Mechanistic Studies of the Corrosion of 2024 Aluminum Alloy in Nitrate Solutions

The material used for the experiments was a 25 mm thick plate of 2024 T351 aluminum alloy. The T351 treatment consists in solution heat-treating at 500°C, water quenching, straining, and tempering the alloy at room temperature for 4 days. Its composition is given in Table I.

The electrochemical behavior of 2024 alloy in nitrate solutions with or without chlorides was studied by plotting potentiokinetic curves. The samples were mechanically polished to 4000 grit SiC paper, ultrasonically cleaned in distilled water, and air dried. The samples consisted of cylindrical rods embedded in an epoxy resin in order to control the sample area exposed to the solution and to mask the electrical connection. A sample area of 0.78 cm² was exposed to a solution in contact with air, at room temperature. A three-electrode electrochemical cell was used including a platinum grid with a large surface area as the auxiliary electrode, the reference electrode being a saturated calomel electrode (SCE) with a Luggin capillary. All potentials quoted are with respect to the SCE reference. Solutions were prepared by dissolving various amounts of with different amounts of NaCl in distilled water. All chemicals used were analytical reagent grade. The electrolyte compositions used in this investigation are given in Table II. The samples were immersed in the electrolyte and the potential was immediately scanned at a rate of 1 mV s⁻¹ from −1000 to 1000 mV/SCE.

Table II. 

Electrolyte compositions (pH 6.6) and weight loss for 2024 T351 alloy exposed to a nitrate solution with or without chlorides for 9 days.
Solution	(M)	NaCl (M)		(mg cm−2)
Distilled water				2.52
1	1			8.47
2	1		0
3				10.97
4			0	0
5
6			0	0
7			1
8			0	7.92
9			∞

Other samples were mechanically polished with 1 μm diamond paste, ultrasonically cleaned in distilled water and air dried. A series of samples was potentiokinetically polarized in nitrate solutions with or without chlorides from −1000 mV to −500, −300, −100, 0 or +150 mV/SCE, and then maintained for 5 min at this final potential prior to SEM, EDS, AFM, or SIMS examination. Another series of samples were potentiostatically polarized for an hour at −500, −300, or −100 mV/SCE in nitrate solutions and then chloride ions were added at a concentration of 0.004 M. The samples were maintained in the chloride-containing nitrate solution at the same potential for 2 min prior to surface analyses.

Corrosion rates were determined by conventional weight loss (WL) measurements. Alloy samples were cut into parallelepipeds with a total surface area exposed to the electrolyte of about 12 cm². All specimens were polished to 4000 grit SiC paper and then with 1 μm diamond paste, ultrasonically cleaned in distilled water, and air dried. Prior to exposure testing, specimens were weighed to the nearest 0.1 mg. They were exposed to the solutions described in Table II for different lengths of time. The volume of the solution was small (60 mL solution volume/sample in order to observe the variation of the chemical species more easily. Moreover, the tests were performed in small closed flasks to avoid the oxygen consumed from being replaced. Then, the samples were removed from the solution, immediately rinsed with distilled water, dried in an air stream, weighed, and optically observed. After the initial examination, the corrosion product layers were removed by exposure to 75% solution for 5 to 10 min in an ultrasonic bath. The specimens were then rinsed with distilled water, dried in an air stream, and reweighed; this procedure was repeated until no weight change was observed. The pH of the solution was measured at the end of the WL test. The solution was filtered and analyzed using UV spectroscopy. The solution ultraviolet absorbance was measured from 400 to 250 nm to detect the presence of and ions. The of the transition is, respectively, 300 and 353 nm for and in aqueous solution.

The corroded zones were observed with SEM using a LEO 435 VP apparatus and the compositional analysis of the intermetallic particles was determined with EDS by averaging the results obtained with about 20 particles. An accelerating voltage of 15 kV was used for both secondary electron imaging and EDS analysis. A few samples were examined by using an atomic force microscope to quantify the dissolution of the intermetallic particles. Other samples were analyzed by using SIMS (CAMECA IMS4F) in the profiling mode with an analyzed zone of 30 μm in diameter. The mapping mode was also used with an analyzed zone of ions were used for abrasion, and the intensity profiles plotted from the recombination of ions with the analyzed chemical elements in order to reduce the matrix effect, i.e., to obtain intensity profiles closer to the concentration profiles.

The potentiodynamic curves of Al 2024 in nitrate solutions are presented in Fig. 1. Whatever the nitrate concentration, the corrosion potential was around −150 mV/SCE. Near in the cathodic part, a current plateau was observed. It was attributed to oxygen reduction. For potentials lower than −400 mV/SCE, the current density increased with decreasing potential and depended on nitrate concentration. The higher the nitrate concentration, the higher the current density. This part of the cathodic domain was related to the reduction of nitrate ions to nitrite ions

In the anodic domain, Al 2024 exhibited stable passivity with a passive current density ranging from to for nitrate concentrations from to 1 M. A weak anodic peak was observed which could be related to the dissolution of a particular phase of the alloy. For the lowest nitrate concentration no passivity plateau was observed. The anodic branch revealed a two-step dissolution mechanism. The potentiodynamic curves thus showed that nitrate ions acted as efficient inhibitors for 2024 alloy only when they were present in solutions at a sufficiently high level.

Zoom In Zoom Out Reset image size

Weight loss (WL) data are summarized in Table II for specimens exposed to a nitrate solution with or without chlorides. No significant weight loss was measured for samples immersed in solutions 4 and 6 for 9 days. After immersion, the specimens only showed slight discoloration (black appearance). Comparison with the sample immersed in distilled water showed that nitrate ions acted as efficient inhibitors for 2024 alloy when they were present in solutions at high concentrations. On the contrary, when immersed in a nitrate solution (solution 8), 2024 alloy exhibited greater weight loss than that observed in distilled water. The severity of attack was exemplified by the presence of a voluminous deposit of corrosion products that surrounded the entire sample. Weight loss tests thus confirmed the results obtained with potentiodynamic scans, i.e., there was a critical nitrate concentration (between and for 2024 alloy protection. Under this critical concentration, nitrate ions did not act as inhibitors but, on the contrary, were aggressive towards 2024 alloy. Moreover, analysis of the solutions by UV spectroscopy after WL tests allowed the detection of nitrite ions in the nitrate solution which showed that the reduction of nitrate to nitrite ions occurred simultaneously with Al alloy dissolution. For and nitrate solution, no nitrate reduction was observed.

Comparison of Fig. 1 and 2 shows that, in chloride-containing nitrate solutions, the corrosion potential was shifted to around −200 mV/SCE whatever the nitrate concentration. This was related to the enhanced aggressiveness of the solutions due to the presence of chloride ions. No significant difference was observed in the cathodic domain between the solutions without chlorides and those with chlorides except for the solution for which the cathodic current density was increased in comparison with the 1 M solution. On the contrary, in the anodic domain, the behavior of 2024 alloy in chloride-containing nitrate solutions was found to be very different from that observed in nitrate-only solutions. For nitrate concentrations equal to or higher than a strong anodic peak with a current density ranging from to was observed before the passivity plateau (passivity current density ranging from to . It was related to the dissolution of a particular phase of the alloy. This phenomenon was also observed in nitrate-only solutions with a weak anodic peak, but it was increased here due to the synergy between chloride and nitrate ions. Moreover, WL tests (Table II) showed that the sample immersed in solution 3 exhibited severe degradation in comparison with that exposed to solution. After the test, the solution was found to contain nitrite ions and a strong smell of ammonia was also noticed which suggested cathodic reduction of to But it can be noted that, after this peak, stable passivity was obtained which again showed the inhibitory effect of nitrate ions. The sample exposed to solution 7 exhibited a behavior similar to that observed in solution with stable pitting appearing at a potential slightly higher than the corrosion potential. This showed once more that nitrate ions acted as efficient inhibitors for 2024 alloy only at sufficiently high concentrations.

Zoom In Zoom Out Reset image size

SEM observations were performed on samples potentiostatically polarized for 1 h at −300 mV/SCE or at −100 mV/SCE in a 0.1 M solution. Figure 3 shows a sample after polarization at −300 mV/SCE, but the same behavior was observed at −100 mV/SCE. Examination of Al-Cu-Mg particles (S-phase particles) revealed that two phenomena can be associated to their reactivity. First, these particles were found to be active with respect to the aluminum matrix since they were partially dissolved. They were surrounded by a black area identified by EDS analyses as a copper deposit. But local dissolution of the surrounding matrix was also observed. This last behavior, characteristic of a particle noble with respect to the aluminum matrix, can be related to the copper enrichment of the particle. EDS analyses showed that after polarization the copper content of the 20 particles analyzed was in the 60 to 88 wt % range with an average content of 70 wt %. The copper content of the particles measured on a just-polished sample was 38 wt % on an average (between 32 and 43 wt %). Al-Cu-Mg particles were initially active phases with respect to the aluminum matrix. In nitrate solution, they dissolved and copper enrichment occurred leading to an ennoblement of the particle. They then behaved as noble sites inducing the dissolution of the matrix at their periphery.

Zoom In Zoom Out Reset image size

Attention was also paid to the reactivity of Al-Cu-Mn-Fe particles. The reactivity of this type of particles was also the same at −300 and −100 mV/SCE. Figure 4 shows a sample polarized at −100 mV/SCE. Heterogeneous dissolution of this type of particle was clearly observed after polarization in 0.1 M nitrate solutions. This result was quite different from those obtained in sulfate solutions² in which Al-Cu-Mn-Fe particles were found not to be dissolved whatever the applied potential. However, this observation was similar to that of Schmutz et al.¹¹ in NaCl solutions. They explained the nonuniform dissolution of this type of particle by referring to their heterogeneous composition. It can also be observed in Fig. 4 that the film on and around the particles had cracked and disappeared. This was not observed for Al-Cu-Mg particles. It seemed that the film grown on Al-Cu-Mn-Fe particles was different and thicker which could explain that these particles are less reactive than the Al-Cu-Mg particles. SEM observations thus showed that nitrate ions acted as aggressive species towards intermetallic particles in 2024 alloy whereas no attack of the matrix was observed, explaining that measured weight losses were negligible.

Zoom In Zoom Out Reset image size

When nitrate and chloride ions were simultaneously present in solution, a more severe attack of the intermetallic particles was observed whereas the rest of the matrix always remained unattacked. For Al-Cu-Mg particles, a more significant copper enrichment occurred. The copper content could reach 80 wt % after polarization for 1 h at −300 mV/SCE in a 0.1 M solution, addition of 0.004 M chloride ions and maintaining for 2 min in this medium. For Al-Cu-Mn-Fe particles, stronger heterogeneous dissolution was observed on addition of chloride ions. Moreover, for a few Al-Cu-Mn-Fe particles, pitting of the matrix at the periphery of the particles was observed (Fig. 5) when the samples were polarized at low potential. The dissolution of the surrounding matrix, which had not been observed in nitrate solutions, was also explained for this type of particles by a copper enrichment. After polarization at −100 mV/SCE and the addition of chloride ions, the copper content of the Al-Cu-Mn-Fe particles reached 57 wt% whereas, on a noncorroded sample, it was of 27 wt %.

Zoom In Zoom Out Reset image size

Surface observations after potentiokinetic experiments were also performed for comparison with potentiostatic tests. An accurate observation of the whole surface of samples polarized from −1000 mV/SCE to different final potentials (between −500 and +150 mV/SCE) in chloride-containing nitrate solution showed that, whatever the final potential, pitting was always associated with the intermetallic particles (Fig. 6). As for potentiostatic tests, dissolution of the surrounding matrix around the particles was observed, but no pits were found on the matrix far from the intermetallic particles due to the inhibitive effect of nitrate ions towards aluminum. The higher the final potential the stronger the dissolution of the intermetallic particles. Moreover, EDS analyses indicated that, for final potentials between −500 and −200 mV/SCE, the copper content measured at the surface of Al-Cu-Mg particles was around 50 wt %. For final potentials higher than −150 mV/SCE, the copper content suddenly increased and became equal to 70-80 wt %. For Al-Cu-Mn-Fe particles, the copper content remained equal to 30 wt % for final potentials between −500 and −200 mV and suddenly increased to 50 wt % for final potentials higher than −150 mV/SCE. The results thus confirmed once more that nitrate ions are very effective inhibitors for the aluminum matrix, but have to be considered as aggressive species towards copper-rich particles. In the presence of chloride ions, there was a synergistic effect of nitrates and chlorides in the pitting of the intermetallic particles which explained the strong anodic peak observed on the potentiokinetic curves for potentials around 50 mV/SCE. For the solution (Fig. 2, curve a), it can be assumed that the beginning of the anodic domain (from −200 to −100 mV/SCE) was related to the dissolution of intermetallic particles and the second part (potentials higher than −100 mV/SCE) to the dissolution of the matrix since, for this concentration, nitrate ions did not protect the aluminum matrix. The anodic part of the curve was close to that observed in the absence of nitrates (Fig. 2, curve e).

Zoom In Zoom Out Reset image size

AFM analysis of samples polarized in chloride-containing nitrate solutions allowed the dissolution of intermetallics to be accurately quantified. Figure 7 shows an ex situ AFM topographic profile obtained for a 2024 alloy sample polarized from −1000 to +150 mV/SCE and maintained for 15 min at +150 mV/SCE in The topographic profile of the selected region in air before immersion showed that intermetallic particles, and mainly Al-Cu-Mn-Fe particles, protruded slightly from the surface because of their greater hardness and lower rate of polishing relative to the matrix which allowed the particles to be localized. After corrosion testing, heterogeneous dissolution of Al-Cu-Mn-Fe particles was observed since the center of the particle was found to be pitted which can be related, as mentioned above, to the heterogeneity of the particle composition. Assuming that no dissolution occurred on the edges of the particles, the hole in the center was 190 nm deep. No dissolution at the periphery of the particle was observed here. Al-Cu-Mg particle behavior was quite different. These particles corroded homogeneously and after the corrosion test, many of them were completely dissolved so that only a hole was observed at the end of the experiment at the locations where S-phase particles had been. For others, not only dissolution of the particle but also considerable dissolution around the edge of the particle was observed as seen by the deep trench in the topographical image (Fig. 7 and Fig. 8). Assuming that there was no dissolution of the matrix at a certain distance from the particle, the dissolution of the S-phase particles reached 320 nm deep (the depth resolution of the apparatus equals to 1 to 2 nm). These observations confirmed SEM results which had shown that S-phase particles were initially active phases with respect to the Al matrix but then behaved as noble sites due to copper enrichment.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Copper enrichment of S-phase particles after corrosion tests can be explained in two ways, selective dissolution of Al and Mg while Cu remains or dissolution of the particle followed by copper redeposition. For various authors,^{1 10 16} dealloying by preferential dissolution of aluminum and magnesium is responsible for copper enrichment on the surface of the intermetallic particles. The former showed that dealloying of Al-Cu-Mg particles produces Cu particles that are no longer attached to the metal surface which results in redistribution of Cu around the intermetallics. Schmutz and Frankel¹⁰ wrote that the simultaneous dissolution of both and ions followed by copper reduction seems unlikely at the corrosion potential. Recently, Suter et al.¹⁶ explained the presence of copper around inclusions by preferential dissolution of Al and Mg in the S-phase particles.

SIMS analyses were performed here to give additional results about intermetallic particle behavior. Figure 9 shows the SIMS profiles obtained for a 2024 alloy sample polarized for 1 h in a 0.1 M solution at −100 mV/SCE, just above Acquisition was performed in a 30 μm diam zone containing S-phase particles. At the initial time, the signal was high; it decreased very significantly during the 200 first seconds of sputtering. This response was related to the copper enrichment of the surface of the analyzed zone. Then, during 1000 further seconds of sputtering, the signal remained constant and then decreased again. This could be seen more easily on the signal. The Cu plateau was related to the sputtering of the S-phase particles. At the end, the matrix underneath was analyzed. The signal showed that the surface layer did not contain high amounts of Mg. Figure 10 shows Al, Mg, and Cu maps of another zone containing S-phase particles before and after sputtering. The sputtering conditions were chosen here to remove a very thin layer on the sample surface. Before sputtering, the Cu map shows a broad copper-enriched zone with several spots of higher copper content and low magnesium content. These spots are identified as sites of S-phase particles. After sputtering, the broad copper-enriched zone had disappeared and the Al-Cu-Mg particles appeared with clearer delimitation. The particle magnesium content is higher after sputtering than before. This confirmed the previous hypothesis, i.e., the formation of a copper layer on the surface of the particles and around them. When Al-Cu-Mg particles were polarized in nitrate solutions at low potentials, Al, Cu, and Mg dissolved but Cu redeposition occurred. The copper deposit covered the particle remnant, hiding the magnesium during analysis. When this copper layer was removed, magnesium from the particle remnant was detected. The homogeneous copper-enriched zone around the particles shows that copper redeposition occurred. The copper redeposition on and around the particles appeared to be very homogeneous, and it seems unlikely that this resulted from selective dissolution of Al and Mg followed by dissolution of detached copper clusters and copper redeposition. Moreover, observations by SEM and EDS analysis of 2024 samples polarized in nitrate solutions at higher potentials (+300 mV/ECS) showed that no copper enrichment of Al-Cu-Mg particles occurred. At high potentials, Al-Cu-Mg particles always dissolved. If dealloying was the mechanism responsible for copper redistribution, a copper enrichment at the surface of the particles would have to be observed also at high potentials because of the presence of not-yet detached copper clusters. This was not the case. It must be noted, that no copper redeposition can occur at the periphery of the particles at high potentials, and this is valid for both mechanisms (particle dissolution or particle dealloying).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

SIMS analysis of several zones containing Al-Cu-Mn-Fe particles was also performed. For only a few particles, a thin layer of copper was observed on and around the intermetallics, but this observation was not clear. This was mainly due to the difficulty to detect the presence of Al-Cu-Mn-Fe particles by SIMS. It was not possible to observe the Al-Cu-Mn-Fe particles without sputtering a thin layer on the sample surface. For Al-Cu-Mg particles, direct observation was possible. This suggested that the oxide film on the intermetallic surface was quite different and maybe thicker on Al-Cu-Mn-Fe particles compared with Al-Cu-Mg particles as suggested by SEM observations (Fig. 3 and 4). Exact determination of the structure and composition of these films will require more detailed surface analysis. This study was not performed here.

These results showed that nitrate ions acted as inhibitors towards the aluminum matrix, but they produced pitting phenomena on the intermetallic particles. In order to obtain more detailed explanations about the interaction between nitrate ions and intermetallic particles, WL tests were performed. Figure 11 shows WL and pH vs. immersion time for 2024 alloy immersed in a solution. The WL curve shows that there was an induction time for 2024 alloy degradation in chloride-containing nitrate solution since, up to 150 h of immersion, WL was very low. During this period, pH increased and stabilized at a value close to 9. Optical microscopy observations of the samples after WL tests showed that the attack was localized on the intermetallic particles during this period. Then, WL and pH increased rapidly and, at the end of the longer test (576 h), WL was about 40.5 mg cm⁻² and pH was stabilized at a value of 13.5. Analysis of the solutions by UV spectroscopy after corrosion tests (Fig. 12) showed that, from 168 h of immersion, nitrite ions became concentrated enough to be detected and, after 336 h, nitrate ions had almost disappeared. The increase of nitrite concentration and the detection of ammonia occurred simultaneously with the WL increase. It must be noticed that the same tests performed in chloride-free nitrate solution showed that WL was not significant and pH remained between 6 and 7 for 30 days of immersion. So, during the first period of immersion in chloride-containing solution, Al-Cu-Mg particles dissolved due to the synergetic effect of nitrate and chloride ions. As explained by McIntyre et al. ,²⁴ nitrate ions can act as cathodic depolarizers by virtue of their kinetically facile reduction. Cathodic reduction of serves to sustain the corrosion process and accelerate it in the presence of oxygen. Reduction of nitrate ions led to the formation of nitrite ions and thus the possible cathodic reactions are

Copper enrichment of the surface of Al-Cu-Mg particles could not occur since, as explained by Foley et al. ,²³ it can be assumed that Cu forms soluble complexes with the produced by nitrite reduction. This might explain the strong dissolution of the Al-Cu-Mg particles. The same type of phenomenon must be assumed for Al-Cu-Mn-Fe particles even though they were less corroded than Al-Cu-Mg particles. Successive reductions of to and then led to an alkalization of the solution at first localized around the intermetallics, which increased their degradation. After 150 h of immersion, the passive film on the whole surface sample was broken down by chloride ions and generalized corrosion occurred which can be seen by a rapid increase in weight loss. Observations by optical microscopy of samples immersed for more than 400 h showed that not only strong dissolution of the intermetallic particles but also generalized degradation of the matrix were observed. If we assume that the electrons produced by the generalized corrosion of aluminum matrix were consumed by reduction to and then the total reduction of nitrate contained in the 60 cm³ of solution would lead to a WL of 36.6 mg cm⁻² and a pH of 13.9. These values were close to those measured for the longer test which indicates that the reaction between aluminum and nitrate was complete.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Immersion tests in chloride-containing nitrite solutions showed that, for nitrite ions (Fig. 13), 2024 alloy degradation began after a longer induction time (about 250 h). However, after this induction time, the pH became higher than 9 and then, WL increased very rapidly. Optical microscopy observations showed that, after this induction time, strong dissolution of the intermetallic particles but also severe pitting of the matrix were observed. However, it can be noticed that a WL of only 7 mg cm⁻² was measured after 500 h of immersion. Complete reduction of nitrite would lead to a WL of 27.4 mg cm⁻². The WL due to nitrite ions was lower than that observed in the nitrate solution because of the localized nature of the attack. Potentiokinetic curves of 2024 alloy in solution (Fig. 14) show that the corrosion potential was about −200 mV/SCE, i.e., similar to the value measured in chloride-containing nitrate solutions but, beyond the corrosion potential, a passivity plateau and then stable pitting were observed. Several tests were performed which showed that, even though the width of the passivity plateau varied randomly, stable pitting was always observed. Scattering of pitting potentials is a common phenomenon. This suggested that the induction time of 250 h for WL increase was just a particular value and that the increase of the WL in nitrite solutions due to the pitting of the matrix (and to the dissolution of the intermetallics) might occur at very randomly scattered immersion times if we performed the experiments many times.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Comparison of Fig. 2, curve e, and Fig. 14 shows that nitrite ions inhibit aluminum matrix pitting but with a lower efficiency than nitrate ions.

Immersion tests were also performed in ammonia in order to obtain more details about the interaction between and the 2024 alloy. HCl solution was added in order to obtain a concentration equal to The initial pH was 8.4. Figure 15 shows that an induction time of 500 h occurred before WL increased rapidly to 1.1 mg cm⁻². Degradation of 2024 alloy in solution thus remained low. Observations by optical microscopy showed that, after the corrosion tests, the matrix at a certain distance from the intermetallic particles was not damaged at all. Al-Cu-Mn-Fe particles were slightly corroded. A stronger dissolution of Al-Cu-Mg was observed, but their degradation was not as significant as in nitrate or nitrite solutions. Moreover, it was noticed that the pH on the plateau, related to the dissolution of intermetallics in chloride-containing nitrate solutions, was about 8, i.e., the pH of the ammonia solution. This showed that the strong degradation of the intermetallics in nitrate solutions was related to the cathodic activity of nitrate ions at the periphery of the intermetallics. The pH of the solution was a significant value since it is well known that aluminum oxide is not stable at high pH.

Zoom In Zoom Out Reset image size

Nitrate ions, which are known to be efficient inhibitors for aluminum, favor the growth of a very protective oxide film on the aluminum matrix. On the contrary, it can be assumed that the oxide films developed on the intermetallic particles is not so protective. Moreover, it is known that nitrate ions are very aggressive species towards copper. Therefore, due to their high copper content and to their less protective oxide film, Al-Cu-Mg particles act as anodic sites and are preferentially dissolved, the cathodic reaction being the reduction of nitrate ions. Al, Cu, and Mg from the particles are dissolved and, when the applied potential is low enough, copper becomes redeposited on and around the particles. The particle with the copper layer becomes more noble in comparison with the adjacent matrix and thus induces the dissolution of the surrounding matrix. However, simultaneously, nitrate ions induce pitting of the copper enriched particles which go on dissolving. Therefore, Al-Cu-Mg particles act successively as anodic and cathodic sites until they are completely dissolved. This mechanism can be checked when the variation of the open-circuit potential (OCP) vs. immersion time for 2024 alloy exposed to a chloride-containing nitrate solution is plotted (Fig. 16). The experiment was performed many times. Two tests are reported in Fig. 16. During the first 100 min, the OCP was quite stable and equal to −200 mV/SCE, i.e., the value of the corrosion potential measured on the potentiokinetic curves. Then fluctuations of the OCP were observed with a potential amplitude up to 250 mV. These fluctuations were related to the activity of Al-Cu-Mg particles. These particles acted successively as anodic and cathodic sites. When the particles were dissolved, fluctuations were not observed any more which occurred for one test after about 600 min. For another test, more than 2000 min were necessary for the OCP to become stable again. This was related to the fact that not all the Al-Cu-Mg particles were reactive at the same time as reported by different authors.^{27 28} Moreover, pitting is a random phenomenon which may explain such a scattering of the results. It can be noticed that, during Al-Cu-Mg particle dissolution, reduction of led to the formation of and hindered copper redeposition and so the Al-Cu-Mg particles became more reactive and dissolved completely. Moreover, because of the nitrate reduction and the subsequent great alkalization of the solution, the passivity was not maintained and generalized corrosion was observed for long immersion times. In nitrite solution, the alkalization was lower due to a lower hydroxyl ion production and only local attack (pitting) of the matrix was observed.

Zoom In Zoom Out Reset image size