Study of the effect of cerium nitrate on AA2024-T3 by means of electrochemical micro-cell technique

Article abstract of DOI:10.1016/j.electacta.2012.02.099

This work evaluates the effect of cerium nitrate as corrosion inhibitor for AA2024-T3 in the view of its introduction in sol-gel coatings able to provide self-healing ability. Since it is well established that deposition of Ce species is activated by the local pH increase, the objective of this paper is to investigate the behavior of AA2024-T3 (open circuit potential and polarization curves) in the presence of Ce species in aggressive solutions by means of a local technique, the electrochemical micro-cell. This technique enables the investigation of small areas with resolution in the micrometer range by the use of glass capillaries to define the working electrode area. The micro-cell results clearly displayed that the entire AA2024-T3 area exposed to the cerium-containing electrolyte was involved in the cerium precipitation mechanism. The heterogeneous electrochemical behavior of the microstructure is minimized by the formation of a cerium-containing layer able to protect the metal substrate.

Full text of DOI:10.1016/j.electacta.2012.02.099

Electrochimica Acta 70 (2012) 25–33
Contents lists available at SciVerse ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Study of the effect of cerium nitrate on AA2024-T3 by means of electrochemical
micro-cell technique
L. Paussaa,∗, F. Andreattaa, N.C. Rosero Navarrob, A. Duránb, L. Fedrizzia
a University of Udine, Department of Chemistry, Physics, and Environment, Via del Cotonificio 108, 33100 Udine, Italy
b Instituto de Ceramica y Vidrio (CSIC), c/Kelsen 5, Campus de Cantoblanco, 28049 Madrid, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
This work evaluates the effect of cerium nitrate as corrosion inhibitor for AA2024-T3 in the view of
its introduction in sol–gel coatings able to provide self-healing ability. Since it is well established that
deposition of Ce species is activated by the local pH increase, the objective of this paper is to investigate
the behavior of AA2024-T3 (open circuit potential and polarization curves) in the presence of Ce species in
aggressive solutions by means of a local technique, the electrochemical micro-cell. This technique enables
the investigation of small areas with resolution in the micrometer range by the use of glass capillaries to
define the working electrode area.
The micro-cell results clearly displayed that the entire AA2024-T3 area exposed to the cerium-
containing electrolyte was involved in the cerium precipitation mechanism. The heterogeneous
electrochemical behavior of the microstructure is minimized by the formation of a cerium-containing
layer able to protect the metal substrate.
Received 22 September 2011
Received in revised form 28 February 2012
Accepted 28 February 2012
Available online 16 March 2012
Keywords:
Micro-electrochemical cell
AA2024-T3
Inhibition
Cerium
© 2012 Elsevier Ltd. All rights reserved.
1. Introduction
can take into account several critical aspects. Boag et al. [15] investi-
gated the local electrochemical behavior of AA2024-T3 by coupling
The heterogeneous microstructure of AA2024-T3 aluminum
alloy strongly affects its corrosion behavior [1–5]. Different kinds
of coarse intermetallic particles have been identified in AA2024-
T3 [6–8]. Buchheit reported a compilation of corrosion potentials
for aluminum-based intermetallic particles found in the litera-
ture [9]. Corrosion potentials for intermetallic particles containing
Al–Cu–Mg are more negative than those detected for intermetal-
lic particles containing Al–Cu–Fe–Mn. As for AA2024-T3 aluminum
alloy, Buchheit et al. [10] further highlighted that Al–Cu–Mg based
intermetallic particles are the most numerous category while
the second largest type is rich in Al–Cu–Fe–Mn. It has been
found that immediately after immersion in a chloride solution,
Al–Cu–Mg-based intermetallic particles interact as anodic sites
with the aluminum matrix. However, magnesium and aluminum
dissolution promotes the formation of Mg–Al-depleted intermetal-
lic particles which can act cathodically for longer immersion
times. Otherwise, Al–Cu–Fe–Mn intermetallic particles constantly
interact as cathodes with the aluminum matrix [11–13]. The elec-
trochemical behavior of second phase particles is affected by the pH
solution and it can become very complex [14]. Therefore, the study
of the electrochemical behavior of aluminum alloys is not easy and
SEM and EDXS analyses, which were used for mapping the alloy
surface at different immersion times. It has been found that inter-
metallic particles, anodic and cathodic with respect to the matrix,
play an important role in the corrosion behavior of the alloy but
they were not able to quantify their influence. Battocchi et al. [16]
used the scanning vibrating electrode technique (SVET) to mea-
sure the local current distribution of a wire beam electrode (WEB)
made of AA2024-T3 wires. It has been found that the local surface
conditions affected the heterogeneous electrochemical behavior of
the alloy. Seegmiller and Buttry [17] used the scanning electro-
chemical microscopy (SECM) to investigate the cathodic activity at
AA2024-T3 surfaces which turned out more intense on the inter-
metallic particles present in the microstructure rather than on the
aluminum matrix.
In this view, the electrochemical micro-cell technique can be
usefully employed for the evaluation of the micro-electrochemical
behavior of metal microstructures [18]. The electrochemical micro-
cell technique was initially employed by Suter et al. [19–21] to
study the influence of sulfide inclusions on pitting initiation in
stainless steels. Suter and Alkire [22], for the first time, employed
the micro-cell technique to evaluate the electrochemical behavior
of AA2024-T3 in the micrometer range. It has been found that the
onset of pitting corrosion depends on the intermetallic particles
category. In the case of Al–Cu–Mg-based intermetallic particles,
it was observed that a fast magnesium and aluminum dissolution
∗
Corresponding author. Tel.: +39 432 558831; fax: +39 432 558803.
E-mail address: luca.paussa@uniud.it (L. Paussa).
0013-4686/$ – see front matter © 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2012.02.099

26
L. Paussa et al. / Electrochimica Acta 70 (2012) 25–33
preferentially starts at the edge rather than at the center of inter-
metallic particles. In case of longer immersion time, the Al–Cu–Mg
intermetallic particle type can become cathodic with respect to the
surrounding matrix. On the other hand, the preferential dissolution
of the matrix was mainly detected at the edge of Al–Cu–Fe–Mn-
based intermetallic particles due to the galvanic coupling. The
micro-cell technique was also employed for the study of the micro-
electrochemical behavior of AA7075 aluminum alloy [23–25]. The
interaction of cathodic intermetallic particles (Al7Cu2Fe) with the
matrix was fully investigated [26]. Al7Cu2Fe-based intermetal-
lic particles promote the dissolution of the surrounding matrix
inducing the formation of trenching corrosion morphologies. Fur-
thermore, it was found that Mg2Si intermetallic particles are anodic
with respect to the matrix [27] and they represent the preferential
site for the pitting corrosion initiation.
Salts of rare-earth elements are able to improve the corrosion
behavior of aluminum alloys [28–30]. Hinton et al. [31–33] proved
that rare-earths, and particularly cerium, inhibit redox reactions
occurring on aluminum alloys. It was highlighted that the improve-
ment of the corrosion resistance is due to the precipitation of
cerium compounds on cathodic regions thus promoting the inhibi-
tion of reduction reactions [34]. Thanks to the first studies on the
mechanism of cerium compounds precipitation, it was assumed
that the process starts on regions where the pH solution tends
to be increased by reduction reactions which produce chemical
conditions favorable to cerium precipitation. It was hypothesized
that the process continues in the surrounding regions via island
growth involving small active regions which have not been covered
yet [35]. Yasakau et al. [36] proved that, in case of the AA2024-
T3 aluminum alloy, cerium precipitation preferentially occurs on
Al–Cu–Mg-based intermetallic particles modified by the local mag-
nesium and aluminum dissolution which leads to the overall
copper-enrichment of the intermetallic particles. Despite the fact
that many studies have focused on the comprehension of cerium
precipitation mechanisms, there are some aspects which still need
clarification such as the precipitation of cerium compounds on the
matrix. In this work, the electrochemical micro-cell approach was
used to evaluate the influence of the AA2024-T3 microstructure on
the corrosion inhibition provided by cerium nitrate.
In this work, micro capillaries of 50 ␮m size were employed. The
counter electrode was a 1.5 mm platinum wire while an Ag/AgCl
electrode was used as reference. The micro-electrochemical cell
employed in this work is of the form reported by Suter [18,19].
All electrodes were connected to a potentiostat/galvanostat (IPS
– Elektroniklabor Peter Schrems) able to measure currents in
the range of fA. Three different kinds of surface regions were
studied by means of micro-cell technique: regions where inter-
metallic particles are not visible (indicated in the paper as matrix),
regions containing a high amount of Al–Cu–Mg-based intermetallic
particles and regions containing a high number of Al–Cu–Fe–Mn-
based intermetallic particles. Open circuit potential measurements
(OCPs) were performed for 5 min in order to evaluate the sta-
bility of the region identified by the micro capillary mouth.
Potentiodynamic polarization curves were then acquired on the
same area exposed during OCP measurements. A scan rate of
1 mV s−1 was used to perform polarization curves. Anodic and
cathodic branches were acquired separately. Anodic branches were
detected starting near the OCP. Cathodic branches were rather
acquired starting from a potential slightly more negative than the
open circuit potential just measured. This procedure was adopted
in order to prevent a possible anodic polarization which could
have affected the cathodic behavior of the aluminum alloy sur-
face covered by cerium compounds. All the micro-electrochemical
measurements were performed in
solution.
a 0.05 M NaCl aqueous
In order to evaluate the effect of cerium on the corrosion behav-
ior of the AA2024-T3 microstructure, OCP and polarization tests
were performed on AA2024-T3 samples before and after immer-
sion in a 0.05 M NaCl solution containing 5 g L−1 Ce(NO3)3·6H2O.
Polished samples were immersed for 12 h in the cerium solu-
tion in order to allow cerium precipitation. In order to obtain
reliable results, for each region of interest and for each surface
condition (before and after interaction with cerium), more than
20 of both OCP measurements and polarization curves were car-
ried out. Following an approach based on the employment of
localized techniques, it is very important to acquire a reason-
able number of measurements able to validate the experimental
results.
2.3. Scanning electron microscopy
2. Materials and experimental procedure
Scanning electron microscopy (SEM) observations were com-
bined with electrochemical measurements. The combination of the
two techniques allowed to associate the electrochemical results
with the surface morphologies observed with SEM. In order to
enable the characterization of the same area of the sample before
and after immersion in the cerium-containing solution, the investi-
gated areas were marked with a grid of indentations prepared using
a Vickers micro-indenter (Struers Micro-Indenter). Square areas of
100 ␮m × 100 ␮m size were marked. This procedure enables to per-
form easily both SEM and micro-cell characterizations on the same
sample area. A Carl Zeiss EVO 40XVP microscope equipped with the
INCA ENERGY 250 microanalysis apparatus was used.
2.1. Materials
AA2024-T3 samples were ground using SiC paper and subse-
quently polished using 6 and 1 ␮m diamond suspensions. Ethanol
was used instead of water in order to avoid or reduce the fast disso-
lution of the magnesium present on aluminum surfaces, especially
as the constituent element of Al–Cu–Mg intermetallic particles.
2.2. Micro-electrochemical investigations
The micro-electrochemical investigation was carried out by
means of the micro-cell technique. The micro-cell apparatus basi-
cally consists of a glass microcapillary embedded in a micro-cell
which is mounted on the revolving nosepiece of an optical micro-
scope. This setup enables to point the measurements towards the
region desired by observing the surface with the optical micro-
scope. The sealing at the glass capillary tip is ensured by a silicone
layer which prevents to avoid solution leakage [18]. A standard
three electrode setup was employed for micro-electrochemical
measurements where the working electrode corresponded to the
exposed area defined by the diameter size of the glass micro
capillary.
3. Results and discussion
3.1. AA2024-T3 microstructure
Fig. 1 shows the surface morphology of AA2024-T3 after polish-
ing detected by means of scanning electron microscopy. The image
was acquired by collecting the back-scattered electron signal emit-
ted by the metal surface. From this picture, it is possible to recognize
the heterogeneous microstructure of the alloy. The dark area cor-
responds to the aluminum matrix while the brighter regions are

L. Paussa et al. / Electrochimica Acta 70 (2012) 25–33
27
Table 1
EDXS chemical composition (wt%) of the regions marked in Fig. 1.
Al (wt%)
Cu (wt%)
Mg (wt%)
Fe (wt%)
Mn (wt%)
Si (wt%)
Matrix (Spectrum 1)
Mg-rich INT (Spectrum 2)
Fe-rich INT (Spectrum 3)
93.84 0.4
88.40 0.4
65.54 0.5
4.30 0.3
8.09 0.4
7.79 0.3
1.86 0.1
3.52 0.1
–
–
–
–
–
–
–
13.06 0.3
9.34 0.2
4.27 0.4
associated to the second phase particles. As explained in Section 2,
in this work, three regions were considered as interesting areas
for the evaluation of cerium precipitation mechanisms: regions
rich in Al–Cu–Mg intermetallic particles (from now on called Mg-
rich), regions rich in Al–Cu–Fe–Mn intermetallic particles (from
now on called Fe-rich) and regions with a very small amount of
intermetallic particles (from now on called matrix). In Table 1 the
chemical composition of the regions identified by the three boxes in
Fig. 1 is reported: the matrix, the iron-rich intermetallic region and
the magnesium-rich intermetallic region. EDXS analysis (Table 1)
performed on the matrix area (Spectrum 1) evidences a chemical
composition in accordance with commercial AA2024-T3 aluminum
alloys [1]. Spectrum 2 can be associated to an intermetallic particle
rich in Al–Cu–Mg because the amount of magnesium and cop-
per is higher than the one detected in the matrix (Spectrum 1).
In Spectrum 3, in addition to copper and aluminum signals, iron,
manganese, and silicon were detected. This chemical composition
can be associated to the Al–Cu–Fe–Mn intermetallic particle type.
It has been further seen and confirmed by a deeper investigation
that intermetallic particles rich in iron are usually larger than those
rich in magnesium as found by Buchheit et al. [10]. While the max-
imum size of the magnesium-rich type is in the range of a few
micrometers, the size of the iron-rich type can vary in a large range
from a few micrometers up to 20 ␮m. Since in this work capillar-
ies with an average diameter of 50 ␮m were used, the area directly
exposed to the electrolyte necessarily includes both intermetal-
lic particles and a fraction of matrix. This aspect has to be taken
into account for the explanation of the results. It is also important
to point out that it was easier to perform micro-electrochemical
measurements on regions containing iron-rich intermetallic par-
ticles than on regions containing magnesium-rich intermetallic
particles. Indeed, Al–Cu–Mg-based intermetallic particles are more
reactive than those rich in iron because immediately after immer-
sion, magnesium dissolution strongly affects the electrochemical
behavior. This aspect will be discussed in detail in Section 3.3.
3.2. Open circuit potential before immersion in cerium nitrate
Fig. 2 shows the open circuit potential in 0.05 M NaCl for pol-
ished AA2024-T3 samples measured on the three areas of interest
defined in Section 2: the matrix region (solid line), the region con-
taining iron-rich intermetallic particles (long-dashed line) and the
region containing magnesium-rich intermetallic particles (dash-
dotted line). The matrix and iron-rich intermetallic regions behave
in a similar way. In the case of the matrix (solid line), immediately
after immersion OCP is −500 mV vs. Ag/AgCl and after 300 s the
value is very close to the one detected at the beginning. However,
some spikes were detected during the OCP measurement on the
matrix. The first small OCP variation can be identified after 50 s,
while after 130 s, the OCP suddenly starts to decrease to −600 mV
vs. Ag/AgCl. In the range between 130 and 200 s, the OCP for the
matrix region appears noisy suggesting that corrosion attacks are
taking place. This evidence can likely be associated to a localized
corrosion attack which is typical of AA2024-T3 aluminum alloys [9].
Immediately after immersion, OCP for the region containing iron-
rich intermetallic particles (long-dashed line) is equal to −510 mV
vs. Ag/AgCl and tends to continuously increase. It reaches −360 mV
vs. Ag/AgCl after 300 s immersion. The OCP shift towards more pos-
itive values might be related to the activation of reduction reactions
at the intermetallic particle surface which behaves cathodically
with respect to the matrix [37]. The OCP trend for the region con-
taining magnesium-rich intermetallic particles (dash-dotted line)
is not as constant as in the previously discussed case. Immedi-
ately after immersion, magnesium strongly affects the OCP which
is −640 mV vs. Ag/AgCl. For longer immersion time, the OCP shifts
from very negative values (the minimum value is in the range of
−1000 mV vs. Ag/AgCl) to values typical of the aluminum matrix
(close to −500 mV vs. Ag/AgCl). For magnesium-rich intermetallic
regions, the unstable OCP trend can be due to the high electrochem-
ical activity of magnesium which quickly dissolves during the first
period of immersion. In many cases (not here represented), imme-
diately after immersion, the OCP of magnesium-rich intermetallic
regions was very active (down to −1100 mV vs. Ag/AgCl) shifting to
Fig. 2. Open circuit potential for the AA2024-T3 microstructure in 0.05 M NaCl
acquired using a glass micro-capillary of 50 ␮m size.
Fig. 1. SEM morphology of the polished AA2024-T3 surface.

28
L. Paussa et al. / Electrochimica Acta 70 (2012) 25–33
Fig. 4. Corrosion morphology of a region containing a coarse Al–Cu–Fe–Mn inter-
metallic particle.
particles (long-dashed line) and the region containing magnesium-
rich intermetallic particles (dash-dotted line). The potentiody-
namic polarization curves were performed on the same area
exposed to the electrolyte during OCP measurements. The anodic
behavior shown in Fig. 3(a) evidences that the matrix and iron-rich
intermetallic regions behave in a similar way. For iron-rich inter-
metallic regions, Ecorr is more positive than that acquired for the
matrix, confirming the trend evidenced by OCP measurements. For
regions containing iron-rich intermetallic particles, icorr is around
1 ␮A cm−2 which is almost one order of magnitude higher than
that evaluated for the matrix region. This result can be explained
considering that in the region containing iron-rich intermetallic
particles, a fraction of the area exposed to the electrolyte is made
of matrix. Hence, aluminum matrix dissolution is promoted by
the galvanic coupling with iron-rich intermetallic particles leading
to higher corrosion rates. Fig. 4 shows the corrosion morphology
of a region containing Al–Cu–Fe–Mn intermetallic particles. The
image was acquired after the carrying out the anodic polarization.
A deep corrosion attack can be seen in the aluminum matrix near
the intermetallic particle boundary. This evidence is related to the
higher corrosion current density detected for iron-rich intermetal-
lic regions with respect to that identified for regions containing
only the matrix. This behavior can also be associated to the dif-
ferent breakdown potentials exhibited by matrix and iron-rich
intermetallic regions. The natural oxide layer grown on the inter-
face between the matrix and the intermetallic particles tends to
be weak. These are the points where the oxide can be dissolved
by polarizing anodically the substrate. Corrosion of the matrix can
be otherwise related to the presence of strengthening particles
which can act as precursor sites for the onset of localized cor-
rosion attacks. However, this aspect has not been considered in
this paper because it will be a part of another work, in which a
combination of different localized techniques has been considered.
In Fig. 3(a), it is possible to observe that for the matrix, the pas-
sive region is wider than for iron rich intermetallic particles. The
breakdown potential related to the localized corrosion attack can
be identified at −290 mV and −330 mV vs. Ag/AgCl for the matrix
and the iron-rich intermetallic regions, respectively. The anodic
behavior of regions containing magnesium-rich intermetallic par-
ticles (Fig. 3(a)) is rather different compared to the anodic behavior
of both the matrix and the iron-rich intermetallics regions. For
magnesium-rich intermetallic regions, Ecorr is more active than
the ones of other regions. This evidence confirms OCP measure-
ment results. Magnesium strongly influences the electrochemical
behavior of every alloy and in the case of AA2024-T3, second phase
Fig. 3. Anodic (a) and cathodic (b) polarization curves for the AA2024-T3
microstructure in 0.05 M NaCl acquired using a glass micro-capillary of 50 ␮m.
values typical of the matrix (close to −500 mV vs. Ag–AgCl) before
the end of the measurement (300 s). In these cases, the results of
potentiodynamic polarization measurements, which will be dis-
cussed in Section 3.3 (Results and Discussion), were very similar
to those obtained on the matrix regions. In those cases, the mag-
nesium content was totally dissolved during free immersion and
the area which underwent polarization was probably consisting of
both the matrix and the intermetallic particles depleted in magne-
sium and rich in copper. This evidence was one of the most critical
points to be considered in the experimental approach for the eval-
uation of the micro-electrochemical behavior of AA2024-T3. From
this point of view, the polishing procedure has had an important
role, because, as explained in Section 2, it was very important to
prevent magnesium dissolution during sample preparation.
3.3. Potentiodynamic polarization curves before immersion in
cerium nitrate
As explained in Section 2.2, the polished AA2024-T3 surface
was electrochemically characterized by acquiring a large number
of anodic and cathodic polarization curves. The results discussed
in the paper are therefore representative of the overall measure-
ments carried out. Table 2 reports the average and the standard
deviation of corrosion and breakdown potentials for all the anodic
polarizations carried out in 0.05 M NaCl. Fig. 3 shows anodic and
cathodic potentiodynamic polarization curves in 0.05 M NaCl for
polished AA2024-T3 on the three areas of interest: the matrix
region (solid line), the region containing iron-rich intermetallic

L. Paussa et al. / Electrochimica Acta 70 (2012) 25–33
29
Table 2
Average and standard deviation of corrosion and breakdown potentials (both vs. (Ag/AgCl)/mV) calculated from anodic polarizations carried out in 0.05 M NaCl.
Ecorr (mV)
Standard deviation (mV)
Breakdown potential (mV)
Standard deviation (mV)
Matrix
Mg-rich INT
Fe-rich INT
−437.1
−748.4
−323.7
34.4
81.6
21.6
−228.2
−564.6
−312.1
25.6
59.2
17.6
particles containing a high magnesium amount have a remarkable
effect on Ecorr. From Fig. 3(a), for magnesium-rich intermetallic
regions, icorr is about 20 ␮A cm−2 which is two orders of magni-
tude higher than for the matrix. The breakdown potential can be
identified at −640 mV vs. Ag/AgCl which is very active with respect
to those exhibited by the matrix and the iron-rich intermetallic
particles. The magnesium effect on the electrochemical behavior of
AA2024-T3 can also be evaluated in Fig. 3(b) where cathodic poten-
tiodynamic polarization curves are reported. The matrix and the
iron-rich intermetallic regions exhibit a similar cathodic behavior
near Ecorr where the oxygen reduction and the hydrogen evolu-
tion reactions take place. At much more negative potentials than
Ecorr, cathodic reactions taking place on both matrix and iron-
rich intermetallic regions are strongly increased. The increase is
most likely related to the water reduction reaction occurring at
these potentials. The cathodic current densities exhibited by the
magnesium-rich intermetallic regions are higher than those mea-
sured on the matrix. For magnesium-rich intermetallic regions,
at a potential that is around −1000 mV vs. Ag/AgCl, the current
densities are in the range of 400 ␮A cm−2 which are one order of
magnitude higher than those measured for the matrix region. At
these potentials, on regions rich in magnesium intermetallic parti-
cles, a strong bubble evolution, which can be likely related to an
intense hydrogen reduction reaction, was observed [38]. At the
end of the micro-electrochemical investigation carried out on bare
AA2024-T3, it is possible to state that magnesium strongly affects
the corrosion behavior of the alloy. Anodic and cathodic reactions
are heavily dependent on the amount of magnesium contained in
the region under investigation. In Section 3.4, the effect of cerium
nitrate on the heterogeneous electrochemical behavior of AA2024-
T3 aluminum alloy will be discussed.
matrix regions (thick solid line), on regions containing iron-rich
intermetallic particles (thick long-dashed line) and on regions con-
taining magnesium-rich intermetallic particles (thick dash-dotted
line). OCP measurements acquired before immersion in the solu-
tion containing cerium (see Fig. 2) are reported here as reference
(thinner lines). By comparing OCP measurements before and after
immersion in cerium solution, it is possible to point out that a
remarkable modification of the electrochemical behavior can be
observed for the entire surface. For the matrix, the iron-rich and
the magnesium-rich intermetallic regions, OCP tends to be more
positive than before immersion in cerium (thinner lines) and for all
three regions, after 50 s in 0.05 M NaCl, a steady state condition is
reached and maintained until the end of the measurement suggest-
ing that a cerium film has been deposited everywhere. Fig. 6 shows
a detail of the AA2024-T3 surface modified by the cerium inter-
action. The figure shows the surface morphology of an iron-rich
intermetallic region (with cathodic behavior) after 12 h of immer-
sion in 0.05 M NaCl containing 5 g L−1 Ce(NO3)3·6H2O. From the
picture, it is possible to recognize the presence of a very thin cerium
film on the iron-rich intermetallic particle only because a film crack
is visible. On the surrounding matrix, it is not possible to recognize
the cerium film. This does not mean that the film is not present
but that the film is only too thin to be observed. SEM-EDXS anal-
yses performed on the region considered in Fig. 6 are not able to
clearly identify the cerium signal because the penetration depth
of the electron beam into the specimen is in the range of microm-
eters while the film thickness is probably in the range of a few
nanometers. Hence, the contribution of the thin cerium film on
the X-ray generation is not enough to guarantee a clear cerium
signal. As well known in the literature [35], cerium precipitation
occurs on regions where cathodic reactions take place. However,
as shown in Fig. 7, cerium compounds were also observed not
only on intermetallic particles with a cathodic behavior (Fig. 6) but
also on intermetallic particles with anodic behavior (Al–Cu–Mg-
based) as intense precipitates, already determined by Yasakau et
al. [36]. Table 3 shows the EDXS micro-chemical composition of
the region containing the magnesium-rich intermetallic particles
which is indicated with the arrows in Fig. 7(a) and (b). Before
3.4. Open circuit potential after immersion in cerium nitrate
Fig. 5 shows the open circuit potential for polished AA2024-T3
in 0.05 M NaCl measured after 12 h immersion in 0.05 M NaCl con-
taining 5 g L−1 Ce(NO3)3·6H2O. Measurements were performed on
Fig. 5. Open circuit potential for the AA2024-T3 microstructure in 0.05 M NaCl, after
12 h of immersion in cerium nitrate, acquired using a glass micro-capillary of 50 ␮m
Fig. 6. Surface morphology of a coarse Al–Cu–Fe–Mn intermetallic particle after
size.
12 h immersion in cerium nitrate.

30
L. Paussa et al. / Electrochimica Acta 70 (2012) 25–33
Table 3
EDXS chemical composition (wt%) of the marked region indicated by the circle in Fig. 7 acquired before (a) and after (b) immersion in cerium nitrate.
Al (wt%)
Cu (wt%)
Mg (wt%)
Ce (wt%)
O (wt%)
Before immersion
After immersion
64.78 0.5
4.81 0.1
26.39 0.4
2.12 0.2
7.46 0.2
–
–
1.11 0.3
37.26 0.4
55.45 0.4
immersion (Fig. 7(a)), a high amount of magnesium was detected
(7.46 wt%), which was not detected on the same region after immer-
sion (Fig. 7(b)). On the other hand, an intense signal of cerium and
oxygen was detected suggesting that a heavy cerium precipita-
tion occurred in this region during immersion in cerium-containing
solution. Since after immersion the aluminum signal drastically
decreases, the precipitate is probably thick enough to shield the
metal substrate. EDXS analysis does not allow to acquire signals
deeper than a couple of micrometers (depending on the material
under investigation), and the low aluminum signal acquired is a
consequence of this limitation.
In the AA2024-T3 aluminum alloy, immediately after immer-
sion, Al–Cu–Mg intermetallic particles behave anodically [39] but
they can behave cathodically after magnesium dissolution. From
Fig. 7(b), it is possible to recognize that on the marked region a
wide precipitation occurred covering the magnesium-rich inter-
metallic particles. The dark matrix region appears smoother than
the bright magnesium-rich intermetallic particles, suggesting that
the precipitation which occurred on the matrix was not as strong
as on magnesium-rich intermetallic particles [36]. Eckermann et
al. [40] studied the micro-electrochemical reactivity of intermetal-
lic particles present in another group of aluminum alloys mainly
containing Al–Mg–Si. It has been found that an increasing of solu-
tion pH can be identified on MgSi intermetallic particles which
should have an anodic behavior with respect to the matrix. Two
different hypotheses have been formulated to explain that evi-
dence. The first explanation is related to the development on MgSi
particles of a strong hydrogen reduction associated to a fast mag-
nesium dissolution [41]. The second mechanism takes its stand on
the hypothesis that after magnesium dissolution, MgSi particles
can change their behavior from anodic to cathodic. On modified
MgSi particles, reduction reactions can occur producing the local
increase in pH. A similar explanation can be given for the micro-
electrochemical behavior of Al–Cu–Mg particles embedded in the
AA2024-T3 microstructure investigated in the present work. The
cerium deposition mechanism can be therefore related to the fast
magnesium dissolution which should be able to supply electrons
for cathodic reactions occurring very close to the anodic region.
This first mechanism considers that magnesium dissolution and
cathodic reactions occur together on intermetallic particles rich in
magnesium which are immediately covered by cerium precipita-
tion. Otherwise, a second cerium precipitation mechanism can be
based on a slower precipitation kinetic. Measurements shown in
Fig. 2 evidence that the OCP for magnesium-rich intermetallic par-
ticles can shift from very negative potentials to potentials typical of
the matrix in a very limited length of time. As already discussed, it
is possible to suppose that a main part of the total amount of mag-
nesium completely dissolves before cerium precipitation quickly
occurs. If this happens, intermetallic particles are depleted in mag-
nesium and are rich in copper and aluminum. Hence, it is possible
that cerium precipitation occurs on the magnesium-depleted inter-
metallic particles which in the meantime might behave similar to
a binary intermetallic particle rich in copper and aluminum. Binary
Al–Cu intermetallic particle is more positive than both the ternary
Al–Cu–Mg and the Al–Cu–Fe–Mn types [9]. It is reported [9] that
in sodium chloride, the corrosion potential of the Al–Cu binary
compound is −0.32 mV vs. SCE while that of the Al–Cu–Mg type
is −0.92 mV vs. SCE. The difference in corrosion potentials might
explain how a preferential cerium precipitation might occur on
magnesium-depleted intermetallic particles where cathodic reac-
tions directly occur promoting the increase of the local pH needed
for cerium compounds precipitation. It is very difficult to prove
what of the mechanism is more probable but a combination of the
two routes may be a reasonable hypothesis.
Fig. 7(a) and (b) allows us to confirm that the preferential sites
for cerium precipitation are those rich in magnesium [36]. How-
ever, since cerium precipitation occurred on the entire AA2024-T3
surface, the film growth can be explained by considering the island-
growth mechanism proposed by Hughes et al. [35] for aluminum
alloys. According to this mechanism, the island-growth considers
the microstructure of metal substrates as a collection of micro-
anodes and micro-cathodes where an increasing in the local pH,
necessary for cerium precipitation, can occur. Cerium precipitation
is then limited when the entire surface is covered by the sponta-
neous film deposition.
3.5. Potentiodynamic polarization curves after immersion in
cerium nitrate
Fig. 8 shows potentiodynamic polarization curves in 0.05 M NaCl
after immersion in the solution containing cerium for the three
Fig. 7. Surface morphology of an AA2024-T3 marked region acquired before (a) and
after (b) immersion in cerium nitrate.

L. Paussa et al. / Electrochimica Acta 70 (2012) 25–33
31
Table 4
Average and standard deviation of corrosion and breakdown potentials (both vs. (Ag/AgCl)/mV), after 12 h immersion in cerium nitrate solution, calculated from anodic
polarizations carried out in 0.05 M NaCl.
Ecorr (mV)
Standard deviation (mV)
Breakdown potential (mV)
Standard deviation (mV)
Matrix
Mg-rich INT
Fe-rich INT
−98.1
−322.5
−224.6
22.4
42.4
28.8
−34.3
+58.2
−84.2
23.2
49.6
12.8
areas of interest: the matrix region (solid line), the region con-
taining iron-rich intermetallic particles (long-dashed line) and the
region containing magnesium-rich intermetallic particles (dash-
dotted line). Potentiodynamic polarization curves were performed
on the same area exposed to the electrolyte during OCP measure-
ments (see Fig. 5). Table 4 reports the average and the standard
deviation of corrosion and breakdown potentials calculated from
all the anodic polarizations carried out in 0.05 M NaCl, after 12 h
immersion in cerium-containing solution. After cerium immer-
sion, for all three regions, Ecorr shifts towards more positive values
confirming OCP results. In the case of the matrix region, Ecorr
shifts from −400 mV to −100 mV vs. Ag/AgCl provoking a ꢀEcorr
equal to 300 mV. For the region containing iron-rich intermetal-
lic particles, Ecorr shifts from −360 mV to −200 mV vs. Ag/AgCl
while it changes from −800 mV to −300 mV vs. Ag/AgCl for the
region rich in Al–Cu–Mg intermetallic particles. Regions containing
a strong reduction. In the case of matrix regions, icorr changes from
0.2 ␮A cm−2 (Fig. 3) to 0.03 ␮A cm−2 (Fig. 8). For the region con-
taining iron-rich intermetallic particles, before immersion, icorr is
2 ␮A cm−2 while it is reduced to 0.03 ␮A cm−2 after immersion.
Regions containing magnesium-rich intermetallic particles evi-
dence the largest ꢀicorr: before immersion icorr is 20 ␮A cm−2 while
after immersion it decreases down to 0.06 ␮A cm−2. It is important
to point out this evidence because regions containing magnesium-
rich intermetallic particles, before immersion in a cerium solution,
tend to corrode faster than both the matrix and the regions con-
taining iron-rich intermetallic particles. After cerium precipitation,
corrosion current densities measured on the entire surface are
closer to a common value than before immersion. Moreover, areas
containing magnesium-rich intermetallic particles display a very
wide passive region (Fig. 8) which is larger than that exhibited
by both the matrix and the iron-rich intermetallic regions. By
considering Fig. 3(a), the anodic behavior of magnesium-rich inter-
metallics regions is considerably improved by the immersion in the
cerium solution (Fig. 8(a)). After immersion in the cerium solution,
the passive current densities are two orders of magnitude lower
than those before immersion. The breakdown potential is around
+50 mV vs. Ag/AgCl which is 100 mV and 150 mV more positive than
for the matrix and the iron-rich intermetallic region, respectively.
The effect of cerium precipitation can also be evaluated considering
the cathodic behavior (Fig. 8(b)). Reduction reactions occurring on
the entire substrate are strongly reduced by cerium precipitation.
In particular, cathodic current densities measured on regions con-
taining magnesium-rich intermetallic particles are strongly limited
since they are reduced more than two orders of magnitude. The
strong current reduction observed for the magnesium-rich inter-
metallic region can be related again to the high amount of cerium
compounds precipitated on this region. However, the inhibition of
cathodic reactions exhibited by the entire surface confirms that
cerium precipitation involves all the microstructure. The electro-
chemical behavior of the three regions of interest is therefore
more homogeneous than before immersion in the electrolyte con-
taining cerium. After cerium precipitation, which clearly occurs
on the entire exposed surface, both the anodic and the cathodic
behaviors appear very similar for the three regions investigated. In
agreement with a recent paper written by Birbilis et al. [42], cerium-
compounds are able to provide anodic and cathodic inhibitions.
Birbilis, however, reported an increase in corrosion potential with
the presence of cerium dibutyl phosphate only for S-phase inter-
metallic particles. A decrease in corrosion potential was detected
for intermetallic particles with a cathodic behavior with respect
to the aluminum matrix. This apparent disagreement is proba-
bly due to the different approaches used because, in the case
of the approach followed by Birbilis, intermetallic compounds of
thousands micrometers in diameter were synthesized and tested
without the influence of the aluminum matrix which was instead
considered by the experimental approach adopted in the present
work.
magnesium-rich intermetallic particles exhibit the largest ꢀEcorr
.
Due to the immersion in cerium solution, icorr is also subjected to
Fig. 9 shows a region containing a large intermetallic particle
with a very complex chemical composition. The three images ((a),
(b) and (c)), were acquired after the anodic polarization carried out
subsequently to the immersion in the solution containing 5 g L−1
cerium nitrate. Table 5 shows the average chemical composition
Fig. 8. Anodic (a) and cathodic (b) polarization curves for the AA2024-T3
microstructure in 0.05 M NaCl, after 12 h of immersion in cerium nitrate, acquired
using a glass micro-capillary of 50 ␮m.

32
L. Paussa et al. / Electrochimica Acta 70 (2012) 25–33
Fig. 9. Surface morphology of a mixed Al–Cu–Fe–Mn/Al–Cu–Mg intermetallic particle after the carrying out anodic polarization in 0.05 M NaCl. Before the measurement, the
AA2024-T3 sample was hold for 12 h in cerium nitrate.
Table 5
EDXS chemical composition (wt%) of the intermetallic regions indicated in Fig. 9. The average composition of the entire intermetallic was acquired before the immersion in
cerium nitrate while localized analyses were performed after the anodic polarization measurement.
Average composition before immersion
Region 1 after immersion
Region 2 after immersion
Al (wt%)
Cu (wt%)
Mg (wt%)
Fe (wt%)
Mn (wt%)
Si (wt%)
Ce (wt%)
O (wt%)
70.62 0.5
10.38 0.3
2.23 0.1
6.71 0.2
4.69 0.2
2.25 0.1
–
18.85 0.5
7.57 0.6
52.33 0.6
8.77 0.5
–
15.46 0.4
10.97 0.4
–
–
–
–
–
41.62 0.8
30.35 0.8
1.29 0.3
7.12 0.6
3.11 0.4
acquired on the whole intermetallic particle before immersion in
the cerium-containing solution. In Table 5, the chemical composi-
tions of Region 1 and Region 2 after immersion in cerium nitrate are
also reported. As shown in Table 5, Mg, Al, Cu, Fe, Mn, and Si signals
were detected in the average composition of the intermetallic par-
ticle before immersion in cerium. The latter three elements, with
Cu and Al, form intermetallic particles with cathodic behavior [9]
while intermetallic particle types containing Mg, Al, and Cu behave
anodically with respect to the matrix. Hence, the intermetallic par-
ticle observed in Fig. 9 is probably made of second phase particles
with different micro electrochemical behavior. The intermetallic
particle is probably formed by the combination of a part rich in
Al–Cu–Mg (Region 1) and another part containing Al–Cu–Fe–Mn
(Region 2). In Fig. 9(a), the silicon mark produced by the contact
between the capillary mouth and the AA2024-T3 surface defines
the area exposed to the electrolyte during the measurement. It cor-
responds to a diameter in the range of 50 ␮m as the size of the glass
capillary. Fig. 9(c) shows the localized corrosion attack at the matrix
near to the interface with the intermetallic particle (Region 2). The
corrosion attack occurred in correspondence with a restricted edge.
In the opposite part of the intermetallic particle (Region 1), no evi-
dence of corrosion is observed (Fig. 9(b)). In Fig. 9(b), relevant to
Region 1, it is also possible to recognize a strong cerium precipita-
tion (see Table 5) which was most likely able to protect the metal
substrate when the anodic polarization measurement was carried
out. The chemical micro-analysis carried out on the intermetallic
region represented in Fig. 9(c) exhibits a weak signal of cerium and
oxygen while intense peaks of iron and manganese were detected.
Then, the pitting corrosion of the matrix visible in Fig. 9(c), occurred
in correspondence with the intermetallic particle interface rich in
iron and manganese. The small amount of cerium detected and
the morphology observed in Fig. 9(c) indicate that a very thin film
covers this part of the intermetallic particle and the surrounding
matrix. The same evidence was observed for the iron-rich inter-
metallic particle of Fig. 6. The corrosion attack identified at Region
2 is probably related to a limited protection provided by cerium
compounds while the Region 1 is completely covered by a thick
cerium precipitate. This aspect evidences one more time that on
intermetallic regions rich in iron and manganese, cerium precipita-
tion is less intense than on regions rich in magnesium-intermetallic
particles.
4. Conclusions
The electrochemical micro-cell technique has been employed
for the micro-evaluation of the effect of cerium nitrate on the corro-
sion behavior of AA2024-T3 aluminum alloy. Magnesium-rich and
iron-rich intermetallic particles have been considered as well as the
matrix region. The corrosion behavior of AA2024-T3 samples after
immersion for 12 h in the solution containing the cerium nitrate
salt is significantly improved. The precipitation of thin cerium com-
pounds covering the entire metal surface improves both the anodic
and the cathodic behavior of the different analyzed regions. Hence,
the electrochemical behavior of the considered aluminum alloy
appears to be quite improved after the precipitation of cerium com-
pounds on both the matrix and the areas containing intermetallic
particles.
Cerium precipitation mechanism depends on the local pH val-
ues controlled by the cathodic reactions occurring on the whole
metal surface for which the kinetics is affected by the specific
cathodic activity of each considered region (matrix or inter-
metallic particles). In particular, regions containing high amounts
of magnesium-rich intermetallic particles are very reactive. The
cerium deposition mechanism can be therefore related to the
fast magnesium dissolution which should be able to supply elec-
trons for cathodic reactions occurring very close to the anodic
region. This mechanism considers that magnesium dissolution and
cathodic reactions occur together on intermetallics rich in mag-
nesium which are immediately covered by cerium precipitation.
Therefore, magnesium-rich intermetallic regions showed to be
sites where a more intense cerium precipitation can occur. Due to
the intense cerium precipitation, these areas become electrochem-
ically more stable and the overall current flow related to Faradic
processes involving the entire microstructure regions is strongly
reduced.

L. Paussa et al. / Electrochimica Acta 70 (2012) 25–33
33
Acknowledgment
[20] T. Suter, T. Peter, H. Bohni, Mater. Sci. Forum 25 (1995) 192.
[21] H. Bohni, T. Suter, A. Schreyer, Electrochim. Acta 40 (1995) 1361.
[22] T. Suter, R.C. Alkire, J. Electrochem. Soc. 148 (2001) B36.
[23] J. Wloka, G. Burklin, S. Virtanen, Electrochim. Acta 53 (2007) 2055.
[24] F. Andretta, M.M. Lohrengel, H. Terryn, J.H.W. de Wit, Electrochim. Acta 48
(2003) 3239.
The authors are grateful to Prof. Marjorie Olivier, University of
Mons, for the valuable and useful contributions on the discussion
of the results.
[25] F. Andreatta, H. Terryn, J.H.W. de Wit, Electrochim. Acta 49 (2004) 2851.
[26] N. Birbilis, M.K. Cavanaugh, R.G. Buchheit, Corros. Sci. 48 (2006) 4202.
[27] N. Birbilis, R.G. Buchheit, J. Electrochem. Soc. 152 (2005) B140.
[28] M. Bethencourt, F.J. Botana, J.J. Calvino, M. Marcos, M.A. Rodriguez-Chacon,
Corros. Sci. 40 (1998) 1803.
[29] B. Davo, J.J. de Damborenea, Electrochim. Acta 49 (2004) 4957.
[30] R.L. Twite, G.P. Bierwagen, Prog. Org. Coat. 33 (1998) 91.
[31] B.R.W. Hinton, D.R. Arnott, N.E. Ryan, Met. Forum 7 (1984) 211.
[32] B.R.W. Hinton, D.R. Arnott, N.E. Ryan, Mater. Forum 9 (1986) 162.
[33] D.R. Arnott, B.R.W. Hinton, N.E. Ryan, Corrosion (Houston, TX, USA) 45 (1989)
12.
[34] D.R. Arnott, B.R.W. Hinton, N.E. Ryan, Mater. Perform. 42 (1987) 47.
[35] A.E. Hughes, S.G. Hardin, T.G. Harvey, T. Nikpour, B. Hinton, A. Galassi, G.
McAdam, A. Stonham, S.J. Harris, S. Church, C. Figgures, D. Dixon, C. Bow-
den, P. Morgan, S.K. Toh, D.D. McCulloch, J. DuPlesis, Proceedings of the Third
International Symposium on Aluminum Surface Science and Technology, Bonn,
Germany, 2003, p. 264.
[36] K.A. Yasakau, M.L. Zheludkevich, S.V. Lamaka, M.G.S. Ferreira, J. Phys. Chem. B
110 (2006) 5515.
[37] M. Iannuzzi, G.S. Frankel, Corros. Sci. 49 (2007) 2371.
[38] J.H. Greenblatt, Can. J. Chem. 36 (1958).
[39] N. Dimitrov, J.A. Mann, M. Vukmirovic, K.J. Sieradzki, J. Electrochem. Soc. 147
(2000) 3283.
[40] F. Eckermann, T. Suter, P.J. Uggowitzer, A. Afseth, P. Schmutz, Electrochim. Acta
54 (2008) 844.
References
[1] R.H. Stevens, Metals Handbooks Manuals, vol. 9, ASM International, United
States of America, 1985.
[2] A. Garner, D. Tromans, Corrosion (Houston, TX, USA) 35 (1979) 55.
[3] I.L. Mueller, J.R. Galvele, Corros. Sci. 17 (1977) 179.
[4] R. Bakish, W.D. Robertson, J. Electrochem. Soc. 103 (1956) 320.
[5] K. Urishino, K. Sugimoto, Corros. Sci. 19 (1979) 225.
[6] Z. Szklarska-Smialowska, Corros. Sci. 41 (1999) 1743.
[7] P. Campestrini, E.P.M. van Westing, H.W. van Rooijen, J.H.W. de Wit, Corros.
Sci. 42 (2000) 1853.
[8] I.T.E. Fonseca, N.E. Lima, J.A. Rodrigues, M.I.S. Pereira, J.C.S. Salvador, M.G.S.
Ferreira, Electrochem. Commun. 4 (2002) 353.
[9] R.G. Buchheit, J. Electrochem. Soc. 142 (11) (1995) 3994.
[10] R.G. Buchheit, R.P. Grant, P.F. Halva, B. Mckenie, G.L. Zender, J. Electrochem. Soc.
144 (8) (1997) 2621.
[11] T.J. Warner, M.P. Schmidt, F. Sommer, D. Bellot, Z. Metallkd. 86 (1995) 494.
[12] V. Guillaumin, G. Mankowski, Corros. Sci. 41 (1999) 421.
[13] C.M. Liao, R.P. Wei, Electrochim. Acta 45 (1999) 881.
[14] N. Birbilis, R.G. Buchheit, J. Electrochem. Soc. 155 (3) (2008) C117.
[15] A. Boag, A.E. Hughes, A.M. Glenn, T.H. Muster, D. McCulloch, Corros. Sci. 53
(2011) 17.
[16] D. Battocchi, J. He, G.P. Bierwagen, D.E. Tallman, Corros. Sci. 47 (2005) 1165.
[17] J.C. Seegmiller, D. Buttry, J. Electrochem. Soc. 150 (2003) B413.
[18] T. Suter, Ph.D. Thesis No. 11962, ETH Zurich, 1997.
[19] T. Suter, H. Bohni, Electrochim. Acta 42 (1997) 3275.
[41] G.L. Song, A. Atrens, Adv. Eng. Mater. 1 (1999) 11.
[42] N. Birbilis, R.G. Buchheit, D.L. Ho, M. Forsyth, Electrochem. Solid-State Lett. 8
(2005) 180.