**3. Results**

#### *3.1. Metallographic Analysis*

The microstructure of the AA 7075 T6 alloy has already been presented in [26]. Coarse precipitates along the grain boundaries (Figure 4) and micrometric precipitates within the grains of α-Al, oriented in the rolling direction [28], can be well demonstrated. The strengthening MgZn2 nanoprecipitates are not visible under the optic or SEM microscope, while macro-precipitates are copper and iron-rich phases as Al2CuMg and Al7Cu2Fe or zinc and magnesium rich phases, i.e., undissolved MgZn2 during alloy solution annealing treatment, or MgSi.

**Figure 4.** Optical image (Keller attack) of the alloy AA 7075 T6 (base material). Elongated grains along the rolling direction and the macro-precipitates are visible.

In the nugget, re-crystalized fine equiaxed grain structure can be noticed. In addition, MgZn2 strengthening precipitates are dissolved (T6 artificial aging temper) into the supersaturated solid solution and reprecipitated after cooling (Figure 5). Venugopol et al. affirmed that the temperature obtained during the process is below the melting temperature of the alloy, but above the solutioning temperature, as demonstrated by the re-crystallization of the weld nugge<sup>t</sup> and the redistribution of the precipitates [21]. For the same authors, the absence of fine precipitates in the weld nugge<sup>t</sup> indicates that the cooling rates are such that larger precipitates could nucleate and grow but not the finer ones. In the TMAZ, the precipitates are quite random, and the coarsening of finer precipitates observed in base material can occur during welding [21].

**Figure 5.** (**a**) SEM image of the nugge<sup>t</sup> of the FWS 7075 alloy (no metallographic attack) the rounded recrystallized grains are visible owing to the precipitates of MgZn2 on their grain borders and EDX spectra in which is visible an enrichment of zinc and copper in the correspondence of the precipitates. In the image the grain border is demonstrated; (**b**) optical image of the nugge<sup>t</sup> with metallographic attack (Keller's reagent) that better demonstrated the rounded recrystallized grains.

The size of the precipitates at the grain boundaries of the HAZ (Figure 6) is unfortunately too small to permit their observation under the SEM, but many transmission electron microscope (TEM) observations were reported in the literature and they demonstrated that these smaller precipitates play a fundamental role in the intergranular corrosion and the SCC of these alloy series [18].

**Figure 6.** Optical image (Keller attack) of the HAZ of the FWS AA 7075 alloy. The grains appear elongated along the rolling direction; presence of macro-precipitates is evident.

#### *3.2. 4PBB Specimens*

The OCP measurements (average values of four specimens) did not show any evident differences between 4PBB loaded and unloaded specimens (Figure 7). The corrosion potentials generally were in the range of −870–−840 mV vs SCE; the corrosion potential showed a sharp decrease during the solution refresh owing to the partial removal of non-adherent scale of corrosion products that exposes the corroded surface on the fresh solution; after this, the corrosion potential increases to the same potential values detected prior to the partial substitution of the testing solution. This effect was more marked for specimens without loading compared to loaded specimens. The corrosion potential of the loaded specimens decreases after 1300 h, then it increases again. Compared to the values obtained by the analysis of the three different zones (nugget, TMAZ/HAZ and base materials—Figure 8) similar results can be outlined in both the TMAZ/HAZ and the nugget. The base metal showed slightly nobler potential values.

**Figure 7.** Open circuit potential (OCP) vs time for loaded and unloaded 4PBB specimens. The decreasing of the OCP of the specimens in the correspondence of the partial refresh of the test solution that breaks the aluminum oxide scale and exposes fresh metal surface to the aggressive solution.

**Figure 8.** OCP vs time of the different zones (base materials, TMAZ and nugget) of the alloy AA7075 T6.

Figure 9a shows the corrosion morphology of the specimens at the end of the 4PBB tests. No differences in the corrosion morphology were observed between the loaded and unloaded specimens. The HAZ is preferentially corroded. The attack propagates along the grain boundaries, following the rolling direction (Figure 9b). The EDX analysis (Figure 9c) demonstrated the depletion of zinc and copper in the corroded zone (Spectrum 1) with respect to the un-corroded zones (Spectrum 2).

**Figure 9.** (**a**) Image of a 4PBB specimens after the test: an intense localized attack in the correspondence of the TMAZ and (**b**) SEM image of the metallographic longitudinal section of the loaded specimen in which is visible the deviation of the attack along the rolling direction, (**c**) particular of the attack and EDX analysis.

(**c**)

#### *3.3. Constant Loading (CL) Tests*

Figure 10 shows the effect of the applied load on the OCP of the CL specimen. The OCP of the unloaded specimens rapidly stabilized between −850 and −800 mV vs SCE. The fluctuations are mainly ascribable to the removal of corrosion products of aluminum formed at very early exposures, due to the recirculation of the testing solution. When the specimen is loaded in elastic field, a decrease in the OCP was noticed but the potential values stayed in the same range measured in unloaded condition. Conversely, at strain level exceeding the yield stress—i.e., in the plastic field—100 mV decrease in the OCP occurred due both to the rupture of the thick corrosion product scale of aluminum and the plastic straining exposes the very active metal to the aggressive environment. The OCP came back to the initial value after about 24 h. This value remained almost constant, until the end of the tests. The specimen did not break during the tests but showed a corrosion attack mainly localized at the HAZ (Figure 11). Differently from the 4PBB specimens, the attack appears wider, and positioned perpendicularly to the loading direction coinciding with the rolling direction (Figures 11c and 12a). Some small ramifications along the grain boundaries were observed (Figure 12b).

**Figure 10.** Open circuit potential measurements of the specimen loaded at different values during CL test. The decrease in the OCP mainly occurs as loading increase due to aluminum corrosion products cracking.

#### *3.4. Slow Strain-Rate (SSR) Tests*

Figure 13 reports the load and OCP measurements performed during the SSR tests of the AA 7075-T6 base materials and the FSW joints. The tests were twice repeated, but only one curve for each condition is reported in the graph for simplicity. The tensile curves of the base material and the weld obtained in air or in NaCl 0.6 M are practically overlapped. The time to failure of the base material is close to the mean value with a very small deviation of 7%. 25% deviation was observed for the welded specimens. No stress-corrosion crack occurrence was noticed for the base metal and the time to failure in 0.6 M NaCl solution is even longer compared to the value measured at air. The role of active corrosion in NaCl solution can be hypothesized to enhance dislocation mobility and thus the plastic deformation, as reported by Jones et al. [28,29]. The authors named this phenomenon anodic attenuation of strain hardening.

**Figure 11.** Aspect of the specimen at the end of the constant load test. (**a**) Macro image: the dark zone in the central part exposed to the NaCl 0.6 M solution; (**b**) close-up of the red circle with the selective attack of the TMAZ; (**c**) metallographic section of localized attack.

**Figure 12.** (**a**) Metallographic section of the enlarged localized attack in the TMAZ zone of the CL specimen. Close-up of the arrow in (**b**) in which is evident the change the morphology of attack that allow the rolling direction.

Figure 14 compares the fracture surface of the SSR specimens after the test in air and in 0.6 M NaCl solution. The fracture surface at air (Figure 14a) showed a shearing failure, typical of the prismatic specimens, while the fracture surface of the specimen after the SSR test in 0.6 M NaCl solution exhibits an initial flat area (Figure 14b) due to the presence of microdefects able to trigger localized corrosion initiation. These microcracks are mainly at macro-precipitates (black arrows in Figure 15a) and have

depth less than 100 μm. In the correspondence of these microcracks, the fracture surface is heavily corroded (Figure 15b) and it is possible to observe the presence of several micro-precipitates. The flat zone of the fracture surface (Figure 16b) is mixed brittle/ductile as some small dimples and quasi cleavage areas were observed, similarly to the fracture surface at air (Figure 16a). The presence of very small zones with typical SCC morphology, indicated with the arrows in Figure 16b [29] was also confirmed; the final shearing fracture of these specimens was similar to that in air.

No macroscopic SCC phenomena were observed probably due to the strain rate (10−<sup>6</sup> s<sup>−</sup>1) value adopted and very aggressive NaCl solution. Results presented in other works [30] confirmed the absence of stress corrosion on this alloy during the SSR tests at the OCP and strain rate of 10−<sup>6</sup> s<sup>−</sup>1. The OCP remains constant in the elastic field, but it increases with the plastic deformation, probably due to the enrichment in iron and copper due to the dissolution of the surrounding aluminum matrix. This observation is supported by back-scattered electron (BSE) image and EDX spectrum (Figure 17).

**Figure 13.** Stress vs time and OCP of the specimens during the slow strain-rate tests.

**Figure 14.** SEM image of the fracture surface of the AA7075 T6 specimen after the SSR test (**a**) macro at air, (**b**) macro in 0.6 M NaCl solution.

**Figure 15.** SEM image of the fracture surface of the AA7075 T6 specimen after the SSR test (**a**) specimen surface with corrosion products of aluminum and several microcracks at macro-precipitates (black arrows); (**b**) particular of fracture initiation zone after SSR test in 0.6 M NaCl solution.

**Figure 16.** SEM image of the fracture surface of the AA7075 T6 specimen after the SSR test: (**a**) particular of the fracture surface at air and (**b**) in 0.6M NaCl solution (the arrows indicate small areas with typical SCC morphology).

**Figure 17.** (**a**) Backscatter electron detector (BSD) image of the external surface of a base material specimen after the SSR test in NaCl 0.6 M, the bright zone indicates precipitates with higher atomic weight than the aluminum matrix; (**b**) EDX spectrum of one of these particles.

The curves of the FSW joints exhibit time-to-failure values lower than the base material. This can be mainly ascribable to the decrease in the tensile properties at the joint TMAZ/HAZ zone, as observed also elsewhere [26,31–33]. Consequently, the hardness and then the UTS of the alloy decrease due to microstructural modifications [34,35].

Figure 18a reports the hardness profile of the longitudinal section of the weld. The hardness sharply decreases at the TMAZ/HAZ and, therefore, the plastic deformation mainly occurs at this point, thus decreasing the total elongation at break and the time of failure. Despite the high strain rate due to plastic strain localization at TMAZ/HAZ, slight decrease in the time to failure in 0.6 M NaCl solution was noticed compared to air—i.e., 10 ± 0.5 h—thus denoting SCC occurrence. Bala Srinivasana et al. reported about SCC phenomena in the HAZ of the AA7075 of mixed FWS joints of AA7075 and AA6056, when the specimens were tested at strain rate 10−<sup>7</sup> s<sup>−</sup><sup>1</sup> in NaCl solutions, but only localized corrosion in the tests carried out at 10−<sup>6</sup> s<sup>−</sup><sup>1</sup> was detected [36]. The SEM observation of the surfaces demonstrate a flat zone at fracture initiation with maximum depth of 1 mm (Figure 18b,c). The presence of non-coherent corrosion products of aluminum and several precipitates was also noticed (Figure 19a). EDX analysis detected the presence of zinc, aluminum, and copper, as well as magnesium depletion (Figure 20). The final fracture is by shearing (Figure 18b) and its morphology is characterized by elongated dimples and a few brittle areas Figure 19b), similarly to the base material (Figure 16).

**Figure 18.** (**a**) Hardness profile along the FSW butt joint; (**b**) SEM image of the fracture surface, a flat zone of about 1 mm of growth (close-up in (**c**) is highlighted in (**d**)) metallographic section of the fracture surface in which is visible the flat zone and the small ramifications.

**Figure 19.** (**a**) Particular of the mixed ductile/brittle fracture zone and (**b**) of the ductile shearing final rupture.

**Figure 20.** EDS spectrum of the precipitates present on the mixed ductile/brittle zone of the fracture surface of the FWS specimen after the SSR test in 0.6 M NaCl solution.

The OCP of the FSW butt joints lies in the range of −820–−810 mV vs SCE and it is less noble than the base material, whose potential was in the range between −725 and −670 mV vs SCE. The OCP of the FSW alloy decreases during the elastic deformation. Conversely, constant values were detected once the specimen was strained in the plastic field. The values of the OCP of the FSW specimens during SSR tests are close to that measured during 4PBB tests and CL tests (Figures 7 and 10), corresponding to the OCP of the TMAZ/HAZ and the nugge<sup>t</sup> (Figure 8). The very low OCP potential indicates that these zones are more susceptible to corrosion than the base material, as confirmed by the intense attack observed on the specimens at the end of the test.
