Corrosion resistance performance of cerium doped silica sol-gel coatings on 304L stainless steel Roohangiz ZANDI ZAND1, Kim VERBEKEN2, Annemie ADRIAENS1,* 1 Department of Analytical Chemistry, Ghent University, Krijgslaan 281-S12, B-9000, Ghent, Belgium 2 Department of Materials Science and Engineering, Ghent University, Technologiepark 903, B-9052 Zwijnaarde (Ghent), Belgium Abstract The aim of this work is the synthesis and investigation of silane based organic–inorganic hybrid coatings, which can be used to improve the corrosion performance of steel structures subjected to a marine environment. The silane based sol-gel coatings were prepared by dip coating 304L stainless steel in a solution of organically modified silica sol made through hydrolysis and condensation of 3-glycidoxypropyl-trimethoxysilane (GPTMS) as precursor and bisphenol A (BPA) as a cross-linking agent in an acid catalysed condition. The influence of the addition of cerium and the use of bisphenol A as a cross-linking agent on the microscopic features and morphology as well as on the corrosion resistance of the coatings were examined using Fourier transform infrared spectroscopy (FTIR), scanning electron microscopy (SEM), neutral salt spray tests, potentiodynamic polarization and electrochemical impedance techniques. Results show that cerium modified nano-hybrid coatings exhibit a superior corrosion inhibition performance to that displayed by silica hybrid coatings. Additionally, data showed that the bisphenol A as a cross-linking agent has a significant effect on the morphology and corrosion resistance of the cerium doped silica coating. Omitting the use of bisphenol A causes the creation of defects/cracks in the coating, thereby promoting diffusion of the aggressive electrolyte towards the substrate and decreasing the corrosion resistance of the coating. Keywords: 304L stainless steel, cerium, bisphenol A, corrosion inhibitor, silane, cross-linking agent * Corresponding Author: Tel: +32 9 264 4826, fax: +32 9 264 4960, Email address: annemie.adriaens@ugent.be (Annemie Adriaens) 1 1. Introduction AISI 304 stainless steel and other corrosion resistant alloys are widely used in different industrial fields because of their mechanical and corrosion inhibition properties [1] but are, nevertheless, susceptible to localized corrosion in the presence of chlorides or other aggressive ions [2-3]. The aggressiveness of the chloride ion is due to its small size, high diffusivity and strong acidic anionic nature [4]. A droplet of salt solution on the surface, for instance, loses water due to a temperature increase or a decrease in the relative humidity (RH). As a result the chloride concentration increases until pitting corrosion is initiated, which is accompanied by a sharp drop in the open circuit potential (OCP) of the metal [5-7]. According to Ibrahim et al [8], the migration of the chloride ions into pits enhances the electrical neutrality and the hydrolysis of the corrosion products inside the pits, which in its turn causes acidification and prevents re-passivation. The mechanism is autocatalytic because the increased acidity accelerates the dissolution rate inside the pits [8]. Several strategies are applicable to increase the corrosion resistance of steel. A first one is to alloy the material. On the one hand, one can opt to add small additions of for example Cu, P, Si or Cr to form a protective rust layer [9]. Weathering steels, i.e. structural high strength alloy steels with improved atmospheric corrosion properties are commonly known as CORTEN steels, which is their original US brand name. On the other hand, a significant addition of chromium to the steel that makes it resist rust, or stain “less” than other types of steel. The chromium in the steel combines with the oxygen in the atmosphere to form a thin, invisible layer of chrome-containing oxide, called the passive film. As an alloying element, molybdenum is more effective in increasing the pitting corrosion resistance as is expressed by the so-called PREN value (pitting resistance equivalent number) [10]: PREN= % Cr + 3.3% Mo + 16% N A second strategy to improve the corrosion resistance is the use of coatings. Metallic coatings are an option as they create a barrier between the base metal and the corrosive environment. Moreover, they might even offer additional resistance by cathodic protection, as is for example realised by galvannealing or galvanizing steel sheet. Non-metallic coatings with the incorporation of corrosion inhibitors are an alternative option. Different inhibitors are possible and chromium (VI) compounds, mainly chromates, are the most common substances used and 2 their efficiency/cost ratio has made them standard corrosion inhibitors. Chromates have been applied in three different ways: incorporated into conversion coatings, as an additive in anodizing baths and as pigments in anti-corrosive primer paints [1, 11-20]. However, the hexavalent chromium species are responsible for several diseases including DNA damage and cancer, which is the main reason for banning chromium (VI) containing anticorrosion coatings in Europe since 2007 [13-14, 16, 21]. For this reason, different initiatives have been launched to develop “green” alternative coatings. The development of hybrid organic–inorganic sol–gel coatings sintered at low temperatures represents a good alternative for coatings containing chromates [22]. They offer increased toughness and are thicker than conventional chemical conversion coatings [1, 16, 21, 23]. Moreover, hybrid organic–inorganic systems are environmentally compliant and compatible with the organic paints and organic top coats used in most of industrial applications [1, 16, 21, 23]. However, the sol–gel coatings may contain pores, cracks and areas of low cross-link densities, through which the corrosion initiators can be diffused to the coating/metal interface. The corrosion process usually begins in such areas. Normally, the sol–gel coatings cannot stop the development of the corrosion process when the defects appear, due to the lack of selfhealing properties [16, 23-25]. Lately attempts have been made to enhance the protective effect of silanes by combining them with other corrosion protection systems such as silica nano-particles or corrosion inhibitors [25-32]. Lanthanides ions, such as Ce3+, Y3+, La3+, Pr3+, Nd3+, seem to fulfil the basic requirements for alternative corrosion inhibitors [33-34]. These elements form insoluble hydroxides and have a low toxicity since their ingestion or inhalation is not considered harmful to health. In addition, lanthanides are economically competitive products, as some of them in particular cerium are relatively abundant in nature [16, 23, 35]. Cerium doped sol–gel coatings have been actively investigated during the last decades [1, 1213, 15, 25, 35]. The first work on cerium conversion layers for aluminium alloys was carried out in the mid 1980’s by Hinton and co-workers [36-37]. Cerium was later tested for the corrosion protection of aluminium and steel based alloys [38-39] as well as galvanized steel in aqueous sodium chloride solution [40-41]. It is claimed that, similarly to the pure sol–gel coatings, the cerium doped ones also can form a dense barrier, which hinders the penetration of electrolyte towards the metallic substrates. The important difference between them is that 3 cerium doped sol–gel coatings can supply self-healing abilities, which can automatically repair the corroded areas, thereby providing a long-term corrosion protection [1, 12-13]. The modification of hybrid sol–gel coatings by adding cerium was successfully tested for improvement of corrosion protection of aluminium [42-43], zinc [44-45] and mild steel substrates [46]. The present work focuses on a detailed study of organic–inorganic hybrid coatings obtained by dip coating 304L stainless steel specimens in an organically modified silica sol made from the hydrolysis and condensation of 3-glycidoxypropyl-trimethoxysilane (GPTMS) as the precursor in acid catalysed condition. Coatings prepared from epoxy silanes and propyltrimethoxy silane using HCl as a catalyst, show a good corrosion protection but are at the same time very brittle and form cracks by bending. For this reason, Kasemann et al. [47] introduced an organic chain to connect the epoxy group. Diols were chosen for this purpose. The epoxy functions can react to polyethylenoxide chains or can be connected by polyaddition reaction with the aromatic diol (bisphenol A) in the presence of a basic catalyst (1-methylimidazole) (see Figure 1). {INSERT Figure 1 HERE} The influence of the addition of cerium on the microscopic features, morphology as well as the corrosion resistance of the coatings is examined using scanning electron microscopy (SEM), optical microscopy, Fourier transform infrared spectroscopy (FTIR), neutral salt spray tests, electrochemical impedance spectroscopy and potentiodynamic polarization. Further, the effect of bisphenol A (BPA) as a cross-linking agent on the microscopic features, morphology and the corrosion performance of the cerium doped silica hybrid coating is performed using the above-mentioned techniques. 2. Experimental Procedure 2.1. Materials Cerium nitrate hexahydrate was purchased from Fluka. All other chemicals and reagents, including 3-glicidoxypropyl-trimethoxysilane (GPTMS), bisphenol A (BPA), hydrochloric acid, sodium chloride and 1-methylimidazol (MI), used were purchased from Merck. 4 2.2. Sol preparation The reference sol was prepared by adding stoichiometric amounts of the silane precursor, GPTMS, and the organic cross-linking agent, BPA, into HCl-acidified water (pH = 2). The H2O/Si mole ratio was 0.5 and was chosen according to a previous study [48]. The solution was stirred at room temperature for 80 minutes at a rate of 240 rpm. Then 1-methylimidazol, MI, was added to accelerate the condensation reaction after hydrolysis. The MI/Si mole ratio was 0.01. The solution was again stirred for 10 minutes at room temperature. The result was a clear and colourless homogenous solution. In what follows, this sol will be called “silica hybrid coating” (A). In the case of cerium doped coating, cerium nitrate hexahydrate was added to the mixture of GPTMS and acid prior to the addition of BPA and 1-methylimidazol. The Ce/Si mole ratio was 0.2. The solution was stirred at room temperature for 10 minutes at a rate of 240 rpm. Then BPA was added to the mixture and dissolved by mixing the solution for 80 minutes. 1methylimidazol was added to the solution and the sol was again stirred for 10 minutes. The result was a brown homogenous solution. This type of coating is referred to as “cerium doped silica hybrid coating” (B). In order to study the effect of BPA on the microscopic features, morphology and corrosion performance of the cerium doped silica hybrid coating, another (B) coating was prepared with the identical chemical composition and procedure, excluding BPA (BPA= 0). The result was a colourless homogenous solution. This coating is referred to as “non BPA cerium doped silica hybrid coating” (C). 2.3. Surface cleaning and silane treatment The substrate material used for the present investigation is a 304L stainless steel alloy. Its chemical composition is given in Table 1. Sheets (7 cm × 15 cm × 0.1 cm) were used for the salt spray tests, while coupons (4.98 cm2 area and 0.1 cm thickness) were used for all other experiments. The substrates were successively abraded with 600 and 1200 grits emery paper, degreased with acetone in an ultrasonic bath for 10 min. Chemical etching was performed by dipping the samples into the 1 molar NaOH solution for 5 min at 50 °C. After rinsing in tap water and then in deionised water, the 304L samples were air-dried. The cleaned substrates were dipped for 1 min in different silane solutions. The coated samples were dried in room temperature for 24 hours, which was followed by a 25-130 °C curing process for 90 minutes to initiate extensive cross-linking in the hybrid films [22]. The coating thickness was measured using profilometry (Talystep, UK). 5 {INSERT Table 1 HERE} 2.4. FTIR measurements FTIR measurements were carried out on the coated samples using a Bio-Rad 575C spectrophotometer in the mid-IR range from 4000 to 400 cm−1. All spectra were taken at an incident angle of 45° normal to the surface of the specimen, with a spectral resolution of 4 cm−1. For each measurement, 64 scans were collected. 2.5. Scanning electron microscopy (SEM) Scanning electron microscopy (SEM) measurements were performed with a XL30 SEM microscope (FEI). The aim was to characterize the surface morphology of the bare and coated substrates. Secondary and backscattered electron images were collected at 15 and 20 kV. 2.6. Anodic and cathodic polarization tests Anodic and cathodic polarization tests were carried out on the 304L stainless steel substrates with and without the silane treatments at 25 °C in a neutral 3.5 % NaCl solution. Prior to the measurements, in order to reach steady state potential, the silane-treated samples were kept in the working solutions for at least 3 hours. An Ag/AgCl/KCl sat electrode and a platinum mesh were used as the reference and counter electrodes, respectively. The potentiodynamic measurements were taken within the range of -1500 to 1000 mV versus Ag/AgCl/KCl sat at a rate of 1 mV/s. The Tafel extrapolation method (conducted according to the ASTM Standard G3-89, 2004) was utilized to determine Icorr and Ecorr. 2.7. Electrochemical impedance spectroscopy measurements (EIS) Electrochemical impedance spectroscopy measurements (EIS) were employed to monitor the corrosion performance of the silane-treated 304L stainless steel substrates in a 3.5% NaCl solution. EIS measurements were carried out at the open circuit potential, using an Autolab PG-STAT 20 potentiostat equipped with a frequency response analyser module. Impedance fitting was performed using appropriate equivalent circuits by means of Zview software (Scribner Associates Inc.). The data were obtained as a function of frequency, using a sine wave of 10 mV amplitude peak to peak. A frequency range of 105 Hz to 10-2 Hz was selected. A three-electrode electrochemical cell arrangement was used, consisting of the working electrode (4.98 cm2 of exposed area), Ag/AgCl/KCl sat electrode as reference and Pt as counter electrode. 6 2.8. Salt spray testing The corrosion performance of the bare and coated substrates was evaluated in a neutral salt spray test, following the ASTM B117 procedure and employing a 5% sodium chloride solution. Prior to exposure, the back and edges of the plates were covered with adhesive tape. An artificial scratch was made in the coating that penetrated to the substrate to examine possible delamination. Visual assessments of the macroscopic surfaces were carried out at various time intervals during the exposure time. 3. Results and discussion 3.1. Characterization Figures 2a and b show FT-IR spectra of the silica hybrid coating (sample A) and the cerium doped silica hybrid coating (sample B) in magnified regions from 500 to 1300 cm-1 and 2800 to 4000 cm-1, respectively. The main features of the spectra include bands corresponding to Si-O-Si sequence, OH group and –CH2 group: The strong peaks between 1000 and 1200 cm-1 are associated with the Si–O–Si/C–O–C asymmetric bond stretching vibration [49], which is the structural backbone of the hybrid material [50]. The high frequency band near 1020 cm-1 can be assigned to the antisymmetric stretching of oxygen atoms. The bands at intermediate frequencies around 750-800 cm-1 can be attributed to symmetric stretching motions of the oxygen atoms. The –CH2 group incorporated into the silica group is indicated by the presence of the two weak absorption peaks at about 2930 cm-1 and 2950 cm-1, which can be assigned to symmetric and asymmetric –CH2 bond stretching modes present on the silicon moiety [51]. The presence of this hydrocarbon unit reveals a certain degree of planarization of the coating macromolecular chains. It can be observed that intensities of the 1087 cm-1 band along with 780 and 2950 cm-1 shoulders increase in the cerium doped sample (sample B), which demonstrates that the incorporation of cerium nitrate leads to the advance of the condensation process. The broad band centered near 3400 cm-1can be assigned to residual Si–OH stretching vibrations and hydrogen-bonded water [52]. This is possibly formed in the hydrolysis reaction of the alkoxy groups of GPTMS. {INSERT Figure 2 HERE} 7 The effect of bisphenol A, as an aromatic diol on cross-linking of organo-silica networks via chemical coupling of aromatic diol and epoxy-functionalities was also studied by FT-IR spectroscopy. Figures 3a and b show the FT-IR spectra of the cerium doped silica hybrid coating with and without BPA (samples B and C) in magnified regions from 500 to 1300 cm-1 and 2800 to 4000 cm-1, respectively. As can be seen in these figures, the bands of the antisymmetric epoxide ring deformation at 758 and 902 cm-1 in sample C appears to be almost as intense and allows monitoring the chemical reaction of the coupling of epoxy functionalities of organo silica networks with diol cross-linkers. These bands almost disappear in the FTIR spectra of sample B, which indicates the chemical bonding of organo-silica networks into a cross-linked network [53]. Also the bands centred at 1100 cm-1 due to Si-O-Si in sample C become more intense and broader as compared to that sample B which can be attributed to direct connection of epoxy silanes and propyl trimethoxy silane in sample C that its resultant is the formation of a very brittle film. {INSERT Figure 3 HERE} SEM was used to investigate the effect of cerium and BPA on the morphology of the sol-gel hybrid coatings. Figure 4a shows typical SEM micrographs of the non-doped silica hybrid coating (sample A) and demonstrates that such coating is defects and cracks free. One, however, sees agglomerations, which may be ascribed to the relatively high viscosity of the non-doped sol. The places around the agglomerations may be prone to corrosion because of the discontinuous coating structure. Results also show that the morphology of the coating doped with cerium nitrate is free of cracks and highly homogenous (Figure 4b). This demonstrates that the cerium nitrate can cause a depression of agglomerations and the appearance of many particles entrapped in the coatings without any micro-scale pores and cracks, suggesting good corrosion resistance. In the case of cerium doped coating, without BPA (sample C), several micro-scale pores and cracks appear in the surface of the coating (Figure 4c) which indicate that omitting the BPA as a cross-linking agent will lead to the appearance of defects. These results suggest a poor corrosion resistance. {INSERT Figure 4 HERE} 8 3.2. Performance in the salt spray test The traditional salt spray coating corrosion resistance evaluation technique was used to evaluate the corrosion performance of the sol–gel films. Photographs of samples A, B and C after 2000 hours of salt spray exposure are shown in Figure 5, along with a photograph of a bare plate as a reference (taped areas were cropped from the picture). The photographs show that the silica hybrid coated (sample A) (average thickness of the coating = 19.2 µm) and the cerium doped silica hybrid coated surfaces (sample B) (average thickness of the coating = 57.71 µm) exhibit not any signs of blistering, delaminating and corrosion and still retains its originally shiny surfaces after exposure for 2000 hours. In contrast, the bare plate and the plate coated with cerium doped silica hybrid coating but without BPA (sample C) (average thickness of the coating= 57.12 µm), show a limited resistance in the neutral salt spray test. As was also demonstrated by the electrochemical tests. Sample C exhibited blistering and delaminating after an exposure for 148 hours and first signs of corrosion in both the bare sample and sample C appeared after an exposure for 1600 hours. {INSERT Figure 5 HERE} 3.3. Electrochemical behaviour Potentiodynamic polarization measurements were carried out to estimate the effect of cerium and BPA on the corrosion resistance of the silica hybrid coatings. Figure 6 presents the potentiodynamic polarization curves of the bare and the A, B and C coated 304L stainless steel substrates recorded after 3 hours of immersion in a neutral 3.5 % NaCl solution. Table 2 shows the summary of the electrochemical parameters obtained from these results. The limiting current density for the oxygen reduction and the anodic current density for all the coated specimens decrease in comparison to those for the uncoated 304L stainless steel substrate. These observations suggest that the corrosion protection offered by these coatings is due to both blocking of parts of 304L stainless steel surface with reduction of oxygen and metal dissolution occurring in the pores of the coating layer. The sample coated with the silica hybrid coating (sample A), shows a decrease in corrosion current density, icorr, around 10-8 A.cm-2. As well, its potential reaches a higher value of about −0.414 V from the corrosion potential (Ecorr) with respect to the bare sample. However, a very limited passive region was present at -0.106 V over the Ecorr, prior to the onset of pitting. With addition of cerium ions (sample B), further reduction in the corrosion current density and its potential is again revealed and a distinct passive region, extending over several 9 hundred mV, is visible (+0.284 V over the Ecorr, before the point of the breakdown potential was reached). The poor performance displayed by the specimens coated with the cerium doped silica hybrid coating but without BPA (sample C), may be due to the permeability of the coating associated with the cracks or porosity which allowed the aggressive electrolyte to reach the metal surface and initiating corrosion. This indicates that omitting BPA as a cross-linking agent, can have a significant effect on the corrosion resistance of the cerium doped silica hybrid coating. {INSERT Figure 6 HERE} {INSERT Table 2 HERE} The SEM images for the bare and coated 304L stainless steel samples obtained after potentiodynamic polarization test in a neutral 3.5% NaCl solution are depicted in Figure 7. After the anodic polarization, a localized corrosion is observed on all samples, consisting of an exfoliation of the corrosion products, cracks (Figures 7a, c and e) and pits (Figure 7g) of different sizes. A closer inspection shows that the region surrounding some of the cracks and pits is damaged (see Figures 7b, d, f and h). This may indicate a preferential localized attack, occurring after the electric potential exceeded the breakdown potential. This localized attack promotes the hybrid film deterioration and delamination, possibly due to the hydrolysis reactions at the interface. Additionally, the diffusion of the oxidant ions speeds up and the corrosion rate increases causing an accumulation of corrosion products at the interface, promoting the formation of defects and micro-cracks as can be seen in Figure 7 (a-f). However, samples A and B show a higher corrosion resistance described by the Ecorr and icorr values (Table 2), as reported before. {INSERT Figure 7 HERE} 3.4. Electrochemical impedance spectroscopy Figure 8 presents the impedance spectra obtained for the bare and coated 304L substrates after 3 hours of immersion in a 3.5 % NaCl solution. For the bare sample, the EIS spectrum is characterized by two time constants. The time constant at higher frequencies can be assigned to the presence of an oxide/hydroxide film existing on the bare 304L substrate, whereas the time constant at lower frequencies can be assigned to corrosion activity. The total impedance 10 of the bare substrate is lower than 106 Ω cm2. At this stage, the bare surface shows a significant amount of corrosion products. Samples A and B display two time constants: one in the high frequency range, attributed to the presence of silane layer and another one in the lower frequencies accounting for the charge transfer reaction associated with oxidation of the substrate. The Bode phase angle plot for the sample C demonstrates the presence of a new time constant in mid range of frequencies that can be attributed to the cracking and corrosion activity at the substrate surface. This is probably due to the brittle structure of the film in absence of cross-linking agent (BPA), which can cause the formation of micro cracks and delamination of the film. In addition, the results show that both the silica hybrid coating (sample A) as well as the cerium doped silica hybrid coating (sample B) increase the low frequency impedance of the system by about one order of magnitude as compared to the bare sample and sample C (Figure 8b). The highest impedance in samples A and B is probably due to an area effect where the coating is blocking the aggressive electrolyte from reaching the reactive metal surface. The lower impedance observed in the case of sample C is due to creation of defects/cracks in the coating promoting diffusion of the aggressive electrolyte to the substrate. {INSERT Figure 8 HERE} Furthermore, electrochemical impedance spectroscopy measurements were performed on both the silica hybrid coating (sample A) and the cerium doped silica hybrid coating (sample B) in order to evaluate the barrier properties as a function of immersion time and the inhibition mechanisms. Figures 9 and 10 show the impedance diagrams (Bode modulus and phase angle plots) for samples A and B, respectively, as a function of immersion time in a 3.5% NaCl solution for a period of 22 days. Impedance measurements carried out on the non-inhibited system (sample A), display a slight evolution of the barrier properties associated to the hybrid coating. This behaviour can be related to the water uptake occurring in the first days of immersion in the electrolyte. After 22 days, the non-inhibited system displays less barrier effects. The system containing cerium species (sample B) is more protective than the non-inhibited (sample A) one and its barrier properties remain constant for longer immersion time (22 days). The inhibited system shows a slower electrochemical evolution than the non-inhibited one, because the inhibition effect of cerium ions slows down the corrosion kinetics. 11 {INSERT Figures 9 and 10 HERE} In order to obtain better insight in the corrosion features of the samples, the EIS spectra were analyzed using equivalent electric circuits. In the simulation of impedance plots, the constant phase element (CPE) was used instead of an “ideal” capacitor to explain the deviations from slope −1 in the modulus Bode plot. CPE can be described by the expression: ZCPE = 1 n Y0 ( jw ) (1) with −1 < n < 1 [1]. In this equation, n is a coefficient associated to system homogeneity (being 1 for an ideal capacitor), ω is the frequency, and Y0 is the pseudo capacitance of the system (expressed in Ωcm-2sn) that can be represented by: Y0 r 0 A d (2) In the Eq. (2), ε0 is the permittivity of free space (expressed in F/m), ε the dielectric constant of the surface film, d the thickness of the coating (expressed in µm), A the exposed area (expressed in cm2) and r the roughness factor [54]. The impedance response for the silica hybrid coating (sample A) after 22 days of immersion suggests a non-ideal dielectric capacitive behaviour, which includes the electrolyte resistance (Rs), polarization resistance (Rpo) of the corroded areas and a non-ideal capacitance of the coating (CPEc) (Figure 11a). An example of the fitted model adopted for sample (A) is included in Figure 11b. {INSERT Figure 11 HERE} The equivalent circuit shown in Figure 12a was used to model the cerium doped silica hybrid coating (sample B) after 22 days of immersion and it consists of the following elements: Rs represents the electrolyte resistance, CPEc is related with the non ideal capacitance of the coating, Rpo is the resistance presented by the porous or defects to the passage of the electrolyte. In this figure, the new parameters included in the circuits (CPEd and Rt) consider the presence of a double layer between the metal surface and the electrolyte which is leading the corrosion products to the pore e.g.: CPEd is related with the non ideal capacitance of the double layer and Rt is related with the charge transfer resistance of the metal. Figure 12b shows the impedance spectrum and the fitting curves obtained using the equivalent circuit for 12 sample B. Table 3 shows the fitting parameters for the models presented in Figures 11a and 12a. The decrease observed in the polarization resistance of the system in sample A can be associated with the water uptake through the pores and/or defects in the coating and is probably connected with an increase of exposed area. {INSERT Figure 12 HERE} {INSERT Table 3 HERE} The electrochemical parameters derived from the polarization curves together with the EIS results, reveal the protective role of cerium doped silica hybrid coatings, which persists with increasing immersion time. Sol–gel coatings used as protection against metal corrosion, do not usually improve with the immersion time. In fact, depending on the composition and structure of the coating (totally inorganic or hybrid), the electrolytic medium and the soaking time, deterioration of the coatings occurs, permitting the contact of the electrolyte with the substrate and thus decreasing the protector effect [55]. However, the application of sol–gel coatings in this work leads to an improvement in the corrosion protection. The increase of immersion time probably leads to the release of cerium in the defects of the coatings. Then, cerium produces insoluble hydroxides when it reacts with hydroxyl groups from cathodic reactions [56-57]. These hydroxides together with corrosion products decrease the cathodic current and, therefore, reduce the overall corrosion rate [12]. 4. Conclusions Silica based hybrid organic–inorganic coatings with the incorporation of cerium ions were synthesized via a sol–gel method to protect 304L stainless steel substrate against corrosion. Three types of samples were produced by means of dip coating technique: a non-inhibited silica hybrid coating, an inhibited silica coating containing cerium and an inhibited silica coating containing cerium but without cross-linking agent (BPA). The latter one was used to study the effect of the cross-linking on the morphology and the corrosion protection properties of prepared samples. Results of FTIR and SEM analyses confirm the formation of crack-free silica hybrid coatings with Si-O-Si structural backbone and −CH2 group incorporated into silica network which results in the formation of a transparent and homogenous film. Cracks are observed in the linear cross-linked cerium doped silica hybrid 13 coating due to lack of cross-linking agent, which causes the formation of very brittle and crack full coating with poor corrosion resistance. The corrosion protection performance of the films was evaluated by natural salt spray, potentiodynamic polarization and EIS techniques. The results show that the presence of the cerium ion has significantly improved the corrosion protection properties of the silica hybrid coatings. This effect is due to both increases in the barrier properties as well as the corrosion inhibition ability. Thus cerium doped silica hybrid coatings protect the 304L stainless steel surface effectively and can be used as a pre-treatment for corrosion protection. Acknowledgements The authors would like to acknowledge Ghent University for financially support of this study. The authors would also like to thank Peter Mast, Michel Moors and Veerle Boterberg for their technical assistance. References [1] A. Pepe, M. Aparicio, A. Durán, S. Ceré, J. Sol-Gel Sci. Techn. 39 (2006) 131–138. [2] H.P. Leckie, H.H. Uhlig, J. Electrochem Soc. 113 (1966) 1262-1267. [3] N.J. Laycock, R.C. Newman, Corros. Sci. 39 (1997) 1771-1790. [4] J.R. Galvele, Corros. Sci. 21 (1981) 551-579. [5] R.P. Vera Cruz, A. Nishikata, T. Tsuru, Corros. Sci. 40 (1998) 125-139. [6] S. Ito, M. Yabumoto, H. Omata, T. Murata, 5th International Symposium of Passivity of Metal and Semiconductors, Bombannes, France, 30 May - 3 June (1983) 637-642. [7] B. Maier, G. S. Frankel, J. Electrochem. Soc. 157 (2010) C302-C312. [8] A.M. Ibrahim. Magdy, S.S. Abd El Rehim, M.M. Hamza, Mater. Chem. Phys. 115 (2009) 80–85. [9] S. Vervynckt, K. Verbeken, B. Lopez, J.J. Jonas, Inter. Mater. Rev. (2011), doi: 10.1179/1743280411Y.0000000013. [10] The salt spray test and its use in ranking stainless steels, International Stainless Steel Forum, May (2008). [11] W.J. Wittke, Met. Finish 87 (1989) 24-26. [12] A. Pepe, M. Aparicio, S. Ceré, A. Durán, J. Non-Cryst. Solids 348 (2004) 162–171. [13] X. Zhong, Q. Li, J. Hu, X. Yang, F. Luo, Y. Dai, Prog. Org. Coat. 69 (2010) 52–56. 14 [14] D. Shchukin, H. Möhwald, M. Ferreira, S. Guerreiro, M. Zheludkevich, Patent WO2007/104457 A1 (2007). [15] M. Garcia-Heras, A. Jimenez-Morales, B. Casal, C.J. Galvan, J. Alloy Compd. 380 (2004) 219–224. [16] C.N. Rosero-Navarroa, L. Paussa, F. Andreatta, Y. Castro, A. Durán, M. Aparicio, L. Fedrizzi, Prog. Org. Coat. 69 (2010) 167–174. [17] A. Pardo, M.C. Merino, A.E. Coy, F. Viejo, R. Arrabal, E. Matykina, Corros. Sci. 50 (2008) 780–794. [18] T. Yamamoto, K. Fushimi, M. Seo, S. Tsuri, T. Adachi, H. Habazaki, Corros. Sci. 51 (2009) 1545–1553. [19] A. Pardo, M.C. Merino, A.E. Coy, F. Viejo, R. Arrabal, E. Matykina, Corros. Sci. 50 (2008) 1796–1806. [20] T. Bore Jegdi, M. Dragutin. T. Draqi, P. Jovan, T. Popi, Corros. Sci. 50 (2008) 1235– 1244. [21] B. P. Mosher, C. Wu, T. Sun, T. Zeng, J. Non-Cryst. Solids 352 (2006) 3295–3301. [22] R. Zandi Zand, K. Verbeken, A. Adriaens, Prog. Org. Coat. 72 (2011) 709-715. [23] N. Pirhady Tavandashti, S. Sanjabi, Prog. Org. Coat. 69 (2010) 384-391. [24] T.L. Metroke, R.L. Parkhill, E.T. Knobbe, Prog. Org. Coat. 41 (2001) 233–238. [25] A. Phanasgaonkar, V.S. Raja, Surf. Coat. Tech. 203 (2009) 2260–2271. [26] W. J. Van Ooij, D. Zhu, M. Stancy, A. Seth, T. Mugada, J. Gandhi, P. Paomi, Tsinghuax, Sci. Tech. 10 (2005) 639-664. [27] M. Quinet, B. Neveu, V. Moutarlier, P. Audebert, L. Ricq, Prog. Org. Coat. 58 (2007) 46-53. [28] V. Palanivel, D. Zhu, W.J. Van Ooij, Prog. Org. Coat. 47 (2003) 384-392. [29] L. S. Kasten, J. T. Grant, N. Grebasch, N. Voevodin, F. E. Arnold, M. S. Donley, Surf. Coat. Tech. 140 (2001) 11-15. [30] M.F. Montemor, A.M. Cabral, M.L. Zheludkevich, M.G. Ferreira, Surf. Coat. Tech. 200 (2006) 2875-2885. [31] M. Menning, G. Jonschker, H. Schmidt, SPIE Sol–Gel Opt .1758 (II) (1992) 125-134. [32] H. Schmidt, G. Jonschker, S. Goedike, M. Menning, J. Sol– Gel Sci. Techn. 19 (2000) 39-51. [33] A. Caglieri, M. Goldoni, O. Acampa, R. Antreoli, M.V. Vettori, M. Corradi, P. Apostoli, A. Mutti, Health Perspect 114 (2006) 542-546. 15 [34] T.P. Anatas, C.T. Williamson, American Chemical Society Symposium Series, Washington DC 626 (1996) 1-17. [35] A. Nazeri, P.P. Trzaskoma-Paulette, D. Bauer, J. Sol-Gel Sci. Techn. 10 (1997) 317– 331. [36] B.R.W. Hinton, D.R. Arnot, N.E. Ryan, Met. Forum 7 (1984) 211-217. [37] L. Wilson, B. R. W. Hinton, Patent WO 88/ 06639 (1998). [38] B. Hinton, Corrosion 89, Apr. 17-21, paper 170, NACE, New Orleans (1989). [39] B. Hinton, K. Shankar, P. Trathen, M. Salagaras, L. Wilson, G. Devereaux, Paper 135, Vol 3, Proceedings of the 13th International Corrosion Congress, Melbourne, Australia, 25-29 November (1996). [40] W. Trabelsi, P. Cecilio, M.G.S. Ferreira, M.F. Montemor, Prog. Org. Coat. 54 (2005) 276-284. [41] M.F. Montemor, A.M.P. Simoes, M.G.S. Ferreira, C. B. Breslin, Prog. Org. Coat. 44 (2002) 111-120. [42] H. Wang, R. Akid, Corros. Sci. 49 (2007) 4491-4503. [43] L.E.M. Palomino, P.H. Suegama, I.V. Aoki, Z. Paszti, H.G. De Melo, Electrochim. Acta 52 (2007) 7496-7505. [44] W. Trabelsi, P. Cecilio, M.G.S. Ferreira, K. Yasakau, M.L. Zheludkevich, M.F. Montemor, Prog. Org. Coat. 59(2007) 214-223. [45] M.F. Montemor, W. Trabelsi, M. Zheludkevich, M.G.S. Ferreira, Prog. Org. Coat. 57 (2006) 67-77. [46] H. Wang, R. Akid, Corros. Sci. 50 (2008) 1142-1148. [47] R. Kasemann, H. Schmidt, New J. Chem. 18 (1994) 1117-1123. [48] R. Zandi-Zand, Investigation of corrosion, abrasion and weathering resistance in hybrid nanocomposite coatings based on epoxy-silica, Thesis, Azad University-Tehran North Branch, (2005). [49] R.J. Bell, P. Dean, Nature 212 (1966) 1354-1355. [50] Y.H. Han, A. Taylor, M.D. Mantle, K.M. Knowles, J. Non-Cryst. Solids 353(2007) 313-320. [51] A. Tiwari, L. H. Hihara, TRI-service corrosion conference 3-7 December (2007), Denver. [52] A. Bertoluzza, C. Fagnano, M.A. Morelli, V. Gotardi, M. Guglielmi, J. Non-Cryst. Solids 48 (1982) 117-128. 16 [53] R. Zandi-Zand, A. Ershad-langroudi, A. Rahimi, J. Non-Cryst. Solids 351 (2005) 13071311. [54] C. Liu, Q. Bi, A. Mattews, Corros. Sci. 45 (2003) 1257-1273. [55] O. de Sanctis, N. Pellegri, A. Durán, J. Non-Cryst. Solids 121 (1990) 338-343. [56] W.G. Fahrenholtz, M.J. O’Keefe, H. Zhou, J.T. Grant, Surf. Coat. Tech. 155 (2002) 208-213. [57] J. Olefjord, C. R.Clayton, ISIJ Int. 31 (1991) 134-141. 17 Table 1. Chemical composition of 304L stainless steel (in mass %). C Si Mn P S Al Cr Mo Ni Cu Fe 0.017 0.41 1.75 0.032 0.0050 0.0049 17.65 0.25 8.59 0.45 Balance Table 2. Summary of the electrochemical parameters obtained from the polarization results in the 3.5% NaCl solution. Sample Ecorr (V vs Ag/AgCl/KCl sat) Icorr (A.cm-2) Bc (V/dec) Ba (V/dec) Bare SS 304 -0.690 3.97×10-6 0.331 0.243 A -0.414 2.304×10-8 0.264 0.103 B -0.389 4.481×10-8 0.297 0.1 C -0.427 5.101×10-7 0.315 0.101 Table 3. Parameters obtained from data fitting of 304L coated with silica hybrid coating (sample A) and Ce-doped silica hybrid coating (sample B) after 22 days of immersion in a 3.5% NaCl solution. Sample Rs (Ω cm2) CPEC (Ω-1cm-2sn) nc Rpo (Ω cm2) CPEdo (Ω-1cm-2sn) nd Rt (Ω cm2) A 22.59 1.4245×10-5 0.83983 3.6751×105 - - - B 22.59 2.6192×10-6 0.82889 7.4422×105 1.238×10-6 0.50668 6.4913×106 18 Figure Captions Figure 1. Simplified schematic of bonding mechanism between bisphenol A and the epoxy function of GPTMS [47]. Figure 2. FTIR spectra of the 304L stainless steel substrates coated with silica hybrid coating (sample A) and Ce-doped silica hybrid coating (sample B):(a) magnified region from 500 to 1300 cm-1 and (b) magnified region from 2800 to 4000 cm-1. Figure 3. FTIR spectra of the 304L stainless steel substrate coated with Ce-doped silica hybrid coating (sample B) and non BPA Ce-doped silica hybrid coating (sample C):(a) magnified region from 500 to 1300 cm-1 and (b) magnified region from 2800 to 4000 cm-1. Figure 4. Scanning electron micrographs of the coated 304L stainless steel samples prior to the polarization test: (a) sample A “silica hybrid coating”, (b) sample B “Cedoped silica hybrid coating” and (c) sample C “non BPA Ce-doped silica hybrid coating”. Figure 5. Photographs of (a) silica hybrid coated, (b) Ce-doped silica hybrid coated and (c) non BPA Ce-doped silica hybrid coated and (d) bare 304L stainless steel samples after 2000 hours of exposure in the salt spray chamber. Figure 6. Potentiodynamic polarization curves of the (A), (B) and (C) coated and bare 304L stainless steel in a 3.5 % NaCl solution. Figure 7. Scanning electron micrographs of the bare and coated 304L stainless steel samples after potentiodynamic polarization in a neutral 3.5% NaCl solution with different magnifications: (a and b) sample A, (c and d) sample B, (e and f) sample C and (g and h) bare 304L stainless steel. Figure 8. EIS Bode plots (a) θ vs log freq and (b) log |Z| vs log freq for bare and (A), (B) and (C) coated 304L stainless steel substrates after 3hours of immersion in a 3.5 % NaCl solution. Figure 9. EIS Bode plots (a) θ vs log freq and (b) log |Z| vs log freq for silica hybrid coated (sample A) 304L stainless steel substrates at different immersion times in a 3.5 % NaCl solution. Figure 10. EIS Bode plots (a) θ vs log freq and (b) log |Z| vs log freq for Ce doped silica hybrid coated (sample B) 304L stainless steel substrates at different immersion times in a 3.5 % NaCl solution. 19 Figure 11. (a): Equivalent circuit used for modelling and (b): bode plots with respective fittings for 304L stainless steel substrate protected with the silica hybrid coating (sample A) after 22 days immersion in 3.5 % NaCl solution. Figure 12. (a): Equivalent circuit used for modelling and (b): bode plots with respective fittings for 304L stainless steel substrate protected with the Ce doped silica hybrid coating (sample B) after 22 days immersion in 3.5 % NaCl solution. 20 O O Si(OCH3)3 O O Si O O O O Si O O O O Si O O O O O O O HCl/H2O 1-Methylimidazol O O O OH O O O OH O O Si O Si + O O Si O Si O O O O O Figure 1 21 90 (a) 80 60 50 1280 40 780 % Transmittance 70 30 10 A 0 1300 1200 1100 1020 B 1087 20 1000 900 800 700 600 500 Wavenumber cm -1 120 (b) 80 60 2950 3350 2930 % Transmittance 100 40 B 20 A 0 4000 3800 3600 3400 Wavenumber cm -1 Figure 2 22 3200 3000 2800 100 (a) 90 80 60 50 40 30 758 902 % Transmittance 70 B 10 C 0 1300 1100 20 1200 1100 1000 900 800 700 600 500 Wavenumber cm -1 120 (b) 80 3350 % Transmittance 100 60 40 B 20 C 0 4000 3800 3600 3400 Wavenumber cm -1 Figure 3 23 3200 3000 2800 (a) (b) (c) Figure 4 24 (a) (b) (c) (d) Corrosion spot Figure 5 25 1 E /V vs Ag/AgCl/KCl sat 0.5 0 A B -0.5 C bare ss 304 -1 -1.5 1.E-09 1.E-08 1.E-07 1.E-06 1.E-05 I/ A cm-2 Figure 6 26 1.E-04 1.E-03 1.E-02 (a) (b) (c) (d) (e) (f) (g) (h) Figure 7 27 90 (a) 80 -phase / deg(+) 70 RC 1 60 RC 2 50 RC 3 40 30 A 20 B C 10 Bare SS304 0 -2 -1 0 1 2 3 4 log(f)(Hz) 8 (b) 7 log(Z)(Ω.cm2) 6 5 4 3 A 2 B C 1 Bare SS304 0 -2 -1 0 1 log(f)(Hz) Figure 8 28 2 3 4 90 (a) 80 70 -phase / deg(+) 60 1 Day 50 2 Days 40 3 Days 4 Days 30 8 Days 20 14 Days 22 Days 10 0 -2 -1 0 1 2 3 4 log(f) (Hz) 7 (b) 6 log(Z) (Ω.cm2) 5 1 Day 4 2 Days 3 days 3 4 Days 2 8 Days 14 Days 1 22 Days 0 -2 -1 0 1 log(f) (Hz) Figure 9 29 2 3 4 90 (a) 80 70 -phase / deg(+) 60 1 Day 50 2 Days 40 3 Days 4 Days 30 8 Days 20 14 Days 22 Days 10 0 -2 -1 0 1 2 3 4 log(f) (Hz) 7 (b) 6 log(Z) (Ω.cm2) 5 1day 4 2 Days 3 Days 3 4 Days 8 Days 2 14 Days 1 22 Days 0 -2 -1 0 1 log(f) (Hz) Figure 10 30 2 3 4 Rs (a) CPEC Rpo (b) 80 6 70 5 -phase / deg(+) 4 50 40 3 30 2 20 1 Experimental data 10 FitFitted Resultresults 0 0 -2 -1 0 1 log(f) (Hz) Figure 11 31 2 3 4 log(Z) (Ω.cm2) 60 (a) Rs CPEc Rpo CPEd Rt (b) 80 7 70 6 5 50 4 40 3 30 2 20 Experimental data 1 10 Fit Result 0 0 -2 -1 0 1 log(f) (Hz) Figure 12 32 2 3 4 log(Z)(Ω.cm2) -phase / deg(+) 60