Electrochemical and time-of-flight secondary ion mass spectrometry analysis of ultra-thin metal oxide (Al2O3 and Ta2O 5) coatings deposited by atomic layer deposition on stainless steel

Article abstract of DOI:10.1016/j.electacta.2011.02.074

Ultra-thin (5-50 nm) layers of aluminium and tantalum oxides deposited by atomic layer deposition (ALD) on a stainless steel substrate (316L) for corrosion protection have been investigated by electrochemical methods (linear scan voltammetry, LSV, and electrochemical impedance spectroscopy, EIS) and time-of-flight secondary ion mass spectrometry, ToF-SIMS. The effects of the deposition temperature (250 °C and 160 °C) and coating thickness were addressed. ToF-SIMS elemental depth profiling shows a marked effect of the organic and water precursors used for deposition and of the substrate surface contamination on the level of C and OH trace contamination in the coating, and a beneficial effect of increasing the deposition temperature. The polarization data show a decrease of the current density by up to four orders of magnitude with increasing coating thickness from 5 to 50 nm. The 50 nm films block the pitting corrosion in 0.8 M NaCl. The uncoated surface fraction (quantified from the current density and allowing a ranking of the efficiency of the coating, also confirmed by the capacitance and resistance values extracted from the EIS data) was 0.03% with a 50 nm thick Al₂O₃ film deposited at 250 °C. The correlation between the porosity values of the coatings and the level of C and OH traces observed by ToF-SIMS points to a marked effect of the coating contaminants on the sealing performance of the coatings and on the corrosion resistance of the coated systems.

Full text of DOI:10.1016/j.electacta.2011.02.074

Electrochimica Acta 56 (2011) 10516–10523
Contents lists available at ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Electrochemical and time-of-flight secondary ion mass spectrometry analysis of
ultra-thin metal oxide (Al O and Ta O ) coatings deposited by atomic layer
2
3
2
5
deposition on stainless steel
a
a
a,∗
a
b
´
Belén Díaz , Jolanta Swiatowska , Vincent Maurice , Antoine Seyeux , Bernard Normand ,
c
c
a
Emma Härkönen , Mikko Ritala , Philippe Marcus
Laboratoire de Physico-Chimie des Surfaces, CNRS (UMR 7045) - ENSCP (Chimie-ParisTech), 11 rue Pierre et Marie Curie, F-75005 Paris, France
Université de Lyon, INSA de Lyon, MATEIS (UMR CNRS 5510), Bât. L. de Vinci, F-69621 Villeurbanne, France
Laboratory of Inorganic Chemistry, University of Helsinki, P.O. Box 55, FIN-00014 Helsinki, Finland
a
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
Ultra-thin (5–50 nm) layers of aluminium and tantalum oxides deposited by atomic layer deposition (ALD)
on a stainless steel substrate (316L) for corrosion protection have been investigated by electrochemical
methods (linear scan voltammetry, LSV, and electrochemical impedance spectroscopy, EIS) and time-
Received 15 November 2010
Received in revised form 17 February 2011
Accepted 18 February 2011
Available online 26 February 2011
◦
of-flight secondary ion mass spectrometry, ToF-SIMS. The effects of the deposition temperature (250 C
◦
and 160 C) and coating thickness were addressed. ToF-SIMS elemental depth profiling shows a marked
effect of the organic and water precursors used for deposition and of the substrate surface contamination
on the level of C and OH trace contamination in the coating, and a beneficial effect of increasing the
deposition temperature. The polarization data show a decrease of the current density by up to four
orders of magnitude with increasing coating thickness from 5 to 50 nm. The 50 nm films block the pitting
corrosion in 0.8 M NaCl. The uncoated surface fraction (quantified from the current density and allowing a
ranking of the efficiency of the coating, also confirmed by the capacitance and resistance values extracted
Keywords:
Oxide coatings
Corrosion protection
ALD
EIS
ToF-SIMS
◦
from the EIS data) was 0.03% with a 50 nm thick Al2O3 film deposited at 250 C. The correlation between
the porosity values of the coatings and the level of C and OH traces observed by ToF-SIMS points to a
marked effect of the coating contaminants on the sealing performance of the coatings and on the corrosion
resistance of the coated systems.
©
2011 Elsevier Ltd. All rights reserved.
1
. Introduction
row holes and trenches. When the deposition parameters such as
temperature, precursor pulse and purge times are properly cho-
sen, each reaction step reaches saturation, making the growth
self-limiting and the amount of material deposited in each step
uniform over the surface. The saturative ALD reactions also ensure
reproducible and precise thickness control simply by counting the
number of reaction cycles. Commercially ALD is used mostly in
microelectronics but also for making invisibly thin antitarnish-
ing coatings on silver jewellery [10]. Despite the excellent film
properties, studies on using ALD for depositing corrosion protec-
tion coatings have remained quite limited. The very first study
[7] showed that a few hundred nanometer thick ALD aluminium,
titanium and tantalum oxide coatings give excellent protection
to stainless steel against corrosion. Ultra thin, about 50 nm thick,
Metallic surfaces can be protected against corrosion by coatings
which separate the metal or alloy from the aggressive environment.
The interest for the application of thin films as protective coatings
hasincreasedrecently. Severaltechniqueshavebeenappliedforthe
deposition of thin layers, including magnetron sputtering, chemical
and physical vapor deposition methods [1–5]. However, the pres-
ence of defects or pinholes in the coatings cannot be completely
avoided and limits the corrosion protection performance [3,6].
Atomic layer deposition (ALD), a thin film deposition method
derived from chemical vapor deposition, is an ideal candidate to
grow advanced ultra-thin and non porous coatings for corrosion
protection. ALD allows the growth of high-quality, thin and uni-
form films with excellent conformality [7–9]. The film uniformity
extends not only over flat substrate surfaces but also into very nar-
TiO layers deposited by ALD, have also been proved to efficiently
2
improve the corrosion performance of stainless steel [11].
In this paper the corrosion resistance properties of ultra-thin
(from 5 to 50 nm) amorphous aluminium and tantalum oxide coat-
∗ Corresponding author.
E-mail address: vincent-maurice@chimie-paristech.fr (V. Maurice).
ings prepared by ALD are presented and discussed. The influence
◦ ◦
of the deposition temperature (250 C vs. 160 C) is addressed.
0
013-4686/$ – see front matter © 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2011.02.074

B. Díaz et al. / Electrochimica Acta 56 (2011) 10516–10523
10517
Table 1
Chemical composition (wt%) of the 316L stainless steel.
Element
wt%
C
Cr
Ni
Mo
2.7
Mn
Si
P
S
N
Fe
Max. 0.03
18–20
13
Max. 2
Max. 0.75
Max. 0.04
Max. 0.03
Max. 0.1
Bal.
The elemental depth distributions in the coatings were analyzed
by means of time of flight secondary ion mass spectroscopy
reflectance (XRR, Bruker AXS D8 Advance) and modelled using
Leptos 7.03.
(
ToF-SIMS). Potentiodynamic polarization measurements were
Table 2 shows the thicknesses and deposition temperatures of
the prepared samples.
performed to analyze the corrosion protection and to quantify the
coating sealing performance, in particular the uncoated surface
fraction. Electrochemical impedance spectroscopy was applied to
characterize the coating quality.
2.3. ToF-SIMS depth profiling
Elemental depth profiles were obtained using a ToF-SIMS 5
spectrometer (IonToF). The spectrometer was run at an oper-
2
. Experimental
ating pressure of 10 mbar. A pulsed 25 keV Bi+ primary ion
source was employed for analysis, delivering 1.8 pA of current
over a 175 ␮m × 175 ␮m area. Depth profiling was performed with
a 2 keV Cs+ sputter beam giving a 100 nA target current over
a 400 ␮m × 400 ␮m area. Data acquisition and post-processing
analyses were performed using the Ion-Spec software. Negative
ion profiles were recorded because they are more sensitive to
fragments originating from oxide matrices. The samples were
examined in exactly the same sputtering and analysis conditions,
allowing a direct comparison of the intensities of the profiles.
The ToF-SIMS analyses were performed on pristine samples not
exposed to the corroding electrolyte. Before the measurements
they were cleaned in ethanol as described below.
−
9
2.1. Material and surface preparation
The chemical composition of the 316 stainless steel substrates is
presented in Table 1. The sample surface was mechanically polished
with successive 9, 3 and 1 ␮m diamond suspension followed by a
final polishing with a SiO2 suspension (0.04 ␮m). The roughness
measured by profilometry was 1.7 nm. Before coating, the samples
were cleaned in an ultrasonic bath of isopropanol for 5 min, rinsed
with ethanol and dried by blowing with compressed air.
2.2. Coating preparation by atomic layer deposition
The coatings were prepared with a Picosun SUNALE R-150
ALD reactor. The precursors employed for the deposition proce-
2.4. Electrochemical measurements
dure were tantalum ethoxide (Ta(OC H5)5), trimethylaluminium
2
◦
(
Al(CH ) ) and H O. Ta(OC H5)5 was evaporated at 140 C while
The electrochemical measurements were carried out in a con-
ventional three-electrode cell using a saturated calomel electrode
(SCE) and a platinum wire as reference and counter electrodes,
3
3
2
2
the other two precursors were kept at room temperature. The tem-
peratures in the deposition chamber were selected to be 160 and
2
0
◦
2
50 C. The deposition rates were about 0.048 nm/cycle and about
.1 nm/cycle for Ta O5 and Al O , respectively. The ALD reactor
respectively. The working electrode area was 0.44 cm . All the
experiments were performed at room temperature in a 0.8 M NaCl
solution prepared with ultra pure Millipore® water (resistivity
>18 Mꢀ cm) and reagent grade chemicals (NaCl Analar Norma-
2
2
3
was operated with a constant nitrogen flow at a pressure of about
mbar. The nitrogen carrier and purge gas flow was 600 sccm with
5
−5
pur analytical reagent VWR® BDH Prolabo ). The electrolyte was
deaerated by bubbling argon gas for 30 min prior to the measure-
ments. The pH was 7.2 ± 0.2 before the corrosion tests. An Autolab
PGSTAT30 was used.
®
impurity content below 10 (impurity flow below 0.006 sccm).
The thickness of the coatings was measured from the sur-
face of a silicon wafer coated simultaneously with the steel
substrates. The thickness of the thicker coatings (≥50 nm) was
measured with an UV-VIS spectrophotometer (Hitachi U-2000)
and simulated using the Ylilammi and Ranta-aho [12] software.
The thickness of the thinner samples was measured with X-ray
The samples were cleaned with ethanol in an ultrasonic bath
for 10 min and dried with compressed air before the electrochem-
ical measurements. Before starting the DC and AC electrochemical
Table 2
Passive current density (jpass) and porosity (PLSV) values for the uncoated and coated samples calculated from the linear scan voltammetry measurements in 0.8 M Naclaq
according to Eq. (1)). The EIS fitting parameters obtained using the equivalent circuit presented in Fig. 6 are also compiled. The capacitance values, Eq. (4), are shown instead
of the CPE values obtained by fitting. The Ccoat values obtained after subtracting the Cnative are also given.
(
Sample
LSV
EIS
jpass (A cm 2
−
)
PLSV (%)
Re (ꢀ)
Rc (ꢀ)
Cc (F)
N
Ccoat (F)
Un-coated
Al2O3 160
8.6E−07
–
22.2
7.8E+06
6.3E−06
0.95
–
◦
C
C
5 nm
3.0E−07
2.7E−07
4.8E−09
4.2E−09
35.48
31.45
0.56
21.1
18.3
20.6
23.5
7.2E+06
1.1E+07
1.4E+08
3.6E+08
3.8E−07
2.6E−07
1.1E−07
8.9E−08
0.92
0.91
0.97
0.97
4.1E−07
2.7E−07
1.1E−07
8.1E−08
1
2
5
0 nm
0 nm
0 nm
0.49
◦
Al2O3 250
5 nm
1.3E−07
15
NA
0.08
0.03
31.2
22.9
22.2
21.5
1.4E+07
1.7E+09
2.8E+09
6.9E+09
7.4E−07
3.3E−07
1.7E−07
7.0E−08
0.95
0.99
0.99
0.99
8.4E−07
3.5E−07
1.8E−07
7.0E−08
a
a
1
2
5
0 nm
0 nm
0 nm
NA
6.9E−10
2.4E−10
◦
Ta2O5 250
C
5 nm
1.2E−07
2.2E−07
2.0E−07
1.2E−09
13.80
26.05
23.01
0.14
23.2
31.6
NA
3.8E+06
4.2E+06
NA
2.4E−06
3.2E−06
NA
1.6E−07
0.92
0.88
NA
3.9E−06
6.4E−06
NA
1.7E−07
1
2
5
0 nm
0 nm
0 nm
20.8
8.2E+07
0.98
a
NA – procedure not applicable.

1
0518
B. Díaz et al. / Electrochimica Acta 56 (2011) 10516–10523
experiments, the open circuit potential (OCP) was measured
in order to reach a stable potential. Then, the electrochemical
impedance spectroscopy (EIS) measurements were performed at
OCP, in a frequency range between 10 and 10 Hz, with the ampli-
tude of the exciting signal set at 10 mV. The linear scan voltammetry
A
coating
AlO2
interface
5
2
0 nm Al O3
substrate
Ni
5
−2
250ºC
4
3
2
1
0
10
10
10
10
10
Al
(
LSV) tests were performed after the EIS measurements with a scan
−1
OH
rate of 1 mV s from −0.9 V up to 0.5 V.
O
CrO2
FeO2
2.5. Porosity evaluation
C
The sealing performance of the coating was quantified by eval-
Cr
Fe
uating the surface fraction of the uncoated substrate exposed to
the electrolyte. It is hereafter defined as the coating porosity.
Defects, pores and/or pinholes, present in the coating are assumed
to connect the bulk electrolyte to the substrate surface where
the corrosion reactions proceed. Several electrochemical methods
based on different estimations have been previously proposed to
evaluate the coating porosity. They are based on a comparison of
the polarization resistances obtained from impedance data [13,14]
or from potentiodynamic tests [1,2,4] of uncoated and coated sam-
ples. A comparison of the current density values for a fixed potential
value has also been proposed by Landolt et al. for passive systems
AlC
NiO2
400
0
100
200
300
500
Sputtering time / s
[
15].
In this study, the evaluation of porosity was based on the linear
coating
B
5
0 nm Al O3
2
interface
160ºC
scan voltammetry measurements. For the passive systems pre-
sented in this work, the porosity is calculated as a ratio of the
passive current density ipass for the coated sample to the passive
AlO2
substrate
4
10
10
10
Al
O
0
current density i
of the uncoated sample (Eq. (1)).
pass
OH
3
2
1
0
ipass
P =
× 100%
(1)
0
i
Ni
pass
C
The passive current was always taken at a potential 500 mV more
anodic than the corrosion potential.
CrO2
FeO2
AlC
Fe
Cr
1
1
0
0
3
. Results and discussion
NiO2
3
.1. Time-of-flight secondary ion mass spectrometry
Fig. 1 shows the ToF-SIMS depth profiles for the 50 nm thick
0
100
200
300
400
500
◦
Al O coatings prepared at 250 and 160 C. The distribution of ions
2
3
Sputtering time / s
+
is plotted versus the sputtering time with Cs ions. The intensity is
plotted in logarithmic scale in order to emphasize the low intensity
signals.
Fig. 1. ToF-SIMS negative ion depth profiles for the 50 nm Al2O3 coatings prepared
◦
◦
at (A) 250 C and (B) 160 C.
Three regions can be identified: (i) the coating, (ii) the coat-
ing/substrate interface, and (iii) the alloy substrate. The coating
−
−
−
region is characterized by a plateau of the O , Al and AlO2
ion profiles. At about 330 s, the intensity of the Fe, Cr and Ni-
containing ions starts to increase marking the beginning of the
coating/substrate interfacial region in which the intensity of the Al-
containing ions drops. The intensity of the Fe, Cr and Ni-containing
ions becomes stable after approximately 450 s of sputtering, indi-
cating the end of the interfacial region and entry in the bulk
substrate. For both samples the width of the interfacial region is
OH profile
5
0 nm Al O3
2
10
160ºC
1
20–130 s. This similar width is expected for coatings with similar
morphology, both being amorphous and conformal, and deposited
on substrates having a similar roughness. The substrate roughness,
by causing shadowing effects of sputtering, enlarges the width of
the interfacial region.
50 nm Al O3
250ºC
2
−
−
−
In the coating region, the intensity of the O , Al and AlO2 ion
profiles recorded with the same primary ion beam current is the
same for the two samples indicating a similar stoichiometry of the
10
0
10
20
30
40
50
◦
−
alumina coatings prepared at 250 and 160 C. However, the OH
Sputtering time / s
ion profiles at the extreme and near surface region, presented in
magnification in Fig. 2, show a higher OH contamination for the
coating prepared at lower deposition temperature. The OH content
−
Fig. 2. ToF-SIMS OH ion depth profiles of the extreme and near surface regions for
the 50 nm Al2O3 coatings.

B. Díaz et al. / Electrochimica Acta 56 (2011) 10516–10523
10519
in the bulk of the coating is higher, by a factor of 2, in the coating
coating
◦
substrate
deposited at 160 C. This is an expected result as it is known that
interface
◦
below about 200 C the impurity contents in ALD oxide films start
1
1
1
0
0
0
OH
TaO2
50 nm Ta O5
to increase quite strongly with decreasing temperature because of
less complete reactions [9,16].
2
250ºC
Fig. 2 also presents the OH contamination at the extreme surface
C
Ni
−
of the coating. The sputtering time needed to reach a stable OH ion
O
intensity (i.e. the bulk of the coating) is longer after deposition at
◦
◦
1
60 C (around 20 s) than after deposition at 250 C (around 12 s).
TaC
The presence of OH-containing species at the extreme surface is
related to the exposure of the sample to ambient air that causes sur-
face hydroxylation and hydration of the alumina film. The longer
time needed to reach the OH bulk level in the aluminium oxide
layer deposited at lower temperature shows that surface hydroxy-
lation/hydration is higher and that the OH ions penetrate deeper in
Cr
CrO2
FeO2
Fe
10
NiO2
150
the sub-surface region of this Al O layer. This difference observed
2
3
1
0
at the coating surface can be related to the coating porosity. Indeed,
with increasing porosity, one can expect that more defects (nano-
or micro-channels) are present in the coating and thus promote sur-
face hydroxylation/hydration and the penetration of OH ions when
exposed to ambient conditions. As a result of this analysis, a higher
0
50
100
200
250
300
350
Sputtering time / s
Fig. 3. ToF-SIMS negative ion depth profile for the 50 nm Ta2O5 coating prepared at
◦
◦
250 C.
porosity is expected for the Al O coating prepared at 160 C.
2
3
−
−
Carbon contamination is also shown by the C and AlC ion pro-
files. It is observed at the extreme surface but decreases after a few
seconds of sputtering, thereby implying its origin being hydrocar-
bon contamination from ambient. The C and AlC ion profiles in
the bulk coating indicate the presence of some unreacted fragments
of the metal organic precursor Al(CH3)3 trapped inside the coat-
ing. This carbon contamination decreases with the deposition time
and is reduced at higher deposition temperature, showing that the
increase of temperature promotes elimination of residues of both
precursors used for deposition of the Al O coatings.
coating/substrate region is contaminated by a surface native oxide
and carbon species.
−
−
3.2. Linear scan voltammetry
Fig. 4 shows the potentiodynamic curves for the bare and
coated samples. For the bare 316L alloy, a peak assigned to the
active/passive transition is observed at around −0.2 V, in very
good agreement with previous measurements [17]. Its presence is
assigned to the reduction (most likely partial) of the native oxide
occurring during cathodic polarization. For all the coated samples,
this reaction is hindered. At the pitting potential of 0.4 V, the sharp
increase in the current marks the initiation and growth of pits for
2
3
A higher amount of carbon contamination is observed in the
interfacial region. This is assigned to the organic contamination
present on the substrate surface before the coating deposition.
Likewise, the OH contamination increases in the interfacial region
indicating hydroxylation of the stainless steel surface prior to depo-
sition.
the uncoated sample. For the samples covered by the thinner (5 and
◦
10 nm) Al O coatings deposited at 250 C (Fig. 4A), current fluctu-
2
3
The Fe, Cr and Ni-containing ion profiles all display a peak in
the interfacial region assigned to the presence of a native oxide on
the substrate surface before deposition. The growth of a spurious
thermal oxide in the initial stages of the deposition, prior to
complete sealing of the substrate surface can also contribute to the
ations related to metastable pitting were observed in the passive
range and preceded stable pitting. However, for the thicker (20 and
50 nm) coatings deposited at 250 C, no metastable nor stable pit-
◦
ting could be observed indicating efficient protection and blocking
of localized corrosion in the explored range of potential. For the
−
−
◦
interfacial oxide. The slightly smaller width of the FeO2 , CrO2
Al O coatings prepared at 160 C (Fig. 4B), metastable pitting and
2
3
−
and NiO2 peaks observed for the sample prepared at the higher
deposition temperature suggests the formation of a thinner oxide,
and thus faster sealing of the substrate surface at higher deposition
temperature. The deposition temperature does not markedly
modify the composition of the interfacial region as indicated by
stable pitting were also observed for the 5 and 10 nm thick films,
indicating a less efficient corrosion protection. Blocking of localized
corrosion was achieved only with the 20 nm thick coating.
For both deposition temperatures, the curves in Fig. 4 show
the decrease of the anodic current density with increasing coat-
ing thickness, showing that the coating not only blocks localized
corrosion but also markedly decreases the corrosion rate in the
passive state. Table 2 compiles the passive current density val-
ues as well as the porosity values calculated from Eq. (1). Due to
−
−
−
the relative intensities of the FeO2 , CrO2 and NiO2 peaks.
Fig. 3 presents the ToF-SIMS depth profile for the 50 nm thick
Ta O5 coating. The profile intensities and width cannot be directly
2
compared to those obtained for the Al O3 coating of the same
2
thickness since the oxide matrices are different, which affects the
sputter yields of all ions. However, marked differences of intensi-
ties (larger than a factor of 10) are observed for the C and OH
the noise observed during the measurement of the 10 nm Al O3
2
film (Fig. 4A), passive current density and porosity values could not
be extracted for this sample. As expected, at both deposition tem-
peratures, the lowest porosity values are obtained for the thickest
coatings prepared. The results presented in Table 2 show a sig-
nificant improvement of the coating sealing performance (lower
porosity) also with the increase of the deposition temperature. A
better sealing performance (porosity of 0.03%) is obtained with the
−
−
ion profiles in the bulk coating region compared to those of the
◦
Al O coating prepared at the same temperature (50 nm, 250 C).
2
3
It has been previously measured by elastic recoil detection anal-
ysis that the carbon and hydrogen contents were <0.2 and 1 at.%,
◦
respectively, in the Al O3 film deposited at 250 C from Al(CH )
2
3
3
◦
and water, and 0.3–1.1 and 0.6–1.5 at.%, respectively, in the Ta O5
50 nm thick coating prepared at 250 C than with the 50 nm thick
2
◦
◦
coating deposited at 300 C from Ta(OC H5)5 and water [16]. This
coating prepared at 160 C (porosity of 0.5%). The porosity order
2
indicates that higher amounts of unreacted fragments of the pre-
cursors are retained inside the oxide matrix during the growth of
the Ta O5 coating. Similarly to the Al O coating, the interfacial
obtained for the two 50 nm thick coatings prepared at 250 and
160 C is in agreement with the ToF-SIMS data that show a deeper
penetration of hydroxide anions assigned to a more defective struc-
◦
2
2
3

1
0520
B. Díaz et al. / Electrochimica Acta 56 (2011) 10516–10523
Fig. 4C shows the potentiodynamic curves obtained with the
bare substrate 316L
Al O 5nm 250ºC
Al O 10nm 250ºC
Al O 20nm 250ºC
Al O 50nm 250ºC
Ta O5 coatings. A significant improvement of corrosion resis-
A
2
1E-3
1E-4
1E-5
1E-6
1E-7
1E-8
1E-9
tance is observed only with the 50 nm thick coating that blocks
metastable and stable pitting and markedly reduces the corrosion
rate in the passive state. For the thinner 20 nm coating, metastable
pitting is again observed in the anodic range but no stable pitting
occurs in the tested potential range. This fact suggests improved
repassivation and higher resistance to stable pitting with the Ta O5
2
layers than with the Al O3 layers. Still the higher current denotes
2
a lower sealing performance than the 20 nm thick Al O coating
2
3
◦
deposited at 250 C. Table 2 allows quantitative comparison of the
coating sealing performance, confirming a drastic difference (fac-
tor of ∼300) between the 20 nm thick Ta O5 and Al O coatings
2
2
3
◦
prepared at 250 C. For the 5 nm thick coatings, this difference is
reduced to a factor of 5. Again the comparison with the ToF-SIMS
data points to a key role of the C-containing precursor fragments
1
E-10
E-11
1
contaminating the coating and in higher concentration in the Ta O5
2
-1.2
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6
than in the Al O coating as shown above. The decreasing difference
2
3
E vs SCE / V
of the sealing performance between the Ta2O5 and Al2O3 coat-
ings with increasing thickness is then explained by the decreasing
contamination by precursor fragments, as shown by the ToF-SIMS
depth profile analysis. It may also indicate lower nucleation density,
and thereby later closure, of Ta O5 compared to Al O .
1
1
1
1
1
1
1
E-3
bare substrate 316L
Al O 5nm 160ºC
Al O 10nm 160ºC
Al O 20nm 160ºC
Al O 50nm 160ºC
B
2
2
3
E-4
E-5
E-6
E-7
E-8
E-9
3.3. Electrochemical impedance spectroscopy
EIS has been widely used in the last two decades to character-
ize the corrosion behavior of coated metallic surfaces, and several
attempts to characterize the coatings quality by EIS have been made
[
18–21]. Fitting with an adequate equivalent electric circuit is nec-
essary for the interpretation of the impedance spectra.
Fig. 5A and B shows the impedance data for the Al O coatings
2
3
presented as Bode plots (impedance module and phase angle vs.
log(freq)). Only one time constant can be distinguished in the spec-
tra for the uncoated as well as for the coated samples. The phase
1E-10
◦
angle is close to 90 which shows a strong capacitive response
either from the native oxide present on the uncoated samples or
from the coated samples [22]. The capacitive response resulting
from the native oxide can be considered as constant for all analyzed
samples if we assume no thickness variation of the native oxide
during coating. However, the data show for both deposition tem-
peratures an increasing capacitive behavior at higher frequencies
as the thickness of the coating increases, which is directly related to
the dielectric properties of the Al O coating. As the coating thick-
-1.2
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6
E vs SCE / V
bare substrate 316L
Ta O 5nm 250ºC
C
1E-3
2
5
Ta O 10nm 250ºC
1E-4
1E-5
1E-6
1E-7
1E-8
1E-9
2
5
Ta O 20nm 250ºC
2
3
2
5
ness increases, the corresponding dielectric capacitance decreases
following the description for a parallel plate capacitor (Eq. (2)):
Ta O 50nm 250ºC
2
5
εε0S
C =
(2)
d
2
where S is the surface area covered by the coating (cm ), d the
coating thickness (cm), ε the dielectric constant for the mate-
rial inside the capacitor (ε = 9 for Al2O3 [23] and ε = 25 for Ta2O5
[
(
24]), and ε0 corresponds to the vacuum dielectric constant
1
E-10
E-11
−14
−1
F cm ). So the time constant, ꢁ, related to this capac-
8.85 × 10
−
1
itance and defined as (RC) where R is the coating resistance will
be shifted to higher frequencies with increasing coating thickness.
An equivalent circuit with two time constants representing the
coating and the substrate is frequently employed in most of the
papers when characterizing the quality of the coating [1,4,19,25].
However, without a clear indication of the existence of two time
constants as shown by the Bode plots presented in Fig. 5, it is not
possible to clearly distinguish the coating and the substrate [18,22].
The equivalent circuit shown in Fig. 6A is presented as the most
reasonable for the fit of the EIS data, either for the uncoated or the
coated samples. Native oxide layers formed on the stainless steel
surface are highly stable [26–28] with high capacitive responses
1
-
1.2
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6
E vs SCE / V
Fig. 4. Polarization curves of the uncoated and Al2O3 coated samples prepared at
◦ ◦ ◦
A) 250 C and (B) 160 C and for the Ta2O5 coated samples prepared at (C) 250 C.
(
ture of the lower temperature coating. The comparison with the
ToF-SIMS data also suggests that the more defective structure of
the coating prepared at lower temperature may be related to the
higher amount of precursor fragments contaminating the coating.

B. Díaz et al. / Electrochimica Acta 56 (2011) 10516–10523
10521
A
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
90
80
70
60
50
40
30
bare 316L
S1H5 Al O 5nm 250ºC
20
S1H6 Al O 10nm 250ºC
S1H7 Al O 15nm 250ºC
S1H8 Al O 50nm 250ºC
10
0
1E-3
0.01
0.1
1
10
100
1000 10000 1000001000000
Frequency / Hz
B
1
1
1
1
1
1
1
1
0
9
7
6
0
5
0
0
0
0
0
0
0
0
Fig. 6. (A) Equivalent electric model employed for fitting the impedance spectra (see
text for details). (B) Diagram describing the ALD coating on the metallic substrate.
priate and the information concerning the coating as a dielectric
material can be extracted [23,32].
For the proposed electric model, Re, the limit at high frequency,
describes the electrolyte resistance, Rc refers to the electric resis-
tance of the layer (oxide coating + native oxide on the substrate) and
the Cc-CPE corresponds to the capacitance not only of the dielec-
tric film deposited on the sample but also of the native oxide on
the substrate [22]. For the thicker oxide coatings with lower dielec-
tric capacitance, the contribution of the native oxide, with higher
dielectric capacitance due to the lower thickness, can be neglected
since both capacitances are in series. In this case, only the lower
capacitance will be considered.
45
bare substrate 316L
Al O 5nm 160ºC
Al O 10nm 160ºC
Al O 20nm 160ºC
Al O 50nm 160ºC
3
1
0
0
5
1E-3
0.01
0.1
1
10
100
1000 10000 1000001000000
Frequency / Hz
Constant phase elements (CPEs) are commonly used in data fit-
ting to describe the non-ideal capacitive behavior of the system.
The CPE has no clear physical meaning but it allows for an evalua-
tion of the complexity of the studied system. The impedance for a
CPE is defined in Eq. (3) [33]:
C
10
10
10
10
10
10
10
9
0
80
70
60
50
40
30
1
ZCPE
=
(3)
n
Q(jω)
where Q is a constant value, independent of the frequency, obtained
directly from the fitting, and the factor n is the CPE power, which
is related to the angle ((1 − n) × 90) that evaluates the Nyquist
plot depression below the x-axis. The factor n is related to the
surface heterogeneity such as irregularity, roughness or porosity
[21,34]. For n = 1 the CPE describes an ideal capacitor and for n close
to 0.5 the CPE represents a Warburg impedance with a diffusion
character. The smoother and cleaner the surface, the closer is the
parameter n to unity [33,35]. The Q and n parameters are obtained
from the fitting and the real equivalent capacitance value can be
easily calculated from the Brug equation [36] (Eq. (4)).
bare substrate 316L
Ta O 5nm 250ºC
Ta O 10nm 250ºC
Ta O 20nm 250ºC
Ta O 50nm 250ºC
20
10
0
-
10
1
E-3
0.01
0.1
1
10
100
1000 10000 1000001000000
Frequency / Hz
◦
Fig. 5. Bode plots for the uncoated and Al2O3 coated samples prepared at (A) 250
and (B) 160 C and for the Ta2O5 coated samples prepared at (C) 250 C.
C
◦
◦
C = Q1/nRe(1−n)/n
(4)
Some other equations have been proposed to obtain the real
capacitance from the Q value [37] but Huang et al. [38] have proven
that the most realistic results were obtained following the Brug
formula [36].
due to their low thicknesses (around 2 nm [29–31]). They are con-
sidered as inert in our EIS measurements performed at OCP without
cathodic pre-treatment. Thus, no elements have been introduced in
the equivalent circuit to account for active corrosion. Moreover the
analyzed frequencies (ω) were not low enough to get information
about the corrosion process and only information about the dielec-
tric behavior of the coating can be obtained [22]. Therefore the
proposed equivalent circuit presented in Fig. 6 is the most appro-
The parameters of the EIS data fit for the Al O coated samples
2
3
are compiled in Table 2. The capacitance values for the coatings
(Cc following Eq. (4)) are presented instead of the CPE-Q obtained
directly from the fitting. The good correlation between the theo-
retical values of the capacitance for the oxide coatings (calculated
from Eq. (2), and using the coating coverage, 1 − P, deduced from the

1
0522
B. Díaz et al. / Electrochimica Acta 56 (2011) 10516–10523
than for the Al O3 coatings having the same thickness and the
same deposition temperature. For the Ta2O5 layers, only the thicker
2
A
Ccoat-fitting
Ccoat-theoretical (Eq. 3)
coating (50 nm) shows a marked improvement of the corrosion
resistance. A phase angle lower than 90 , indicating a less capacitive
◦
2
Al O3
behavior and thus a more corroding system [4,19], is observed for
the 20 nm coating. This shows an even higher degradation for the
coated sample than for the bare substrate after the immersion in the
aggressive solution. Actually the equivalent circuit proposed above
has no physical meaning and it cannot be employed in this particu-
lar case. For the other samples, the fitting parameters are presented
in Table 2. Fig. 7B shows the comparison of the expected theoretical
capacitance values (calculated from Eq. (2)) and the values obtained
by fitting (after correction following the Brug equation, Eq. (4), and
removing the contribution of the native oxide capacitance) for the
10
10
10
Ta O5 coatings. A significant data dispersion can be observed com-
2
pared to the Al O coatings. A good agreement is obtained only for
2
3
uncoated 5 nm 10 nm 20 nm 50 nm 5 nm 10 nm 20 nm 50 nm
the thicker Ta O5 layer (50 nm). The poor and inconsistent results
2
2
50ºC
160ºC
obtained with the thinner Ta O5 layers imply some problems in
2
nucleation of Ta O5 and thereby point to a need to improve the sur-
2
face preparation. Whatever the reason for the nucleation problems,
B
the Al O3 process remains clearly much less affected. Indeed, the
2
Ta O5
2
Al(CH ) –H O process is known as the best nucleating ALD process
3
3
2
250ºC
and has been successfully exploited in growing high quality Al O3
2
films also on polymer substrates [40].
1
0
0
0
The lowest values of the CPE power, n, obtained from the data
fit for the Ta O5 films are consistent with the large scatter between
2
the expected and the fitted capacitance values. In particular for the
1
0 nm Ta O5 coating, for which the lowest corrosion resistance was
2
obtained, the n parameter shows the highest deviation from the
ideal (pure capacitive) behavior (n < 0.9) among all samples, which
corresponds to a higher heterogeneity of the coating. Comparing
all samples, the highest n values, close to 1, corresponding to the
coatings having the lowest porosity values, were obtained with the
1
1
Ccoat-fitting
Ccoat-theoretical (Eq. 2)
◦
Al O films deposited at 250 C. The better electrochemical perfor-
2
3
mance obtained with these samples indicates less defects and/or
pinholes relevant for corrosion. Moreover the parameter n close to
unity shows that the surface is smoother and freer of defects, which
corresponds to an ideal, higher corrosion resistant system. Liu et al.
[19] have reported that a decrease of n value with an increase of
immersion time of the coated sample was related to a lower capac-
itive behavior since the coating becomes more conductive due to
the presence of a higher number of holes. Thus, and as confirmed by
our data, the n parameter can be used as an indicator of the coating
quality and its performance.
uncoated
5 nm
10 nm
50 nm
Fig. 7. Comparison of the experimental dielectric capacitance (obtained from the
impedance simulation shown in Figs. 4 and 5 and removing the contribution of the
native oxide capacitance) with theoretical capacitance (calculated from Eq. (2)) for
(
A) the Al2O3 coatings and (B) the Ta2O5 coatings.
porosity values obtained from the LSV data) and the values obtained
by fitting (Ccoat, where the contribution of the native oxide was
neglected) justifies the selection of the equivalent circuit employed
for the simulation. Fig. 7A shows the good agreement between the
4. Conclusions
expected capacitances and those obtained from fitting for the Al O3
2
coatings.
Surface analysis by ToF-SIMS and electrochemical analysis in a
0.8 M NaCl solution by LSV and EIS of ultra thin (5–50 nm) oxide
Although no quantitative analysis of the coatings in terms of
defects can be performed from the impedance data, some qual-
itative information can be extracted from the comparison of the
resistance values. A higher resistance points to the growth of a less
defective coating. Thus, considering the coating resistance Rc as
the electric resistance through the existing pores and pinholes (as
described in Fig. 6B), an increase of Rc, after taking into account the
different coating thickness, is associated to a reduction in either
◦
(Al O and Ta O5) coatings deposited by ALD at 250 and 160 C on
2
3
2
stainless steel (316L) have been performed.
ToF-SIMS depth profiling allows the elemental characterization
of the bulk coating, interfacial and substrate regions. The bulk coat-
ings are contaminated by traces of C and OH species originating
from unreacted fragments of the precursors used for deposition and
trapped in the films. A higher level of C contamination is observed
in the Ta O5 than in Al O owing to more complete reactions
◦
the size and/or density of these defects. Both at 160 and at 250 C,
2
2
3
the higher resistances obtained for the thicker Al O3 coatings are
in the Al O3 deposition. Increasing the deposition temperature
2
2
in agreement with the lower porosities obtained from the LSV data.
Moreover a better sealing (less defective) coating is confirmed by
the higher resistance obtained after growth at higher temperature.
The growth of a denser and more compact coating by ALD at higher
deposition temperature agrees with previous data [9,39].
allows to decrease the C and OH contamination of the bulk coat-
ing. The increase of C and OH contamination at the approach of
the interfacial region points to an influence of the substrate surface
contamination in the first deposited layers. The interfacial region
contains also a native oxide grown on the stainless steel surface,
possibly modified by thermal oxidation prior to complete sealing
by the coating.
The Bode plots corresponding to the Ta O5 coatings are pre-
2
sented in Fig. 5C. The impedance results show a worse performance

B. Díaz et al. / Electrochimica Acta 56 (2011) 10516–10523
10523
The potentiodynamic polarization data show that the deposited
coatings can block pitting of the 316L substrate in the tested range
of potential (up to 0.5 V/SCE) in 0.8 M NaCl at room temperature.
[3] K.L. Chang, S. Han, J.H. Lin, J.W. Hus, H.C. Shih, Surf. Coat. Technol. 172 (2003)
2.
4] V.K.W. Grips, V.E. Selvi, H.C. Barshilia, K.S. Rajam, Electrochim. Acta 51 (2006)
461.
7
[
3
This is achieved with the 50 nm thick films deposited at 250 and
[5] D. Turcio-Ortega, S.E. Rodil, S. Muhl, Mater. Sci. 14 (2008) 15.
[6] P. Bardy, K. Weisbrod, I. Paulauskas, R.A. Buchanan, K.L. More, H. Wang, M.
Wilson, F. Garzon, L.R. Walker, Scripta Mater. 50 (2004) 1017.
◦
1
60 C. A drastic decrease of the current density is measured both
in the cathodic and in the anodic branches. The current reduction
becomes more pronounced with increasing coating thickness and
reaches four orders of magnitude in the best case (50 nm Al O
[
7] R. Matero, M. Ritala, M. Leskela, T. Salo, J. Aromaa, O. Forsen, J. Phys. IV Fr. 9
(1999) 493.
[8] B.S. Lim, A. Rahtu, R.G. Gordon, Nat. Mater. 2 (November) (2003) 749.
[9] M. Ritala, M. Leskelä, in: H.S. Nalwa (Ed.), Handbook of Thin Film Materials, vol.
2
3
◦
deposited at 250 C). The sealing performance of the coating was
assessed from the drop of the passive current and quantified as
the coating porosity, equivalent to the uncoated surface fraction.
The porosity decreases with the thickness for all coatings, a better
1, Academic Press, San Diego, 2001, p. 103 (Chapter 2).
[10] M. Ritala, J. Niinistö, ECS Trans. 25 (2009) 641.
[11] C.X. Shan, X. Hou, K.L. Choy, Surf. Coat. Technol. 202 (2008) 2399.
12] M. Ylilammi, T. Ranta-aho, Thin Solid Films 232 (1993) 56.
13] B. Elsener, A. Rota, H. Bohni, Mater. Sci. Forum 44–45 (1989) 29.
[14] C. Liu, Q. Bi, A. Leyland, A. Matthews, Corros. Sci. 45 (2003) 1257.
[15] W. Tato, D. Landolt, J. Electrochem. Soc. 145 (1998) 4173.
[
[
sealing performance being obtained with Al O than with Ta O5
2
3
2
◦
and at a deposition temperature of 250 C. The lowest porosity value
of 0.03% was achieved in this study for a 50 nm Al O film.
[
16] R. Matero, A. Rahtu, M. Ritala, M. Leskelä, T. Sajavaara, Thin Solid Films 368
2000) 1.
17] I. Olefjord, B. Brox, U. Jelvestam, J. Electrochem. Soc. 132 (1985) 2854.
2
3
(
The EIS Bode plots show one time constant for the coated and
uncoated systems. For the Al O coatings, the experimental values
[
[18] S. Rudenja, C. Leygraf, J. Pan, P. Kulu, E. Talimets, V. Mikli, Surf. Coat. Technol.
14 (1999) 129.
2
3
1
of the coating capacitance obtained from the fit of the EIS data show
a good agreement with values calculated assuming a flat parallel
plate capacitor. The n parameter of the constant phase element
used for the evaluation of the capacitive behavior of the coating
efficiently allows for a qualitative ranking of the coatings in terms
of presence of pores and defects, in line with the porosity data.
The ranking of the coating resistances is also in agreement with
the classification obtained from the potentiodynamic polarization
measurements.
[
[
19] C. Liu, Q. Bi, A. Matthews, Corros. Sci. 43 (2001) 1953.
20] J. Zhang, V. Desai, Surf. Coat. Technol. 190 (2005) 98.
[21] Y.-Y. Chang, D.-Y. Wang, Surf. Coat. Technol. 200 (2005) 2187.
22] C. Liu, Q. Bi, A. Leyland, A. Matthews, Corros. Sci. 45 (2003) 1243.
23] A.I. de Sá, C.M. Rangel, Q. Lu, P. Skelton, G.E. Thompson, Corros. Sci. 48 (2006)
[
[
2203.
[24] K. Kukli, M. Ritala, M. Leskelä, J. Electrochem. Soc. 142 (1995) 1670.
[25] R.C. Barik, J.A. Wharton, R.J.K. Wood, K.R. Stokes, R.L. Jones, Surf. Coat. Technol.
199 (2005) 158.
[
26] I. Olefjord, B. Brox, in: M. Froment (Ed.), Passivity of Metals and Semiconductors,
Elsevier, Amsterdam, 1983, p. 561.
The correlation between the porosity values extracted from the
CV data and their ranking confirmed by EIS on one hand and the
level of C and OH traces observed by ToF-SIMS on the other hand
points to a marked effect of the coating contaminants on the seal-
ing performance and on the corrosion resistance of the coated
systems.
[27] P.A. Schweitzer, Fundamentals of Metallic Corrosion: Atmospheric and Media
Corrosion of Metals, vol. 4, 2nd ed., CRC Press, 2006, p. 110.
[
28] R. Baboian, Corrosion Tests and Standards: Application and Interpretation
ASTM Manual Series MNL 20, 2nd ed., ASTM International, West Conshohocken,
PA, USA, 2005, p. 133.
[29] V. Maurice, W. Yang, P. Marcus, J. Electrochem. Soc. 143 (1996) 1182.
[30] V. Maurice, W. Yang, P. Marcus, J. Electrochem. Soc. 145 (1998) 909.
[31] C.M. Abreu, M.J. Cristóbal, R. Losada, X.R. Nóvoa, G. Pena, M.C. Pérez, J. Elec-
troanal. Chem. 572 (2004) 335.
[
[
32] G.E. Cavigliasso, M.J. Esplandiu, V.A. Macagno, J. Appl. Electrochem. 28 (1998)
Acknowledgements
1213.
33] E. Barsoukov, J.R. Macdonald, Impedance Spectroscopy, Theory, Experimental
and Applications, 2nd ed., John Wiley & Sons Inc., 2005, p. 495.
34] E.M.A. Martini, I.L. Muller, Corros. Sci. 42 (2000) 443.
The research leading to these results has received funding
from the European Community’s Seventh Framework Programme
[
[
35] G.E. Cavigliaso, M.J. Esplandiu, V.A. Macagno, J. Appl. Electrochem. 28 (1998)
◦
(
FP7/2007-2013) under grant agreement n CP-FP 213996-1 (COR-
1213.
RAL). Region Ile-de-France is acknowledged for partial support for
the ToF–SIMS equipment.
[36] G.J. Brug, A.L.G. van den Eeden, M. Sluyters-Rehbach, J.H. Sluyters, J. Electroanal.
Chem. 176 (1984) 275.
[
[
37] C.H. Hsu, F. Mansfeld, Corrosion 57 (2001) 747.
38] V.M.-W. Huang, V. Vivier, M.E. Orazem, N. Pébère, B. Tribollet, J. Electrochem.
Soc. 154 (2007) C99.
References
[
39] M.D. Groner, F.H. Fabreguette, J.W. Elam, S.M. George, Chem. Mater. 16 (4)
(
2004) 639.
[
[
1] S.H. Ahn, Y.S. Choi, J.G. Kim, J.G. Han, Surf. Coat. Technol. 150 (2002) 319.
2] S.H. Ahn, J.H. Yoo, Y.S. Choi, J.G. Kim, J.G. Han, Surf. Coat. Technol. 162 (2003)
[
40] C.A. Wilson, R.K. Grubbs, S.M. George, Chem. Mater. 17 (2005) 5625.
2
12.