Electrochemical approach to repassivation kinetics of Al alloys: gaining insight into environmentally assisted cracking

3.1 The width of the hysteresis loop and the inflection in the reverse curve of a PS

Figure 2 reports ΔE values for the investigated Al alloys, determined from PS in 0.6 m NaCl (pH 6.5) using different experimental conditions: as-received and polished surfaces, N₂ and air saturated solutions, and extent of promoted corrosion (i_rev of 1 and 5 mA/cm²) (Comotti et al., 2013; Trueba & Trasatti, 2010). For a given Al alloy, ΔE increases with i_rev regardless of the surface state condition and the presence of additional cathodic reactant, more importantly for Al 2024-T3 and Al 7075-T6. The comparison of different materials indicates similar ΔE values in the presence of dissolved oxygen for 1000, 6000, and 5000 series Al alloys, in agreement with the findings of Nilsen and Bardal (1977). ΔE is more sensitive to the experimental conditions for Al 2024-T3 and Al 7075-T6, conversely to the other series alloys, indicating some correlation with the susceptibility to localized corrosion. Nevertheless, the mere analysis of ΔE without considering the features of the reverse curve does not provide explanation of different behaviors, also as a function of the experimental conditions. The increase of this magnitude may be caused by the occurrence of an inflection during polarization into the active region, with the consequent shift of E_prot to more negative values (Figure 1) (Comotti et al., 2013; Trueba & Trasatti, 2010).

Figure 2:

ΔE values for different Al alloys determined from PS (10 mV/min) in 0.6 m NaCl (pH 6.5) using different experimental conditions: as-received and polished surfaces, N₂ and air saturated solutions, and extent of promoted corrosion (i_rev of 1–5 mA/cm²).

Figure 3A–C shows the average PS of polished Al 6083-T6, Al 2024-T3, and Al 7075-T6, obtained in 0.6 m NaCl (pH 6.5) using different current density limits for scan reversal (i_rev between 0.1 and 5 mA/cm²) (Comotti et al., 2013). Note the similarity of the forward curves, characterized by a rapid increase of the current at E_pit. The inflection, instead, is more or less pronounced depending on Al alloy and on the extent of promoted corrosion (i_rev). Al 7075-T6 shows two breakdown potentials (E_pit1≅E_corr and E_pit2>E_pit1), typically reported for the peak-aged condition of this alloy using polished surfaces (Zhao & Frankel, 2007a,b) as well as a second inflection at more negative potentials for i_rev≥3.25 mA/cm². The variation of i_ptp with i_rev seems to be more important than that of E_ptp, whereas this potential is always close to E_pit, as reported in other works (Moore et al., 2008; Yasuda et al., 1990). For commercially pure Al 1050, the inflection was always indistinguishable, regardless of the surface state condition (polished and covered with the native Al oxide layer) and the temper (O and H24), although the morphology of attack consistently differs (Figure 4A–C). Crystallographic pitting prevails for polished surfaces, whereas hemispherical etched pits at the metal/film interface were promoted for as-received Al 1050, respectively. For the other alloys, SEM analysis showed constricted localized attack other than pitting (Figure 4D–K). In the case of Al 6082-T6, IG corroded paths in the alloy bulk beneath a corrosion film are promoted with i_rev=1 mA/cm² (Figure 4D) because of the selective dissolution of unbalanced Mg₂Si (Comotti et al., 2013) along grain boundaries, leaving behind discrete silicon-rich particles (Figure 4E). This may explain the knife-edge type of IG attack for higher i_rev (Figure 4D, inset), suggesting environmentally induced IGSCC due to the weakening of the grain boundaries as Si-rich particles formed. In the case of high-strength Al 2024-T3 and Al 7075-T6, transitions from IGC to microstructural pitting (Figure 4G–I) and to brittle fracture (Figure 4J and K), respectively, are promoted with i_rev. For the latter substrate, the transition correlates with the occurrence of the second inflection for i_rev≥3.25 mA/cm² (Figure 3C) (Comotti et al., 2013). Fissures along grain boundaries indicate TGSCC induced by stress concentration at the tip of narrow IG corroded paths that acts to pull the metal apart. Increasing stress concentration with fissures growth promotes plastic deformation at the tip bottom and brittle failure.

Figure 3:

Average PS (10 mV/min) as a function of i_rev for (A) polished Al 6082-T6, (B) polished Al 2024-T3, and (C, D) polished and as-received Al 7075-T6, in air-saturated 0.6 m NaCl (pH 6.5). The reverse curves are indicated with arrows. Numbers correspond to the anodic current limit (i_rev) in mA/cm² (Comotti et al., 2013, reproduced with permission of John Wiley & Sons).

Figure 4:

Corrosion morphology of Al alloys after PS (10 mV/min) in air-saturated 0.6 m NaCl (pH 6.5): (A, B) Al 1050-H24 (polished), i_rev of 4 and 2.5 mA/cm², respectively; (C) Al 1050-H24 (as-received), i_rev=5 mA/cm²; (D–F) Al 6082-T6 (polished), i_rev of 1, 2.5, and 5 mA/cm²; (G–I) Al 2024-T3 (polished), i_rev of 1, 2.5, and 5 mA/cm²; (J, K) Al 7075-T6 (polished), i_rev of 2.5 and 5 mA/cm²; (I) Al 7075-T6 (as-received) i_rev of 5 mA/cm² (Comotti et al., 2013, reproduced with permission of John Wiley & Sons).

The above-mentioned analyses indicate that the inflection is not only determined by the degree of constriction of pit cavities. Interdependent microstructural and local environment interfacial interactions play a more important role. This consideration is supported by the similar features of the reverse curves of polished and as-received Al 7075-T6 (Figure 3C and D). Although the latter substrate shows only one breakdown in the forward curve, i.e. in the presence of the native Al oxide layer, the potential values at the inflections are closely equal, and fissures along grain boundaries are promoted with i_rev (Figure 4L). Furthermore, for as-received 5000 series Al alloys, Al 5754-H111 showed a marked inflection in both N₂ and air-saturated solutions, different from Al 5083-H111 with indistinguishable inflection in the presence of dissolved oxygen (Trueba & Trasatti, 2010). Corrosion morphology analysis indicated that IGSCC is related to the β-phase (Al₃Mg₂) at grain boundaries, which acts as a catalyst for crack growth through anodic dissolution, reduction, and ingress pathway for H atoms (Jones, Baer, Danielson, & Vetrano, 2001; Oguocha, Adigun & Yannacopoulos, 2008). β-Phase does not need to be continuous to provide a path for crack propagation, which is the case of Al 5754-H111 presenting fine-grained structure with more discontinuous β-phase (Al₃Mg₂) precipitates along gain boundaries as compared to Al 5083-H111 (Figure 5A and B) (Trueba & Trasatti, 2010). For this alloy, oxygen-induced alkaline corrosion promotes crevice-like pits and also the precipitation of discrete rodlike β-phase particulate (Figure 5C) (Lyndon et al., 2013; Scamans, Holroyd, & Tuck, 1987), which is beneficial for hydrogen discharge.

Figure 5:

Microstructure and corrosion morphology of 5000 series Al alloys. (A, B) Images at the metallographic microscope of as-polished Al 5754-H111 and Al 5083-H111 surfaces after etching with 10% H₃PO₄ at 60°C; (C) SEM image showing discrete rodlike β-phase particulate (Al₃Mg₂) on the surface of Al 5083-H111 after PS in air-saturated 0.6 m NaCl (pH 6.5), using as-received specimens and i_rev=5 mA/cm², inset – cross-sectional SEM image showing a crevice-like pit (Trueba & Trasatti, 2010, reproduced with permission of Elsevier).

3.2 Effect of corrosion extent

From the characteristic electrochemical parameters of the PS (Figure 3), empirical relationships as a function of corrosion extent (i_rev) were determined with statistical significance (Comotti et al., 2013). Concerning the characteristic potentials E_pit, E_ptp, and E_prot, Figure 6A shows an i_rev-independent E_ptp in the case of Al 6082-T6 that differs from E_pit by approximately 35 mV, whereas E_prot decreases linearly with i_rev with 27 V/Acm^-2 slope. Conversely, E_ptp shifts linearly with 5 V/Acm^-2 slope toward E_pit, and E_prot decreases exponentially with i_rev in Al 2024-T3 (Figure 6B). A linear increase of E_ptp1 with similar slope (8 V/Acm^-2), in between E_pit1 and E_pit2, and an exponential decay of E_prot with i_rev are obtained also for Al 7075-T6 (Figure 6C). The significant variation of E_prot between i_rev of 2.5 and 3.25 mA/cm² is due to the occurrence of the second inflection at E_ptp2 (Figure 3C), which is however i_rev independent (914±1 mV) (Comotti et al., 2013). A similar result was obtained for as-received Al 7075-T6. In addition, reverse curves recorded for Al 7075-T6 after potentiostatic prepolarization at -700, -600, and -500 mV (versus SCE) (Figure 7A) resemble closely those shown in (Figure 3C). E_ptp shifts linearly to less negative potentials as a function of i_rev with similar slope values (i_rev is taken as the current density value at the beginning of potentiodynamic polarization) (Figure 7A, inset) (Melilli, 2013). Because the second transition becomes important with brittle fracture, additional resistance due to proton reduction in occluded cavities at very negative potential is more likely. The closely equal behavior of Al 7075-T6, as well as the small difference of E_ptp values (≅10 mV) for Al 6082-T6 and Al 2024-T3, regardless of the surface state condition, provides the confirmation of E_ptp as a characteristic of Al alloy matrix and not as the conventional repassivation potential (Little et al., 2007; Moore et al., 2008). Surface state condition principally influences the inflection position along the current density axis. More importantly, the variation of E_ptp with i_rev for Al 2024-T3 and Al 7075-T6, in contrast to Al 6082-T6, supports Cu depletion controlling crack tip electrochemistry as opposed to an active precipitate phase as the critical feature (Figure 4) (Knight, Birbilis, Muddle, Trueman, & Lynch, 2010; Wall & Stoner, 1997).

Figure 6:

Relationships between average characteristic potentials and i_rev for polished substrates of: (A) Al 6082-T6, (B) Al 2024-T3, and (C) Al 7075-T6; determined from PS (10 mV/min) in stagnant, air-saturated 0.6 m NaCl (pH 6.5) (Comotti et al., 2013, reproduced with permission of John Wiley & Sons).

Figure 7:

Effect of potentiostatic polarization and scan rate on the inflection during potential scan into the active region. (A) Reverse curves (10 mV/min) in air-saturated 0.6 m NaCl (pH 6.5) of polished Al 7075-T6 after potentiostatic prepolarization at E≥E_pit2 in the test solution. Inset – relationships between E_ptp and i_rev (≅current at the beginning of the potentiodyamic polarization into the active region). (B) PS (i_rev=2.5 mA/cm²) of polished Al 2024-T3 recorded at different scan rates in stagnant, air-saturated 0.1 m NaCl (pH 6.5).

The kinetics of repassivation are evaluated from semilogarithmic relationships between the steepness below the inflection and i_ptp (Figure 8). The steepness increases linearly for Al 6082-T6 and Al 2024-T3. However, a slope change is evident for the latter alloy for i_rev≥2.5 mA/cm². In the case of Al 7075-T6, the steepness reaches a threshold value (≅200 mV/dec). The plateau-like response correlates with stress-independent corrosion due to stress relaxation by brittle fracture (Holroyd, 1990), which was produced at higher i_rev (Figure 4K). A constant steepness of approximately 200 mV/dec with fissure growth (i_rev≥2.5 mA/cm²) was obtained also for as-received Al 7075-T6 (Comotti et al., 2013) (Figure 4L). By extrapolation to 0 mV/dec of steepness, an indistinguishable transition with E_ptp≅E_prot (Figure 1) is indicated for i_ptp≅10^-7–10^-6 A/cm² (arrows in Figure 8), wherein these values are typical of the passive region. It must be pointed out that the semilogarithmic plots of the steepness as a function of i_rev are meaningless because extrapolation to i_rev=0 gives either positive or negative steepness with no physical meaning (Comotti et al., 2013). The similarity of the steepness variation with log i_ptp to that of crack growth rate as a function of stress intensity factor for Al 6151-T6 (Li, Wang, Ren, Chen, & Mu, 2012), Al 2024-T3 (Henaff, Menan, & Odemer, 2010), and Al 7075-T6 (Holroyd, 1990) in 3.5% NaCl is noteworthy. The Tafel-type behavior of the polarization response below E_ptp can thus be associated with a periodic active/passive sequence involving dissolution/passivation/film rupture because of the local action of the corrosive medium with possible competition of hydrogen production/uptake. Differently from Al 6082-T6 and Al 7075-T6, as well as Al 5083-H111, the PS response of Al 2024-T3 is more sensitive to experimental conditions when the amount of promoted corrosion is small (Figure 2) (Comotti et al., 2013; Trueba & Trasatti, 2010). In addition, some effect of local deformation on localized attack propagation was indicated by experiments conducted using different scan rates (Henaff et al., 2010; Melilli, 2013). In particular, the inflection is more pronounced as the scan rate increases, more importantly in diluted NaCl (Figure 7B), and contrary to the behavior of Al 7075-T6 and Al 5083-H111. The effect of the scan rate could be related to the selective dissolution of S-phase (Al₂CuMg) precipitates, leading to the formation of Cu-rich layer. If dealloying of S-phase is significant, local cathodic corrosion of the matrix would be enhanced because of Cu enrichment, thus encouraging the formation of solid corrosion products at the bottom of cavities. A peak in the current in the reverse curves of Al 6082-T6 in alkaline NaCl (pH 11.5) under continuous bubbling of O₂ showed a correlation with the onset of crevice corrosion by O₂ depletion after cavity repassivation (Cicolin et al., 2014).

Figure 8:

Relationships between average steepness and log i_ptp for polished Al 6082-T6, Al 2024-T3, and Al 7075-T6; determined from PS (10 mV/min) in stagnant, air-saturated 0.6 m NaCl (pH 6.5) using different i_rev (Figure 3) (Comotti et al., 2013, reproduced with permission of John Wiley & Sons).

3.3 Effect of test solution composition

From the systematic studies of polished Al 6082-T6 repassivation behavior (i_rev=2.5 mA/cm²) as a function of chloride concentration [Cl^-] and solution pH (2–11.5), as well as under continuous bubbling of oxygen (Cicolin et al., 2014), E_ptp is pH independent and tends to E_pit by decreasing solution pH and [Cl^-] concentration (Figure 9), which correspond closely to acidified pit-like solution. Although E_pit shifts to more positive values with log [OH^-], the variation is negligible as compared to that reported for iron with pH>10 (slope of 18 mV) because of the external pH contribution buffered by the various hydrolysis steps of Al³⁺ (Galvele, 1976, 1981). Similarly, i_ptp changes mainly with [Cl^-] (Figure 10A), but values are much higher in 0.1 m NaCl for pH between 4 and 9, corresponding to the pH range of aluminum hydroxide stability. Because the critical pit chemistry is always concentrated, the mass flux out of the pit is promoted in concomitance with cavities opening toward the outer surface by increasing [Cl^-]. Thus, the critical concentration for sustaining occluded chemistry decreases, and the dilution of pit cavity solution is more favored (Cicolin et al., 2014). According to these results, E_ptp can be considered as the thermodynamic driving force of Al dissolution on freshly created surface rather than on initially passivated surface (E_pit). Similarly, i_ptp can be considered proportional to the rate at which hydrolysis equilibrium is reached, represented by the following general equation to account for the contribution of the OH^- at pH higher than 7 (Galvele, 1981):

Figure 9:

Relationships between average E_pit and log [OH^-] for polished Al 6082-T6 determined from PS in stagnant, air-saturated NaCl solutions with different concentrations (A–C). Average E_ptp values are included for clarity (see the text) (Cicolin et al., 2014, reproduced with permission of Elsevier).

Figure 10:

Relationships between (A) i_ptp and log [OH^-] for different [Cl^-] and (B) log i_ptp and log [Cl^-] for different pH levels, determined from PS (10 mV/min) in stagnant, air-saturated solutions for polished Al 6082-T6 (Cicolin et al., 2014, reproduced with permission of Elsevier).

In the presence of Cl^-, the reaction of H₂O with AlCl²⁺ and Al(OH)Cl⁺ intermediates produces relatively stable Al(OH)₂Cl, which is the initial solid phase leading to the formation of mononuclear and polynuclear species by Al³⁺ hydrolysis (Galvele, 1976, 1981; Guseva et al., 2009). The nature of the prevailing species depends on Al³⁺ concentration and solution pH.

Linear relationships with negative slopes for log (i_ptp) as a function of log [Cl^-] at all pH are shown in Figure 10B. The absolute slope values are similar to reaction orders (n) with respect to Cl^- for pit initiation of Al alloys in 0.1–3 m KCl solutions at different pH levels (n between 1 and 2), which were determined from logarithmic relationships between the reciprocal of the induction time (rate of pit initiation) and [Cl^-] (Natishan & O’Grady, 2014). Such reaction orders suggest transitory stable covalent compounds like Al(OH)₂Cl and Al(OH)Cl₂, as in diluted acidified pit-like solution. The presence of these species weakens the passive film and promotes further pit nucleation or localized attack reactivation at the transition onset.

The effect of [Cl^-] on the repassivation behavior (i_rev=2.5 mA/cm²) of polished Al 5754-H111, Al 5083-H111, Al 2024-T3, and Al 7075-T6 is summarized in Figure 11 (Melilli, 2013; Melilli et al., 2014). Points missing in the figure correspond to conditions with indistinguishable inflection in the reverse curve. Al 7075-T6 substrates showed two breakdown potentials and only one inflection at potentials close to E_pit1 (Figure 3C, i_rev=2.5 mA/cm²), independent of the [Cl^-]. In the case of Mg-rich Al 5754-H111 and Al 5083-H111, two inflections were detected in the reverse curves (Melilli, 2013), the second inflection being more marked. For both alloys, the first transition at E_ptp1≅E_pit and the second one at E_ptp2 just below E_pit shifted in concomitance with E_pit to less negative values by increasing [Cl^-] (Figure 11A). The same result was obtained for E_ptp and E_pit of Al 2024-T3 and Al-7075-T6 (Figure 11A), as well as for Al 6082-T6 (Cicolin et al., 2014). Although the corrosion morphology evolution with [Cl^-] of the former alloys was similar to that shown in Figure 4G–K, transitions from IGC to crystallographic pitting and to crevice corrosion, respectively, were obtained with Mg-rich alloys (Melilli, 2013). Pitting corrosion was promoted with [Cl^-] in the case of Al 6082-T6 as well (Cicolin et al., 2014). The quite similar electrochemical behavior of 5000 and 6000 series alloys is indicated further by the comparison of E_ptp and E_pit (Cicolin et al., 2014; Melilli, 2013). E_ptp2 is shifted below the corresponding E_pit by 30–40 mV (Figure 11A). In addition, E_pit and E_ptp2 decrease linearly with log [Cl^-] with slope values between 60 and 120 mV/dec. The higher absolute slope values (≅200–300 mV/dec) estimated for Al 2024 and Al 7075 (Figure 11A) indicate more important changes in local surface and solution compositions because of microstructural evolution with corrosion (Knight et al., 2010; Wall & Stoner, 1997). Notice that E_ptp tends to differ from E_pit with the dilution of the test solution.

Figure 11:

Empirical relationships between (A, B) electrochemical parameters (see Figure 1 and text for clarity) and log [Cl^-]; (C) steepness and log [Cl^-]; (D) steepness and log i_ptp (labels near the symbols indicate the corresponding [Cl^-]), determined from PS (10 mV/min) in stagnant, air-saturated solutions (pH 6.5) for polished Al 5754-H111, 5083-H111, Al 2024-T3, and Al 7075-T6. The values near the lines in panels A–C indicate to slopes values.

From the relationships between log i_ptp and log [Cl^-] (Figure 11B), absolute slope values close to 1 are obtained for the second transition of Mg-rich alloys, indicating that mononuclear Al(OH)₂Cl is involved. Values <1 for the first transition of Al 5083-H111 suggest the participation of less stable complex ions such as AlCl²⁺ and Al(OH)Cl⁺, which is supported by the indistinguishable inflection in 0.1 m NaCl (Figure 11B). Similarly, the [Cl^-]-independent i_ptp for Al 5754 (n≅0) suggests the participation of loosely held chemisorbed species. These differences may be related to the alloy microstructural features and the type of corrosion attack (Figure 5). Crevice corrosion was promoted with [Cl^-] for Al 5083-H111 and crystallographic pitting for Al 5754-H111 (Melilli, 2013). Higher i_ptp values and fractional reaction orders (n between 0.4 and 0.7) are obtained for Al 2024-T3 and Al 7075-T6 (Figure 11B). Fractional n indicates chain reactions or complex reaction mechanisms because reaction orders higher that 2 should be obtained with the present test conditions if more stable hydrolysis products are formed. Accordingly, consecutive reactions involving consumption/generation of H⁺ in the stepwise propagation of localized attack are more likely. This is not excluded but contributes in a lesser extent in the case of Mg-rich alloys.

The plots of the steepness as a function of log [Cl^-] (Figure 11C) show similar slope values for Al 5754-H111, although the steepness after the first transition is less marked. This suggests similar reaction processes are involved but that concentration polarization occurs at later stages of alloy repassivation as a consequence of the onset of corrosion front propagation on a narrower front at less negative potentials (E≅E_ptp1). Considering crystallographic pitting promoted with [Cl^-], the linear decrease of the steepness with log [Cl^-] is consistent with repassivation by the dilution of occluded cavity solution (Cicolin et al., 2014). For Al 5083-H111 (Figure 11C), steepness versus log[Cl^-] plot for the second inflection shows a change of the slope for [Cl^-]≥0.3 m, which suggests a change in the repassivation mechanism because of the onset of crevice corrosion as salt precipitation at the bottom of damaged sites occurs with [Cl^-] (Melilli et al., 2014). A similar result is obtained for Al 7075-T6 (Figure 11C), but the steepness is even more important and increases more markedly. These differences show a correlation with fractional n values (Figure 11B) and brittle fracture induced by [Cl^-] (Melilli et al., 2014), in support of proton reduction controlling the anodic dissolution of Al. Accordingly, the magnitude of the steepness can be considered proportional to local acidity removal for full hydrolysis to be reached at E_prot. If this occurs, the steepness tends to decrease while i_ptp increases. Notice that both electrochemical properties increase with i_rev but for constant [Cl^-] (Figure 6) (Comotti et al., 2013). Notice also that, for the first inflection of Al 5754-H111, the steepness changes more importantly than i_ptp with [Cl^-], in correlation with the activity of small discontinuous β-phase precipitates along grain boundaries (Figure 5). Thus, i_ptp is influenced by the corroded area, while the steepness is principally determined by the kinetics of dissolution processes after the transition onset at E_ptp. Although the previous analyses cannot be done for Al 2024-T3 with the present test conditions due to indistinguishable inflection in 0.1 m NaCl (Figure 11C), high steepness values ([Cl^-]≥0.3 m) suggest important dissolution kinetics after the transition onset.

Additional experiments conducted with specimens positioned on one side of the O-ring cell showed less marked inflection, indicating the influence of mass transfer between the occluded cavity and the outer bulk solutions. Thus, gravity assists in the formation of metastable salt layers regulating the transport of ions in and out of occluded cells, i.e. with face-up specimens. On the basis of the reaction orders with respect to [Cl^-] during repassivation (Figure 11B), the previous observation correlates with the formation of metastable Al hydroxychloride compounds, although local cell reactions depend also on the alloy composition as well as on microstructural evolution during corrosion. For vertically positioned specimens, the inflection was almost indistinguishable for Al 2024-T3 and Al 7075-T6, but the width of the hysteresis loop was still important (Melilli, 2013). In the case of Mg-rich alloys, a distinguishable inflection (between E_ptp1 and E_ptp2 of faced-up positioned specimens, Figure 11A) was still detected.

Table 2 reports the results of the effect of [Cl^-] (0.1 and 0.6 m, pH 6.5), working area (1 and 0.1 cm²), and scan rate (ν of 0.2 and 10 mV/s), and of their interactions, on the electrochemical parameters in experiments using faced-up positioned specimens and a statistical analysis (95% confidence level). The p-values (significant if p<0.05), which enabled the estimation of the effect of the above-mentioned factors on the electrochemical properties, were determined by multifactor ANOVA. The regression coefficients, R², that provide an estimation of the linear model that includes the variation of the experimental data as a function of all the factors investigated were determined by linear multiple regression (LMR). A quantitative evaluation of multifactor effects generated by experimental observations indicates a significant effect of [Cl^-] on E_pit and E_ptp (p=0.000). Conversely, the characteristic parameters of the kinetics of repassivation such as i_ptp and steepness are not influenced by [Cl^-] (p≥0.05). In addition, the steepness does not depend on the working surface area. This further supports the view that i_ptp is related to the critical Al³⁺ concentration at the transition onset. Similarly, the scan rate has a major effect on the steepness, as compared to i_ptp, which provides the confirmation of electrochemical events at the bottom of the occluded cell governing the sequence of active tip-passive walls and consequent microstructural evolution. It is interesting to note the similar effect of the working area on E_ptp for Al 5083-H111 and Al 7075-T6, both prone to the formation of surface corrosion films. The analysis of R² values indicates less important nonlinear, and thus nonadditive, effects of the investigated factors in the case of Al 5754-H111, which exhibits crystallographic pitting as the main type of localized attack. A more significant nonlinearity is estimated for Al alloys prone to geometrically constricted localized attack other than pitting.

3.4 Effect of age hardening

To evaluate the effect of microstructure evolution induced with aging on the repassivation kinetics Al 2024 and Al 7075-based alloys, different thermal treatments were followed using Al 2024-T3 and Al 7075-T6 as starting materials (Guastafferro, 2013; Pogliana, 2013). Figure 12 schematically shows the heat treatment conditions used for each alloy, based on reported procedures (Little et al., 2007; Zhao & Frankel, 2007a,b). These were chosen because of the similarity of the electrochemical approaches and alloy systems, although the reverse scan in the case of Al 7075 was not recorded. In the case of Al 2024, annealed specimens (O temper) were also investigated. PS was recorded at 10 mV/min for i_rev=2.5 mA/cm² in 0.6 m NaCl (pH 6.5) using faced-up positioned specimens. Figure 13A reports the representative PS of Al 2024 for different thermal treatments. The inflection in the reverse curve is always detected at E_ptp values close to the corresponding E_pit, despite the remarkable shift of this electrochemical parameter to more negative potentials with age hardening. This provides additional confirmation that these potentials are closely related, as previously indicated. The width of the hysteresis loop changes more importantly than E_prot (between -850 and -900 mV). The quantitative analysis of E_pit and E_ptp (Figure 13B) indicates good agreement with the trend obtained for Al-Cu-Mg-Ag alloys in alkaline, deaerated 0.6 m NaCl (pH 10) as a function of the approximate wt% Cu in solid solution (XRD) (Little et al., 2007). An improvement in the IGSCC behavior on overaging, evaluated by CERT, was consistent with a decrease in E_ptp with increasing aging time and showed a correlation with the decrease in global Cu content in solid solution toward the Cu solid solution concentration in equilibrium with grain boundary precipitates. Diffractograms in Figure 13C also indicate a decrease of global Cu content with aging, according to the decrease of the peak at approximately 25° assigned to AlCu phase (Biswas, Siegel, Wolverton, & Seidman, 2011). Nevertheless, annealed surfaces (O temper in Figure 12A) show a quite intense AlCu peak, but E_pit and E_ptp values are in between those obtained for T8 and OA aging conditions (Figure 13B and C). Surface SEM/EDX analysis of as-polished surfaces after T8 and O treatments (Figure 14) indicated the precipitation of Cu-rich phases at grain boundaries, more importantly for the latter temper state. This suggests that E_ptp alone is not a sufficient parameter for evaluating the resistance against IGSCC of Cu-rich alloys. Figure 15 reports steepness and log i_ptp values for the different thermal treatments (Figure 13A). No correlation is obtained between these electrochemical characteristics and the intensity of AlCu peak, in contrast to the localized attack morphology (cross-sectional SEM images in Figure 15). The smallest values of steepness and log i_ptp are obtained for OA treatment, which exhibit wide open pits with some IG attack at the pit bottom. Both parameters increase as IG attack tends to prevail (T8), where the increase of log i_ptp could be related to the decrease of liquid-to-solid ratio in narrow corroded paths. The highest steepness is obtained for SHT and T3 tempers, being both susceptible to TGSCC, notwithstanding the different intensities of the AlCu peak as well as the closely similar E_ptp values (Figure 13B and C). Results suggest that Cu global content in solid solution and, consequently, Cu solid solution concentration in equilibrium with grain boundary precipitates have little effect on the transition onset and localized attack reactivation. The associated interfacial reactions seem to be determined by interdependent electrochemical and metallurgical events.

Figure 12:

Heat-treatment processes for (A) Al 2024 and (B) Al 7075 (SHT, solution heat treated; UA, underaged; T3, naturally aged; T6 and T8, artificially aged; OA and T7, overaged; O, annealed).

Figure 13:

Electrochemical behavior and composition of Al 2024 as a function of the thermal treatment. (A) PS of Al 2024 after different thermal treatments (Figure 12A) recorded at 10 mV/min with i_rev=2.5 mA/cm² in stagnant, air saturated 0.6 m NaCl (pH 6.5). (B) Variation of E_pit and E_ptp as a function of the thermal treatment. (C) Diffractograms of as-polished Al 2024 surfaces; symbols correspond to (*) Al, (•) Al₂Cu, (♦) AlCu, (■) AlCuMg, and (∇) Al₇Cu₂Fe.

Figure 14:

Surface SEM BSE/SE images of as-polished Al 2024 after annealing (A) and T8 (B) thermal treatments.

Figure 15:

Relationship between steepness and log i_ptp of Al 2024 (Figure 13A). Cross-sectional SEM images of corroded substrates for all treatment conditions but UA are included.

In the case of Al 7075, PS shows major differences in the forward curves with two breakdowns detected for peak-aged (T6) and overaged (T7) treatment conditions (Figure 16A), in agreement with previous findings (Zhao & Frankel, 2007a,b). E_pit1 and E_pit2 shift to less negative potentials, whereas E_ptp tends to E_pit2 with aging (Figure 16B). According to XRD analysis (Figure 16C), thermal treatment favors the precipitation of silicon- and chromium-rich phases. Unlike Al 2024, log i_ptp increases as cavities open, while the steepness slightly changes (Figure 17). Common to both alloys, however, is the remarkable increase of the steepness in the presence of TGSCC (T6). The comparison of Al 7075-T6 and Al 7075-T7 indicates similar log i_ptp (as well as E_ptp) but smaller steepness for the latter specimen, which correlates with a lower degree of damage of the alloy bulk (Figure 17). Thus, the steepness seems to be more adequate for estimating the degree of localized attack reactivation as a function of complex factors, as governed by local electrochemical reactions.

Figure 16:

Electrochemical behavior and composition of Al 7075 as a function of the thermal treatment. (A) PS of Al 7075 after different thermal treatments (Figure 12B) recorded at 10 mV/min with i_rev=2.5 mA/cm² in stagnant, air saturated 0.6 m NaCl (pH 6.5). (B) Variation of E_pit and E_ptp as a function of the thermal treatment. (C) Diffractograms of as-polished Al 7075 surfaces; symbols correspond to (*) Al, (•) Mg₂Si, (∇) MgZn₂, and (♦) Al₁₇Cr₂Fe.

Figure 17:

Relationships between steepness and log i_ptp of Al 7075 (Figure 16A). Cross-sectional SEM images of corroded substrates for all treatment conditions are included.

To consider also the effect of mechanical properties, microhardness measurements by microindentation (ISO 14577, DIN 50359) were conducted, using as-treated surfaces (Figure 12). Among the properties derived from the load-indentation depth plot, Vickers hardness was always higher for Al 7075 substrates. No correlation was found for electrochemical properties with the exception of the steepness. As shown in Figure 18, the percentage of plastic deformation in the case of Al 2024 and the elastic modulus in the case of Al 7075 changes with the thermal treatment as the steepness. Good match is also indicated for as-received alloys after 10 years of storage. Considering that steepness is principally determined by the kinetics of dissolution processes after the transition onset, results suggest that residual stresses interact with the local environment, influencing interfacial electrochemical kinetics in the periodic active-passive sequence involving dissolution-passivation-film rupture. The residual stress effect manifested by different mechanical properties correlates with the higher strength of Al 7075 and the greater toughness and ductility of Al 2024.

Figure 18:

Correlation between steepness and mechanical properties for (A) Al 2024 and (B) Al 7075 for different thermal treatments (T3* and T6*, respectively, correspond to as-received alloy after 10 years of storage).

According to these findings, an investigation of the repassivation kinetics in the presence of externally applied mechanical load has been undertaken to implicitly consider the four interrelated major variables of EAC: material, environment, electrochemical state, and stress. Fatigue studies combined with critical electrochemical potentials provide criterion for cracking but do not shed light on the features that dominate crack growth kinetics, which can be based on different mechanisms for controlling propagation rates (Little et al., 2007; Wall & Stoner, 1997). Thus, the electrochemical approach described herein could provide an understanding of non-steady state crack tip processes and on the fundamental mechanisms of environmental embrittlement.