Electrochimica Acta 51 (2005) 208–217

Nanostructured sol–gel coatings doped with cerium nitrate as

pre-treatments for AA2024-T3
Corrosion protection performance
M.L. Zheludkevich a,∗ , R. Serra a , M.F. Montemor b , K.A. Yasakau a ,
I.M. Miranda Salvado c , M.G.S. Ferreira a,b
a University of Aveiro, Department of Ceramics and Glass Engineering, 3810-193 Aveiro, Portugal
b Instituto Superior Técnico, ICEMS, Dep. Chem. Eng., Av. Rovisco Pais 1049-001 Lisboa, Portugal
c University of Aveiro, CICECO, Department of Ceramics and Glass Engineering, 3810-193 Aveiro, Portugal

Received 21 December 2004; received in revised form 12 February 2005; accepted 16 April 2005
Available online 17 May 2005

Abstract

Nanostructured hybrid sol–gel coatings doped with cerium ions were investigated in the present work as pre-treatments for the AA2024-
T3 alloy. The sol–gel films have been synthesized from tetraethylorthosilicate (TEOS) and 3-glycidoxypropyltrimethoxysilane (GPTMS)
precursors. Additionally the hybrid sol was doped with zirconia nanoparticles prepared from hydrolyzed tetra-n-propoxyzirconium (TPOZ).
Cerium nitrate, as corrosion inhibitor, was added into the hybrid matrix or into the oxide nanoparticles.
The chemical composition and the structure of the hybrid sol–gel films were studied by XPS (X-ray photoelectron spectroscopy) and
AFM (atomic force microscopy), respectively. The evolution of the corrosion protection properties of the sol–gel films was studied by EIS
(electrochemical impedance spectroscopy), which can provide quantitative information on the role of the different pre-treatments. Different
equivalent circuits, for different stages of the corrosion processes, were used in order to model the coating degradation. The models were
supported by SEM (scanning electron microscopy) measurements.
The results show that the sol–gel films containing zirconia nanoparticles present improved barrier properties. Doping the hybrid nanos-
tructured sol–gel coatings with cerium nitrate leads to additional improvement of the corrosion protection. The zirconia particles present in
the sol–gel matrix seem to act as nanoreservoirs providing a prolonged release of cerium ions. The nanostructured sol–gel films doped with
cerium nitrate can be proposed as a potential candidate for substitution of the chromate pre-treatments for AA2024-T3.
© 2005 Elsevier Ltd. All rights reserved.

Keywords: AA2024; EIS; Corrosion; Nanostructured coating; Sol–gel; Inhibitor; Cerium

1. Introduction Si O Al conversion layer providing a stable alumina/sol–gel

The development of environmentally friendly pre-
treatments for metallic substrates is a filed of growing
interest due to banning the use of chromates as protective
pre-treatments [1].
Among the possible candidates for environmentally
friendly pre-treatments for aluminium alloys are the silica-
based sol–gel ones [2]. Such pre-treatments are able to form a
∗ Corresponding author. Tel.: +351 234 378146; fax: +351 234 425300.
E-mail address: mzheludkevich@cv.ua.pt (M.L. Zheludkevich).
film interface, which impairs the onset of corrosion [3].
Schmidt et al. [4] reports that the alumina–silica mixed
compounds present Gibbs energy lower than the boehmite
(AlO(OH)), which is the product of the first stage of alu-
minium oxidation in the presence of moisture. The lower
value of the Gibbs energy leads to thermodynamical stability
of the silica/aluminium interface in corrosive environments.
The chemical bonding of the sol–gel films to the aluminium
substrate also contributes to improved adhesion. Moreover
sol–gel coatings, especially the hybrid films, provide a dense
barrier against electrolyte uptake [2]. However, the coating,

0013-4686/$ – see front matter © 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2005.04.021
 M.L. Zheludkevich et al. / Electrochimica Acta 51 (2005) 208–217
by itself, is not sufficient to effectively protect for long term
the metal substrate against corrosion. The sol–gel coatings
contain micro-pores, cracks and areas of low cross-link den-
sity that provides pathways for diffusion of corrosive species
to the coating/metal interface [2]. The corrosion processes
start in these zones. The crack ability and porosity of the
sol–gel films can be decreased by incorporation of oxide
nanoparticles into the hybrid matrix [5]. However, even such
films cannot offer an adequate protection when the coating is
damaged due to the lack of self-healing properties.
Incorporation of corrosion inhibitors into the sol–gel films
can improve the protective ability of the coatings, suppressing
the corrosion process in the defects. Phosphates, vanadates,
borates, cerium and molybdenum compounds were found to
have inhibiting action on the corrosion processes [6–8]. It has
been reported [6,9–11] that some of the most effective and
environmentally friendly corrosion inhibitors for aluminium
alloys are derived from cerium salts. This seems to result
from deposition of hydrated cerium oxide on the cathodic
intermetallic particles, existing in the aluminium alloy,
thus suppressing the cathodic reaction [12]. Voevodin et al.
[13] investigated the corrosion protection properties of the
epoxy-zirconia sol–gel coatings containing inhibitor ions
salts such as Ce(NO3 )3 , NaVO3 , and Na2 MoO4 . The sol–gel
films with NaVO3 and Na2 MoO4 did not provide adequate
corrosion protection due to the decrease of the sol–gel
network stability. However, the sol–gel coatings with cerium
dopants performed at least as good as the undoped epoxy-
zirconia films [13]. Additional improvement of the corrosion
protection was also revealed in the case of silane pre-
treatments when cerium nitrate was introduced in the coating
[14].
Organic inhibitors can be also incorporated in the
sol–gel matrix in order to improve the corrosion protection
of metallic substrates. Khramov et al. [15] studied the
protection properties of SNAP (Self-assembled NAnophase
Particles) films with incorporated mercaptobenzothiazole
and mercaptobenzimidazole as corrosion inhibitors. The
␤-cyclodextrin was also included in some of these coatings.
The corrosion protection of sol–gel derived coatings was
sufficiently improved when the encapsulation of the organic
inhibitors in the presence or absence of ␤-cyclodextrin
took place. However, the SNAP films with ␤-cyclodextrin
exhibited superior corrosion protection due to the prolonged
release of the inhibitor from the cyclodextrin–inhibitor com-
plexes. As a result the incorporation of corrosion inhibitors
absorbed in nanoparticles could be a good way to enhance
the protection properties of these films. This procedure can
provide an additional reinforcement of the hybrid matrix by
the nanoparticles [5] and a prolonged release of inhibitor.
Table 1
Composition of aluminium alloy 2024-T3
Element Cu Cr Fe Mg
Concentration 3.8–4.9 0.1 0.5 1.2–1.8
209
The present work aims at producing and assessing the
corrosion protection conferred to the aluminium alloy
AA2024-T3 by hybrid sol–gel films with incorporated
zirconia nanoparticles and doped with cerium nitrate as
corrosion inhibitor. The cerium inhibitor was added to the
hybrid matrix or to the oxide nanoparticles at the respective
hydrolysis steps. The hybrid film and the oxide nanopar-
ticles were prepared by the controllable hydrolysis of 3-
glycidoxypropyltrimethoxysilane (GPTMS)/tetraethylortho-
silicate (TEOS) and tetra-n-propoxyzirconium (TPOZ),
respectively. Electrochemical impedance spectroscopy (EIS)
was used to assess the corrosion performance of the devel-
oped pre-treatments. The surface morphology and the evo-
lution of the coatings structure were investigated by atomic
force microscopy (AFM) and scanning electron microscopy
(SEM) techniques. The chemical composition of the pre-
treated surfaces was assessed by X-ray photoelectron spec-
troscopy (XPS) and energy dispersive spectroscopy (EDS).
2. Experimental
2.1. Sol–gel pre-treatments
The aluminium alloy 2024-T3 was used as substrate. The
nominal composition of the used alloy is given in Table 1.
The metallic samples were immersed in an alkaline aqueous
TM
solution containing 60 g/l of TURCO 4215 for 15 min at
◦
60 C followed by immersion for 15 min at room temperature
in a 20 wt.% solution of nitric acid before coating application.
The nanostructured hybrid films were synthesized using
the controllable sol–gel route. An organosiloxane alcosol
with another alcosol containing the zirconia precursor were
combined in order to provide the hybrid sol–gel solution.
The first organosiloxane sol was prepared by combining
GPTMS, TEOS and 2-propanol in a 1:1:2 volume ratios and
then stirring for 30 min at room temperature; 5 vol.% of the
acidified water or the solution of Ce(NO3 )3 ·6H2 O (in the case
of the hybrid matrix doped with inhibitor) was added to the
mixture at the beginning of the stirring.
The second sol containing the zirconia nanoparticles was
obtained hydrolyzing TPOZ precursor. The sol was prepared
by the addition of ethylacetoacetate to the TPOZ, 70% in
n-propanol, in a 1:1 volume ratio. This solution was stirred
with the synchronous ultrasonic agitation at room temper-
ature for 80 min. Acidified water (pH ∼ 1) or solution of
Ce(NO3 )3 ·6H2 O was added to the sol 20 min after the begin-
ning of the solution stirring. The amount of acidified water
was about 10 vol.%. Concentration of the nitrate in acidified
Mn Si Ti Zn Other Al
0.3–0.9 0.5 0.15 0.25 0.15 Balance
210 M.L. Zheludkevich et al. / Electrochimica Acta 51 (2005) 208–217
Table 2
Types of sol–gel coatings
Coating reference A B C D
ZrO2 nanoparticles + + + +
Cerium in the hybrid matrix − + − −
Cerium in the nanoparticles + − + −
Concentration of cerium in the 0.5 0.5 1 0
final coating (wt.%)
water was varied in order to achieve 0.5 wt.% or 1 wt.% of
cerium ions in final sol–gel coating.
The two prepared sol–gel solutions were mixed in 2:1 vol-
ume ratio, respectively. The final sol–gel system was stirred
and ultrasonically agitated for 60 min and then aged for 1 h at
room temperature before the pre-treatment of the substrates.
A reference coating using only the first sol stirred for
90 min was prepared in order to compare its performance
with that of the doped coatings. The different types of coat-
ings used in this work are shown in Table 2.
The sol–gel films were produced by a dip-coating proce-
dure conducted by immersion of the clean substrate in the
final sol–gel solution for 100 s, followed by controlled with-
drawal with a speed of 18 cm/min. After coating application,
the samples were cured at 130 ◦ C for 1 h in an oven.
2.2. Experimental techniques
Electrochemical impedance spectroscopy (EIS) was used
to estimate the electrochemical parameters associated with
the corrosion process occurring on the AA2024-T3 pre-
treated with the different hybrid sol–gel coatings during im-
mersion in 0.005 M NaCl solution. Low concentration of
the chloride ions was used in order to decrease rate of the
corrosion processes since the sol–gel film is not a complete
coatings and they are working only as a pre-treatment. The
decreased rate of the corrosion allows more correct esti-
mation of the processes on the early stages. However, as
shown below, such concentration of chlorides is perfectly
enough to cause drastic corrosion impact to the untreated
AA2024-T3 substrate or the alloy coated with less protective
Fig. 1. SEM micrographs for AA2024-T3 alloy coated with film A: (a) plane view, (b) cross-sectional view at 40◦ .
pre-treatments. Use of such corrosive environments to study
early stages of pre-treated metallic substrates is described
E elsewhere [16,17]. All the measurements were carried out
− at room temperature in a Faraday cage. A three-electrode
−
−
arrangement was used, consisting of a saturated calomel ref-
0 erence electrode, a platinum foil as counter electrode and
the exposed sample as a working electrode. The working
area was 3.4 cm2 . The EIS measurements were performed
using a Gamry FAS2 Femtostat with a PCI4 Controller. The
impedance measurements were carried out in a 100 kHz down
to 10 mHz frequency range with 10 steps per decade. All the
spectra were recorded at open circuit potential with applied
10 mV sinusoidal perturbation. At least two samples prepared
for the same conditions were tested in order to ensure re-
producibility of the results. The impedance plots were fitted
using different equivalent circuits in order to simulate the
different stages of the pre-treated system during immersion.
The structure of the sol–gel coatings before and after im-
mersion was studied by SEM (Hitachi S-4100 system with
electron beam energy of 25 keV) and EDS (energy disper-
sive spectroscopy). The chemical composition of the sol–gel
films was determined by XPS using a Microlab 310 (VG
Scientific) equipped with a Mg (non-monochromated) anode
and a concentric hemispherical analyzer. The spectra were
taken at constant analyzer mode (CAE = 30 eV).
The morphology of the sol–gel films was assessed by
atomic force microscopy using a nanoscope digital instru-
ments fitted with a NanoScope III controller and a silicon tip
covered with PtIr5 in taping mode to prevent scratching of
the pre-treated surface.
3. Results and discussion
3.1. Structure and composition of the films
The SEM and AFM techniques were used in order to as-
sess porosity and structure of the different coatings. Fig. 1
presents the SEM micrograph of the AA2024-T3 treated with
the sol–gel film A. The surface of the sol–gel film seems to
 M.L. Zheludkevich et al. / Electrochimica Acta 51 (2005) 208–217
Fig. 2. Topography of the hybrid sol–gel film A (AFM).
be crack-free, containing few pinhole-like defects (Fig. 1a).
The thickness of the film was estimated by the cross section
analysis (Fig. 1b) and is about 400–500 nm.
The AFM scan presented in Fig. 2 demonstrates the sur-
face morphology of the sol–gel coated substrate. The nano-
sized particles incorporated into the film matrix are clearly
defined in this scan. The distribution of the nanoparticles is
relatively uniform and the particles have an average diame-
ter around 50 nm. Several particles in the hybrid matrix have
a bigger diameter and seem to result from agglomerates of
smaller ones.
The composition of the hybrid films was investigated us-
ing the XPS technique. Fig. 3 depicts the XPS photoelectron
spectra for Zr 3d and Ce 3d obtained on different coatings.
The spectra were corrected for the C 1s at 285.0 eV. The re-
sults evidence that the binding energy of the Zr 3d5/2 peaks
depends on the concentration of cerium ions. For the coatings
without cerium nitrate addition the binding energy of the Zr
3d5/2 photoelectron spectra is around 182.8 eV, being char-
acteristics of the presence of ZrO2 . However, the Zr 3d5/2
ionization shifts towards lower binding energy (182.1 eV)
for the coatings containing 1% of cerium nitrate. This result
may be due to adsorption and/or absorption of cerium ions
on zirconia nanoparticles. Simultaneously the Ce 3d5 photo-
electron spectrum becomes evident. The well-defined cerium
peak appears on the XPS spectrum only in the case of the sam-
ple with higher concentration of inhibitor ions. The Ce 3d5/2
peak is composed by two components corresponding to Ce3+
and Ce4+ species, with predominance of the former species.
The peak at 916 eV, which is normally associated with the
Ce4+ oxidation state, was not revealed on spectra probably
due to the cerium species concentration lower than detection
limit of the XPS technique. The absence of the well-defined
916-eV photoelectron peak in the case of several Ce (IV)
compounds was found and discussed elsewhere [18]. The Ce
3d5/2 peaks are very similar to those from the correspond-
ing reference salts. Therefore, no significant changes in the
oxidation state of the cerium occur during sol–gel synthesis
211
Fig. 3. Zr 3d (a) and Ce 3d (b) photoelectron spectra for different sol–gel
coatings.
[18]. Cerium could not be detected after immersion, showing
that cerium ions are released from the sol–gel coating.
3.2. EIS measurements
3.2.1. Equivalent circuit models
The impedance spectra can be used to provide adequate
modeling of the physicochemical processes on the coated
substrate during corrosion tests. Fig. 4 presents the evolution
of the impedance spectrum of coating D (without cerium) af-
ter different periods of immersion in the NaCl solution. After
7 h of immersion the impedance spectra reveal the presence
of two time constants at 4 × 103 and 0.1 Hz, which can be
ascribed to the sol–gel film capacitance, Ccoat and than in-
termediate layer with capacitance, Coxide , respectively. The
intermediate conversion layer is formed due to interaction
of Al OH groups with Si OH forming Al O Si covalent
bonds as described elsewhere [4]. This oxide film is intact and
almost does not show a resistive response at low frequencies
after 7 h of immersion. At this stage only two resistive parts
are observed in the impedance plot. These appear at 105 and
50 Hz, describing the resistance of the solution (Rsolut ) and
the pore resistance of the sol–gel film (Rcoat ), respectively.
The equivalent circuit used for numerical fitting of the exper-
212 M.L. Zheludkevich et al. / Electrochimica Acta 51 (2005) 208–217
Fig. 4. Bode plots for AA2024 alloy coated with film D immersed for dif-
ferent periods.
imental Bode plots during the initial stages is presented in
Fig. 5A. The increase of the immersion time leads to the de-
crease of the pore resistance of sol–gel film due to formation
and growth of new cracks and pores.
A new well-defined resistive part (Roxide ) appears at low
frequencies after approximately 50 h resulting from cracks
that are formed in the intermediate oxide layer as schemat-
ically exhibited in Fig. 5B. Formation of cracks in the in-
termediate oxide layer is clearly demonstrated in the SEM
micrograph (Fig. 6). Defects in the oxide layer appear inside
the cracks of the hybrid coating. The EDS measurements (not
shown) prove that cracks in intermediate layer appears in the
places of intermetallic particles. In result the electrolyte can
ingress directly to the substrate forming conductive pathways
and causing the low-frequency resistive response (Roxide ).
The first signs of the corrosion process were detected by
a developing additional time constant at 0.1 Hz after 60 h of
immersion. This time constant is clearly shown in Fig. 4 af-
ter 196 h of immersion. Thus, two additional elements were
added to the equivalent circuit in order to account for the
double layer capacitance (Cdl ) and polarization resistance
(Rpolar ) of the corroded areas. The main corrosion process
of AA2024-T3 in chloride-containing medium is the forma-
tion and growth of pits as exhibited in Fig. 7a. The pits usually
start at zones around the intermetallic particles which have
a nobler potential than the potential of the aluminium alloy
matrix [19]. Dealloying of these intermetallics occurs during
the initial corrosion stages due to oxidation of Mg and Al.
The result is the formation of honeycomb-like metallic ag-
gregates inside the pits. The EDS measurements presented in
Fig. 7b clearly demonstrate that these metallic aggregates are
very rich in copper. The copper aggregates in the pits have
sufficiently nobler potential and can act as effective cathodes
promoting pitting initiation and growth. Formation of such
large copper aggregates can be achieved only due to copper
redeposition processes. This statement is in good accordance
with previous results described the process of pits forma-
tion in aluminium alloys [20]. The honeycomb-like structure
of copper aggregates (Fig. 7a, insert) provides a diffusion-
controlled pathway for the corrosion products and corrosive
agents; thus, a new time constant in the impedance spectrum
at low frequency become visible (Fig. 4, 214 h). Addition of
a Warburg element (W) to the equivalent circuit (Fig. 5D)
can provide an adequate model for the impedance spectrum
at this stage. The mass transfer limitations during pit growth
in the case of the aluminium alloys were described elsewhere
[21].
3.2.2. Influence of the type of coating on the different
components of the coating/metal interface
The low frequency impedance is dependent on the coat-
ing and on the immersion time. During the initial stage of
immersion all the coatings showed identical impedance val-
ues. However the impedance decreases with time. The small-
est changes were revealed for the coating with Ce-doped
nanoparticles. Fig. 8 presents EIS spectra of these coatings
after 250 h of continuous immersion in 0.005 M NaCl solu-
tion. Impedance spectra in Fig. 8 were fitted using the equiv-
alent circuits 5B and 5D for coatings A and D, respectively.
Constant phase elements (CPE) instead of capacitances were
used in all fittings presented in the work. Such modification
is obligatory when the phase angle of capacitor is different
from −90◦ .
The impedance of the CPE depends on frequency accord-
ing to the following equation:
1
= Q(jω)n , (1)
Z
where Z is the impedance, Q a parameter numerically equal to
the admittance (1/|Z|) at ω = 1 rad s−1 , ω the frequency, and
n ≤ 1 a power coefficient calculated as ratio of phase angle
at maximum of corresponding time constant to −90◦ . The
capacitance values for the different elements in the equivalent
circuit were calculated using the following equation:
C = Q(ωmax )n−1 , (2)
where ωmax is the frequency at which the imaginary
impedance reaches a maximum for the respective time con-
stant [22]. The equivalent circuits were chosen for fittings
basing on number of time constants and analyzing the good-
ness of fits. Table 3 presents parameters of the coatings A
and D after 5 and 250 h of continuous immersion. The used
equivalent circuits provide the adequate fitting of the exper-
imental results and the good estimation of parameters. Fit-
ting the experimental results with an incompatible equiva-
lent circuit gives very high value of deviation (last column of
 M.L. Zheludkevich et al. / Electrochimica Acta 51 (2005) 208–217 213

Fig. 5. Schematic representation of the different stages of defects formation and respective equivalent circuits.

Table 3) and cannot be used for adequate estimation of sol–gel
films.
The impedance of the aluminium alloy coated with the hy-
brid film containing Ce-doped zirconia nanoparticles (coat-
ing A) still shows high impedance value (above 106  cm2 )
at low frequencies. Two maxima (two time constants), one
at 5 × 104 Hz and another at 1 Hz can be seen in the phase
angle plot, which can be ascribed to the sol–gel film and
an intermediate mixed oxide layer formed on the alloy/film
interface (as described above), respectively. Contrasting to
this behaviour the coating without cerium (coating D) shows
impedance values at a frequency of 10−2 Hz about five times
214 M.L. Zheludkevich et al. / Electrochimica Acta 51 (2005) 208–217

Fig. 8. Bode plots with respective fittings for coatings A and D after 250 h
of immersion in chloride solution.

Fig. 6. Micrograph of typical crack in the intermediate layer.
lower. The impedance of the film with the Ce-doped ma-
trix (not shown) reveals, at the same frequency, intermedi-
ate values of impedance. Additionally in the case of the last
two films, the low-frequency (0.1 Hz) time constant appears
due to starting of the corrosion activity as described above.
These results show that the hybrid sol–gel films prepared with
cerium-containing zirconia nanoparticles confer higher cor-
rosion protection than the undoped ones and those with the
doped hybrid matrix.
The evolution of parameters of the coated systems was
analyzed in order to assess the corrosion protection prop-
erties of the different pre-treatments. The parameters of the
electrolyte/sol–gel film/substrate system were obtained us-
ing the fits of experimental spectra with equivalent circuits

Fig. 7. Micrograph of pit (a) and EDS spectra (b) of two different zones.

Table 3
Parameters of the sol–gel film/substrate systems obtained from fitting of the experimental impedance spectra with different equivalent circuits
Coating A Coating D

Immersion time (h) 5 250 5 250 250

Equivalent circuit 5B 5B 5B 5D 5C
Rsolut ( cm2 ) 1135 ± 17 227 ± 12 307 ± 8 231 ± 16 256 ± 9
Rcoat. ( cm2 ) 1424 ± 256 421 ± 20 14230 ± 172 375 ± 25 293 ± 14
Qcoat (nS cm−2 ) 371 ± 30 475 ± 180 253 ± 13 487 ± 266 81 ± 36
ncoat 0.755 ± 0.009 0.745 ± 0.036 0.743 ± 0.005 0.744 ± 0.052 0.903 ± 0.042
Roxide (k cm2 ) 2392 ± 66 1170 ± 80 28790 ± 4800 128 ± 10 258 ± 4
Qoxide (␮S cm−2 ) 3.95 ± 0.05 4.30 ± 0.15 4.30 ± 0.15 5.01 ± 0.24 7.76 ± 0.09
noxide 0.938 ± 0.005 0.958 ± 0.006 0.913 ± 0.036 0.913 ± 0.008 0.829 ± 0.003
Rpolar (k cm2 ) – – – 95 ± 32 –
Qdl (␮S cm−2 ) – – – 18.62 ± 2.88 –
ndl – – – 1.001 ± 0.150 –
W (␮S cm−2 ) – – – 36 ± 6 –
Goodness 1.896e−3 949.2e−6 2.402e−3 327.4e−6 12.46e−3
 M.L. Zheludkevich et al. / Electrochimica Acta 51 (2005) 208–217 215

Fig. 9. Evolution of the capacitance for the sol–gel film for the different Fig. 10. Evolution of the pore resistance of the sol–gel film for the different
coatings during immersion. coatings during immersion.

presented in Fig. 5. Only models obtained highest goodness Only after 150 h the pore resistance of the coating B (with
were used to fit experimental results and to extract parameters doped matrix) is decreasing faster than coatings A and D,
of the investigated systems. that is in a good accordance with the evolution of the capac-
itance for film B. The pore resistance of the sol–gel film C is
3.2.2.1. Sol–gel film. The change of the sol–gel film capac- dropping very fast showing that high concentration of cerium
itance during immersion in the chloride solution is presented nitrate leads to formation of a fragile film with poor barrier
in Fig. 9. Generally the capacitance of the dielectric film properties. Therefore the addition of the cerium nitrate seems
depends on the amount of absorbed water. Coating D pre- to negatively affect the barrier properties of the sol–gel layer.
pared without addition of cerium inhibitor shows very stable However incorporation of cerium into zirconia nanoparticles
capacitance for 350 h. The sol–gel film A with Ce-doped can reduce this negative effect.
nanoparticles is also relatively stable; however the capac-
itance increases after around 200 h probably due to water 3.2.2.2. Intermediate oxide layer. The compactness of the
absorption. The capacitance of the coating B (with Ce-doped intermediate oxide layer formed between the sol–gel film and
matrix) exhibits significantly faster growth. Increase of the the metallic substrate is also very important from the point
sol–gel film water uptake can be due to higher concentration of view of corrosion protection, since breakdown of this film
of the cerium inhibitor in the sol–gel matrix, since cerium provide direct ingress of the corrosive agents to the metallic
nitrate was introduced directly into the organosiloxane sol surface. Fig. 11 presents the evolution with time of the resis-
in contrast to the film A, where the inhibitor was added to tance of the intermediate mixed oxide layer (Roxide ). The ref-
the sol with zirconia nanoparticles. Probably the cerium ions erence coating E prepared using only the hybrid organosilox-
can decrease stability of the hybrid film leading to faster ane sol without addition of ZrO2 nanoparticles and cerium-
hydrolytic destruction in the aqueous solutions. based inhibitor shows very fast decrease of the resistance of
Another important parameter is the pore resistance (Rcoat ) the intermediate oxide layer from 4 × 106 to 5 × 103  cm2
of the sol–gel layer that characterizes the crack ability and
porosity of the hybrid film. The pore resistance consists of
the resistance of electrolyte in pores, cracks and pits con-
nected in parallel to each other. Fig. 10 shows the evolution
of the pore resistance of sol–gel films. The undoped film D
has the highest pore resistance for the beginning of immer-
sion; however the resistance decreases during the first 25 h
two orders of magnitude showing formation of cracks in the
sol–gel layer. The initial pore resistance of the cerium-doped
coatings A, B and C is lower in comparison to coating D es-
pecially in the case of the film C with higher concentration
of inhibitor ions. However, short after immersion the pore re-
sistances show fast increase of about one order of magnitude
probably due to blocking of the pores as described in previ-
ous works [5,23]. Coatings A and B after the initial increase
achieve pore resistance values close to the ones for undoped Fig. 11. Evolution of the resistance of the intermediate oxide layer for the
coating and keep similar values during further weathering. different coatings during immersion.
216 M.L. Zheludkevich et al. / Electrochimica Acta 51 (2005) 208–217
Fig. 12. Evolution of the polarization resistance for the different coatings
during immersion.
after only 24 h of immersion. The AA2024 samples pre-
treated with the sol–gel system C with high concentration of
cerium show slower decrease of the resistance. The maximum
initial resistance of the intermediate layer was found for the
Ce-free coating D. However the resistance is dropping rela-
tively fast and achieves 105  cm2 after 330 h of immersion.
The coating B (with cerium-doped organosiloxane matrix)
shows a stable value for the intermediate layer impedance
during the first 100 h of weathering tests. After the initial
stable period the resistance of the oxide conversion layer is
decreasing relatively fast from 107 to 105  cm2 that is close
to the value of the undoped D coating. The stable resistance
of the intermediate layer for 100 h can be obtained due to
the inhibiting and self-healing action of the cerium ions. The
cerium can suppress the cathodic reaction [12] preventing
the propagation of cracks in the intermediate layer. When the
cerium ions are completely released from the sol–gel matrix
the resistance value decreases rapidly.
The resistance of the oxide layer for coating A (with
cerium-doped nanoparticles) is about three times lower at the
beginning of immersion than for film B and behaves differ-
ently with immersion. The resistance slowly decreases dur-
ing 350 h of immersion from 5 × 106 to 5 × 105  cm2 and
finally reaches a five times higher value than coatings B and
D. This behaviour can be explained in terms of prolonged
release of the cerium ions from the zirconia nanoparticles.
As a result the coating A prepared with zirconia nanoparti-
cles doped with cerium inhibitor shows higher stability of the
oxide intermediate layer and confers better corrosion protec-
tion properties from the point of view of long term corrosion
protection.
3.2.2.3. Corrosion process. The rate of the corrosion pro-
cesses can be estimated measuring the polarization resis-
tance (Rpolar ). Fig. 12 shows the evolution of the polariza-
tion resistance during the immersion tests for the different
hybrid sol–gel coatings. The impedance spectra of AA2024-
T3 coated with coating A shows the first signs of third time
constant ascribed to the corrosion processes only at the final
stage of the immersion tests. Thus, the polarization resistance
in this case can be measured only after 275 h of immersion.
The value of the polarization resistance is at least one order
of magnitude higher than in the case of the other coatings.
The impedance spectra of coating B (cerium in matrix) show a
measurable response of the corrosion processes after 150 h of
immersion, which is in a good agreement with the evolution
of the resistance of both the intermediate layer and the sol–gel
film. The polarization resistance for the alloy coated with film
B is around 106  cm2 at this stage and fast decreases with
immersion. Decrease of the polarization resistance can start
when the most part of the cerium ions is already released from
the coating and the inhibition of the corrosion processes is
discontinued. In the case of the cerium-free coating D the
corrosion processes appear twice faster after 75 h. The coat-
ing C shows the first signs of corrosion after only 24 h of
immersion. The polarization resistance in this case is suffi-
ciently lower due to the poor barrier properties of this film.
The reference coating E has higher corrosion activity and
consequently low polarization resistance immediately after
the beginning of the immersion tests.
The above results demonstrate significant improvement of
the corrosion protection due to incorporation of the cerium-
inhibitor ions into the zirconia nanoparticles.
4. Conclusions
The EIS method can be used to model the metal/coating
interface of the sol–gel pre-treatments on aluminium alloys.
The corrosion resistance of the coating and the kinetics of the
corrosion process in chloride solution were evaluated for the
AA2024-T3 coated with hybrid sol–gel films.
Incorporation of zirconia nanoparticles leads to improve-
ment of the barrier properties of the organosiloxane hybrid
sol–gel coatings. Additional corrosion protection is conferred
by doping the sol–gel film with cerium-based inhibitor. How-
ever the too high concentration of cerium nitrate leads to
degradation of the barrier properties of the sol–gel matrix.
The hybrid sol–gel coatings with incorporated zirconium
oxide nanoparticles doped with cerium inhibitor provide
long term corrosion protection and can be prospective candi-
dates for development of new environmentally friendly pre-
treatments.
The nanostructured zirconia particles play the role of
nanoreservoirs for storage and controllable release of the in-
hibitor.
References
[1] J.H. Osborne, Prog. Org. Coat. 41 (2001) 280.
[2] T.L. Metroke, R.L. Parkhill, E.T. Knobbe, Prog. Org. Coat. 41 (2001)
233.
 M.L. Zheludkevich et al. / Electrochimica Acta 51 (2005) 208–217
[3] G.P. Thim, M.A.S. Oliveira, E.D.A. Oliveira, F.C.L. Melo, J. Non-
Cryst. Solids 273 (2000) 124.
[4] H. Schmidt, S. Langenfeld, R. Naß, Mater. Des. 18 (1997)
309.
[5] M.L. Zheludkevich, R. Serra, M.F. Montemor, I.M. Miranda Salvado,
M.G.S. Ferreira, Corrosion protective properties of nanostructured
sol–gel hybrid coatings to AA2024-T3, Surf. Coat. Technol., in press.
[6] R.L. Twite, G.P. Bierwagen, Prog. Org. Coat. 33 (1998) 91.
[7] R.G. Buchheit, H. Guan, S. Mahajanam, F. Wong, Prog. Org. Coat.
47 (2001) 174.
[8] E. Almeida, L. Fedrizzi, T.C. Diamantinio, Surf. Coat. Technol. 105
(1998) 97.
[9] B.F. Rivera, B.Y. Johnson, M.J. O’Keefe, W.G. Fahrenholtz, Surf.
Coat. Technol. 176 (2004) 349.
[10] M. Dabala, E. Ramous, M. Magrini, Mater. Corros. 55 (2004)
381.
[11] J.H. Osborne, K.Y. Blohowiak, S.R. Taylor, C. Hunter, G. Bierwagen,
B. Carlson, D. Bernard, M.S. Donley, Prog. Org. Coat. 41 (2001)
217.
[12] A.E. Hughes, J.D. Gorman, P.R. Miller, B.A. Sexton, P.J.K. Paterson,
R.J. Taylor, Surf. Interface Anal. 36 (2004) 290.
217
[13] N.N. Voevodin, N.T. Grebasch, W.S. Soto, F.E. Arnold, M.S. Donley,
Surf. Coat. Technol. 140 (2001) 24.
[14] M.F. Montemor, W. Trabelsi, A.M. Cabral, M.G.S. Ferreira, The
corrosion resistance of hot dip galvanised steel and aluminium al-
loys pre-treated with bis-[triethoxysilylpropyl] tetrasulfide solutions
doped with Ce(NO3 )3 , Corros. Sci., in press.
[15] A.N. Khramov, N.N. Voevodin, V.N. Balbyshev, M.S. Donley, Thin
Solid Films 447 (2004) 549.
[16] M.G.S. Ferreira, R.G. Duarte, M.F. Montemor, A.M.P. Simões, Elec-
trochim. Acta 49 (2004) 2927.
[17] W. Trabelsi, L. Dhouibi, E. Triki, M.G.S. Ferreira, M.F. Montemor,
Surf. Coat. Technol. 192 (2005) 284.
[18] L.S. Kasten, J.T. Grant, N. Grebasch, N. Voevodin, F.E. Arnold,
M.S. Donley, Surf. Coat. Technol. 140 (2001) 11.
[19] R.G. Buchheit, R.P. Grant, P.F. Hlava, B. McKenzie, G.L. Zender,
J. Electrochem. Soc. 144 (1997) 2621.
[20] Z. Szklarska-Smialowska, Corros. Sci. 41 (1999) 1743.
[21] D.W. Buzza, R.C. Alkire, J. Electrochem. Soc. 142 (1995) 1104.
[22] C.S. Hsu, F. Mansfeld, Corrosion 57 (2001) 747.
[23] X.F. Yang, D.E. Tallman, V.J. Gelling, G.P. Bierwagen, L.S. Kasten,
J. Berg, Surf. Coat. Technol. 140 (2001) 44.