Diez 1993
Diez 1993
Diez 1993
9
International
Among commercially available powders for plasma spraying, precipitation-calcined processed (PCP)
powders have not been used due to the presence of agglomerates, which produce coatings with closed
pores. This investigation studied the behavior of PCPaggiomerates during plasma spraying. Two PCFyttria powders were selected and air plasma sprayed. Splat morphology and coating microstructure were
characterized under various conditions of plasma parameters (i.e., plasma jet enthalpy, nozzle design,
spraying distance). A model for agglomerate behavior on impact is proposed to explain microstructural
variations and to enable the development of spray parameters that yield dense coatings.
1. Introduction
In the manufacture of engineered plasma sprayed ceramic coatings, characteristics of the starting powder are critical. As recently reported,Ill powder characteristics have a significant influence on the final properties of the coating. Commercially
available powders are processed by different routes, each yielding specific powder characteristics. The most commonly used
ceramic powders in plasma spraying are
9
body
9
2. Experimental Procedure
Two PCP powders were selected for this study. Each was
characterized for mean diameter, size distribution, and particle
morphology, using scanning electron microscopy (SEM) observation and laser-scattering measurements (MicrotraC M). Size
distribution results obtained with each measurement technique
were compared. The relevance of this comparison was reported
in a previous work, I4] in that a powder composed of weak agglomerates exhibited an apparent mean particle diameter measured by MicrotraO 'M much lower than the one measured by
SEM observation. This experimental evidence resulted from the
breakup of agglomerates caused by the turbulent aqueous media
used for particle dispersion in the Microtrac TM measurement
technique. A significant difference between the two measuremerit techniques was therefore considered as an indication of the
mechanical strength of the agglomerates.
The next step consisted of an investigation of splat and coating microstmcture with each powder. Spraying was performed
at atmospheric pressure in air. The powder carrier gas flow rate
was selected to achieve optimum coating substrate targeting and
a uniform spraying pattern,
The following plasma processing parameters were varied
(Table 1):
Dissolution of an organic compound in yttrium-bearing solution (for example urea in yttrium nitrate)D]
Table 1 Plasma processing parameters used to spray PCP powders in current study
Sample
No.
A1 .........
A2 .........
A3 .........
A4 .........
A5 .........
A6 .........
A7 .........
A8 .........
A9 .........
AI0 ........
AI 1........
A12........
B 1..........
B2 ..........
B3 ..........
B4 ..........
B5 ..........
EPI (a)
nozzle
93
120
93
93
93
120
120
93
120
120
93
120
93
93
93
93
93
Particle
residence time
High
Low
High
High
High
Low
Low
High
Low
Low
High
Low
High
High
High
High
High
Power,
kW
64
63
60
60
64
64
72
72
54
60
75
72
57
60
64
68
68
Plasma enthalpy,
kJ/I
29.9 (medium)
29.9 (medium)
18.2 (low)
24.1 (low)
29.9 (medium)
29.9 (medium)
33.2 (high)
33.2 (high)
21.4 (low)
26 (medium)
34.4 (high)
33.1 (high)
14.3 (low)
18.8 (low)
29.9 (medium)
33.8 (high)
33.8 (high)
Spraying distance,
cm
7.5 (short)
7.5 (short)
8.75 (medium)
8.75 (medium)
8.75 (medium)
8.75 (medium)
8.75 (medium)
8.75 (medium)
10 (long)
10 (long)
10 (long)
10 (long)
8.75 (medium)
8.75 (medium)
8.75 (medium)
7.5 (short)
10 (long)
tion powder ports. This latter nozzle was designed to provide a longer particle residence time in the hotter part of the
plasma jet.
3. Results
3.1 Powder Characterization
Both powders were produced by similar processes and differed only in size distribution. The mean diameter (dm) of powder A was 51 [am, whereas powder B had a mean diameter of 34
particle
splat
R?~)A
J~|
t1~
I~ll
nilrd
/
Fig. 5 Typical closed pore St structure in as-sprayed etched microstructure of PCP yttria powder A (dm = 51 lUm)coatings sprayed with
EPITM 120 nozzle and medium plasma jet enthalpy (condition A6 in Table 1).
J X720
0515
10. O"U"DUM
Maximum coating density was reached when the entire coating microstructure consisted of lamellar, well-melted splats, and
when closed, entrapped porosity was minimum. This occurred
when spraying with the EPI TM 93 nozzle, high plasma jet enthalpy, and short spraying distance (75 mm). In such conditions,
coating density was higher for powder B (din = 34 gm) than for
powder A (din = 51 gm). Visual observations of specimens metallography prepared in an identical fashion are shown in Fig. 8.
4. Discussion
lower plasma enthalpy conditions, resulted in a larger grain (5 to
10 Jam) equiaxed microstructure in which a lamellar orientation
was difficult to distinguish (Fig. 4b). This structure retained
much of the agglomerated character of the initial powder.
Powder A and B agglomerate stability was the dominant factor determining coating microstructure. A model for agglomerate behavior in the plasma jet is proposed in Fig. 9 and 10 and
discussed below. The material travel distance inside the plasma
(a)
(a)
(b)
Fig. 6 Typical aspect of hollow sphere yielding closed pore $2 structure in microstructure of PCP yttria powder B (dm = 34 ~m) coatings
sprayed with EPF M 93 nozzle at high plasma jet enthalpy and long
spraying distance (condition B5 in Table 1). (a) SEM aspect of hollow
sphere with melted and resolidified shell ($2 structure). (b) Entrapped
hollow spheres ($2 structure) in as-sprayed etched microstructure.
jet from its injection to its impact on the substrate was divided
into (1) a melting zone, where the material heated up, melted,
and eventually superheated; (2) a cooling zone, between the jet
tip and the substrate, where the material leaving the melting
zone at a higher temperature than the surroundings started to
Liquid
g.
Resolidified shell
Unmelted
Solid
Liquid
Gas
Case 2 : Thin
resolidified shell
Case 3 : Thicker
resolidified shell
Case 4 : Fully
resolidified shell
Case 1
No resolidification
Conical splat
Case 2 :
Small resolidified shell
thickness
I
L i q u i ~ ~
Liquid
Resolidified
shell
Case 3 :
Large resolidified shell
thickness
Resolidified solid
r
Bounce off
.....
re
Unmelted solid
Gluing liquid
Unmelted solid
I
Small gram size
Lamellar structure
Largergrain size
Closed porosity
Moreequiaxed structure
(Nucleation on pieces of
broken resolidified shell )
Low
deposition
efficiency
agglomerates could stick on the substrate by entrapment by another liquid splat. In most cases, these bounced off, especially at
long spraying distances, for which low deposition efficiencies
were observed.
Agglomerates that were not fully molten in the melting zone
arrived on the substrate with a solid inner core. If their liquid
shell was thick enough, it deformed readily on impact, adhering
the solid core to the substrate (case 1 in Fig. 10). In cross sections, this mechanism corresponded to the closed pore structure
S 1 (defined in Fig. 5) with an outer recrystallized structure and
an inner core composed of unmelted original powder particles
and closed pores between them. This structure was typical of
powder A when spraying with medium plasma jet enthalpy and
medium spraying distance conditions.
For incompletely melted agglomerates, after the melting
zone, that were partially resolidified in the cooling zone, the deformation mode on impact depended on the solid shell thickness
and on the amount of remaining liquid (cases 2 and 3 in Fig. 10).
If they adhered to the substrate, then such agglomerates resulted
in closed pores of type S 1.
5. Conclusions
This investigation confirmed that yttria powders derived
from the precipitation-calcination process were composed of
cubic crystal l ire agglomerates, which were mechanically stable
enough to behave as single particles during the spraying proc-
ess. Some agglomerates were made of sufficient primary particles to form a closed structure with some degree of inner porosity. A model of agglomerate behavior in the plasma jet was
proposed to explain the observed coating microstructure.
Smaller powder B agglomerates (din = 34 ~tm) were easily
melted even when spraying with low plasma enthalpy conditions, but because of their size quickly resolidified. Hence, a
short spraying distance (<7 cm) was required to achieve high
densities (>90%) and avoid partial inflight resolidification,
which caused formation of closed pores. At long spraying distances (>10 cm), agglomerates completely resolidified, and
most of them bounced off, yielding low deposition efficiencies.
Large powder A agglomerates (din = 51 ~tm) were often incompletely melted even when spraying with a nozzle providing
a long particle residence time in the plasma jet and at high
plasma jet enthalpy. These resulted in coating microstructures
with closed pores of S 1 structure, exhibiting remnants of initial
powder particles. When complete resolidification occurred,
they deformed, as shown in case 3 in Fig. 10, and usually
bounced off. This resulted in the low deposition efficiencies observed for large spraying distance conditions.
This detailed investigation provided a better understanding
of agglomerate behavior in the plasma jet and enabled high coating densities. PCP powders, usually of a lower cost, are acceptable for plasma spraying applications despite the presence of
agglomerates.
Acknowledgments
The authors would like to thank P6chiney and Uranium
P6chiney Corporation for providing support and funding for this
investigation.
References
1. E. Kubel, Powders Dictate Thermal Spray Coating Properties, Adv.
Mater. Proc., Vol 138 (No. 6), 1990, p 24-32
2. M. Vardelle,A. Vardelle, A. Denoirjean,and E Fauchais, "Heat Treatmentof ZirconiaPowders with DifferentMorphologiesunder Thermal
Spray Conditions,"MRS Spring Meeting 1990 Proc., Vol 190, 1990,
p55
3. M. Ciftcioglu, M. Akinc, and L. Burkhart, Effect of Agglomerate
Strength on SinteredDensityfor Yttria Powders ContainingAgglomerates of MonosizeSpheres,J. Am. Ceram. Soc., Vo170(No. 11), 1987,
C329-C334
/
4. E Diez, "Microstructure and Interface Bonding Studies in Plasma
Sprayed Yttria Coatings on Graphite," MS thesis, Drexel University,
1991
5. H.S.InghamandA.J. Fabel,ComparisonofPlasma SprayGases, Weld.
J., Vo154, 1975, p 101-105
6. E. Bourdin,E Fauchais,and M. Boulos,TransientHeat Conductionunder Plasma Conditions,Int. J. Heat Mass Transfer, Vol 26 (No. 4),
1983, p 567-582