Electrochemical and Salt Spray Testing of Hybrid Coatings Based on Si and Zr Deposited on Aluminum and Its Alloys

Hybrid coatings based on Si and Zr were assessed for their protection of aluminum and its alloys 2024-T3 and 7075-T6 against corrosion in 0.5 M NaCl. Coatings were prepared from sol synthesized from mixtures of tetraethyl orthosilicate and 3-methacryloxypropyl trimethoxysilane to which various concentrations of zirconium tetrapropoxide and methacrylic acid were added. Shortand long-term electrochemical behavior was assessed. The degree of protection was dependent on the sol composition, ageing time and curing temperature. A coating containing Zr/Si ratio of 0.41, aged for 48 h and cured at 100◦C exhibited the best combination of resistance towards general and localized corrosion. Electrochemical parameters obtained under polarization (corrosion current density of 4.4 nA/cm2 and stability up to 7 V), and under open circuit conditions (impedance in the few tens of Mohm cm2 after one week immersion) prove that the protection was high. At other conditions, such as higher Zr/Si ratios, shorter ageing time or curing at room temperature, protection was excellent early after immersion but then lessened rapidly. Coatings prepared under optimal conditions provided a high degree of protection over 500 hours salt spray testing, especially on aluminum metal. These coatings have the potential to provide good protection in chloride environments. © The Author(s) 2015. Published by ECS. This is an open access article distributed under the terms of the Creative Commons Attribution 4.0 License (CC BY, http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse of the work in any medium, provided the original work is properly cited. [DOI: 10.1149/2.0801510jes] All rights reserved.

While aluminum has good corrosion resistance in seawater, its high-strength alloys in the 2xxx (Al-Cu-Mg) and 7xxx series (Al-Zn-Mg) are normally not recommended for use under these conditions. 1,2 Since there is an ongoing need for light-weight, high-strength materials, the possibility of using alloys other than those of the series 5xxx (Al-Mg-Mn) and 6xxx (Al-Mg-Si) normally recommended for sea-coast applications has to be considered. 2,3 Aluminum and its alloys are highly susceptible to oxidation, resulting in the passivation of the surface by a 0.01 to 10 μm thick oxide layer when exposed to air. This naturally formed oxide protects aluminum under atmospheric conditions. 2 Passivation of the alloys is, however, less effective. Due to the presence of unevenly distributed intermetallic particles (IMPs), the passive film is, overall, thinner and less uniform. IMPs such as Al 2 CuMg, Al-Cu-Mn-Fe, Al 2 Cu, Al 3 Fe, MgSi 2 , MgZn 2 , create local galvanic cells that promote corrosion on the surface. 4-6 Al 2 O 3 is a good insulator but, when containing IMPs, is a semiconductor that allows limited passage of electrons. 2 High concentrations of chloride ions allows localized breakdown of the aluminum oxide film, most commonly observed as pitting, crevice or galvanic corrosion.
Al alloys have been protected for decades using chromate conversion coatings. In the last decade these coatings have been widely prohibited within 7 and outside 8 the EU due to the fact that hexavalent chromium is a toxic, carcinogenic and environmentally hazardous compound. [9][10][11] Other environmentally hazardous compounds, such as volatile organic compounds (VOCs) used, for example, in the painting industry, have also been subject to stricter regulations. The need therefore exists for non-toxic, environmentally friendly surface treatments capable of providing corrosion protection. 2 Sol-gel systems are one of the options for replacing chromatebased pre-treatments and reducing the use of VOCs. 2 The sol-gel process is convenient for synthesising oxide films from alkoxysilyl containing materials via continuous reaction steps of hydrolysis and (poly)condensation. [12][13][14][15][16] The sol-gel process allows incorporation of organic molecules inside the inorganic network, thereby forming organic-inorganic hybrid materials. These have several advantages such as curing at relatively low temperatures of less than 200 • C, 15 and the fact that the process is inexpensive and easy, especially when compared to deposition techniques involving vacuum. Low temperature curing of the coating is important also for the underlying substrate which undergoes artificial ageing at temperatures between 100 and 200 • C. 2 Sol-gel coatings based specifically on the precursors tetraethyl orthosilicate (TEOS) and 3-methacryloxypropyl trimethoxysilane (MAPTMS) show good corrosion protection. 15,17,18 Their anticorrosion properties can be improved by adding non-silicon, metal alkoxides such as zirconium tetrapropoxide (ZTP). Control of the reactivity of the latter precursors enables the structure of the resulting materials to be tailored. Chemical additives, such as organic acids, acetylaceton and 2,2-bipyridine, have often been used as chelating ligands to slow the hydrolysis and (poly)condensation reactions of nonsilicon metal alkoxides. 19 The effect of organic chelates on the performance of hybrid sol-gel coatings deposited on the alloy AA2024-T3 has been studied. 19 Good corrosion properties were achieved when ZTP was chelated with methacrylic acid (MAA). 20 Similar behavior was observed when hybrid coatings based on TEOS, MAPTMS, ZTP and MAA were applied to AA7075-T6 alloy. 20,21 The chelating reagents also influence the polymerization of the sol constituents, especially the acrylic group that polymerizes in the presence of UV light. To improve the cross-linking by UV radiation, an organic resin, photoinitiator or diluent/monomer is often added. [22][23][24][25][26][27][28][29][30][31] The same effect can be achieved by adding MAA and ZTP to a siloxane sol. 20,21,24,[31][32][33] Coatings cured by UV radiation at room temperature showed better resistance to corrosion of AA7075-T6 under simulated aircraft conditions (so called Harrison's solution) than coatings cured thermally, as shown in our previous study. 32 The good corrosion resistance of the coatings containing zirconium precursors was ascribed to the presence of ZrO 2 nanoparticles that appear to have a pore blocking effect. 15,34 The corrosion resistance was shown to depend on the concentration and size of the particles. 15 The corrosion of AA2024-T3 in 3% NaCl was reduced when coatings containing 20 mol% ZrO 2 were applied. 34 ZrO 2 nanoparticles were also formed with other chelating reagents, such as acetylacetone, 2,2bipyridine and 3,4-diaminobenzoic acid. 19,20,23 The degree of corrosion protection depends on the type of chelating agent used. 20,35 The hybrid silane coatings based on TEOS, MAPTMS, ZTP and MAA are amorphous and do not contain ZrO 2 particles seen by SEM, AFM and TEM techniques. 33 This was ascribed to the chelation of ZTP with MAA and the polymerization of the acrylic group. In addition to these two factors, the presence of Si−O−Si and Si−O−Zr bonds within the coating is the main basis for their strong corrosion resistance under simulated aircraft conditions. 32 The relative concentrations of Zr/Si, ZTP/MAA and MAPTMS/ MAA influence the performance of coatings deposited on AA7075-T6 in dilute Harrison's solution, as shown in our previous study. 32 The degree of corrosion protection increases with increase in the Zr/Si ratio that is responsible for establishing Si−O−Si and Si−O−Zr bonds and for increasing the ZTP/MAA ratio that determines the degree of chelation. In order to test the possibilities for the use of these hybrid coatings we continue the study, using the same type of Si and Zr hybrid coatings, but expanding their composition and ageing/curing conditions and examining other substrates and another solution. In the present study the coatings were deposited on aluminum and two alloys, AA2024-T3 and AA7075-T6, and studied electrochemically in 0.5 M NaCl solution as an approximation to seawater conditions. All coatings were aged under daylight to enable UV-induced polymerization. The effects of ZTP and MAA contents, ageing time and curing temperature were studied. In addition to electrochemical potentiodynamic measurements and electrochemical impedance spectroscopy (EIS), salt spray chamber testing was employed to assess the long-term stability to 500 h exposure.
Chemicals sodium chloride (NaCl), sodium hydroxide (NaOH) and HCl used for electrochemical and salt spray experiments were of p.a. quality and supplied by AppliChem (Darmstadt, Germany).
Synthesis of sols and deposition of coatings.-Organic-inorganic hybrid coatings were synthesized according to previous studies 20,21,32 by combining two, separately prepared, alkoxide sols (Sol 1 and Sol 2). Sol 1 is a mixture of TEOS and MAPTMS hydrolysed under acid conditions. Sol 2 is a mixture of ZTP and MAA. Sol 1 and Sol 2 were combined to give a hybrid sol-gel solution denoted as TMZ (TEOS, MAPTMS and ZTP). Compositions of TMZ sols are summarized in Table I.  Table II.
Discs of diameter 14 mm were cut from the sheets, then ground successively with 2400-and 4000-grit SiC emery papers (Struers, United Kingdom), rinsed thoroughly with deionized water and cleaned ultrasonically in ethanol for 10 minutes.
Deposition and curing of the coatings.-After synthesis the TMZ sols were aged, applied in the form of coating on the substrate using a spin coater (Laurell WS-650-23NPP/LITE/IND) and then cured as described. 20,21,32,33 Different ageing and curing treatments were used (Table I). Sols TMZ-1 to TMZ-5 were aged for 48 h before application in the form of coating and cured at 100 • C for 1 h. Curing at room temperature was also examined (TMZ-5 * ). TMZ-6a differed from other sols in that the ageing time was only 1 h and samples were cured at 100 • C (TMZ-6a) or room temperature (TMZ-6a * ). For the sake of comparison, the TMZ-5 coating was aged for 1 h and cured at room temperature (TMZ-5a * ) or at 100 • C on a preheated hot plate (TMZ-5a). All coatings were cured under daylight to stimulate polymerization and densification due to presence of UV component compared to curing in the dark. 32 Electrochemical measurements.-A three-electrode standard corrosion cell (Corrosion Cell Kit, model K0047, volume 1 L, EG&G) was used, at 25 • C, for electrochemical measurements. The working electrode was embedded in a Teflon holder (model K0105 Flat Specimen Holder Kit, EG&G), leaving an area of 0.95 cm 2 exposed to the solution. A saturated calomel electrode (SCE, 0.241 V vs. Saturated Hydrogen Electrode), placed in a Luggin capillary, was used as reference electrode. All potentials in the text refer to the SCE scale. Carbon rods served as counter electrode. Electrochemical experiments were carried out with an Autolab PGSTAT 12 (Metrohm Autolab, Utrecht, The Netherlands) potentiostat/galvanostat controlled by Nova 1.10 software.
The electrochemical measurements were made in 0.5 M NaCl, pH = 5.9 prepared using Milli-Q Direct water. Linear polarization and potentiodynamic polarization measurements.-Prior to measurements, samples were allowed to stabilize under open circuit conditions for approximately 1 hour to bring them to equilibrium. During that time, the open circuit potential (E oc ) was measured as a function of time. Electrochemical measurements were carried out following stabilization.  Table I Linear polarization was measured vs. stable E oc over a potential range of ±10 mV, at a scan rate of 0.1 mV/s. Values of polarization resistance, R p , were deduced from the slopes of fitted current density vs. potential lines using Nova software. The values of R p reflect the resistance of the aluminum to general corrosion. Potentiodynamic curves were recorded at 1 mV/s potential scan rate, starting at 250 mV negative to the stable E oc . The potential was then increased in the anodic direction.

Table III. Electrochemical corrosion parameters measured in 0.5 M NaCl for uncoated aluminum and coated with TMZ coatings after 1 h stabilization at the open circuit potential: polarization resistance (R p ), corrosion current density (i corr ), cathodic and anodic Tafel slopes (β c , β a ), corrosion potential (E corr ), pitting potential (E pit ) and difference in potentials ( E = E pit -E corr ). Results are presented as mean value ± standard deviation. Composition of the coatings and conditions of ageing and curing are given in
Corrosion current density (i corr ), corrosion potential (E corr ), anodic and cathodic Tafel slopes (β a , β c ) were obtained from polarization curves by Tafel approximation. For each sample, measurements were performed in at least triplicate. In tables, results are presented as mean values ± standard deviation. Representative measurements were chosen and presented in graphs. Possible inhibition of the cathodic reaction, due to the deposition of a layer on the surface, was assessed from the cathodic polarization curves. Barrier properties, such as breakdown of the film, pitting potential (E pit ), and difference between pitting and corrosion potentials, E ( E = E pit -E corr ), were obtained from the anodic polarization curves. Electrochemical impedance spectroscopy.-Selected samples, i.e. TMZ-3, TMZ-5 * and TMZ 6a * , deposited on aluminum substrate were tested by electrochemical impedance spectroscopy. Samples were immersed in 0.5 M NaCl solution open to air at room temperature for up to 1 week. Electrochemical impedance spectra were recorded between 10 mHz and 100 kHz using a Frequency Response Analyzer (FRA) module (Autolab PGSTAT 12, Eco-Chemie, The Netherlands). The amplitude of the sinusoidal voltage signal was 10 mV. Impedance responses were recorded at E oc after different times of immersion.
For each sample, measurements were performed at least in triplicate.
Salt spray test.-Selected samples, i.e. TMZ-3 and TMZ 6a * , deposited on aluminum and aluminum alloys substrates were tested by salt spray testing. The salt spray chamber had 0.17 m 3 capacity (ASCOTT, Staffs, Great Britain) and was operated according to the standard. 36 The pH of NaCl solution (50 ± 1% g/L) was set between 6.0 and 6.5 at room temperature to give values of pH between 6.5 and 7.2 after heating the solution to 35 • C. pH was adjusted with 0.1 M NaOH or HCl solutions.
The device for spraying the salt solution comprised a supply of clean air of controlled pressure and humidity, a reservoir to contain the solution to be sprayed and a single sprayer. The compressed air was passed through a filter to remove all traces of oil or solid matter. The temperature of the hot water in the saturation tower was 46 • C and the overpressure 85 kPa. The temperature in the salt spray chamber was set to 35 • C ± 2 • C. Each test lasted for 21 days.
The samples were taken from the chamber every 24 hours and photographed (Canon digital camera).

Results
The corrosion protection of aluminum and its two alloys, AA2024-T3 and AA7075-T6, was examined in terms of coating composition, ageing and curing conditions aiming to explore the use of these materials in chloride environment.

Electrochemical potentiodynamic measurements on aluminum.-
The effect of composition of the coating.-The values of R p of only a few k cm 2 and of i corr in the range of several hundreds of nA cm -2 for uncoated aluminum confirm its poor corrosion resistance (Table III). At potentials more positive than E corr a constant increase in current density was observed indicating pitting of aluminum (Fig. 1). The aluminum surface is thus not stable and requires additional corrosion protection. First, coatings containing different amounts of ZTP (0.06, 0.12 and 0.48 mol) at constant MAA content (0.12 mol) were considered (TMZ-1, TMZ-2 and TMZ-3). The coatings were prepared from aged sols and cured for 1 h at 100 • C. Coated samples exhibited R p values three orders of magnitude greater than that for uncoated aluminum, reaching between 6.9 and 9.6 M cm 2 . The shape of the polarization curves ( Fig. 1) of coated samples shows that the coatings act as a barrier between substrate and chloride medium, with values of i corr smaller by four orders of magnitude (from 3.25 to 5.75 nA cm -2 )   (Table III). At the same time the value of E corr did not change significantly (Fig. 1). At potentials more positive than the Tafel range, the current density plateau was established at approximately 10 -6 A cm -2 . It remained independent of potential up to the high electrode potential of 7 V. This behavior was observed for all three coatings. Coatings TMZ-1 to TMZ-3 thus provide excellent protection of aluminum in chloride solution. The differences between the coatings are small and almost insignificant when compared to that from the uncoated substrate. The behavior is comparable to that observed for corrosion protection of AA7075-T6 in Harrison's solution. 32 The effect of different amounts of MAA (0.12, 0.48 and 0.96 mol) was investigated at constant content of ZTP (n = 0.48 mol) in coatings TMZ-3, TMZ-4 and TMZ-5 respectively. The coatings were prepared from aged sols and then cured for 1 h at 100 • C after deposition. Coated samples exhibited R p values from 9 to 17.4 M cm 2 , the value approximately doubling as the content of MAA was increased (Table III). The value of i corr was lower for larger amounts of MAA (Fig. 2). At the same time the E corr value shifted to slightly more positive potentials. The most significant difference between these coatings is the breakdown potential, E pit . While the TMZ-3 coating remained stable up to 7 V, coatings TMZ-4 and TMZ-5 exhibited localized breakdown, resulting in E pit values of about 0.3 V (Fig. 2). Thus, despite larger R p value and a smaller i corr value, increased amounts of MAA led to coatings inferior to that of TMZ-3, due to their susceptibility to localized breakdown. Effects of ageing time and curing temperature.-The effects of ageing time and curing temperature on the corrosion properties of coated aluminum substrates were investigated. Coatings were aged for 48 h then cured at 100 • C (TMZ-5) or at room temperature (TMZ-5 * ), or aged for only 1 h and then cured at 100 • C (TMZ-5a) or at room temperature (TMZ-5a * ) (Table III and Fig. 3). A longer ageing time is shown to be essential to achieving good protection, especially for curing at room temperature, at which the best results, i.e. the largest R p value and smallest values of i corr and E pit > 7.9 V, were achieved (TMZ-5 * ). Curing at 100 • C exhibited only slightly poorer R p and i corr values but the coating started to break down already at 0.29 V (TMZ-5). Coatings prepared from fresh sol (TMZ-5a and TMZ-5a * ) both exhibited results that were poorer than those from aged sols; the values of current density being greater and breakdown of the coating appearing earlier at 0.04 V and -0.55 V. These results underline that the ageing process is very important for effective anti-corrosion characteristics of these hybrid coatings in chloride solution. Similar behavior was observed in Harrison's solution for AA7075-T6 coated with TMZ-5 coating. 32 Optimal results were obtained by longer ageing times at room temperature; coatings aged for shorter times could not compete with coatings aged longer times regardless of the curing temperature. The effect of larger amounts of ZTP and MAA (sol TMZ-6a).-In order to decrease the time of preparation of the coating, ageing of the sol was accelerated by adding larger amounts of ZTP and MAA (sol TMZ-6a). TMZ-6a was aged for only 1 h as opposed to the 48 h used for sols TMZ-1 to TMZ-5. The relative effectiveness of corrosion protection of the coatings TMZ-5a, TMZ-5a * , TMZ-6a and TMZ-6a * was determined from the effects of ageing time (1 h and 48 h) and curing temperature (room temperature and 100 • C) (Table III, Fig. 4). For the coatings cured at 100 • C and aged 1 h, increased contents of ZTP and MAA  resulted in values of R p from 6.2 M cm 2 (TMZ-5a) to ≈100 M cm 2 (TMZ-6a) ( Table III). The values of i corr were accordingly smaller. However, the breakdown potential for the TMZ-6a coatings occurred at −0.21 V, ∼200 mV more negative than for TMZ-5a. Further improvement in R p and i corr values was observed for the coatings cured at room temperature and aged 1 h, TMZ-5a * and TMZ-6a * . For TMZ-6a * , extreme values of almost 260 M cm 2 and 0.1 nA cm -2 were observed. At the same time the value of E pit shifted in a more positive direction compared to values obtained after curing at 100 • C (−0.21 V for TMZ-6a) and for coatings containing smaller amounts of ZTP and MAA and cured at room temperature (−0.55 V for TMZ-5a * ). However, none of the coatings prepared from sols aged for only 1 h (TMZ-5a, TMZ-5a * , TMZ-6a, TMZ-6a * ) achieved the stability of coatings prepared from sols aged for 48 h (TMZ-5, TMZ-5 * , Fig. 3), regardless of the curing temperature.
Electrochemical potentiodynamic measurements on aluminum alloys 2024-T3 and 7075-T6.-Selected TMZ coatings were applied to AA2024-T3 and AA7075-T6 substrates. Electrochemical corrosion parameters for the most effective coatings on the former are summarized in Table IV. Anti-corrosion properties of AA2024-T3 in 0.5 M NaCl were significantly improved by protection with sol-gel coatings. For coatings TMZ-2 and TMZ-3, values of R p and i corr similar to those on coated aluminum were obtained. The effectiveness of the protection was also confirmed by potentiodynamic polarization measurements (Fig. 5). Regardless of the ZTP content, coatings exhibited no localized pitting corrosion up to the high electrode potential of 7 V and the current density remained below 10 -6 A/cm 2 . The hybrid coatings described here achieve better corrosion protection of AA2024-T3 than other hybrid sol-gel coatings containing ZrO 2 . 19,34 Curing of coatings containing larger ZTP content (TMZ-5 * ) was also optimal at room temperature for corrosion protection of AA2024-T3 substrate, a high R p value in the vicinity of E corr along with the broad range of stability at high electrode potentials being achieved (Table IV, Fig. 5). Compared to TMZ-5 * , TMZ-6a and TMZ-6a * coatings exhibit even larger values of R p and smaller i corr values, but were susceptible to localized breakdown already at -0.37 V. As with aluminum, coating TMZ-6a on AA2024-T3 substrate is highly resistant to general corrosion in the vicinity of E corr but susceptible to breakdown once polarized to anodic potentials.
The TMZ coatings deposited on AA7075-T6 substrates exhibit similar behavior as on AA2024-T3 substrate, except for the TMZ-6a (aged 1 h and cured at 100 • C) and TMZ-6a * (aged 1 h and cured at room temperature) coatings (Table V, Fig. 6). In contrast to aluminum and AA2024-T3 substrates, where extremely high values of R p (100-200 M cm 2 ) were achieved, the TMZ-6 coatings exhibit values of 0.24 and 1.8 M cm 2 for TMZ-6a and TMZ-6a * , respectively. The shape of the potentiodynamic curves follows that of uncoated substrate.
The presented results show that the TMZ coatings offer barrier protection not only on aluminum but also on two alloys. Comparable performance to coated aluminum is achieved for the 2024-T3 and 7075-T6 alloys coated with TMZ-2, TMZ-3 and, especially, TMZ-5 * coatings. It seems therefore that the effect of substrate is largely suppressed due to the good protection properties of the coatings, mainly based on homogeneous coverage of the metal surface and small porosity averaging only 2.3 × 10 −4 %. 32 Metal inclusions therefore do not affect significantly the effectiveness of the corrosion protection of the TMZ-1-TMZ-5 coatings. For the TMZ-6a coating, which is less stable upon polarization, the influence of substrate is more pronounced.
Electrochemical impedance measurements on aluminum.-EIS was used to study the time dependence of the coating protectiveness during long-term immersion of aluminum in the corrosive chloride medium. Coatings were selected to reflect different synthesis and curing conditions, i.e. composition, ageing time and curing temperature. The effectiveness of the coatings on aluminum substrate immersed in chloride solution was followed after 1 hour, and 1, 2, 3 and 7 days of immersion. Bode plots of the impedance, Z, and phase angle, φ, are presented for the coatings TMZ-3, TMZ-5 * and TMZ-6a * (Figs. 7, 8 and 9). At frequencies (f) between 10 5 and 10 1 Hz, log |Z| for the TMZ-3 coating increased linearly as f decreased, with a slope close to 1 (Fig. 7). The maximum value of φ of almost −90 • spanned over a broad range of frequencies indicates the highly capacitive character of the coating. In the range between 10 1 and 10 −2 Hz, the impedance reached a plateau. Within one week of immersion the magnitude of this plateau decreased from 10 7 to 2 × 10 6 cm 2 and φ shifted to higher frequencies. These changes indicate the progressive development of a corrosion process through the pores of the coating and at the coating/metal interface. The latter process may be related to the increase observed in the Bode φ vs. f curve in the range between 1 (β c , β a ), corrosion potential (E corr ), pitting potential (E pit ) and difference in potentials ( E = E pit -E corr ). Results are presented as mean value ± standard deviation. Composition of the coatings and conditions of ageing and curing are given in Table I Table I Figure 6. Potentiodynamic polarization curves for uncoated AA7075-T6, and for AA7075-T6 coated with TMZ-2 coating (aged 48 h aged and cured 1 h at 100 • C), TMZ-5 * coating (aged 48 h and cured 1 h at room temperature), and TMZ-6a * coating (aged 1 h and cured 1 h at room temperature). Samples were cured under daylight. dE/dt = 1 mV/s. and 10 -2 Hz. It is notable, however, that the extent of the impedance plateau remained unchanged, indicating that the corrosion process did not significantly affect the coating structure. Thus, despite signs of slow deterioration of the coating the impedance value still remained in the M cm 2 range after one week of immersion.  The general shape of the Bode plots of the coating TMZ-5 * was similar to that for the TMZ-3 coating (Fig. 8). However, after already 2 days, the Z values decreased below 10 6 cm 2 and decreased further. The φ shifted to higher frequencies and the shape of the φ vs. f curve changed, indicating the progressive deterioration of the coating. At short immersion time, i.e. after 1 h, the TMZ-6a * exhibited Z values higher by more than one order of magnitude ( Fig. 9) compared to TMZ-3 and TMZ-5 * coatings. The maximum φ of almost −90 • spanned over an even broader range of frequencies compared to the TMZ-3, up to 10 2 Hz. With increasing immersion time the value of impedance decreased and φ shifted to higher frequencies. Despite these changes, after 1 day of immersion was still more than 10 7 cm 2 , i.e. slightly higher than for the TMZ-3 coating (Fig. 7). After one day, however, Z dropped to values below 10 6 cm 2 , and the shape of the φ vs. f curve changed, indicating the beginning of the corrosion process. Therefore, despite extremely high values of impedance, the coating with high ZTP and MAA contents deteriorated faster than that with smaller ZTP and MAA contents. Typically, coatings with initial resistance above 10 8 cm 2 provide excellent corrosion protection, while those with resistance lower than 10 6 cm 2 provide poor corrosion protection. 37,38 The porosity of the coatings based on Si and Zr is very low 32 and its effect on the coating degradation with immersion time is relatively small. In addition to the coating porosity, another parameter involved in the degradation of the sol-gel coating may be the hydrolysis of the polysiloxane network. For example, coatings containing higher concentration of TEOS were hydrophilic and degraded faster due to the formation of Si(OH) 4 as a product of hydrolysis of SiO 2 . 39 In the present work this process has been minimized due to high degree of chelation and polymerization, as well as the hydrophobic character of the coating. 33,39 Schematic presentation of the network formation.-Optimal conditions are shown to exist that ensure the highest resistance towards both general and localized corrosion. While increased contents of ZTP and MAA and shorter ageing times (TMZ-6a) resulted in high stability to general corrosion, the coating, once polarized, was more susceptible to localized breakdown, regardless of the curing temperature (Fig. 4).

Table V. Electrochemical corrosion parameters measured in 0.5 M NaCl for uncoated AA7075-T6 and coated with TMZ coatings after 1 h stabilization at the open circuit potential: polarization resistance (R p ), corrosion current density (i corr ), cathodic and anodic Tafel slopes (β c , β a ), corrosion potential (E corr ), pitting potential (E pit ) and difference in potentials ( E = E pit -E corr ). Results are presented as mean value ± standard deviation. Composition of the coatings and conditions of ageing and curing are given in
Similar results are obtained by EIS measurements under open circuit conditions after only 2 days immersion ( Fig. 9). At smaller contents of ZTP and MAA and longer ageing time, the coating cured at room temperature (TMZ-5 * ) exhibited good stability under polarization (Figs. 3, 5 and 6) but deteriorated rapidly during long-term immersion (Fig. 8). The optimal conditions are achieved for coatings with Zr/Si ratio of 0.41, longer ageing time and curing at 100 • C (TMZ-3), both under polarization (Figs. 1 and 2) and long-term immersion conditions (Fig. 7).
The reason for this behavior may be found in the nature of the processes taking place at different temperatures. MAA is involved in three important steps: chelation of ZTP, polymerization of the acrylic group between MAA-MAA or MAA-MAPTMS and passivation of the metal surface by formation of the protective complex. 21,32 ZTP affects both inorganic network structure with Si−O−Si and Si−O−Zr bonds, as well as organic structure through coordination with methacrylic acid. Steps of hydrolysis, condensation and network formation including polymerization were explained and schematically presented in our previous study. 21 The scheme of formed network is taken herein as a starting point for explanation of role of zirconium in the process of radical polymerization (Fig. 10).
Molecular oxygen acts as an inhibitor of free-radical polymerization. It scavenges the radicals formed during the initiation of polymerization, leading to peroxyl radicals, ROO • , that are unreactive towards double bonds and thus inhibit the polymerization. Conventional routes to overcoming oxygen inhibition include the use of irradiation, inert gases or special oxygen scavengers. It was reported that ZTP constitutes an alternative to the conventional routes taken to overcome O 2  inhibition. 24,25,31 ZTP reacts with peroxyl radicals ROO • (Fig. 10) to give a pentavalent alkoxide (ROO • Zr(OR) 4 ). 24 In the succeeding step, ROO • are converted to propoxyl radical PrO • , which are the initiating species of radical polymerization. 31 The reason why the degree of protection varies with Zr/Si content and curing temperature lies in the variation of network formation. Coatings with large ZTP and MAA contents (TMZ-6a, TMZ-6a * ) undergo radical polymerization at room temperature, which occurs mainly through polymerization of acrylic group (Fig. 10). However, these coatings readily degrade once exposed to chloride solution. In contrast, coatings with smaller ZTP and MAA contents require higher curing temperatures (TMZ-3) to induce polymerization, which proceeds mainly by condensation where the formation of Si−O−Si and Si−O−Zr bonds are dominant. When coatings with smaller ZTP and MAA contents cured at room temperature (TMZ-5 * ) the polymerization of organic network again prevails over condensation of inorganic network leading to worse protective properties during longer immersion in chloride solution.
Salt spray testing on aluminum and alloys AA2024-T3 and AA7075-T6.-AA2024-T3 (Fig. 12) and AA7075-T6 (Fig. 13) alloys, bare and coated with TMZ-3 and TMZ-6a * , were compared with aluminum ( Fig. 11) after different periods of time in a salt spray chamber. The coatings TMZ-3 and TMZ-6a * are chosen as two extremes in curing and ageing conditions. Aluminum corroded very rapidly and, after only 16 h, corrosion products were clearly observed on the metal surface. The corrosion process continued, resulting in increasing amounts of corrosion products. On the other hand, aluminum coated with TMZ-3 coating was protected against corrosion, even after 21 days (Fig. 11). No corrosion products were observed, thus confirming good corrosion resistance of the sol-gel coating. The protection of aluminum by TMZ-6a * coating is less effective. The protection is stable up to 7 days but, at longer exposure times, corrosion products were formed on the metal substrate. Therefore, despite good electrochemical results (Table III), protection gradually decreased at longer exposure times, probably due to dissolution of the coating through the pores, as confirmed by EIS (Fig. 9). Nonetheless, after 7 days the surface appearance of the substrate coated with TMZ-6a * coating is comparable to the appearance of uncoated substrate after 16 h exposure, which accounts for the reasonable substrate protection.
AA2024-T3 is poorly resistant to chloride containing media and corroded fast in the salt spray chamber (Fig. 12). Corrosion products were observed after only a few hours and, after 16 h, covered almost the whole surface. In contrast, AA2024-T3 coated with TMZ-3 coating was protected, even after 21 days. Corrosion products were not observed, demonstrating that the sol-gel coating provided good corrosion protection in the salt spray environment. As for Al, the TMZ-6a * coating was less effective than the TMZ-3 coating.
AA7075-T6 alloy was even more susceptible to corrosion in chloride media than AA2024-T3 being visibly corroded after only 2 h, with corrosion progressing with time (Fig. 13). The TMZ-3 coating protected the substrate though showing some corrosion defects after 21 days. The TMZ-6a * coating was less effective, with corrosion products seen after only one day's exposure.

Conclusions
The corrosion protection of aluminum-based materials by the hybrid sol-gel coatings is affected by the composition of the sol, as well by the conditions of ageing and curing temperature. For constant conditions of 48 h ageing and curing at 100 • C, the greatest degree of protection was achieved at a Zr/Si ratio of 0.41 (TMZ-3).
When curing at 100 • C, ageing improved the corrosion protection but the effect was much more important for curing at room temperature. Under these conditions the best short-term results for Al and both alloys were obtained for coatings prepared from aged sols (TMZ-5 * ). However, coatings cured at room temperature cannot withstand longer immersion in chloride solution.
The addition of larger amounts of ZTP and MAA (TMZ-6a) is of interest due to the shortening of ageing time to only 1 h. However, the protective ability deteriorates rapidly, the coating showing poorer long-term performance than coatings prepared with smaller ZTP contents, longer ageing and cured at 100 • C. The presented results show that differences in the process of network formation have an important impact on corrosion properties. Condensation resulting in the formation of inorganic network of Si−O−Si and Si−O−Zr bonds and increased cross-linking and polymerization produced by MAA which affects the coating density are both essential for good shortand long-term protection. In terms of R p and i corr values, the degree of coating protection achieved by hybrid coatings on Al, AA2024-T3 and AA7075-T6 in 0.5 M NaCl is comparable to protection on AA7075-T6 in Harrison's solution, 32 even though that chloride solution is a more aggressive medium.
Salt spray testing showed aluminum and AA2024-T3, both coated with TMZ-3 coating, to be resistant to corrosion for 500 h operating service. When deposited on AA7075-T6 somewhat poorer results were obtained but still much better than for the uncoated alloy. Overall it has been demonstrated that the described hybrid coatings are applicable for the use of aluminum-based materials in chloride environments.