Corrosion Behavior of AA2055 Aluminum-Lithium Alloys Anodized in the Presence of Sulfuric Acid Solution

: The aim of this work was to evaluate the corrosion behavior of the AA2055 Aluminum-lithium alloy anodized in a sulfuric acid (H 2 SO 4 ) bath, varying the current density of 0.19 and 1 A · cm − 2 and why the sealing solution was water (H 2 O) and sodium dichromate (Na 2 Cr 2 O 7 ). Anodized samples were exposed to a 10 vol.% H 2 SO 4 solution and the electrochemical technique used was electrochemical impedance spectroscopy. Scanning electron microscopy and X-ray photoelectron spectroscopy were employed to characterization of the anodizing layer, determinate morphology and thickness of coatings. The Na 2 Cr 2 O 7 sealing solution tends to increase the charge transfer resistance and produces a more homogeneous and compact passive oxide layer, and imparts a corrosion inhibition protection to the AA2055. SEM observations indicated that the morphology and thickness of the anodic ﬁlms formed on AA2055 aluminum-lithium alloy anodized have the best results for both current densities.


Introduction
Anodizing is an electrochemical process used to increase the thickness of the oxide layer on the surface of aluminum and its alloys. Anodizing treatment is a method of conversion coatings in an electrolytic cell for surface protection on aluminum and its alloys in aerospace and aeronautical applications [1][2][3][4][5][6].
The development of the aluminum-lithium alloys (known as the third generation) has been used by the aeronautical and other applications due to its good properties, replacing conventional Al-Cu and Al-Zn alloys, resulting in an increase in payload and fuel efficiency [7,8]. Al-Li alloys have a characteristic that a reduction in density is to their advantage, and lithium, being the lightest metal, would have the greatest influence on reducing the density of aluminum. Another advantage of this alloy is that it has a lower flat mechanical anisotropy, good resistance to SCC and fracture toughness [9][10][11][12][13][14][15][16]. Therefore, low density, high strength and rigidity are one of the main alloys in aircraft manufacturing; for example, reinforcing the fuselage, floor beams and wing stringers The aluminum samples were mechanically polished using metallographic techniques according to ASTM E3 [30]. The material was sequentially polished using different SiC grit papers with 400, 600, and 800 grades, followed by ultrasonic cleaning in ethanol (C 2 H 5 OH) and deionized water.

Anodizing Treatment
The anodizing treatment consisted of the following stages [31]: The anodized specimens were immersed in H 2 O or in 6 wt.% Na 2 Cr 2 O 7 solution (analytical grade reagents (J.T. Baker)) at 95 • C for 25 min. Table 2 shows the nomenclature of different samples (S).

SEM Characterization
Microstructural analysis of the surface and in cross sections was done using scanning electron microscope (SEM, JEOL-JSM-5610LV, Tokyo, Japan), operating at 20 KV and a work distance of 12 mm. The surface micrographs by SEM was done using a secondary electrons (SE) detector. Observations of micrographs of the oxide layer thickness in cross sections were done using a backscattered electrons (BSE) detector.
EIS measurements parameters were used in the frequency range from 100 kHz to 1 mHz, with an amplitude of 10 mV sinewave [35,36]. The impedance data were collected and modelled using ZPlot and ZView software (Zview 2.0 software), respectively (Scribner Associates, Southern Pines, NC, USA). Corrosion tests were conducted in triplicate.

XPS Characterization
XPS analyses were performed to determine the chemical composition of the aluminum oxide layer and were done using the Thermo Fisher Scientific ESCALAB 250 Xi equipment (Waltham, MA, USA) at a pressure of approximately 10 mBar; the excitation of the analyzed photoelectrons was carried out with a monochromatic Al KAlpha X-ray source (1486 eV), analysis radius of µm, as well as the analysis conditions for the high-resolution zones (eV step energy, 45 • of "take-off angle" with 0.1 eV step).

SEM Microstructural Analysis
The SEM-analyzed micrographs of the oxide layer of AA2055 alloy anodized is of great importance for the corrosion behavior. Figure 1 shows the surface micrographs by SEM using a secondary electrons (SE) detector for the AA2055 anodized samples (S1, S2, S3, and S4) at different current densities (0.19 and 1 A·cm −2 ) and sealing in H 2 O and Na 2 Cr 2 O 7 solutions. Figure 1a shows a heterogeneous surface compared to the superficial micrographs of Figure 1b-d. In the surface micrographs of sample S1, there are areas with a series of bumps and some bright precipitations derived from substrate elements and impurities. Herrera et al. [37] mentioned that these impurities affect the homogeneity of the passive layer and are present because they were not completely removed during mechanical polishing or in the stage prior to anodizing. Some authors [17,38,39] have reported that the micrometric scale defects present on the surface of the anodic film have had a similar behavior in the AA2099 alloy due to the high copper content. Figure 1c,d shows samples S3 and S4 with some imperfections (cavity) on the surface; this may be due to varying the current density, the sealing solutions and the preferential dissolution of intermetallic phases, during the growth of the Al 2 O 3 layer in the H 2 SO 4 bath, resulting in a no-uniform surface [40].    Figure 3 shows Nyquist diagrams, for anodized and sealed Al-Li AA2055 alloy immersed in 10 vol.% H 2 SO 4 solution. In all samples, a capacitive loop is observed at high frequencies, at intermedium frequencies, it observed a capacitive behavior, and finally, in low frequency, we see an inductive loop having origin at frequencies between 0.20 and 0.01 Hz, which is presented in the fourth quadrant. The two capacitive loops associated at high and intermediate frequencies can be related to the ingress of the corrosive medium through the surface of the anodic coating. This is due to the surface defects seen in Figures 1 and 2, which in some cases extend to the surface of the metallic substrate. For the inductive loop observed at low frequencies, some authors have proposed different origins, among which it may be attributed to the disturbance induced by redissolution of the oxide layer formed in the anodizing process, to the species adsorption-desorption processes and related to the existence of a passive current [42][43][44]. According to Bastidas, Jirón, Jinlong, and Alexe [45][46][47][48][49], the adsorption-desorption processes on the anodized surface origin negative values of capacitance and resistance. Information provided by the diameter of the semicircle corresponding to resistance of the process is occurring [50][51][52][53][54][55]. Because of the presence of heterogeneities in barrier and porous layers, their capacitive behavior is better simulated by constant phase elements (CPE) than by a simple capacitance (C) of the electrochemical mechanism by interfacial charge storage. Fitting of the impedance data was accomplished using the equivalent circuit shown in Figure 4. Using this software (Zview 2.0 by Scribner Associates, Southern Pines, NC, USA), all the circuit parameters were adjusted simultaneously in order to fit the measured data and to obtain resistances, capacitances, relaxation frequencies, and the equivalent circuit of the materials under study. The impedance response provided by the equivalent circuit models simulates the experimental data accurately. As a result of the impedance measurements obtained, threetime constants are identified, which were adjusted using the equivalent electrical circuit model (EEC). Using CPE generally, increases in fit [56][57][58][59] and in corrosion studies, CPE are more often used to describe the frequency dependence of non-ideal capacitive behavior. Below is a CPE defined as according to the mathematical definition by Equation (1):

Electrochemical Impedance Spectroscopy Analysis
where Y = admittance (the reciprocal of electrochemical impedance), j n = (−1) = imaginary number, ω = angular frequency, and n = dimensionless fraction exponent (−1 < n < +1). CPE is an ideal capacitor (n = +1) and is an inductor (n = −1) [45,54,[60][61][62][63][64][65][66][67]. The electrical processes at layers in anodized samples have been related to the EEC model that contains the electrolyte resistance (Rs), resistance due to the sealing effect is a parallel combination (R1//CPE1); R1 is in series with a second R2//CPE2 for resistance of the porous layer; and, finally, resistance (R3) and a constant phase element (CPE3) correspond to resistance of the barrier layer.  Table 3 shows the EIS measurements obtained by EEC simulation of AA2055 Aluminum-Lithium alloys anodized and exposed to 10 vol.% H 2 SO 4 solutions. In the samples (S1-S4), the χ 2 values are in the interval of 1.48 × 10 −2 -9.49 × 10 −3 represented, and the simulations carried out by the proposed equivalent circuit can be used. The variation of resistance solution (Rs) value in the interval of 4.6490-520.9189 Ω·cm 2 is due in most cases to the morphology of the anodic layer, which is associated with the surface charge and the double electrochemical layer generated between the electrolyte and the surface [68]. Y 1 , Y 2 , and Y 3 stand for sealing effect admittance, porous layer admittance, and barrier layer admittance, respectively in the ECC. The values of the resistance R 1 , which are in an interval of 14.78 to 4121 Ω·cm 2 , correspond to the sealing effect; R 2 have a value of 128 to 208,920 Ω·cm 2 , which represents the resistance of the porous layer, and finally, R 3 with values of −66.77 to −67,814 Ω·cm 2 is the resistance of the barrier layer. The resistance to charge transfer (Rct) is the sum of the absolute value of the resistances R 1 , R 2 , and R 3 , which can be directly related to the corrosion resistance of the anodic layers formed during the anodizing process and sealing solution. In samples S3 and S4, the high resistance values (R 1 , R 2 , and R 3 ) were presented, which indicates that sealing in Na 2 Cr 2 O 7 solution produces anodic layers more resistant to corrosion. This behavior is related in some articles to the formation of compounds between the aluminum oxide and the chromium present in the sealing solution, benefiting the corrosion resistance properties of the film formed. Compounds formed in the sodium dichromate sealing process can be aluminum oxydichromate [Al(OH)CrO 4 ], or aluminum oxychromate [(AlO) 2 CrO 4 )], which are formed in the micropores; additionally the residual hexavalent chromium acts as a corrosion inhibitor [69][70][71]. The highest values of Rct were present in sample S3 (anodized with 0.19 A·cm −2 and sealed with Na 2 Cr 2 O 7 solution) and the lowest values were for sample S2 (H 2 O sealing solution), which can be due to the porosity of the aluminum oxide layer, which does not completely reduce the diameter of the pore when it is sealed with water. The n 1 value is medium (0.6290-0.7961), the significance of a homogeneous surface of the anodic oxide layer. The n 2 value is high (0.5585-0.9032), indicating a producing depressed semicircle, which may be attributed to the homogeneity of the porous layer, and the n 3 (0.5390-0.7795) parameter value is medium-high, which may be attributed to the homogeneity of the barrier layer [69][70][71][72][73][74].
The negative values of R 3 and Y 3 that correspond to the barrier layer can be explained as follows: negative capacitance occurs when the surface layer generated by different species on the surface decreases and the negative resistance is due to the interaction between the adsorption and electrodissolution of the oxide layer; negative resistances can also occur in a passive-active transition of electrochemical systems, and these negative values of capacitance and resistance occur when the systems are unstable [75][76][77][78][79].

XPS Surface Analysis
To determine the surface chemical composition of the oxide layer and valence states in the Al-Li AA2055 alloy, anodized the samples were investigated by XPS, as shown in Figures 5 and 6. Spectral analysis was performed using Avantage software of Thermo Fisher Scientific equipment (Waltham, MA, USA). To carry out chemical analysis, it is necessary to make the calculations of the deconvolution or contributions that the experimental spectrum contains, and assignment of the elements or compounds is through the peaks binding energy (they may be simple, doublets/satellites) using the database NIST (National Institute of Standards and Technology; Gaithersburg, MD, USA). Figures 5 and 6 show high-resolution XPS spectra that were collected for Al 2p and O 1s. X-ray photoelectron spectra for Al 2p are shown in Figure 5a-d. Five chemical species found Al 2 O 3 , Al 2 S 3 , Al 2 SiO 5 , Al(OH) 3 , Al(OH), and Cr 2 O 3 . The binding energy range for the high-resolution spectra for Al 2p where this behavior occurs is 73.85 to 76.88 eV; see the table of results of the parameters obtained from the peak binding energy and full width at half maximum (FWHM) fitting in Figure 5a'-d', respectively.
According to some authors, the peak of the binding energy of metallic aluminum is presented at 71.6 eV, which was not detected in the outer layer of the surface of the anodized aluminum of this work. However, the peaks of oxides and aluminum hydroxides ( Figure 5) whit values approximate of 74.5 eV [80]. For this reason, it can be considered the FWHM with a value of 0.99 in Figure 5b' related to the binding energy of 73.85 and 74.19 eV, corresponding to aluminum oxides and hydroxides, since the separation between the peaks of metallic aluminum and oxides and hydroxides have been reported with binding energy values of 2.5 and 3.7 eV, which depend on the thickness and the growth process of the oxide [81]. In the XPS spectra for Al 2p, a displacement of the binding energy peaks is observed (Figure 5b-d); when there is an increase in the current of the anodizing process influences, the structure of the oxide of the samples under study is independent of the sealing solution [82]. In this investigation, the presence of Cr 2 O 3 was detected in the samples sealed with Na 2 Cr 2 O 7 , this may be due to a migration process of the chromate ions into the interior of the Al oxide layer shown in Figures 5c and 6c,d. According to the literature [81][82][83][84][85], the non-detection of compound suggests that their concentration has not reached the necessary critical level, probably because they integrate very low solubility second-phase particulates instead of being in solid solution.
The contributions of O 1s can be observed in the XPS spectra, where they reveal the typical binding energies for O 1s, from 530.62 to 532.8 eV according to the Gauss-Lorentz peak (Figure 6a-d). The presence of oxygen in the form of aluminum oxide and hydroxides on the surface could be attributed to the presence of OH-and water groups by the sealing solutions, respectively.
The presence of SiO 2 in the XPS spectra of the O 1s orbital, whose binding energy is between 531.28 and 533.59 eV, Figure 6a'-d', can be attributed to different reasons such as: mechanical polishing, impurity in the anodizing bath or the sealing solution. In some works, different species have been found in the oxides produced by the anodizing process from aged baths; however, this difference was not significant in these works [40].
See table results of the parameters obtained, Figure 6a-d [41,[76][77][78][79][80]. XPS spectra for Al-Li AA2055 alloy: Figure 5-(a) element Al 2p for S1 sample, (b) Al 2p-S2, (c) Al 2p-S3, (d) Al 2p-S4, and Figure 6-(a) element O 1s-S1 sample, (b) O 1s-S2, (c) O 1s-S3, (d) O 1s-S4, of the chromate ions into the interior of the Al oxide layer [65]. The results of AA2055 specimens sealed H 2 O presented a surface film rich Al 2 O 3 and Al(OH); this is attributed to the anodizing and sealing process, while for the Na 2 Cr 2 O 7 sealing, the Al 2 O 3 was the main aluminum oxide found on the surface [86][87][88][89][90][91][92][93][94].  The variations in the FWHM indicate changes in chemical state and physical influences. For example, if a peak broadening, corresponding to a possible change in the number of chemical bonds, that contributes to the shape of a peak, a change in sample conditions due to X-ray damage, or localized charge differences in the surface. Therefore, metals normally have lower FWHM values than metallic compounds [95]. There exist some substantial differences in the FWHM of the Al 2p peak, being narrower in the case of α-Al 2 O 3 (FWHM of 1.35 eV), wider for hydroxides with FWHM values between 1.59 and 1.75 eV, and abnormally larger for γ-Al 2 O 3 (FWHM of 2.46 eV) [96]. In this sense, we can see in Figure 5a' sample S1 that the FWHM values are close to the peak of the α-Al Anodizing current density influences the compounds that can be formed during the process, since, as can be seen in Figure 5a',c', the compounds formed with current densities of 0.19 A·cm −2 correspond to α-Al 2 O 3 , as observed in the FWHM values of samples S1 and S3. In Figure 5b',d', which were the anodized samples with a current density of 1 A·cm −2 , there is the combination of α-Al 2 O 3 compounds, hydroxides, and γ-Al 2 O 3 since the FWHM for samples S2 and S4 present values of 0.99, 1.13, 1.8, and 2.45 eV. It is quite difficult to predictively calculate the O 1s interval between oxide and hydroxide, but it is interesting to note that, in aluminum compounds (Al-O), the bond is approximately the same length for oxide and hydroxide bonds, while in most other elements, the hydroxide bond is higher. One way to distinguish between oxide, hydroxide, and water compounds is through analysis in the O 1s region.
The O 1s peaks of aluminum oxides generally have values between 529 and 530 eV, which gives an interval of approximately 1 eV to distinguish between aluminum oxide and hydroxide compounds. Aluminum hydroxides have higher binding energy than oxides, which are in a binding energy range for the O 1s orbital between 0.5 and 0.75 eV. The usual values for the separation between oxide and hydroxide for most metals are in the interval of 1.5 to 2 eV, which makes it difficult to assign oxides and hydroxides in aluminum compounds [97]. In this work, O 1s peaks with binding energies from 530.62 to 532.34 eV were presented, which limits the possibility of distinguishing between oxides and hydroxides in this orbital. For aluminum compounds, the differences between oxides and hydroxides cannot be easily distinguished, since any of the oxides may have a peak in the O 1s orbital (529 to 530 eV). As can be seen in Figure 6a'-d', all binding energies fall outside the range proposed in the literature for the oxide region [97]. The FWHM of all the O 1s peaks is between 1.58 and 1.92 eV, giving a difference of 0.34 eV, which makes it difficult to assign aluminum oxide-hydroxide to the O 1s peaks.

•
The results reveal that the EIS technique is a good tool for obtaining detailed information on the influence of sealing and the process on anodized aluminum. • EIS results show that R 2 increase for the sample S 3 in presence of the Na 2 Cr 2 O 7 sealing solution has a higher charge transfer resistance when it has a less homogeneous and less compact layer of anodic oxide.

•
The inductive loop present in Nyquist diagrams for all samples indicates species on the surface decreases, adsorption and electrodissolution of the oxide on the AA2055 alloys anodized surface.

•
The SEM and EIS results indicated that the formation of a more compact and homogeneous anodic oxide layer is due to an increase in charge transfer resistance (R CT ), by preventing Cl − ions ingress into the anodized layer.
• SEM characterization indicated that, in both current densities, the thickness is homogeneous for the oxide films formed on anodized AA2055 aluminum-lithium alloy. The surface micrographs indicated that samples have a heterogeneous surface with bumps, some bright precipitations, and cavity. This may be due to varying the current density and the sealing solutions during the growth of the Al 2 O 3 layer in the H 2 SO 4 bath. • XPS characterization verified that the surface film of AA2055 alloy in this study consisted of a mixture of chemical compounds, such as Al 2 O 3 and AlO(OH), respectively.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.