- Email: [email protected]
Synthesis and characterisation of thin cerium oxide coatings elaborated by cathodic electrolytic deposition on steel substrate
Surface & Coatings Technology 200 (2006) 4636 – 4645 www.elsevier.com/locate/surfcoat
Synthesis and characterisation of thin cerium oxide coatings elaborated by cathodic electrolytic deposition on steel substrate J. Creus*, F. Brezault, C. Rebere, M. Gadouleau Laboratoire d’Etudes des Mate´riaux en Milieux Agressifs, L.E.M.M.A. (EA 3167), Baˆtiment Marie Curie, Avenue Michel Cre´peau, F-17042 La Rochelle cedex 01, France Received 11 February 2005; accepted in revised form 7 April 2005 Available online 13 June 2005
Abstract Cerium oxide films are widely studied as a promising alternative to Cr(VI) based pre-treatments for the corrosion protection of different metals and alloys. This paper deals with the deposition and characterisation of cerium oxide films deposited onto mild steel substrates by cathodic electrolytic deposition via Ce(III) chloride aqueous and mixed water – ethyl alcohol solutions with and without hydrogen peroxide as a precursor. The major role of this precursor is to increase the deposition rate of the oxide films allowing smooth and adherent hydrated cerium oxide films to be obtained. However, the evolution of the composition of the bath with time leads to a non-reproducibility of the morphology and composition of the thin films. The deposited oxide films are composed of small particles of hydrated Ce(IV) oxide with a significant amount of amorphous phase. In the absence of a precursor, only the mixed water – ethyl alcohol solution permits cerium oxide films to be deposited. The influence of the deposition parameters such as applied cathodic current density, deposition time and composition of the bath was also investigated. The deposits exhibit a quite developed crack network, which is highly dependent on the deposition parameters and the drying process. Moreover, the incorporation of chloride ions during the deposition has been observed. The protection efficiency of these films was examined during extended immersion tests in 3% NaCl solution. The deposits allow the cathodic reaction kinetics to be reduced by acting as a cathodic inhibitor. However, the presence of large defects or cracks induces an accelerated local degradation of the steel substrate, removing partially the oxide deposit, whereas a thin crack network can rapidly be sealed by corrosion products without deterioration of the deposit. D 2005 Elsevier B.V. All rights reserved. Keywords: Cerium oxide film; Cathodic electrodeposition; Protection efficiency
1. Introduction Chromium(VI) compounds are largely used to create a conversion layer that provides an improved corrosion resistance for many metallic alloys. The use of rare earth (RE) salts is nevertheless studied as an environmentally friendly alternative to the use of Cr(VI) based pre-treatments. An enhancement of the corrosion resistance for different alloys such as aluminium alloys [1 –3], zinc alloys [4 –8] or steels [9– 11] has been reported. The different mechanisms suggested in literature show that the role of the rare earth conversion films is not well understood. The main * Corresponding author. Tel.: +33 5 46 45 72 94; fax: +33 5 46 45 72 72. E-mail address: [email protected] (J. Creus). 0257-8972/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.surfcoat.2005.04.027
mechanism of action for the RE elements seems to be based on the reduction of the rate of the cathodic process acting as cathodic inhibitors [1,6]. However, Montemor et al. [5] recently showed that RE elements act as mixed inhibitors, reducing both cathodic and slightly anodic activity on hot dip galvanised steel substrates. Ceria (CeO2) has a considerable interest as a promising material in corrosion protection. Different techniques such as laser beam evaporation ion [12], chemical vapour deposition, sol – gel processing, immersion [1 – 9] and electrodeposition [10,11,13 –15] have been used to obtain cerium oxide films. Electrodeposition is an attractive method for the preparation of thin films since it offers the advantage of low processing temperature, controlled thickness of the film and low cost process. Cerium oxide films
J. Creus et al. / Surface & Coatings Technology 200 (2006) 4636 – 4645
have been obtained by cathodic or anodic deposition on different substrates [1,10,11,13– 18]. In the cathodic electrolytic deposition, hydroxide ions are formed at an electrode surface by a cathodic process and metal ions or complexes are then hydrolysed by the electrogenerated base. Zhou and Switzer [11], Zhitomirsky and Petric [13,14] and Arurault et al. [9] have used a solution of Ce(III) nitrate in order to deposit thin oxide films. The reduction of nitrate ions into nitrite or ammonia ions also produces hydroxide ions, which promote a rapid increase of the pH at the interface. The electrogenerated base induces the formation of colloidal particles of Ce(III) hydroxide or hydrous oxide [11,13,14], or also the formation of a ionic species [Ce(OH)22+], that can be hydrolysed into ceria. Different authors [13,14,19] suggest that the use of a precursor like hydrogen peroxide (H2O2) prevents the accumulation of nitrite or ammonia ions that provokes deposition instabilities. Zhitomirsky and Petric [13] have shown that, using a Ce(III) chloride solution, associated to hydrogen peroxide precursor, leads to more adherent and uniform oxide films than Ce(III) nitrate solution. The aim of this study was to characterise ceria coatings elaborated on a carbon steel substrate by cathodic electrolytic deposition from Ce(III) chloride aqueous and mixed water –ethyl alcohol solutions. The influence of hydrogen peroxide precursor is discussed and the corrosion resistance of coated steel is investigated in saline solution through extended immersion tests.
2. Experimental procedure Electrochemical experiments were carried out in a classical three-electrode glass cell. The working electrode was a 36NiCrMo16 (EN 1083-1) steel disk with an area of 0.78 cm2. Before each experiment, the steel surfaces were polished with silicon carbide (particle size 5 Am), sonicated for 2 min, rinsed thoroughly with Milli-Q water and dried. The reference electrode was a saturated calomel electrode (SCE). The counter electrode was a large platinum grid. The experiments were performed at room temperature with an AMETEK 263 A potentiostat system, driven by Softcorr 3 software. The steel electrodes were left at the open circuit potential for 30 s prior to the potential scans. The cerium oxide films were deposited by cathodic electrolytic deposition (CELD) via two electrolytic baths of CeCl3I7H2O. Their compositions are presented in Table 1. Table 1 Composition of the (A) aqueous and (B) mixed water – ethyl alcohol (1:9 volume ratio) solution for the deposition of thin solid films Composition
CeCl3I7H2O H2O2 (30 %)
Aqueous solution
Mixed water – ethyl alcohol solution
A1
A2
B1
B2
B3
0.8 mM
0.8 mM 25 mM
0.8 mM
10 mM
0.8 mM 25 mM
4637
Table 2 Deposition conditions of thin ceria films from both A and B solutions Conditions
Aqueous solution A1
A2
B1 and B3
B2
Duration (min)
10 20 0.005 0.01 0.1
10 20 0.005 0.01 0.1
10 20 0.01 0.1 0.25 1 5
10
j (mA/cm2)
Mixed water – ethyl alcohol solution
0.1 0.25
All the reagents were Sigma-Aldrich maximum purity chemicals. Cyclic voltammetry, performed from 0.55 V/SCE to 1.8 or 2.5 V/SCE at the scan rate of 20 mV/s, was used to characterise the electrochemical reactions. The deposits were obtained in galvanostatic mode in unstirred solutions at room temperature. The deposition conditions are summarised in Table 2. After deposition, the samples were carefully washed in order to prevent the detachment of the thin films and dried during 24 h in air at room temperature. The morphology of the samples was studied using optical microscope and scanning electron microscope (SEM JEOL 5410 Low Vacuum), coupled to X-ray energy dispersive spectroscopy (EDS). The cerium oxides were also analysed by X-ray diffraction (XRD) with a classical powder diffractometer (Bruker AXS D8-Advance), using Cu-Ka1 wavelength (k = 1.5406 nm) in Bragg-Brentano geometry. Raman spectra were recorded with a LabRam HR spectrometer, equipped with a confocal microscope, using an incident beam of 632.82 nm emitted by an argon laser. Differential scanning calorimetry (DSC TA Instrument Q100) was carried out in air between 40 and 550 -C at a heating rate of 20 -C/min. UV – visible absorption spectra of the tested Ce(III) chloride aqueous and mixed water –ethyl alcohol-based solutions containing H2O2 as a precursor were obtained using a Helios h-UNICAM spectrometer. The protection efficiency of the coatings was investigated via experiments performed in aerated and stirred 3% NaCl solution (pH = 6). The open circuit potential (OCP) and polarisation resistance (R p) of the coated substrates versus time were recorded during long immersion tests (around 65 h). The polarisation resistance measurements were estimated via a linear polarisation experiment performed at a scan rate of 0.16 mV/s about T20 mV/OCP. Polarisation curves i(t) were recorded at a scan rate of 0.16 mV/s after 1 h of immersion in saline solution.
3. Results and discussion 3.1. Synthesis and characterisation of cerium oxide powder The synthesis of cerium oxide by precipitation was performed by addition of sodium hydroxide solution in
4638
J. Creus et al. / Surface & Coatings Technology 200 (2006) 4636 – 4645
Ce(III) chloride aqueous solution. The precipitate was filtrated and dried in air during 24 h at room temperature. The same procedure was performed with a Ce(III) chloride solution containing hydrogen peroxide (H2O2) as a precursor. This precursor had been added for complete oxidation of Ce(III) into Ce(IV). The results of the various analyses performed on the precipitate obtained from CeCl3 and NaOH are presented in Fig. 1. The precipitate obtained in the aqueous solution containing H2O2 presents the same characteristics. The XRD pattern (Fig. 1a) is typical of ceria (CeO2) as it is reported in literature [9]. No diffraction lines associated to Ce(III) oxide are present, which seems to be in agreement with [20]. We observe that the line width is large, which is
typical of a small particle size. Raman spectrum (Fig. 1b) presents an intense band at 461 cm1, which corresponds to CeO2 [21 – 23]. Kosacki et al. [21] remarked that the width of the intense band increased when the size of ceria particles decreased. Therefore, the precipitation technique via the Ce(III) chloride solution leads to a precipitate composed of small particles of ceria. Moreover, the bands around 1660 cm1 in the Raman spectra suggest the presence of hydration water molecules. Indeed, the DSC diagram (Fig. 1c), performed between 40 and 550 -C, presents an intense exothermic band centred on 100 -C during the heating step that is attributed to a dehydration mechanism, in agreement with literature [9,18]. Therefore, the precipitates obtained in both solutions correspond to hydrated ceria. The presence of hydrogen peroxide precursor allows to completely oxidise Ce(III) into Ce(IV) leading to a stoichiometric cerium(IV) oxide. Without this precursor, Ce(III) precipitates into hydrated Ce(OH)3 that is then oxidised into hydrated cerium(IV) oxide either during the increase of the pH of the solution [2] or during the drying process [10,11,18]. 3.2. Elaboration and characterisation of cerium oxide films by CELD Electrochemical mechanism of base electrogeneration during cathodic deposition has been widely discussed in the literature [9 – 18,24 – 26]. In the case of chloride baths, where anions do not participate in any cathodic reaction, the hydroxyl ions are produced by the reduction of dissolved oxygen or water (Eqs. (1) and (2)). O2 þ 2H2 O þ 4e Y4OH
ð1Þ
2H2 O þ 2e YH2 þ 2OH
ð2Þ
The addition of hydrogen peroxide leads to the formation of supplementary hydroxyl ions during the cathodic reaction (Eq. (3)): H2 O2 þ 2e Y2OH
Fig. 1. Characterisation of the precipitate obtained by alkalisation of aqueous Ce(III) chloride solution. (a) XRD pattern, (b) Raman spectrum and (c) DSC diagram of the precipitate dried in ambient air during 24 h at room temperature.
ð3Þ
Zhitomirsky and Petric [13] showed that the role of H2O2 as a precursor was to prevent the formation of a nonstoichiometric cerium oxide film by oxidising Ce(III) into Ce(IV), the latter being more easily hydrolysed. Moreover, this precursor was shown to enhance the adherence and uniformity of the oxide films. The concentration of chloride ions in the solution is progressively decreased by the action of H2O2, but this precursor does not avoid the incorporation of chloride ions into the coating. The formation of hydroxyl ions at the cathode leads to a local increase of the pH at the surface that promotes the formation of a Ce(OH)3 precipitate or/and the formation of a soluble ionic complex Ce(OH)22+ (Eqs. (4) or (5a) and (5b)). Ce(OH)22+ appears preferentially when H2O2 is added in the solution.
J. Creus et al. / Surface & Coatings Technology 200 (2006) 4636 – 4645
Ce3þ þ 3OH YCeðOHÞ3
ð4Þ
Ce3þ þ 2OH YCeðOHÞ2þ 2 þe
ð5aÞ
2Ce3þ þ 2OH þ H2 O2 Y2CeðOHÞ2þ 2
ð5bÞ
The cerium oxide film then results either from the oxidation of Ce(OH)3 (Eq. (6)) or from the hydrolysis of Ce(OH)22+ (Eq. (7)) [9,18]. CeðOHÞ3 YCeO2 þ H3 Oþ þ e
ð6Þ
CeðOHÞ2þ 2 þ 2OH YCeO2 þ 2H2 O
ð7Þ
Recently, Stefanov et al. [10] showed that the electrodeposition of cerium oxide on stainless steel from an aerated Ce(III) chloride solution leads to a mixture of Ce2O3 and CeO2. The electrochemical mechanism of formation of these oxide films is really complex. The electrogenerated base induces the formation of colloidal particles of Ce(III) hydroxide or hydrous oxide near the surface [11,13,14,18]. Li and Thompson [18] suggested that the deposition of the solid film proceeded through a nucleation and growth process. Fig. 2 presents the cyclic voltammetry curves obtained with steel substrates in freshly prepared Ce(III) chloride aqueous (Fig. 2a, baths A1 and A2) and mixed water – ethyl
Fig. 2. Cyclic voltammograms performed in aqueous (a) and mixed water – ethyl alcohol (b) Ce(III) chloride solutions with (- - - -) and without (—– ) H2O2.
4639
alcohol (Fig. 2b, baths B1 and B3) solutions with and without hydrogen peroxide. In aqueous solution (bath A1), the variations of the current density are due to the reactions (1) and (2). The current hysteresis, observed during the reverse sweep, indicates a modification of the electrode surface during the cathodic polarisation. This surface modification is more important when H2O2 is added to the solution (bath A2). The shape of the voltammogram is completely changed due to reaction (3). We can see that water reduction (2) is strongly shifted towards more negative potentials probably due to an important increase of the interfacial pH. Therefore, it appears that the hydrogen evolution reaction (HER) prevents the deposition of cerium oxide films in aqueous solution. In mixed water – ethyl alcohol solution (bath B1), the voltammogram is different from solutions A due to the low amount of water in the solution. The presence of water is necessary to electrogenerate the OH ions that are essential for the formation of the Ce(III) hydroxide particles. The current densities are lower. The addition of H2O2 induces an important increase of the current densities and an important current hysteresis is observed which suggest a beneficial effect of this precursor on the modification of the electrode surface. Therefore, in both baths A and B, the addition of the H2O2 promotes the formation of the cerium oxide films onto the steel substrate as expected. Once dried, the coatings present a dark yellowish colour. The SEM observations reveal that the films are covering and adherent and EDS analyses underline the presence of low Ce amounts. Fig. 3 shows the surface morphology, as observed by SEM, of cerium oxide films deposited after 20 min in aqueous solution containing H2O2 (bath A2) at 0.005 (Fig. 3a) and 0.01 mA/cm2 (Fig. 3b). Many surface defects such as important crack networks and porosities appear in all the coatings. Circular cracks whose diameters can even reach 20 –30 Am are also observed regardless of the applied cathodic current. These circular cracks or craters are associated to either the existence of inclusions or to the formation of gas bubbles, which is promoted during the process on the steel surface when current densities are increased. Raman spectra performed either in the centre of a large circular crater (Fig. 4a) or near the extremity of a crack (Fig. 4b) are typical of corrosion products of steel, namely ferrihydrite 5Fe2O3I9H2O (Fig. 4a) and lepidocrocite g-FeOOH (Fig. 4b). Cerium oxide bands are hindered by the important degradation of the substrate, which may occur during the drying process. Moreover, coating morphology and composition are not reproducible due to the evolution of bath during time, this evolution being more important in aqueous solution than in mixed water –ethyl alcohol solution. For freshly prepared baths (A2 or B3), the films obtained during the cathodic polarisation are essentially composed of cerium oxide, but after 1 day, the appearance of the coating changes and an important degradation of the substrate is observed.
4640
J. Creus et al. / Surface & Coatings Technology 200 (2006) 4636 – 4645
Fig. 4. Raman spectra of cerium oxide films deposited at 0.005 (a) and 0.1 (b) mA/cm2 during 20 min in aqueous Ce(III) chloride solution containing H2O2.
Fig. 3. SEM observations of cerium oxide films deposited at 0.005 (a) and 0.01 (b) mA/cm2 during 20 min in aqueous Ce(III) chloride solution containing H2O2.
Fig. 5 presents the evolution of the pH of the aqueous and mixed water – ethyl alcohol solutions with and without H2O2 precursor. The addition of H2O2 in colourless Ce(III) chloride solutions induces an orange colour associated to the formation of the hydroxide species Ce(OH)22+ as it was reported by Aldykiewicz et al. [26] and Golden and Qi Wang [27]. The pH of the solutions gradually decreases after the addition of H2O2. For the bath A2, after 1 day, the pH remains constant and the solution change progressively into a yellowish colour. This colour modification is associated to the Ce(IV) species, which is confirmed by UV – visible spectroscopy. After a few days, the solution becomes progressively colourless whereas pH remains constant. The UV – visible spectroscopy analyses reveal that the amount the Ce(IV) species is reduced. This suggests that the Ce(IV) species are progressively reduced into Ce(III) into the aqueous media. For the bath B3, concerning the mixed water – ethyl alcohol solution, the same solution evolution is observed when H2O2 is added. The evolution of the pH and the nature of the ionic species in solution explain the non reproducibility of the coating morphology and composition.
In aqueous solution (bath A1), whatever the applied cathodic current densities between 0.005 and 0.6 mA/cm2 and whatever the duration of the experiment, no cerium oxide films are detectable. In mixed water – ethyl alcohol solution (bath B1), the cathodic current densities were varied between 0.01 and 5 mA/cm2. Only cathodic current densities above 0.1 mA/cm2 allowed to obtain cerium oxide films. Fig. 6 presents the SEM surface morphologies of the cerium oxide films deposited in mixed water – ethyl alcohol solution (bath B1) for different cathodic current densities and time deposition. The oxide films present a ribbed morphology, with a crack network that develops during the
Fig. 5. Evolution versus time of the pH of aqueous solutions with (h) or without (r) H2O2 and mixed water – ethyl alcohol solution with (?) or without (4) H2O2.
J. Creus et al. / Surface & Coatings Technology 200 (2006) 4636 – 4645
4641
Fig. 6. SEM observations of ceria films deposited at 0.25 mA/cm2 during 20 (a) and 10 min (b), at (c) 0.5 mA/cm2 during 20 min and (d) 1 mA/cm2 during 10 min in mixed water – ethyl alcohol Ce(III) chloride solution.
drying process, as also reported in the literature [9,13,14]. This crack network increases with the applied current density or with time deposition as observed for coatings deposited at 0.25 mA/cm2 during 10 min (Fig. 6b) and 20 min (Fig. 6a). Therefore, the increase of the coating thickness promotes crack formation during the drying process. For current densities around 0.5 mA/cm2, the coatings become rather inhomogeneous combining zones with large cracks and coating detachments and adherent zones (Fig. 6c). When current densities are above 1 mA/ cm2, the oxide films are not adherent and present large clusters and large cracks (Fig. 6d). Locally, the detachment of the oxide coating during the drying process leaves the steel surface uncovered. A Raman spectrum typical of all the coatings deposited in bath B1 is shown in Fig. 7. It presents a large intense band at 454 cm1 that may correspond to a small particle size [21,23]. The size of the hydroxide or hydrous oxide particles strongly depends on the applied cathodic current density. Zhou and Switzer [11] reported that, during the deposition from aqueous Ce(III) nitrate solution onto stainless steel, the grain size decreased from 18 nm to approximately 6 nm when applied current densities varied from 0.5 to 3 mA/cm2. The enlargement of the intense band can also be attributed to an amorphous character of the
films. The XRD pattern (not presented) of the deposit obtained at 0.1 mA/cm2 during 20 min exhibits a very broad line near 2h å 27-, which suggests that the deposit contains a significant amount of amorphous phase [14]. Therefore, it seems that the deposits are composed of low size particles with a significant amount of amorphous phase. In the following, the influence of the concentration of Ce(III) species in solution is discussed. Fig. 8 presents the SEM surface morphology of an oxide film deposited at 0.1 mA/cm2 during 10 min in bath B2. The coating presents few
Fig. 7. Raman spectra of ceria films deposited at 0.25 mA/cm2 during 10 min in mixed water – ethyl alcohol Ce(III) chloride solution.
4642
J. Creus et al. / Surface & Coatings Technology 200 (2006) 4636 – 4645
a
20 µm 325.5 391.0
b 139.9
716.3
5000
1362.7
4000 1070.1
Intensity (a.u.)
6000
500.1
7000
3000 2000 1000 200
400
600
800
1000
1200
1400
Wavenumber (cm-1)
Fig. 8. SEM Observation (a) and Raman spectrum (b) of ceria film deposited at 0.1 mA/cm2 during 10 min in Ce(III) chloride (10 mM) mixed water – ethyl alcohol solution.
dark spots (Fig. 8a) scattered at the coating surface and the number of these dark spots increases with deposition time. The morphology of these dark zones is totally different from the smooth and adherent oxide film. The Raman spectrum (Fig. 8b) performed at the centre of a dark spot is typical of a corrosion product of steel (ferrihydrite), which suggests a degradation of the steel substrate during the drying process. The corrosion of steel seems to be promoted by the incorporation of chloride ions during the formation of the hydrous oxide film. This incorporation is less pronounced in less CeCl3 concentrated solutions. Thus, the incorporation of chloride ions may depend on the rate of the oxide film formation and the electrical field at the electrode interface. Increasing the concentration of Ce(III) in mixed water – ethyl alcohol solutions promotes the formation of the oxide film but also the incorporation of chloride ions which is harmful for the steel substrate during the drying process. 3.3. Corrosion behaviour of oxide films deposited in mixed water –ethyl alcohol solutions Fig. 9 gathers the results of the experiments performed to estimate the protection efficiency of oxide films compared to bare steel in 3% NaCl solution. The polarisation curves of
Fig. 9. Corrosion behaviour in saline solution of cerium oxide films deposited onto steel substrate compared to bare steel. (a) Polarisation curves of (—– ) bare steel, (- - -) Ce oxide film (bath B1, 0.25 mA/cm2 during 10 min) and (IIII) Ce oxide film (bath B2, 0.25 mA/cm2 during 10 min). (b) Polarisation resistance versus immersion time of (†) bare steel, (?) Ce oxide film (bath B1, 0.25 mA/cm2 during 10 min) and (r) Ce oxide film (bath B2, 0.25 mA/cm2 during 10 min).
two oxide films deposited at 0.25 mA/cm2 during 10 min respectively in baths B1 and B2 are compared to that of bare steel (Fig. 9a). The corrosion potentials and current densities are gathered in Table 3. The current densities are associated to the limiting current densities deduced from the cathodic branch of the polarisation curves. Both oxide coatings reduce the corrosion current densities. The oxide coating deposited in bath B1 induces the lowest decrease of the corrosion current density and we can note that the corrosion potential is shifted towards more negative values. The cathodic curve, corresponding to the reduction of dissolved oxygen is slightly shifted, indicating a slight slowing down of the cathodic reaction kinetics. Moreover, the anodic part of the polarisation
Table 3 Electrochemical characteristics of coated steel compared to bare steel after 1 h of immersion in 3% NaCl solution Bare steel
E corr (mV/SCE) i corr (AA/cm2)
388 190
Oxide film deposited at 0.25 mA/cm2 during 10 min Bath B1
Bath B2
495 120
361 65
J. Creus et al. / Surface & Coatings Technology 200 (2006) 4636 – 4645
curve is quite similar to that of bare steel, which suggests that the oxide film has a limited effect on the dissolution mechanism of steel. For the coating deposited in bath B2, the polarisation curve is quite different. The corrosion current density is reduced and the corrosion potential is shifted towards more positive values. A slight reduction of the cathodic curve is observed which suggests a blocking effect on oxygen reduction reaction. In the anodic part, a rapid increase of current density occurs for very small anodic overpotential, which suggests a local corrosion degradation of the coated steel. Therefore, these oxide films, deposited via a mixed water –ethyl alcohol bath, exhibit an effective reduction of the corrosion current density by acting as a physical barrier against the aggressive media and reducing the kinetic of the reduction of dissolved oxygen [1,5,6]. Open circuit potential (not shown) and polarisation resistance R p (Fig. 9b) of these coatings during extended immersion tests in 3% NaCl solution were also monitored. A linearity relationship of potential/current in the range of T20 mV/OCP was observed which allowed to measure the polarisation resistance for different immersion duration. The open circuit potential of steel rapidly decreases to reach a constant value around 630 mV/SCE after 20 h of
4643
immersion. The open circuit potentials of cerium oxide films are quite similar to that of bare steel. The potential values are nobler, in particular for the coating deposited in bath B2. For bare steel, the R p values are close to 250 V cm2 during the first hours of immersion, and then increase slightly due to the formation of the corrosion products of steel limiting the diffusion of dissolved oxygen. This induces a shift of the open circuit potential towards negative values. For the oxide deposits, the R p values during the first hours of immersion are higher than those of bare steel. The highest values correspond to the deposit obtained in bath B2, which is in agreement with the polarisation curves. The R p value of this coating decreases until 20 h of immersion and then remains constant, whereas the R p evolution of the deposit obtained in bath B1 is similar to that of bare steel but the values are lower. This is more likely due to a local degradation of the substrate. Fig. 10 presents SEM observations of oxide films obtained either in bath B1 (Fig. 10a) or in bath B2 (Fig. 10b) after extended immersion tests. Raman spectra of typical corrosion products observed in both oxide films are presented in Fig. 10c and d. The SEM observations reveal a local attack of the substrate through defects like cracks or porosities in each case. The oxide films are still present and locally prevent the localised degradation of the substrate (Fig. 10a), even if a
b
c
d 453.5
a
3000
375.6
1297.2
648.3
5000
525.5
10000 306.9 347.0
1500
15000
141.9 215.7
1068.4
2000
1366.1
685.1
Intensity (a.u.)
530.7
133.7 177.8 243.9
Intensity (a.u.)
376.3
20000 2500
1000 0 500
1000 Wavenumber (cm-1)
1500
500
1000
1500
Wavenumber (cm-1)
Fig. 10. SEM observations of cerium oxide films deposited onto steel substrate after extended immersion tests. (a) Deposit at 0.25 mA/cm2 during 10 min (bath B1 0.8 mM), (b) deposit at 0.25 mA/cm2 during 10 min (bath B2 10 mM) and Raman spectra of typical corrosion products (c and d) developed on the surface of the samples.
4644
J. Creus et al. / Surface & Coatings Technology 200 (2006) 4636 – 4645
crack network exists. These cracks are rapidly sealed with corrosion products from the steel substrate as far as the crack width is small. Then, the corrosion of the steel substrate is negligible. The oxide coating obtained in bath B2 contains many dark spots that are identified as ferrihydrite. This degradation occurs during the drying process and this corrosion product partially seals the cracks and defects of the oxide film. This explains why this oxide coating presents during the first hour of immersion the highest polarisation resistance values. The SEM observation performed after the extended immersion tests shows the presence of corrosion products that covered again the oxide film and an expansion and modification of the dark spots observed before the immersion. The Raman spectra (Fig. 10c and d) present the main corrosion products observed on both oxide coatings. The spectrum in Fig. 10c is typical of ferrihidrite that is mainly observed in the centre of the cracks, whereas the spectrum in Fig. 10d is quite typical of lepidocrocite gFeOOH that is observed either on the top of the cracks or on the surface. Finally, the broad band at 453 cm1 (Fig. 10c) confirms the presence of the cerium oxide films. It is interesting to underline that certain zones of the coated steel remain unaffected even after extended immersion tests (62 h) in saline solution, in particular for the oxide films deposited in bath B1. So these coatings may ensure a protection of the steel surface but the protection effectiveness strongly depends on the morphology and composition of the oxide films. The crack network that appears during the drying process is harmful for the protection of the substrate during extended immersion tests as it can be observed in Fig. 9b. Therefore, it seems that the deposition conditions should be accurately adjusted to control the morphology of this oxide film. Another important parameter is the incorporation of chloride ions in the deposited hydrous oxide film. In solution B2, the negative effect of this incorporation that induces a local attack of the substrate during the drying process is clearly demonstrated. Incorporation of chloride ions in the hydrous oxide films is assumed to play also a role during the extended immersion tests, in particular by promoting a local acidification at the substrate – oxide interface. A more detailed investigation should therefore be performed to underline the influence of the incorporation of chloride ions in the cerium oxide films on the corrosion mechanism in saline solution and to improve the quality of the coatings.
4. Conclusions Two deposition baths were compared so as to produce cathodic electrodeposits of cerium oxide. Whatever the deposition parameters performed in aqueous Ce(III) chloride solutions, no cerium oxide film could be obtained. It seems that the reduction of dissolved oxygen does not produce a sufficient increase of the interfacial pH necessary to the formation of colloidal Ce(III) hydroxide particles. At
higher applied current densities, the hydrogen evolution reaction seems to prevent the formation of the cerium oxide films. The addition of H2O2 is essential in aqueous baths. In a freshly prepared solution, the rate of deposition is increased via the reduction of H2O2 that produces hydroxyl ions at the interface. Adherent, uniform and homogeneous hydrated cerium oxide films were then obtained. These oxide films are characterised by an intense crack network that develops during the drying process. It strongly depends on the deposition parameters such as applied current density, deposition time and bath composition. However, the ageing of baths containing H2O2 leads to a non-reproducibility of the morphology and composition of the oxide films, and also to local degradations of the steel substrate. So, the addition of H2O2 in electrolytic baths of cerium has to be avoided in steel protection applications and alternative processes are under investigation. With mixed water – ethyl alcohol solutions, cerium oxide films could be obtained without the addition of H2O2. The influence of H2O2 has been studied, leading to results similar to those obtained with aqueous solutions This mixed solution allows the pH increase close to the electrode interface to be localised promoting the formation of colloidal Ce(III) hydroxide particles. The cerium oxide coatings are composed of small particles with a significant amount of amorphous phase. At low current densities, the oxide films are adherent, dense and present a thin crack network. The crack width increases with the deposition time. The increase of the coating thickness promotes the formation of large cracks, which locally induces a delamination of the coating. High current densities or extended deposition times induce non-adherent films, whereas increasing the Ce(III) chloride concentration in solution should promote the incorporation of chloride ions in the hydrous cerium oxide films. This induces localised degradations of steel during the drying process. The best morphologies are obtained at low current densities and low deposition times. The corrosion behaviour of the coated steel was examined in aerated and stirred 3% NaCl solution. The protection efficiency strongly depends on the shape of the crack network and the adherence of the cerium oxide films. The films obtained at low current densities or short deposition times present the best protection efficiency due to the thin crack network. The incorporation of chloride ions into the hydrous cerium oxide films seems to affect the protection efficiency. The behaviour of samples submitted to extended immersion tests is however promising and the cerium oxide films should be considered as a potential alternative for steel protection.
Acknowledgements B. Peraudeau and M. Jeannin are gratefully acknowledged for their technical support during this project.
J. Creus et al. / Surface & Coatings Technology 200 (2006) 4636 – 4645
References [1] B.R.W. Hinton, N.E. Ryan, D.R. Arnott, Mater. Australas. (1987) 18. [2] A.J. Davenport, H.S. Isaacs, M.W. Kendig, Corros. Sci. 32 (1991) 653. [3] W.G. Fahrenholtz, M.J. O’Keefe, H. Zhou, J.T. Grant, Surf. Coat. Technol. 155 (2002) 208. [4] M.G.S. Ferreira, R.D. Duarte, M.F. Montemor, A.M.P. Simoes, Electrochim. Acta 49 (2004) 2927. [5] M.F. Montemor, A.M.P. Simoes, M.G.S. Ferreira, Prog. Org. Coat. 43 (2001) 274. [6] M.A. Arenas, J.J. De Damborenca, Surf. Coat. Technol. 187 (2004) 320. [7] K. Aramaki, Corros. Sci. 46 (2004) 1565. [8] K. Aramaki, Corros. Sci. 43 (2001) 2201. [9] L. Arurault, P. Monsang, J. Salley, R.S. Bes, Thin Solid Films 466 (2004) 75. [10] P. Stefanov, G. Atanasova, D. Stoychev, T.S. Marinova, Surf. Coat. Technol. 180 – 181 (2004) 446. [11] Y. Zhou, J.A. Switzer, J. Alloys Compd. 237 (1996) 1. [12] S. Kanakaraju, S. Mohan, A.K. Sood, Thin Solid Films 305 (1997) 191.
4645
[13] I. Zhitomirsky, A. Petric, Ceram. Int. 27 (2001) 149. [14] I. Zhitomirsky, A. Petric, Mater. Lett. 40 (1999) 263. [15] M. Balasubramaniam, C.A. Melendres, A.N. Mansour, Thin Solid Films 347 (1999) 178. [16] A. Qi Wang, T.D. Golden, J. Electrochem. Soc. 150 (2003) C616. [17] T.D. Golden, A. Qi Wang, J. Electrochem Soc. 150 (2003) C621. [18] E.B. Li, G.E. Thompson, J. Electrochem. Soc. 146 (1999) 1809. [19] Th. Pauporte, D. Lincot, J. Electroanal. Chem. 517 (2001) 54. [20] Y.X. Li, X.Z. Zhou, Y. Wang, X.Z. You, Mater. Lett. 58 (2003) 245. [21] I. Kosacki, T. Suzuki, H.U. Anderson, P. Colomban, Solid State Ion. 149 (2002) 99. [22] J.E. Spanier, R.D. Robinson, Z. Zhang, S.W. Chan, I.P. Hernan, Phys. Rev., B, Condens. Matter Mater. Phys. 64 (2001). [23] S. Wang, W. Wang, J. Zuo, Y. Qian, Mater. Chem. Phys. 68 (2001) 246. [24] J.M. Brossard, J. Balmain, J. Creus, G. Bonnet, Surf. Coat. Technol. 185 (2004) 275. [25] G.H. Annal Therese, P. Vishnu Kamath, J. Appl. Electrochem. 28 (1998) 539. [26] A.J. Aldykiewicz, A.J. Davenport, H.S. Isaacs, J. Electrochem. Soc. 143 (1996) 147. [27] T.D. Golden, A. Qi Wang, J. Electrochem. Soc. 150 (2003) C621.
Synthesis and characterisation of thin cerium oxide coatings elaborated by cathodic electrolytic deposition on steel substrate J. Creus*, F. Brezault, C. Rebere, M. Gadouleau Laboratoire d’Etudes des Mate´riaux en Milieux Agressifs, L.E.M.M.A. (EA 3167), Baˆtiment Marie Curie, Avenue Michel Cre´peau, F-17042 La Rochelle cedex 01, France Received 11 February 2005; accepted in revised form 7 April 2005 Available online 13 June 2005
Abstract Cerium oxide films are widely studied as a promising alternative to Cr(VI) based pre-treatments for the corrosion protection of different metals and alloys. This paper deals with the deposition and characterisation of cerium oxide films deposited onto mild steel substrates by cathodic electrolytic deposition via Ce(III) chloride aqueous and mixed water – ethyl alcohol solutions with and without hydrogen peroxide as a precursor. The major role of this precursor is to increase the deposition rate of the oxide films allowing smooth and adherent hydrated cerium oxide films to be obtained. However, the evolution of the composition of the bath with time leads to a non-reproducibility of the morphology and composition of the thin films. The deposited oxide films are composed of small particles of hydrated Ce(IV) oxide with a significant amount of amorphous phase. In the absence of a precursor, only the mixed water – ethyl alcohol solution permits cerium oxide films to be deposited. The influence of the deposition parameters such as applied cathodic current density, deposition time and composition of the bath was also investigated. The deposits exhibit a quite developed crack network, which is highly dependent on the deposition parameters and the drying process. Moreover, the incorporation of chloride ions during the deposition has been observed. The protection efficiency of these films was examined during extended immersion tests in 3% NaCl solution. The deposits allow the cathodic reaction kinetics to be reduced by acting as a cathodic inhibitor. However, the presence of large defects or cracks induces an accelerated local degradation of the steel substrate, removing partially the oxide deposit, whereas a thin crack network can rapidly be sealed by corrosion products without deterioration of the deposit. D 2005 Elsevier B.V. All rights reserved. Keywords: Cerium oxide film; Cathodic electrodeposition; Protection efficiency
1. Introduction Chromium(VI) compounds are largely used to create a conversion layer that provides an improved corrosion resistance for many metallic alloys. The use of rare earth (RE) salts is nevertheless studied as an environmentally friendly alternative to the use of Cr(VI) based pre-treatments. An enhancement of the corrosion resistance for different alloys such as aluminium alloys [1 –3], zinc alloys [4 –8] or steels [9– 11] has been reported. The different mechanisms suggested in literature show that the role of the rare earth conversion films is not well understood. The main * Corresponding author. Tel.: +33 5 46 45 72 94; fax: +33 5 46 45 72 72. E-mail address: [email protected] (J. Creus). 0257-8972/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.surfcoat.2005.04.027
mechanism of action for the RE elements seems to be based on the reduction of the rate of the cathodic process acting as cathodic inhibitors [1,6]. However, Montemor et al. [5] recently showed that RE elements act as mixed inhibitors, reducing both cathodic and slightly anodic activity on hot dip galvanised steel substrates. Ceria (CeO2) has a considerable interest as a promising material in corrosion protection. Different techniques such as laser beam evaporation ion [12], chemical vapour deposition, sol – gel processing, immersion [1 – 9] and electrodeposition [10,11,13 –15] have been used to obtain cerium oxide films. Electrodeposition is an attractive method for the preparation of thin films since it offers the advantage of low processing temperature, controlled thickness of the film and low cost process. Cerium oxide films
J. Creus et al. / Surface & Coatings Technology 200 (2006) 4636 – 4645
have been obtained by cathodic or anodic deposition on different substrates [1,10,11,13– 18]. In the cathodic electrolytic deposition, hydroxide ions are formed at an electrode surface by a cathodic process and metal ions or complexes are then hydrolysed by the electrogenerated base. Zhou and Switzer [11], Zhitomirsky and Petric [13,14] and Arurault et al. [9] have used a solution of Ce(III) nitrate in order to deposit thin oxide films. The reduction of nitrate ions into nitrite or ammonia ions also produces hydroxide ions, which promote a rapid increase of the pH at the interface. The electrogenerated base induces the formation of colloidal particles of Ce(III) hydroxide or hydrous oxide [11,13,14], or also the formation of a ionic species [Ce(OH)22+], that can be hydrolysed into ceria. Different authors [13,14,19] suggest that the use of a precursor like hydrogen peroxide (H2O2) prevents the accumulation of nitrite or ammonia ions that provokes deposition instabilities. Zhitomirsky and Petric [13] have shown that, using a Ce(III) chloride solution, associated to hydrogen peroxide precursor, leads to more adherent and uniform oxide films than Ce(III) nitrate solution. The aim of this study was to characterise ceria coatings elaborated on a carbon steel substrate by cathodic electrolytic deposition from Ce(III) chloride aqueous and mixed water –ethyl alcohol solutions. The influence of hydrogen peroxide precursor is discussed and the corrosion resistance of coated steel is investigated in saline solution through extended immersion tests.
2. Experimental procedure Electrochemical experiments were carried out in a classical three-electrode glass cell. The working electrode was a 36NiCrMo16 (EN 1083-1) steel disk with an area of 0.78 cm2. Before each experiment, the steel surfaces were polished with silicon carbide (particle size 5 Am), sonicated for 2 min, rinsed thoroughly with Milli-Q water and dried. The reference electrode was a saturated calomel electrode (SCE). The counter electrode was a large platinum grid. The experiments were performed at room temperature with an AMETEK 263 A potentiostat system, driven by Softcorr 3 software. The steel electrodes were left at the open circuit potential for 30 s prior to the potential scans. The cerium oxide films were deposited by cathodic electrolytic deposition (CELD) via two electrolytic baths of CeCl3I7H2O. Their compositions are presented in Table 1. Table 1 Composition of the (A) aqueous and (B) mixed water – ethyl alcohol (1:9 volume ratio) solution for the deposition of thin solid films Composition
CeCl3I7H2O H2O2 (30 %)
Aqueous solution
Mixed water – ethyl alcohol solution
A1
A2
B1
B2
B3
0.8 mM
0.8 mM 25 mM
0.8 mM
10 mM
0.8 mM 25 mM
4637
Table 2 Deposition conditions of thin ceria films from both A and B solutions Conditions
Aqueous solution A1
A2
B1 and B3
B2
Duration (min)
10 20 0.005 0.01 0.1
10 20 0.005 0.01 0.1
10 20 0.01 0.1 0.25 1 5
10
j (mA/cm2)
Mixed water – ethyl alcohol solution
0.1 0.25
All the reagents were Sigma-Aldrich maximum purity chemicals. Cyclic voltammetry, performed from 0.55 V/SCE to 1.8 or 2.5 V/SCE at the scan rate of 20 mV/s, was used to characterise the electrochemical reactions. The deposits were obtained in galvanostatic mode in unstirred solutions at room temperature. The deposition conditions are summarised in Table 2. After deposition, the samples were carefully washed in order to prevent the detachment of the thin films and dried during 24 h in air at room temperature. The morphology of the samples was studied using optical microscope and scanning electron microscope (SEM JEOL 5410 Low Vacuum), coupled to X-ray energy dispersive spectroscopy (EDS). The cerium oxides were also analysed by X-ray diffraction (XRD) with a classical powder diffractometer (Bruker AXS D8-Advance), using Cu-Ka1 wavelength (k = 1.5406 nm) in Bragg-Brentano geometry. Raman spectra were recorded with a LabRam HR spectrometer, equipped with a confocal microscope, using an incident beam of 632.82 nm emitted by an argon laser. Differential scanning calorimetry (DSC TA Instrument Q100) was carried out in air between 40 and 550 -C at a heating rate of 20 -C/min. UV – visible absorption spectra of the tested Ce(III) chloride aqueous and mixed water –ethyl alcohol-based solutions containing H2O2 as a precursor were obtained using a Helios h-UNICAM spectrometer. The protection efficiency of the coatings was investigated via experiments performed in aerated and stirred 3% NaCl solution (pH = 6). The open circuit potential (OCP) and polarisation resistance (R p) of the coated substrates versus time were recorded during long immersion tests (around 65 h). The polarisation resistance measurements were estimated via a linear polarisation experiment performed at a scan rate of 0.16 mV/s about T20 mV/OCP. Polarisation curves i(t) were recorded at a scan rate of 0.16 mV/s after 1 h of immersion in saline solution.
3. Results and discussion 3.1. Synthesis and characterisation of cerium oxide powder The synthesis of cerium oxide by precipitation was performed by addition of sodium hydroxide solution in
4638
J. Creus et al. / Surface & Coatings Technology 200 (2006) 4636 – 4645
Ce(III) chloride aqueous solution. The precipitate was filtrated and dried in air during 24 h at room temperature. The same procedure was performed with a Ce(III) chloride solution containing hydrogen peroxide (H2O2) as a precursor. This precursor had been added for complete oxidation of Ce(III) into Ce(IV). The results of the various analyses performed on the precipitate obtained from CeCl3 and NaOH are presented in Fig. 1. The precipitate obtained in the aqueous solution containing H2O2 presents the same characteristics. The XRD pattern (Fig. 1a) is typical of ceria (CeO2) as it is reported in literature [9]. No diffraction lines associated to Ce(III) oxide are present, which seems to be in agreement with [20]. We observe that the line width is large, which is
typical of a small particle size. Raman spectrum (Fig. 1b) presents an intense band at 461 cm1, which corresponds to CeO2 [21 – 23]. Kosacki et al. [21] remarked that the width of the intense band increased when the size of ceria particles decreased. Therefore, the precipitation technique via the Ce(III) chloride solution leads to a precipitate composed of small particles of ceria. Moreover, the bands around 1660 cm1 in the Raman spectra suggest the presence of hydration water molecules. Indeed, the DSC diagram (Fig. 1c), performed between 40 and 550 -C, presents an intense exothermic band centred on 100 -C during the heating step that is attributed to a dehydration mechanism, in agreement with literature [9,18]. Therefore, the precipitates obtained in both solutions correspond to hydrated ceria. The presence of hydrogen peroxide precursor allows to completely oxidise Ce(III) into Ce(IV) leading to a stoichiometric cerium(IV) oxide. Without this precursor, Ce(III) precipitates into hydrated Ce(OH)3 that is then oxidised into hydrated cerium(IV) oxide either during the increase of the pH of the solution [2] or during the drying process [10,11,18]. 3.2. Elaboration and characterisation of cerium oxide films by CELD Electrochemical mechanism of base electrogeneration during cathodic deposition has been widely discussed in the literature [9 – 18,24 – 26]. In the case of chloride baths, where anions do not participate in any cathodic reaction, the hydroxyl ions are produced by the reduction of dissolved oxygen or water (Eqs. (1) and (2)). O2 þ 2H2 O þ 4e Y4OH
ð1Þ
2H2 O þ 2e YH2 þ 2OH
ð2Þ
The addition of hydrogen peroxide leads to the formation of supplementary hydroxyl ions during the cathodic reaction (Eq. (3)): H2 O2 þ 2e Y2OH
Fig. 1. Characterisation of the precipitate obtained by alkalisation of aqueous Ce(III) chloride solution. (a) XRD pattern, (b) Raman spectrum and (c) DSC diagram of the precipitate dried in ambient air during 24 h at room temperature.
ð3Þ
Zhitomirsky and Petric [13] showed that the role of H2O2 as a precursor was to prevent the formation of a nonstoichiometric cerium oxide film by oxidising Ce(III) into Ce(IV), the latter being more easily hydrolysed. Moreover, this precursor was shown to enhance the adherence and uniformity of the oxide films. The concentration of chloride ions in the solution is progressively decreased by the action of H2O2, but this precursor does not avoid the incorporation of chloride ions into the coating. The formation of hydroxyl ions at the cathode leads to a local increase of the pH at the surface that promotes the formation of a Ce(OH)3 precipitate or/and the formation of a soluble ionic complex Ce(OH)22+ (Eqs. (4) or (5a) and (5b)). Ce(OH)22+ appears preferentially when H2O2 is added in the solution.
J. Creus et al. / Surface & Coatings Technology 200 (2006) 4636 – 4645
Ce3þ þ 3OH YCeðOHÞ3
ð4Þ
Ce3þ þ 2OH YCeðOHÞ2þ 2 þe
ð5aÞ
2Ce3þ þ 2OH þ H2 O2 Y2CeðOHÞ2þ 2
ð5bÞ
The cerium oxide film then results either from the oxidation of Ce(OH)3 (Eq. (6)) or from the hydrolysis of Ce(OH)22+ (Eq. (7)) [9,18]. CeðOHÞ3 YCeO2 þ H3 Oþ þ e
ð6Þ
CeðOHÞ2þ 2 þ 2OH YCeO2 þ 2H2 O
ð7Þ
Recently, Stefanov et al. [10] showed that the electrodeposition of cerium oxide on stainless steel from an aerated Ce(III) chloride solution leads to a mixture of Ce2O3 and CeO2. The electrochemical mechanism of formation of these oxide films is really complex. The electrogenerated base induces the formation of colloidal particles of Ce(III) hydroxide or hydrous oxide near the surface [11,13,14,18]. Li and Thompson [18] suggested that the deposition of the solid film proceeded through a nucleation and growth process. Fig. 2 presents the cyclic voltammetry curves obtained with steel substrates in freshly prepared Ce(III) chloride aqueous (Fig. 2a, baths A1 and A2) and mixed water – ethyl
Fig. 2. Cyclic voltammograms performed in aqueous (a) and mixed water – ethyl alcohol (b) Ce(III) chloride solutions with (- - - -) and without (—– ) H2O2.
4639
alcohol (Fig. 2b, baths B1 and B3) solutions with and without hydrogen peroxide. In aqueous solution (bath A1), the variations of the current density are due to the reactions (1) and (2). The current hysteresis, observed during the reverse sweep, indicates a modification of the electrode surface during the cathodic polarisation. This surface modification is more important when H2O2 is added to the solution (bath A2). The shape of the voltammogram is completely changed due to reaction (3). We can see that water reduction (2) is strongly shifted towards more negative potentials probably due to an important increase of the interfacial pH. Therefore, it appears that the hydrogen evolution reaction (HER) prevents the deposition of cerium oxide films in aqueous solution. In mixed water – ethyl alcohol solution (bath B1), the voltammogram is different from solutions A due to the low amount of water in the solution. The presence of water is necessary to electrogenerate the OH ions that are essential for the formation of the Ce(III) hydroxide particles. The current densities are lower. The addition of H2O2 induces an important increase of the current densities and an important current hysteresis is observed which suggest a beneficial effect of this precursor on the modification of the electrode surface. Therefore, in both baths A and B, the addition of the H2O2 promotes the formation of the cerium oxide films onto the steel substrate as expected. Once dried, the coatings present a dark yellowish colour. The SEM observations reveal that the films are covering and adherent and EDS analyses underline the presence of low Ce amounts. Fig. 3 shows the surface morphology, as observed by SEM, of cerium oxide films deposited after 20 min in aqueous solution containing H2O2 (bath A2) at 0.005 (Fig. 3a) and 0.01 mA/cm2 (Fig. 3b). Many surface defects such as important crack networks and porosities appear in all the coatings. Circular cracks whose diameters can even reach 20 –30 Am are also observed regardless of the applied cathodic current. These circular cracks or craters are associated to either the existence of inclusions or to the formation of gas bubbles, which is promoted during the process on the steel surface when current densities are increased. Raman spectra performed either in the centre of a large circular crater (Fig. 4a) or near the extremity of a crack (Fig. 4b) are typical of corrosion products of steel, namely ferrihydrite 5Fe2O3I9H2O (Fig. 4a) and lepidocrocite g-FeOOH (Fig. 4b). Cerium oxide bands are hindered by the important degradation of the substrate, which may occur during the drying process. Moreover, coating morphology and composition are not reproducible due to the evolution of bath during time, this evolution being more important in aqueous solution than in mixed water –ethyl alcohol solution. For freshly prepared baths (A2 or B3), the films obtained during the cathodic polarisation are essentially composed of cerium oxide, but after 1 day, the appearance of the coating changes and an important degradation of the substrate is observed.
4640
J. Creus et al. / Surface & Coatings Technology 200 (2006) 4636 – 4645
Fig. 4. Raman spectra of cerium oxide films deposited at 0.005 (a) and 0.1 (b) mA/cm2 during 20 min in aqueous Ce(III) chloride solution containing H2O2.
Fig. 3. SEM observations of cerium oxide films deposited at 0.005 (a) and 0.01 (b) mA/cm2 during 20 min in aqueous Ce(III) chloride solution containing H2O2.
Fig. 5 presents the evolution of the pH of the aqueous and mixed water – ethyl alcohol solutions with and without H2O2 precursor. The addition of H2O2 in colourless Ce(III) chloride solutions induces an orange colour associated to the formation of the hydroxide species Ce(OH)22+ as it was reported by Aldykiewicz et al. [26] and Golden and Qi Wang [27]. The pH of the solutions gradually decreases after the addition of H2O2. For the bath A2, after 1 day, the pH remains constant and the solution change progressively into a yellowish colour. This colour modification is associated to the Ce(IV) species, which is confirmed by UV – visible spectroscopy. After a few days, the solution becomes progressively colourless whereas pH remains constant. The UV – visible spectroscopy analyses reveal that the amount the Ce(IV) species is reduced. This suggests that the Ce(IV) species are progressively reduced into Ce(III) into the aqueous media. For the bath B3, concerning the mixed water – ethyl alcohol solution, the same solution evolution is observed when H2O2 is added. The evolution of the pH and the nature of the ionic species in solution explain the non reproducibility of the coating morphology and composition.
In aqueous solution (bath A1), whatever the applied cathodic current densities between 0.005 and 0.6 mA/cm2 and whatever the duration of the experiment, no cerium oxide films are detectable. In mixed water – ethyl alcohol solution (bath B1), the cathodic current densities were varied between 0.01 and 5 mA/cm2. Only cathodic current densities above 0.1 mA/cm2 allowed to obtain cerium oxide films. Fig. 6 presents the SEM surface morphologies of the cerium oxide films deposited in mixed water – ethyl alcohol solution (bath B1) for different cathodic current densities and time deposition. The oxide films present a ribbed morphology, with a crack network that develops during the
Fig. 5. Evolution versus time of the pH of aqueous solutions with (h) or without (r) H2O2 and mixed water – ethyl alcohol solution with (?) or without (4) H2O2.
J. Creus et al. / Surface & Coatings Technology 200 (2006) 4636 – 4645
4641
Fig. 6. SEM observations of ceria films deposited at 0.25 mA/cm2 during 20 (a) and 10 min (b), at (c) 0.5 mA/cm2 during 20 min and (d) 1 mA/cm2 during 10 min in mixed water – ethyl alcohol Ce(III) chloride solution.
drying process, as also reported in the literature [9,13,14]. This crack network increases with the applied current density or with time deposition as observed for coatings deposited at 0.25 mA/cm2 during 10 min (Fig. 6b) and 20 min (Fig. 6a). Therefore, the increase of the coating thickness promotes crack formation during the drying process. For current densities around 0.5 mA/cm2, the coatings become rather inhomogeneous combining zones with large cracks and coating detachments and adherent zones (Fig. 6c). When current densities are above 1 mA/ cm2, the oxide films are not adherent and present large clusters and large cracks (Fig. 6d). Locally, the detachment of the oxide coating during the drying process leaves the steel surface uncovered. A Raman spectrum typical of all the coatings deposited in bath B1 is shown in Fig. 7. It presents a large intense band at 454 cm1 that may correspond to a small particle size [21,23]. The size of the hydroxide or hydrous oxide particles strongly depends on the applied cathodic current density. Zhou and Switzer [11] reported that, during the deposition from aqueous Ce(III) nitrate solution onto stainless steel, the grain size decreased from 18 nm to approximately 6 nm when applied current densities varied from 0.5 to 3 mA/cm2. The enlargement of the intense band can also be attributed to an amorphous character of the
films. The XRD pattern (not presented) of the deposit obtained at 0.1 mA/cm2 during 20 min exhibits a very broad line near 2h å 27-, which suggests that the deposit contains a significant amount of amorphous phase [14]. Therefore, it seems that the deposits are composed of low size particles with a significant amount of amorphous phase. In the following, the influence of the concentration of Ce(III) species in solution is discussed. Fig. 8 presents the SEM surface morphology of an oxide film deposited at 0.1 mA/cm2 during 10 min in bath B2. The coating presents few
Fig. 7. Raman spectra of ceria films deposited at 0.25 mA/cm2 during 10 min in mixed water – ethyl alcohol Ce(III) chloride solution.
4642
J. Creus et al. / Surface & Coatings Technology 200 (2006) 4636 – 4645
a
20 µm 325.5 391.0
b 139.9
716.3
5000
1362.7
4000 1070.1
Intensity (a.u.)
6000
500.1
7000
3000 2000 1000 200
400
600
800
1000
1200
1400
Wavenumber (cm-1)
Fig. 8. SEM Observation (a) and Raman spectrum (b) of ceria film deposited at 0.1 mA/cm2 during 10 min in Ce(III) chloride (10 mM) mixed water – ethyl alcohol solution.
dark spots (Fig. 8a) scattered at the coating surface and the number of these dark spots increases with deposition time. The morphology of these dark zones is totally different from the smooth and adherent oxide film. The Raman spectrum (Fig. 8b) performed at the centre of a dark spot is typical of a corrosion product of steel (ferrihydrite), which suggests a degradation of the steel substrate during the drying process. The corrosion of steel seems to be promoted by the incorporation of chloride ions during the formation of the hydrous oxide film. This incorporation is less pronounced in less CeCl3 concentrated solutions. Thus, the incorporation of chloride ions may depend on the rate of the oxide film formation and the electrical field at the electrode interface. Increasing the concentration of Ce(III) in mixed water – ethyl alcohol solutions promotes the formation of the oxide film but also the incorporation of chloride ions which is harmful for the steel substrate during the drying process. 3.3. Corrosion behaviour of oxide films deposited in mixed water –ethyl alcohol solutions Fig. 9 gathers the results of the experiments performed to estimate the protection efficiency of oxide films compared to bare steel in 3% NaCl solution. The polarisation curves of
Fig. 9. Corrosion behaviour in saline solution of cerium oxide films deposited onto steel substrate compared to bare steel. (a) Polarisation curves of (—– ) bare steel, (- - -) Ce oxide film (bath B1, 0.25 mA/cm2 during 10 min) and (IIII) Ce oxide film (bath B2, 0.25 mA/cm2 during 10 min). (b) Polarisation resistance versus immersion time of (†) bare steel, (?) Ce oxide film (bath B1, 0.25 mA/cm2 during 10 min) and (r) Ce oxide film (bath B2, 0.25 mA/cm2 during 10 min).
two oxide films deposited at 0.25 mA/cm2 during 10 min respectively in baths B1 and B2 are compared to that of bare steel (Fig. 9a). The corrosion potentials and current densities are gathered in Table 3. The current densities are associated to the limiting current densities deduced from the cathodic branch of the polarisation curves. Both oxide coatings reduce the corrosion current densities. The oxide coating deposited in bath B1 induces the lowest decrease of the corrosion current density and we can note that the corrosion potential is shifted towards more negative values. The cathodic curve, corresponding to the reduction of dissolved oxygen is slightly shifted, indicating a slight slowing down of the cathodic reaction kinetics. Moreover, the anodic part of the polarisation
Table 3 Electrochemical characteristics of coated steel compared to bare steel after 1 h of immersion in 3% NaCl solution Bare steel
E corr (mV/SCE) i corr (AA/cm2)
388 190
Oxide film deposited at 0.25 mA/cm2 during 10 min Bath B1
Bath B2
495 120
361 65
J. Creus et al. / Surface & Coatings Technology 200 (2006) 4636 – 4645
curve is quite similar to that of bare steel, which suggests that the oxide film has a limited effect on the dissolution mechanism of steel. For the coating deposited in bath B2, the polarisation curve is quite different. The corrosion current density is reduced and the corrosion potential is shifted towards more positive values. A slight reduction of the cathodic curve is observed which suggests a blocking effect on oxygen reduction reaction. In the anodic part, a rapid increase of current density occurs for very small anodic overpotential, which suggests a local corrosion degradation of the coated steel. Therefore, these oxide films, deposited via a mixed water –ethyl alcohol bath, exhibit an effective reduction of the corrosion current density by acting as a physical barrier against the aggressive media and reducing the kinetic of the reduction of dissolved oxygen [1,5,6]. Open circuit potential (not shown) and polarisation resistance R p (Fig. 9b) of these coatings during extended immersion tests in 3% NaCl solution were also monitored. A linearity relationship of potential/current in the range of T20 mV/OCP was observed which allowed to measure the polarisation resistance for different immersion duration. The open circuit potential of steel rapidly decreases to reach a constant value around 630 mV/SCE after 20 h of
4643
immersion. The open circuit potentials of cerium oxide films are quite similar to that of bare steel. The potential values are nobler, in particular for the coating deposited in bath B2. For bare steel, the R p values are close to 250 V cm2 during the first hours of immersion, and then increase slightly due to the formation of the corrosion products of steel limiting the diffusion of dissolved oxygen. This induces a shift of the open circuit potential towards negative values. For the oxide deposits, the R p values during the first hours of immersion are higher than those of bare steel. The highest values correspond to the deposit obtained in bath B2, which is in agreement with the polarisation curves. The R p value of this coating decreases until 20 h of immersion and then remains constant, whereas the R p evolution of the deposit obtained in bath B1 is similar to that of bare steel but the values are lower. This is more likely due to a local degradation of the substrate. Fig. 10 presents SEM observations of oxide films obtained either in bath B1 (Fig. 10a) or in bath B2 (Fig. 10b) after extended immersion tests. Raman spectra of typical corrosion products observed in both oxide films are presented in Fig. 10c and d. The SEM observations reveal a local attack of the substrate through defects like cracks or porosities in each case. The oxide films are still present and locally prevent the localised degradation of the substrate (Fig. 10a), even if a
b
c
d 453.5
a
3000
375.6
1297.2
648.3
5000
525.5
10000 306.9 347.0
1500
15000
141.9 215.7
1068.4
2000
1366.1
685.1
Intensity (a.u.)
530.7
133.7 177.8 243.9
Intensity (a.u.)
376.3
20000 2500
1000 0 500
1000 Wavenumber (cm-1)
1500
500
1000
1500
Wavenumber (cm-1)
Fig. 10. SEM observations of cerium oxide films deposited onto steel substrate after extended immersion tests. (a) Deposit at 0.25 mA/cm2 during 10 min (bath B1 0.8 mM), (b) deposit at 0.25 mA/cm2 during 10 min (bath B2 10 mM) and Raman spectra of typical corrosion products (c and d) developed on the surface of the samples.
4644
J. Creus et al. / Surface & Coatings Technology 200 (2006) 4636 – 4645
crack network exists. These cracks are rapidly sealed with corrosion products from the steel substrate as far as the crack width is small. Then, the corrosion of the steel substrate is negligible. The oxide coating obtained in bath B2 contains many dark spots that are identified as ferrihydrite. This degradation occurs during the drying process and this corrosion product partially seals the cracks and defects of the oxide film. This explains why this oxide coating presents during the first hour of immersion the highest polarisation resistance values. The SEM observation performed after the extended immersion tests shows the presence of corrosion products that covered again the oxide film and an expansion and modification of the dark spots observed before the immersion. The Raman spectra (Fig. 10c and d) present the main corrosion products observed on both oxide coatings. The spectrum in Fig. 10c is typical of ferrihidrite that is mainly observed in the centre of the cracks, whereas the spectrum in Fig. 10d is quite typical of lepidocrocite gFeOOH that is observed either on the top of the cracks or on the surface. Finally, the broad band at 453 cm1 (Fig. 10c) confirms the presence of the cerium oxide films. It is interesting to underline that certain zones of the coated steel remain unaffected even after extended immersion tests (62 h) in saline solution, in particular for the oxide films deposited in bath B1. So these coatings may ensure a protection of the steel surface but the protection effectiveness strongly depends on the morphology and composition of the oxide films. The crack network that appears during the drying process is harmful for the protection of the substrate during extended immersion tests as it can be observed in Fig. 9b. Therefore, it seems that the deposition conditions should be accurately adjusted to control the morphology of this oxide film. Another important parameter is the incorporation of chloride ions in the deposited hydrous oxide film. In solution B2, the negative effect of this incorporation that induces a local attack of the substrate during the drying process is clearly demonstrated. Incorporation of chloride ions in the hydrous oxide films is assumed to play also a role during the extended immersion tests, in particular by promoting a local acidification at the substrate – oxide interface. A more detailed investigation should therefore be performed to underline the influence of the incorporation of chloride ions in the cerium oxide films on the corrosion mechanism in saline solution and to improve the quality of the coatings.
4. Conclusions Two deposition baths were compared so as to produce cathodic electrodeposits of cerium oxide. Whatever the deposition parameters performed in aqueous Ce(III) chloride solutions, no cerium oxide film could be obtained. It seems that the reduction of dissolved oxygen does not produce a sufficient increase of the interfacial pH necessary to the formation of colloidal Ce(III) hydroxide particles. At
higher applied current densities, the hydrogen evolution reaction seems to prevent the formation of the cerium oxide films. The addition of H2O2 is essential in aqueous baths. In a freshly prepared solution, the rate of deposition is increased via the reduction of H2O2 that produces hydroxyl ions at the interface. Adherent, uniform and homogeneous hydrated cerium oxide films were then obtained. These oxide films are characterised by an intense crack network that develops during the drying process. It strongly depends on the deposition parameters such as applied current density, deposition time and bath composition. However, the ageing of baths containing H2O2 leads to a non-reproducibility of the morphology and composition of the oxide films, and also to local degradations of the steel substrate. So, the addition of H2O2 in electrolytic baths of cerium has to be avoided in steel protection applications and alternative processes are under investigation. With mixed water – ethyl alcohol solutions, cerium oxide films could be obtained without the addition of H2O2. The influence of H2O2 has been studied, leading to results similar to those obtained with aqueous solutions This mixed solution allows the pH increase close to the electrode interface to be localised promoting the formation of colloidal Ce(III) hydroxide particles. The cerium oxide coatings are composed of small particles with a significant amount of amorphous phase. At low current densities, the oxide films are adherent, dense and present a thin crack network. The crack width increases with the deposition time. The increase of the coating thickness promotes the formation of large cracks, which locally induces a delamination of the coating. High current densities or extended deposition times induce non-adherent films, whereas increasing the Ce(III) chloride concentration in solution should promote the incorporation of chloride ions in the hydrous cerium oxide films. This induces localised degradations of steel during the drying process. The best morphologies are obtained at low current densities and low deposition times. The corrosion behaviour of the coated steel was examined in aerated and stirred 3% NaCl solution. The protection efficiency strongly depends on the shape of the crack network and the adherence of the cerium oxide films. The films obtained at low current densities or short deposition times present the best protection efficiency due to the thin crack network. The incorporation of chloride ions into the hydrous cerium oxide films seems to affect the protection efficiency. The behaviour of samples submitted to extended immersion tests is however promising and the cerium oxide films should be considered as a potential alternative for steel protection.
Acknowledgements B. Peraudeau and M. Jeannin are gratefully acknowledged for their technical support during this project.
J. Creus et al. / Surface & Coatings Technology 200 (2006) 4636 – 4645
References [1] B.R.W. Hinton, N.E. Ryan, D.R. Arnott, Mater. Australas. (1987) 18. [2] A.J. Davenport, H.S. Isaacs, M.W. Kendig, Corros. Sci. 32 (1991) 653. [3] W.G. Fahrenholtz, M.J. O’Keefe, H. Zhou, J.T. Grant, Surf. Coat. Technol. 155 (2002) 208. [4] M.G.S. Ferreira, R.D. Duarte, M.F. Montemor, A.M.P. Simoes, Electrochim. Acta 49 (2004) 2927. [5] M.F. Montemor, A.M.P. Simoes, M.G.S. Ferreira, Prog. Org. Coat. 43 (2001) 274. [6] M.A. Arenas, J.J. De Damborenca, Surf. Coat. Technol. 187 (2004) 320. [7] K. Aramaki, Corros. Sci. 46 (2004) 1565. [8] K. Aramaki, Corros. Sci. 43 (2001) 2201. [9] L. Arurault, P. Monsang, J. Salley, R.S. Bes, Thin Solid Films 466 (2004) 75. [10] P. Stefanov, G. Atanasova, D. Stoychev, T.S. Marinova, Surf. Coat. Technol. 180 – 181 (2004) 446. [11] Y. Zhou, J.A. Switzer, J. Alloys Compd. 237 (1996) 1. [12] S. Kanakaraju, S. Mohan, A.K. Sood, Thin Solid Films 305 (1997) 191.
4645
[13] I. Zhitomirsky, A. Petric, Ceram. Int. 27 (2001) 149. [14] I. Zhitomirsky, A. Petric, Mater. Lett. 40 (1999) 263. [15] M. Balasubramaniam, C.A. Melendres, A.N. Mansour, Thin Solid Films 347 (1999) 178. [16] A. Qi Wang, T.D. Golden, J. Electrochem. Soc. 150 (2003) C616. [17] T.D. Golden, A. Qi Wang, J. Electrochem Soc. 150 (2003) C621. [18] E.B. Li, G.E. Thompson, J. Electrochem. Soc. 146 (1999) 1809. [19] Th. Pauporte, D. Lincot, J. Electroanal. Chem. 517 (2001) 54. [20] Y.X. Li, X.Z. Zhou, Y. Wang, X.Z. You, Mater. Lett. 58 (2003) 245. [21] I. Kosacki, T. Suzuki, H.U. Anderson, P. Colomban, Solid State Ion. 149 (2002) 99. [22] J.E. Spanier, R.D. Robinson, Z. Zhang, S.W. Chan, I.P. Hernan, Phys. Rev., B, Condens. Matter Mater. Phys. 64 (2001). [23] S. Wang, W. Wang, J. Zuo, Y. Qian, Mater. Chem. Phys. 68 (2001) 246. [24] J.M. Brossard, J. Balmain, J. Creus, G. Bonnet, Surf. Coat. Technol. 185 (2004) 275. [25] G.H. Annal Therese, P. Vishnu Kamath, J. Appl. Electrochem. 28 (1998) 539. [26] A.J. Aldykiewicz, A.J. Davenport, H.S. Isaacs, J. Electrochem. Soc. 143 (1996) 147. [27] T.D. Golden, A. Qi Wang, J. Electrochem. Soc. 150 (2003) C621.