Zeolites as reservoirs for Ce(III) as passivating ions in anticorrosion paints

2.1.1 Minerals characterization

Two different samples of zeolitic rocks were used in this research. Their qualitative mineralogical composition was obtained by X-ray diffraction (XRD) once the rock was ground. After grinding, the minerals morphology was examined by scanning electron microscopy (SEM) and their surface composition was obtained by energy dispersive X-ray analysis (EDX). The ground natural minerals density was determined by pycnometric method according to ASTM G 153 standard specification. Zinc phosphate (PZ) powder was also observed by SEM, and its elemental composition, obtained by EDX. A FEI Quanta 200-s microscope and a RX microanalyzer with the EDX detector Apollo 40 were employed.

Two zeolites minerals were tested, and, for the sake of clarity, from now on, the two minerals will be referred to as Z1 and Z2 (Table 1).

2.1.2 Preparation of cerium-exchanged zeolites

The zeolitic rocks were ground to obtain fine-grain powder whose average particle size was <10 μm. After being ground and washed twice with distilled water (DW), they were placed for 1 h in a beaker with boiling 0.2 m HNO₃ to solubilize ferric compounds. The acid solution was added from time to time to keep the solution volume constant. The solids were separated from the supernatant by centrifugation at 2200×g for 10 min and washed with DW several times. Then, they were placed in a beaker with 2 m NaCH₃COO for 3 h under continuous stirring to put the zeolite back into the Na form. Afterwards, the ground minerals were separated by centrifugation and washed twice with DW. Finally, they were exchanged with Ce(III) ions by bringing them into contact for 24 h with a 1 m Ce(NO₃)₃ under constant stirring. At last, the exchanged minerals were separated by centrifugation, washed four times with DW, and dried in an oven at 90°C until constant weight. From now on, the exchanged zeolites will be labeled as ZCe.

A scheme of this procedure is presented in Figure 1.

Figure 1:

Scheme of the exchanged-zeolites preparation and testing.

2.1.3 Cation exchange capacity

In order to determine the minerals capability of being a reservoir for the passivating ions, the cation exchange capacity of the zeolites was measured. In this sense, sodium exchange capacity and that for cerium ions were determined for both zeolites.

The ground mineral was put into contact for 3 h with 2 m NaCH₃COO under continuous stirring. Then, the sorbed sodium was back extracted into solution with 100 ml of 1 m NH₄CH₃COO, per each gram of the corresponding mineral, by continuous stirring during 24 h. The solid was then separated by centrifugation, and Na ions were quantified in the supernatant by atomic absorption spectroscopy.

In order to determine the cerium exchange capacity of the zeolitic minerals, the sorbed ion was back extracted into solution with 100 ml of 1 m NH₄CH₃COO, from 1 g of the corresponding modified zeolites by continuous stirring for 24 h. The solid was then separated by centrifugation, and Ce was quantified in the supernatant by a gravimetric technique (Welcher, 1948). A 70-ml aliquot of the solution containing Ce(III) was acidified with 10 ml of 2 m acetic acid and treated with a slight excess of 3% weight/volume (w/v) 8-hydroxyquinoline (C₉H₇NO) in ethyl alcohol. Cerium “oxinate”, Ce(C₉H₆ON)₃, was then precipitated by adding 20 ml of 10% w/v ammonium hydroxide and heating the solution to boiling point. The purple-brown precipitate was separated by centrifugation at 2200×g for 5 min, washed with hot water, dried at 110°C, and weighed in an analytical balance (precision 0.1 mg). The cerium exchanged per gram of zeolite may be calculated from the amount of precipitated (W_solid) by the following equation:

Percentage of cerium ( % ) = 24.5 ∗ W solid 1 g

The control was a cerium solution containing 0.1 g of Ce(III) in 70 ml. In every case, determinations were done in triplicate.

2.1.4 Electrochemical tests

The protective behavior of Ce(III) was assessed by electrochemical techniques. The corrosion potential (Ecorr) evolution of commercial SAE 1010 steel electrodes, immersed in Ce(III) solutions, was measured as a function of time, with respect to the saturated calomel electrode (SCE) as reference. The cerium nitrate concentration varied from 1.15×10⁻⁵ to 4.6×10⁻⁴m. These solutions were employed to simulate the Ce(III) concentration leached from the exchanged zeolites in the supporting electrolyte (NaCl 0.025 m). For the sake of comparison, Ecorr of steel immersed in 0.025 m NaCl was also measured. Measurements were performed during 4 h and in duplicate. The same measurements were carried out using a suspension of the anticorrosion pigments.

Linear polarization tests were also done, employing the Ce(III) solutions whose concentration varied from 1.15×10⁻⁵ to 4.6×10⁻⁴m. The SCE was used as reference, and a platinum grid, as the counter electrode. Measurements were done after 2 and 4 h of immersion. Polarization resistance (Rp) and corrosion rate (Icorr) were calculated from these curves. The swept amplitude was ±20 mV from the open circuit potential, and the scan rate, 0.166 mV · s⁻¹. The exposed area was 0.28 cm². Measurements were carried out with the 273A EG&G PAR Potentiostat/Galvanostat in combination with model SOFTCORR 352 software (EG&G PAR). The same measurements were carried out using suspensions of the anticorrosion pigments. Values were compared with the Rp and Ic obtained for PZ and a control solution (without inhibitor).

The inhibitive efficiency (E) was calculated with the equation

E = I 0 − I I 0 × 100

I₀=Icorr for the blank; I=Icorr for the tested specimen.

Potentiodynamic scans were done, employing 0.5 m NaCl, SCE as reference and a platinum grid as the counter electrode and SAE 1010 steel as working electrode. Three different suspensions were tested: one containing zeolite 1 exchanged with cerium ions (Z1Ce), another containing cerium-exchanged zeolite 2 (Z2Ce), and the third one containing zinc phosphate. The swept amplitude was ±250 mV from the open circuit potential, and the scan rate was 0.2 mV · s⁻¹. The exposed area was 0.28 cm². Measurements were performed after 2 h of immersion.

Measurements were carried out with the 273A EG&G PAR Potentiostat/Galvanostat in combination with model SOFTCORR 352 software (EG&G PAR).

In polarization measurements, the concentration of the supporting electrolyte was increased to improve both the cell conductivity and the Ce(III) exchange from the zeolite.

In Table 1, the nomenclature of the samples is shown, while in Table 2, a summary of the electrochemical tests is presented.

2.1.5 Observation and characterization of the exposed surface

Steel panels in contact with the cerium nitrate solutions, the suspensions of the exchanged zeolites (ZCe), and the mixtures ZCe+PZ were observed by SEM after Ecorr measurements were finished.

3.1.1 Characterization and properties of the zeolitic minerals

The mineralogical composition of the zeolitic rocks obtained by XRD showed that Z1 contained a high proportion of mordenite (Na₈){Al₈Si₄₀O₉₆} · 24H₂O (>80%), while Z2 presented a mixture of crystalline phases being mordenite (50–80%) and heulandite, (Ca₄){Al₈Si₂₈O₇₂} · 24H₂O (30–50%), the major constituents. The natural minerals density was similar (Table 4).

SEM micrographs and EDX of both zeolites and zinc phosphate are shown in Figure 2. In every case, the particle size was <10 μm. This feature is important as the size of the pigments should be <20 μm to obtain a uniform dry paint film. The particle and particle size distribution of both zeolites did not differ significantly from those of PZ. The zeolitic minerals also contained iron compounds, which were eliminated by the nitric acid washing (del Amo et al., 2003).

Figure 2:

SEM micrographs (6000×) and EDX analysis of (A) zeolite 1, (B) zeolite 2, and (C) zinc phosphate powder.

The zeolites exchange capacity is shown in Table 4. The cation exchange capacity of both minerals towards Ce(III) was quite low and that of Z2 was four times higher than the capacity of Z1.

3.1.2 Electrochemical tests

At the beginning of the test period, no significant differences in steel Ecorr could be observed between the tested Ce(III) solutions, but after 50 min, differences became perceivable. The Ecorr of SAE 1010 steel immersed in Ce(III) solutions was displaced to values more positive than the blank. The shift towards more positive values depended on the Ce(III) concentration – Ecorr of the steel electrode immersed in the most diluted solution being the one with the most positive values. At the end of the test, steel immersed in Ce(III) solutions had corrosion potential values around −450 mV, while the blank fluctuated around −620 mV (Figure 3). These results showed that steel was less active in the presence of Ce(III) salts and that increasing Ce(III) concentration did not lead necessarily to an increase in Ecorr. This was confirmed by the determination of Rp and Icorr (Table 5). Ce(III) reduced steel corrosion rate. The anticorrosion efficiency may be higher than 98.5%. Cecílio and co-workers reported that cerium nitrate is preferable to cerium chloride because nitrate anion possess certain protective action (Cecílio et al., 2007).

Figure 3:

Corrosion potential of the samples immersed in 0.025 m NaCl at different concentrations of Ce(NO₃)₃.

In pigment suspensions, at the beginning of the exposure period, Ecorr was displaced to more positive values and dropped off as time elapsed (Figure 4). At the end of the immersion time, the steel Ecorr value was the most negative one and corresponded to steel undergoing corrosion. Instead, the steel Ecorr values in pigment suspension were at least 100 mV more positive because some kind of protection was achieved. Differences among the pigments mixtures were not significant (≤50 mV). The best inhibitive pigment mixture seemed to be ZCe+PZ.

Figure 4:

Corrosion potential of the samples immersed in 0.05 m NaCl with different anticorrosive pigment mixtures.

Ic and Rp values of steel in the pigments suspensions (Z, ZCe, and ZCe+PZ) were reported in Table 6. All the inhibitive pigments diminished corrosion rates when compared to the blank. The zeolites, by themselves, reduced Icorr according to their exchange capacity. Steel in the mixtures zeolite+zinc phosphate exhibited lower Ic values. Moreover, these values were similar to those obtained when PZ was used alone, thus indicating that PZ was responsible for this further reduction in Icorr.

The potentiodynamic scans showed that all the anticorrosion pigments diminished the anodic dissolution current in a similar way, with respect to the blank. Cathodic current was also diminished. These facts account for the inhibitive properties of the tested pigments (Figure 5).

Figure 5:

Potentiodynamic scans (Tafel mode) of the samples immersed in 0.5 m NaCl for 2 h with different anticorrosion pigments.

3.1.3 Observation and characterization of the exposed surface

The morphology of the protective layer was observed by SEM; after 24 h of immersion in the above-mentioned solutions, steel was covered with a uniform film.

Globular oxides were observed for the lowest Ce(III) concentrations (Figure 6A), and the protective film presented a tendency to form pits, particularly at higher concentrations of Ce(III) (Figure 6B). Pits may be covered by corrosion products, which had a cerium content of 12.7%. Some stick-like formations had grown on the protective layer when the Ce(III) concentration was 4.6×10⁻⁵m (Figure 6A). These sticks had 56.0% of Ce. The “over precipitation” of Ce-enriched particles has been reported elsewhere (Kesavan et al., 2012). Small amounts of Ce (2.5%) were also detected in certain globular formations (Figure 6B).

Figure 6:

SEM micrograph of the protective film formed on the steel panels in contact with cerium (III) nitrate solution (A) 4.6×10⁻⁵m and (B) 4.6×10⁻⁴m in 0.025 m NaCl.

Figures 7–11 show the morphology of the protective coating formed on steel in contact with suspensions of PZ and of the exchanged zeolites (ZCe) or the mixture ZCe+PZ.

Figure 7:

SEM micrograph and EDX of the protective film formed on the steel panels in contact with zinc phosphate suspensions in 0.05 m NaCl.

Figure 8:

SEM micrograph and EDX of the protective film formed on the steel panels in contact with Z1Ce suspensions in 0.05 m NaCl.

Figure 9:

SEM micrograph and EDX of the protective film formed on the steel panels in contact with Z1Ce+PZ suspensions in 0.05 m NaCl.

Figure 10:

SEM micrograph and EDX of the protective film formed on the steel panels in contact with Z2Ce suspensions in 0.05 m NaCl.

Figure 11:

SEM micrograph and EDX analysis of the protective film formed on the steel panels in contact with Z2Ce+PZ suspensions in 0.05 m NaCl.

In the case of PZ, a homogeneous film was formed on the metal substrate, basically composed by iron oxyhydroxides, some agglomerations grown on it. These agglomerations were composed of P, Fe, and Zn, and it is thought that they plug the pores of the film (Figure 7A and B).

A homogeneous thin film was also observed in the case of Z1Ce (Figure 8A). At higher magnification (12,000×), corrosion products accumulations may be appreciated, which followed almost straight lines. These products contained small amounts of Ce (0.9%) (Figure 8B). As reported previously in the literature (Zhu et al., 2013), Ce(III) is incorporated in the oxide film, thus improving its protective ability. A low amount of zeolite was also incorporated in the protective film, as it could be deduced from the presence of Si (0.6%). The presence of Ce in the protective film was interpreted to come from an ionic exchange between the ZCe and cations like Na(I) associated to the aggressive chloride; this process may be represented by the following equation:

Ce 3+ /3 Z + 3 Na 1+ ↔  3 Na 1+ /Z + Ce 3 +

As ZCe was put in contact with the aggressive electrolyte NaCl, penetrating the coating, Ce(III) was released from the zeolite particle. Other cations will provoke the same exchange.

The film formed on steel exposed to the Z1Ce+PZ suspension was homogeneous (Figure 9A). According to EDX spectrum, it was composed mainly by iron oxyhydroxides, as Fe and O were detected. The protective film also contained small amounts of zeolite (0.19% of Si and 0.18% of Al) and phosphorus (0.33% of P). The incorporation of phosphate to the protective film was considered beneficial (Kozlowski & Flis, 1991). This base film appeared to be collapsed in some areas, exhibiting pits covered with corrosion products (Figure 9B). EDX analysis done inside these pits showed that the corrosion products contained P, Cl, Si, Fe, Al, and, in some cases, Zn (Figure 9B). The presence of P and Zn in the pits was supposed to favor pit repassivation as a consequence of their inhibitive properties. Some zeolite particles seemed to be deposited on the film (Figure 9C).

The protective film formed on steel in Z2Ce and Z2Ce+PZ suspensions presented similar morphologies. A more or less uniform film with a granule-like appearance was formed onto the metallic surface in contact with the Z2Ce suspension (Figure 10A and B). At higher magnification (12,000×), a cracked film was observed under these agglomerations (Figure 10B). The elemental analysis of these agglomerations revealed that they were mostly composed by silicon and cerium in a higher proportion than in the case of Z1Ce (Figure 10B). These agglomerations may be the Z2Ce particles depositing onto the metallic substrate. A greater amount of cerium (26%) was detected in the film underneath, which seemed to be constituted by iron oxyhydroxides (Figure 10B).

Cerium was not detected in the film formed on the steel in contact with the mixture Z2Ce+PZ, which was composed mainly of iron and oxygen (Figure 11A). Some pits appeared on the surface; several of them were covered by products with similar composition to the zeolitic mineral (Figure 11B). On the other hand, some pits were not covered, and the composition on the edge consisted of iron, sodium, silicon, zinc, and oxygen. Inside the pit, high content of iron was found, indicating that it may go all through the film (Figure 11C).

3.2.1 Salt spray and humidity chambers

Results obtained in the salt spray chamber are shown in Table 7. According to the standard practice, the degree of rusting was evaluated using a 0–10 scale based on the percentage of visible surface rust. The rust distribution is classified as spot rust (S), general rust (G), pinpoint rust (P), or hybrid rust (H). Spot rust refers to rusting concentrated in a few localizes areas. General rust is concerned with rust spots of different sizes randomly distributed across the surface. Pinpoint rusted surfaces present rust distributed across the surface as very small individual specks.

Paint 2A, pigmented with Z2Ce, showed a disappointing protective behavior because it could not surpass 192 h of exposure. Corrosion spots appeared during the first week of testing; on the other hand, the other paints showed, at that moment, no rust. After 500 h of exposure, paints 1A and 2B had good qualification (rusting degrees 8P and 7G, respectively), while paint 1B was qualified with 5G, showing poorer protective performance. Paint 1 had the same anticorrosion behavior compared with the control paint; so, it is possible to totally replace PZ by Z1Ce. The use of Z2Ce alone led to poorer performance, but in combination with a reduced amount of PZ, it led to acceptable results. The employment of modified zeolites allowed partial or full replacement of PZ, which depended on the selected zeolite.

The standard test method for evaluating the degree of paints blistering refers to the size and the density of blisters on the painted surface. The evaluation was conducted by comparison with photographic standards. All paints blistered after 24 h of exposure in the humidity chamber except the control that showed no blisters (Table 8). The size and/or frequency of the blisters increased with time. Paints formulated with Z1Ce developed smaller blisters (qualification 6 after 168 h) compared with paints containing Z2Ce (qualified as 4 after 72 h). In every case, the blister density was medium. The development of blisters in the humidity chamber is expected when alkyd resins are used, but this feature does not constitute a problem because anticorrosion paints are top coated in service conditions. After 168 h of exposure, panels were taken out the chamber due the onset of rusting after blistering.

3.2.2 Corrosion potential

At the beginning of the immersion time, the corrosion potential of painted panels was displaced towards more positive values except for paint 2A. This sample showed the worst anticorrosion behavior because it exhibited the most negative values from the beginning of the test (Figure 12). This more noticeable displacement to positive values for paints 1A, 1B, and 2B may be attributed to the presence, in the paint formulation, of Z1Ce or to the combination of the exchanged zeolites with PZ. However, as time elapsed, Ecorr values shifted towards more negative values, and after 2 weeks, most paints exhibited Ecorr values below −500 mV. The control paint had an almost constant value, close to −500 mV, along the test period.

Figure 12:

Time dependence of the corrosion potential (Ec) of painted panels in NaCl 0.5 m.

3.2.3 Electrochemical impedance spectroscopy

EIS is a valuable technique for acquiring information about processes taking place at the coating/substrate interface (Scully & Hensley, 1994; Grundmeier et al., 2000; Sorensen et al., 2009). The analysis of Bode’s plots revealed a resistive-capacitive response of all tested paints; two of them are presented in Figure 13. However, the point of view adopted in this paper was that of Amirudin and Thierry (1995), i.e. the visual observation of the spectra could not indicate the exact number of time constants involved in the degradation of the organic coating subjected to a corrosive environment; in change, the number of these constants must be determined by data analysis by suitable procedures (Boukamp, 1989).

Figure 13:

Bode’s plots of selected painted panels in NaCl 0.5 m.

The equivalent circuit models used to fit experimental data were that shown in Figure 14, where Re is the electrolyte resistance, Rpo is the ionic resistance of the protective coating, Cc is the coating capacitance; Rt is the charge transfer resistance of the corrosion process, and Cdl is the double layer capacitance.

Figure 14:

Equivalent circuit to interpret impedance data of a metal/paint-with-defects systems.

Distortions observed in the resistive-capacitive contributions indicate a deviation from the theoretical models due to either lateral penetration of the electrolyte at the steel/paint interface (usually started at the base of intrinsic or artificial coating defects), underlying steel surface heterogeneity (topological, chemical composition, surface energy), or diffusional processes that could take place along the test. As all these factors cause the impedance/frequency relationship to be non-linear, they are taken into consideration by replacing the capacitive components (Ci) of the equivalent circuit transfer function by the corresponding constant phase element Qi (CPE), thus obtaining a better fit of data. The CPE is defined by the following equation:

Z = ( j ω ) − n Y 0

where

• Z=impedance of the CPE (Z=Z′+Z″) (Ω),

• j=imaginary number (j²=−1),

• ω=angular frequency (rad),

• n=CPE power n=α/(π/2) (dimensionless),

• α=constant phase angle of the CPE (rad), and

• Y₀=part of the CPE independent of the frequency (Ω⁻¹).

The accuracy of the fitting procedure was measured by the χ² parameter obtained from the difference between experimental and fitted data; the most probable circuit was selected providing that χ²<10⁻⁴.

In the present work, the fitting process was mainly performed using the phase constant element Q_i instead of the dielectric capacitance C_i. However, this last parameter was used in the plots in order to facilitate results visualization and interpretation. As an example, one of the fitted is presented in Figure 15.

Figure 15:

Fitting curve and experimental data of one of the painted panels.

The measured values of Rpo are shown in Figure 16. According to current literature (Grundmeier et al., 2000), the barrier effect is satisfactory when Rpo≥10⁸ Ω·cm² and a barrier residual effect is present if Rpo is between 10⁶ and 10⁸ Ω·cm². Paint C presented the highest values of Rpo, which oscillated around 10⁷ Ω·cm². The observed oscillations in all the measurements may be attributed to temporary pore plugging by corrosion products. The incorporation of the modified zeolites to the paints reduced the barrier effect of the coatings almost from the beginning of the test. No significant barrier effect could be observed in the paints pigmented with the modified zeolites. The capacitance of the different coatings varied concomitantly with Rpo. Capacitance values in Figure 17 revealed that all coatings, except the control one, are not intact coatings because Cc was higher than 1×10⁻⁹ F·cm⁻². In this sense, the control paint also showed signs of incipient deterioration.

Figure 16:

Time dependence of the electrolyte resistance (Rpo) of the painted panels systems in NaCl 0.5 m.

Figure 17:

Time dependence of the capacitance (Cc) of the painted panels systems in NaCl 0.5 m.

The Rt values of the control paint C oscillated, as a trend, around 10⁶ and 10⁷ Ω·cm². Instead, during the first 11 days of immersion, Rt of all zeolite pigmented paints showed strong fluctuations between 10⁴ and 10⁷ Ω·cm² while the Rt of paint 2A maintained below 10⁴ Ω·cm² (Figure 18). From this period on, Rt of all zeolite paints fluctuated between 10³ and 10⁴ Ω·cm². The protection afforded by the control paint relayed on its barrier properties in conjunction with the inhibitive action of the pigment, but in the other cases, the protection was derived from the inhibitive properties of the exchanged zeolites. This protection was enough to achieve a good behavior of the paints in the salt spray test.

Figure 18:

Time dependence of the charge transfer resistance (Rt) of the painted panels systems in NaCl 0.5 m.

The double layer capacitance values (Cdl) of the different coatings (Figure 19) showed an initial increase due to certain deterioration as water uptake by the film increased, followed by certain fluctuations of one or more orders of magnitude. The abnormal values observed for most paints (Cdl>10⁻⁶ F·cm⁻²) suggested that Cdl may be regarded as a pseudo-capacitance more than a true one.

Figure 19:

Time dependence of the of the double layer capacitance (Cdl) painted panels systems in NaCl 0.5 m.