Corrosion behavior of a cerium-based conversion coating on alumina borate whisker-reinforced AA6061 composite pre-treated by hydrogen fluoride

S.W. Tang,J. Hu,X.H. Zhao

DOI: https://doi.org/10.1016/j.corsci.2011.04.026

IF: 7.72

2011-01-01

Corrosion Science

Abstract:Cerium-based conversion coatings have been made on alumina borate whisker-reinforced AA6061 composite pre-treated with hydrogen fluoride. The pre-treatment prior to coating significantly affects the microstructure and the corrosion behavior of the ultimate coated composites. The coating deposited on composite pre-treated for 20 s provides better corrosion protection than that pre-treated for 180 s. A self-healing effect is observed on the coated composite with pre-treatment for 20 s in the early stage of immersion, while no self-healing effect appears with pre-treatment for 180 s. It is largely because the morphologies of the coatings are remarkably different under the two pre-treatment conditions. Keywords A. Aluminum A. Metal matrix composite B. EIS B. Polarization C. Passive film 1 Introduction Aluminum metal matrix composites (AMMCs) have great superiority in applications where low density and high mechanical performance are required, such as in the motorcar and aerospace industry [1–3] . Though the mechanical performance can be improved by introducing reinforcements, the presence of reinforcements is detrimental to corrosion resistance of the matrix, which has been reported in the literature [4–7] . Therefore, corrosion is one of the most critical problems for the composites, which need to be solved before the application of AMMCs in aggressive environment. Many protective methods, such as anodizing [8] and laser melting [9] , have been used to improve the corrosion resistance of AMMCs. Cerium-based conversion coatings (CeCC) have drawn more and more attention in anticorrosion area for over 20 years, as it is one of the most prospective replacements of carcinogenic chromium-based conversion coatings for its environmentally friendly property, low cost and easy operation. In most of the studies, CeCC were mainly deposited on aluminum alloys [10–12] , and effective protection was achieved. Cerium can also provide additional corrosion resistance for AMMCs. Neila and Garrard [13] studied the effect of cerium inhibition on silicon carbide fibers reinforced AA6061 composite in 3.5% NaCl aqueous solution, and it was shown that the pit initiation was retarded for approximately 1–2 days. Pardo et al. [14] studied the effect of Ce protection on pitting corrosion behavior of silicon carbide particles (SiC p ) reinforced aluminum composite. They found that Ce-treated surfaces presented better corrosion behavior than uncoated one. The protection mechanism is blocking of the cathodic reaction of the corrosion process by forming a layer of Ce 2 O 3 over intermetallic compounds, SiC p and Si eutectic. Many factors affect the microstructure and the corrosion resistance of conversion coatings [15–17] , such as the condition of the conversion solution, the coating methods and the post-treatment. Johnson et al. [17] studied the microstructure of CeCC obtained on 7075-T6 by using transmission electron microscope (TEM) and electron diffraction analysis. The results revealed that the electrolytic and spray coatings developed the same crystal structure, possibly Ce 7 O 12 , while the dip-immersion coating developed a different structure that had not been reported in the literature. Heller et al. [18] found that the duration and temperature of phosphate post-treatment affected the corrosion performance of CeCC on 2024-T3. Moreover, the concentrations of Ce salt and H 2 O 2 were found to have a strong effect on the corrosion behavior of CeCC deposited on AMMCs [19,20] . Pre-treatment has also proved to be a very important procedure in time of preparing CeCC. Johnson et al. [15] characterized the microstructure of cerium oxide conversion coatings prepared under different deposition conditions using TEM. The cleaning temperature of the alkaline cleaner was found to greatly affect the microstructure of CeCC. For coatings obtained from the same coating bath, cracks were more extensive for samples cleaned at 43 °C than those for samples cleaned at 55 °C. Brunelli et al. [21] found that CeCC on magnesium and magnesium alloy exhibited better adherence and higher corrosion resistance when HCl pre-treatment was performed. Maddela et al. [22] studied the effects of acid (sulfuric acid) and alkaline (sodium carbonate) pre-treatment on the morphology and corrosion resistance of CeCC on AZ91D magnesium alloy. They found that the surface pre-treatment significantly influenced the CeCC, especially coating morphology and crack density. A combination of acid and alkaline cleaning resulted in a better salt fog performance compared with individual acid or alkaline cleaning. Hamdy et al. [23] studied effects of different pre-treatments on the corrosion resistance of the CeCC on AMMCs. The coating deposited on the sample treated with boiling water after alkaline etching exhibited higher corrosion resistance. de Frutos and co-workers [24] investigated the role of pre-treatment in the formation of CeCC for the protection of 7075-T6 and 2024-T3 alloy, and they found that the formation of a uniform coating was dependent upon the time of de-smutting. Previously, the effects of hydrogen fluoride (HF) pre-treatment on the microstructure and the formation mechanism of CeCC on the alumina borate whisker-reinforced AA6061 composite (Al 18 B 4 O 33w /AA6061) were studied by Hu et al. [25] . It was found that the pre-treatment had a strong influence on the formation, growth and morphology of coatings. Relatively, little is known about the effects of HF pre-treatment on the effectiveness of corrosion protection of Al 18 B 4 O 33w /AA6061 in NaCl solution. In the present work, we will report on the microstructure and electrochemical characteristics of CeCC coating on Al 18 B 4 O 33w /AA6061 with HF pre-treatment for different time. The corrosion mechanisms of coated composites will be also discussed in detail. 2 Experimental Al 18 B 4 O 33w /AA6061 composite used in this research was fabricated by squeeze casting. Squeeze casting is a simple fabrication method with which the wet ability between matrix and reinforcement can be increased, and thus good mechanical performance of MMCs can be obtained. The volume content of whiskers in the composite was 20%. After the surfaces of samples were polished with diamond paste and washed with ultrasonic acetone, HF pre-treatment of the samples was carried out. The samples were respectively immersed in HF solution at room temperature for 10, 20, 60 and 180 s. Then the pre-treated samples were rinsed with deionised water. HF diluted from analytic reagent for pre-treatment was 2.5 mol/L. The pre-treated samples were soaked in conversion solution for 10 min at 30 °C. The conversion contained CeCl 3 (0.02 mol/L) and H 2 O 2 (100 ml/L). The solution pH was adjusted to 2.85 with NaAc/HAc buffer solution and was kept constant during the coating process. CeCC obtained on the composite with different pre-treatment time were abbreviated as H10, H20, H60 and H180, respectively. The coated samples were rinsed with distilled water and dried in air for more than 3 days. The microstructures of the CeCC were studied by SEM (Hitachi S-4700 scanning electron microscope operated at 20 kV) and TEM (Philips CM-12 system with electron beam energy of 120 kV). The cross-sectional sample of TEM was first prepared by a ‘sandwich’ method [26] , and then fabricated by means of mechanical polishing, dimpling and ion milling. The phases of the coating were identified by grazing incidence X-ray diffract meter (Philips X’pert) using Cu Kα radiation with a fixed grazing incidence angle of 3° (working at 40 kV, 35 mA). Polarization measurements were performed in aerated 3.5% NaCl aqueous solution at room temperature and recorded with a M273A potentiostat (Princeton EG&G). Three-electrode cell configuration was used: working electrode was the composite sample, auxiliary electrode was graphite rod and reference electrode was saturated calomel electrode. Before starting the scanning, the sample was kept at open circuit potential (OCP) for 300 s. Then the measurement started from −200 mV vs. OCP at a sweep rate of 0.5 mV/s, and ended when the anode current density was close to 0.01 A/cm 2 . EIS measurements were performed by using a Princeton EG&G M273A potentiostat (with a Perkin Elmer M5210 lock in amplifier). In order to decrease the corrosion rate, NaCl solution with a low concentration (0.005 mol/L) was selected in EIS measurement. So the corrosion process could be monitored in detail. The impedance measurements were carried out from 10 kHz to 0.01 Hz frequency with four steps per decade of frequency, and all the data points were recorded at open circuit potential with applied 10 mV sinusoidal perturbation. EIS results were analyzed by using ZSimpWin software. 3 Results 3.1 Microstructure of the CeCC obtained on the pre-treated composites Fig. 1 shows the morphologies of composites pre-treated by HF. As shown in Fig. 1 a, c, e and g, more and more whiskers are exposed on the surface of the composite with the increase of pre-treatment time. The composite shows many etch pits on the matrix when it is pre-treated for 10 and 20 s, as shown in Fig. 1 b and d. The matrix is seriously etched and the whiskers are exposed more completely when the composite is pre-treated for 60 and 180 s ( Fig. 1 f and h). Fig. 2 shows the morphologies of CeCC obtained on the pre-treated surfaces of Al 18 B 4 O 33w /AA6061. From Fig. 2 , it can be observed that surface morphologies of the coatings are very different. Typical appearance of CeCC can hardly be observed in Fig. 2 a and c. Substrate topographies left by chemical pre-treatment, such as whiskers and etch pits, can be clearly seen on the surfaces of H10 and H20. A typical dry-mud-like morphology appears in Fig. 2 e and g. Whiskers can be clearly seen on the surfaces of H60 and H180. The corresponding magnified morphologies of coatings obtained on the pre-treated composites are shown in Fig. 2 b, d, f and h. As is seen, all the coatings are not only deposited over aluminum matrix, but also deposited over whiskers. It is revealed that all the pre-treated surfaces are covered by granular coatings, and the coating particles are ultrafine. Moreover, it is found that the coating particles formed on H60 and H180 are much bigger than those formed on H10 and H20, which indicates that the surface pre-treatment for a longer time can accelerate the growth of coating particles. The microstructures of CeCC obtained on the surfaces of Al 18 B 4 O 33w /AA6061 pre-treated for 20 and 180 s were investigated by TEM [25] . The TEM images and the selected area electron diffraction patterns clearly indicated that under both pre-treatment conditions the deposited coatings were composed of CeO 2 nanoparticles with a sphere-like morphology. Fig. 3 displays the XRD patterns measured under the grazing incidence condition. For H20, only diffraction peaks of Al and Al 18 B 4 O 33 appear in the XRD pattern. For H180, a widen diffraction peak (2 θ = 28.5°) corresponding to the position of (1 1 1) reflection of CeO 2 appears at the XRD profile. These phenomena suggest that the thickness of the coatings on the pre-treated surfaces increases with the pre-treatment time owing to the same X-ray penetration depth with the same incident beam angle. Since the X-ray penetration is deep, hence the information of the coatings can be covered by that of the composite. When the coating is thick enough, it can be just detected by X-ray. Therefore, a thick coating may form on the surface pre-treated for 180 s. Moreover, the presence of broad diffraction peak in XRD pattern may be corresponding to nano-crystalline or amorphous CeO 2 . Fig. 4 presents cross-sectional TEM characterization of the coatings on H20 and H180 samples. As is expected, a thin coating (about 100 nm) formed on the sample pre-treated for 20 s. There is neither defect at the interface between the coating and the substrate nor cracks in the coating ( Fig. 4 a). A correspondingly thick coating (about 200 nm) appears on the sample pre-treated for 180 s. Cracks also appears on the surface of H180, as shown in Fig. 4 b. It indicates that H180 has a higher depositing rate. Since the longer the pre-treatment time is, the higher the activity of surfaces is and the quicker the nucleation rate of coatings on pre-treated surfaces is. Thus the longer the pre-treatment time is, the thicker the coating forms on the pre-treated surface under the same deposit condition are. 3.2 Potentiodynamic polarization behavior of the coated composites Fig. 5 shows the potentiodynamic polarization curves of the pre-treated composites. As reference, the curve corresponding to the as-cast composite is also presented in this figure. The pre-treated composites do not show any obvious difference on their potentiodynamic polarization curves. It can be found that the corrosion potential was decreased by about 40 mV after HF etched. The cathode current density of the pre-treated composite is a little higher than that of as-cast composite. These phenomena indicates that the pre-treatment decreases he corrosion resistance of the composite. Potentiodynamic polarization curves for coated composites are presented in Fig. 6 . As reference, a curve corresponding to an uncoated composite is also included in this figure. The corrosion potential of the uncoated composite is around −685 mV, and there is no passivation area on its anode polarization curve. The cathode current density of this composite is between 3 × 10 −7 and 6 × 10 −6 A/cm 2 . From Fig. 6 , it is found that the position and the shape of polarization curves of the coated composites change. These curves show that the pre-treatment time prior to CeCC has a strong influence on the corrosion behavior of the coated composites. The anodic current density increases slowly with the potential, which causes a passivation region in the anodic polarization curves of H10 and H20. Higher pitting potentials of H10 and H20 are presented. The pitting potential is increased to −600 mV for H10 and −380 mV for H20. The results show that the pitting resistance of the composite can be improved by either of the coatings. No passivation region appeared in the anodic polarization curves of the H60 and H180, and the pitting potentials of H60 and H180 do not almost change relative to that of the as-cast composite. It is also interesting to note that the current density shift to lower value compared with the current density obtained from the uncoated sample. It indicates that the electrochemical activity of the system decreases. Moreover, the most coated samples show lower cathode current densities, about one order of magnitude lower than that of the uncoated one. It confirms the cathode inhibiting effect of CeCC [14] . The corrosion morphologies of the uncoated and coated composites after polarization measurement are shown in Fig. 7 . Fig. 7 a is the corroded surface morphology of the uncoated sample. It is found that a lot of pits, as shown on the top of Fig. 7 a, exist on the surface of the uncoated sample. At higher magnifications a large number of whiskers have been found in the pits. Fig. 7 b and c are the corroded surface morphologies of H20 and H180, respectively. The density of pits on H20 is lower than that on the uncoated sample, which suggests that the susceptibility of the coated composite to localized corrosion is lower. Some pits can be seen on the corroded surface of H180. Though whiskers can also be found in the pits, it is not corroded as severely as the uncoated sample. The protection effect of H180 is lower than that of H20. 3.3 EIS of the coated composites Fig. 8 depicts the evolution of the impedance spectra of H20 and H180 which are immerged in 0.005 mol/L NaCl. As is seen in Nyquist plot, H20 only shows a capacitance loop before 100 h. In the early stage of immersion (⩽10 h), the radius of the capacitance loop increases with the immersion time. Subsequently, the radius of the capacitance loop decreases continuously. The radius of the loop is about 0.25 MΩ cm 2 at 2.5 h, and it reaches a higher value (more than 1 MΩ cm 2 ) at 10 h. Then the capacitance loop shrinks after 10 h till two capacitance loops exhibit on the Nyquist plot at 100 h. Bode plot of H20 shows that the phase angles do not almost change in the immersion process. Two time constants can be clearly seen in the spectra at 100 h. It is worthwhile noting that H180 shows a capacitance loop just in the early stage of immersion (⩽10 h), and the radius of the capacitance loop obviously decreases with the increase of the immersion time. Two capacitance loops appears on the Nyquist plot of H180 at 24 h. With the immersion time increasing from 24 to 100 h, the two loops show almost no change either in size or in shape. Bode plot of H180 shows that the phase angles have decreased obviously from 2.5 to 24 h, and the shape of the plot distinctly changes. A distinct time constant appears at low frequency at 24 h, and this time constant can also be seen from 48 to 100 h. Bode plot sustains its shape from 24 to 100 h. Fig. 9 shows the impedance modulus of H20 and H180 at 0.01 Hz. H20 presents increased modulus value at 10 h which is followed by a successive decrease till 100 h. However, the modulus value of H180 keeps on decreasing until the immersion time is 24 h. During the succeeding immersion, the impedance values almost maintain a certain level. The EIS of H20 and H180 can be fitted by the equivalent circuit in Fig. 10 , and the fitting parameters are shown in Fig. 11 . The simulation curves (solid lines as shown in Fig. 8 ) present good fitting results of the experimental data. R sol is the resistance of the solution, C coat is coating capacitance, R coat is coating resistance, C dl is double layer capacitance, and R pol is polarization resistance. CPE (constant phase angle element, Q ) is employed to simulate C coat and C dl , because the pure capacitance cannot give a perfect fitting result in some cases. As is seen in Fig. 11 , H20 has higher values of both R pol and R coat than that of H180 in the immersion process. The differences of R coat are mainly caused by the different numbers of defects on CeCC. H180 has more paths for the penetration of Cl − , so it presents a lower value of R coat . The lower value of R pol of H180 is probably caused by the longer pre-treatment. The sharp increase of C dl (from 10 to 24 h for H180 and from 72 to 100 h for H20) indicates the corrosion of the substrate. Evidently, the corrosion of the substrate can appear on the H180 earlier. 4 Discussion The investigation indicates that the difference in the number of cathodic sites present on the substrate surface is caused by different pre-treatment time. The number of cathode sites influences the formation, growth, morphology and granular size of the conversion coating on the composite surface. There is no denying that the pre-treatment prior to CeCC will affect the corrosion behavior of the coated composites without failure. When the composite is pre-treated by HF, matrix aluminum can be dissolved. The whiskers and defects (crevices or pores) in matrix appear on the surface of the composite. As the pre-treatment time increases, more aluminum can be etched and more whiskers and defects of matrix can emerge on the pre-treated surfaces. The longer the pre-treatment time prior to the coating is, the larger the number of the micro-cells in the composite becomes. Then a higher activation degree of the surface is obtained. The deposition of Ce-coatings is possible because of the chemical heterogeneity of the substrate, which establishes the local electrochemical cells. When the pre-treatment time is short (10 or 20 s), CeCC can uniformly precipitate on the pre-treated surface, and then ‘defect-free’ CeCC appears on the pre-treated surface. A compact and uniform conversion coating morphology can be seen on the surfaces of H10 and H20 (as shown in Fig. 2a–d ). The substrate topographies left by chemical pre-treatment can also be seen because of the thin coating which forms on the surface ( Fig. 4 a). When the pre-treatment time is long (60 or 180 s), cathode sites of the whole surface of the composite is highly activated. With the number of cathodic sites increasing, the precipitation reaction occurs more rapidly, and thus a thick coating appears on the surface ( Fig. 4 b). In this case, the corrosion resistance of the coated composites should increase with the pre-treatment time, because the substrate surfaces have been covered with the thicker coatings already. But in practice, H60 and H180 exhibit poor corrosion resistance (as shown in Fig. 6 ), and it seems that the poor corrosion resistance is caused by cracked ‘dry mud-like’ structure of CeCC presented on the surfaces of H60 and H180 ( Fig. 2 e–h). A faster deposition rate of the coating results in the increase of both the thickness and the unevenness of the coating. In addition, the faster deposition rate of the coating leads to larger inner stress of the coating, so more cracks appear on the surface of H180. The cracks can also be caused by the heterogeneous nucleation and the growth of coatings, and thus more defects will appear in the coatings. Hence the corrosion resistance of the composite can be improved by ‘defect-free’ CeCC. It is also confirmed by the corrosion morphologies in Fig. 7 . H180 shows more corrosion areas on its corroded surface, while H20 shows only one. The improved corrosion resistance for H20 is attributed to the existence of a ‘defect-free’ CeCC on the surface of the composite. Though a ‘defect-free’ CeCC also appears on the surface of H10, the corrosion resistance of H10 is lower than that of H20. According to the analysis above, one can deduce that a thinner coating may form on the surface of H10 because the substrate surface is pre-treated for a short time. Thus the protection effect decreases. H60 exhibits poor corrosion resistance owing to cracked CeCC forming on the surface. EIS results indicate that the corrosion mechanism of H20 and H180 are different. Though their Nyquist plots exhibit the same evolution character (from one capacitance loop to two capacitance loops) with the increase of the immersion time, their evolution rules are different. For H20, the capacitance loop increases in the beginning and decreases later. For H180, the capacitance loop decreases all the while. With the increase of the immersion time, two loops can be clearly observed in both coated samples (as shown in Fig. 8 ). One is at high frequency corresponding to the CeCC, the other is at low frequency corresponding to the double-layer. It is worth noting that two capacitance loops appears on Nyquist plots at 100 h for H20 and 24 h for H180, respectively, which represents that H20 has a higher corrosion resistance in corrosion media. When the coated sample is immerged into the corrosion media, the dissolution of the coating and the penetration of Cl − along the defects of the coating start at the same time. With the dissolution of the coating, Ce ions can be released to the coating surface. The free Ce ions have self-healing effect, which has been proved by many researchers [27–29] . Ce–silane bi-layer containing free Ce 4+ ions and silica nanoparticles showed some self-healing effect, which was investigated by scanning vibrating electrode technique [27] . Paussa et al. [28] evaluated the behavior of cerium as corrosion inhibitor in silane-based hybrid films on AA2024. They provided strong evidence that the behavior of inhibited system could be related to migration of Ce species to the substrate/coating interface. Rosero-Navarro et al. [29] studied the hybrid sol–gel coatings with Ce 3+ incorporated by using EIS. They found that the increase of impedance at 0.01 Hz with immersion time from the very beginning could be attributed to the self-healing effect provided by cerium ions. Electrochemical impedance spectroscopy has been demonstrated as a useful technique for the study of self-healing process in the protective coatings on metallic substrates by Zheludkevich [30] . Though no artificial defects are made in our experiment to study the self-healing effect, the inherent defects inevitably exist not only on cracked CeCC (H60 and H180) but also on ‘defect-free’ CeCC (H10 and H20). In the immersion process, some free Ce ions diffuse into the defects of the coatings to form the cerium-containing deposits. Therefore, the impedance modulus of H20 increases at the very beginning of the immersion (from 2.5 to 10 h), which is caused by the self-healing effect of CeCC around micro defects. Subsequently the impedance modulus decreases successively, which suggests that the corrosion of the substrate cannot be inhibited effectively. Though the free Ce ions have self-healing effect, once the coating is pierced by Cl − , the protection ability of the coating decreases. The impedance modulus decreases and the capacitance loop at low frequency appears, because H20 is too thin to provide enough self-healing media during the long-term immersion. The impedance modulus of H180 decreases continuously in the whole immersion period, and no self-healing effect is observed on H180. H180 cannot repair itself due to a mass of cracks present on the surface, which provides more paths for the penetration of Cl − , so the capacitance loop at low frequency appears earlier. 5 Conclusions The CeCC improves the corrosion resistance of the composite because it inhibits the anodic and cathodic reaction rate in a chloride-containing environment. The pre-treatment prior to coatings has a significant effect on the microstructure and the corrosion behavior of CeCC-coated Al 18 B 4 O 33w /AA6061 composite. EIS measurements clarify the change of the corrosion mechanism. Nyquist plots evolve from one capacitance loop to two capacitance loops with the increase of immersion time. Two capacitance loops appear on Nyquist plots at 100 h for H20 and 24 h for H180. It also indicates that H20 has higher values of both R pol and R coat than that of H180 in the immersion process. The increase of impedance modulus due to a self-healing effect is observed in the early immersion stage of H20. No self-healing effect is observed on H180. The corrosion resistance of the composite can be improved by a ‘defect-free’ CeCC. Cracked CeCC forming on the surface of composite is responsible for the poor corrosion resistance of the composite. References [1] L. Bertolini M.F. Brunella S. Candiani Corrosion behavior of a particulate metal–matrix composite Corrosion 55 1999 422 431 [2] B. Dikici M. Gavgali C. Tekmen Corrosion behavior of an artificially aged (T6) Al–Si–Mg-based metal matrix composite J. Compos. Mater. 40 2006 1259 1269 [3] I.A. Ibrahim F.A. Mohamed E.J. Lavernia Particulate reinforced metal matrix composites – a review J. Mater. Sci. 26 1991 1137 1156 [4] L.H. Hihara R.M. Latanision Galvanic corrosion of aluminium–matrix composites Corrosion 48 1992 546 552 [5] A.J. Trowsdale B. Noble S.J. Harris I.S.R. Gibbins G.E. Thompson G.C. Wood The influence of silicon carbide reinforcement on the pitting behaviour of aluminium Corros. Sci. 38 1996 177 191 [6] B. Dikici C. Tekmen O. Yigit M. Gavgali U. Cocen Detrimental effect of particle sol–gel coating on the corrosion behavior of A380–SiC composite Corros. Sci. 51 2009 469 476 [7] A. Pardo M.C. Merino A. Urena S. Merino F. Viejo M. Campo Effect of reinforcement coating on corrosion behavior of AA6061/SiC/20p composite in high relative humidity environments Corrosion 60 2004 945 953 [8] C.L. He Q. Zhou J.T. Liu X.W. Geng Q.K. Cai Effect of size of reinforcement on thickness of anodized coatings on SiC/Al matrix composites Mater. Lett. 62 2008 2441 2443 [9] X.M. Zhang H.C. Man T.M. Yue Corrosion properties of excimer laser surface treated Al–SiC metal matrix composite Scr. Mater. 35 1996 1095 1100 [10] D.R. Arnott N.E. Ryan B.R.W. Hinton B.A. Sexton A.E. Hughes Auger and XPS studies of cerium corrosion inhibition on 7075-aluminium alloy Appl. Surf. Sci. 22–3 1985 236 251 [11] A.J. Aldykiewicz H.S. Isaacs A.J. Davenport The investigation of cerium as a cathodic inhibitor for aluminium–copper alloys J. Electrochem. Soc. 142 1995 3342 3350 [12] M. Dabala E. Ramous M. Magrini Corrosion resistance of cerium-based chemical conversion coatings on AA5083 aluminium alloy Mater. Corros. 55 2004 381 386 [13] W. Neila C. Garrard The corrosion behaviour of aluminium–silicon carbide composites in aerated 3.5% sodium chloride Corros. Sci. 36 1993 837 851 [14] A. Pardo M.C. Merino R. Arrabal F. Viejo M. Carboneras J.A. Munoz Influence of Ce surface treatments on corrosion behaviour of A3xx.x/SiCp composites in 3.5 wt.% NaCl Corros. Sci. 48 2006 3035 3048 [15] B.Y. Johnson J. Edington M.J. O’Keefe Effect of coating parameters on the microstructure of cerium oxide conversion coatings Mater. Sci. Eng. A-Struct. 361 2003 225 231 [16] A. Pardo M.C. Merino R. Arrabal F. Viejo M. Carboneras Improvement of corrosion behavior of A3xx.x/SiCp composites in 3.5 wt.% NaCl solution by Ce surface coatings J. Electrochem. Soc. 153 2006 B52 B60 [17] B.Y. Johnson J. Edington A. Williams M.J. O’Keefe Microstructural characteristics of cerium oxide conversion coatings obtained by various aqueous deposition methods Mater. Charact. 54 2005 41 48 [18] D.K. Heller W.G. Fahrenholtz M.J. O’Keefe The effect of post-treatment time and temperature on cerium-based conversion coatings on Al 2024-T3 Corros. Sci. 52 2010 360 368 [19] J. Hu X.H. Zhao S.W. Tang W.C. Ren Z.Y. Zhang Corrosion resistance of cerium-based conversion coatings on alumina borate whisker reinforced AA6061 composite Appl. Surf. Sci. 253 2007 8879 8884 [20] J. Hu X.H. Zhao S.W. Tang M.R. Sun Corrosion protection of aluminium borate whisker reinforced AA6061 composite by cerium oxide-based conversion coating Surf. Coat. Technol. 201 2006 3814 3818 [21] K. Brunelli M. Dabala I. Calliari M. Magrini Effect of HCl pre-treatment on corrosion resistance of cerium-based conversion coatings on magnesium and magnesium alloys Corros. Sci. 47 2005 989 1000 [22] S. Maddela M.J. O’Keefe Y.M. Wang H.H. Kuo Influence of surface pretreatment on coating morphology and corrosion performance of cerium-based conversion coatings on AZ91D alloy Corrosion 66 2010 1 8 [23] A.S. Hamdy A.M. Beccaria P. Traverso Corrosion protection of aluminium metal–matrix composites by cerium conversion coatings Surf. Interface Anal. 34 2002 171 175 [24] A. de Frutos M.A. Arenas Y. Liu P. Skeldon G.E. Thompson J. De Damborenea A. Conde Influence of pre-treatments in cerium conversion treatment of AA2024-T3 and 7075-T6 alloys Surf. Coat. Technol. 202 2008 3797 3807 [25] J. Hu S.W. Tang Z.Y. Zhang Microstructure and formation mechanism of cerium conversion coating on alumina borate whisker-reinforced AA6061 composite Corros. Sci. 50 2008 3185 3192 [26] C.S. Lin C.Y. Lee W.C. Li Y.S. Chen G.N. Fang Formation of phosphate/permanganate conversion coating on AZ31 magnesium alloy J. Electrochem. Soc. 153 2006 B90 B96 [27] L.M. Palomino P.H. Suegama I.V. Aoki M.F. Montemor H.G. De Melo Electrochemical study of modified cerium–silane bi-layer on Al alloy 2024-T3 Corros. Sci. 51 2009 1238 1250 [28] L. Paussa N.C. Rosero-Navarro F. Andreatta Y. Castro A. Duran M. Aparicio L. Fedrizzi Inhibition effect of cerium in hybrid sol–gel films on aluminium alloy AA2024 Surf. Interface Anal. 42 2010 299 305 [29] N.C. Rosero-Navarro S.A. Pellice A. Duran M. Aparicio Effects of Ce-containing sol–gel coatings reinforced with SiO 2 nanoparticles on the protection of AA2024 Corros. Sci. 50 2008 1283 1291 [30] M.L. Zheludkevich K.A. Yasakau A.C. Bastos O. Karavai M.G.S. Ferreira On the application of electrochemical impedance spectroscopy to study the self-healing properties of protective coatings Electrochem. Commun. 9 2007 2622 2628