This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright
Transcript
This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies areencouraged to visit:
Materials Chemistry and Physics 113 (2009) 650–657
Contents lists available at ScienceDirect
Materials Chemistry and Physics
journa l homepage: www.e lsev ier .com/ locate /matchemphys
Cathodic electrodeposition of cerium-based oxides on carbon steel fromconcentrated cerium nitrate solutionsPart I. Electrochemical and analytical characterisation
Y. Hamlaouia,1, F. Pedrazaa,∗, C. Remazeillesa, S. Cohendoza, C. Rébéréa, L. Tifoutib, J. Creusa
a Laboratoire d’Etudes des Matériaux en Milieux Agressifs (LEMMA), Pôle Sciences et Technologie, Université de La Rochelle, Avenue Michel Crepeau,17042 La Rochelle Cedex 1, Franceb Laboratoire de Génie de l’Environnement, Université Badji Mokhtar, BP 1223, 23020 El Hadjar-Annaba, Algeria
a r t i c l e i n f o
Article history:Received 8 January 2008Received in revised form 19 June 2008Accepted 3 August 2008
In this work the elaboration by cathodic electrodeposition of cerium-based oxides on carbon steel fromrelatively concentrated cerium nitrate solutions is investigated. In particular, the study presented here(Part I) focuses on the electrochemical and analytical characterisation of the films and on the correlationsbetween the electrochemical features and the characteristics of the layers. The effect of other parame-ters such as concentration, temperature, pH and additives to improve the behaviour of the film againstcorrosion will be investigated in part II of the study.
The electrochemical characterisation will reveal that Ce(IV)–steel interactions can be responsible forsome weak electrochemical waves appearing in the cyclic voltammograms that often are attributed tooxygen or nitrates reduction. This results from the oxidation of Ce(III) solutions to Ce(IV) in contact withair. Furthermore, the deposits strongly depend on the applied current density. Low current densities donot render fully covering deposits on the steel and a carbonated green rust will appear. On the contrary, theincrease of the current density leads to denser layers of relatively small crystallite size that readily coversthe steel surface. The deposits have a needle-like morphology and the Ce content achieves a plateauof about 20–22 at.%. However, a significant network of cracks appears probably occurring during thedeposition process itself. The differential scanning calorimetry (DSC) results indicate that the deposits arenot fully crystalline after 550 ◦C in contrast with the X-ray diffraction (XRD) patterns that unambiguouslyshow a fluorite-type CeO2 phase whose crystallite size decreases with increasing the current density. Therinsing medium also brings about different features of the films. Rinsing with water allows to incorporatemore nitrates and to adsorb CO2 than when rinsing with ethanol. However, R-OH bonds will be trappedin the latter.
Cerium containing salts have been long investigated as envi-ronmentally friendly alternatives to the use of Cr(VI) baths forthe protection of steels [1,2]. Cathodic electrodeposition has beenclaimed to become an important method on the processing ofthis ceramic because of the low cost of equipment and of pre-cise control of deposited thickness [3–6]. Also, the films can beobtained on a vast variety of metallic substrates as aluminiumand aluminium alloys [7–9], magnesium alloys [10,11], stainlesssteel [5,6,12–14], and galvanised and Zn-plated steels [15–22]. To
1 On research leave from the Institut des Sciences et Sciences de l’Ingénieur, CentreUniversitaire de Souk-Ahras, BP 1553, 41000 Souk-Ahras, Algeria.
this end, low precursor concentrations (between 10−3 and 10−2 M)seem to be required but the use of relatively high concentrationshas only been investigated by Arurault et al. [13] using an elec-trochemical cell with a diaphragm. Similarly, very few works havebeen devoted to the deposition of Ce-based coatings on low carbonsteel [23,24]. This is however widely employed as the substratefor Zn-based coatings (galvanised or Zn-plated). Therefore, someprotection should be conferred upon the storage period prior tocoating or in the case the coating showed significant defaults. More-over, the rinsing medium and its viability to eliminate water andincorporated nitrates in the film is not yet understood.
The aims of this study are therefore to elaborate and tocharacterize the films of cerium-based oxides electrodepositedfrom relatively high precursor concentrations using a classicthree-electrode cell. The correlation between the electrochemicalfeatures in the cerium nitrate solutions and the characteristics ofthe layers developed on an A366 low carbon steel is established
Y. Hamlaoui et al. / Materials Chemistry and Physics 113 (2009) 650–657 651
through polarisation experiments, Raman spectroscopy, X-raydiffraction (XRD), Fourier transformed infrared spectroscopy(FT-IR) and scanning electron microscopy (SEM) coupled to energydispersive spectrometry (EDS). As the rinsing medium may play arole on the morphology and compactness of the deposits [25,26]the differences between ethanol and water as rinsing media willbe also addressed in this work. Furthermore, the appearance ofa corrosion product will raise the question on how to inhibit itsdevelopment and growth. The effect of other parameters suchas concentration, temperature, pH and additives to improvethe behaviour of the film against corrosion will be thereforeinvestigated in part II of the study.
2. Experimental
2.1. Materials
Plates of 2-mm thick of A366 cold-rolled steel were cut to produce discs of14 mm of diameter. The low carbon steel nominal composition – according to SAE1008/1010 – is: 0.13 C, 0.041 Mn, 0.040 S, 0.012 N, 0.55 Cu (wt.%) and Fe balance.The samples were mechanically abraded from coarser to finer 2400 grit SiC emerypaper, rinsed with water then cleaned in an ultrasonic bath of ethanol and finallydried with pulsed hot air immediately before the deposition of the films.
2.2. Experimental set-up for cathodic electrodeposition
The electrochemical baths were composed of 0.01, 0.1 and 0.25 MCe(NO3)3·6H2O solutions at room temperature. The electrodeposition wasrealised using a classical three-electrode experimental set-up, with the steelsample being the cathode, a platinum grid as the counter electrode and a saturatedcalomel electrode (SCE) served as the reference one. The deposition experimentswere carried out in the galvanostatic mode without stirring the solution. Cyclicvoltammetry was recorded from −0.6 to −2.5 V under various scanning rates. Afterelectrodeposition, the samples were rinsed in ethanol and dried in a desiccatorovernight before any further subsequent analysis.
2.3. Characterisation of the films
The morphologies of the films were first investigated by optical microscopy thenmore thoroughly by scanning electron microscopy (JEOL 5410 LV) coupled to EDSanalysis (Rontec detector). The X-ray patterns were obtained in a Bruker AXS D8-Advanced diffractometer using Cu K� radiation (� = 1.5406 nm) at a scan rate of0.04◦ s−1 in the symmetric configuration. The crystallite size “D” was calculatedfrom the Scherrer equation (D = 0.9�/ˇ cos �) applied to the diffraction peak cor-responding to the (1 1 1) reflections (where � is the wavelength of the X-ray, � is thediffraction angle and ˇ is the full width at half maximum of the peak). The Ramanspectra were recorded with a LabRam HR8000 spectrometer equipped with a con-focal microscope using an incident beam of 632.82 nm emitted by a HeNe laser.The deposits were also scraped then milled and further analysed by XRD and dif-ferential scanning calorimetry (DSC) in a Q100 of TA Instruments between 30 and550 ◦C with a heating rate of 10 ◦C min−1 under nitrogen. This atmosphere confersto the deposit a better microstructural order and a longer oxidation time than inoxygenated atmosphere [27]. The infrared measurements were performed with aThermo Nicolet FT-IR Nexus spectrometer using a KBr beamsplitter, a DTGS detec-tor and a diffuse reflectance accessory. The spectra were recorded with the Omnicsoftware at a resolution of 4 cm−1, with 128 scans and a gain of 8. The samples werediluted with KBr.
3. Results and discussion
3.1. Electrochemical characterisation
Fig. 1 presents the cyclic voltammetry responses of the A366electrodes immersed in increasing (0.01, 0.1 and 0.25 M) concen-trations of fresh aqueous cerium nitrate solutions. The variationof the current density can be in principle related to the followingreactions:
O2 + 2H2O + 4e− → 4OH− (1)
NO3− + H2O + 2e− → NO2
− + 2OH− (2)
2H2O + 2e− → H2 + 2OH− (3)
Fig. 1. Cyclic voltammograms performed on the uncoated steel at different concen-trations of Ce(III) nitrate at room temperature (scanning rate: 20 mV s−1).
The hysteresis observed during the reverse scan indicates thatthe surface of the electrode is modified during the cathodic polar-isation especially in the more concentrated solutions. As theconcentration is increased, the solution pH lowers (4.5, 3.8 and 3.3for the 0.01, 0.1 and 0.25 M concentrations, respectively). Therefore,the corresponding thermodynamic potentials for the reaction (1)should be −1.06, −1.01 and −0.96 V/SCE. However, the experimen-tal curves show that the potential values have moved towards lesscathodic values with overpotentials of about 200 mV. This wouldsuggest that the reduction reaction occurs rapidly.
Furthermore, during the cathodic scanning, a peak of weakintensity is observed at −0.86 and −0.89 V/SCE for the 0.1 and0.25 M concentrations, respectively, but it does not appear at0.01 M, hence suggesting that it depends on the cerium nitrate con-centration. The same peak was also observed in previous works[28,29] and was attributed to the nitrate reduction, the reductionof oxygen or to a catalytic reaction. In principle, three interactionscould explain the origin of this peak: reduction of nitrates, reduc-tion of oxygen and evolution of the cerium species. The reductionof nitrates has been investigated by comparing the cyclic voltam-mograms of the substrate in 0.1 M KNO3 and in 0.1 M CeCl3·7H2Osolutions (Fig. 2). It can be observed in the enlarged area of Fig. 2that the peak appears in the cerium chloride solutions but not inthe potassium nitrate ones. Therefore, the role of the reduction ofnitrates can be disregarded.
Fig. 2. Cyclic voltammograms performed on the uncoated steel in 0.1 M CeCl3·7H2Oand in KNO3 solutions at room temperature (scanning rate: 20 mV s−1).
Author's personal copy
652 Y. Hamlaoui et al. / Materials Chemistry and Physics 113 (2009) 650–657
Fig. 3. Cyclic voltammograms performed on the uncoated steel at different scanningrates and room temperature from 0.1 M Ce(III) nitrate solutions.
The oxygen reduction can occur in a single step involving fourelectrons (reaction (1)) or in two steps involving two electronseach (reactions (4) and (5)). In the case of oxygen reduction, whichshould be under mass transport control, the cyclic voltammogramsobtained at different scanning rates in 0.1 M Ce(NO3)3 solutions(Fig. 3) show no linear relationship between the current densityand the
√� (� = scanning rate). This implies that the weak peak
cannot be either related to the reduction of oxygen.
O2 + 2H2O + 4e− → 4OH− E0 = 0.4 V/SHE (1)
or
O2 + H2O + 2e− → HO2− + OH− E0 = −0.07 V/SHE (4)
HO2− + H2O + 2e− → 3OH− E0 = 0.87 V/SHE (5)
The third possibility regarding the origin of the weak peak isrelated to the role that cerium species may play. At higher solu-tion pH, O2 tends to oxidize Ce(III) into Ce(IV) species like Ce(OH)4,Ce(OH)2
2+ and CeO2·xH2O [30,31]. For the O2–Ce(III)–Ce(IV) sys-tem, the major species during oxidation is Ce(OH)4 at pH > 4(reaction (6)):
4Ce3+ + O2 + 14H2O → 4Ce(OH)4 + 12H+ (6)
This means that under the presence of oxygen the solutionmay evolve towards an equilibrium in which both the Ce(III) andCe(IV) species exist. Moreover, Creus et al. [24] showed that the pHdropped with increasing the exposure of the solution to air. Indeed,for fixed 0.1 M concentrations, the solution can be stabilised byreducing the pH. This brings about a shift of the weak peak towardsmore cathodic values (Fig. 4) until its disappearance at pH 1.5.
The linear relationship established between the reductionpotential “Er” and log(Ce3+) characterized by a slope of 0.07 (sign oftransfer of one electron) and the fact that cerium deposition seemsto be controlled by the mass transfer [32] therefore suggest that thepeak is attributed to the interaction between the substrate (steel)and the ions Ce4+ free in the solution. The absence of the peak onthe voltammogram established in 0.01 M which should theoreti-cally appear at −0.7 V/ECS is probably due to the weak free Ce4+
concentration.
3.2. Film electrodeposition
Electrodeposition under the present experimental conditionsleads to the formation of a film over the steel substrate surface, evenat low current densities. The electrochemical mechanism of base
Fig. 4. Cyclic voltammograms performed in 0.1 M Ce(III) nitrate solution at differentpH and room temperature.
electro-generation during cathodic deposition has been widely dis-cussed in the literature [3–6,12,13,24]. In the case of nitrate bathswere the anion participates in the cathodic reaction, the deposit for-mation can be accompanied by the following steps [3,6,12,32–36]:
(1) nitrate reduction
NO3− + 10H+ + 8e− → NH4
+ + 3H2O (7)
NO3− + H2O + 2e− → NO2
− + 2OH− (2)
(2) O2, H2O and H3O+ reduction
2H3O+ + 2e− → H2 + 2H2O (8)
2H2O + 2e− → H2 + 2OH− (3)
O2 + 2H2O + 4e− → 4OH− (1)
O2 + 2H2O + 2e− → 2OH− + H2O2 (9)
(3) H3O+consumption and OH− production favour the formation ofCe(OH)3 and/or Ce(OH)2
2+.
Ce3+ + 3OH− → Ce(OH)3 (10)
4Ce3+ + O2 + 4OH− + 2H2O → 4Ce(OH)22+ (11)
2Ce3+ + 2OH− + H2O2 → 2Ce(OH)22+ (12)
These two last reactions are rather unlikely because of thesignificant number of species involved.
(4) Oxidation of Ce3+ to Ce4+
Ce(OH)3 → CeO2 + H3O+ + e− (13)
Ce(OH)22+ → CeO2 + 2H2O (14)
The effect of the current density on the morphology has beeninvestigated in order to obtain homogeneous and crack-free thinfilms. Fig. 5 gives the variation of potential versus time during depo-sition. For the current densities higher than 0.5 mA cm−2, a quickand monotonous decrease of potential is observed in the first fewseconds followed by a stabilisation of the potential. Such stabili-sation appears sooner by increasing the applied current density.However, at current densities lower than or equal to 0.5 mA cm−2,the electrodeposition passes by several stages before the potentialstabilises.
The change of the shape of the polarisation curves with theapplied current densities can be related to the reduction reactionstaking place at the electrode. A continuous variation of potentialwith time is indicative of a single reaction, whereas sudden changes
Author's personal copy
Y. Hamlaoui et al. / Materials Chemistry and Physics 113 (2009) 650–657 653
Fig. 5. Evolution of potential with deposition time for the different applied currentdensities in 0.1 M Ce(III) nitrate solutions and at room temperature.
Fig. 6. Evolution of the potential with the final pH of the 0.1 M Ce(III) nitrate solu-tions at room temperature.
are indicative of a change in the dominant reaction taking place atthe electrode [37]. It therefore appears that at weak current densi-ties the electrogeneration of base is due to the nitrate reduction andthe hydrogen evolution reactions (HER) for the first period of depo-sition (t < 200 s) then the HER may become predominant. However,at high current densities, the only reaction responsible for elec-trogeneration of base is related to the HER. In the initial stage ofdeposition, the nucleation and the growth of nuclei compete witheach other. In the case of high current density, the generation reac-tion rate is high, and the rate of nucleation exceeds that of growth ofnuclei, while, in the case of low current densities, the rate of growth
Fig. 7. DSC curves of the powders scraped off from the deposits elaborated at anapplied current density of 3 mA cm−2 for 20 min in 0.1 M Ce(NO3)3 solutions and atroom temperature.
Fig. 8. FT-IR spectra of the powders scraped off from the deposits elaborated at anapplied current density of 3 mA cm−2 for 20 min in 0.1 M Ce(NO3)3 solutions and atroom temperature.
of nuclei predominates [38]. Thus, at high current densities, the firstperiod (t < 200 s) is attributed to the formation of a deposit of fineparticle size that fully covers the surface. Therefore, the potentialstabilises rapidly. The second period can be related to the stackingof new layers (three-dimensional growth) [39,40]. However, at lowcurrent densities, the nucleation and subsequent growth of largecrystallites delay the total surface covering until 600 s.
Table 1Thermal evaluation of the melting effects observed for the indicated systems
Transformation Rinsed with the water Rinsed with ethanol
654 Y. Hamlaoui et al. / Materials Chemistry and Physics 113 (2009) 650–657
Regardless of the current density, after each electrodepositionthe solution is acidified. Fig. 6 shows the equilibrium potential ver-sus the final pH of the solution. This is mainly due to the OH−
consumption necessary for hydroxide Ce(OH)3 deposition (Eq. (12))and to proton release (Eq. (15)) in the case of CeO2 formation. Zhouet al. [41] reported that a strong OH− concentration facilitates theoxidation of Ce(III) to Ce(IV), but also this oxidation reaction isfurther accompanied by the production of hydronium ions, whichdecreases the solution pH according to the following equation:
Ce(H2O)x(OH)y(4−y)+ + H2O → CeO2·nH2O + H3O+ (15)
3.3. Effect of the rinsing media
Most of the works devoted to the cerium oxide depositedby CELD have shown that the film is composed of cerium oxy-hydroxide mixtures [7,33,42,43]. Moreover, the presence of foreignanions is not excluded when the anions also participate in the for-mation of the deposit process. Andreescu et al. [25] showed that the
CeO2 obtained by chemical precipitation from a cerium acetate pre-cursor solution and rinsed with water was more crystalline than theone rinsed with ethanol. In addition, using the same process, Raneet al. [44] found that rinsing with water eliminates less nitrates thanrinsing with ethanol. For a better comprehension of the influence ofthe rinsing method on the quality of the deposits, the incorporationof foreign anions and the nature of the obtained coating, a series ofdeposits was then realised at 3 mA cm−2 for 20 min. The films werethen scraped off, rinsed with ethanol or water and dried for at least24 h in a desiccator before analysis. The scraping of the depositis realised at least after 1 h of the deposition to avoid expansionof the lattice parameter due to non-stoichiometric cerium oxidephase transformation [45]. Then, DSC measurements were carriedout between 30 and 550 ◦C under pure N2(g) flow. The results areshown in Fig. 7 and summarised in Table 1. Three important thermalsteps during heating can be underlined.
First, the two endothermic peaks at 130 and 145 ◦C related to therelease of free water and the crystallization of the amorphous por-tion –Ce(OH)4– to crystalline CeO2 [44,46] appear in the ethanol
Fig. 9. SEM surface morphologies of the films deposited by applying (a) 0.25 mA cm−2, (b) 0.5 mA cm−2, (c) 1 mA cm−2, (d) 1.5 mA cm−2 and (e) 3 mA cm−2 for 20 min in 0.1 MCe(NO3)3 solutions and at room temperature.
Author's personal copy
Y. Hamlaoui et al. / Materials Chemistry and Physics 113 (2009) 650–657 655
rinsed deposits. In contrast, in the water rinsed deposits the sameoccurs at lower temperatures (about 75 and 104 ◦C). This may indi-cate that some R-OH bonds from ethanol are found within thedeposit, which elevate the temperature of transition and requiremore energy (higher �H). The second significant domain concernsthe weak peak at 249 ◦C that can be attributed to the decompositionof nitrates [44]. However, the peak is not observed in the depositsrinsed with ethanol and suggests that rinsing with ethanol couldeliminate more nitrates than water or that their decomposition istoo weak to be recorded. Finally, the peaks situated at tempera-tures above 300 ◦C are related in both cases to the transformation ofthe hydroxide into cerium oxide and to the subsequent crystalliza-tion or the formation of the fluorite (FCC) crystalline structure [47].Nevertheless, the presence of a weak endothermic peak during thereverse scanning indicates that the deposit crystallization is incom-plete. Therefore, from the DSC analyses, it seems that the depositsconsist of a hydrated cerium oxide with unbounded nitrates whenrinsing is carried out with water.
Fig. 8 shows the IR spectra of the deposits scraped and milled.There are no significant differences in what water concernsbetween both rinsing methods (stretching modes at 3440 cm−1 andbending mode at 1630 cm−1). The peak concerning the CO2 from theair situated at 2344 cm−1 [44] is more intense when the deposit isrinsed with water. The contact of trapped water in the deposit withCO2 can bring about the formation of carbonate ions characterizedby a peak situated at 1473 cm−1 [50–52]. This point will be dis-cussed further in the text. IR spectra also show residual nitrates inthe deposits regardless of the rinsing media with the peaks appear-ing at 1385 and 830 cm−1. The peak at 1038 cm−1 can be attributedeither to nitrate or carbonate ions. For the deposits rinsed in ethanola weak peak situated around 1140 cm−1 attributed to the vibrationband (–CH2–) of ethanol is observed [48,49]. This confirms the DSCresults that indicated partial ethanol bonding in the deposits.
From the above results, ethanol was retained as the rinsingmedium as the deposits seem to contain less nitrates and carbon-ates than the water rinsed films. Once rinsed, the deposits weredried under a hot air jet for a few seconds and finally dried in adesiccator for 24 h.
3.4. Characterisation of the deposits
Fig. 9 shows the SEM surface morphology of the cerium oxidefilms deposited at different current densities. It can be observedthat the films fully cover the surface of the steel. The microstruc-ture evolves significantly with the increase of the current densitiesas does the composition. The Ce and O contents increase suddenlywith the current density contrary to the Fe amount as depicted inFig. 10. Despite EDS is not a sensitive enough technique for lightelements, the nitrogen content is shown to decrease with increas-ing the current density. The presence of nitrogen can be related tothe residual nitrates shown by FT-IR (Fig. 8). These curves also sug-gest that there is a saturation level of the Ce concentration of about20–22 at.%. Furthermore, at higher magnifications the surface mor-phology does not change significantly with the increase of appliedcurrent density (Fig. 11). Indeed, the deposits are mostly needle-likeregardless of the applied current density but a few clusters formedby tangling of the needles appear at the lowest current densitieswhereas they are not present at higher currents. This morphologyseems to be mainly due to the interaction of hydrogen bounds [53].
However, all the deposits are cracked at applied current den-sities higher than 0.5 mA cm−2. These cracks are associated witheither the formation of gas bubbles during the process on the steelsurface when the current densities are increased, to the dehy-dration process itself [24] or from the shearing stresses betweenthe substrate and the deposit [13]. Underneath the coatings and
Fig. 10. Evolution of the surface composition of the electrodeposits with increasingthe applied current density.
between these cracks, the substrate is nevertheless coated as indi-cated by the EDS analysis as observed in the backscattered electronmode (Fig. 12) hence suggesting that cracking occur upon the depo-sition process itself. On the other hand, at current densities equalto or lower than 0.5 mA cm−2, the electrode surface shows manycorrosion products indicating an important degradation of the sub-strate.
Regardless of the applied current density, the Raman spectra onthe deposits clearly highlight the symmetric vibration of the F2gmode of Ce–O at 449 cm−1 (Fig. 13). The shift of this band can be
Fig. 11. SEM surface morphology of the ceria films deposited at applied currentdensities of (a) 0.25 mA cm−2 and (b) 3 mA cm−2, respectively, during 20 min at roomtemperature from 0.1 Ce(NO3)3 solutions.
Author's personal copy
656 Y. Hamlaoui et al. / Materials Chemistry and Physics 113 (2009) 650–657
Fig. 12. Backscattered electron image of the deposits obtained at −1.5 mA cm−2 for20 min and room temperature in 0.1 M Ce(NO3)3 solutions showing a new depositgrowing between the cracks.
Fig. 13. Raman spectra of the cerium oxide deposits elaborated at increasing appliedcurrent densities of 0.25, 0.5, 1, 1.5 and 3 mA cm−2 (bottom to up) for 20 min in 0.1 MCe(NO3)3.
strongly influenced by the particle size and the oxygen vacanciesin the deposits [24,50,54,55]. In the present study, it seems thatthe likely reduction of Ce4+ to Ce3+ in the CeO2 layer as depositioncontinues could bring about the development a large amount ofoxygen vacancies. This hypothesis is further supported by the evo-lution of the band at 600 cm−1 and the strong asymmetry observedin the Ce–O peak. These spectra also show the typical internal vibra-tion modes of the nitrate anions at 1049 and 740 cm−1 [56,57] andtherefore confirm that these have not been fully removed from thefilms in agreement with the FT-IR observations but in contrast to theDSC measurements. The XRD patterns (Fig. 14) also confirm that thecubic fluorite CeO2 phase is the major compound diffracting from
Table 2Crystallite size of CeO2 calculated by the Scherrer equation from the (1 1 1) reflec-tions of the XRD patterns
Current density (mA cm−2) Particle size (nm)
0.25 10.840.5 5.881.0 5.481.5 4.823.0 4.21
Fig. 14. XRD patterns of the cerium oxide films elaborated at increasing appliedcurrent densities of 0.25, 0.5, 1, 1.5 and 3 mA cm−2 (bottom to up) for 20 min in 0.1 MCe(NO3)3.
the deposits. No diffraction line associated with Ce(III) hydroxideis present, which is in agreement with the work of Li et al. [58].However, with the increase of the applied current density the fullwidth at half maximum of the (1 1 1) CeO2 peaks increases hencethe crystallite size of the deposits drops (Table 2) in agreement withthe studies by Wang et al. [59] on the anodic electrodeposition ofCeO2 on different substrates.
The X-ray diffraction patterns of the deposits obtained at0.25 mA cm−2 in Fig. 14 also indicate the appearance of a broadpeak near 2� = 10.7◦ and a very weak one at 23◦ that suggest thata carbonated green rust has developed as these reflections havebeen associated with the (0 0 3) and (0 0 6) crystallographic planes,respectively [60]. Corrosion of the steel substrate may have beeninitially induced by the acidity of the solution, which coupled tothe carbonation of the ceria films can be responsible for the forma-tion of the green rust. The presence of this corrosion product hasbeen confirmed by SEM/EDS and �Raman spectroscopy and hasmade the object of a separate study [61].
4. Conclusions
Relatively concentrated cerium nitrate solutions have beenemployed in this study to elaborate films of cerium-based oxideson carbon steel. The cyclic voltammetry and the potential valueshave clearly indicated that the reduction reactions occur rapidlytherefore leading to significant modification (covering) of the steelsubstrate, especially with increasing the precursor concentration.The electroanalytical behaviour suggest that an evolution fromCe(III) to Ce(IV) occurs within the solution in contact to air. Theappearance of extra peaks in the voltammograms has been there-fore ascribed to Ce(IV)–steel interactions and not to the reductionof nitrates nor of oxygen.
The current density plays a major role in the formation of thedeposits. At low current densities, the nucleation and growth of thecrystallites compete with each other. However, at higher currentdensities the growth rate seems to be faster therefore developinga film that allows stabilisation of the potential. At a later stage,stacking of new layers can occur without any significant alterationof the potential (plateau).
Rinsing with water or with ethanol changes markedly the com-position but not the morphology of the films. The deposits rinsedwith water seem to contain more nitrates and adsorbed CO2 than
Author's personal copy
Y. Hamlaoui et al. / Materials Chemistry and Physics 113 (2009) 650–657 657
the ones rinsed with ethanol. The presence of CO2 and of waterwill induce the formation of a carbonated green rust –GR(CO3
2−)–in the case of films elaborated at low current densities. However,the deposits rinsed with ethanol seem to trap R-OH bonds andthe temperature of water release is therefore increased. Regardlessof the rinsing medium and according to the DSC measurements,crystallization would be incomplete up to 550 ◦C and thereforethe deposits would still contain hydrated cerium oxides. The XRDdiffraction patterns however have shown that the deposits seem tohave the fluorite structure and are of nanometre size. The increaseof the current density would reduce the size of the crystallites.
The morphology of the deposits and the composition variesdepending on the applied current density. At low values the surfaceis not fully covered with a dense deposit. As the current den-sity increases, the surface becomes progressively covered and thedeposits show a significant network of cracks. Very likely crack-ing occurs upon the deposition process itself as the film/substrateinterface contains variable amounts of Ce. The Ce content on thescales otherwise stabilises at about 20–22 at.% with the increase incurrent density. Within the scales, the deposits have a needle-likemorphology and a few clusters appear formed by tangling of theseneedles.
Acknowledgements
This work has been carried out under during the research place-ment of Y. Hamlaoui at the LEMMA laboratory under the frameworkof the PROFAS Programme (Programme Franco-Algérien de For-mation Supérieure en France). Their financial support is greatlyindebted.
References
[1] B.R.W. Hinton, J. Alloy Compd. 180 (1992) 15.[2] A.E. Hughes, D. Ho, M. Forsyth, B.R.W. Hinton, Corros. Rev. 25 (2007) 591.[3] Y. Zhou, J.A. Switzer, J. Alloy Compd. 237 (1996) 1.[4] I. Zhitomirsky, A. Petric, Mater. Lett. 40 (1999) 263.[5] I. Zhitomirsky, A. Petric, Ceram. Int. 27 (2001) 149.[6] P. Stefanov, G. Atanasova, D. Stoychev, Ts. Marinova, Surf. Coat. Technol. 180/181
(2004) 446.[7] A.J. Davenport, H.S. Issac, J. Electrochem. Soc. 136 (1989) 1837.[8] Y. Xingwen, C. Chunan, Y. Zhiming, Z. Derui, Y. Zhongda, Corros. Sci. 43 (2001)
(2002) 208.[10] M. Dabala, K. Brunelli, E. Napolitani, M. Magrini, Surf. Coat. Technol. 172 (2003)
227.[11] C.S. Linz, S.K. Fang, J. Electrochem. Soc. 152 (2005) B54.[12] M. Balasubramaniana, C.A. Melendresa, A.N. Mansour, Thin Solid Films 347
(1999) 178.[13] L. Arurault, P. Monsang, J. Salley, R.S. Bes, Thin Solid Films 446 (2004) 75.[14] C. Wang, F. Jiang, F. Wang, Corros. Sci. 46 (2004) 75.[15] M.A. Arenas, J.J. de Damborenea, Surf. Coat. Technol. 187 (2004) 320.[16] M.A. Arenas, J.J. de Damborenea, Electrochim. Acta 48 (2003) 3693.
[17] A. Amedeh, B. Pahlevani, S. Heshmati-Manesh, Corros. Sci. 44 (2002) 2321.[18] M.F. Montemor, A.M. Simoes, M.G.S. Ferreira, Prog. Org. Coat. 44 (2002) 111.[19] C. Motte, N. Maury, M.-G. Olivier, J.-P. Petitjean, J.-F. Willem, Surf. Coat. Technol.
200 (2005) 2366.[20] Y. Kobayashi, Y. Fujiwara, Electrochim. Acta 51 (2006) 4236.[21] K. Aramaki, Corros. Sci. 48 (2006) 766.[22] M.G.S. Ferreira, R.G. Duarte, M.F. Montemor, A.M.P. Simoes, Electrochim. Acta
49 (2004) 2927.[23] S. Berna, F.J. Botana, J.J. Calvino, M. Marcos, J.A. Pérez-Omil, H. Vidal, J. Alloy
Compd. 225 (1995) 638.[24] J. Creus, F. Brezault, C. Rebere, M. Gadouleau, Surf. Coat. Technol. 200 (2006)
4636.[25] D. Andreescu, E. Matijevic, D.V. Goia, Colloid Surf. A 291 (2006) 93.[26] H.S. Kang, Y.C. Kang, H.Y. Koo, S.H. Ju, D.Y. Kim, S.K. Hong, J.R. Sohn, K.Y. Jung,
S.B. Park, Mater. Sci. Eng. 127B (2006) 99.[27] R.C. Korosec, I.K. Skofic, N. Bukovec, Thermochim. Acta 411 (2004) 211.[28] D. Sylla, Electrodéposition d’alliages Zn-Mn en milieu alcalin, Ph.D. Thesis, Uni-
versité de La Rochelle, 2004 (in French).[29] E.J. Ruiz, R. Ortega-Borges, L.A. Godinez, Electrochim. Acta 52 (2006) 914.[30] S.A. Hayes, P. Yu, T.J. O’Keefe, M.J. O’Keefe, J.O. Stoffer, J. Electrochem. Soc. 149
(2002) C623.[31] P. Yu, T.J. O’Keefe, J. Electrochem. Soc. 153 (2006) C80.[32] T.D. Golden, A.Q. Wang, J. Electrochem. Soc. 150 (2003) C621.[33] A.J. Aldykiewicz Jr., A.J. Davenport, H.S. Isaacs, J. Electrochem. Soc. 143 (1996)
147.[34] F.-B. Li, R.C. Newman, G.E. Thompson, Electrochim. Acta 42 (1997) 2455.[35] S. Kanakaraju, S. Mohan, A.K. Sood, Thin Solid Films 305 (1997) 191.[36] F.-B. Li, G.E. Thompson, J. Electrochem. Soc. 146 (1999) 1809.[37] G.H.A. Therese, P.V. Kamath, J. Appl. Electrochem. 28 (1998) 539.[38] Y. Matsumoto, J. Hombo, J. Electroanal. Chem. 279 (1990) 331.[39] J.O’M. Bockris, G.A. Razumney, Fundamental Aspects of Electrocrystallization,
Plenum Press, New York, 1967, p. 27.[40] J.W. Dini, Electrodeposition, in: The Materials Science of Coatings and Substrate,
Noyes Publications, 1992, p. 195.[41] X.D. Zhou, W. Huebner, H.U. Anderson, Appl. Phys. Lett. 80 (2002) 139.[42] P.J.K. Paterson, Corros. Sci. 38 (1996) 1957.[43] M.F. Montemor, A.M. Simões, M.G.S. Ferreira, Prog. Org. Coat. 43 (2001) 274.[44] N. Rane, H. Zou, G. Buelna, J.Y.S. Lin, J. Membr. Sci. 256 (2005) 89.[45] A. Kossoy, J.P. Nair, E. Wachtel, I. Lubomirsky, J. Fleig, O. Maie, J. Electroceram.
13 (2004) 605.[46] B.N. Sivasankar, J.R. Sharmila, R. Saratha, S. Govindarajan, Thermochim. Acta
417 (2004) 107.[47] R.C. Korosec, I.K. Skofic, N. Bukovec, Thermochim. Acta 411 (2004) 211.[48] S. Nagarajan, S. Sudhakar, K.S.V. Srinivasan, Pure Appl. Chem. 70 (1998) 1245.[49] M. Rozenberg, A. Loewenschuss, Y. Marcus, Spectrochim. Acta A54 (1998)
1819.[50] A. Siokou, S. Ntais, V. Dracopoulos, S. Papaefthimiou, G. Leftheriotis, P. Yianoulis,
Thin Solid Films 514 (2006) 87.[51] B. Djuriclica, S. Pickering, J. Eur. Ceram. Soc. 19 (1999) 1925.[52] C. Li, Y. Sakata, T. Arai, K. Domer, K. Maruya, T. Onishi, J. Chem. Soc. 85 (1989)
929.[53] H.Y. Chang, H.I. Chen, J. Cryst. Growth 283 (2005) 457.[54] S. Wang, W. Wang, J. Zuob, Y. Qian, Mater. Chem. Phys. 68 (2001) 246.[55] I. Kosacki, T. Suzuki, H.U. Anderson, P. Colomban, Solid State Ionics 149 (2002)
99.[56] M.I. Cabac, M.I. de Barros Marques, M.A. Marques, A.M. Gaspar, M.L. de Almeida,
J. Mol. Liquids 117 (2005) 69.[57] M.K. Marchewka, A. Pietraszko, J. Phys. Chem. Solids 66 (2005) 1039.[58] Y.X. Li, X.Z. Zhou, Y. Wang, X.Y. Zou, Mater. Lett. 58 (2004) 245.[59] A.Q. Wang, P. Punchaipetch, R.M. Wallace, T.D. Golden, J. Vac. Sci. Technol. B23
(2003) 1169–1175.[60] Ph. Refait, M. Reffass, J. Landoulsi, R. Sabot, M. Jeannin, Colloid Surf. A 299 (2007)
29.[61] Y. Hamlaoui, F. Pedraza, L. Tifouti, Corr. Sci. 50 (2008) 2182.
