Galvanic Couple Behavior Between AA7050-T7451 and Stainless Steel in a Fastener Arrangement Assessed with Coupled Multi-Electrode Arrays Under Atmospheric Exposure Conditions

Aluminum alloy (AA) 7050-T7451 (UNS A97050⁽¹⁾) is a precipitation hardened Al-Zn-Mg-Cu alloy that is often used in aerospace structures for fuselage and wing applications.^1-17  AA7050 is precipitation hardened through the precipitation and growth of MgZn₂.^18-23  MgZn₂ is often formed at the grain boundaries and becomes enriched with Cu upon overaging.²⁴  This is known to affect the local corrosion properties of AA7050-T7451 by depleting Cu in regions adjacent to the grain boundaries resulting in intergranular corrosion (IGC).²⁵  There are two coarse intermetallic particles often observed in AA7050, Al₇Cu₂Fe (β phase) and Al₂CuMg (S phase). These coarse intermetallic particles are distributed randomly throughout the AA matrix.²³  Both coarse constituent particles are anodic to the AA7050 matrix as they contain Cu and Fe. Cu-rich intermetallic particles play an important role in the local corrosion electrochemistry. Cu-containing particles are known to undergo incongruent dissolution which may leave a Cu-rich surface that is cathodic to the matrix.^24-30  In some cases, dealloying of Al₂CuMg occurs where the Al and Mg dissolve leaving behind a Cu-rich remnant which may become mechanically detached and dissolve into solution leading to Cu replating on the surface of the alloy.

AA7050-T7451 is often coupled with stainless steels in some rivet-plate assemblies for structural aircraft components. These joining locations are inherently susceptible to corrosion and fatigue because of wicking and trapping of electrolyte in the tight crevice, forming an aggressive occluded localized environment in a confined space. As a result, galvanic interactions between Al alloys and steel fasteners can often lead to failure mechanism in the aircraft applications. While the dangers of galvanic corrosion are recognized, little work has been done to study the galvanic interactions inside a crevice in detail. Young studied AA2024-T3 (UNS A92024) panels galvanically coupled to stainless steel fasteners by exposures followed by metallography, and demonstrated that corrosion damage did not occur by a single or incessant corrosion mode. The corrosion damage often transitioned from crevice corrosion and IGC to stress corrosion cracking.³¹  Matzdorf, et al., studied the behavior of AA2024 and stainless steel fasteners following ASTM B-117 exposures and long-term seacoast exposures. Analysis was conducted by scanning Kelvin probe measurements after exposure.^32-33  It was shown that galvanic corrosion of the AA2024 was controlled by the available cathodic current and that localized corrosion often occurred near the stainless steel fastener.³²  While the scanning Kelvin probe can provide valuable information on Volta potentials relatable to electrode potential, it does not capture local corrosion events that likely occur during galvanic attack.³⁴  Lastly, Lillard developed model predictions for Al-Cu galvanic couples and demonstrated the limitations of traditional polarization data to predict galvanic corrosion in complicated geometries.³⁵  However, this approach and analysis was limited to finite element method which elucidated static current and potential distributions under a limited number of E-log i boundary conditions. Therefore, none of the aforementioned studies captured the local corrosion events occurring during galvanic coupling.

In this work, coupled multi-electrode arrays (CMEAs) are used to interrogate the time dependent and local galvanic current interaction in both a simple flat galvanic geometry and a fastener geometry under full immersion, thin electrolyte film, and atmospheric conditions. CMEAs are often used to simulate common real-life configurations in terms of material, geometry, physical arrangement, and environment including atmospheric and thin film conditions. Aerospace applications typically involve atmospheric conditions as a result of the operational environments.^36-44  Little work has been conducted to understand the time dependent effects of thin films and atmospheric conditions on the localized galvanic corrosion kinetics in fastener geometries. Consequently, mechanistic insights are limited. A major advantage in using CMEAs is the ability to collect concurrent and instantaneous currents enabling measurements of real-time local corrosion processes.

Previous work on open-circuit corrosion of heterogeneous materials has been studied in the atmosphere with coupled microelectrode arrays in novel configurations.^45-49  For example, CMEAs have been used to study crevice corrosion to establish a critical crevice depth for stainless steels.^48,50  CMEAs have also been used to investigate IGC on sensitized Type 304 (UNS S30400) stainless steel.⁵⁰  Lastly, CMEAs have been utilized to study the effect of immersion conditions on the “galvanic throwing power” of a magnesium-rich primer system for AA2024-T351 substrates.⁴⁶  It was demonstrated that the throwing power was limited by the reduction in length of ionically conductive pathways during wet/dry cycles when the electrolyte became distorted but increased significantly in the shrinking zone of galvanic interaction just before drying.

Very little work has been conducted on the galvanic current interactions inside a fastener geometry of precipitation age hardened Al alloys such as AA7050-T7451. Moreover, the time-dependent local nature of these processes has not been elucidated. The rivet-plate arrangement is an interesting crevice lacking anode/cathode separation. The objective of this work is to investigate the galvanic current interactions in complicated geometries such as a galvanic fastener in both full immersion and atmospheric conditions including wet/dry cycling. CMEAs were used to elucidate galvanic current interactions between AA7050-T7451 and Type 316 (UNS S31600) stainless steel under bulk, thin films, and atmospheric wet/dry cycles with different geometric arrays, providing a 2D representation of a fastener plate galvanic couple. Galvanic current measurements taken concurrently and instantaneously facilitated the measurement of real-time, local corrosion processes in this crevice arrangement. Additionally, an understanding of local current interactions inside a simulated fastener array as a function of electrolyte thickness and physical geometrical aspects of a fastener were developed.

The wire used to construct a flush mounted CMEA was drawn from a rod machined from an AA7050 alloy plate provided by Alcoa. The wire drawing process was conducted by DOE Ames Laboratory. The Type 316 stainless steel wires (diameter 250 μm, insulated with heavy polyamide) was obtained from California Fine Wire Company. The compositions of AA7050-T7451 and Type 316 stainless steel can be found in Table 1.

Two CMEAs were constructed with different geometries to understand the galvanic current interactions of AA7050-T7451 coupled to Type 316 stainless steel. The first CMEA was designed to simulate a simple planar galvanic couple geometry. This CMEA was constructed using an AA7050-T7451 rectangular panel (2.7 mm × 10.5 mm) with 12 holes (0.25 mm) spaced 0.15 mm apart. Six AA7050-T7451 and six Type 316 stainless steel electrodes (diameter 0.25 mm) were embedded in the AA panel to be flush with the rectangular panel. The AA panel of flush mounted electrodes was embedded in epoxy. Figure 1 shows a schematic to this CMEA, which is referred to as the flat galvanic CMEA. The second CMEA was constructed to simulate a 2D fastener-plate geometry where 22 AA7050 electrodes were embedded in an AA panel and 20 Type 316 stainless steel electrodes were embedded in a stainless steel panel and arranged in a 2D geometric rendition of fastener geometry (Figure 2). This CMEA is referred to as the galvanic fastener CMEA. The CMEA simulates a fastener where the gap between the AA7050-T7451 and Type 316 stainless steel may be adjusted using plastic set screws in the epoxy in combination with shims of various thicknesses to set the gap describing this confined space. A thin film of electrolyte was placed on the AA7050-T7451 fastener mouth, and wicked inside the fastener hole. It is important to note that the electrolyte layer only was placed on the Al alloy and not on the stainless steel. In all cases, the electrodes were constructed from electrically isolated wires of 0.25 mm diameter with a spacing of 0.15 mm. Insulating the electrodes allowed electrodes to contact each other and the panel while ensuring electrical isolation. Before each exposure the arrays were freshly ground with successively finer grit SiC abrasive paper to 1200 grit.

CMEA exposure testing was conducted in a controlled relative humidity (RH) chamber set to 98% RH or cyclic wet/dry cycles (Figure 3). The ribbon cable connections were made to a coupled multi-electrode analyzer (CMMA) through an aperture in the chamber. The RH controlled cabinet collected RH and temperature data throughout exposures. The two CMEAs described were tested under 70 μm thin films, 500 μm thin films, and full immersion in 0.6 M NaCl at room temperature (25°C). The metal panel was used in this study to illustrate the effects of the high and more realistic wettability of Al and its oxides compared to epoxy. The equilibrium concentration of 0.6 M NaCl solution at 98% RH remains near the 0.6 M NaCl concentration. At 95% RH, the equilibrium concentration of 0.6 M NaCl solution is 1.4 M NaCl. Thin films were placed on the CMEA using a controlled area masked with polyamide tape and a micropipette to place a water layer of controlled volume on the surface. Wet/dry cycling was also conducted in the controlled RH chamber. The duration of the wet/dry cycle was 24 h with 30 min at 95% RH and 3.5 h at 30% RH. Wet/dry cycles were tested at an elevated temperature of 50°C. Wet to dry cycles occurred over the 3.5 h period and all data were recorded. While not shown in this paper, in separate exposures, the CMEAs was also tested under a 70 μm thin film of 4 M NaCl and under a 500 μm thin film of 0.6 M NaCl.⁵¹  The CMEA arranged in a fastener geometry was also tested with a larger gap (500 μm) between the AA7050-T7451 and the Type 316 stainless steel.

A multi-electrode analyzer was used to provide data acquisition of each microelectrode current. The CMMA galvanically couples each microelectrode via zero resistance ammeter (ZRA) with a measureable current range of 3.3 nA to 100 μA per channel. The minimum and maximum current densities of the MMA given the small cross-sectional area of the flush mounted microelectrodes are 2 × 10⁻⁶ A/cm² and 2 × 10⁻¹ A/cm², respectively. In each color map, dark red indicates an anodic current (positive) and dark blue indicates a cathodic current (negative). Microelectrodes which are freely corroding pass a net current of zero and are color coded white. In this work, the data acquisition rate was one point per second.

The flat geometric CMEA, shown in Figure 1, provided a reference condition for a geometrically simple galvanic couple between AA7050-T7451 and Type 316 stainless steel. The CMEA was exposed to different atmospheric conditions to understand the effect of thin films (70 μm) relative to full immersion and in wet/dry cycling conditions. The ZRA net galvanic current density for the flat geometric CMEA in 0.6 M NaCl full immersion can be observed in Figure 4. Two AA7050 anodes of high magnitude prevailed over the 24 h exposure, A5 and A6. These two electrodes were located the closest to the stainless steel electrodes. 100% of the net anodic charge was contributed from A5 and A6 (Table 2). The rest of the AA7050 electrodes had net cathodic charge. Figure 5 shows the color maps after every 10,000 s in the exposure. Initially, all AA7050 electrodes were red (indicating anodic current) and all of the Type 316 stainless steel electrodes were blue (indicating cathodic current). After 20,000 s, electrodes A1 through A4 switched from net anodic to net cathodic behavior. The total net anodic charge density over the 24 h exposure was 0.92 C/cm². The cathodic charge density was equal but opposite in sign. In previous work, ZRA corrosion measurements (Pt mesh and saturated calomel electrode [SCE]) on planar electrodes (1 cm²) of AA7050-T7451 coupled to Type 316 stainless for 24 h in 0.6 M NaCl yielded a net anodic charge density of 0.84 C/cm², suggesting excellent agreement between the CMEA and planar electrodes in full immersion.

Figure 6 shows the ZRA net galvanic current density over a 24 h exposure under a 70 μm continuous thin film of 0.6 M NaCl at 98% RH. It can be observed that two anodes prevailed throughout the 24 h exposure, A4 and A5. The charge densities of A4 and A5 were 16.3 C/cm² and 0.8 C/cm² (Table 3). Electrode A4 accounted for 95% of the total net anodic current observed. The cathodic electrodes all maintained similar low current values through the 24 h exposure. The color maps in Figure 7 show that at the start of the exposure all AA7050 electrodes were initially anodic (red). However, by 10,000 s four of the six AA7050 electrodes switched from anodic to cathodic current (blue). Therefore, by the end of the 24 h exposure under a 70 μm thin film, 85% of the electrodes were supporting oxygen reduction reaction (ORR) helping to sustain the two anodic sites. The total net anodic charge density over the 24 h exposure was 16.4 C/cm². Under a 70 μm thin film, the charge density increased by over one order of magnitude when compared to full immersion conditions.

A 24 h wet/dry cycle consisting of 30 min at 95% RH followed by 3.5 h at 30% RH for 24 h at a higher temperature of 50°C was conducted on the simple flat CMEA (Figure 8). Five anodic sites prevailed over the exposure. A2 was the strongest anode, supplying 39% of the anodic current (Table 4). Anodic and cathodic current spikes were observed immediately at the end of dry cycle as the RH began climbing to the wet cycle (Figure 9). Interestingly, the strongest cathode was AA7050 electrode A1, supplying 90% of the total net cathodic charge. The cathodic current of A1 became more negative during wetting periods, supplying the cathodic reaction which supported the anodic sites A2 through A6. Figure 10 shows the color maps relative to the wet/dry cycles. It can be observed that at around 4 h, as the dry cycle ended and the RH began to climb, A1 was a strong cathodic site (blue), supplying the current for A2 through A5, which were all dark red indicating strong anodic sites. While the six stainless steel electrodes were also cathodic sites, the stainless steel electrodes only supplied 10% of the total cathodic current. The total anodic charge over 24 h for the 0.6 M NaCl 70 μm thin film wet/dry cycle was 49.12 C/cm².

While the galvanic current data are not shown in this paper, the CMEA was also exposed to a thicker film of 500 μm of 0.6 M NaCl at 98% RH. During the 24 h exposure, three anodic sites prevailed. Under the 500 μm thin film, 50% of the AA7050 electrodes were anodic sites. The total net anodic charge density over the 24 h exposure was 12.9 C/cm². The exposure under a 500 μm thin film experienced a lower charge density by 4.3 C/cm² relative to the exposure under a 70 μm thin film. The CMEA was also exposed under a 4 M NaCl 70 μm thin film. A3 was the strongest anodic site, accounting for 87% of the total net anodic current. Interestingly, AA7050 electrode A2 sharply switched from anodic current to the strongest cathodic current for the remaining 8 h. The net charge density of A2 after the 24 h exposure was −1.27 C/cm². AA7050 electrodes contributed approximately 20% of the total net cathodic current density. The total net anodic charge density over the 24 h exposure was 19 C/cm². The 4 M NaCl exposure exhibited a faster corrosion rate by 18% over the 0.6 M NaCl exposure.

The geometric fastener array shown in Figure 2 was exposed under thin films of various thicknesses and other atmospheric conditions such as wet/dry cycling. As a reference point, Figure 11 shows a 10,000 s exposure in 0.6 M NaCl full immersion with a 100 μm gap between the AA7050-T7451 and Type 316 stainless steel. It can be observed that AA7050 electrode A11 was the dominant anodic site throughout the exposure. A11 was located inside the crevice but near the mouth. Figure 12 shows the color maps at different scales allowing the electrode with the highest magnitude currents to be distinguished. Part (a) shows color maps at a higher current scale and (b) shows the same maps at a lower current scale. It can be observed that in this exposure, higher currents were observed at the mouth of the fastener and also at the bottom of the fastener. The dominant cathode was observed at the bottom of the Type 316 stainless steel. The middle region of the crevice had the slowest kinetics. The net anodic charge density over the 10,000 s was 0.71 C/cm².

Figure 13 shows the ZRA current measurements over 10,000 s under a 70 μm thin film of 0.6 M NaCl with a 100 μm gap between the AA7050-T7451 and Type 316 stainless steel. Under a thin film many more anodic and cathodic sites were activated relative to the full immersion exposures. The strongest anodic sites were AA7050 electrodes A8 and A22. AA7050 A8 corresponds with the electrode at the edge of the mouth of the fastener and A22 corresponds to last electrode in the bottom of the crevice. The stainless steel rivet had the highest cathodic currents inside the fastener. In general, higher magnitude currents were observed at the mouth and at the bottom of the crevice (Figure 14). The highest anodic and cathodic currents were seen at the edge of mouth of the fastener. The net anodic charge density over the 10,000 s was 7.12 C/cm². The anodic charge density was higher by one order of magnitude under a thin film than in full immersion. Moreover, the fastener geometry creating a confined space resulted in currents four times higher than the geometric flat array, after 10,000 s.

The fastener array was exposed to a 24 h wet/dry cycle, with a 70 μm thin film of 0.6 M NaCl solution. ZRA current measurements can be observed in Figure 15. Currents increased upon the onset of wet and dry cycles. A1 was the dominant anode throughout the exposure. During the wet/dry cycles many changes in polarity were observed (Figure 16). The color maps showed that the highest currents were observed at the bottom of the fastener (Figure 17). Electrodes at the mouth of the electrode often switched from anodic to cathodic sites (Figure 18) throughout the cycle. The net anodic charge density over the 24 h was 72.4 C/cm². The net anodic charge of the cyclic environment was higher than the same exposure under constant 98% RH.

The CMEA arranged in a 2D fastener configuration was used to assess the galvanic interaction under a thicker film (500 μm) and also in a separate experiment, with a larger gap between the AA7050-T7451 and Type 316 stainless steel.⁵¹  However, these data are not shown in this paper. Under thicker film of height 500 μm the highest anodic currents were observed at the mouth of the crevice.⁵¹  Furthermore, these anodic sites were supported by three strong cathodic sites inside the crevice on the stainless steel. The net anodic charge density over the 10,000 s was 5.9 C/cm². Overall, lower currents were observed under 500 μm than in under a 70 μm thin film of 0.6 M NaCl.

Under a 70 μm thin film and a larger gap between the AA7050 and the stainless steel (500 μm), the highest anodic currents were observed on AA7050 electrodes at the mouth of the AA7050. Interestingly, after 4,000 s the anodic current increased at the AA7050 electrode located at the bottom of the crevice. This was a result of the behavior of AA7050 electrode A2, at the mouth of the crevice, which switched from a strong anodic to a strong cathodic site. This can be likely attributed to S-phase dealloying or copper replating on the surface, as discussed elsewhere.⁵²  The mouth of the crevice had the highest anodic currents. Moreover, AA7050 electrodes inside the crevice were mostly cathodic sites, which may be attributed to copper replating upon corrosion inside the crevice. The net anodic charge density over the 10,000 s was 5.6 C/cm². Overall, lower currents were observed with a larger gap of 500 μm compared to the exposure with a narrower gap of ∼100 μm.

When AA7050-T7451 was coupled to Type 316 stainless steel in both a simple geometry and in a fastener geometry, inhomogeneous persistent anodes prevailed throughout the exposures (Figures 4, 6, and 15). Concerning the simple geometric CMEA, typical anodic sites initiated on the AA7050-T7451 with no tendency toward favoring AA7050 electrodes that were the closest to the Type 316 stainless (Figure 7). This suggests that while the Type 316 stainless steel consistently supports cathodic ORR on all electrodes, the inhomogeneity of the AA7050-T7451 underlying microstructure is a key factor. AA7050-T7451 contains course intermetallic particles such as Al₂CuMg and Al₇CuFe that are anodic to the AA7050 matrix which may serve as initiation regions for localized corrosion to occur.^22-24,51  These sites later may transition to other forms of corrosion such as intergranular or fissures.^53-54  Moreover, Al₂CuMg may undergo dealloying of Al and Mg, leaving behind a Cu-rich particle readily supporting ORR.²⁷  This copper remnant may exhibit fast electron transfer for ORR and hydrogen evolution reaction (HER).^55-56  In many cases (Figures 6, 10, and 13), only 30% of the AA7050 electrodes exhibited a net anodic current density. Typically, one to two anodic sites dominated the net galvanic anodic currents, while the remaining AA7050-T7451 electrodes as well as Type 316 stainless steel electrodes supported ORR to enable these locally intense anodic reaction rates (Figure 7). Therefore, it can be speculated that the anodes were dictated by susceptible local sites where Al initiated localized corrosion. This is consistent with the observations of local corrosion in the form of fissures reported from x-ray tomography and cross-sectioning in previous work.^12-13,17  Therefore, it can be speculated that the anode locations were dictated by local sites where AA7050 initiated localized corrosion and not some “macro current distribution effect” as seen in galvanic couples on homogeneous electrodes with strong polarity differences. It is also important to note that in many exposures, the dominant net anodic sites often switched to net cathodic sites (Figures 8 and 13). This could be an indication of S-phase (Al₂CuMg), initially a net anodic site, switching to a net cathodic site after dealloying of the Al and Mg, leaving behind a cathodic high surface area remnant, or may result in Cu release which replates on the surface of the alloy.^27,45,52  In future work, further inspection of electrodes by scanning electron microscope is warranted. However, cyclic voltammograms on selected electrodes indicated copper replating on AA7050.⁵²  In some cases the AA7050-T7451 supported over 50% of the net cathodic reaction in the galvanic couple (Table 4). The main conclusion here is that AA7050-T7451 electrodes can become significant cathodes that support isolated AA7050 fissure growth, even when coupled to Type 316 stainless steel. Moreover, AA7050 electrodes that formed strong anodes at variable sites are likely based on the metallurgically driven propensity for localized breakdown as opposed to proximity to the stainless steel as usually in a macrogalvanic couple formed between active homogeneous materials.

Figure 19 shows comparisons of the net anodic charge from each exposure on the simple CMEA. The charge density for the flat geometric CMEA was calculated over the 24 h exposure at room temperature, unless noted otherwise. For each exposure, experiments were conducted twice to ensure reproducibility. The net anodic charge on the flat geometric CMEA was 0.92 C/cm² (Figure 4), under full immersion 0.6 M NaCl solution. Comparatively, on planar electrodes consisting of galvanically coupled AA7050-T7451 and Type 316 stainless steel in a conventional flat cell on regular planar electrodes, the net anodic current was 0.84 C/cm² under full immersion. This shows good agreement and ensures the CMEA method is a valid method to compare factors such as thin film thickness, geometry, and wet/dry cycling.

Under a 70 μm thin film at 98% RH, the anodic charge density was 16.4 C/cm² in 0.6 M NaCl and 19.1 C/cm² in 4 M NaCl solution (Figure 6). This demonstrated that under thin films, the galvanic corrosion rate increased by over one order of magnitude relative to full immersion. The increase in charge density from the change from full immersion to thin films can be attributed to the higher concentration gradient for mass transport limited ORR of O₂ resulting from shorter diffusion lengths. This increases the rate of mass transport limited ORR typical on Al alloys.^57-58  While the data are not shown in this paper, under a thicker electrolyte layer, 500 μm (0.6 M NaCl), the net anodic charge density was just 11.9 C/cm², demonstrating that under a thicker film the corrosion rate was reduced by over 25% compared to a thinner 70 μm electrolyte film. The results for 500 μm are between those on 70 μm thin films and in full immersion. The charge density under a 500 μm thin film was less than one order of magnitude higher than under full immersion. These results were as expected as increasing the water layer thickness increases the O₂ diffusion length resulting in ORR at a slower rate than for the 70 μm thin film. The effect may not scale by a direct inverse relationship with electrolyte thickness, as ORR on the heterogeneous Al alloy does not necessarily obey Fick’s first law as shown by modeling of rotating disk electrode studies on AA2024.⁵⁹ 

The CMEA arranged in a fastener geometry was exposed to four atmospheric conditions for 10,000 s and compared to full immersion on both the simple galvanic CMEA. Figure 20 shows the charge density summary plots for the fastener CMEA after 10,000 s of exposure. Under full immersion conditions in 0.6 M NaCl solution, the net anodic charge density was 1.87 C/cm² (Figure 11). For comparison purposes, the charge density on the geometric flat CMEA under a 70 μm thin film after 10,000 s was interpolated from the 24 h exposure (Figure 6). Under a thicker film of 500 μm, the charge density was slightly lower than in the case of the 70 μm exposure, 5.9 C/cm². In summary, the effect of water layer thickness on the galvanic corrosion demonstrated that (1) net anodic charge density during galvanic corrosion increased under thin electrolyte films relative to full immersion, and (2) the net anodic charge density during coupling of AA7050 and Type 316 stainless steel decreased when changing from 70 μm to 500 μm thick electrolyte films.

In a fastener geometry between AA7050-T7451 and Type 316 stainless steel, anodic currents were highest at the mouth and inside the crevice, where there was a confined electrolyte resulting from the geometry of a fastener (Figures 11 through 18). The increase in current at the mouth of the crevice can be speculated to occur as a result of oxygen being readily available for ORR to occur on the Type 316 stainless steel. The increase in ORR in thin films and during cycling is well established.^57-58 

Because of the geometry of the fastener, it is very difficult for oxygen to diffuse to the inside of the crevice sufficiently to maintain the bulk O₂ levels. Hence, a different explanation must be provided to explain behavior of the crevice at deep depths. The ORR rate at 8 ppm dissolved oxygen concentration would be approximately 1.9 × 10⁻⁶ A/cm² at δ = 500 μm and 9.6 × 10⁻⁶ A/cm² at δ = 100 μm. This is consistent with i_CMEA (Figure 14). The question arises as to how fast these ORR rates deplete the available oxygen. Oldfield and Sutton developed mathematical models for crevice corrosion.⁶⁰  The first step during this process is O₂ depletion. Acidification occurs after anode-cathode separation which relies on the depletion step.

The time for O₂ depletion inside a crevice, t_depletion (s), is given by Equation (1):

where n is the number of electrons per mol for oxygen reduction (4 equiv/mol), F is Faraday’s constant (96,485 C/equiv), c is the bulk concentration of oxygen (∼2.5 × 10⁻⁵ mol/cm³), x is the average crevice gap (micrometers), and I_p is the passive current (μA/cm²). Furthermore, the time for oxygen diffusion, t_diffusion (s), into a crevice can be expressed by Equation (2):

where l_diff is the mean distance traveled by O₂ and D is the diffusion coefficient of O₂ in the crevice solution (2.1 × 10⁻⁵ cm²/s). Given these expressions, the time elapsed until oxygen was depleted if the gap was 100 μm was approximately 1 s and 4.8 s for the exposure with a larger gap of 500 μm for I_p = 10⁻³ A/cm². In a stagnant unstirred solution, (I_p = 10⁻⁵ A/cm²), the time elapsed before oxygen was depleted if the gap was 100 μm was approximately 96.4 s and 482 s for the exposure with a larger gap of 500 μm. Considering a stagnant unstirred solution where I_p = 10⁻⁶ A/cm², the time elapsed before oxygen was depleted if the gap was 100 μm was approximately 964.9 s and 1.3 h for the exposure with a larger gap of 500 μm. Moreover, the time for oxygen diffusion to the very bottom of the crevice (15 mm from the mouth of the crevice), assuming depletion at the bottom, was found to be 29.7 h, which is longer than the time of exposure. The time for oxygen to diffuse nearly halfway through the crevice to A11, after the corner of the crevice (6 mm from the start of the crevice) was 4.6 h. Lastly, the time for oxygen to diffuse from the mouth of crevice (2 mm from the mouth of the crevice) was 0.5 h. This type of analysis demonstrated that oxygen was likely depleted deep inside the crevice but available near the mouth. These results are summarized in Table 5.

This further validates the notion that ORR was likely not the main cathodic reaction occurring on either material inside the fastener hole, albeit the likely dominant reaction near the mouth. Without ORR the stainless steel cathode can easily support HER if it is assumed that it is polarized to the critical potential for local corrosion of AA7050 and if anode/cathode separation is at least partially attained such that some acidification is experienced as a result of corrosion of AA7050 near the stainless steel. It is noted that the AA7050 itself can support significant hydrogen evolution on replated copper (Figure 12). It was shown that for anode to cathode ratio 1∶1, the pH of a corrosion cell with Al must be above 6, where at pH 6 the reversible hydrogen potential is E = −0.59 V_SCE.⁶¹  However, inside a fissure, more anode area is available, decreasing the pH to approximately 4, where the reversible hydrogen potential is E = –0.48 V_SCE. The critical potential for local corrosion of AA7050 is −0.66 V, which is below the reversible hydrogen potential at this pH.^51,62  Thus Al, replated copper on Al, or any stainless steel surface polarized to this potential, which is likely given the nonpolarizable nature of the local corrosion site, likely can support HER during galvanic coupling.

Cathodic polarization data for stainless steel at pH 4 and 11 further supports the notion of galvanic corrosion supported by HER.^12,51  The critical potential of AA7050 (E = −0.66 V_SCE) is below the reversible hydrogen potential in the case of pH 4 but not in the case of pH 11. The CMEA cathodic current densities observed on stainless steel are 10⁻⁵ A/cm² and 10⁻⁷ A/cm², respectively, which can certainly be supported based on inspection of polarization data.^12,51 

Furthermore, the Al³⁺ concentration increases within the crevice as AA7050-T7451 undergoes dissolution. Therefore, the higher magnitude in current at the bottom of the fastener can be speculated to be caused by an increase in Al³⁺, which has been shown to increase HER kinetics.⁶³  Acidification in the crevice, resulting from Al³⁺ hydrolysis, leads to chloride ions moving into the crevice to preserve electroneutrality. This increases local currents potentially enough to enable HER on stainless steel. Therefore, a local occluded environment with low pH may be created at the bottom of the crevice, with HER occurring locally on Cu-replated AA7050 near or within the fissures. The stainless steel must also undergo HER as 10⁻⁵ A/cm² at faster rates than the ORR rate can support given O₂ depletion (0.1 ppm O₂). The HER cathodic reactions generate one hydroxyl per electron released by the cation, independent of the pH. The lack of anode and cathode separation limits the drop in pH. In the case of a crevice, Al³⁺ would accumulate in the confined space if the transport rate out is slower than the production rate resulting from dissolution. Assuming a 70 μm boundary layer thickness, the limiting current density for ORR with no oxygen depletion (8 ppm O₂) was 1.3 × 10⁻⁵ A/cm², while it was 1.7 × 10⁻⁷ A/cm² when oxygen was depleted (0.1 ppm O₂). This level of ORR cannot be supported deep inside the crevice. The magnitude of current at the bottom of the fastener in Figure 13 was of the order of 10⁻⁵ A/cm² for the stainless steel. Therefore, it can be further speculated based on these current values, which are higher than possible by mass transport limited ORR, that HER is supported deep inside the crevice and controlling the cathodic reaction rates during galvanic attack deep in crevices. Inspecting a polarization scan on stainless steel confirms that HER occurs near the breakdown potential expected on AA7050.⁵¹ 

The fastener CMEA corrosion was also exposed under a thin film with a larger gap between the AA7050 and the stainless steel. For all other exposures, the gap between AA7050-T7451 and Type 316 stainless was approximately 100 μm. With a larger gap, 500 μm the charge density was 5.6 C/cm². This was 1.5 C/cm² less than the exposure with a gap of 100 μm. It is speculated that this can be attributed to the increased separation between the AA7050 and stainless steel.

The impact of the fastener geometry was investigated by comparing the flat planar galvanic CMEA exposure to the fastener CMEA exposures. After 10,000 s in the planar geometry under a 70 μm thin film at 98% RH the charge density was 0.7 C/cm² (Figure 6), under the exact same conditions, the fastener geometry yielded 7.1 C/cm² (Figure 13). This demonstrated that the fastener geometry contributed to an increase in corrosion rate by almost 10 times. This further demonstrated that the geometry of a fastener is inherently more susceptible to corrosion as a result of higher magnitude currents at the mouth of the fastener and the local chemistry developed inside the crevice as discussed above.

Under cyclic wet/dry conditions the corrosion rate approximately doubled compared to thin film exposures at constant RH. Figures 8 and 15 show the RH cycle CMEA exposure for the geometric flat CMEA and the CMEA arranged in a faster arrangement. The equilibrium concentration of 0.6 M NaCl varies with ambient RH.³⁶  After drying during subsequent wetting, the NaCl on the surface of the CMEA deliquesces over time and forms an electrolyte layer or droplet that may achieve equilibrium salt concentrations reported. Figure 21 shows the salt concentration and water layer thickness as a function of one wet to dry cycle used in this work. This calculation was conducted with experimental RH data, hence showing noise in the data. As the RH in an exposure environment changes, the calculated equilibrium salt concentration and water layer thickness changed as well. As the RH decreases, the equilibrium salt concentration becomes more concentrated as the water layer thickness decreases. This increases the galvanic current through enhanced ionic conductivity and decreases ORR diffusion length even though O₂ solubility is lower in concentrated salts. NaCl deliquesces at 75% RH; therefore below 75% RH, no galvanic coupling should occur.⁶⁴  However, Schindelholtz showed evidence that corrosion may occur down to 33% RH for NaCl, further extending the acceleration of the corrosion rate under atmospheric conditions.⁵⁸  In this work, cycling increases the corrosion rate by as much as an order of magnitude. During the wet to dry cycle, it was observed in Figure 21 that the salt concentration was as high at 6.3 M, assuming NaCl deliquesces at 75% RH. This concentration could be even higher if the droplet remained liquid to 33% RH. Throughout the exposure, the resulting theoretical water layer thickness varied from 1 μm to 80 μm. The combination of the increase in salt concentration and resulting thin electrolyte layers led to increased corrosion rates as observed elsewhere.

• •

  Two galvanic dissimilar metal CMEAs were constructed with AA7050-T7451 and Type 316 stainless steel electrodes to determine galvanic current interaction under atmospheric conditions such as static thin films and wet/dry cycling. Galvanic corrosion anodic charge was increased by at least one order of magnitude under a 70 μm thin film of 0.6 M NaCl when compared to full immersion utilizing a simple flat geometric CMEA. Furthermore, a water layer thickness of 500 μm reduced galvanic corrosion kinetics relative to the 70 μm thick electrolyte film under a flat CMEA but galvanic attack was greater than in the case of full immersion.

• •

  Wet/dry cycling resulted in an increase in galvanic charge density relative to a fixed static film thickness and RH. Anodic charge increased by a factor of three in 24 h on a flat CMEA. During the wet/dry cycle, sharp current increases were observed upon the onset of wetting attributed to the high Cl⁻ concentration and the thin electrolyte layer thickness of the incipient droplet.

• •

  The CMEA arranged in a fastener geometry showed higher currents at the mouth of the crevice where Type 316 and AA7050 were geometrically close and O₂ was available, and also at the bottom of the crevice. Galvanic corrosion at the mouth was attributed to oxygen being readily available to enable ORR on Type 316 stainless at the mouth. Corrosion deep in the dissimilar metal crevice was attributed to partial acidification within the crevice and HER where O₂ was likely depleted. The net anodic charge associated with galvanic corrosion was 3.5 times greater on the fastener geometry than the flat geometric CMEA exposure under the same conditions.

• •

  Galvanic corrosion in AA7050-T7451 was shown to occur at isolated anode sites controlled by the heterogeneous microstructure of this alloy. Galvanic corrosion initiated at locations that were not always near the stainless steel. In many of the CMEA exposures, AA7050 anodic sites were often found to abruptly switch to cathodic sites. This suggested that that changes in electrochemistry on single electrodes controlled by dealloyed S-phase (Al₂CuMg) and/or copper replating on the surface were the cause of the increased cathodic kinetics of AA7050-T7451 electrodes. It was observed that when coupled to stainless steel, some of the AA7050-T7451 electrodes participated as significant cathodes supporting ORR. In some cases, AA7050-T7451 electrodes contributed 90% of the total net cathodic reaction. This enabled the support and growth of corrosion fissures in AA7050-T7451.^12-13,17 

This work was supported by the United States Office of Naval Research (N00014-14-1-0012) with W. Nickerson as the program monitor.