Chemical, morphological and nano-mechanical characterizations of Al 2O3 thin films deposited by metal organic chemical vapour deposition on AISI 304 stainless steel

Article abstract of DOI:10.1016/j.electacta.2004.10.097

Amorphous alumina coatings of different thickness have been deposited on AISI 304 stainless steel substrates by MOCVD in a hot wall reactor at 380 °C under O₂/H₂O atmosphere. The used aluminium precursor was the high volatile and easy to prepare dimethyl-aluminum-isopropoxide. Selected films were annealed in N₂ and O₂ atmosphere at 500 and 700 °C to evaluate the effects of the thermal treatments on the morphology and on the nano-mechanical properties of the coatings. X-ray diffraction and Rutherford backscattering spectroscopy measurements indicated that both the as grown and annealed films were amorphous and very pure with the correct Al ₂O₃ stoichiometry. The surface morphology, investigated by atomic force microscopy, was free of cracks with a roughness of the films that increases with deposition time and with annealing in oxygen atmosphere. The hardness and the elastic modulus of the films and of the AISI 304 stainless steel substrate were measured by load-depth nano-indentation tests. The results highlighted a significant increase in the Berkovich hardness of the coated samples compared to that of the bulk AISI 304 stainless steel.

Full text of DOI:10.1016/j.electacta.2004.10.097

Electrochimica Acta 50 (2005) 4615–4620
Chemical, morphological and nano-mechanical characterizations
of Al2O3 thin films deposited by metal organic chemical
vapour deposition on AISI 304 stainless steel
M. Natalia, G. Cartaa,∗, V. Rigatob, G. Rossettoa, G. Salmasob, P. Zanellaa
a
Istituto di Chimica Inorganica e delle Superfici, Consiglio Nazionale deller Ricerche, CNR – ICIS Corso Stati Uniti 4, 35127Padova, Italy
b
INFN, Laboratori Nazionali di Legnaro, 2 – 35020 Legnaro PD, Italy
Received 15 February 2004; received in revised form 14 September 2004; accepted 15 October 2004
Available online 24 June 2005
Abstract
Amorphous alumina coatings of different thickness have been deposited on AISI 304 stainless steel substrates by MOCVD in a hot wall
reactor at 380 ◦C under O2/H2O atmosphere. The used aluminium precursor was the high volatile and easy to prepare dimethyl-aluminum-
isopropoxide. Selected films were annealed in N2 and O2 atmosphere at 500 and 700 ◦C to evaluate the effects of the thermal treatments
on the morphology and on the nano-mechanical properties of the coatings. X-ray diffraction and Rutherford backscattering spectroscopy
measurements indicated that both the as grown and annealed films were amorphous and very pure with the correct Al2O3 stoichiometry. The
surface morphology, investigated by atomic force microscopy, was free of cracks with a roughness of the films that increases with deposition
time and with annealing in oxygen atmosphere. The hardness and the elastic modulus of the films and of the AISI 304 stainless steel substrate
were measured by load-depth nano-indentation tests. The results highlighted a significant increase in the Berkovich hardness of the coated
samples compared to that of the bulk AISI 304 stainless steel.
© 2005 Elsevier Ltd. All rights reserved.
Keywords: Alumina coatings; MOCVD; Wear protection; Hardness; Nano-indentation
1. Introduction
they can be obtained by CVD [6] at high substrate tem-
peratures (>900 ◦C). Instead, amorphous Al2O3 coatings can
be obtained at lower temperatures using different deposition
processes such as r.f. magnetron sputtering [7], plasma depo-
sition [8], physical vapour deposition [9], ion beam assisted
deposition [10], sol–gel [11] and MOCVD. The last is par-
ticularly interesting because of its large potential for mass
production and the ability to coat complex shaped samples.
However, for its successful implementation the develop-
ment of new highly volatile non-pyrophoric liquid sources is
advisable to make the whole process more controllable and
safer. So far, several aluminium precursors have been pro-
posed for MOCVD of alumina films [12], however, the most
common aluminium sources are homoleptic aluminum com-
pounds such as trimethylaluminum [13], aluminium-tri-iso-
propoxide [14] and aluminum-tri-sec-butoxide [15]. These
sources present several important drawbacks; the alkyls are
Alumina has gained a great attention as a coating material.
This interest is based on its high performance as functional
material in wear [1] and corrosion [2] protection, in ther-
mal barrier [3] applications and in microelectronics [4]. In
particular, because of the excellent properties such as chemi-
cal inertness, corrosion resistance, good mechanical strength,
high hardness, and transparency, Al2O3 is a good material
for the wear protection of metal surfaces [5]. In the case
of stainless steel substrates, the deposition temperature must
not exceed the tempering temperature in order to avoid the
softening process of the steel. For that reason hard ␣-Al2O3
films are not suitable for the coating of stainless steel as
∗
Corresponding author. Tel.: +39 049 8295942; fax: +39 049 8702911.
E-mail address: carta@icis.cnr.it (G. Carta).
0013-4686/$ – see front matter © 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2004.10.097

4616
M. Natali et al. / Electrochimica Acta 50 (2005) 4615–4620
highly pyrophoric while the alkoxides are solid and have
low volatility. On the contrary, a good compound with a
high volatility and a lower dangerousness is the heterolep-
tic dimethyl-aluminium-isopropoxide. With this precursor,
amorphous Al2O3 depositions occur at moderate tempera-
ture (<400 ◦C) in a atmosphere of oxygen mixed with water
vapour [16]. The obtained films are crack and pore free and
present good surface uniformity.
For the wear protection of the alumina coatings it is also
important to characterize the nano-hardness of the obtained
deposit as the surface and near-surface mechanical proper-
ties of thin films can be critical for their final performances.
In this regard, several studies concerning the deposition and
the characterization of protective alumina coatings on stain-
less steel substrates have been reported [11,17–20], but most
investigations concern alumina used as anticorrosion coat-
ing and, to our knowledge, only few data have been reported
about the micro- and nano-hardness of the obtained coatings
of the system Al2O3/stainless steel [21,22].
Therefore, in the present work, we report the results of
an investigation of the compositional, morphological and
nano-mechanical properties of alumina films obtained by
MOCVD on AISI 304 stainless steel, using the volatile
dimethyl-aluminium-isopropoxide as aluminum source. The
films were characterized by X-ray diffraction (XRD), atomic
force microscopy (AFM), Rutherford backscattering spec-
troscopy (RBS) and profilometer scans. The nano-hardness
and elastic properties of both the film and substrate were eval-
uated using the load-depth curve measured by depth sensing
indentation technique and finally, the study of the effects of
different annealing conditions on the surface morphology and
nano-mechanical properties of the coatings have been con-
sidered.
the RUMP software. ERDA measurements were performed
using the HVEC 2.5 MeV AN 2000 accelerator, with a
2.2 MeV He+ beam, in order to investigate Hydrogen con-
tent. AFM characterization of surface morphology was per-
formed with a DME DualscopeTM instrument using silicon
cantilevertipswithnominaltipradiusof10 nminnon-contact
mode. X-ray diffraction measurements were made with a
Philips PW1830 powder diffractometer using Cu K␣ radia-
tion. Film thicknesses were determined using a Tenkor P-10
surface profiler by analysing the film height steps on par-
tially masked samples. The obtained thickness values were
cross-checkedforselectedwiththevaluesobtainedfromRBS
spectra.
Nano-indentation was used to determine film hardness
and elastic modulus using a Nano-test 600 instrument from
Micromaterials Ltd, with a Berkovich (three-sided pyrami-
dal) diamond indenter. The following peak loads were used,
for each sample: 2.5, 5, 7.5, 10, 15, 17.5, 20, 30, 40, 50,
65, 80, 100, 120, 160, 200 mN with loading rate = unloading
rate that were varied in proportion to the peak loads start-
ing at a value of 0.1 mN/s for the 2.5 indentations, while
common experimental conditions as initial (contact) load
0.05 mN and holding period at peak load 10 s were used
for all the measurements. The indentations were repeated
at least five times at each load on different regions of the
sample surface apart 50 ␮m. The hardness and reduced mod-
ulus have been determined from these indentation curves
using a method originally proposed by Oliver and Pharr
which fits a power-law function to the unloading curve [23].
The Young’s modulus of the samples has been calculated
from the reduced modulus via the equation which includes
the effects of non-rigid indenters on the load-displacement
behaviour: 1/Er = (1 − ν2)/E + (1 − νi2)/Ei where E and
ν are the Young’s modulus and Poisson’s ratio for the spec-
imen, respectively; Ei (1141 Gpa) and νi (0.07) the corre-
sponding indenter quantities; and Er is the reduced modulus
which has been obtained by the initial slope of the unloading
curve.
2. Experimental
Alumina coatings were grown in a horizontal hot wall
reactor at a reduced pressure (2 Torr) at a temperature of
380 ◦C using dimethyl-aluminium-isopropoxide. The carrier
gas was nitrogen (10 scc/min) flowing through the bubbler
containing the aluminium source, thermostatically set to
a temperature of 10 ◦C, while the reactant gas (O2 + H2O
vapour mixture, 200 scc/min) was introduced into the main
flow, after the precursor evaporation zone, in close proxim-
ity to the entry of the reaction chamber, in order to avoid
the decomposition of the precursor. The substrate used in the
experiments was AISI 304 stainless steel. The stainless sam-
ples were ground on SiC paper with a final size of 4000 grit
and polished, and all samples were cleaned and degreased
with hot trichloethylene. Annealing of the samples was per-
formed in the same reactor in an N2 and O2 atmosphere.
RBS measurements were performed at the CN accel-
erator at LNL-Legnaro, Italy, using a 2 MeV He2+ beam.
Fixed random spectra were recorded and the film stoichiom-
etry and thickness obtained by simulating the spectra with
3. Results and discussion
3.1. Coating characterization
Both the as grown and the annealed films were amorphous
as determined by X-ray diffraction. From RBS analysis (see
Fig. 1) only Al and O film signals, floating on a background of
Fe and Cr from the substrate, were detected and a stoichiom-
etry of Al:O ∼ 2:3 was found. No carbon contamination was
detected within the RBS detection limit that we estimate to
be about 2 at.%. Moreover, from ERDA measurements also
the hydrogen content of the films was found to be below
2 at.%. As well together with the absence of C impurities this
indicates that organic residues that may be produced during
the precursor decomposition process, fully desorbs from the
growth surface.

M. Natali et al. / Electrochimica Acta 50 (2005) 4615–4620
4617
Fig. 3. RMS roughness value for alumina coating at different thickness.
Fig. 1. RBS spectrum of Al2O3 film on AISI 304 stainless steel, collected
with a 2.2 MeV He+ beam. Al and O signals are seen, floating on a back-
ground of Fe and Cr from the steel substrate. The film thickness is 176 nm
assuming a density of 3.42 g cm−3 while the stoichiometry, determined from
the ration of the Al and O signals, is about Al:O ∼ 2:3.
that increased with the film thickness. The root mean square
(RMS) roughness values were determined by analysing sev-
eral areas of the AFM images, excluding from the calculation
the deep trenches visible as vertical black lines in Fig. 2(a–d),
that correspond to scratches of the substrate remaining after
the polishing process. The standard deviation of the rough-
ness values obtained from several such determinations are
also reported. In Fig. 3 the root mean square roughness val-
ues are plotted against the film thickness and a linear fit to
the data is reported. Clearly the surface roughness increases
In Fig. 2 AFM images of Al2O3 films with different thick-
ness (b–d) are shown as well as the surface morphology
of the polished steel substrate (a). The surface morphol-
ogy of the films presents globular aggregates with diameter
Fig. 2. AFM morphology of: (a) AISI 304 stainless steel substrate; (b) alumina coating thickness 480 nm; (c) alumina coating thickness 1420 nm; (d) alumina
coating thickness 6080 nm. Root mean square (RMS) roughness values are reported for each sample determined as described in the text.

4618
M. Natali et al. / Electrochimica Acta 50 (2005) 4615–4620
Fig. 4. AFM imagines of alumina coating: (a) as grown; (b) annealing 30 min at 500 ◦C in O2; (c) annealing 30 min at 700 ◦C in O2.
with the film thickness, i.e., with deposition time. In Fig. 4
AFM images of an alumina film are shown in its as grown
state (a) and after annealing in O2 atmosphere for 30 min at
2 Torr at 500 ◦C (b) and 700 ◦C (c). Again the formation of
a globular texture, similar to that of thick films is observed,
accompanied by an increase of the surface roughness with
annealing temperature.
the unloading segment is almost vertical typical of a ductile
plastic material such as most metals. This indicates that the
elastic strains during indentation are small compared with
the plastic strains, so that most of the indenter displacement
is accommodated plastically giving little elastic recovery of
the indentation depth (5.8%). For the coated sample a larger
indenter depth is recovered as the load is removed. This indi-
cates that the composite material presents a larger degree of
elastic recovery during unloading (20.7%). In any case it must
be noted that to determine the intrinsic hardness value for the
alumina film the indentation depth should be considerably
smaller than the film thickness since for larger indentation
depths a composite hardness that also includes a contribu-
tion from the substrate will be measured. This behaviour
is well observed in Fig. 6 that shows the variation in hard-
ness with depth calculated from the indentations, covering the
loading range 2.5–200 mN, for uncoated and alumina coated
(thickness = 480 nm) AISI 304 surfaces. Both samples show
a decrease in the hardness as the indentation depth increases;
this trend, for uncoated sample, is observed for indenting up
to about 500 nm depth, while for larger depths the hardness
reaches a constant value. Probably the higher hardness near to
the surface represents a change in the mechanical properties
of the near surface region due to a previous spontaneous oxi-
dation of the alloy or work hardening due to polishing. The
decrease of the hardness observed in the coated AISI 304
is much more pronounced and the hardness values quickly
approach that of the substrate. This behaviour is typical for a
hard film on a softer substrate. The measured hardness of the
3.2. Nano-indentation measurements
The nano-indentation measurements were performed on
the surface of both uncoated AISI 304 stainless steel and
on alumina coatings in order to determine their hardness
and elastic properties. The values obtained for coatings, as
previously reported [22], depend on the film thickness and
on the indentation depth reached at the maximum load. In
any case the alumina coated samples present higher hardness
values. In Fig. 5 it is shown a comparison of indentation
load-depth curves for bare AISI 304 and for an alumina
film (thickness 2640 nm) on AISI 304 for a single inden-
tation carried out to peak loads of 65 mN. The difference
in hardness of the two samples is immediately clear as evi-
denced from the large difference in the depth obtained at
the maximum load. The uncoated AISI 304 sample exhibits
a larger indentation depth and an almost double amount of
creep deformation during the 10 s hold period at maximum
load. It is also interesting to observe the slope of the unload-
ing segment of the load-displacement graphs and to assess
the elastic recovery on unloading. For AISI 304 the slope of
Fig. 5. Load-displacement curves for: (a) AISI 304 stainless steel substrate;
(b) alumina coating thickness = 2640 nm.
Fig. 6. Variation of hardness with depth for: (a) AISI 304 steel substrate; (b)
alumina coating thickness = 480 nm.

M. Natali et al. / Electrochimica Acta 50 (2005) 4615–4620
4619
Fig. 8. Variation of hardness with depth for (a) alumina coating as grown
and alumina coating after annealing at 500 ◦C (b) and 700 ◦C (c).
Fig. 7. Variation of hardness with depth for (a) AISI 304 stainless steel
substrate and alumina coating, thickness: (b) 480, (c) 1420, (d) 2640, (e)
6080 nm.
4. Conclusions
coating can be very close to the true hardness of the coating
only for loads that give indentation depths about 10 times
smaller than the film thickness with the contribution of the
substrate being less than 2%. On the basis of these consider-
ations we have grown several samples of different thickness
whose hardness values at different indentation depth are
reported in Fig. 7. Each curve shows a first region, at low
indentation depth, where the coating hardness is almost con-
stant and a second region at greater indentation depths where
a progressively increasing substrate effect can be observed.
The range of the hardness plateau depends on the thickness
of the coating and the average of the values obtained in this
plateau region were used to calculate the intrinsic value of the
hardness of the alumina coatings reported in Table 1 together
with the values of reduced modulus, determined by unload-
ing curves, and Young’s modulus, calculated by the reduced
modulus assuming the values of ν = 0.3 for stainless steel and
ν = 0.25 for the amorphous alumina coatings. The obtained
values are in good agreement with those previously reported
on Al2O3 films deposited by RF sputtering [22] and electron
beam evaporation [8,21]. The hardness seems not to depend
on the annealing at 500 ◦C both in N2 and O2 atmosphere,
where only a small increase has been observed. A decreasing
of the composite hardness has been observed increasing the
temperature to 700 ◦C as shown in Fig. 8 and Table 1. It is
probable that the reduction of the hardness revealed by the
sample annealed at 700 ◦C could be related to the increas-
ing surface roughness observe in the AFM images. These
observations indicate that the number of defects in the coat-
ing increases with the increasing annealing temperature and
such behaviour can be due to a beginning of conversion of
amorphous alumina into ␥-alumina at higher temperature.
• MOCVD technique was able to prepare alumina coatings
on stainless steel at moderate low temperature.
• The obtained films were stoichiometric, amorphous, only
slightly contaminated by carbon and hydrogen and had
good mechanical properties.
• The coated stainless steel resulted harder and less stiff than
the uncoated substrate.
• The annealing of the alumina coating at 500 ◦C in N2 or
O2 atmosphere for 30 min did not change significantly the
hardness, while annealing at 700 ◦C slightly degraded the
surface morphology and decreased the hardness by a small
amount.
References
[1] S. Ruppi, in: Proceedings of the EUROCVD 13, Athens, August
26–31, 2001, J. Phys. IV 11 (2001) 847.
[2] R. Emmerich, B. Enders, H. Martin, F. Stippich, G.K. Wolf, P.E.
Andersen, J. Ku´delha, P. Luka´c, H. Hasuyama, Y. Shima, Surf. Coat.
Technol. 89 (1997) 47.
[3] G. Wahl, Ch. Metz, S. Samoilenkov, in: Proceedings of the
EUROCVD 13, Athens, August 26–31, 2001, J. Phys. IV 11 (2001)
835.
[4] W.S. Yang, Y.K. Kim, S.-Y. Yang, J.H. Choi, H.S. Park, S.I. Lee,
J.-B. Yoo, Surf. Coat. Technol. 131 (2000) 79.
[5] H.O. Pierson, Handbook of chemical vapour deposition, Noyes Pub-
lications, Westwood, NJ, 1992.
[6] B. Lux, C. Colombier, H. Altena, K. Stjernberg, Thin Solid Films
138 (1986) 49.
[7] B.G. Segda, M. Jacquet, J.P. Besse, Vacuum 62 (2001) 27.
[8] J.C. Barbour, J.A. Knapp, D.M. Follstaedt, T.M. Mayer, K.G. Minor,
D.L. Linam, Nucl. Instrum. Meth. Phys. Res. B 166/167 (2000)
140.
[9] R.G. Frieser, J. Electrochem. Soc. 113 (1966) 357.
[10] I. Shimizu, Y. Setsuhara, S. Miyake, M. Kumagai, K. Ogata,
M. Kohata, K. Yamaguchi, Surf. Coat. Technol. 131 (2000)
187.
Table 1
Summary of evaluated mechanical properties of Al2O3 coatings and 304
stainless steel
[11] J. Masalski, J. Gluszek, J. Zabrezeski, K. Nitsch, P. Gluszek, Thin
Solid Films 349 (1999) 186.
[12] A.R. Barron, in: W.S. Rees Jr. (Ed.), CVD of Non-metals, VCH
Weinheim, 1996, Chapter 5.
[13] T.H. Hua, M. Armgarth, J. Electron. Mater. 16 (1987) 27.
[14] J. Saraie, J. Kwon, Y. Yodogawa, J. Electrochem. Soc. 132 (1985)
890.
Material
H (GPa)
Er (GPa)
E (GPa)
AISI 304 stainless steel
Alumina
Alumina after annealing at 500 ◦C
Alumina after annealing at 700 ◦C
3.84
10.89
11.88
10.56
211.11
168.78
183.05
163.60
233.33
184.29
198.29
178.91
The calculated values have little error (within 5%).

4620
M. Natali et al. / Electrochimica Acta 50 (2005) 4615–4620
[15] D.H. Kuo, B.-Y. Cheung, R.J. Wu, Thin Solid Films 198–399 (2001)
35.
[19] V.A.C. Haanappel, D.V.D. Vendel, H.D. van Corbach, T. Fransen,
P.J. Gellings, Thin Solid Films 256 (1995) 8.
[16] D. Barreca, G.A. Battiston, G. Carta, R. Gerbasi, G. Rossetto, E.
Tondello, P. Zanella, in: Proceedings of the EUROCVD 13, Athens,
August 26–31, 2001, J. Phys. IV, France 11 (2001) 5394.
[17] V.A.C. Haanappel, J.B. Rem, H.D. van Corbach, T. Fransen, P.J.
Gellings, Surf. Coat. Technol. 72 (1995) 1.
[20] E. C¸ elik, A.S. Demirkiran, E. Aver, Surf. Coat. Technol. 116–119
(1999) 1061.
[21] A.K. Dua, V.C. George, R.P. Agarwala, Thin Solid Films 165 (1988)
163.
[22] J.-H. Ahn, D. Kwon, Mater. Sci. Eng. A 285 (2000) 172.
[23] W.C. Oliver, G.M. Pharr, J. Mater. Res. 7 (1992) 1564.
[18] V.A.C. Haanappel, H.D. van Corbach, T. Fransen, P.J. Gellings, Surf.
Coat. Technol. 75 (1995) 13.