3.1. Lap Shear Strength (LSS)
The values obtained after the lap shear strength test are presented in
Table 3. The experiments conducted with the aluminum samples treated by PEO (01 to 17) show average lap shear strength values ranging from 5.7 to 9.5 MPa. Considering the lap shear values, it is observed that the PEO treatment resulted in increases of 2.5 to 4.3× compared to the set without surface treatment. This is due to the costing’s morphological characteristics, such as micro-pores, protrusions, and other microstructures that appear in the oxide coating, providing a larger surface area for the sample and allowing greater interaction with the thermoplastic matrix [
1,
2,
21,
26,
27,
28].
The obtained lap shear values are promising compared to other studies conducted on both metallic and polymeric materials.
Table 4 presents some comparisons of lap shear values and the materials used.
With the development of a statistical model, the significance of parameters (independent variables) was evaluated using analysis of variance (ANOVA). This model was adjusted by the coefficient of determination (R
2), which quantitatively helps assess how much each variable influences the response variable (LSS) [
1,
36,
37].
The ANOVA analysis results, using the independent variables immersion time, duty cycle, and electrolyte concentration, to the dependent variable (LSS), are presented in
Table 5.
It was found that the F value (responsible for checking whether a term or the model itself is associated with the response variable) for the model was 2.03, implying that the adopted model is not significant in the response variable (LSS), and also showing that there is a 20% probability of the F value being due to noise. This study observed that the only factor significantly influencing the dependent factor is the electrolyte concentration (X3). As the “
p-value” is less than 0.05, it indicates that this factor has a significant effect on the response variable. Furthermore, when analyzing the coefficient of determination (R
2), the model explains 75.15% of the PEO + welding process variability. However, the statistical model did not demonstrate significance in explaining the variability, considering its
p-value (0.200), which is greater than 0.05 [
23,
29,
37].
It can be inferred that the model is not significant, due to the complex mechanisms of the PEO process added to the oxy-fuel welding process (which still requires optimization), as already presented in a previous study [
35,
38].
To evaluate the influence of each independent variable on the dependent variable, Equation (2), called effect size (η
2), can be used. This equation considers the sum of squares of factors, blocks, and/or the statistical model itself to quantify variability within a data set [
39].
where “SS” is the sum of squares of the specific factor, which quantifies the variability attributed to a specific factor within the statistical model, and “SST” is the total sum of squares of the model, which represents the total variability observed in the data, accounting for both the explanatory variables and the error [
39]. Based on Equation (2), the effect of each factor was 14.29% for immersion time, 5.11% for duty cycle, and 28.96% for concentration. It is observed that the pure model error influences only 1.16% of the total process variability.
3.3. Fractographic Evaluation of the Aluminum/Composite Interface
After obtaining the PEO process parameters and conducting the lap shear test, the fractured surfaces were visually characterized following the ASTM 5573-99 (2019) standard [
40] to analyze the failure modes at the interface.
Figure 5 shows the surfaces of the aluminum and thermoplastic composite samples, highlighting the joined regions. It is observed that there were mixed-mode fractures, with three different types. The yellow arrows indicate a small region with adhesive failure, as there was no polymer matrix in that region in the aluminum sample after the assembly rupture, representing 16.64%. The blue arrows indicate a polymer matrix’s cohesive failure, where the matrix remained on the aluminum sample, representing 34.02%. Finally, the red arrows indicate a light fiber failure, with the reinforcement of the composite exposed both outside the laminate sample and on the aluminum sample, representing 49.34% of the total fracture [
41]. It is also possible to observe the edge effect at the boundaries of the composite sample, indicating an excessive tightening in the assembly during the welding process, with the matrix reaching the glass transition temperature and flowing to the edges, a phenomenon previously observed in other studies [
19,
24,
30,
39].
In
Figure 6, the different stages of deformation in the samples during the lap shear test are shown, and on the right, a real image after fracture is observed. Noticeably, the aluminum sample exhibited considerable deformation after the test, which is attributed to the material condition. In this study, AA2024-O aluminum was used, where the “O” designation signifies annealed material, characterized by high ductility and significant plastic deformation (purple arrow in
Figure 6) [
42,
43,
44].
Although the PEI/glass fiber sample did not show deformation like the aluminum alloy, it exhibited an edge effect (green arrow), which is attributed to excessive tightening and the thermoplastic matrix softening phase, leading to flow towards the edges of the sample, as shown in
Figure 6 and observed in other studies [
19,
24,
30]. Studying this effect is crucial because it can result in delamination in the composite material, impacting the integrity of structures and/or substructures [
45,
46,
47].
3.4. Coating Corrosion Resistance
The data from the linear polarization measurements are presented in
Table 6, and the polarization curves of treated and untreated aluminum samples are shown in
Figure 7. In the figures, E
corr and j
corr represent the corrosion potential and corrosion current density, respectively. Additionally, E
pit indicates the potential where a breakdown of the anodic layer occurs and localized corrosion (pitting) begins. The corrosion rate (CR) in mm/year is presented in
Table 6 and is calculated using Equation (3).
where CR is the corrosion rate (mm/year); j
corr is the corrosion current density (A/cm
2); W is the equivalent mass of the material studied (according to ASTM G102: 2015 [
48]), and
is the specific mass of the material studied (g/cm
3).
In general, it is observed that the PEO treatment made the aluminum sample much more corrosion-resistant compared to the untreated sample, both in terms of nobility (corrosion potential, E
corr) and corrosion rate (current density, j
corr). A higher potential on the
y-axis indicates a greater nobility of material, with less interaction with the aggressive environment. Moreover, being more shifted to the left on the
x-axis (Log (i)) of the graph indicates higher corrosion resistance [
2,
49,
50].
The data from
Table 6 and
Figure 7 demonstrate that the PEO treatment significantly increased the corrosion resistance of the AA2024 aluminum alloy. The pitting potential (E
pit) indicates the oxide layer’s durability until its breakdown. In the untreated sample, the rupture points were subtle, around −0.55 V, with subsequent attempts at passivation. In contrast, the PEO-treated sample exhibited a more pronounced pitting potential at +0.85 V, characterized by a clear breakdown of the anodic curve and a sudden increase in corrosion current density [
2,
19,
51].
The PEO treatment on the AA2024-O alloy significantly increased its corrosion resistance, with a reduction in corrosion rate of approximately 99.84% compared to the untreated sample (Equation (4)).
This improvement in resistance is attributed to the formation of an oxide layer acting as a barrier, consistent with findings from other studies demonstrating that PEO treatments indeed decrease the corrosion rate in aluminum alloys [
52,
53].
The noise observed at the end of the polarization curve corrosion test may indicate an attempt at repassivation of the alloy, reflecting changes in the surface state as it tries to form a protective layer. However, the test interruption due to the equipment reaching its “overlay” suggests that the continuous scanning method may have introduced inaccuracies. Continuous sweeps can distort the half-sweep curve, affecting corrosion potential and current density. Furthermore, traditional analysis techniques can introduce errors, especially in small ranges of superpotentials. Therefore, although noise may suggest repassivation, it is crucial to consider experimental limitations when interpreting results [
54,
55].
Figure 8 shows scanning electron microscopy images of the AA2024 aluminum alloy: an untreated sample (a, b) and a sample treated with the PEO process in experiment 3 (c, d).
The treated sample (c, d) exhibits a surface with a compact morphology and few pores, featuring a thin coating with a thickness of 1.7 ± 0.5 µm. This is attributed to both the applied electrical regime, with a final voltage of around 320 V and a constant current in 60 mA, and the electrolyte used (Na
2B
4O
7·10H
2O), which allowed the development of a compact and thin layer. Other authors have observed that morphology strongly depends on the applied electrical potential; higher potentials result in thicker oxide coatings due to the dielectric layer breakdown [
2,
11,
56].
The PEO coating showed a low number of pores, which can be explained by the electron avalanche that occurs during the process. This phenomenon generates rapid ionization and an intense electrical discharge, leading to the uniform melting and solidification of the material, minimizing the formation of pores on the coating surface. Thus, the high energy density provided by the electron avalanche contributes to a more compact and homogeneous structure. This phenomenon causes localized temperatures to briefly reach between 10,000 K and 25,000 K, followed by rapid cooling with the aqueous solution [
8,
57].
It is observed that the untreated sample in
Figure 8a,b exhibits pores with distinct morphologies across the surface, indicated by the yellow arrows. Within these pores, microcracks are visible, which could potentially allow the substrate to be exposed to a chloride-rich environment, thereby increasing the corrosion rate, as illustrated in
Figure 9.
As already reported in the literature, Cl
− ions interact with AlOH functional groups, reacting to form soluble aluminum chloride, as shown in Equation (5). This process leads to an increase in surface acidity due to the hydrolysis of Al
3+ ions [
58,
59,
60].
The passive layer of the alloy allows Cl
− ions to penetrate easily, reaching the base material and promoting localized corrosion. However, the PEO treatment facilitated the growth of a thicker oxide layer with lower porosity, effectively mitigating the interaction of the alloy with Cl
− ions. This phenomenon has also been observed in other studies [
61,
62].