 101

conductivity. The cracks also increase the overall 2.1. Processing parameters
compliance of the coating and, hence, enhance the
thermal shock resistance [9
/11]. Due to thermal cycling Initially, a study was carried on the effects of feed-
conditions encountered by TBCs in service, the influence stock particle size on the properties of YSZ coatings.
of thermal cycling on the thermal protection and Coatings were deposited using a Sulzer Metco 3 MB
spallation resistance has been studied in great detail plasma gun at a 100 mm standoff distance. The spray
[12
/15]. In this paper, we examine processing-micro- parameters are listed in Table 3. Splats were collected on
structure-property correlations in plasma sprayed YSZ polished stainless steel substrates. Freestanding deposits
coatings. Studies involving the influence of feedstock were also obtained for characterization. Some of the
particle size, particle conditions (particle temperature, coatings were subjected to ten cycles of heating at
velocity and flux), deposition conditions (substrate 1150 8C for 30 min, followed by cooling in air for 15
temperature and angle-of-impact) are related to micro- min to observe resultant deposit properties on thermal
structure development and properties. The experimental cycling.
results are presented as follows: particle
/plasma inter- The second part of the study was an investigation of
actions to show the inflight behavior of feedstock in the the effects of particle velocity and substrate temperature
plasma flow; correlations of particle state to splat on coating properties. The particle velocity was changed
morphologies on impact with the substrate; splat
/splat by varying the plasma conditions during spraying (e.g.
interactions leading to the microstructure development; using nozzles of different internal diameter, 6 and 8 mm,
correlation of microstructure to porosity and properties for high and low velocities, respectively). The particle
of the coatings. Results are also given of microstructure conditions used are shown in Table 4. In-flight particle
and property alterations that occur upon thermal diagnostics were performed using the Tecnar DPVTM
cycling. Such studies will allow tailoring of coating system to monitor particle temperature and velocity, the
microstructure to meet future design needs of prime results of which are shown in Table 5. Previous work
reliant TBCs. [16,17] has shown the substrate temperature to have a
dramatic effect on splat morphology, thus influencing
the adhesion and bonding between splats and, hence, the
physical and mechanical properties of the coatings. With
2. Experimental this in mind, the splats and coatings were also produced
at two substrate temperatures, designated ‘‘low’’ and
In order to examine particle behavior in the plasma ‘‘high’’. By controlling the preheating and air-cooling
flow, in-flight diagnostic studies were carried out on during spraying, it was possible to vary substrate
feedstock powders with different manufacturing meth- temperatures from 100 8C (Cold) to 300
/350 8C (Hot).
ods and shaped morphologies. Three powders were Lastly, a study was carried out on the influence of
chosen, made by different manufacturing methods and spray angle on the microstructure. The splats were
two different particle size distributions. Table 1 lists the collected and coatings were sprayed at two different
characteristics of powders that were studied. spray angles, normal and 458 to the substrate. The spray
In order to explore the effects of particle size process parameters are given in Table 6. Coatings were
distribution, three different particle sizes (classified as sprayed using a fused-and-crushed powder at very high
‘Fine’, ‘Medium’ and ‘Coarse’ were sieved from the as- power to develop dense microstructure [18]. The influ-
received plasma densified powder and compared to the ence of particle flux on the microstructure was also
as-received/ensemble powder. Two of the powders were investigated by changing the traverse speed of the
chosen for the diagnostic study */‘Fine’ and ‘Medium’ plasma gun during spraying; 70 and 30 m s 1 for
size distributions: Table 2. standard and slow (high particle flux) traverse, respec-

Table 1
Feedstock characteristics

Powder type Processing Manufacturer Mean particle size-d50 S.D. Surface area (m2
(mm) (mm) g 1)a

Fused and crushed (F&C) Cast material, subsequently MSM ZY7 45 22 0.08
crushed
Agglomerated and sintered Spray dried and sintered Sulzer-Metco AE 45 28 0.4
(A&S) 7216
Plasma densified (HOSP) Plasma processed/densified Sulzer-Metco 61 33 0.48
AE7593
a
Characterized using BET surface area technique at NIST, Gaithersburg, MD.
102 A. Kulkarni et al. / Materials and Engineering A359 (2003) 100
/111

Table 2
Particle size distributions for sieved plasma densified powder

Powder size Manufacturer Sulzer-Metco Mean particle Size-d50 (mm) S.D. (mm) Surface area (m2 g 1)a

Fine AE7592 35 18 0.91

Medium AE7591 54 21 0.44
Coarse AE7590 103 24 0.32
a
Characterized using BET surface area technique at NIST, Gaithersburg, MD.

tively. Deposition with high particle flux produce

Table 3 thermal stresses to generate vertically cracked structure
Processing parameters for particle size effects in the coatings [19,20], the preferential orientation of
Plasma gun Sulzer-Metco 3 MB these cracks can be observed on a macroscale with
Current 650 A respect to spray angle.
Voltage 66
/68 V
Primary gas 40 SLM Ar
Secondary gas 8 SLM H2 2.2. Deposit characteristics
Carrier gas 3000 SCCM Ar
Feed rate 20
/40 g min1 Splats were observed under SEM, since splat
Standoff distance 100 mm morphologies and interactions between them control
deposit integrity and properties. To quantify the surface
profile and dimensionality of these splats, they were
observed under a ZygoTM surface interferometer. Free-
Table 4 standing YSZ deposits were evaluated for porosity
Processing parameters for particle velocity and substrate temperature content and thermal conductivity. Surface-connected
effects porosity in as-sprayed and thermal-cycled specimens
was measured by mercury intrusion porosimetry (MIP)
Plasma gun PT F4 torch
Current 600 A using a Quantachrome Autoscan 33 porosimeter. Ther-
Voltage 66
/68 V mal conductivity measurements were carried out on
Primary gas 20 SLM Ar disk-shaped specimens, coated with carbon, using a
Secondary gas 8 SLM H2 Holometrix laser flash thermal properties instrument.
Secondary gas 30 SLM He
Depth-sensitive indentation studies were carried out
Carrier gas 3000 SCCM Ar
Feed rate 20
/40 g min 1 with a Nanotest 600 instrument with a 1/16-in. WC-
Standoff distance 100 mm Co spherical indenter with a maximum load of 10 N.
Phase and microstructural characterization were also
carried out. Microstructural changes upon thermal
cycling/sintering were monitored through thermal ex-
Table 5 pansion measurements on a Netzsch 402CTM Dilat-
In-flight diagnostics ometer at a heating rate of 10 K min1 up to 1400 8C.
Low velocity High velocity

Temperature (8C) 33469/200 32279/200 3. Results and discussions

Velocity (m s 1) 1809/30 2359/30

3.1. Particle
/plasma interactions

The results from diagnostic studies, summarized in

Table 6 Fig. 1, show that the particle size and shape of the
Processing parameters for angle-of-impact studies feedstock powder play key roles in determining the
Plasma gun PT F4 torch velocity and temperature of the particles in the plasma
Current 650 A flow. The Fig. 1 indicates that the coarser powder
Voltage 66
/68 V (‘‘medium’’) travel slower through the plume as com-
Primary gas 40 SLM Ar pared with the finer distribution. This can be under-
Secondary gas 10 SLM H2
Carrier gas 3000 SCCM Ar
stood from the fact that the momentum transfer from
Feed rate 20
/40 g min 1 the gas to the particle is strongly dependent on the mass
Standoff distance 100 mm of the injected particle. Thus, the coarser (heavier)
particles attain a lower velocity as compared with the
 A. Kulkarni et al. / Materials and Engineering A359 (2003) 100
/111 103

Fig. 1. Diagnostic data for different powders shows differences in particle
/plasma interactions measured at the standoff distance.

finer ones. On the other hand, the average surface factor, which is significant with regards to heat and
temperature recorded for the coarser particles is higher momentum transfer, is the particle shape. The powders
than that recorded for the finer particles. This difference with particles of different shape*/dense spherical (ag-
arises due to the different trajectories of the injected glomerated and sintered) and angular (fused and
particles and the longer residency time. Fig. 2 represents crushed) result in different particle velocities and
a cross-sectional snapshot of the plasma plume showing temperatures.
the particle size distribution within the flow at two
different distances from the nozzle, 60 and 120 mm. The
center of plasma plume corresponds to coordinates (0, 3.2. Droplet
/substrate interactions
0) and the particles are injected from the top-above the
coordinate (0, 10). As can be seen from the two contour It is clear that droplet impact, spreading and splat
maps, the larger particles travel closer to the center of solidification largely determine microstructure develop-
the flow and the finer ones are on the periphery. The ment. The intrinsic properties of the individual splats
coarser particles, which have a larger momentum than and the intricate correlations among them are affected
the finer ones, penetrate the plasma plume to a greater by the conditions of the impacting particles, such as size,
extent. The coarser particles also travel much slower temperature and velocity. Substrate condition (or that
than the fines and hence spend a longer time in the of previously solidified particles) affect wetting, thermal
plume. This results in a trajectory closer to the high contact resistance, roughness/splat profile, chemical
temperature zones of the plasma. Hence, the coarser interactions, etc. The governing interrelations of droplet
particles attain a higher surface temperature for similar impact and solidification history to process parameters
injection parameters. For the particle size distribution and substrate condition are examined here. The results
chosen in this study, the particles were in a range demonstrate the role of particle energy and substrate
suitable for plasma spraying and not so coarse as to temperature in determining the nature of splat spreading
cause the particles to ‘fall’ through the flow. The other and solidification.
104 A. Kulkarni et al. / Materials and Engineering A359 (2003) 100
/111

Fig. 3 shows the microstructure of a single YSZ splat.

Fig. 3A shows a disk-shaped splat, collected on a
polished stainless steel substrate at 100 8C. ‘‘Mud
cracks’’ due to stress relief of YSZ are also seen together
with the core and rim regions. Fig. 3B shows the
droplet
/substrate interaction. Columnar grain struc-
ture, having an average grain diameter of 50 nm, is
observed in regions where the particle droplet first
comes in contact with the substrate. The columns
grow in a direction perpendicular to the substrate, since
heat is extracted through the substrate and the solid/
liquid interface moves vertically. Also seen is the
microcrack inside the splat.

3.3. Splat
/splat interactions/coating buildup

Fig. 4 shows a fractured cross-section of a plasma

sprayed YSZ coating showing layers of splats, along
with other features such as interlamellar pores, cracks
and globular pores. Fig. 4B shows columnar epitaxy
preserved in a splat
/splat interaction along with glob-
ular pores from gas and interlamellar pores due to
improper localized adhesion. The interlamellar pores
result from poor wetting/adhesion between the splats as
they accumulate to form the coating. Although they
provide significant reduction in through-thickness ther-
mal conductivity, these poorly bonded regions can result
in delaminations, causing premature failure during
thermal cycling. The microcracks or ‘‘mud-cracks’’
present within splats result from the release of ‘‘quench-
ing’’ stresses during deposition. These stresses arise due
to constrained shrinkage of a solidifying splat and
generally have a large tensile magnitude for materials
Fig. 2. This figure represents a cross-sectional snapshot of the plasma such as YSZ [8]. In the case of most ceramics, the large
plume showing the particle size distribution within the flow at two
different distances from the gun (A) 60 and (B) 120 mm.
quenching stresses cannot be retained by the splat
during deposition and cooling and result in micro-
cracking (Fig. 3A). These micro-cracks can provide a
small degree of in-plane compliance. The globular pores

Fig. 3. Microstructure development of a single splat, (A) shows single splat morphology collected on a polished substrate and (B) shows columnar
structure along with microcracks (M).
 A. Kulkarni et al. / Materials and Engineering A359 (2003) 100
/111 105

Fig. 4. Cross-section micrograph of a plasma sprayed YSZ coating, (A) shows splats along with other features with interlamellar pores, cracks and
globular pores and (B) shows splat
/splat interface along with globular pores and interlamellar pore (improper adhesion).

result from lack of filling around the undulations of that in thermal conductivity is 9/3%. The porosity
splats and debris. In general, globular pores are increases with increased particle size. This can be
uncontrollable and undesirable since they provide sites attributed to two factors: reduced melting efficiency of
for failure initiation. Another aspect in the TBC coarser particles in the plasma plume compared with
microstructure is the existence of the large vertical fine particles and, secondly, due to the increased
macro-cracks , which form due to the relief of thermal fragmentation of splats resulting in poor adhesion and
stresses associated with the mismatch between thermal porosity. The thermal conductivity obeys the inverse
expansion coefficients of the substrate and the coating. relationship, decreasing with increasing particle size.
When properly controlled, these cracks can provide the The modulus results, plotted versus porosity, are shown
desired in-plane compliance, such as those in EB-PVD along with the microstructures in Fig. 7. Vast differences
deposits. are observed, showing fragmented splats leading to poor
splat
/splat contact and formation of pores. It is evident
3.4. Effect of feedstock characteristics that the coating prepared from fine powder shows well-
adhered splats, while the unmelted and poorly adhered
Fig. 5 shows the different splat morphologies ob- particles can be seen in the case of the coating prepared
served for the three different particle sizes. It is evident from coarse powder. In-plane and out-of-plane modulus
that the morphology of the splats changes from a disk- measurements were carried out to show the anisotropy
like shape to fragmented shape with increasing particle in these coatings. The modulus, both in through-
size. These shapes are highly representative of the vast thickness and in-plane direction, decreases with increas-
number of splats for each size distribution. This ing particle size, obeying the same trend as the thermal
difference in splat morphology manifests itself in coating conductivity. However, the well-adhered splats along
microstructure and properties. with the interlamellar pores and intrasplat cracks
These differences in microstructure show themselves generate greater anisotropy in the case of coating
as wide variations in the porosity levels in the coatings. made from fine powder particles. This anisotropy
Fig. 6 shows particle size effects on porosity and thermal reduces with increasing particle size owing to the
conductivity. The scatter in porosity values is 9/5% and microstructure developed from the unmelted particles.

Fig. 5. Particle size effects on splat morphologies observed, collected on polished stainless steel substrate at 100 8C.
106 A. Kulkarni et al. / Materials and Engineering A359 (2003) 100
/111

Fig. 6. Influence of feedstock particle size on porosity and thermal conductivity of YSZ coatings.

3.5. Particle velocity and substrate temperature effects case. Certain regions of unmelts are also observed in the
high velocity condition, owing to shorter flight times in
Fig. 8 shows the splat morphologies observed under the plasma. Also, the thickness of the splats is different;
an interferometer for the low and high particle velo- thicker for the low velocity condition than for the high
cities: 180 and 235 m s 1, respectively. The splats velocity conditions, thus fewer interfaces are observed in
appear to be fragmented for both the conditions, the the former case. MIP porosity shows an increase in
extent of fragmentation appearing greater for the high surface-connected porosity with increasing velocity,
velocity case. The surface topography indicates a arising from the interlamellar and globular pores due
disturbed/rough surface profile for the high velocity to improper adhesion and unmelted particles in the
condition vis-à-vis the low velocity condition. As microstructure (Fig. 9). As a consequence of the higher
anticipated, the dimensionality of the splats show the velocity (at almost similar particle temperatures (Table
splats to be thinner for the higher velocity case. The 5) and a reduced residence time of the particle in the
splat morphologies observed for the different substrate plasma), it can be concluded that the microstructure of
temperatures were similar to those observed in the the coating is influenced to a greater degree by particle
previous study [16], confirming its role in enhancing velocities than it is owing to a change in substrate
contact area and wetting between the splat and the temperature.
substrate. The thermo-mechanical properties observed for the
Fig. 9 shows the significantly different microstructural two different velocity conditions at two substrate
features observed for these particle conditions. The temperatures are shown in Fig. 10. As indicated earlier,
coating sprayed under low velocity condition shows the scatter in thermal conductivity is 9/3% and that for
well-adhered splats, compared with the high velocity elastic modulus values is 9/1%. A lower porosity is

Fig. 7. Microstructural variations explaining anisotropy in coatings with respect to elastic modulus.
 A. Kulkarni et al. / Materials and Engineering A359 (2003) 100
/111 107

Fig. 8. Surface topography of splats observed under interferometry.

observed for the high substrate temperature, leading to alumina), Fig. 12, showing cracks tilted by 70
/808 to the
increased thermal conductivity and elastic modulii, substrate plane. In order to determine the effect of spray
consistent with prior studies [16]. Also, the increase in angle and flux on physical properties, a study was
porosity combined with the large number of splat carried out on thermal conductivity and modulus as a
interfaces (as a result of thinner splats for the higher function of two spray angles.
velocity case) decrease thermal conductivity and elastic The observed thermal conductivity and elastic moduli
modulus. values are shown in Fig. 13. The effect of particle flux is
controlled by traverse speed of the plasma gun during
3.6. Effect of particle flux and angle-of-impact spraying; 70 and 30 m s 1 for standard (low particle
flux) condition and slow (high particle flux) condition,
In this section are discussed the splat morphologies respectively. The variation in surface connected porosity
and properties with reference to the particle flux effects was not significant for each case owing to large crack
and the angle-of-impact of impinging particle. Fig. 11 network contributing to porosity. However, the proper-
shows the splat morphologies and dimensions observed ties will depend on the intrinsic structure of the coating.
for two (45 and 908) impinging angles. The thickness of The values of thermal conductivity and elastic modulus
the splats is nominally 1.29/0.3 mm for both cases. It is are higher for coatings sprayed by slow traverse speed
observed that the splats are disk-shaped for normal because of higher particle flux impinging on the
incidence and elongated for the 458 impact angle. This is substrate leading to lower porosity coatings. The pre-
expected to influence microstructure development, ferential crack orientation for the angle-of-impact
which is shown in Fig. 12. Fused-and-crushed powder studies show lower thermal conductivity for coatings
was selected, given its ability to produce dense coatings sprayed at 458 impact angle as opposed to coatings
as observed in previous studies [18]. The microstructures sprayed at 908 impact angle.
show segmented crack networks due to the biaxial stress An investigation of microstructural changes observed
developed during deposition at high power and particle upon heating and thermal cycling would improve our
flux [20]. Earlier work on the effect of spray angle on current understanding of the sintering-related failure
gray alumina coatings showed the tendency for cracks to mechanisms in these coatings. The change in micro-
orient preferentially with the spray direction [19]. This structure reflects changes in specimen dimensions, which
surprising effect is seen for YSZ coatings (tougher than are given in the thermal expansion curve in Fig. 14.
108 A. Kulkarni et al. / Materials and Engineering A359 (2003) 100
/111

Fig. 9. This figure shows the fractured cross-sections observed for different particle conditions, the table shows the variation in surface-connected

Fig. 10. This figure shows the experimental thermal conductivity and elastic modulus of the YSZ coatings for the varying velocity and temperature
conditions.
 A. Kulkarni et al. / Materials and Engineering A359 (2003) 100
/111 109

Fig. 11. The figure shows single splat morphologies for different angle-of-impacts.

Fig. 12. The figure shows effect of angle-of-impact on coating microstructures (cracks seen preferentially oriented).

Curves (1) and (2) are for measurements in the through- cases as reflected in the coefficient of thermal expansion
thickness and the in-plane direction, respectively. It is values. It can be inferred that the dimensional change is
seen that the dimensional changes are different for both larger for the through-thickness direction than the in-

Fig. 13. The figure shows the effect of particle flux and angle-of-impact on the coating properties.
110 A. Kulkarni et al. / Materials and Engineering A359 (2003) 100
/111

Fig. 14. The figure shows thermal expansion curves observed for YSZ coatings heated to 1400 8C.

plane direction, because of the larger dimensions of which are shown in Fig. 15. This change in thermal
interlamellar pores compared with the intrasplat cracks. conductivity can be attributed to sintering upon heating
Another interesting result is the behavior of the coating [22]. The change is greater for the lower porosity
in the temperature range of 450
/800 8C, which likely is coatings (fine particle size) as compared with high
due to the monoclinic-to-tetragonal transformation [21]. porosity coatings (large particle size) because of differ-
Further shrinkage above 1100 8C is due to sintering. ences of the mean opening dimensions of pores [23].
The change in microstructure reflects itself as changes
in porosity and thermal conductivity, the results of
4. Summary and conclusions

There is a significant influence of processing condi-

tions on the properties of plasma sprayed YSZ coatings.
Feedstock particle size, particle condition in the plasma
(particle temperature, velocity and particle flux) and
deposition conditions (substrate temperature and angle-
of-impact) have been identified as critical parameters
that influence the microstructure and properties. Results
show that the particle state in the plasma is strongly
influenced by feedstock size and shape. This in turn
dictates microstructure development to a large extent.
The particle size and substrate temperature studies show
that from among the different microstructural features,
the major factors influencing coating properties are
porosity and interlamellar contact. While wide variation
Fig. 15. The figure shows porosity thermal conductivity relationships
in microstructures and resultant properties are observed,
observed in YSZ coatings. depending on processing conditions, the vertical macro-
 A. Kulkarni et al. / Materials and Engineering A359 (2003) 100
/111 111

cracks seem to provide an ideal combination of lower [5] O. Unal, T.E. Mitchell, A. Heuer, J. Am. Ceram. Soc. 77 (4)
(1994) 984.
thermal conductivity with high strain tolerance for
[6] H. Herman, Sci. Am. 259 (3) (1988) 112.
TBCs in service. [7] R. McPherson, Thin Solid Films 83 (1981) 297.
Sintering during thermal cycling results in a decrease [8] J. Matejicek, S. Sampath, P.C. Brand, H.J. Prask, Acta Mater. 47
in porosity leading to an increase in thermal conductiv- (2) (1999) 607.
ity and a loss of the strain tolerant behavior. The level of [9] R.H.J. Hannink, P.M. Helly, B.C. Muddle, J. Am. Ceram. Soc. 83
(3) (2000) 461.
sintering depends on the initial deposit density realized
[10] M.V. Swain, L.F. Johnson, R. Syed, D.P.H. Hasselman, J. Mater.
from various powder morphologies and particle sizes. Sci. Lett. 5 (1986) 799.
Quantitative analysis on the interlamellar and globu- [11] R. McPherson, Thin Solid Films 112 (1984) 89.
lar pores and intrasplat cracks are being carried out [12] J. Gutleber, S. Usmani, S. Sampath, in: C.C. Berndt (Ed.),
using multiple small-angle neutron scattering for a Thermal Spray: A United Forum for Scientific and Technological
Advances, ASM International, Metals Park, OH, 1997, pp. 285
/
better insight into the porosity-property relationships.
289.
This analysis is also being used to develop a predictive [13] G. McDonald, R.C. Hendricks, Thin Solid Films 73 (1980) 491.
capability of the microstructural property correlations [14] B.C. Wu, E. Chang, S.F. Chang, C.H. Chao, Thin Solid Films 172
[18]. (1989) 185.
[15] C.C. Berndt, H. Herman, Thin Solid Films 108 (1983) 427.
[16] S. Sampath, X. Jiang, Mater. Sci. Eng. A304
/306 (2001) 144.
[17] S. Sampath, X. Jiang, J. Matejicek, A.C. Leger, A. Vardelle, Mat.
Acknowledgements Sci. Eng. A272 (1999) 181.
[18] A. Kulkarni, Z. Wang, T. Nakamura, S. Sampath, A. Goland, H.
Herman, A. Allen, J. Ilvasky, G. Long, J. Frahm, R. Strinbrech,
This research work was sponsored by the MRSEC
Comprehensive microstructural characterization and predictive
program of National Science Foundation under award property modeling of plasma-sprayed zirconia coatings, Acta
DMR-0080021. Mater., 51 (2003) 2457.
[19] J. Ilavsky, A.J. Allen, G.G. Long, S. Krueger, C.C. Berndt, H.
Herman, J. Am. Ceram. Soc. 80 (3) (1997) 733.
[20] P. Bengtsson, T. Ericsson, J. Wigren, J. Therm. Spray Technol. 7
References (2) (1998) 340.
[21] Q. Liu, S. An, W. Qui, Solid State Ionics 121 (1999) 61.
[1] R.A. Miller, Surf. Coat. Technol. 30 (1987) 1. [22] J. Ilavsky, G.G. Long, A.J. Allen, C.C. Berndt, Mater. Sci. Eng.
[2] W.J. Brindley, R.A. Miller, Adv. Mater. Proc. 8 (1989) 29. A272 (1999) 215.
[3] W. Mannsmann, H.W. Grunling, J. Phys. IV 3 (1993) 903. [23] A. Kulkarni, A. Goland, S. Sampath, H. Herman, A. Allen, Small
[4] R.L. Jones, in: K.H. Stern (Ed.), Metallurgical and Protective angle neutron scattering studies to investigate particle size effects
Coatings, Chapman and Hall, London, 1996, pp. 194
/235. in plasma sprayed zirconia coatings, unpublished work.