Introduction

High strength-to-weight ratio aluminium alloy (AA) 2024-T3 is an indispensable material in many transport industries like aerospace as it increases loading capacity and reduces energy consumption1. Although nano-to-micro scale intermetallic particles (IMPs) are the main contributors to mechanical properties, they are also the main culprits for the most insidious site-specific local degradation events at the surface of AA20242. IMPs such as Al2CuMg and Al2Cu tend to dealloy at the open circuit potential (OCP), leading to the creation of catalytic high surface area copper enriched intermetallic remnants and associated copper redistribution (i.e., copper replating) on the alloy surface3,4,5,6,7,8. Both of these processes significantly support increased oxygen reduction reaction (ORR) rates leading to increased local alkalinity due to the production of OH ions, causing aluminium matrix dissolution in the form of localised trenching around the IMPs and pit initiation and growth9,10,11. Therefore, an effective way to increase the corrosion resistance of AA2024 is to suppress the dissolution of IMPs. One aspect of current research in corrosion inhibition has focused on organic molecules with heteroatoms, such as nitrogen, oxygen, sulphur and phosphorus, that contain a lone pair of electrons and/or delocalised π electrons due to the presence of multiple bonds or aromatic rings12.

Our research group recently evaluated, using conventional electrochemical techniques, the efficacy of 2-mercaptobenzimidazole (MBI) and octylphosphonic acid (OPA) as single inhibitors13 and their mixtures to achieve possible synergy14 on copper and aluminium. MBI is an efficient inhibitor of copper corrosion due to its adsorption as deprotonated thiolate and the formation of a stable inhibitor film with the copper surface13,14. OPA, on the other hand, does not form strong bonds with Cu, but it forms strong metal-phosphonate bonds with Al13,14. Therefore, the MBI is an efficient inhibitor for Cu but not for Al, and OPA shows just the opposite behaviour in relation to Cu and Al substrates. These two compounds show a strong synergistic effect on Cu; when used in a molar mixture MBI: OPA = 9 : 1, a bonding mechanism changes from thiolate to thiolate and thione forms of MBI14. Furthermore, a thicker inhibitor film was formed owing to the change in pH to a slightly acidic range where the formation of cuprous ions is stimulated. It was discovered that chloride ions, acting simultaneously as promoters and reactants in forming a thick inhibitory film on copper15, caused the establishment of the aforementioned synergistic effect. No synergy of MBI and OPA was observed on aluminium, but MBI did not obstruct the inhibitor action of OPA, which is a good inhibitor for Al corrosion14. These previous results were the starting points for the next step of the research hypothesising that MBI and OPA would act complementary on Cu-rich IMPs and Al matrix, respectively, and thus mitigate the corrosion of AA2024.

The classical electrochemical techniques provide limited information on the electrochemical behaviour at corrosion initiation sites and lack sufficient spatial resolution (i.e., they only provide information on the corrosion rates averaged over the entire surface). Thus, we used localised electrochemical research techniques which offer in situ monitoring of the electrochemical activity at an early stage of corrosion and the progression of corrosion processes with time, with and without the presence of inhibitors16,17. Firstly, scanning electrochemical microscopy (SECM) and scanning vibrating electrode technique (SVET) were explored to study the Al/Cu galvanic pair in chloride solution and discern local anodic and cathodic sites18. The electrochemical activity in the galvanic pair was greatly reduced when the Cu surface was covered by MBI, thus blocking the cathodic sites18. In contrast, OPA did not provide sufficient barrier properties against the penetration of chloride ions and did not prevent pitting corrosion18. Continuing our approach further and translating to the AA2024 Cu-rich Al alloy, we have focused here on a localised microelectrochemical approach using the SVET to study the actions of organic compounds MBI and OPA.

Numerous localised electrochemical studies on corrosion inhibition of AA2024 were reported. Lamaka et al.19 showed, by applying the scanning Kelvin probe force microscopy (SKPFM), that quinaldic acid, salicylaldoxime and 8-hydroxyquinoline (8-HQ) suppress the dissolution of Mg, Al and Cu from intermetallic particles. Marcelin and Pébère20 investigated the 8-HQ and benzotriazole (BTA) mixture on Al/Cu couple and AA2024 by localised electrochemical impedance spectroscopy (LEIS). The corrosion inhibition effect of BTA was mainly limited to the copper-rich IMPs, whereas 8-HQ acted on the Al matrix due to its chelating properties. Williams et al.21 used SVET to quantify the effect of several copper complexing/precipitation reagents on the kinetics of pitting corrosion. The results showed that 2,5-dimercapto-1,3,4-thiadiazolate (DMTD), BTA and N,N-diethyl-dithiocarbamate (DEDTC) effectively inhibit pitting corrosion by the formation of a metal-inhibitor film on Cu-rich IMPs. In addition, they revealed that ethylenediaminetetraacetic acid (EDTA) prevented Cu replating without decreasing localised corrosion currents. This finding suggests sufficient cathodic oxygen reduction can be sustained on Cu-rich and dealloyed IMPs to drive the observed pitting corrosion without any copper replating. Cross-sectional analytical transmission and scanning electron microscopies were used to reveal a dealloying-driven local corrosion initiation at the studied IMPs that have been considered as cathodic phases traditionally (such as Al7Cu2Fe(Mn)), thus suggesting a possible initial (local) anodic character in most IMPs due to local compositional variations22,23. Independently, Olgiati24 showed that dealloying of the S-phase occurs twice as fast as the dealloying of AlCuFeMn.

Based on previously shown individual and synergistic inhibition actions of MBI and OPA on Cu and Al metals13,14,15,18, this paper is devoted to exploring the detailed interaction mechanism of MBI and OPA, as potential corrosion inhibitors for aluminium alloy 2024, with different Al- and Cu-rich intermetallic particles. The local electrochemical activity was evaluated in situ and quasi-real-time using SVET at the free corrosion potential in an inhibitor containing aggressive electrolyte such as 10 mM NaCl. Conventional electrochemical characterisation by potentiodynamic polarisation was performed in 0.5 M NaCl. Here, we also sought to understand the viability of formed inhibitory films and their effectiveness against corrosion in the NaCl solution that does not contain an MBI or OPA reservoir. Finally, we attempted to provide a more comprehensive morphological and surface chemical composition, both on top and cross-section of the substrate, using ex-situ focused ion beam/scanning electron microscopy coupled with energy-dispersive X-ray spectroscopy (FIB-SEM-EDS).

Results and discussion

Localised activity of AA2024 in chloride solution

The visualisation of ionic current flows was determined by repetitive in situ SVET scanning as a function of time over the surface of an unpolarised AA2024 sample, corroding freely in an aerated 10 mM NaCl solution without added inhibitor. Less concentrated solutions were selected for SVET measurements compared to polarisation curves and SEM-EDS experiments (where 0.5 M NaCl was chosen) because the sensitivity of SVET for measuring ionic current density is compromised by high electrolytic conductivity. Figure 1 shows representative ionic current density maps taken immediately after immersion (i.e., after 1 min), after 30 min, 6 h and 12 h. In the interests of better spatial resolution of the distribution of electrochemical activity on the surface of the AA2024, where the small IMPs concentrate the anodic activity while the corresponding cathodic activity is distributed over the larger matrix alloy, SVET images are constructed using two different scales; one is employed for the highly localised positive ionic (anodic) current densities and another for the more homogeneously distributed negative ionic (cathodic) current density distributions over the larger alloy matrix. An SVET scan recorded immediately after immersion revealed local anodic activity at numerous locations with a maximum ionic current density of jz ≤ 85 µA cm−2, indicating that localised corrosion begins within 1 min of exposure to the chloride-containing solution (Fig. 1a). The observed positive ionic currents are due to the flux of Al3+ ions released from the IMPs due to dealloying in the presence of chloride ions. Apart from the dealloying of the IMPs, the dissolution of the Al matrix at the periphery of the IMPs could also contribute to positive ionic currents, leading to pit initiation and growth. Successive scans show that some individual pits propagated and remained active over 12 h. In contrast, many others tended to deactivate or even became passive, which is observable by a reduced magnitude of the ionic current densities in the latter case (Fig. 1a-d). These changes with immersion time are related to the stochastic nature of the pitting process, including different stages such as pit nucleation, initiation, metastable regime, and stable pit growth2,25. Some nucleated pits do not propagate but repassivate2,25.

Fig. 1: SVET maps of AA2024 in NaCl.
figure 1

SVET maps recorded over AA2024 during immersion in 10 mM NaCl at selected exposure times: a 1 min, b 30 min, c 6 h and d 12 h. Areas of positive (anodic) and negative (cathodic) ionic current densities are given according to the colour scale bars from red to blue, respectively. Some pits (denoted by the red colour of high anodic current densities) tend to disappear at longer immersion time.

The visual appearance of the sample surface after 1 min and at the end of the immersion period of 12 h is shown in Fig. 2. The pit openings that coincide with the locations of the high anodic ionic current densities were especially pronounced during prolonged immersion; however, the ionic current densities decrease, which can be attributed to the deposition of corrosion products over time and changed electrochemical activity of some IMPs during immersion. Whilst each pit can be considered a point anode, the cathodic reaction, which is assumed to be the ORR, takes place in the surrounding areas (Fig. 1). It is likely that due to the excessive copper redistribution around the pit openings, as well as on the greater part of the surface of the Al matrix, new cathodic sites are formed over time. This process stimulates even more intensive oxygen reduction, reflected in more pronounced cathodic ion fluxes (i.e., negative ionic current densities) (Fig. 1a, d). It should be noted that the pits shown in Fig. 2b are much larger in size compared to the pits in Fig. 2a due to the increase in local alkalinity resulting from the generation of OH ions during oxygen reduction at the cathodic sites, which causes increased dissolution of the surrounding matrix at the periphery of the Cu-containing IMPs leading to localised corrosion and trenching23. One promising strategy to increase the corrosion resistance of AA2024 is to suppress the selective dissolution of Cu-rich IMPs (e.g., S-phase). For this purpose, the effectiveness of two corrosion inhibitors, MBI and OPA, in mitigating the pitting corrosion of AA2024 was evaluated.

Fig. 2: Optical micrographs of AA2024 in NaCl.
figure 2

Optical micrographs corresponding to the SVET images given in Fig. 1a, d.

Localised corrosion inhibition of AA2024 in MBI-containing solution

Figure 3 shows the maps of the ionic current density distributed over the surface of an AA2024 sample corroding spontaneously at the free corrosion potential during in situ immersion in a solution containing MBI (i.e., 10 mM NaCl + 1 mM MBI). An ionic current density map taken immediately after immersion (i.e., after 1 min) did not reveal significant local electrochemical activity indicating that the MBI inhibitor is instantly very effective against pitting corrosion (Fig. 3a). Ionic currents were homogeneously distributed over the sample surface with extremely low values. After 24 h immersion, an additional decrease in the ionic current density was observed, with anodic activity jz ≤ 0.5 µA cm−2, which may be a consequence of preventing the dissolution of S-phase particles since the MBI is a well-known inhibitor for Cu26,27,28,29,30,31,32 (Fig. 3b). The visual appearance of the sample surface after 1 min and 12 h of immersion is shown in Supplementary Fig. 1a, b. After that, the cell was rapidly washed with distilled water to remove traces of inhibitor and then filled with 10 mM NaCl solution to define its effectiveness against corrosion in a chloride environment without an MBI reservoir. In the early phase of immersion in NaCl solution (i.e., after 1 min) (Fig. 3c), one localised anodic site with activity not exceeding 5 µA cm−2 was observed, which was deactivated after 12 h (Fig. 3d). The visual appearance of the sample surface is shown in Supplementary Fig. 1c, d. Meanwhile, the remaining sample surface was slightly active due to the reported cathodic ionic current densities (jz ≤ −4.5 µA cm−2 in Fig. 3c) originating from the reduction of dissolved oxygen. This implies moderate retention of inhibitory properties in the environment without MBI. We have previously reported that MBI adsorbs only on Cu but not on Al13,14,18. Therefore, we assume here for AA2024 that the Cu-MBI film is formed on Cu-containing IMPs in the MBI solution. The Cu-MBI film provides significant in situ protection of the alloy substrate even after being removed from the MBI inhibitor solution, i.e., inhibitor reservoir, and replaced with only aggressive NaCl solution. Compared to Fig. 1, where the ionic currents in 10 mM NaCl solution could reach 85 µA cm−2, in Fig. 3c, d the exposure of AA2024 surface to MBI solution for 24 h before immersion in inhibitor-free NaCl solution hampered the localised activity to exceed over 5 µA cm−2 due to the stable Cu-MBI film, which prevented or diminished the dealloying of IMPs.

Fig. 3: Ionic current maps of AA2024 in NaCl + MBI.
figure 3

Ionic current density maps recorded above an AA2024 during immersion in 1 mM MBI + 10 mM NaCl at selected exposure times: a 1 min and b 24 h. The inhibitor solution was then replaced with 10 mM NaCl, and the maps were taken after c 1 min and d 12 h immersion. Areas of positive and negative ionic current densities are given according to the colour scale bars from red to blue, respectively. Optical micrographs corresponding to the SVET images are given in Supplementary Fig. 1.

Localised corrosion inhibition of AA2024 in OPA-containing solution

In contrast to the MBI inhibitor treatment, when AA2024 was exposed to a solution containing the OPA inhibitor (i.e., in 10 mM NaCl + 1 mM OPA), localised corrosion was observed at several spots at a very early stage of immersion with an anodic ionic current density of 28 µA cm−2 (Fig. 4). The optical micrographs of the sample surface after 1 min and 24 h of immersion are shown in Supplementary Fig. 2. The anodic activity increased dramatically during prolonged immersion (i.e., after 24 h, jz ≤ 180 µA cm−2). Interestingly, cathodic ionic currents were perhaps very weak over the entire surface but significantly increased near the pits, indicating that OPA is a good inhibitor for the Al matrix but does not prevent pitting corrosion. Due to the appearance of pitting in the presence of OPA, further experiments in NaCl solution without an OPA reservoir were not conducted.

Fig. 4: Ionic current maps of AA2024 in NaCl + OPA.
figure 4

Ionic current density maps recorded above an AA2024 during immersion in 1 mM OPA + 10 mM NaCl at selected exposure times: a 1 min and b 24 h. Areas of positive and negative ionic current densities are given according to the colour scale bars from red to blue, respectively. Optical micrographs corresponding to the SVET images are given in Supplementary Fig. 2.

Conventional polarisation measurements

Figure 5 shows the anodic and cathodic polarisation curves for AA2024 (i.e. blank or control sample) after 24 h ex-situ immersion in 0.5 M NaCl solution with and without adding 1 mM MBI or OPA. Measurements were performed in a classical (macro)-eletrochemical set-up using the AA2024 surface as the working electrode. Deduced electrochemical parameters are given in Supplementary Table 1. The cathodic ionic current densities of the sample in the presence of MBI are reduced for more than one decade together with the lower value of the cathodic Tafel slope (−bc = 193 mV dec−1), compared to the blank (−bc = 307 mV dec−1). In addition, little to no anodic inhibition was observed despite the shift of Ecorr to more positive values. This indicates that MBI is an effective cathodic corrosion inhibitor for AA2024 in an aqueous electrolyte containing chloride ions, resulting in the creation of a cathodically controlled system through the chemisorption of MBI-inhibitor molecules to the substrate (IMPs) via S and/or N atom(s), thus reducing the corrosion rates13,14,15,18. In contrast, the OPA sample has no corrosion inhibitor characteristics, i.e., it neither affects cathodic (especially on ORR kinetics) nor anodic reactions. The polarisation curves measurements are consistent with SVET results, except that the lack of anodic activity observed in SVET can be attributed either to significantly lower NaCl concentrations (10 mM) compared to polarisation curves experiments (0.5 M) or to the fact that the localised corrosion is highly stochastic and difficult to measure.

Fig. 5: Potentiodynamic polarisation curves of AA2024 in NaCl with and without MBI and OPA.
figure 5

Surface-averaging polarisation curves recorded for AA2024 after 24 h immersion in 0.5 M NaCl solution with and without adding 1 mM of MBI or OPA. The scan rate was 1 mV s−1. The bc values represent the corresponding cathodic Tafel slopes. Deduced electrochemical parameters are given in Supplementary Table 1.

Surface characterisation

Ex-situ surface analysis of AA2024 was performed by SEM-EDS after 24 h of immersion in 0.5 M NaCl electrolyte with and without added MBI or OPA (Fig. 6). The surface of the control sample (i.e., in NaCl solution without inhibitor) showed drastic changes due to localised corrosion processes, possibly, including Cu deposits over the entire surface of the Al matrix originating from the well-known copper redistribution process4 (Fig. 6a). This behaviour aligns with the proposed dealloying mechanism of S-phase producing predominantly Cu metal particles that are no longer attached to the metal surface and are mobile under the influence of moving solution or corrosion product3,4.

Fig. 6: SEM images of AA2024 in NaCl with and without MBI and OPA.
figure 6

SEM secondary electron images for AA2024 immersed for 24 h in a 0.5 M NaCl solution, b 0.5 M NaCl containing 1 mM MBI and c 0.5 M NaCl containing 1 mM OPA.

The sample pretreated with the MBI inhibitor had an insignificant number of these Cu deposits (Fig. 6b). This may imply that MBI inhibits the dealloying of IMPs and thus prevents copper redistribution or mitigates corrosion of Cu remnants. Both strategies would lead to an improvement in the kinetics of cathodic activity at local sites. In contrast, Fig. 6c shows that the OPA acted as a corrosion activator, resulting in many Cu particles surrounding the dealloyed IMPs.

The SEM images of the denoted area in Fig. 6a are recorded at higher magnification in SE and BSE modes (Fig. 7). Related EDS elemental maps and point analysis provide information on the elemental distribution after corrosion attack in NaCl solution without inhibitor (Fig. 7 and Table 1). There was evidence of a relatively high Cu content due to preferential dealloying of Al and Mg in the particle and dissolution of the Al matrix, which subsequently caused trenching around these particles. Semi-quantitative EDS chemical analysis of bare AA2024 (i.e. sample before immersion in NaCl solution) is shown in Supplementary Fig. 3 for comparison. It reveals that after immersion, the concentrations of Al and Mg are largely reduced due to dealloying. Further, the high O counts around the particle may indicate a thicker oxide than the surrounding matrix (Table 1).

Fig. 7: SEM-EDS analysis of Al2CuMg IMP in NaCl.
figure 7

SEM-EDS analysis of local corrosion induced by S-phase (Al2CuMg) particles (in Fig. 6a) recorded in secondary electron (SE) mode (left) and back-scattered mode (right) after 24 h immersion in 0.5 M NaCl solution. Locations denoted by a cross mark, square and rectangle indicate the areas/spots where the EDS analysis was conducted (Table 1). The elements were labelled in the EDS maps.

Table 1 EDS analysis of AA2024 immersed for 24 h in NaCl. Analyses were performed at locations 1 to 5, as denoted in Fig. 7.

For the sample treated with MBI, the EDS elemental maps (Fig. 8) show the distribution of S and N, which are an integral part of MBI. This suggests that MBI molecules form precipitates on copper-rich intermetallic particles via stable Cu–S and Cu–N bonds, thus blocking ORR sites, as previously shown using XPS and DFT14,29,30. In contrast, MBI was not precipitated on the Al matrix (Fig. 8, location 5, Table 2), consistent with our previous observations on a bare aluminium surface13. However, the detected sulphur at location 5 indicates the presence of MBI, most likely on Cu domains in solid solution, but this is negligible. It should be emphasised that the copper/iron-rich IMP (most likely Al7Cu2Fe(Mn)) was significantly attacked but still not completely dealloyed due to the remaining Al in the particle (locations 3 and 4, Table 2) and the lack of a bright ring around the particle corresponding to O counts, i.e., oxide deposits. MBI was also detected on the S-phase (Al2CuMg) particle as shown in Supplementary Fig. 3.

Fig. 8: SEM-EDS analysis of Al7Cu2Fe IMP in NaCl + MBI.
figure 8

SEM-EDS analysis of local corrosion induced by Al7Cu2Fe particle (in Fig. 6b) recorded in secondary electron (SE) mode (left) and back-scattered mode (right) after 24 h immersion in 0.5 M NaCl solution containing 1 mM MBI. Locations denoted by a cross mark, square and rectangle indicate the spots where the EDS analysis was conducted (Table 2). The elements were labelled in the EDS maps.

Table 2 EDS analysis of AA2024 immersed for 24 h in MBI+NaCl. Analyses were performed at locations 1 to 5, as denoted in Fig. 8.

The cross-section shown in Fig. 9 determines the dealloying penetration depth of the S-phase particle of Supplementary Fig. 4. The preferentially attacked particle resulted in a spongy morphology of the dealloyed S-phase with a continuous solid network of Cu-enriched remnants. Cross-sectional EDS analysis of the S-phase particle revealed that dissolution gradually evolved from the surface inwards. However, the remaining Mg with significant amounts of S and N (Fig. 9, locations 4 and 5, Table 3) indicates that the Al2CuMg was not completely corroded because MBI adsorbed on the porous copper surface resulting from the dealloyed S-phase particle. Likely, MBI does not entirely prevent S-phase dissolution. Still, it does reduce the dissolution rate and the rate of oxygen reduction on the porous copper remnant, as shown by potentiodynamic polarisation measurements (cf. Figure 5). Therefore, this behaviour is believed to be similar to the interaction of MBI with bare copper13,14. It is interesting that chlorine is observed where S and N are detected (Table 3, Supplementary Fig. 4), which is consistent with our recent finding15, where the role of chloride ions in the formation of corrosion protective films on copper is described in detail.

Fig. 9: Cross-section analysis of Al2CuMg IMP in NaCl + MBI.
figure 9

FIB-SEM (cross-section) analysis of local corrosion induced by Al2CuMg phase after 24 h immersion of AA2024 in 0.5 M NaCl containing 1 mM MBI. Yellow horizontal rectangles denote spots where the EDS analysis was performed (Table 3). Top-view of the particle is given in Supplementary Fig. 4.

Table 3 EDS analysis of AA2024 immersed for 24 h in MBI+NaCl. Analyses were performed at locations 1 to 5, as denoted in Fig. 9. The results are presented as atomic percentages (at.%).

On the other hand, the S-phase of the OPA-treated sample is characterised by complete corrosion where Al and Mg are selectively dissolved in the first stage, leaving the Cu-rich remnants exposed to the corrosion environment (Fig. 10). The IMP is almost entirely dealloyed and undercut from its adjacent matrix. EDS point analysis showed that copper redistribution occurs in the IMP environment, i.e., that redeposited Cu particles are trapped within the oxide deposit around the intermetallics (Fig. 10, locations 3 and 4, Table 4). In addition, the phosphorus map showed that OPA has not adsorbed on the IMP or Cu remnant but only on the Al matrix (Fig. 10, locations 3, 4 and 5, Table 4), which is consistent with the preference for OPA to readily form bonds with aluminium, but not with Cu. However, the adsorption of OPA was not sufficient to alleviate the local dissolution of AA2024. Our previous study on Al metal showed that phosphonic acids act as strong inhibitors of Al corrosion33. The prolongation of the alkyl chain from eight to eighteen carbon atoms further increases the inhibitory action of the phosphonic group33. Further study to reveal the effect of increased concentrations of OPA and/or prolonged chain lengths of phosphonic acids in the case of AA2024 remains to be continued.

Fig. 10: SEM-EDS analysis of Al2CuMg IMP in NaCl + OPA.
figure 10

SEM-EDS analysis of local corrosion induced by S-phase (Al2CuMg) particles (in Fig. 6c) recorded in secondary electron (SE) mode (left) and back-scattered mode (right) after 24 h immersion in 0.5 M NaCl containing 1 mM OPA. Locations denoted by a cross mark, square and rectangle indicate the spots where the EDS analysis was conducted (Table 4). The elements were labelled in the EDS maps. The Mg map was not shown because Mg is present only in the matrix.

Table 4 EDS analysis of AA2024 immersed for 24 h in OPA+NaCl. Analyses were performed at locations 1 to 5, as denoted in Fig. 10.

In summary, MBI is a very effective cathodic inhibitor for AA2024 in a chloride medium, exhibiting a lower cathodic Tafel slope (−bc = 193 mV dec−1) than the blank (−bc = 307 mV dec−1). In addition, SVET showed no localised activity on AA2024, even in a chloride environment without added MBI. On the other hand, OPA induces localised corrosion of AA2024, and its cathodic Tafel slope (−bc = 405 mV dec−1) is higher than that of the control sample, indicating higher kinetics of the cathodic reaction (particularly the ORR). Qualitative EDS analysis at the cross-section revealed that MBI does not entirely prevent the dealloying of Al and Mg, but reduces the dissolution rate and the rate of oxygen reduction by forming an insoluble Cu–MBI film on the partially dealloyed AlCu-based particles, i.e., on the porous copper remnants. Therefore, MBI has the potential to be used as a corrosion inhibitor of Cu-based aluminium alloys, while OPA, although it acts inhibitory on Al matrix, is a corrosion activator of Cu-rich sites leading to the propagation of localised corrosion. This study also reveals that the transition of the mechanism of inhibition of pure metals (Cu and Al) to alloy (AA2024) is not straightforward and requires further adjustments to achieve complete protection.

Methods

Materials, substrate preparation and chemicals

Flat sheets of AA2024 clad, 0.5 mm thick, distributed by Kaiser Aluminum, USA, were cut into discs 15 mm in diameter. The alloy contains Cu (4.4 wt.%), Mg (1.3 wt.%), Mn (0.5 wt.%), Fe (0.3 wt.%), Si (0.1 wt.%), and the rest is Al. The sample’s surface was water-ground first with P600-grit SiC emery paper to remove the alclad layer, then further abraded with P2400- and P4000-grit, rinsed with deionised water, and ultrasonically cleaned in absolute ethanol. The chemicals utilised for AA2024 modification were sodium chloride (NaCl, purity 99.5%, supplied by Fisher Scientific), 2-mercaptobenzimidazole (MBI, purity 98%, supplied by Sigma Aldrich) and octylphosphonic acid (OPA, purity 98%, supplied by Ark Pharm, Inc.).

The molecular formula of MBI is C7H6N2S; it is a planar organosulfur azole derivative with a C = S (thione) double bond (in aqueous solutions). MBI involves two N atoms in the ring and an S atom in the thiocarbonyl group available for coordination. The molecular formula of OPA is C8H19O3P. The OPA is an organophosphorus compound containing C-PO(OR)2 groups (where R = alkyl, aryl or hydrogen). Structural formulae of MBI and OPA are given in Supplementary Fig. 5.

SVET instrumentation and experimental procedure

The scanning vibrating electrode instrumentation was manufactured by Applicable Electronics Inc. (Forestdale, MA, USA). The vibrating probe consisted of Pt/Ir (80%/20%) wires, insulated with the parylene C® except at the tip, which was platinised to produce a spherical platinum black deposit at the tip with a diameter of 20 µm, which served as a sensor. Two Pt wires were employed as signal and reference electrodes. Mapping of electrochemical activity was performed in a constant height regime with a mean distance between the microelectrode and the sample surface of 100 µm. A video camera connected to an optical microscope was introduced in the system both for determining the tip-sample distance and for monitoring the movement of the vibrating electrode over the sample during operation, recording images of the substrate surface in real time. The measurements were made with the electrode tip vibrating perpendicular and parallel to the sample embedded horizontally in the epoxy resin and facing upwards. Microelectrochemical measurements were performed at room temperature under naturally aerated conditions in either 10 mM NaCl-containing 1 mM MBI or OPA inhibitors or just 10 mM NaCl solution. In-situ measurements of the corrosion inhibition consisted of two stages: first, the measuring cell was filled with 1 mM MBI or OPA + 10 mM NaCl solution (i.e., with an inhibitor reservoir) for 24 h, and then the solution was replaced by 10 mM NaCl (i.e., without inhibitor reservoir) for 12 h. The electrochemical activity was monitored throughout the immersion period. In addition, the corrosion of a control sample (blank) was monitored in 10 mM NaCl for 12 h. The 1 mM inhibitor concentration was selected to align with our previous studies on Al and Cu with and without galvanic coupling13,14,15,18.

Conventional electrochemical measurements

Electrochemical measurements were carried out in a conventional three-electrode cell assembly using PGSTAT M204 Autolab (Metrohm, Utrecht, Netherlands) multichannel potentiostat/galvanostat controlled by NOVA software. The working electrode (WE) was an AA2024 disc with a 1 cm2 exposed surface area. A Pt mesh and an Ag/AgClsat. (+0.197 V vs. standard hydrogen electrode) were employed as the counter and reference electrodes, respectively. The latter was used with a Luggin capillary to minimise IR drops at the WE surface. The polarisation curves were recorded starting in the cathodic range 250 mV below EOCP and increasing the potential in the anodic direction until reaching the cut-off ionic current density of 10−3 A cm−2 at a scan rate of 1 mV s−1. All electrochemical measurements were carried out at 298 K in NaCl solution containing inhibitor, except the blank solution containing only 0.5 M NaCl.

SEM-EDS-FIB characterisation

Scanning electron microscopy combined with energy-dispersive X-ray spectroscopy (SEM-EDS) was applied to reveal the morphology and semi-quantitative composition of the surface at areas of interest. SEM images were recorded in secondary electron (SE) and back-scattered electron (BSE) imaging modes at beam energies of 5 keV and 15 keV, respectively, using a field emission (FE) SEM JSM 7600 F, JEOL, Japan, equipped with EDS (Inca Oxford 350 EDS SDD). Before analysis, the samples were coated with a thin carbon layer to reduce the charging effect.

Focused ion beam microscopy (FIB) coupled with EDS revealed the composition and thickness of layers at the cross-section of the S phase (Al2CuMg) particle on the AA2024 substrate. First, an FEI Helios Nanolab 650 microscope, equipped with EDS (Oxford Instruments Aztec system with X-max SDD detector), was employed to deposit two Pt protection layers – electron-beam-assisted and ion beam (Ga+)-assisted ones with thicknesses of 0.2 and 1 µm, respectively. A trench was then opened by milling under rough conditions with Ga+ ions of 2.5 nA at 30 kV, followed by rough polishing (0.77 nA, 30 kV) and fine polishing (0.4 nA, 30 kV). Within the trench, the cross-section across the S-phase and the substrate was analysed by SEM-EDS.