Nanostructured alumina coatings manufactured by air plasma spraying: Correlation of properties with the raw powder microstructure

Abstract

High energy ball milled nanostructured Al2O3 (“N”), fused/crushed conventional Al2O3 (“C”) and sintered nanostructured Al2O3 (“S”) powders were air plasma sprayed on 304 stainless steel. The nanostructured powder was composed of nanoparticle agglomerates, whereas the conventional powder consisted of solid granules. The average crystal size of the powders was estimated by X-ray diffraction based methods (the Scherrer equation and the Williamson Hall plot). Deviations between the crystal sizes calculated by the two methods indicated high lattice strain induced by the nanopowder production technique. Sintering of the nanopowder did not cause any considerable grain growth; moreover, the strain was alleviated. The melting degree of the powders, reflected by the γ-Al2O3 content of the coatings, depended on their porosities. Coatings “N” presented the lowest melting degree due to the inherent porosity of the agglomerated nanoparticles composing powder “N”. As a result, their microstructure was characterized by high porosity and extensive microcracking. The “S” coatings exhibited higher melting degree than that of the “N” coatings (similar to that of the “C” coatings), due to a tighter microstructure attained by sintering. At the same time, part of the initial nanostructure had been preserved during sintering and spraying. The “S” coatings presented the highest adhesion because they combined a high melting degree with pockets of retained nanostructure; the latter could act as crack arresters. Increasing the spray power led to an increase in the melting degree and, consequently, a decrease in the coating porosity and an increase in the porosity affected properties (adhesion and hardness).

Full text

Journal of Alloys and Compounds 495 (2010) 611–616
Contents lists available at ScienceDirect
Journal of Alloys and Compounds
journal homepage: www.elsevier.com/locate/jallcom
Nanostructured alumina coatings manufactured by air plasma spraying:
Correlation of properties with the raw powder microstructure
D. Zoisa,b, A. Lekatoub,∗, M. Vardavouliasa, A. Vazdirvanidisc
aPyrogenesis SA, Technological Park of Lavrion, 19500 Lavrion, Greece
bUniversity of Ioannina, Department of Materials Science & Engineering, 45110 Ioannina, Greece
cHellenic Centre for Metal Research, Pireos 252, 17778 Athens, Greece
article info
Article history:
Received 3 July 2008
Received in revised form 9 October 2009
Accepted 12 October 2009
Available online 20 October 2009
Keywords:
Nanostructured alumina
Coating materials
Sintering
X-ray diffraction
SEM
abstract
High energy ball milled nanostructured Al2O3(“N”), fused/crushed conventional Al2O3(“C”) and sintered
nanostructured Al2O3(“S”) powders were air plasma sprayed on 304 stainless steel. The nanostructured
powder was composed of nanoparticle agglomerates, whereas the conventional powder consisted of
solid granules. The average crystal size of the powders was estimated by X-ray diffraction based methods
(the Scherrer equation and the Williamson Hall plot). Deviations between the crystal sizes calculated
by the two methods indicated high lattice strain induced by the nanopowder production technique.
Sintering of the nanopowder did not cause any considerable grain growth; moreover, the strain was
alleviated. The melting degree of the powders, reflected by the ␥-Al2O3content of the coatings, depended
on their porosities. Coatings “N” presented the lowest melting degree due to the inherent porosity of the
agglomerated nanoparticles composing powder “N”. As a result, their microstructure was characterized
by high porosity and extensive microcracking. The “S” coatings exhibited higher melting degree than
that of the “N” coatings (similar to that of the “C” coatings), due to a tighter microstructure attained
by sintering. At the same time, part of the initial nanostructure had been preserved during sintering
and spraying. The “S” coatings presented the highest adhesion because they combined a high melting
degree with pockets of retained nanostructure; the latter could act as crack arresters. Increasing the spray
power led to an increase in the melting degree and, consequently, a decrease in the coating porosity and
an increase in the porosity affected properties (adhesion and hardness).
© 2009 Elsevier B.V. All rights reserved.
1. Introduction
Nanostructured bulk materials exhibit considerably higher ten-
sile strength, fracture toughness, hardness and wear resistance
when compared to their microstructured counterparts. These
improvements are associated with grain sizes smaller than 100 nm
[1–4]. Various thermal spraying techniques have been employed to
prepare nanostructured oxide-based coatings, such as: high veloc-
ity oxygen fuel-HVOF (TiO2)[5,6], vacuum plasma spray-VPS (TiO2)
[7], plasma spraying of liquid precursors (ZrO2, Al2O3, TiO2)[8],
atmospheric plasma spraying-APS (TiO2)[5],Al
2O3–TiO2[9–11],
ZrO2[12]. TiO2and Al2O3–TiO2nanostructured coatings have been
reported to possess superior wear resistance, adhesion, tough-
ness and spallation resistance than their conventional counterparts
[6,10,11].
Amongst nanoparticle production techniques, such as sol–gel
synthesis, inert gas condensation and high energy milling [13–15],
∗Corresponding author. Tel.: +30 26510 97309; fax: +30 26510 97034.
E-mail address: alekatou@cc.uoi.gr (A. Lekatou).
the latter is commonly used to produce large quantities of
nanopowders, because it is a low cost process and is applicable to
a variety of materials [15]. High energy ball milling is, by far, more
effective than conventional ball milling in reducing particle size;
however, it induces high strains to the milled particles.
Lin et al. [16] reported that sintering of alumina–3 wt% titania
nanopowder can increase the density of the granules and assist
spherical shape retention during plasma spraying. By sintering,
fragmentation of the granules is avoided; very fine particles would
not possess the momentum to infiltrate the flame stream or, in
the case of entering the flame, they would rapidly get melted [17],
eliminating any nanostructure. No considerable growth has been
noticed after sintering of nano- and sub-micron alumina in the
temperature range 1150–1275 ◦C[18,19]. Chinelatto et al. [20] sin-
tered high energy milled nanostructured ␣-Al2O3powder in two
stages: At the first stage (950 ◦C), surface diffusion occurred; it did
not cause any significant densification, but led to smoothing of
the agglomerate surface and coalescence of the finest and coars-
est particles. At the second stage (1300 ◦C), the density and grain
size were significantly increased. The sintering temperature for ␣-
Al2O3agglomerate-free nanopowder should be about 300–400 ◦C
0925-8388/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2009.10.055

612 D. Zois et al. / Journal of Alloys and Compounds 495 (2010) 611–616
Fig. 1. The feedstock: (a) Conventional crushed and fused powder. (b) High energy ball milled nanopowder. (c) The nanopowder after sintering (1100 ◦C, 2h). Agglomerates
of nanoparticles, (d) as received, (e) after sintering.
lower than that for Al2O3micropowder [21,22], because of a lower
activation energy in comparison to the alumina micropowder.
In the present work, the effect of three different types of raw
powders (high energy ball milled nanostructured Al2O3, sintered
nanostructured Al2O3and conventional fused/crushed conven-
tional Al2O3) on characteristic properties of air plasma sprayed
coatings, is studied. This work has been conducted within the
framework of investigating the potential replacement of conven-
tional alumina coatings in thread and fibre high temperature
applications with advanced coatings for improved wear and erosion
performance.
2. Experimental
The nanopowder, of particle size (−53 + 10) ␮m, was produced by a labora-
tory high energy ball mill. The conventional powder, of nominal particle size
(−45 + 15) ␮m, was Sulzer Metco’s 105NS. Spraying was performed by a MiniGun
torch (Pyrogenesis Inc) upon 304 stainless steel coupons. Coating porosities and
microstructures were evaluated by optical microscopy (Leica DMLM) and SEM-EDX
Table 1
Crystal sizes and strains of the feedstocks.
Material Crystal size—Scherrer (nm) Crystal size—Williamson Hall (nm) Strain—Williamson
Hall (%)
Conventional powder (“C”) 100 100 0.00
Nano powder (“N”) 24 93 1.15
Sintered nano powder (“S”) 122 >100 0.20

D. Zois et al. / Journal of Alloys and Compounds 495 (2010) 611–616 613
(Philips XL 40 SFEG and JEOL JSM 5600). Microhardness (14 measurements per coat-
ing) was measured by a Shimadzu tester. Adhesion was determined on the principle
of applying a controlled pneumatic force to a piston stud glued to the coating. Three
measurements per coating were conducted by a portable elcometer (110 P.A.T.T.I.),
according to the ASTM C633-01 standard. Sintering was carried out in a Nabertherm
box furnace (air, 1100 ◦C, 2h). X-ray diffraction patterns were obtained by a Bruker
D8 Advance diffractometer. The nanocrystalline size was calculated by two XRD-
based methods, the Scherrer equation and the Williamson Hall plot. The former is
only capable of defining crystal sizes, while the latter is also capable of calculating
lattice strain [23]. Phase percentages in powders and coatings were determined by
the XRD-based Rietveld analysis [24].
3. Results and discussion
3.1. Feedstock characterization
Fig. 1a–c illustrates the morphologies of the conventional (“C”),
nanostructured (“N”) and sintered (“S”) powders, respectively.
Powder “C” consists of solid irregular granules, whereas pow-
ders “N” and “S” consist of agglomerates of nanoparticles having
smooth shapes (a typical morphological result of the milling pro-
cess [15]). Agglomerates of nanoparticles “N” are discerned in
Fig. 1d. Macroscopically, both sintered and “green” nanopowders
present similar morphologies and particle sizes.Therefore, it is
considered that the main sintering mechanism is material redis-
tribution, aiming at decreasing the surface energy of particles by
reducing their available surface area. Indeed, on closer examina-
tion (and in compatibility with previous works [20]), smoothing of
the agglomerate surface and a decrease in the finer particles can be
discerned. The latter have either been consolidated or embodied
to the coarser ones (Fig. 1e); as finer particles have a higher area-
to-volume ratio, the spontaneity to reduce their surface energy is
higher, causing their consolidation. Coalescence occurs by neck-
ing (pointed by arrows in Fig. 1e), which is typical for alumina
[25].
The XRD patterns of Fig. 2 reveal that the main phase in the three
powders is ␣-Al2O3.Aland␣-Fe are also detected in powder “N”
(1.1 and 1.14 wt%, respectively, as determined by Rietveld analy-
sis). The stoichiometry used in the powder production is the main
reason for the aluminium presence. The iron presence suggests
powder contamination by the tool steel balls and vials of the ball
mill. Powder “N” presents the widest peaks amongst the feedstocks,
suggesting nanocrystalline dimensions and possible non-uniform
strain exertion. Powder “S” presents narrower peaks, though wider
than the “C” powder peaks, indicating considerable presence of sub-
micron crystals. Sintering has caused oxidation of Fe to Fe2O3,in
free form and in solution with Al2O3.The latter is indicated by the
␣-phase peak displacement at lower angles (Fig. 2).
Fig. 2. XRD patterns of the three powders and their coatings. ␣:␣-Al2O3,␥:␥-Al2O3.
Table 1 lists the nanopowder crystal sizes calculated by the
Scherrer equation and Williamson Hall plot. These methods are
considered quite accurate for particle sizes less than 100 nm;
beyond this limit, accuracy gradually decreases with size [26,27].
Therefore, they do not provide any information about the crys-
tal size of powder “C”. The quite different crystal sizes of powder
“N”, estimated by the two methods, are owing to lattice strain.
The consecutive high energy impacts during milling induced high
strain and distortion in the grains, which were imprinted on the
peak widths. The Williamson Hall method, being capable of distin-
guishing between size and strain contribution, yields more accurate
crystal size values. Table 1 also shows that sintering has caused
strain relief, serving as an annealing stage. The grain sizes calcu-
lated by Scherrer and Williamson Hall would possibly be similar
for the “S” powder, if accurate Williamson Hall calculations in the
sub-micron range were feasible.
3.2. Characterization of the deposited coatings
A major issue in spraying nanoparticles is the retention of part
of the original nanostructure in the coating, in order to maintain
Fig. 3. Overview cross-sections of the coatings: (a) “C3”, (b) “N3”, (c) “S3”.

614 D. Zois et al. / Journal of Alloys and Compounds 495 (2010) 611–616
Table 2
Quantitative phase analysis and properties of the coatings.
Designation Spraying power (kW) ␣-Al2O3(wt%) ␥-Al2O3(wt%) Porosity (%) Adhesion (MPa) Hardness (HV100)
C125 18.1–23.2 76.8–81.9 6.6 ±0.6 17.6 ±0.9 1063 ±145
C232 9.1–10.8 89.2–90.9 6.1 ±0.7 18.1 ±1.5 1078 ±259
C339 8.3–9.9 90.1–91.7 5.5 ±0.8 19.8 ±1.6 1127 ±105
N125 51.9–54.6 45.4–48.1 15.3±5.8 10.5 ±0.8 972 ±120
N232 48.0–52.0 52.0–48.0 13.2±2.0 11.4 ±1.3 1070 ±166
N339 17.1–23.3 76.7–82.9 10.0±0.1 17.7 ±2.4 1131 ±128
S125 14.9–19.7 80.3–85.1 7.3 ±1.5 24.1 ±2.1 1034 ±152
S232 12.8–15.3 84.7–87.2 7.7 ±2.6 23.4 ±1.8 1065 ±179
S339 10.3–11.5 88.5–89.7 5.9 ±0.5 26.3 ±2.2 1109 ±122
its beneficial effect on the properties. On the other hand, coat-
ings must exhibit integrity, cohesion and adhesion, which can only
be achieved by a sufficient degree of particle melting; the latter
eliminates nanostructure.
Fig. 4. Cross-sections of coating “N” showing: (a) a pocket of semi-molten nanopar-
ticles associated with an asperity in “N1”, (b) hollow spherical particles in “N1”,
(c) a fractured surface of semi-molten particles in “N3”; low in-particle cohesion
generates their splitting.
Cross-sections of the manufactured coatings are illustrated in
Fig. 3. The coatings seem uniform and fairly adherent to the sub-
strate. Table 2 quantifies the microstructure related properties and
the phase composition of the coatings.
Fig. 5. Coating properties vs. torch power: (a) porosity, (b) adhesion, (c) hardness.

D. Zois et al. / Journal of Alloys and Compounds 495 (2010) 611–616 615
3.2.1. Correlation of the powder microstructure to the coating
microstructure
The coatings consist of ␣-Al2O3and ␥-Al2O3, as shown in the
XRD patterns of Fig. 2.␥-Al2O3results from rapid quenching of
molten Al2O3.␣-Al2O3corresponds to the unmelted ␣-Al2O3frac-
tion in the powder [28].␥-Al2O3has a lower energy barrier for
nucleation than ␣-Al2O3[29] leading to its preferential nucleation.
Table 2 shows that at low and medium spraying power, the ␣-
Al2O3content of the “N” coatings is notably higher than that of the
“C” and “S” counterparts. This suggests that, during spraying, a large
fraction of the initial nanoparticles “N” remained unmelted. The
lower melting degree of feedstock “N” is attributed to the inherent
porosity of the “green” agglomerated nanoparticles: Being compact
bodies, conventional particles (“C”) or even sintered nanoparti-
cles (“S”) exhibit higher thermal conductivity than their porous
counterparts (“N”) [30]. From a microstructure standpoint, the
low melting degree of the “N” coatings is reflected by: (a) a high
percentage of partially deformed semi-molten particles, which
are incapable of filling any asperities and gaps of the previously
deposited layers (Fig. 4 a); and (b) a high presence of semi-molten
spheroid particles with large internal voids (Fig. 4b and c). Their
formation is attributed to air impelling through internal capillaries
towards the core of the agglomerated particles (“N”) and subse-
quent heat induced expansion [31]. This phenomenon was also
encountered in the “S” coatings but at a lower extent, as the tighter
structure prevented the impelling of large air quantities.
Table 2 shows that the “S” particles have attained slightly lower
melting degrees than their “C” counterparts. The tighter structure
Fig. 6. Coating-substrate interfaces: (a) “N3”, (b) “S3”. The “N3” coating presents a
looser interface with notably more intensive microcracking in its vicinity.
of the “S” particles due to sintering consequences, such as (limited)
nanoparticle coalescence and increased intraparticle cohesion, has
increased their thermal conductivity, leading to melting degrees
that are comparable to those of the “C” particles.
3.2.2. Correlation of the coating microstructure to the coating
properties
Fig. 5 presents the trends of porosity, adhesion strength and
hardness vs. torch power for the three types of coatings.
As a consequence of their lower melting degree, the “N” coat-
ings present higher porosities than the “C” and “S” counterparts
(Figs. 3 and 5). Fig. 5b demonstrates that the adhesion strength
of the coatings increases with decreasing porosity; thus, the “N”
coatings present the lowest adhesion. Fig. 4c illustrates the frac-
tured surface of semi-molten deposited particles “N”: Cohesion
and not adhesion (to the substrate) failure was the actual rea-
son for the coating failure. Two factors contributed to this: (i) The
increased porosity of the “N” coatings and (ii) the loose structure
of the unmolten “N” agglomerates that was unable to maintain its
integrity under tensile forces.
Coatings “S” present significantly higher adhesion in compar-
ison not only to the nanostructured but also to the conventional
counterparts, regardless of the spraying power. This increase
in adhesion is attributed to: (a) sintering consequences, such
as increased intraparticle cohesion, particle coalescence to sub-
micron particles (Fig. 1e), as well as strain relief (Table 1); the
positive effect of sintering on the intraparticle cohesion and the
Fig. 7. (a) A semi-molten pocket in coating “S3” showing that the initial nanostruc-
ture has been retained. (b) The nanozone can hinder crack propagation in coating
“S3”.

616 D. Zois et al. / Journal of Alloys and Compounds 495 (2010) 611–616
consequent enhancement of the mechanical anchoring between
consecutive splats is suggested by the reduction of microcracking
in the respective coating structure (Fig. 6) and (b) nanostructure
retaining; despite the similar fractions of molten phase in coatings
“S” and “C” (Table 2), coatings “S” have retained part of the powder
nanostructure, as illustrated in Fig. 7a. The retained nanostructure
can hinder crack propagation, as demonstrated in Fig. 7b.
Table 2 shows that coatings “N” exhibit similar hardnesses with
coatings “C” and “S”. This can be explained by the employment of
a small indentation load, so that the hardness values depended to
a major extent on the intrinsic properties of the materials and to a
less extent on their extrinsic characteristics (e.g. pores, defects).
3.2.3. Correlation of the spray power to the coating
microstructure and properties
The properties of the coatings are affected by the high incidence
of semi-molten/unmelted particles, as demonstrated in Table 2.
Porosity is higher at the lower spraying energies, where the semi-
molten particles are less deformed. At higher spraying energies,
molten material of high diffusivity would fill the asperities and
gaps of the previously deposited layers, leading to lower porosi-
ties. Hardness increases with spraying power (Fig. 5c), possibly as a
result of the porosity decrease [32]. The high increase in the adhe-
sion of the “N” coatings at the highest energy, is attributed to a
boost in particle melting, demonstrated by the large decrease in
the ␣-phase content (Table 1).
4. Conclusions
Characterization of air plasma sprayed nanostructured Al2O3
(“N”), sintered nanostructured Al2O3(“S”) and fused/crushed con-
ventional Al2O3(“C”) powders has led to the conclusions:
1. Powder “N” consisted of agglomerates of nanoparticles of aver-
age size 93 nm and lattice strain 1.15%.
2. Sintering of the nanopowder at 1100 ◦C for 2 h, caused surface
smoothing, limited coalescence and strain relief.
3. The low melting degree of the nano-feedstock (“N”) is attributed
to the inherent porosity of the “green” agglomerated nanoparti-
cles.
4. The relatively high melting degree of the “S” coatings is
attributed to their tighter microstructure obtained by sintering.
5. The high porosity of the “N” coatings is partly related to the large
voids in the interior of the semi-molten particles and partly to
their low deformation ability.
6. The “N” coatings presented the lowest adhesion due to the high-
est porosity and the weakest mechanical anchoring, both direct
consequences of the low melting degree.
7. The “S” coatings presented the highest adhesion because they
combined a high melting degree with zones of retained nanos-
tructure that can act as crack arresters.
8. Increasing the spray power led to a higher melting degree of the
powder and a consequent improvement of the coating proper-
ties.
Acknowledgement
D. Zois is grateful for the financial support of the “Hellenic State
Scholarships” foundation.
References
[1] H. Gleiter, Nanostruct. Mater. 1 (1992) 1–19.
[2] H. Hahn, Nanostruct. Mater. 2 (1993) 251–265.
[3] R.W. Siegel, Mater. Sci. Forum 235–238 (1997) 851–860.
[4] K. Jia, T.E. Fischer, Wear 203–204 (1997) 310–318.
[5] C.C. Berndt, E.J. Lavernia, J. Therm. Spray Technol. 7 (1998) 411–420.
[6] R.S. Lima, B.R. Marple, Mater. Sci. Eng. A 395 (2005) 269–280.
[7] Y. Zhu, M. Huang, J. Huang, C. Ding, J. Therm. Spray Technol. 8 (1999)
219–222.
[8] J. Karthikeyan, C.C. Berndt, J. Tikkanen, J.Y. Wang, A.H. King, H. Herman, Nanos-
truct. Mater. 9 (1997) 137–140.
[9] E.H. Jordan, M. Gell, Y.H. Sohn, D. Goberman, L. Shaw, S. Jiang, M. Wang, T.D.
Xiao, Y. Wang, P. Strutt, Mater. Sci. Eng. A301 (2001) 80–89.
[10] Y. Wang, S. Jiang, M. Wang, S. Wang, T.D. Xiao, P.R. Strutt, Wear 237 (2000)
176–185.
[11] M. Gell, E.H. Jordan, Y.H. Sohn, D. Goberman, L.L. Shaw, T.D. Xiao, Surf. Coat.
Technol. 146–147 (2001) 48–54.
[12] H. Chen, C.X. Ding, Surf. Coat. Technol. 150 (2002) 31–36.
[13] S. Shukla, S. Seal, Nanostruct. Mater. 11 (1999) 1181–1193.
[14] L.L. Shaw, D. Goberman, R. Ren, M. Gell, S. Jiang, Y. Wang, T.D. Xiao, P.R. Strutt,
Surf. Coat. Technol. 130 (2000) 1–8.
[15] C. Suryanarayana, Prog. Mater. Sci. 46 (2001) 1–184.
[16] X. Lin, Y. Zeng, S.W. Lee, C. Ding, J. Eur. Ceram. Soc. 24 (2004) 627–634.
[17] A. Agarwal, T. McKechnie, S. Seal, J. Therm. Spray Technol. 12 (2003)
350–359.
[18] L.C. Lim, P.M. Wong, M. Jan, Acta Mater. 48 (2000) 2263–2275.
[19] G.R. Karagedov, N.Z. Lyakhov, Nanostruct. Mater. 11 (1999) 559–572.
[20] A.S.A. Chinelatto, E.M.J.A. Pallone, V. Trombini, R. Tomasi, Ceram. Int. 34 (2007)
2121–2127.
[21] W. Zeng, L. Gao, L. Gui, J. Guo, Ceram. Int. 25 (1999) 723–726.
[22] J.G. Li, X. Sun, Acta Mater. 48 (2000) 3103–3112.
[23] H.P. Klug, L.E. Alexander, X-Ray Diffraction Procedures, 2nd ed., Wiley, New
York, 1974.
[24] M.L. Gualtieri, M. Prudenziati, A.F. Gualtieri, Surf. Coat. Technol. 201 (2006)
2984–2989.
[25] H. Schaper, E.B.M. Doesburg, P.H.M. De Korte, L.L. Van Reijen, Solid State Ionics
16 (1985) 261–266.
[26] G.K. Williamson, W.H. Hall, Acta Metall. 1 (1953) 22–31.
[27] B.D. Cullity, Elements of X-Ray Diffraction, Addison-Wesley, 1956.
[28] M. Wang, L.L. Shaw, Surf. Coat. Technol. 202 (2007) 34–44.
[29] J. Zhang, J. He, Y. Dong, X. Li, D. Yan, Rare Metals. 26 (2007) 391–397.
[30] X. Lin, Y. Zeng, X. Zheng, C. Ding, Surf. Coat. Technol. 105 (2005) 85–90.
[31] Z. Karoly, J. Szepvolgyi, Powder Technol. 132 (2003) 211–215.
[32] H. Luo, D. Goberman, L. Shaw, M. Gell, Mater. Sci. Eng. A346 (2003)
237–245.