Corrosion of Synthetic Intermetallic Compounds and AA7075-T6 in Dilute Harrison's Solution and Inhibition by Cerium(III) Salts

Electrochemical measurements on IMCs were performed using the electrochemical microcell approach. ^55,57 Cathodic and anodic polarization curves were measured separately at ambient temperature in the presence or absence of cerium salts in DHS. The cathodic measurements started +20 mV above the open circuit potential (E_oc) to −250 mV vs E_oc. The anodic measurements started –20 mV below E_oc in the upward direction until the current density reached 10⁻⁴ A cm⁻². The microcell consisted of a three-electrode setup: a working electrode, a platinum wire as a counter electrode, and a saturated calomel electrode (SCE, Hg/Hg₂Cl₂, 0.242 V vs saturated hydrogen electrode) as a reference electrode. The IMC working electrode was contacted by a borosilicate glass capillary with a silicone skirt/seal applied on the open end to prevent solution leakage. The measurement locations were selected randomly. The exposed area was approximately 70–80 μm in diameter, as determined after each experiment by observation using optical profilometry (Veeco Contour GT-K). The measurements were performed using a Gamry reference 600 potentiostat/galvanostat and controlled by Gamry Echem Analyst software (Gamry Framework^TM, Version 4.35).

Cathodic and anodic polarization curves were measured at a scanning rate of 10 mV s⁻¹ after a short initial stabilization time of 30 s following the contact of IMP with the solution. The short stabilization time was chosen to avoid any issues with cell leaking, the establishment of concentration gradients in the capillary, and clogging of the capillary with corrosion products. For the same reasons, a higher scan rate was used than in the macrocell setup. The polarization measurements were repeated at least 5 times; representative curves are shown, but all results are given in the Supplemental Information. The corrosion current density i_corr was obtained by extrapolating the linear portion (from about 50 mV more negative than the corrosion potential E_corr) of the cathodic part to E_corr. The following electrochemical parameters were determined: E_corr, i_corr, pitting potential E_pit and the span of the passive region calculated as ΔE = ∣E_pit − E_corr∣.

Electrochemical measurements on bulk AA7075-T6 were performed using a standard three-electrode electrochemical corrosion cell (Parstat, Corrosion Cell Kit, model K0047, volume 1 L). Substrate disks embedded in a Teflon holder (model K0105 flat sample holder kit), leaving an area of 0.95 cm² (diameter 1.1 cm) exposed to the corroding solution, served as a working electrode. An o-ring between the sample and holder was used to mask the area. Crevice corrosion under the o-ring was not observed. A SCE placed in a Luggin capillary salt bridge was used as a reference electrode, and a carbon rod served as a counter electrode. Electrochemical experiments were carried out with an Autolab PGSTAT 12 (Metrohm Autolab, Utrecht, The Netherlands) potentiostat/galvanostat and controlled by Nova 2.1 software.

The samples were allowed to stabilize at ambient temperature under open-circuit conditions for 3 min, 1 h or 6 h before initiation of polarization. The shortest stabilization time was chosen to assess the initiation process and correlate data with microcell experiments. Following stabilization, potentiodynamic measurements (cathodic and anodic curves) were performed separately using a 1 mV/s scan rate, starting at +20 mV above the E_oc and going to −250 mV vs E_oc in the downward direction, or –20 mV below E_oc and scanning in the upward direction until the current density reached 10⁻⁴ A cm⁻². For each sample, measurements were performed at least in triplicate, and a representative measurement was plotted in graphs. The electrochemical parameters described above were determined: E_corr, i_corr, E_pit and ΔE = |E_pit −E_corr|.

The corrosion behavior during long-term immersion (up to 2 weeks) of bulk AA7075-T6 was monitored using electrochemical impedance spectroscopy (EIS). The frequency range was from 100 kHz to 10 mHz at a sinusoidal voltage amplitude of 10 mV (rms). The sample was immersed in DHS without or with 3 mM cerium salt at ambient temperature for two weeks. The EIS spectra were recorded at E_oc at selected time intervals.

Synthesized Al₂Cu, Al₂CuMg, Al₇Cu₂Fe and MgZn₂ IMCs were metallographically polished to analyze their microstructure. First, the optical profilometry images in Fig. S1 were used to define the grain sizes, shapes and distribution. Additionally, these images were used to select the area to be tested by microcell. The grains in the IMCs differ in type (Fig. S1, left column). The Al₂Cu grain size was a few hundred micrometers, and the grain boundaries could not be distinguished well, but a eutectic network was seen. In the Al₂CuMg IMC, the grains were larger, and two different types of grains were noticed: the bright area representing the IMC was selected for testing. The grains were smaller, better defined, and well distributed in Al₇Cu₂Fe and MgZn₂ IMCs. The diameter of the individual grains was ∼50 μm.

Optical microscopy cannot determine the IMC composition so FE-SEM/EDS analyses were used to investigate further the grains and phases in the IMCs (Fig. 1). These IMCs have already been characterized by different techniques in the literature. ^3,6,35 Herein, we focus on surface characterization using the back-scattered SEM mode to distinguish the size and composition of the phases, where darker areas represent lighter and bright areas represent heavier elements.

Zoom In Zoom Out Reset image size

The BSE SEM images in Fig. 1 show the heterogeneous microstructure of the IMCs. Inset images at higher magnification confirm that different phases were present along the grain borders. The Al₂Cu IMC consisted of two constituents differentiated as dark and bright phases (Fig. 1a, Table I). The bright phase (1) contained Al and Cu with an approximate atomic ratio of 2:1, indicating that they are probably the θ-Al₂Cu phase. The dark grey phase (2) had an atomic ratio of Al:Cu, approximately 5:1, indicating that these are Al₅Cu.

The S-phase (Al₂CuMg) included three phases (Fig. 1b, Table I). The grey phase (3) contained Al, Cu, and Mg with an approximate atomic ratio of 2:1:1, suggesting the Al₂CuMg phase. The darker phase (4) had an atomic ratio of Al:Cu:Mg at approximately 9:1:6, indicating Al₇CuMg₆. The brightest phase (5) had an atomic ratio of approximately 1:1:1, corresponding to AlCuMg.

The Al₇Cu₂Fe sample consisted of four phases (Fig. 1c, Table I). The brightest grey phase (6) contained Al and Cu with an approximate atomic ratio of 2:1 and only a small amount of Fe, indicating they were probably the Al₂Cu phase. The grey phase (7) contained Al, Cu and Fe with an approximate atomic ratio of 7:2:1, indicating that this was Al₇Cu₂Fe (Fe-bearing phase). Structure (8) was a mixture of dark and bright phases and contained Al and Cu and a small amount of Zn with an approximate atomic ratio of Al:Cu of 4:1. The dark phase (9) contained predominantly Al with a small amount of Cu.

The MgZn₂ samples consisted of three phases (Fig. 1d, Table I). The light grey phase (10) contained a ratio Mg:Zn of 1:2, indicating that inclusions were MgZn₂ phase. The darker phase (11) contained Mg and Si as major elements with an approximate atomic ratio 2:1, along with a small amount of zinc; this phase was probably Mg₂Si (Si-bearing phase). The brightest phase (12) had an atomic ratio of Mg:Zn, approximately 1:6, indicating that these were Mg₆Zn.

The presented data confirmed that all IMCs had a well-defined composition without other metallic impurities, but each IMC contained several grains with different phases. The sizes of the phase in IMCs were less than the diameter of the microcell capillary tip, which had a diameter between 70–80 μm (Fig. S1, right column). The electrochemical measurements therefore reflected the average response measured on several grains and their boundaries within the particular IMC samples (Fig. S1), which affected the reproducibility, as described in the Experimental section. The electrochemical measurements presented below represent a typical response of these samples.

The metallographically polished AA7075-T6 surface was characterized using SEM imaging with a CBS detector (Fig. 2). The polishing makes the intermetallic particles well-defined and seen as bright or dark features incorporated in the aluminum matrix. They were differently sized or shaped and randomly distributed. The composition determined by EDS of the whole 85 × 55 μm area in Fig. 2a, which is marked with a dashed rectangle, is representative of the average alloy composition and is shown in the figure in atomic percentage. This bulk alloy composition agrees with the composition given by the producer's quality certificate as stated above. The matrix contains predominantly aluminum and a small amount of oxygen originating from the native passive Al oxide formed on the alloy surface. Within the matrix, very fine MgZn₂ precipitates were formed during solution heat treatment. ⁵ These IMPs can strongly affect the corrosion behavior of the alloy because MgZn₂ particles are active relative to the matrix and dissolve preferentially. ^13,58

Zoom In Zoom Out Reset image size

The size of the other IMPs is from a few tens of nm up to tens of μm depending on their composition (Fig. 2a). The darker, smaller particles ranging from 5 to 10 μm are the Mg-rich phase (marked with yellow arrows), and the lighter, larger particles ranging from 5 to 30 μm, were the α(Al–Fe–Cu) phase (Fe-bearing inclusions) and are marked with black arrows.

The Al₇Cu₂Fe IMPs were larger than other IMPs and irregularly shaped. The EDS composition of the selected Al–Fe–Cu particle in the dotted rectangle in the lower right of Fig. 2a was analyzed. EDS elemental maps and the average composition in Fig. 2b reveal different phases, consisting mainly of Al, Cu, Fe, Zn and Mg. Within the IMP, there was a region with more Al and Cu and darker contrast. Overall, the composition of this IMP correlates with the Al₇Cu₂Fe IMC (Fig. 1d). Zn and Mg were dispersed in the surrounding aluminum matrix.

The FIB/SEM/EDS analysis of the cross-section of the selected IMP in Fig. 2a (at the location marked with a yellow dashed line within the dotted rectangle) is shown in Fig. S2a. The Al–Cu–Fe IMP reached ∼4.9 μm into the aluminum alloy structure. The grain boundaries in the AA7075-T6 are marked with dashed lines. Voids within the alloy structure were probably formed during the production or thermal processing. Figure S2b presents the EDS maps of the visible area along the cross-section to analyze the distribution of the elements. Similar to the top-view imaging (Fig. 2b), Cu, Fe and Al were present in the IMP and Al, Zn and Mg in the surrounding matrix. The native oxide layer on the alloy surface was only a few nm thick and thinner on IMPs, marked with white arrows (Fig. S2a). Such a thin native oxide layer on IMPs represents the weak points for corrosion in the alloy. ⁵⁹

The comparative analyses of synthesized intermetallic compounds and the bulk alloy with intermetallic particles are the basis for further investigation of the corrosion mechanism in DHS and inhibition by Ce presented in the second and third parts of the study. The top view and cross-section SEM/EDS analyses of bulk AA7075-T6 reflected the heterogeneous distribution of the IMPs in the aluminum alloy. Among IMPs, the main focus is on Al–Cu–Fe since our recent study showed that these IMPs play a major role in forming Ce inhibitor layers. ⁴⁰

Separately measured anodic and cathodic curves were recorded using the microcell approach for the IMCs following 30 s immersion in DHS (Fig. 3). Due to the grain size and comingling of phases in each IMC sample (Fig. 1), it is challenging to carry out electrochemical measurements on a single phase. As a result, the electrochemical measurements were not performed only on a single grain/phase (Figs. 1a and S1), which impacted the reproducibility. All recorded curves are provided in Fig. S3. For comparison, the most typical curves were plotted in Fig. 3.

Zoom In Zoom Out Reset image size

The microcell experiments with a short immersion time (30 s) reflect the initial electrochemical response of the metal to contact with the solution compared to the behavior after longer exposure to E_oc as shown below. The curves show that each IMC, due to different compositions, behaves differently in DHS. The i_corr, E_corr, E_pit, and ΔE values are presented in Fig. 4 and Table SI. The less corrosion-resistant in DHS are IMCs containing Mg (MgZn₂ and Al₂CuMg), and the more resistant contain Cu (Al₇Cu₂Fe and Al₂Cu), as evidenced by decreasing i_corr (Fig. 4a). The difference in the i_corr between MgZn₂ (0.31 μA cm⁻²) and Al₂Cu (0.33 nA cm⁻²) or Al₇Cu₂Fe (0.4 nA cm⁻²) was approximately three orders of magnitude due to the high tendency of Mg to dissolve. ^35,58

Zoom In Zoom Out Reset image size

In DHS, E_corr became less negative in the given order: MgZn₂ < Al₂CuMg < Al₇Cu₂Fe < Al₂Cu (Fig. 4b, Table SI). The trend in pitting potential was similar: MgZn₂ < Al₂CuMg < Al₇Cu₂Fe < Al₂Cu (Fig. 4c, Table SI). More negative values of Al₂CuMg and especially MgZn₂ resulted from electrochemically active Mg in the IMC. The magnesium in the Al₂CuMg strongly determines its dissolution behavior in chloride solutions; it is selectively dissolved, leaving the IMP enriched in Cu. ^48,60 In DHS, MgZn₂ was the only IMC that did not have a potential range that established a passive region (ΔE) in the anodic curve. For Al–Cu–containing IMCs ΔE spanned between 230 mV and 380 mV (Fig. 4d).

The corrosion behavior of IMCs has been previously studied in 0.1 M NaCl at different pHs. ^6,7 Herein, we compared the behavior in DHS with that in the NaCl solution (Fig. 3, Table SI). In DHS, the i_corr values for Al₂Cu IMC were almost two orders of magnitude, and for Al₇Cu₂Fe and Al₂CuMg, one order of magnitude smaller than in 0.1 M NaCl. A decreased i_corr in DHS can be ascribed to the lower aggressiveness of the solution due to lower (∼12-times) chloride concentration (0.05 wt% (or ∼0.0085 M) in DHS and 0.1 M in NaCl). In contrast, i_corr for MgZn₂ was 6 times higher in DHS than in NaCl. The effect of corrosion media was also reflected in E_corr; the values for Al₂CuMg and Al₂Cu 0.31 V and 0.25 V more negative, respectively, in NaCl solution. For Al₇Cu₂Fe, the shift was only 0.03 V, whereas E_corr of MgZn₂ remained unchanged. Important differences could also be noticed in E_pit, with more noble values in DHS for the Al₇Cu₂Fe, Al₂CuMg, and Al₂Cu. The main difference was again MgZn₂, for E_pit was 0.11 V more negative in DHS than in NaCl solution.

These results undoubtedly confirmed the change in the corrosion behavior of IMCs in DHS compared to NaCl solution. The main difference between electrolytes can be explained by the lower concentration of Cl⁻, but also by the effect of NH₄ ⁺ and SO₄ ²⁻ ions in DHS. The pH values of the two solutions are similar: 5.2 (DHS) and 5.5 (0.1 M NaCl). Let us first consider the effect of ammonium ions. While all three Al–Cu-based IMCs showed less corrosion susceptibility in DHS than in NaCl, the MgZn₂ IMC exhibited the opposite behavior. Below a pH of approximately 10, the overall reaction of Mg in an aqueous solution can be expressed as the dissolution of magnesium accompanied by hydrogen evolution, reaction 1: ⁶¹

$$\begin{eqnarray}&&{\rm{Mg}}+2{{\rm{H}}}_{2}{\rm{O}}\rightleftharpoons {{\rm{Mg}}}^{2+}+{{\rm{H}}}_{2}+2{{\rm{OH}}}^{-}\end{eqnarray} \tag{ 1 }$$

In an aqueous solution, the ammonium ion acts as a weak acid because it dissociates to form hydroxyl ions and ammonia, reaction 2.

$$\begin{eqnarray}&&{{\rm{NH}}}_{4}^{+}+{{\rm{OH}}}^{-}\rightleftharpoons {{\rm{NH}}}_{3}+{{\rm{H}}}_{2}{\rm{O}}\end{eqnarray} \tag{ 2 }$$

This is the reaction characteristic of ammonium buffer, which maintains a pH between 10 and 11, depending on the concentration of ammonium salt. Consequently, it maintains the pH at a value below that for the precipitation of Mg(OH)₂. ²⁶ It was reported that Mg showed remarkable higher reactivity in the presence of NH₄ ⁺ ions. ²⁶ Uniform dissolution was observed shortly after immersion in 1 M (NH₄)₂SO₄ with vigorous H₂ evolution. This is due to the reaction of magnesium with ammonium ions, giving the overall reaction 3: ⁶²

$$\begin{eqnarray}&&{\rm{Mg}}+{{\rm{NH}}}_{4}^{+}+{{\rm{H}}}_{2}{\rm{O}}\rightleftharpoons {{\rm{Mg}}}^{2+}+{{\rm{OH}}}^{-}+{{\rm{NH}}}_{3}+{{\rm{H}}}_{2}\end{eqnarray} \tag{ 3 }$$

The reactivity of Mg in the presence of ammonium ions may be used to explain the behavior noticed for MgZn₂ in DHS (Figs. 3 and 4), resulting in stimulation of dissolution compared to that in NaCl. Dissolved Mg²⁺ ions can further react with SO₄ ²⁻ ions to form soluble magnesium sulfate MgSO₄ (or less soluble complex (NH₄)₂MgSO₄). ^20,26,63,64

The second issue that needs to be considered when discussing the behaviour in DHS is the formation of a protective layer of aluminum oxide in the sulfate-containing solution. ⁶⁵ It is known that the SO₄ ²⁻ ions displace chloride ions from the surface. ⁹ The presence of SO₄ ²⁻ reduces the pitting corrosion of aluminum, and corrosion occurs only locally on the pre-existing surface voids and cracks. ^20,26,66 At low sulfate concentrations, the formation of Al hydroxide is preferred; however, aluminum sulfate or oxysulfate is favored at higher concentrations. The reaction between SO₄ ²⁻ and Al ⁶⁷ may proceed according to reactions 5 and 6. ⁴³

$$\begin{eqnarray}&&{{\rm{Al}}}^{3+}+{{\rm{SO}}}_{4}^{2-}\rightleftharpoons {\rm{Al}}{\left({{\rm{SO}}}_{4}\right)}^{+}\end{eqnarray} \tag{ 5 }$$

$$\begin{eqnarray}&&{\rm{Al}}{\left({{\rm{SO}}}_{4}\right)}^{+}+{{\rm{OH}}}^{-}\rightleftharpoons {\rm{AlOH}}\left({{\rm{SO}}}_{4}\right)\end{eqnarray} \tag{ 6 }$$

The formation of ${\rm{AlOH}}\left({{\rm{SO}}}_{4}\right)$ can contribute to the more extensive passivation of Al₂Cu, Al₂CuMg, and Al₇Cu₂Mg observed in DHS compared to NaCl (Figs. 3 and 4). The composition of the surface layer is further addressed below.

Separately measured anodic and cathodic polarization curves recorded on AA7075-T6 after 3 min, 1 h and 6 h exposure at E_oc in DHS are given in Fig. 5. Overall, according to the position of E_corr, the curves for bulk alloy resemble those of Al–Cu-based IMCs (Fig. 3). However, the passive region was much smaller. After 3 min, the alloy was more susceptible to corrosion than Al₂Cu, Al₂CuMg and Al₇Cu₂Fe after 30 s immersion. The E_corr and E_pit were –0.52 and –0.42 V SCE, respectively, giving ΔE of 100 mV. When comparing the micro and macro measurements it should be considered that the area investigated largely differs (as explained below), which likely affects the number of breakdown events captured by each type of measurement.

Zoom In Zoom Out Reset image size

After 1 h, the curves were shifted slightly in a positive direction, but the passive region disappeared, indicating pitting at E_oc. After 6 h, the curves were shifted in the negative direction, and the rapidly increasing current density also suggested active pitting corrosion. The polarization curve recorded after 1 h of immersion of AA7075-T6 in 0.1 M NaCl is given for comparison (Fig. 5). In NaCl, i_corr was ∼16-times higher (Table SII) than in DHS because of the less noble value of E_corr, which was pinned at E_pit. ^40,46,68 Thus, similar to Cu-containing IMCs, the bulk AA7075-T6 also had better corrosion resistance in DHS compared to the NaCl solution.

The anodic and cathodic curves recorded in DHS with added cerium salts are presented in Fig. 6. The electrochemical response of all IMCs at the initial stage, after 30 s of immersion, changed in the presence of cerium salts, as shown by changed electrochemical parameters (Fig. 4, Table SI).

Zoom In Zoom Out Reset image size

For the Al₂Cu IMC, the shapes of the curves recorded in the presence of various cerium salts were similar. The cathodic curve was almost not affected by the addition of Ce salts, but there were noticeable differences in the i_corr, E_corr and E_pit. A pronounced increase in i_corr, about one order of magnitude, was noticed for Ce(OAc)₃ owing to an increase in the passive current density. E_corr remained similar in the presence of CeCl₃ and Ce(NO₃)₃, but for Ce(OAc)_3, it shifted more active by 50 mV (Figs. 6 and 4). The passive range became narrower in the presence of CeCl₃ and Ce(NO₃)₃.

Almost no effect of CeCl₃ added to DHS was noticed on Al₂CuMg, whereas Ce(OAc)₃ affected only the cathodic part (Fig. 6b). These two salts did not enhance the preferential dissolution of Mg. In contrast, Ce(NO₃)₃ caused a more significant change in the curve shape of Al₂CuMg. The E_corr and E_pit were shifted to less noble values, and ΔE became narrower. It seems that the addition of NO₃ ⁻ stimulates the dissolution of Al₂CuMg.

More pronounced effects of Ce salts were observed on the Al₇Cu₂Fe IMC (Figs. 6c and 4). Both cathodic and anodic curves were affected. All Ce salts caused both the anodic and cathodic portions of the polarization curves to shift to greater current densities, but the shape of the curves for Ce(OAc)₃ and CeCl₃ remained similar to that in DHS. As was the case for Al₂CuMg, the most significant effect was observed for DHS+Ce(NO₃)₃; i_corr was almost two orders of magnitude higher than the other two Ce salts and E_corr was shifted to more active values.

All Ce salts seem to have a similar effect on the electrochemical behavior of MgZn₂ IMC, primarily increasing the current density in the cathodic curve (Figs. 6d and 4). The E_corr remained similar. Enhancement of the cathodic reaction is the opposite of the effect Ce salts are known to have on Al alloys. ^43,46 This will be addressed below.

To sum up, the results confirmed that different Ce(III) salts affected the IMC electrochemical activity differently. At the initial stage of the process (first 30 s), the dissolution of almost all IMCs was promoted in DHS with added Ce(III) salts. The effect was most pronounced for Al₂Cu and Al₇Cu₂Fe IMCs for all three Ce salts and Al₂CuMg in DHS+Ce(NO₃)₃. MgZn₂ IMC was less influenced by adding Ce(III) salts. The subsequent stages of cerium inhibitor layer precipitation and growth examined by macrocell measurements are described in the next section.

The variation of E_oc of bulk AA7075-T6 as a function of the immersion time in DHS with and without added Ce(III) salts is shown in Fig. 7a. In DHS, E_oc increased abruptly after immersion from the very negative value of −0.95 V and reached a steady-state at −0.60 V after only 90 s. The more active metals (Mg and Al) are dissolved, and the native oxide layer also interacts with the corrosion medium. However, the oxide layer is quickly rebuilt, leading to the stabilization of E_oc. A scattering of values within ±10 mV reflected the continuous corrosion and repassivation processes, and no pitting was observed during 1 h of immersion.

Zoom In Zoom Out Reset image size

Differences in the E_oc curves were observed in DHS containing Ce salts, although the general shape was similar, two regions can be recognized. In the first region, a few seconds after immersion, all Ce salts exhibited more negative E_oc relative to DHS, most for DHS+Ce(NO₃)₃. E_oc shifted more positively but slower than in DHS, suggesting that the initial dissolution of active metals was promoted in the presence of Ce salts. A somewhat longer period was required to reach a steady state with less noisy potential. After 5 min, the E_oc stabilized around −0.70 V for DHS+Ce(OAc)₃ and DHS+CeCl₃. For DHS+Ce(NO₃)₃, E_oc was around −0.46 V after ∼40 min. Only in the latter solution was E_oc more positive than in DHS. Final E_oc values after 1 h were: −0.67 V, −0.63 V, −0.60 V and −0.54 V in DHS+Ce(OAc)₃, DHS+CeCl₃, DHS and DHS+Ce(NO₃)₃, respectively. Note that the scatter of these measurements was within 10 mV. After 6 h, E_oc in DHS+Ce(NO₃)₃ remained similar, whereas in all other solutions became more positive and stabilized around −0.56 V. Shifting potential to more noble values is related to the anodic inhibition or a lessening of cathodic inhibition.

Figures 7b−7d and S4 present the separately measured cathodic and anodic polarization curves of AA7075-T6 in DHS with and without added CeCl₃, Ce(NO₃)₃, and Ce(OAc)₃ after 3 min, 1 h and 6 h of stabilization time at E_oc. The curves were recorded to study the effect of immersion time on the course of corrosion in DHS and inhibition in DHS+Ce salts, respectively. Deduced electrochemical parameters are presented in Fig. 8. Note that the E_oc values were quite reproducible, so the overlaid cathodic and anodic curves appear to be one continuous curve.

Zoom In Zoom Out Reset image size

With adding cerium salts, significant differences observed in the polarization curves confirm that Ce salts act like corrosion inhibitors on AA7075-T6 because they reduce the corrosion rate. Already after 3 min at E_oc, curves were shifted to smaller current densities and more negative potentials. The negative shift was most pronounced for DHS+Ce(OAc)₃, resulting in a large ΔE. The i_corr was slightly lower or remained similar to DHS (Table SI). After 1 h of immersion, the i_corr decreased further in the solutions with Ce(III) salts as the result of decreases in both the anodic and cathodic kinetics. The net result caused a positive shift of E_corr. This trend continued at 6 h of immersion when all curves exhibited similar E_corr but considerably different i_corr and ΔE compared to DHS. The DHS curve indicated pitting at E_oc, whereas the curves in inhibited solution showed extended passive regions with 60 to 95-times smaller i_corr and ΔE between 220 and 330 mV (Fig. 8). The inhibition performance of cerium salts on Al alloys was reported to improve with prolonged time at the E_oc in chloride solution. ^41,43,46 This trend is now confirmed in DHS as well. DHA+Ce(OAc)₃ and DHS+Ce(NO₃)₃ showed similar values of i_corr and ΔE, both being smaller and broader, respectively, than for DHS+CeCl₃.

Based on the above data, all three Ce salts act initially (after 3 min) as cathodic inhibitors for the alloy, shifting E_corr in the negative direction and establishing a passive region with ΔE of a few hundred mV. In contrast to DHS, where the corrosion process is promoted with immersion time, in DHS+Ce salts, the corrosion of AA7075-T6 was considerably suppressed (Fig. 4). The shift of polarization curves with prolonged immersion time shows that all three Ce salts follow the transition from primarily cathodic inhibition at first to anodic inhibition later on. After 6 h, DHS+Ce(NO₃)₃ and DHS+Ce(OAc)₃ showed similar curves, i.e. similar level of protection, whereas CeCl₃ was somewhat lesser. This is related to the corresponding anions (NO₃ ⁻ and OAc⁻), which were involved in the inhibition mechanism in DHS solution, as already shown for chloride solution, ⁴³ in contrast to the etching behavior of CeCl₃. ¹⁶ The beneficial effect of the associated NO₃ ^– and OAc^– ions on the inhibition process is thus confirmed along with the synergetic effect between cerium ions and SO₄ ^2– ions in DHS. The beneficial effect of incorporating SO₄ ^2– ions in the Ce conversion layer was reported. ^43–45

The correlation of inhibition in DHS with and without added CeCl₃, Ce(NO₃)₃ and Ce(OAc)₃ was further evaluated during a prolonged immersion test (up to two weeks) using EIS measurements.

Bode plots of the impedance magnitude (|Z|) and phase angle (ϕ), as a function of frequency (f) are compared after 6 h (Fig. 9a) and 2 d of immersion (Fig. 9b). In DHS, bulk AA7075-T6 exhibited primarily capacitive behavior at 6 h of immersion, but then a large decrease in low-frequency impedance to 1.1 × 10⁵ Ω cm² after 2 d (Fig. 9b). The phase shift ϕ exhibits a rapid decrease at lower frequencies after 2 d. This behavior indicates an enhancement of corrosion, likely associated with pitting corrosion, which was observed on the specimens. The corrosion process in DHS is, however, less aggressive than in the NaCl solution, in which the low-frequency impedance was reported to be one order of magnitude lower, ∼1 × 10⁴ Ω cm². ^43,68

Zoom In Zoom Out Reset image size

The EIS results substantiate significant corrosion inhibition of AA7075-T6 in DHS containing Ce salts. Even after 6 h, the |Z| values increased continually with decreasing f, indicating the highly capacitive character of the protective layer formed at the metal surface during immersion. The greatest inhibition resulted from the addition of Ce(OAc)₃. At the lowest frequency of 10 mHz, the |Z| in DHS+Ce(OAc)₃ was ∼0.9 MΩ cm² compared to ∼0.6 MΩ cm² in DHS (Fig. 9a). After 2 d, the differences among Ce salts became more pronounced. |Z| for DHS+CeCl₃ was ∼0.6 MΩ cm², but still remained above values for DHS (Fig. 9b). However, the |Z| in DHS+Ce(NO₃)₃ and DHS+Ce(OAc)₃ increased up to 2.2 MΩ cm² and a broad plateau of ϕ was established in the frequency range from 10² to 1 Hz. The ϕ exhibited a maximum at −85°, thus reflecting high capacitive properties of the interfacial layers.

The EIS measurements were performed after extended immersion times of 14 d. Only the impedance values at low frequency (|Z_{10 mHz}|) are shown. |Z_{10 mHz}| values were taken as the representative parameter related to the corrosion process at the metal/layer interface ^43,45 and are presented as a function of immersion time (Fig. 10). In DHS, |Z_{10 mHz}| increased in the first 16 h and reached 1.0 MΩ cm², but then decreased progressively reflecting the corrosion process. In the presence of CeCl₃, the |Z_{10 mHz}| also initially reached a maximum at 16 h, but it then increased again, reaching another maximum of 0.6 MΩ cm² after 2 d of immersion. At longer immersion times, the |Z_{10 mHz}| decreased slowly but remained higher than that in the non-inhibited DHS (Fig. 10). Several pits were noticed after the end of the experiment.

Zoom In Zoom Out Reset image size

Measurements in DHS+Ce(NO₃)₃ and DHS+Ce(OAc)₃ showed considerably different curves. After 2 d, the |Z_{10 mHz}| in DHS+Ce(NO₃) reached a maximum of 2.2 MΩ cm² followed by a rapid and then slower decrease, reaching a steady value of 0.6 MΩ cm² after 9 d. The layer formed in the presence of Ce(NO₃)₃ exhibited better inhibition properties compared to that in CeCl₃, but the rapid drop was probably related to the formation of some pits or defects in the layer formed. The best and most durable inhibition performance was achieved in a solution containing Ce(OAc)₃. The |Z_{10 mHz}| values slightly fluctuated but remained between 2.0 and 2.5 MΩ cm² during the immersion period. These results indicate that the layer formed in the DHS+Ce(OAc)₃ solution can resist localized corrosion attack. The increase in low-frequency impedance was related to the precipitation of the protective Ce oxide layer and inhibition of the corrosion process. ⁶⁹

These EIS results demonstrated the importance of choosing the appropriate cerium salt to be added to the DHS due to the synergistic effect between Ce³⁺ cations and corresponding anions (NO₃ ^– and CH₃COO^–), promoting passivation and buffering effects. ^43–45 The addition of Ce(OAc)₃ to NaCl produces a buffering effect, maintaining the solution pH at 8.3, so that it is less sensitve to local pH changes driven by the cathodic reaction. ⁴³ In contrast, in unbuffered NaCl solution, pH quickly increases above pH 10, where aluminum is prone to dissolution.

Based on results obtained with electrochemical measurements, the XPS analyses were performed on the AA7075-T6 samples after 6 h of immersion in DHS with and without added CeCl₃, Ce(NO₃)₃ and Ce(OAc)₃ aiming to reveal the chemical composition and speciate the elements in the surface layers. XPS senses the surface layer of ca. 10 nm and does not have sufficient lateral resolution to study individual IMPS so it provides an average composition of the matrix and IMPs.

The survey XPS spectrum of AA7075-T6 immersed in DHS identified Al, O, S, Cl, and C (Fig. S5, Table II). The Al and O were related to aluminum alloy and the native oxide film. S (2.7 at%) and Cl (0.7 at%) originated from the corrosive medium (DHS); the sulfate ions (SO₄ ²⁻) were probably adsorbed (forming a complex) on the alloy surface, ^70,71 while Cl⁻ ions reacted with aluminum alloy forming AlCl₃. ⁴³ Carbon can be characterized as adventitious and was not analyzed further.

In solutions with DHS+Ce salts, the main differences in the survey XPS spectra were the presence of peaks related to cerium and a decrease in Al content. These results aligned with the electrochemical data, confirming the formation of the cerium conversion layer after 6 h of immersion in cerium-containing media. The content of Ce ranged between 0.1 and 1.0 at%, with lower values for Ce(NO₃)₃ and higher for Ce(OAc)₃ and CeCl₃ (Table II); however, it should be noted that the quantification below 1 at% is unreliable in XPS but still there was a comparative difference between the measured spectra. For the sample immersed in DHS+Ce(NO₃)₃, a small N 1s peak was detected related to N−O/Al bond (Figure S5).

The composition of the aluminum alloy surface was further analyzed with high-resolution XPS spectra of the selected elements (Fig. 11). The broad peak in Al 2p spectra at 74.9 eV can be associated with the native oxide layer (Al₂O₃ at 75.7 eV) and/or Al(III) oxyhydroxide (AlOOH at 74.2 eV) formed during immersion in DHS (Fig. 11a). The peak related to Al metal at 71.0–71.5 eV was not identified, indicating that the surface oxide was greater than ca. 10 nm in thickness. In DHS+Ce salts, the spectra were similar, with a broad peak centered at somewhat higher binding energy, 74.9 eV. It can be hypothesized that this peak is associated with the native oxide layer and very thin mixed Al–Ce oxide/hydroxide deposits.

Zoom In Zoom Out Reset image size

The O 1s spectra showed a similar broad peak for all immersed samples (Fig. 11b). Aluminum oxide and hydroxide appeared at similar binding energy (Al₂O₃ 531.9 eV and Al−OH 532.7 eV). In DHS with added Ce salts, the contribution of Ce oxide/hydroxide was expected as well, Ce–O at 529.5 eV and Ce–OH at 532 eV, due to the presence of Ce 3d peaks. ^72,73

The high-resolution XPS Ce 3d spectra (Fig. 11c) showed peaks related to the mixture of cerium species Ce(III)/Ce(IV) with five peaks corresponding to Ce(III) and Ce(IV) features at 882.7, 886.6, 898.3, 901.3 and 907.3 eV and a Ce(IV) satellite at 916.7 eV. The latter peak appeared for all three Ce salts indicating that both Ce(III) and Ce(IV) species were present in the layers formed. It seems that the ratio Ce(III)/Ce(IV) may vary with the type of Ce salt, but a more detailed analysis was beyond the scope of the present study.

The N 1 s spectra showed a broad peak only for DHS+Ce(NO₃)₃ immersed sample (Fig. 11d). The peak appeared at similar binding energy to most other oxygen species (N−O/Al 396.6 eV). As reported previously, ^17,39 nitrate ions could passivate the Al surface, consequently reducing the possibilities for cerium film deposition, which requires a cathodic reaction to proceed and causes a rise in pH. This process affects the course of inhibition by cerium cations as well.

The sulfur S 2p spectra confirmed the presence of sulfur at the surface (Fig. 11e). The S 2p spectrum comprises S 2p_3/2 and 2 p_1/2 peaks that differ by only 1.1 eV and are not usually differentiated as two separate peaks. SO₄ ²⁻ ions added in the form of (NH₄)₂SO₄ are dissociated in an aqueous solution as SO₄ ²⁻ and HSO₄ ⁻. Both ions are capable of adsorption on the aluminum surface or forming a complex with other ions, such as Ce(OH)SO₄. ⁴⁴ The peak center at 168.6 eV is consistent with the formation of SO₄ ²⁻ which is expected in the range of 168.6–169.5 eV. ⁷³

The XPS data confirmed that the deposition of Ce(III)/Ce(IV) layers occurs in DHS. The action of additional anions present in the solution (ammonium and sulfate ions) affects the course of the deposition, resulting in their reaction with the Al substrate and incorporation in the layers formed. These results correlate with the electrochemical data described above.

Bulk AA7075-T6 samples before and after 6 h of immersion in DHS with and without added CeCl₃, Ce(NO₃)₃ and Ce(OAc)₃ were further characterized by low-voltage SEM/EDS. The aim was to elaborate on the layer characteristics as a function of a cerium salt. Usually, higher electron voltages (10 kV) have been adopted for SEM/EDS analyses because the incident electrons excite the K-shell emissions for light elements to 3d transition metals, and high electron beam currents are easily achieved. In this study, the spatial resolution was increased by reducing the size of the X-ray generation volume at a lower accelerated voltage (Fig. S6). According to Monte Carlo calculations, the imaging and EDS analysis can be performed at a minimum of 5 kV because EDS analysis requires using L lines for first-row transition elements (Fig. S6).

The BSE SEM images of a selected IMP in the bulk AA7075-T6 alloy surface before and after 6 h of immersion are presented in Fig. 12. The same spots on selected Al-Cu-Fe-rich IMPs were imaged but analyzed only after immersion (to prevent the effect of surface activation during surface analysis). The impurities seen before and after immersion on the alloy surface were polishing residuals identified when imaging using a low voltage but were otherwise not identified at 10−15 kV.

Zoom In Zoom Out Reset image size

After 6 h of immersion in DHS solution, there was no evidence of corrosion damage near IMPs, such as pitting or other local types of corrosion, which is typical for immersion in 3.5 wt% NaCl solution. ^34,68 The EIS data in this solution indicated a decrease in low f impedance only after about 16 h. The selected Al–Cu–Fe-rich IMPs (denoted by spot x₁) remained virtually unchanged, but several dark-colored spots of corrosion products (denoted by spot x₂) were noted. These could be related to less corrosion-resistant MgZn₂ IMPs in the alloy matrix. The presence of a great amount of O, Na and S reflected the reaction of the matrix/IMPs with a corrosive medium resulting in the formation of corrosion products and complexes (Fig. 12a₂).

In DHS containing Ce salts, the situation changed (Figs, 12b₂, 12c₂, 12d₂). Comparing the SEM images before and after 6 h of immersion shows that the Ce layers of various thicknesses and morphologies were formed. The cerium layer formed in DHS+CeCl₃ (Fig. 12b₂) was detected on the IMP, spot x₃ (3.1 wt%) and the surrounding matrix, spot x₄ (4.3 wt%). It should be noted that due to the overlapping of Ce M peak and Cu L_α peaks recorded at the low voltage the EDS quantification is less reliable (Fig. S6).%.

The cerium layer formed in DHS+Ce(NO₃)₃ did not uniformly cover the surface after 6 h because several other defects were noted. The amount of Ce was lower on IMPs, spot x₅ (1.2 wt%) and matrix, spot x₆ (2.2 wt%) than in DHS+CeCl₃. In DHS+Ce(OAc)₃, the content of Ce on IMPs and matrix was the largest among the Ce salts, i.e. 6.4 wt% and 2.9 wt%, spots x₈ and spot x₇, respectively (Fig. 12d₂).

These results confirmed that after 6 h of immersion, some minor differences in the formed cerium layers vs cerium salts could be noticed, with Ce(OAc)₃ forming the most compact layer.