3.3.1. Electrochemical Impedance Spectroscopy
Electrochemical impedance spectroscopy (EIS) has been extensively used in the investigation of the metal-electrolyte interface, surface responses, oxides films formation, passivation, corrosion kinetics and mechanism, and coating protective effectiveness to metallic substrates [
37,
38]. Thus, in order to investigate the effects of FSP parameters on the kinetics and characteristics of the electrochemical process, EIS measurements were carried out on the base sample and all processed samples.
Figure 7a–d displays the Nyquist curves, phase angle, and modulus of the base sample and all processed samples. The Nyquist plots depicted in
Figure 7a,b show a relatively large incomplete semicircle at the high-frequency region, while the low-frequency region demonstrates a diffusion tail. The larger incomplete semicircle at the high-frequency region in the Nyquist plot of processed samples may be ascribed to a charge transfer reaction at the boundary between the thin film formed on the aluminum surface and the NaCl solution [
39]. However, the occurrence of the small tail at the low-frequency region of the spectra is attributed to diffusion-controlled processes [
40]. The depression phenomenon under the x-axis can be clearly identified through the incomplete semicircle behavior of the arcs, which is probably caused by surface heterogeneities resulting from possible corrosion products and frequency dispersion [
41]. As can be observed in
Figure 7a, the arc diameter rises with increasing traverse speed. In addition, the smallest diameter was observed for the base sample, while the sample processed with the highest traverse speed (FP-25) demonstrates the highest semicircle diameter. Moreover, the variation in the semicircle diameter between the base sample and the sample processed at the lowest traverse speed (FP-10) was remarkably low compared to the deviation in semicircle diameter between the sample processed at moderate velocity (FP-15) and that fabricated at the highest velocity (FP-25). Interestingly, the variance in semicircle diameter between the base sample and (FP-25) sample was not even distinguishable without zooming the base sample area in the Nyquist plot, as seen in
Figure 7b. A huge increment in the semicircle diameters revealed the significant impact of process conditions on corrosion behavior. Furthermore, the increase in the radius of the semicircle with the processing can be interpreted by an enhancement in the surface protection due to the rising stability and compactness of the passive film formed [
42,
43]. Therefore, the enhancement in the corrosion resistance was found to follow the rank of FP-10 < FP-15 < FP-25, and demonstrated a markedly improved resistance compared to that of the unprocessed sample.
The bode plots for all samples are presented in
Figure 7c. It is evident from the figure that the total impedance for all the processed samples is remarkably higher than that of the base sample, which reflects the increased modulus impedance of the processed samples. In addition, the total impedance revealed a significant increment by increasing the traverse speed. Moreover, the deviation in the total impedance was at its lowest when comparing the base material with the sample processed at the lowest traverse speed; however, with increasing traverse speed, the deviation between samples processed at the intermediate and lowest speed was much higher. Furthermore, a huge variation was observed between the intermediate-speed sample (FP-15) and the sample processed at the highest speed (FP-25), which indicates a superior corrosion resistance for the sample processed at 250 mm/min compared to all other samples.
Figure 7d displays the typical Phase angle-frequency plots for both the processed and base samples. The frequency of the phase angle maxima of the base sample is 25 Hz at −78°. In contrast, the (FP-10) sample demonstrated two time constants, the first shifted toward a frequency of 6.4 Hz at approximately −71° and the second shifted toward a frequency of 241 Hz at −69°. Additionally, samples (FP-15) and (FP-25) exhibited phase angle maxima that are shifted toward frequencies of 8 and 0.78 Hz at −81 and −83°, respectively. It is worth noting that observing the phase angle maxima for the base sample at higher frequency suggests weakness in the protective barrier [
44]. Moreover, it can be noticed that the base sample demonstrated a single time constant and upon increasing the traverse speed two time-constant characteristics can be observed. Similar behavior was also observed in [
45]. Furthermore, more peak broadening indicates an enhancement in the passive protective barrier over a wide frequency range and a consequent corrosion resistance improvement [
44,
46].
All extracted impedance curves were fitted to appropriate corresponding circuits for further numerical estimation of the barrier characteristics of processed surfaces and properties of active corrosion protection. Accordingly, two equivalent circuits were proposed; the first one (Circuit a) was used to stimulate the corrosion behavior at the base surface as shown in
Figure 8a. This circuit included a charge transfer resistance (R
ct) that serially connected to a Warburg element (W) and was in parallel to a constant phase element for double-layer capacitance (CPE
dl), where both (W, R
ct//CPE
dl) components were connected in series with the solution resistance (R
S) component between the reference and working electrodes. The second circuit illustrated in
Figure 8b (Circuit b) was used to simulate all the electrochemical processes at the surface of processed samples and contained an R
ct that linked in series to a Warburg element (W) and both components were in parallel with a constant phase element accounting for the double-layer capacitance of the inner barrier layer (CPE
dl), and the aforementioned three components were connected in series with a film resistance (R
f) component. Further, W, R
ct, CPE
dl, and R
f were connected in parallel with the constant phase element of the outer passive film (CPE
f), and the previous circuit was connected in series with the solution resistance between the reference and working electrodes (R
S).
The charge transfer resistance (R
ct) depicted in
Table 3 exhibited a rapid increment for all processed samples over the base material. In addition, it is evident from the fitted results that the charge transfer resistance increases with increasing traverse speed, which can be ascribed to the passive film formed at the metal/electrolyte interface [
47]. The presence of this film isolated the aluminum metal surface from the corrosive ions attacks and thereby obstructing any further transfer of charge or mass. It should be noted that there is a remarkable difference between the charge transfer resistance (R
ct) and film resistance (R
f) particularly if the processing was carried out at high speeds and this may be attributed to the key role of the formed oxide film on the metal surface in mitigating the corrosion process at a particular speed [
48]. Moreover, the huge variation between the outer and inner resistances indicates that the resistance provided by the inner barrier is much higher when compared with the outer film [
49]. Generally, the Warburg element is added to simulate the diffusion effect as indicated by the straight line of a slope close to 45° to the impedance real axis. This can be observed with the unprocessed sample and gradually reduces for samples processed at 100 and 150 mm/min as seen in
Figure 7a,b. It is expected that a further decrease will be experienced at the highest traverse speed sample i.e., FP-25 if the test was allowed for a much lower frequency. Considering the Warburg impedance, it is obvious from the table that the lowest Warburg impedance was obtained at the highest traverse speed of 250 mm/min and, by reducing the traverse speed to 150 mm/min, the Warburg impedance demonstrated a slight increment. Moreover, with a further reduction in the traverse speed down to 100 mm/min, a considerable increase in Warburg impedance was observed, which is relatively comparable with the base metal. A lower Warburg value indicates a reduction in the diffusion of chloride ions through an oxide passive film and thereby improving the corrosion resistance [
50,
51].
It is worth noting that the constant phase element was utilized as an alternative to pure capacitance because the non-ideal capacitive behavior indicated by the deviation in phase shifts from 90°, which might be attributed to a heterogeneity property on the surface of the metal. Additionally, several models have been developed to correlate the passive film capacitance knowing the constant phase element component (admittance). Hsu and Mansfeld introduced a model (Equation (3)) to show the relationship between capacitance and admittance [
52].
where,
is the frequency when the imaginary part of the impedance is at its max value, Y
0 is the admittance of the CPE
dl, and n is the value corresponding to the surface roughness.
It can be seen from
Table 4 that the double-layer capacitance (C
dl) exhibits a significant reduction for all processed samples when compared to the unprocessed sample and the values are in the order of (FP-15) > (FP-10) > (FP-25). The electrochemical behavior between the charged aluminum surface and 3.5 wt.% NaCl is reflecting an electrical dual-layer capacitance behavior [
53]. Accordingly, the reduction in the double-layer capacitance of the processed surfaces indicates a corrosion resistance enhancement with increasing traverse speed, which can be attributed to a rapid decrease in surface activeness resulting from the formation of a thicker passive film [
54,
55]. Accordingly, the thickness of the passive film (t) can be calculated using the Helmholtz model [
56] (Equation (4)).
where C
dl is the double-layer capacitance (µF),
is the vacuum permittivity (8.85 × 10
−14 F cm
−1 [
57]), and
is the passive layer dielectric constant (for aluminum,
) [
57].
3.3.2. Potentiodynamic Polarization (PDP)
The typical PDP plots of all samples after exposure to the NaCl electrolyte are presented in
Figure 9. Various electrochemical parameters, such as the anodic Tafel slope (ꞵ
a), cathodic Tafel slope (ꞵ
c), corrosion current density (I
corr), and corrosion potential (E
corr) for the base sample as well as the processed samples, were derived from the potentiodynamic polarization curves and are listed in
Table 5. The polarization curves exhibited significant differences between all samples. As can be noted from the data in
Table 5, there is a remarkable difference in corrosion potential between the base and processed samples. It is clear from
Figure 9 and
Table 5 that the corrosion potential (E
corr) is shifted to a more positive noble potential with increasing tool traverse speed. Moreover, the FP-10 sample exhibited the lowest corrosion potential i.e the most negative, whereas the FP-25 sample showed the noblest behavior among all processed and base samples. This shift to a noble direction is an indication of an improvement in corrosion resistance. Further, as can be inferred from
Figure 9 and
Table 5, the corrosion current densities (I
corr) of all processed samples were markedly lower than that of the unprocessed alloy. Additionally, a further reduction in I
corr can be observed as the traverse speed increases. The reduction in I
corr. indicates that the friction stir processing for 6061 AA has enhanced its corrosion resistance. Moreover, the corrosion current density for the base sample is approximately 600 times higher than that of the FP-25 sample, which indicates the outstanding corrosion protectiveness of the FP-25 sample. Also, as compared to a base sample, the pitting potential of the FP-25 sample raised from −780 mV to −600 mV, which suggested an enhancement in the pitting resistance. Furthermore, the pitting resistance of the other two processed samples is also high despite their lower corrosion potential. Enhancing pitting resistance was also reported after the FSP of AA 7075 [
58].
A remarkable variation in the cathodic and anodic Tafel slopes can be detected through the inspection of
Table 5. In particular, the anodic behavior revealed that the unprocessed sample had less tendency to be oxidized compared to all processed samples, where ꞵ
a value for the unprocessed sample was the highest, and a further reduction in the anodic slope can be noticed with increasing tool traverse speed, which suggests an enhancement in the oxidation tendency [
59]. In addition, the cathodic Tafel slope ꞵ
c revealed a remarkable reduction upon deploying FSP. Moreover, the cathodic slope is further reduced with increasing traverse speed, which suggests a diffusion of oxygen molecules to form (OH
−) in the processed samples [
60].
3.3.3. Linear Polarization Resistance (LPR)
Linear polarization plots for different samples (processed and base) in NaCl solution are presented in
Figure 10. The polarization resistance (R
P) was obtained as a slope of potential versus current. To calculate the corrosion current density based on the linear polarization resistance approach, the Stern–Geary equation was utilized, which correlates the current density (I
corr) with polarization resistance (R
P) (Equation (5)) [
61].
where R
p is the polarization resistance obtained from the LPR slope, and
and
are the cathodic and anodic Tafel slopes derived from PDP curves, respectively.
Subsequently, to calculate the corrosion rate (mpy), Equation (6) was used:
where ρ is the sample density and EW is the sample equivalent weight.
As indicated in
Table 6, the corrosion current density of the unprocessed sample is significantly higher than that of all processed samples. Additionally, the highest polarization resistance was obtained for the sample processed at the highest speed and a further reduction can be observed with a reduction in the traverse speed. It should be emphasized here that the polarization resistance is directly related to the corrosion resistance, therefore, a sample with the highest polarization resistance exhibited the highest corrosion resistance.
Results obtained from the different electrochemical techniques (LPR, EIS, and PDP) demonstrate a strong agreement. In addition, the base sample shows the highest corrosion susceptibility, while, after surface processing by employing FSP, a further increment in the corrosion resistance was observed on increasing the tool traverse speed. It should be emphasized that many studies have indicated a direct relationship between grain size, the passive film formed, and the corrosion resistance of aluminum-based alloys [
62,
63]. Due to the unique properties of grain boundaries compared to bulk materials, such as diffusion rates, atomic coordination, and reactivity, it is expected that a reduction in grain size will result in a substantial change in electrochemical behavior [
63]. Additionally, breaking the intermetallic phases is another consequence of grain refinement [
64]. These intermetallic compounds have a cathodic nature relative to the aluminum matrix and therefore their refinement and dissolution will remarkably improve the corrosion resistance. Ralston et al. [
62] reported that a surface with a higher density of grain boundaries and finer grains is more likely to attract an intact passive protective film. Jilani et al. [
65] have also studied the impact of grain refinement on the corrosion resistance of 1-XXX aluminum alloy. Their results confirmed that grain size reduction along with precipitate redistribution markedly improved the corrosion resistance by forming a continuous protective passive film. Likewise, the influence of severe plastic deformation of 2099 Al-Li on the formation of passive film was studied by Jinlong et al. [
66]. The results illustrated that the corrosion susceptibility reduces due to refinement in both grain and precipitates, where grain refinement leads to electron work function decline and, thus, a thicker passive film.
Generally, various studies in the literature have shown that FSP resulted in an increase in the fraction of high-angle grain boundaries [
67,
68,
69]. For instance, Li et al. [
69] studied the impact of heat input during FSP on Mg-Li alloy on the fraction of high-angle grain boundaries. Their results indicated that the reduction in grain size and the increase in the fraction of high-angle grain boundaries are associated with lower heat input.
The consequence of this shift to a higher fraction of high-angle grain boundaries (HAGBs) on corrosion behavior was intensively investigated by Argade et al. [
70], Dan et al. [
71], and Rao et al. [
60]. For example, Argade et al. [
70] have concluded that the corrosion enhancement after friction stir processing of AA 5083 was attributed to grain refinement, which improved the polarization resistance, passivation, and pitting potential. Moreover, high-angle grain boundaries provided by the process raise the corrosion resistance by accelerating the passivation re-passivation phenomenon. Dan et al. [
5] have studied the impact of grain refinement on the corrosion behavior of pure aluminum and have reported a significant enhancement in corrosion and pitting resistance for the sample with finer grain, which was attributed to a denser passive film. Further, it was also suggested that the oxide film favored a higher grain boundary, which protects the surface of the processed sample against chloride attack, thereby enhancing the corrosion resistance. Rao et al. [
60] have demonstrated that the transformation from low-angle grain boundaries to high-angle counterparts after friction stir processing of Al–30Si alloy is a fundamental reason for the stability of the oxide film formed. Therefore, the results obtained in the current study may follow the same behavior, whereby employing FSP significantly diminishes grain size and raises the fraction of high-angle grain boundaries. Moreover, a further reduction in grain size was detected with increasing tool traveling speed, which resulted in superior corrosion resistance due to the more adhered and compacted passivation film.