Effect of Barrier Layer on Corrosion Resistance of Porous-Type Anodic Films Formed on AA2055 Al–Cu–Li Alloy and Pure Aluminum

It is known that typical porous-type anodic ﬁ lms formed on aluminum and its alloys consist of two layers, i.e. a thick porous layer and thin barrier layer. However, the effect of the two layers on corrosion resistance of the anodic ﬁ lm is still not clear, which prevents a thorough understanding of the protection mechanism of the anodic ﬁ lm, thereby limiting the potential to further optimize the anodizing treatment. Herein, an electrochemical barrier layer thinning (EBLT) process was employed to reduce the thickness of the barrier layer of the porous-type anodic ﬁ lms formed on an AA2055 Al – Cu – Li alloy and pure aluminum. Then the structure and composition of the anodic ﬁ lms before and after immersion in a NaCl solution were studied in a comparison manner. It was found that the EBLT process evidently decreased the corrosion resistance of the ﬁ lms formed on AA2055 alloy and pure aluminum. Thus, it was conclusively validated that the barrier layer of the porous-type anodic ﬁ lm played a decisive role in controlling the corrosion resistance of anodized aluminum and aluminum alloys prior to post sealing.

Porous-type anodic film, formed through anodizing in acid solutions, has been employed to improve the corrosion resistance of aluminum and its alloys for decades. Typical porous-type anodic films formed on aluminum and its alloys commonly consist of two layers, i.e. a thick porous outer layer and a thin compact inner barrier layer. [1][2][3] To date, although much work has been devoted to understanding the effect of post treatments, 4-6 electrolyte additives [7][8][9] and substrate microstructure 10-15 on corrosion resistance of anodized aluminum and aluminum alloys, little attention has been paid to the respective role of the porous and barrier layers of the anodic film in a corrosion event. 16,17 Lack of such knowledge, however, prevents a thorough understanding of the protection mechanism of anodic films, thereby limiting the potential to further optimize the anodizing treatment in order to improve the performance of the final products.
Recently, a crucial effect of the barrier layer of the porous-type anodic film on the corrosion resistance of anodized aluminum alloys has been suggested. Veys-Renaux et al. 18 investigated the structure and electrochemical behavior of anodic films formed on different aluminum alloys (1050, 7175 and 2618), revealing that the barrier layer structure of the porous-type anodic films is affected by intermetallic phases precipitated within the aluminum matrix. Previous work of the authors on corrosion behavior of anodized Al-Cu-Li alloys [19][20][21][22] revealed that intermetallic particles introduce film defects and affect locally the structure of barrier layer, resulting in preferred sites for localized corrosion. However, a direct study of the effect of barrier layer on the corrosion resistance of anodized aluminum and aluminum alloys is still awaited and, because of this, the role of the porous layer in determining the corrosion resistance of the anodic film cannot be differentiated.
An electrochemical barrier layer thinning (EBLT) process has been used to remove the barrier layer of the porous-type anodic film in the fields where anodic film is applied as membranes 23 or templates. 24 The principle of the EBLT is schematically illustrated in Fig. 1, where the thickness of the barrier layer is gradually reduced by applying linearly decreasing voltages to the specimen immediately after a normal anodizing process. Since the thickness of the barrier layer is proportional to applied voltage, 2,25 an anodic film with extremely thin barrier layer is produced if the final applied voltage approaches zero. With precise control of the barrier layer thickness, the effect of the barrier layer on the corrosion resistance of anodized aluminum and aluminum alloys can be assessed. Additionally, the effect of porous layer on the corrosion resistance of the anodized aluminum and aluminum alloys can be distinguished.
To date, it is rare to see researchers who use EBLT to control the barrier layer thickness of the porous-type anodic films in order to study the corrosion mechanism of anodized aluminum and its alloys. Therefore, in this work, the EBLT process was employed to reduce the thickness of the barrier layer of the porous-type anodic films formed on AA2055 Al-Cu-Li alloy and pure aluminum; then the structure and composition of the anodic films before and after immersion in a NaCl solution were studied in a comparison manner. The present work provides an insight into the effect of anodic film structure on corrosion resistance of anodized aluminum and its alloys.

Experimental
A cold rolled AA2055 alloy plate (containing, in wt%, 3.46 Cu, 1.28 li, 0.38 Ag, 0.37 Mg, 0.38 Zn, 0.28 Mn, 0.1 Zr and remainder Al) and a cold rolled high purity aluminum sheet (99.999 wt% Al) were used in the present study. The AA2055 alloy was solution treated at 520°C for 40 min and then quenched in icy water to exclude the influence of precipitates on the structure and corrosion resistance of the anodic film. 21 The solution treated alloy specimens, of 10 × 10 × 2 mm dimensions, were mechanically ground with 400, 600, 800, 1200 and 2500 grit silicon carbide papers. Pure aluminum specimens, of 30 × 15 × 0.4 mm dimensions, were electropolished in a mixed solution consisting of 80 vol% ethanol and 20 vol% perchloric acid, under 20 V, at 5°C, for 3 min. The AA2055 alloy and pure aluminum specimens after mechanical polishing or electropolishing were rinsed in deionized water and dried in a cool air stream.
Normal anodizing was performed in an aqueous solution containing 0.53 M tartaric acid and 0.46 M sulfuric acid (TSA), under a constant voltage of 14 V vs the saturated calomel electrode (SCE), at 37 ± 1°C, for 1500 s. For EBLT, the specimens after normal anodizing were further anodized by applying linearly decreasing voltage from 10 to 0 V vs SCE, at a sweep rate of 10 mV s −1 . A large pure aluminum sheet and the specimen were set as cathode and anode, respectively, with the specimen surface facing the pure aluminum sheet cathode. A KR50003-500 V/3 DC power supply and a GAMRY Interface 1000 were used for potentiostatic and potentiodynamic anodizing, respectively. The specimens prior to anodizing were masked with epoxy to expose an area of 1 cm 2 and the epoxy was allowed to cure in air for 24 h. The specimens after anodizing were rinsed in deionized water and dried in a cool air stream.
Selected specimens after anodizing were immersed in a 3.5% NaCl aqueous solution at room temperature (∼20°C) for up to 96 h to introduce corrosion attack (referred to as immersion test), with the open circuit potential (OCP) of each specimen being recorded. After the immersion test, the specimens were cleaned in deionized water and then dried in a cool air stream. The surface and the cross section of the specimens before and after the immersion test were examined using a Zeiss Sigma HD scanning electron microscope, fitted with EDS facilities, and operated at an accelerating voltage of 2 kV. The cross sections of the anodized specimens before and after the immersion test were prepared by ultramicrotomy (Leica Ultracut) using a diamond knife. 26 Ultramicrotomed slices of ∼25 nm thickness from selected specimens were also examined using transmission electron microscopy (TEM). The composition of the anodic films before and after the immersion test was characterized using glow discharge optical emission spectroscopy (GDOES), operated with argon atmosphere of 635 Pa, radiation frequency of 13.56 MHz, power of 35 W and a sampling time interval of 0.005 s.

Results
The current density-time and voltage-time responses during anodizing of the alloy under 14 V for 1500 s and subsequent EBLT by potentiodynamic anodizing from 10 to 0 V (SCE) are shown in Fig. 2. Under such anodizing conditions, an anodic film with two porous layers was produced, with the outer porous layer of ∼3.7 μm thickness formed during normal anodizing (at constant voltage of 14 V) and an inner porous layer of ∼0.5 μm formed during EBLT, as shown in Fig. 3a. The boundary between the two layers was indicated by the dashed line in the figure. It is evident that the inner layer is less porous than the outer layer as a result of reduced anodizing voltages. Figure 3b shows a high angle annular dark field (HAADF) TEM image of the anodic film taken from inner regions of the anodic film, at increased magnification, revealing gradual reduction in pore size and pore wall thickness from the upper region to the film/alloy interface due to the linear reduction of the anodizing voltage. Besides, bright features are revealed at the film/alloy interface and within the anodic film about 100 nm above the film/alloy interface (with typical ones indicated by the dashedline and solid-line arrows, respectively). As reported previously, [27][28][29][30][31] such bright features are related to copper enrichment at the film/alloy interface and incorporation of the copper enrichment layer into the anodic film during anodizing of copper-containing aluminum alloys at voltages below 3 V. With further increase of the magnification (Fig. 3c), it is clearly shown that the barrier layer beneath the pore base was reduced to a value close to zero and good adhesion of the anodic film to the alloy substrate was maintained. Figure 4 compares the open circuit potential-time responses of the bare AA2055 alloy, the alloy after normal anodizing (TSA) and the alloy after normal anodizing and EBLT (TSA-EBLT), when immersed in 3.5 wt% NaCl solution at 20°C. As revealed in previous work, 21 the severe fluctuation of the OCP on bare AA2055 alloy was related to severe localized corrosion in the alloy, and the mild fluctuation of the OCP on the TSA sample was associated with initiation of localized corrosion in the alloy matrix beneath the anodic film. Similar to bare AA2055 alloy, the OCP on the TSA-EBLT sample fluctuated severely in the late stage of the  immersion test, suggesting development of severe localized corrosion in the alloy matrix beneath the anodic film. Evidently, EBLT greatly compromised the corrosion resistance of the anodic film. Figure 5 compares the surface and cross section morphology of the anodized alloy with and without EBLT, after immersion in 3.5% NaCl solution for 24 h. Individual circular cracks appeared on the specimen without EBLT (Fig. 5a). Nanopores in the anodic film free of circular cracks appeared unchanged (inset in Fig. 5a), suggesting that the anodic film only failed locally at the location of the circular cracks. Cross-sectional view of the specimen without EBLT shows localized corrosion of the alloy substrate below the circular cracks (Figs. 5c and 5e). The initiation and propagation of localized corrosion in anodized AA2055 alloy were discussed previously. 21 In contrast, network of cracks appeared on most surface of the specimen with EBLT (Fig. 5b) after immersion in the same solution for the same periods of time. Besides, the nanopores in the anodic film disappeared (inset in Fig. 5b) due to the deposition of corrosion products. Cross-sectional view of the specimen indicated that the anodic film almost completely detached from the alloy substrate (Fig. 5d) and the alloy substrate beneath the anodic film was corroded (Fig. 5f). Figure 5 demonstrates that the specimen with EBLT has much lower corrosion resistance than the specimen after normal anodizing, consistent with the trend of OCP changes during the immersion test.
Although it is demonstrated that EBLT reduced the corrosion resistance of the anodized AA2055 alloy, the effect of the barrier layer is not exclusively validated since copper-rich nanoparticles were introduced as a consequence of anodizing under voltages below 3 V during EBLT and such copper-rich nanoparticles might have affected the corrosion process of the anodized alloy. Thus, pure aluminum was anodized and tested as had been done on the AA2055 alloy. No corrosion features were observed on the surface of normally anodized pure aluminum specimen after immersion in 3.5% NaCl solution for 96 h (the SEM images are not included here). Figure 6 shows surface morphology of EBLT pure aluminum specimen after immersion in 3.5% NaCl solution for 96 h. Except for some surface contamination/deposition (with some indicated by arrows), no other features such as thick corrosion products or cracks were found on the specimen surface. Typical porous-type anodic film morphology was revealed at increased magnification (inset in Fig. 6), suggesting little attack of the film surface during EBLT and the immersion test. Figures 7a and 7b compare the cross sections of the normally anodized and EBLT pure aluminum specimens after immersion in 3.5% NaCl solution for 96 h. For the normally anodized aluminum, again, no evident corrosion features were observed in the aluminum matrix beneath the anodic film (Fig. 7a). However, for the EBLT specimen, a dark line feature (indicated by the arrow in Fig. 7b) was revealed at the film/ aluminum interface after the immersion test (Fig. 7b). TEM analysis of EBLT specimen after the immersion test shows corrosion attack of the aluminum matrix beneath the anodic film (Fig. 7c), suggesting that the dark line feature in Fig. 7b is related to corrosion products.
The corrosion behavior of the normally anodized pure aluminum and EBLT specimen was further confirmed by glow discharge optical emission spectroscopy (GDOES) analysis, as shown in Fig. 8. Figures 8a and 8c show the distribution of O, S and Al elements through the film thickness of the normally anodized and EBLT specimens, respectively. Figures 8b and 8d show the distribution of O, S and Al elements through the film thickness of the corresponding specimens after the immersion test. The signal intensity of O and S is multiplied by a factor of 100. The relatively strong signals of O and S in the early stage of sputtering are indicative of the anodic film while the relatively strong signal of Al signal in the late stage of sputtering is indicative of the aluminum substrate. The film/metal interfaces of the specimens were determined from the half height of the Al profile in the transitional region between the anodic film and the aluminum substrate, and were indicated by the dashed-lines. For the normally anodized pure aluminum (Figs. 8a and 8b), no evident change of the elemental profiles was shown after the immersion test. In contrast, significant change of the elemental profiles was shown for the EBLT pure aluminum specimen after the immersion test (Figs. 8c and 8d). Specifically, the transition for Al signal intensity from the anodic film level to substrate level is significantly longer for the tested specimen (as indicated by the arrow in Fig. 8d), which indicates that the interface became relatively rough as a consequence of corrosion attack of the substrate. Figures 7 and 8 jointly demonstrate that thinning or removal of the barrier layer effectively decreased the corrosion resistance of the anodized pure aluminum.

Discussion
Since the thickness of the barrier layer is proportional to the applied voltage, 2,25 the barrier layer thickness of the anodic film after EBLT approached zero (Fig. 3c). Compared with normally anodized specimens, the specimens with EBLT have an extra inner porous layer of ∼0.5 μm thickness. As shown in Figs. 3 and 7, the inner porous layer is more compact than the outer porous layer and it adheres well to the underneath substrate. Therefore, it is reasonable to believe that the extra inner porous layer would not decrease the corrosion resistance of the EBLT specimens, at least for the pure aluminum specimen. Besides, it takes only 17 min to finish the EBLT treatment under the selected conditions and, therefore, the effect of the electrolyte on the outer surface of the anodic film during EBLT process is weak. Thus, the lower corrosion resistance of the EBLT specimens compared with normally anodized specimens could only be ascribed to the reduced thickness of the barrier layer.
It should be noted that, with the same EBLT process and immersion test, the AA2055 alloy specimen (Figs. 5b, 5d and 5f) showed much more severe corrosion than pure aluminum specimen (Figs. 7b and 7c). This is because the AA2055 Al-Cu-Li alloy is more susceptible to corrosion attack in the NaCl solution than pure aluminum. Besides, the copper-rich nanoparticles formed at the film/ alloy interface would serve as active cathodes, further intensifying corrosion attack of the alloy matrix. With the development of corrosion in the metal substrate, oxides/hydroxides corrosion products are accumulated at the film/substrate interface. Since the corrosion products are insoluble in neutral NaCl solution and have larger volume than the consumed metal substrate, the anodic film above the corrosion products may be cracked and even detach from the substrate when the stress associated with the increased volume of corrosion products exceeds a critical value. Once the anodic film cracks and/or detaches from the substrate, the corrosive electrolyte get access easily to the metal substrate, leading to complete loss of protection by the anodic film.
As the barrier layer of the porous-type anodic film plays an important role in determining the corrosion resistance of the anodized aluminum alloys, one may ask whether it is possible to improve the corrosion resistance of the porous-type anodic film by increasing the barrier layer thickness, for instance, anodizing the alloy at higher voltages or initially anodizing at normal voltage and eventually at a higher voltage. Actually, simply anodizing the alloy at high voltages is normally impractical. Firstly, higher anodizing voltages not only mean thicker barrier layer but also higher porosity; also, higher anodizing voltages usually mean higher energy consumption and higher risk of burning. Segmented anodizing by ending at a higher voltage can be beneficial for pure aluminum due to increased barrier layer thickness and unchanged porosity in the outer region of the anodic film. However, for aluminum alloys, such beneficial effect may be compromised by coarse intermetallic particles (IMPs). As reported in previous work, 21,22 for anodized aluminum alloys, there always exist oxidized IMPs at the film/alloy interface and the barrier layer of the anodic film beneath the oxidized IMPs is usually absent or incomplete; consequently localized corrosion preferentially initiates in the regions where oxidized IMPs are present at the film/alloy interface when exposed to corrosive environment. Now that it is impractical to improve the corrosion resistance of anodized aluminum alloys by increasing the barrier layer thickness of the porous-type anodic film, then the possible option is limited to the design of the porous layer of the anodic film. The most common strategy is to increase the thickness of the porous layer. When the anodic film (the porous layer) is sufficiently thick, the probability of a through-thickness defect is zero, thus, the continuity and therefore the corrosion resistance of the anodic film is improved. However, this strategy is effective only when the pores in the anodic film are properly sealed. Otherwise, the corrosive electrolyte can easily penetrate the anodic film at the locations where anodized IMPs are present at the film/alloy interface, initiating localized corrosion. 21 Besides, the anodic film cannot be too thick because thick anodic film may compromise fatigue performance of the anodized aluminum alloy components. 32 Another strategy is to introduce complex ions into the pores of the anodic film, which allows complexing reaction between the complex ions and cations trapped in the film material, providing additional sealing to the anodic film. This is one of the theoretical basis for designing new anodizing processes such as TSA 6,33,34 and boric-sulfuric acid anodizing (BSA). 9 The third strategy involves introducing corrosion inhibitors into the anodic film. 35,36 With the introduction of corrosion inhibitors, the anodic film can not only serve as a barrier to the corrosion medium but also provide active protection by releasing corrosion inhibitors at the occurrence of corrosion.

Conclusions
An electrochemical barrier layer thinning (EBLT) process was employed to reduce the barrier layer thickness of the porous-type anodic films formed on an Al-Cu-Li alloy and pure aluminum. Compared with normally anodized alloy and pure aluminum, additional EBLT process greatly decreased the corrosion resistance of the counterparts. Consequently, it was conclusively validated that the barrier layer of the porous-type anodic film played a decisive role in controlling the corrosion resistance of anodized aluminum and aluminum alloys prior to post sealing. and Research Foundation of AVIC Manufacturing Technology Institute (KS911608114).