Ureasilicate Hybrid Coatings for Corrosion Protection of Galvanized Steel in Chloride-Contaminated Simulated Concrete Pore Solution

Organic-inorganic hybrid (OIH) sol-gel coatings based on ureasilicates (U(X)) have promising properties for use as eco-friendly coatings on hot dip galvanized steel (HDGS) and may be considered potential substitutes for pre-treatment systems containing Cr(VI). These OIH coatings reduce corrosion activity during the initial stages of contact of the HDGS samples with highly alkaline environments (cementitious media) and allow the mitigation of harmful effects of an initial excessive reaction between cement pastes and the zinc layer. However, the behavior of HDGS coated with U(X) in the presence of chloride ions has never been reported. In this paper, the performance of HDGS coated with ﬁve different U(X) coatings was assessed by electrochemical measurements in chloride-contaminated simulated concrete pore solution (SCPS). U(X) sol-gel coatings were produced and deposited on HDGS by a dip coating method. The coatings performance was evaluated by electrochemical impedance spectroscopy, potentiodynamic polarizationcurvesmeasurements,macrocellcurrentdensityandpolarizationresistanceincontactwithchloride-contaminatedSCPS.TheSEM/EDSanalysesofthecoatingsbeforeandafterthetestswerealsoperformed.TheresultsshowedthattheHDGSsamplescoatedwiththeOIHcoatingsexhibitedenhancedcorrosionresistancetochlorideionswhencomparedtouncoatedgalvanizedsteel.©TheAuthor(s)2015.PublishedbyECS.ThisisanopenaccessarticledistributedunderthetermsoftheCreativeCommons

Corrosion of the reinforcing steel is one of the most important causes of degradation in reinforced concrete structures (RCS). 1 The presence of chloride, carbonation of the concrete or the low quality of the concrete cover causes serious damage to RCS. The entrance of chloride ions through the concrete from marine environments or deicing salts can cause rupture of the passive layer, allowing the steel surface to act as a coupled anodic and cathodic reaction cell in which corrosion processes take place. [1][2][3] Therefore, the presence of chlorides in concrete can seriously shorten the service life of RCS. 3,4 The expanded volume of corrosion products deposited in the interface between the concrete and the steel imposes expansive stresses, leading to delamination, cracking or spalling and eventually to the collapse of the RCS. 1 Several methods to improve the corrosion resistance of RCS have been proposed. 1,[5][6][7] In the last few years, the use of hot-dip galvanized steel (HDGS) has been recognized as an effective measure to increase the service life of RCS. [8][9][10][11] The galvanized coating (zinc layer) acts as a physical barrier, hindering the contact of aggressive agents with the steel substrate and the zinc acts as a sacrificial anode protecting the steel against corrosion. In addition, the formed zinc corrosion products have a smaller volume than those produced from iron, thus reducing the corrosion-induced delamination, cracking or spalling of the concrete. 8 Additionally, the galvanized steel reinforcement can withstand exposure to chloride ion concentrations several times higher than the chloride level that causes corrosion in steel reinforcement. 8,12 However, when the HDGS is in contact with concrete pore solution, (which is essentially a Ca(OH) 2 saturated solution containing significant amounts of KOH and NaOH) the zinc corrodes. [8][9][10]13,14 The pH of this medium is typically above 12.5. This initial corrosion process, which may vary from hours to days, may lead to zinc consumption until either the formation of a passivation layer or until all the zinc layer is consumed. 8,[13][14][15][16] To minimize this corrosion process, measures such as increasing the chromate content by adding water-soluble chromates into the concrete mixture or the use of chromate conversion layers have been widely implemented. Nevertheless, due to the adverse health effects of the hexavalent chromium ions (Cr(VI)), hard restrictions on the use of chromates have been imposed. As a result, the growing interest in developing new materials to replace chromate conversion layers has led to the synthesis and testing of several organic-inorganic hybrid (OIH) materials, generally referred as * Electrochemical Society Student Member. z E-mail: rmfigueira@lnec.pt; rita@figueira.pt ORMOCER (Organic Modified Ceramics) or ORMOSILS (Organic Modified Silanes), as corrosion protective coatings on several metallic alloys, such as aluminum, [17][18][19][20] carbon steel, [21][22][23] stainless steel, 21,[24][25][26] galvanized steel [27][28][29] and magnesium. 30,31 These OIH films are reported to be environmentally friendly alternatives to replace chromate conversion layers. 32 Their synthesis is based on the sol-gel process which involves the hydrolysis of metal alkoxides to produce hydroxyl groups in the presence of stoichiometric water (generally catalyzed by an acid or base) followed by polycondensation of the resulting hydroxyl groups and residual alkoxyl groups, forming thin, dense and chemically inert films on the metallic substrates with controlled physical shape and dimensions. Precursors containing non-hydrolysable groups are used to incorporate the organic part in the coating. This organic component provides flexibility and reduces defects while the inorganic constituent provides superior adhesion to the metal surface and improves mechanical resistance. [33][34][35][36] The sol-gel method is a versatile and simple method to produce silicate based gel materials under room temperature. The gel is aged for a period of time to allow the gel network to strengthen and then it is dried under atmospheric conditions. Materials prepared in this way are called xerogels. 37 Small molecules resulting from the condensation process remain trapped within the formed network even after the post-gelation drying/curing process is completed. 38 The OIH gels exhibit a biphasic structure that combines a rigid and hydrophilic silicate backbone linked to amorphous, malleable and hydrophobic organic (polymeric) spacer. The coexistence of the two distinct phases enhances the dispersion of a large variety of hosted species and allows control of the physical properties of the gel (transparency, porosity, wettability, hydrophobicity, etc.). The term ureasilicate (U(X)) refers to OIH sol-gel coatings obtained from a reaction between a functionalized siloxane (3-isocyanate propyltriethoxysilane) and a di-amino functionalized polyether (Jeffamine) with different molecular weights (ranging from 230 to 2000 g mol −1 ). 27 Since urea bonds are formed between the two precursor molecules, the term U(X) has been used to identify this type of OIH network where "U" refers to the type of bond established and "X" to the molecular weight of Jeffamine used. U(X) coatings have been extensively studied in contact with cementitious media 27,39 showing that these coatings provide barrier properties that withstand the high pH of the electrolyte, protecting the HDGS when it first comes into contact with cementitious media. The lowest corrosion rates obtained, after 127 days of contact with a mortar, are given by U(230) and U(400) and are respectively 30 and 31 times lower than the control samples. 39 Later and considering these promising results, 27,39 studies on the influence of the residence time, curing process between each dipping step and thickness of U(230) and U(400) 40 were performed in highly alkaline environments. The glow discharge optical emission spectroscopy results show that the coating thickness of U(400) globally increases both with the number of dipping steps and with residence time. 40 However, the SEM/EDS results point to the conclusion that full coverage is seldom achieved even when three layers are deposited on HDGS with residence time with the curing process between each deposition. 40,41 The present work is focused on study the corrosion performance of five U(X) sol-gel coatings (U(230), U(400), U(600), U(900) and U(2000)) in chloride-contaminated alkaline solutions which simulate the aqueous phase present in concrete pores. This paper describes the modifications that the presence of chloride ions has on the corrosion behavior of coated HDGS samples in highly alkaline solutions. Bare HDGS samples were used for comparison purposes. The OIH coatings were prepared according to the literature. 27,39,40 The coatings were deposited on HDGS surfaces by dip-coating using one or three consecutive dip steps. The novelty of the proposed work relies on the fact that, to the best of the authors' knowledge, no study is available in the literature where U(X) OIH sol-gel coatings deposited on HDGS have been tested in chloride-contaminated SCPS. The morphology of the OIH films was assessed by scanning electron microscopy (SEM). The corrosion properties of the OIH coatings were evaluated by macrocell current density (i gal ), polarization resistance (R p ), electrochemical impedance spectroscopy (EIS) and potentiodynamic polarization methods in chloride-contaminated simulated concrete pore solutions (SCPS). i gal and R p measurements were carried out since they offer several advantages for corrosion monitoring in new and repaired RCS. 7,42-45 i gal measurements allowed for real-time and continuous monitoring, providing semi-quantitative information about the corrosion rate and detected the instant where the construction steel depassivates. 44,46 The R p measurements were performed periodically since these allowed for measurement of instantaneous corrosion current density in order to assess the condition of the embedded steel reinforcement related to its corrosion. 47 EIS is a reliable, non-destructive and fast method providing accurate results about corrosion protection behavior of the OIH sol-gel coatings. EIS data were interpreted on the basis of electrical equivalent circuits consisting of a combination of resistances and capacitances associated in series or in parallel providing the same electrical response as the studied electrochemical interface. 23,[48][49][50][51] The potentiodynamic polarization curves were used to compare the corrosion resistance of coated HDGS samples. 52
Sol-gel synthesis procedure of OIH ureasilicate coatings.-The experimental steps involved in the synthesis of ureasilicate matrices to produce the coatings on HDGS samples were performed according to the literature. 27,39,40 Five different materials were prepared as coatings on HDGS, i.e. U(230), U(400), U(600), U(900) and U(2000). The matrices were obtained using 1:2 stoichiometric molar ratio of each Jeffamine molecular weight and 3-isocyanate propyltriethoxysilane. These two reagents were mixed and stirred at 700 rpm for 20 min in a glass container. In the second step, 0.22 M of citric acid ethanolic solution was added, setting the citric acid/3-isocyanate propyltriethoxysi-lane molar ratio equal to 0.094. The mixture was stirred and after 15 min distilled water was added until the total volume of reaction media equaled 8 mL. The final mixture was left to react for a further 15 min.
Preparation of the coatings.-The HDGS samples were obtained from commercially available plates and cut with dimensions of 5.0 × 1.0 × 0.1 (in cm). The HDGS samples with an average Zn thickness of 16 μm on both sides were degreased with acetone. Coated HDGS samples were prepared by dipping the metallic plates of HDGS in the synthesized sol mixture using a dip coater (Nima, model DC Small). The OIH coatings were deposited by one and three consecutive dip steps at a withdrawal speed of 10 mm min −1 without residence time. Previous studies have shown that the thickness of the U(X) deposited on HDGS by one and three dip steps ranged between 2.5-12.7 μm and 3.6-24 μm, respectively 39 depending on the MW of Jeffamine used. Therefore, the choice in producing samples coated by one and three consecutive dip steps was mainly to assess the corrosion performance of thinner and thicker coatings in chloride contaminated simulated concrete pore solutions. Coated HDGS samples were subsequently placed in an incubator-compressor (ICP-400, Memmert) and kept at 40 • C for 15 days.
Preparation of chloride-contaminated simulated concrete pore solution.-The corrosion behavior of HDGS coated samples with the different OIH coatings was studied in solutions simulating the concrete interstitial electrolyte (simulated concrete pore solutions -SCPS) and contaminated with 1 wt% of chlorides (SCPS + 1 wt% Cl − ). SCPS were prepared according to the literature 10,53 by the addition of analytical reagent grades 0.2 M KOH to a Ca(OH) 2 saturated solution previously prepared with distilled water. A final solution with a pH = 13.2 was obtained and 1 wt% of chlorides was added as sodium chloride. This medium was prepared in order to induce the corrosion of the substrate. According to the literature, 54 the critical chloride concentration reported to induce corrosion of reinforcing steel in SCPS, with pH values of 12.5 and 13.9, was of 0.02 and 1 wt%, respectively. Considering that the pH of the SCPS prepared in this work was above 12.5, a value of 1 wt% was used to ensure that the chloride content was above the critical chloride concentration.
Electrochemical studies.-In this work different electrochemical techniques were used, namely: electrochemical impedance spectroscopy (EIS), potentiodynamic polarization curve, polarization resistance (R p ) and macrocell current density (i gal ). Two distinct approaches were used to assess the electrochemical behavior of the OIH sol-gel coatings applied on HDGS. In the first approach, short-term studies on HDGS coated with the different OIHs were conducted in chloride-contaminated simulated concrete pore solutions (SCPS + 1 wt% Cl − ) using EIS and potentiodynamic polarization measurements. In the second approach, the barrier stability of the OIH coatings was monitored through i gal and R p measurements during 16 days in SCPS and on the 8 th day 1 wt% of Cl − was added as sodium chloride. For comparison purposes, uncoated samples (control) were also studied.
Electrochemical impedance spectroscopy (EIS) and potentiodynamic polarization curves measurements.-EIS and potentiodynamic polarization curves measurements were carried out on HDGS coated samples by one and three consecutive dip steps without residence time in SCPS + 1 wt% Cl − . All the measurements were made in triplicate to check data reproducibility. These measurements were conducted in order to study the influence of the number of dip steps used to coat the HDGS samples on the barrier protection in the presence of aggressive species such as chloride ions. The EIS measurements on HDGS coated samples were performed in the first instants of immersion in SCPS + 1 wt% Cl − . The potentiodynamic curves measurements were performed after 2h of exposure to SCPS + 1 wt% Cl − .
The EIS and potentiodynamic polarization curve measurements were performed at room temperature in a Faraday cage. A glass cell with a saturated calomel electrode (SCE) and a platinum foil (exposed area ≈ 8 cm 2 ) were used respectively as reference electrode (RE) and counter electrode (CE). The exposed surface area of the working electrode (WE) (HDGS coated sample) in the electrolyte was ≈2 cm 2 . EIS studies were accomplished by applying a 20 mV (peak-to-peak, sinusoidal) electrical potential within a frequency range from 1 × 10 5 Hz to 0.01 Hz (10 points per decade) at open circuit potential (OCP). Measurements were performed using an Impedance/Gain-Phase Analyzer (Model 1260A, Solartron-Schlumberger) and a potentiostat/galvanostat (Model 1287A, Solartron-Schlumberger) controlled by a PC using Zplot software (Solartron-Schlumberger, version 2.9c). The frequency response data of the studied electrochemical cells were displayed in a Nyquist plot, using ZView software (Solartron-Schlumberger, version 2.9c) that was also used for data fitting purposes. For comparison purposes, cells prepared with uncoated HDGS WE electrodes were used as the control.
Polarization resistance (R p ).-The R p values were estimated by the potentiostatic method using a potentiostat/galvanostat (Voltalab PGZ 301) using a three-electrode electrochemical cell system. 45,46,55 The R p values were calculated from the transients due to the application of a 10 mV anodic potential step for 100 s. A stainless steel (SS, type 316L) plate was used as a CE and HDGS coated with the different OIHs was used as WE. Both electrodes had an active average area of 2 cm 2 . The edges of both the CE and WE plates, as well the non-active areas and connecting zones were protected with a dualcomponent epoxy resin (Araldite). Titanium-activated wire (Ti/TiO 2 ) with a length of 1 cm was chosen to be used as RE due to its low cost and suitability to be embedded in real RCS. The electrodes were connected to an isolated copper cable and the cutting zone of the tip of the RE was covered with epoxy resin. For comparison purposes, cells with non-coated HDGS WE electrodes were prepared and used as the reference (hereafter referred generically as control) and to check data reproducibility triplicate cells were assembled. The R p measurements were performed once a day during 16 days on samples coated by one dip step of U(X). After the first eight days of exposure to SCPS, 1 wt% of chloride ions were added as sodium chloride.
Macrocell current density (i gal ).-The i gal measurements were performed using a system based on two parallel electrodes (rectangular metal plates with dimensions of 5.0 × 1.0 × 0.1 cm - Figure 1) according to the literature. 27,44 For comparison purposes, cells prepared with uncoated HDGS WE electrodes were used as control and triplicate cells were assembled to check data reproducibility. To assemble the electrochemical cells, the SCPS was transferred to a 100 mL polyethylene flask. The electrodes were subsequently im-mersed and the flask closed ( Figure 1). Using an automatic data acquisition system (Datataker DT505, series 3), i gal measurements of the prepared cells, were performed by reading the potential difference to the terminals (shunted with a 100 resistor) according to ASTM G109. 56 Measurements were performed at one-minute intervals during 16 days on samples coated by one dip step. After the first eight days of exposure to SCPS, 1 wt% of chloride ions were added as sodium chloride. The measurements were performed at the same periodicity for more eight days.
Scanning electron microscopy (SEM/EDS).-The morphology of the OIH sol-gel coating surface applied on HDGS specimens was analyzed with a scanning electron microscope (SEM, JEOL JSM-6400) coupled with an EDS detector (Inca-xSight Oxford Instruments). The surface of the samples was covered with an ultrathin coating of gold deposited by sputter coating. SEM investigations of the surfaces were carried out by using the back-scattered electron (BSE) detector in order to emphasize the contrast for the different metallic phases. The SEM/EDS studies of the HDGS coated samples were performed on the substrate before and after 16 days in contact with chloridecontaminated SCPS.

Results and Discussion
Electrochemical impedance spectroscopy (EIS).-EIS measurements were performed in order to assess the electrochemical behavior of HDGS coated by one (1 layer) and three (3 layers) consecutive dip steps of OIH sol-gel coatings in chloride-contaminated SCPS (highly alkaline media, pH = 13.2). Figure 2 shows representative Bode plots obtained for control, U(230) and U(400) in the instant of exposure (≈ 10 minutes once the OCP has been established in SCPS + 1 wt% Cl − ).
The Nyquist plot obtained for control samples (Figure 3a) in the instant of exposure consist of one depressed capacitive loop (one timeconstant). The high frequency (HF) capacitive loop is associated to the charge transfer process of the metal corrosion and the double layer behavior. 48,49 For the coated HDGS samples, the impedance spectra consist of two partially overlapped capacitive semicircles (two timeconstants). Samples coated by one or three layers, with the lowest MWs of Jeffamine (U(230) and U(400) - Figures 2 and 3) reveal the presence of two time-constants in first instants of immersion. The loop at HF, between 10 2 and 10 5 Hz, is generally assigned to the OIH sol-gel film capacitance 19,57 and the one at low frequency (LF) can be assigned to the corrosive process by charge transference. 21,22 The Nyquist plots obtained for samples coated with OIHs synthesized using higher MWs of Jeffamine, namely U(600), U(900) and U(2000) exhibited also two time-constants at high and medium frequency indicating at least two electrochemical processes with similar relaxation time constants, while one inductive loop was observed at the LF end ( Figure 4). This type of impedance spectra have been reported 58-62 for galvanized surfaces due to the surface relaxation processes of adsorbed species (such as corrosion products) on the HDGS surface. Considering the high pH of the chloride-contaminated SCPS and the mechanism proposed by Liebau,8,63 the results suggest that the adsorbed species (such as Zn 2+ , Zn(OH) 2   separately from the cathodic. [64][65][66] A possible cause could be the adsorption or desorption of activating or blocking species with increasing overpotential of each process. Two situations can be considered: (i) If the anodic adsorption (therefore cathodic desorption) of a blocking species occurs the inductive loop appears due to the dominant effect in the cathodic process; or (ii) if the anodic desorption (therefore cathodic adsorption) of a blocking species for the reactions (OIH layer) occurs then the observed inductive loop results as the dominant effect in the anodic component. 64 This behavior at LF range suggests that these OIH coatings are not entirely protective and may change when in contact with the electrolyte, thus exhibiting defects. Barely coated regions also allow the electrolyte to penetrate across coating cracks and reach the substrate. Additionally, in a chloride environment, corrosion products such as zinc hydroxychlorides are formed on the surface of zinc. 67 The LF capacitive loop [67][68][69] can be also attributed to the diffusion of the electrolyte in the pores of the coatings and the dissolution of the zinc layer. For samples coated with U(230) and U(400), no inductive loops in the LF region were found. This indicates that these OIH coatings confer good corrosion protection and that the corrosion processes in the first instants of contact with chloride-contaminated SCPS are effectively delayed comparatively to U(600), U(900) and U(2000) coatings. This behavior can be related either to the breakdown of the former protective surface film or relaxation of adsorbed species in the mechanism of zinc dissolution. 65  Detailed information on the behavior of the different OIH solgel coated samples can be extracted from the fitting of the spectra using electrical equivalent circuits (EEC). The selected EECs have been widely used for analysis of impedance spectra of metals coated with sol-gel films. 23,30,33,70 The interpretation of impedance spectra was based on the EECs that include contributions from the sol-gel coating and the corrosion process itself. Capacities were replaced by CPEs to improve the quality of the fitting. The physical origin of the CPE behavior is not completely understood, however, it is generally accepted that it is assigned to inhomogeneous physical properties of the system. 71,72 Two different EECs (Figure 3) were used to fit the experimental data. Figure 3e shows the EEC used to fit the control data results and Figure 3f shows the EEC used to fit the results obtained for samples coated. The solid lines represent the results of the fitting. For samples coated by three consecutive dip steps only the fitting obtained for U(230) is displayed. The Bode plots obtained for samples coated with 1 layer of U(600), U(900) and U(2000) and 3 layers of U(600) are shown in Figure 5. To avoid overloading of Figure 5 only the fitting line for U(600) 1 layer and the results for the sample coated with 3 layers of U(600) are shown. The fitted region for these samples has been obtained by excluding the range of frequencies associated to the inductive process and the EEC used is the same as depicted in Figure 3f. For all the EECs, the electrical elements R s , CPE dl R dl, R coating and CPE coating, are associated respectively, with: the electrolyte resistance; double layer capacitance at the metal-electrolyte interface; the charge transfer resistance of zinc; resistance of the sol-gel coating and capacitance of the sol-gel coating. The fitting parameters are shown in Tables I and II and considering the χ 2 obtained for each one it can be assumed that a good fit is shown in Figures 3 and 5.
All the coatings show comparable responses immediately after their immersion in chloride-contaminated SCPS, exhibiting low capacity and high resistance (Table I and Figure 6a), showing a beneficial resistive behavior compared to control. Figure 6a shows that as the MW of Jeffamine increases the R coating and the CPE coating decrease and increase respectively. Generally, samples coated by three consecutive dip steps show higher resistances and lower capacities than samples coated by one dip step ( Figure 6). The highest R coating was given by samples coated with U(230) and the lowest by samples coated with U(2000) (Table I and Figure 6). In general, the increase of the MW of Jeffamine induced an increase of CPE dl and a decrease of the R dl values (Figure 6b). These variations confirm the poorer barrier proper- ties provided by the OIH coatings when higher MWs of Jeffamine are used. The reason for this behavior may be explained considering the structure of these materials. These OIHs can be described as a rigid inorganic (silicate based backbones) that provide enhanced mechanical properties which are spaced by flexible organic polyether chains linked by urea bonds. 27,39 These polyether chains increase when the MW of Jeffamine increases. Therefore, the organic component of the Table I Figures 4e) and 4f).

R
Potentiodynamic polarization studies.-The potentiodynamic polarization methods can provide a direct measure of the corrosion current, which can be related to corrosion rate and can be used to examine the passivation of a metal in an electrochemical system. Figure 7 shows the polarization curves for uncoated and coated HDGS with one (1 layer) and three (3 layers) dip steps of U(230), U(400), U(600), U(900) and U(2000) recorded after 2 h of immersion in chloride-contaminated SCPS (1 wt% of Cl − was added to SCPS).
The potentiodynamic polarization curves obtained for HDGS coated samples with U(230) and U(400) are appreciably different from the control (Figure 7). First, the OCP of this OIH sol-gel coatings is significantly higher than that of the bare HDGS. Secondly, a passivation region with a rather low passivation current densities is observed, which indicates that U(230) and U(400) coatings provide a physical barrier by blocking the electrochemical process. HDGS coated by one or three consecutive dip steps of U(600), U(900) and U(2000) show OCP values very similar to the control, yet low passivation current densities were recorded. Moreover, the control exhibited a polarization curve with a passive region comparatively narrower than all the OIH-coated HDGS.
Quantitative information on corrosion currents and corrosion potentials can be extracted from the slope of the curves, using the Stern-Geary equation, 73 as follows : Table III shows representative electrochemical parameters obtained namely the Tafel slopes (βc and βa) and the polarization resistance (R p ). The corrosion current density (i corr ) and corrosion potential (E corr ) were determined by analysis of Tafel curves and are shown in Figures 8 and 9, respectively. The protection efficiency (PE%) was calculated by using the following equation: where i corr and i * corr are the corrosion current densities obtained for uncoated and coated HDGS, respectively and is shown in Figure 8.
Generally, coated HDGS samples in the presence of 1 wt% of Cl − show improved results when compared to the control (Table III and Figure 7).    Figure 8 also shows that the highest and lowest PEs were given by samples coated with U(230) and U(2000), respectively. Figure 9 clearly shows that the E corr of samples coated with U(230) and U(400), either by one or three layers, was significantly higher than that of the bare HDGS. This might be due to the effective suppression of the cathodic reaction from water hydrolysis. It can also be observed that the E corr reduces as the MW of Jeffamine increases which indicates that coatings synthesized with higher MWs of Jeffamine are not so effective in suppressing the cathodic reaction when compared to coatings produced with lower MWs of Jeffamine. Although the Cl − is a strong anodic activator due to its small radius and volume the results  . Variation of the E corr values obtained for HDGS coated by one dip step (one layer) and three consecutive dip steps (three layers) of the different OIH U(X) after being exposed in SCPS + 1 wt% Cl − during 2h. The dashed red line represents the E corr value obtained for uncoated HDGS (control), which was included for comparison purposes.
obtained for samples coated with U(230) and U(400) indicate that the Cl − cannot reach the surface of the substrate after 2h of immersion. These findings suggest that the diffusion of Cl − ions across the coating layer is prevented and the Cl − ions remain entrapped within the OIH network being unable to reach the surface of the metallic substrate. Furthermore, the potentiodynamic polarization curves are in agreement with EIS measurements. Figure 10 shows the i gal collected from the different prepared electrochemical cells involving the HDGS coated samples and the control that were immersed in SCPS during 16 days. After 8 days of immersion, 1 wt% of Cl − was added into the SCPS. Generally, the coated samples showed high i gal values in the first two days due to the corrosion of zinc. An exception to this tendency can be seen for the i gal results obtained in the first day of immersion for samples coated with U(230) and U(400). These samples, when compared to U(600), U(900) and U(2000) showed the lowest i gal values. After two, and in certain cases almost three days (U(600)), an oxide layer was formed in the working electrode surface and a decrease in the i gal values was recorded, which is consistent with the literature. 27 The zinc in contact with SCPS (pH = 13.2), was oxidized and the cathodic reaction arisen from water hydrolysis with hydrogen evolution took place on the galvanized surface: 9,12,15,74,75 Anodic dissolution of zinc:

Macrocell current density (i gal ).-
Cathodic reaction from water hydrolysis: The global process can be described as: After 5 days of immersion, samples coated with U(600) and U(2000) dropped to lower i gal values compared to samples coated with U(230) and U(900) which showed steady behavior until the addition of chloride ions. Lower values for samples coated with U(230) 27 were expected after 5 days of immersion. This behavior suggests that the oxide layer formed did not fully cover the HDGS leading to the passivation of the working electrode surface; however no further conclusion can be drawn. The comparison of the i gal values before and after Cl − addition shows that for the coating U(2000) the i gal values increased remarkably compared to the other coated samples. A similar behavior, however, less obvious, was found for samples coated with U(400) and U(600). Samples coated with U(230) and U(900) showed a slight decrease of the i gal values after Cl − addition. Excluding samples coated with U(400), which reached lower i gal values on the 12 th day increasing after that to higher values, the i gal measurements ( Figure 10) showed that after Cl − addition a steady behavior along time was found for each OIH coating. Lower i gal values represent improved corrosion behavior. Coated HDGS samples showed lower values than the control during all the study period, therefore showed improved corrosion behavior. The results also indicated that the i gal cells in SCPS were sensitive to the external laboratory temperature variation, to the presence of chloride ions and to the composition of the OIH material deposited which is also according to the literature. 44,46 Polarization resistance (R p ).- Figure 11 shows the R p values obtained for the different electrochemical cells prepared with HDGS coated by one dip step (1 layer), that were immersed in SCPS during 16 days. Results for control samples are also presented. After 8 days of immersion, 1 wt% of Cl − was added into the SCPS. The R p measurements were performed once a day. Figure 11 shows that the R p data are generally in agreement with i gal measurements. The results also indicate that the R p cells in SCPS are sensitive to the presence of chloride ions and to the composition of the OIH material deposited which is also according to the literature. 7,46,76 HDGS coated samples during all the period of study, in general, displayed higher values than the control. Therefore, showed improved corrosion behavior. During the first days of immersion, the R p values decreased suggesting that zinc corrosion occurred, which is also in agreement with the i gal results and literature. 27 After chloride addition (on the 8 th day) the R p values of samples coated with U(600), U(900) and U(2000) dropped to lower values, however, remained always above the control. Improved results were obtained by samples coated with U(230) and U(400) when compared to the other samples and the R p values are also in agreement with the EIS, potentiodynamic measurements, and i gal data. Morphology of the coatings.-Surface morphology of coated HDGS samples was assessed by SEM/EDS analysis before ( Figure 12) and after contact with SCPS chloride-contaminated (Figures 14 and  15). Preliminary observations of the HDGS surface after being in chloride-contaminated SCPS during 16 days (1 wt% of chloride was added on the 8 th day) were performed using a stereo-zoom microscope and are presented in Figure 13. SEM analysis ( Figure 12) revealed that OIH coatings covered the HDGS substrate and as the MW of Jeffamine increased an improved coverage was obtained. This is due to the increase of the sol viscosity since as the MW of the Jeffamine increases the viscosity of the obtained OIH coatings increases.  Additionally, during the curing process the action of gravity led to the displacement of the gel from the top to the bottom. This is particularly significant for samples coated with U(230) and U(400). Figure 12a clearly demonstrates that samples coated by one dip step of U(400) have several regions, barely coated within the same sample, when compared to samples coated by one dip step of U(2000). The EDS analysis shows that the OIH coating correspond to high peaks of C, Si and O. Uncoated areas only show the presence of high peaks of Zn ( Figure 12).
The preliminary observations of the HDGS surface (working electrode used for i gal measurements) after being in chloride-contaminated SCPS during 16 days (1 wt% of chloride ions were added on the 8 th day) were performed using a stereo-zoom microscope. The images obtained for the control sample, and samples coated by one dip step of the different OIH coatings are shown in Figure 13. The control samples were HDGS samples where no OIH coating was deposited. The visual observation clearly shows a difference between coated and uncoated (control) samples. The control sample shows severe corrosion regions where the presence of iron oxide (rusty deposits) is clear. White deposits (zinc oxide) and calcium hydroxide zincate (CaHZn) crystals are also visible. 8 However, due to the high pH of SCPS the size of CaHZn crystals are so large that they cannot completely cover the HDGS surface. As a consequence of this, small regions of the substrate are left exposed and thus without protection. Under these conditions, the passivation of the HDGS surface is not possible. At high pH values the concentration of Ca 2+ ions in solution is depleted, therefore the dissolution of the zinc is not mitigated and as a result the galvanized coating may totally dissolve 8 leading to the corrosion of the steel underneath.
The appearance of high CaHZn crystals on the surface of HDGS samples coated by one dip step of U(230), U(400) and U(600) may be explained by the poor coverage obtained when these OIHs are deposited. The barely or uncoated areas are attacked by the electrolyte leading to the formation of CaHZn crystals with higher dimensions than the ones formed on the surface of HDGS coated with U(900) and U (2000). In spite of the presence of a few crystals of CaHZn on the surface of coated HDGS, all the coated samples show compact corrosion product films. Furthermore, the presence of iron oxide was not observed. The corrosion product film formed, consisting of many needle-like particles, is compact and covers the HDGS surface completely. The formation of these films is due to the dissolution reac-tion of the zinc layer in chloride-contaminated SCPS. 8 Figure 13 also shows that the corrosion product film becomes more compact when the coatings used were synthesized with higher MWs of Jeffamine (U(900) and U (2000)). This may be explained by the high coverage that was achieved with these OIHs (Figure 12b). Once the passive film of calcium hydroxyzincate is formed, its stability is not altered. 8 EDS analysis for control samples (Figure 14) show, after 16 days in SCPS (1 wt% of Cl − was added on the 8 th day) high intensity peaks of iron. These results indicate that the substrate was attacked by the electrolyte and in certain areas the entire zinc layer was completely eroded (due to zinc corrosion). Coated HDGS samples showed improved results and in the same experimental conditions the presence of U(230) and U(400) was found which is justified by the existence of high intensity peaks of C, Si and O ( Figure 14). Moreover, the presence of iron was not found on the surface of the coated HDGS. The absence of Si peaks in the EDS data obtained for U(600) and U(2000) ( Figure 15) (similar results were obtained for U(900)) suggest that the coatings synthesized with higher MWs of Jeffamine suffer partial dissolution and or destruction after 16 days of immersion in SCPS (1 wt% of Cl − was added on the 8 th day). This behavior may be explained by the rupture of the OIH coatings during the formation of the corrosion product. Nevertheless, the formation of a compact corrosion product film is allowed and the surface of the HDGS is passivated. In spite of these OIHs (U(600), U(900) and U(2000)) suffering partial dissolution/destruction, the results suggest that all coated samples mitigate the zinc corrosion and the hydrogen evolution in the conditions studied. The EDS obtained for coated samples also show (Figures 14 and  15) the presence of chloride suggesting that the Cl − ions remained entrapped within the OIH network, stopping them from reaching the surface of HDGS and thus protecting the substrate from Cl − attack.
In summary, the electrochemical results obtained (i gal , R p , EIS and potentiodynamic curves) are generally in agreement and point to the same outcome showing that improved anti-corrosion performance was given by samples coated with U(230) and U(400) (produced with lower MWs of Jaffamine). Inferior results were provided by samples coated with U(600), U(900) and U(2000) (produced with high MWs of Jeffamine). This behavior may be explained by the increase of the organic chains, which increase when the MW of Jeffamine increases. In highly alkaline environments, as mentioned previously the organic chains are partially damaged by the electrolyte leading to rupture of the coating in certain areas. This rupture is significantly higher in  samples coated with OIHs prepared with higher MWs of Jeffamine leading to inferior results compared to samples coated with lower MWs of Jeffamine. The SEM/EDS results are also in agreement with these findings.

Conclusions
The present work reported the electrochemical study of U(X) solgel based coatings on HDGS in chloride-contaminated SCPS. It was demonstrated that the U(X) coatings prevent chloride ions reaching by diffusion to the surface of the metallic substrate by immobilizing them within the most exterior regions of the OIH network of the coating.
The analysis of the results obtained from electrochemical studies (i gal , R p , EIS and potentiodynamic curves) allowed to conclude that improved anti-corrosion performance was given by samples coated with U(230) and U(400). The poorer results, still better than the control, were given by samples coated with U(2000). SEM/EDS results, in agreement with the i gal and R p results, pointed to the conclusion that full coverage was seldom achieved which is consistent with the literature. 39,40 In conclusion, besides the barrier effect introduced by U(X) coatings by hindering the zinc corrosion activity during the initial stages of contact of the HDGS samples with SCPS (highly alkaline environment), these coatings protect the HDGS from Cl − attack and may be considered potential substitutes for chromate conversion layers and systems containing Cr(VI). Therefore, these U(X) sol-gel coatings can be employed as pre-treatments to reduce the corrosion in the first instants of immersion in SCPS and protect the HDGS from a further attack of Cl − .