Corrosion Resistance of Hard Coat Anodized AA 6061 in Citric–Sulfuric Solutions

: Aluminum is a material widely used in aeronautical and transport industries due to its excellent mechanical and corrosion resistance properties. Unfortunately, aluminum alloys are susceptible to corrosion, which limits their use in some corrosive environments. The aim of this work is to characterize hard coat ﬁlm fabricated by anodizing in a citric–sulfuric acid system using electrochemical techniques. The anodization process was carried out using an aluminum alloy AA 6061 anodization bath: a mix of citric and sulfuric acid solutions were used. For the anodizing process, two current densities were used, 1 and 7.2 A · cm − 2 . Anodized specimens obtained under di ﬀ erent conditions were exposed to a 3.5 wt.% NaCl solution, and their electrochemical behavior was studied by electrochemical impedance spectroscopy (EIS) and cyclic potentiodynamic polarization (CPP) according to ASTM G106-15 and ASTM G5-13, respectively. Scanning electron microscopy (SEM) was employed to determinate the morphology and thickness of coatings. The results showed improved corrosion resistance in 6061 aluminum anodized in citric–sulfuric acid electrolyte compared to those anodized in sulfuric acid solution.


Introduction
Aluminum and its alloys are widely used in different industries, such as construction, machinery, equipment and transportation. When these alloys are exposed to the atmosphere, a thin aluminum oxide layer of 2 to 4 nm thickness naturally forms on the surface [1][2][3]. This natural oxide layer imparts improved hardness and excellent corrosion resistance characteristics to aluminum in certain atmospheric and chemical environments [4]. Anodizing is a surface treatment widely used for aluminum alloys due to its reasonable costs compared to the benefits it can offer in terms of enhanced corrosion resistance and mechanical, functional and aesthetic properties. The anodizing process involves first aluminum metal oxidation (Equation (1)) and its subsequent reaction with water to form alumina (Equation (2)) [5].

Electrochemical Techniques
A conventional three-electrode cell configuration was used for the electrochemical studies, which consisted of a working electrode (un-anodized AA 6061 and anodized aluminum), a saturated calomel electrode (SCE) and a platinum mesh, which were used as a reference and counter electrode, respectively. Electrochemical measurements were carried out using a Gill-AC potentiostat/galvanostat/ZRA from ACM Instruments (Manchester, UK). Corrosion experiments were performed by immersion of the AA 6061 and anodized aluminum specimens, with an exposed surface area of 1.0 cm 2 , in 3.5 wt.% NaCl solution, at a temperature of 25 • C. The electrochemical techniques of cyclic potentiodynamic polarization (CPP) and electrochemical impedance spectroscopy (EIS) were used to determine the corrosion kinetics of AA 6061 with the anodizing treatment. The CPP was recorded at a sweep rate of 0.06 V/min; a potential scan range was applied between −0.3 and 1.0 V vs. SCE, from the corrosion potential (E corr ), using a complete polarization cycle, according to ASTM G5-11 standard [33]. In CPP curves, analysis of the cathodic and anodic branches and the hysteresis curve can yield information about the corrosion process in the system, as well as the corrosion rates. Electrochemical impedance spectroscopy measurements were recorded at the corrosion potential over a frequency range from 1 mHz to 100 kHz, obtaining 10 points per decade and applying 10 mV RMS amplitude according to ASTM G106-15 standard [34]. The results were interpreted through the development of typical impedance models for the electrode surface and curve fitting based on an equivalent circuit using the Zview Impedance program.

Microstructural Characterization
Analysis of surface morphology and composition was done using a scanning electron microscope (SEM, Jeol JSM 6510LV, Tokyo, Japan). Observations of the cross-sections micrographs of the sample were also made to determine the thickness of the oxide layer at magnification of 2000×, operating at a voltage of 20 kV and a working distance (WD) of 12 mm. The chemical composition of the cross-sections structures was obtained by energy dispersive X-ray spectroscopy (EDS).

Chemical Composition
The chemical composition of the material under study obtained by X-ray fluorescence is presented in Table 2. The composition of aluminum corresponds to that presented by different authors [24].  Figure 1 shows the cyclic potentiodynamic polarization curves realized for the un-anodized AA 6061 and anodized material. The absence of a hysteresis loop during the potential scan in un-anodized AA 6061 and SA (sulfuric acid)/1 A·cm −2 means that the localized corrosion did not occur, but it could be a sign of an active surface and general corrosion. The results show a positive hysteresis for the following samples, CA (citric acid)/1 A·cm −2 and CA (citric acid)/7A·cm −2 . This behavior is indicative of localized corrosion on the material surface, and the slow decrease of current density in the reverse scan in positive hysteresis is indicative of difficulty in surface repassivation or stopping the growth of pits.

Cyclic Potentiodynamic Polarization (CPP)
Coatings 2020, 10, x FOR PEER REVIEW 5 of 14 The samples (un-anodized AA 6061 and anodized sample) did not present protection or repassivation potential (ERep), which is indicative of pitting corrosion, since the potential for anodic to cathodic transition was lower in all cases than the corrosion potential (Ecorr). For the non-anodized 6061 alloy and CA (citric acid)/7.2 A·cm −2 , the pitting potential (EPit) coincided with the corrosion potential (Ecorr), which occurred when there was an oxide film on the material surface prior to the polarization. Due to the intersection of cathodic branch with the transpassive region of anodic branch, the value of the pitting potential is the same as the corrosion potential [40,41]. These results are presented in Table 3.    The behavior of the samples CA (citric acid)/1 A·cm −2 and CA (citric acid)/7 A·cm −2 ( Figure 1) was observed in AA 2024 and AA 2099, with anodizing treatment evaluated in NaCl [35,36]. Such behavior is attributed to the presence of Cl − ions in the solution [37,38]. The sample anodized only with sulfuric acid and a current density of 7.2 A·cm −2 (SA (sulfuric acid)/7.2 A·cm −2 ) presented a negative hysteresis cycle, indicating uniform corrosion. In this sample, active-passive transition potential was more noble than corrosion potential (E corr ). This occurred for alloys that are susceptible to passivation and hence are able to restore the damaged oxide film or the alloys that are not susceptible to pitting corrosion [39].
Regarding the corrosion potential of the different samples, an increase in the corrosion potential of the anodized samples in sulfuric acid solution can be observed, presenting potentials of around −0.629 V vs. SCE in the sample SA (sulfuric acid)/1 A·cm −2 and −0.566 V vs. SCE in the sample SA (sulfuric acid)/7.2 A·cm −2 (Table 3). However, the corrosion potential of the CA (citric acid)/1 A·cm −2 sample anodized with the citric-sulfuric acid solution presented more noble potentials of around −0.343 V vs. SCE. Only the CA (citric acid)/7.2 A·cm −2 sample presented corrosion potentials similar to the corrosion potential of the non-anodized material of −0.693 V vs. SCE (Table 3). This may be indicative that although the aluminum oxide layer is formed in the citric-sulfuric acid solution, it does not possess protective characteristics, such as the layers generated in sulfuric acid solution. In addition, it may be due to the porous surface generated during anodization in citric-sulfuric acid solution.
The samples (un-anodized AA 6061 and anodized sample) did not present protection or repassivation potential (E Rep ), which is indicative of pitting corrosion, since the potential for anodic to cathodic transition was lower in all cases than the corrosion potential (E corr ). For the non-anodized 6061 alloy and CA (citric acid)/7.2 A·cm −2 , the pitting potential (E Pit ) coincided with the corrosion potential (E corr ), which occurred when there was an oxide film on the material surface prior to the polarization. Due to the intersection of cathodic branch with the transpassive region of anodic branch, the value of the pitting potential is the same as the corrosion potential [40,41]. These results are presented in Table 3.
In relation to the effect of the current applied in sulfuric acid solution, it was observed that the sample anodized with 1 A/cm 2 (SA (sulfuric acid)/1 A·cm −2 ) had a higher current demand in the anodic section of the CPP and a higher passivation current density (i pass ), greater than the anodized sample with 7.2 A/cm 2 (SA (sulfuric acid)/7.2 A·cm −2 ), with both values being 52.2 and 42.9 nA/cm 2 , respectively. This represents the formation of a more stable aluminum oxide film with better properties in the sample made with a higher current, in this case 7.2 A/cm 2 (see Table 3). In the case of the samples anodized in the citric-sulfuric acid solution, both samples presented similar behaviors in the hysteresis, presenting positive hysteresis ( Figure 1).
The anodized sample in the citric-sulfuric acid solution with an anodizing current density of 1 A/cm 2 had a passivation current density (i pass ) of 65.2 nA/cm 2 , while the anodized sample with 7 A/cm 2 did not show passive behavior. Instead, a considerable current demand was observed in the anodic branch. This behavior was due to the penetration of ions Cl − through the porous oxide layer on the surface, which can also attack the barrier layer. The reference sample, AA 6061 alloy, presented a current demand during the entire cyclic potentiodynamic polarization, and it did not tend Coatings 2020, 10, 601 6 of 14 to passivation. This behavior is indicative that the corrosion process was caused by pitting on the surface of this material (Table 3) [42].
Samples SA (sulfuric acid)/1 A·cm −2 , SA (sulfuric acid)/7.2 A·cm −2 and CA (citric acid)/1 A·cm −2 presented higher corrosion potentials (E corr ) and lower corrosion current densities (i corr ), indicating that with these anodizing conditions the corrosion resistance of AA 6061 aluminum alloy was increased.
For the samples of SA (sulfuric acid)/7.2 A·cm −2 and CA (citric acid)/1 A·cm −2 , the protective layer formed during anodizing hindered the entry of Cl − ions into the pores of the films. This most positive corrosion potential (E corr ) was from −0.566 and −0.343 V, and the smallest corrosion current density (i corr ) was in the order of nanoamperes/cm 2 (1 × 10 −9 amperes), which indicate a better corrosion resistance with those conditions. Figure 2a shows the Nyquist plot obtained for the un-anodized and anodized samples in sulfuric acid after exposure to the 3.5 wt.% NaCl electrolyte. The un-anodized specimens show a capacitive loop across the whole frequency range, and for the anodized SA (sulfuric acid/1 A·cm −2 ) and SA (sulfuric acid/7.2 A·cm −2 ) there is a high-frequency semi-circle and capacitive behavior at lower frequencies, which correspond to the characteristics of the porous layer and the barrier layer, respectively. In a similar way, Figure 3a shows the Nyquist plot obtained for the un-anodized and anodized samples in citric-sulfuric acid mixture after exposure to the 3.5 wt.% NaCl electrolyte. In this case, the CA (7.2 A cm −2 ) specimens show a capacitive loop across the entire frequency range, and for the un-anodized sample and CA (citric acid/1 A·cm −2 ) the high-frequency range of the diagram reflects the porous layer characteristics, while the low-frequency range characterizes the barrier layer properties.  Several studies have shown that the low-frequency range corresponds to the properties of the barrier layer and the high-and medium-frequency ranges reflect the properties of the porous layer [43][44][45]. It has been mentioned previously by some authors that the anodic film on aluminum consists of a very thin compact barrier layer and a thicker porous layer, and that the latter layer is composed of pores and walls of hexagonal-shaped cells. To describe the electrochemical behavior of this system, the thin barrier layer and the porous layer are considered independent of each other [46,47]. To describe these behaviors, several equivalent circuits have been proposed to model the response of porous anodic films [43,48]. It is accepted that different equivalent circuits can model the porous Coatings 2020, 10, 601 7 of 14 anodic film; these models are successfully applied to explain the properties of the compact barrier and the porous layers of many materials [44,[49][50][51][52][53][54].  The variation of EIS parameters for the un-anodized AA 6061 and samples anodized under different conditions exposed in 3.5 wt.% NaCl solution is presented in Table 4. The EIS response of the un-anodized AA 6061 sample in the 3.5 wt.% NaCl solution can be represented by the equivalent circuit in Figure 2b, which consists of a resistance RB and a capacitance CB in parallel. CB can be treated as a constant phase element, and because the nB exponent is 0.94 this means that this constant phase element can be treated as a capacitor. The behavior of anodized SA (sulfuric acid/1 A·cm −2 ) and SA (sulfuric acid/7.2 A·cm −2 ) samples can be represented by the equivalent circuit in Figure 2c. From the values of Rp and RB in Table 4, it can be seen the beneficial effect of anodizing in sulfuric acid compared with the un-anodized AA 6061 sample. The higher Rp value of the SA (sulfuric acid/1 A·cm −2 ) sample compared to the SA (sulfuric acid/7.2 A·cm −2 ) sample could indicate that the porous layer in the former has a more homogeneous morphology than for the latter.

Electrochemical Impedance Spectroscopy (EIS)
From Figure 3, it is interesting to note the behavior of the CA (citric acid/1 A·cm −2 ) sample, which had a porous layer Rp value of 3.16 × 10 9 Ω·cm 2 and a barrier layer RB value of 522,070 Ω·cm 2 (the highest RB value measured under all conditions). To some extent, these data and the nB value obtained indicate the formation of a homogeneous anodized layer. After testing different equivalent circuits for the EIS data obtained, the circuit proposed by Suay et al. [55] gave suitable fits (see Figure 2b,c and Figure 3b,c). In these figures, R s is the electrolyte resistance, R B is the barrier layer resistance, CPE B is the barrier layer constant phase element, n B is the barrier layer impedance exponent, R 1 is the resistance of ionic pathways in pores and defects, CPE w is the constant phase element of the cells walls with uniform or nearly uniform dielectric properties, n w is the cell wall impedance exponent and n p is the pore barrier impedance exponent. The electrochemical properties of the porous layer are characterized by the capacitance C p and the resistance R p , while the properties of the barrier layer are defined by the capacitance R B and the resistance R B .
The variation of EIS parameters for the un-anodized AA 6061 and samples anodized under different conditions exposed in 3.5 wt.% NaCl solution is presented in Table 4. The EIS response of the un-anodized AA 6061 sample in the 3.5 wt.% NaCl solution can be represented by the equivalent circuit in Figure 2b, which consists of a resistance R B and a capacitance C B in parallel. C B can be treated as a constant phase element, and because the n B exponent is 0.94 this means that this constant phase element can be treated as a capacitor. The behavior of anodized SA (sulfuric acid/1 A·cm −2 ) and SA (sulfuric acid/7.2 A·cm −2 ) samples can be represented by the equivalent circuit in Figure 2c. From the values of R p and R B in Table 4, it can be seen the beneficial effect of anodizing in sulfuric acid compared with the un-anodized AA 6061 sample. The higher R p value of the SA (sulfuric acid/1 A·cm −2 ) sample compared to the SA (sulfuric acid/7.2 A·cm −2 ) sample could indicate that the porous layer in the former has a more homogeneous morphology than for the latter.
From Figure 3, it is interesting to note the behavior of the CA (citric acid/1 A·cm −2 ) sample, which had a porous layer R p value of 3.16 × 10 9 Ω·cm 2 and a barrier layer R B value of 522,070 Ω·cm 2 (the highest R B value measured under all conditions). To some extent, these data and the n B value obtained indicate the formation of a homogeneous anodized layer.
The value of resistance R 1 (resistance of ionic pathways in pores and defects) can be related to the quality of pore sealing and varies from about 43,337 to 44,223 Ω·cm 2 for samples anodized with sulfuric acid at 1 A·cm −2 and 7.2 A·cm −2 , respectively. An R 1 value of 284,410 Ω·cm 2 was recorded for the CA (citric acid/1 A·cm −2 ) sample. This could be due to a better sealing of pores under this last condition. On the other hand, the higher values of C B for un-anodized AA 6061 and CA (citric acid/1 A·cm −2 ) samples as compared with other conditions are attributed to a thinner barrier layer. Last, the better corrosion resistance for the CA (citric acid/1 A·cm −2 ) sample might not be related to a thicker anodized oxide layer. Indeed, this was not the case. Thus, a more homogeneous and organized oxide layer in the CA (citric acid/1 A·cm −2 ) sample probably accounts for its good corrosion resistance. Overall, the EIS data showed better corrosion behavior for the CA (citric acid/1 A·cm −2 ) samples in comparison with the other anodizing conditions. Table 4. Electrochemical impedance spectroscopy (EIS) data obtained by equivalent electric circuit simulation of the un-anodized AA 6061 and anodized samples at different conditions exposed in 3.5 wt.% NaCl solution. The combination of anodizing solutions and currents applied in the anodizing process for the two cases presented in this work are indicative that these treatments can be effective against electrolyte penetration, retarding corrosion and degradation of the anodized layer. Therefore, a homogeneous and less porous surface improves the corrosion behavior by restricting the entry of chlorides towards the metallic substrate [56,57]. Figure 4 shows the surface morphology obtained by SEM in plan-view for samples anodized with (a) sulfuric acid solution and (b) a mixture of citric-sulfuric acid solution at two currents densities, i.e., 1 and 7.2 A·cm −2 . As can be observed, irrespective of the anodizing solution used, the samples processed using a current density of 7.2 A·cm −2 presented porosity and superficial cracking (see Figure 4b,d). The anodized samples in both solutions with a current density of 1 A·cm −2 , presented a more homogeneous (less superficial defects) and similar morphology (see Figure 4a,c). In fact, the CA (citric acid)/1 A·cm −2 sample shows a more regular and smoother surface than that for the SA (sulfuric acid)/1 A·cm −2 sample, which leads to a better corrosion resistance, since in this condition Cl − ions do not have an easy access path towards the barrier layer and the metallic substrate. For these conditions, the R p and R B values from Table 4 confirm the above observations. On the other hand, the sample anodized with citric acid and a current density of 7.2 A·cm −2 presented the most defective and heterogeneous surface morphology under all conditions, giving the worst performance against corrosion. This type of structure was also observed in a previous study with a mixed tertiary acid anodization process [50]. Other authors have observed similar morphologies on the surface of anodized AA 1050 and AA 2024 T3 aluminum alloys in boric acid solutions [58].

SEM Microstructural Analysis
Cross-sectional micrographs obtained by SEM of the anodized AA 6061 alloy under different conditions is shown in Figure 5a-d. A noticeable difference in the thickness of the coatings formed is observed, with the greatest thicknesses being those produced with a current of 7.2 A/cm 2 in the two anodizing solutions, i.e., 20.7 µm for the anodizing produced in sulfuric acid solution and 14.3 µm for the anodizing produced in the citric-sulfuric acid solution. On these coatings, pores and cracks are visible, in particular for the CA (citric acid)/1 A cm −2 . An important reason for the nucleation of pores and cracks on the anodized layer could be related to the development of thermal stresses caused by the differences in thermal expansion coefficients of the anodic coating and the metallic substrate during the sealing operation. The observed trend was different from the expected trend of fewer cracks at lower anodization current densities in terms of thermal stresses. Another reason could be related to mechanical stresses induced during the metallographic cross-sections preparation of the samples [59].
Coatings 2020, 10, x FOR PEER REVIEW 9 of 14 defective and heterogeneous surface morphology under all conditions, giving the worst performance against corrosion. This type of structure was also observed in a previous study with a mixed tertiary acid anodization process [50].  Cross-sectional micrographs obtained by SEM of the anodized AA 6061 alloy under different conditions is shown in Figure 5a-d. A noticeable difference in the thickness of the coatings formed is observed, with the greatest thicknesses being those produced with a current of 7.2 A/cm 2 in the two anodizing solutions, i.e., 20.7 μm for the anodizing produced in sulfuric acid solution and 14.3 μm for the anodizing produced in the citric-sulfuric acid solution. On these coatings, pores and cracks are visible, in particular for the CA (citric acid)/1 A cm −2 . An important reason for the nucleation of pores and cracks on the anodized layer could be related to the development of thermal stresses caused by the differences in thermal expansion coefficients of the anodic coating and the metallic substrate during the sealing operation. The observed trend was different from the expected trend of fewer cracks at lower anodization current densities in terms of thermal stresses. Another reason could be related to mechanical stresses induced during the metallographic cross-sections preparation of the samples [59].
The thinnest anodized coating (3.5 μm) was obtained with the sulfuric acid solution and a current density of 1 A·cm −2 , followed by the coating obtained with the citric acid solution and a current density of 1 A·cm −2 (7.7 μm). In the former, its good behavior against corrosion could be due to the fact that the porous oxide layer is thick enough and is not being affected by the electrolyte. However, the latter presented the best corrosion performance. This indicates that not only thickness but also other factors such as a more organized and less defective layer structure, are important to explain the behavior found. The thinnest anodized coating (3.5 µm) was obtained with the sulfuric acid solution and a current density of 1 A·cm −2 , followed by the coating obtained with the citric acid solution and a current density of 1 A·cm −2 (7.7 µm). In the former, its good behavior against corrosion could be due to the fact that the porous oxide layer is thick enough and is not being affected by the electrolyte. However, the latter presented the best corrosion performance. This indicates that not only thickness but also other factors such as a more organized and less defective layer structure, are important to explain the behavior found.
In the anodizing process, the coating is created from the surface of the metal. As the coating gets thicker, the current has to pass through the growing layer to reach the clean metal surface. As the coating grows from the bottom up, it may simultaneously be dissolved at the surface. The aggressiveness of the electrolyte is dependent on the properties of the additives in the acid. Thin coatings are produced when the additive in this case citric acid inhibits the dissolution of the oxide coating [31].
In Figure 6, the cross-section of the anodized AA 6061 alloy under different conditions and the element mapping performed by EDS is shown. The cross-section mapping of elements (Figure 6e-h), shows a higher concentration of oxygen (red) in the anodized layer formed in the anodizing solutions and different current densities, which confirms the formation of the anodized layer.
In the anodizing process, the coating is created from the surface of the metal. As the coating gets thicker, the current has to pass through the growing layer to reach the clean metal surface. As the coating grows from the bottom up, it may simultaneously be dissolved at the surface. The aggressiveness of the electrolyte is dependent on the properties of the additives in the acid. Thin coatings are produced when the additive in this case citric acid inhibits the dissolution of the oxide coating [31]. In Figure 6, the cross-section of the anodized AA 6061 alloy under different conditions and the element mapping performed by EDS is shown. The cross-section mapping of elements (Figure 6e-h), shows a higher concentration of oxygen (red) in the anodized layer formed in the anodizing solutions and different current densities, which confirms the formation of the anodized layer.

Conclusions
• In this work, samples of AA 6061 were anodized in sulfuric acid and citric-sulfuric acid baths and exposed in a 3.5 wt.% NaCl solution. Their electrochemical behavior was studied by cyclic potentiodynamic polarization and electrochemical impedance spectroscopy. • In general, cyclic potentiodynamic polarization results indicated that the sample anodized SA (sulfuric acid)/7.2 A·cm −2 presented negative hysteresis, mainly indicating generalized corrosion, while the rest of the samples presented positive hysteresis indicating a localized corrosion. • EIS results show that there are two time constants: the first referring to the surface porous layer and the second related to a non-porous and protective barrier layer. • For the samples anodized with a current density of 1 A·cm −2 , the sample anodized in citricsulfuric acid solution exhibited the best corrosion resistance, followed by the sample anodized in sulfuric acid solution. The worst performance was observed for the sample anodized in citricsulfuric acid solution with a current density of 7.2 A·cm −2 .

Conclusions
• In this work, samples of AA 6061 were anodized in sulfuric acid and citric-sulfuric acid baths and exposed in a 3.5 wt.% NaCl solution. Their electrochemical behavior was studied by cyclic potentiodynamic polarization and electrochemical impedance spectroscopy.

•
In general, cyclic potentiodynamic polarization results indicated that the sample anodized SA (sulfuric acid)/7.2 A·cm −2 presented negative hysteresis, mainly indicating generalized corrosion, while the rest of the samples presented positive hysteresis indicating a localized corrosion.
• EIS results show that there are two time constants: the first referring to the surface porous layer and the second related to a non-porous and protective barrier layer.

•
For the samples anodized with a current density of 1 A·cm −2 , the sample anodized in citric-sulfuric acid solution exhibited the best corrosion resistance, followed by the sample anodized in sulfuric acid solution. The worst performance was observed for the sample anodized in citric-sulfuric acid solution with a current density of 7.2 A·cm −2 . • SEM observations indicated that the morphology and thickness of the anodic films formed on AA 6061 aluminum in sulfuric acid solution and citric acid-sulfuric acid solution depend upon the current density used. At high-current densities (7.2 A·cm −2 ), there is a considerable increase in the thickness of the anodized layer, compared to anodized samples at low-current densities (1 A·cm −2 ). In the former case, the anodized film presented porosity and cracking, whereas for the latter case, the anodized film was homogeneous and compact.