Layered double hydroxide based active corrosion protective sealing of plasma electrolytic oxidation/sol-gel composite coating on AA2024

This work reports a novel approach for growing layered double hydroxide (LDH) films on any plasma electrolytic oxidation (PEO) coated AA2024 independently of the nature of the PEO coating. The specific PEO coating chosen to carry out this work is considered to be not suitable for direct LDH growth because of phase composition and morphological features. In this paper, we describe a new methodology that consists of covering the PEO coating with a thin layer of aluminum oxide based xerogel as the source of aluminate ions for subsequent in-situ LDH growth. X-ray diffraction (XRD) and scanning electron microscope (SEM) images showed a successful formation of LDHs on the surface. An improvement in terms of active corrosion protection was also demonstrated by electrochemical impedance spectroscopy (EIS) and scanning vibrating electrode technique (SVET).


Introduction
Plasma electrolytic oxidation is a green anodizing process using micro-discharges to produce hard and dense ceramic-like coatings with a good wear and corrosion resistance [1][2][3].
However, the discharges are also responsible for creating defects and pores that can lead to a decrease of the barrier properties provided by the coatings and their failure in long-term corrosion protection. Different methods such as optimization of the process parameters (variation of voltage/current magnitude, mode, frequency and duty cycle) [4][5][6] on one hand and the addition of various particles into the electrolyte on the other hand, have been investigated to overcome the problem [8]. Yet, this optimization do not entirely prevent the presence of the porosity, besides it could significantly affect the PEO process (voltage breakdown, discharge intensity and density) which in turn can influence the final microstructure of the PEO coating [7][8].
Several approaches based on non-toxic post-treatments to seal the pores of PEO coated Mg and Al alloys were developed. Among these approaches is the conventional hydrothermal treatment, where anodized substrates are immersed in hot (96°C -100°C) deionized water to induce the formation of a hydrated oxide and close the pores [9][10]. Narayanan et al. relied on a similar process but improved the post-treatment by decreasing the temperature to 60°C and using an alkaline solution. The process helped to seal the smaller pores, to reduce the size of the bigger sized pores, and to provide a better homogeneity to the coating [11]. Along with these methods, sealing by means of an organic top-coat such as epoxy resins and hybrid solgel layers has been also disclosed in a number of works [12][13][14][15][16]. Although the described methods can seal the pores efficiently, they fail to maintain a long-term anti-corrosion performance when subjected to damages or degradation and become permeable to corrosive species.
In this respect, other authors resorted to the use of active anti-corrosion agents to reinforce the corrosion protection of PEO coatings. Sun et al. have recently discussed the utilization of cerium nitrate and benzotriazole in an immersion post-treatment of PEO coated commercially pure Mg. The results revealed an enhanced corrosion protection owed to the sealing of the pores by Ce(NO 3 ) 3 whereas the benzotriazole has not shown any positive outcome [17]. Posttreatment relying on the use of rare earth metals such as cerium [18][19] for sealing purposes have been previously tested and proven to be effective. However, the method remains unsustainable since in most cases the corrosion inhibitors absorbed into the pores can be easily detached or react with the environment even in absence of corrosion activity, hence leading to the loss of the corrosion inhibiting capacity of the PEO coating.
In this respect, a growing interest has been accorded to smart delivery systems that can act as reservoirs for corrosion inhibitors and release them when triggered by relevant changes in the environment such as the presence of corrosive species, changes in pH or even mechanical damage. Layered double hydroxide (LDH) are promising nanocontainers that belong to the class of anionic-exchangers which structure is that of brucite Mg(OH) 2 . Part of the divalent metal cations are replaced by trivalent ions giving positively charged layers that are balanced by the intercalation of anions in the hydrated interlamellar region [20]. Their aptitude to be loaded with corrosion inhibitors and promote effective active corrosion protection for bare [21][22][23] and anodized/PEO coated metals [24][25][26][27] have been intensively exploited especially for aluminum alloys.
Nonetheless, it was found that in the case of PEO coated aluminum alloys the direct LDH growth is highly contingent on the nature of the PEO coating. Two important criteria were reported to determine the LDH growth on PEO coated AA2024: i) the γ-Al 2 O 3 and α-Al 2 O 3 crystalline phases of the PEO coating cannot be dissolved, hence used as source for LDH formation [28] and ii) the LDH growth depends on the tortuosity of the PEO coating and the accessibility of Al(OH) 4 anions through the PEO pores. For instance, a thicker and more compact PEO coating obtained at higher voltages (~ 400-500V) has proven to be challenging or even unsuitable for LDH formation [28].
In this paper, a solution to enable the growth of LDH independently on the nature and thickness of the PEO coating is presented. An additional sol-gel based post-processing was applied in order to create an aluminum oxide based xerogel layer [29][30]  The AA2024 coupons (30mm x 20 mm x 2.5 mm) were grinded with SiC 1200 abrasive paper and rinsed with deionized water prior to PEO treatment.

PEO treatment
The PEO process was carried out in an electrolyte composed of 3g/l KOH, 20g/l Na 2 SiO 3 and 20g/l Na 3 PO 4 . A pulsed DC power supply in constant voltage mode was used. In this case a voltage of 450V was applied for 10min with a pulse ratio of t on /t off =1ms/9ms at maximum current of 0.5A. The electrolyte was constantly stirred and maintained at 20±2 °C by a water cooling system.

Sol-gel post-treatment (PEO-X)
The obtained PEO samples were immersed in aluminum oxide sols that were prepared according to [29]. Briefly, 0.0123 mol of Al(NO 3 ) 3 9H 2 O was dissolved in 20 ml of isopropanol and then 0.135 mol of propylene oxide was added. The mixture was stirred for 2 min before immersing the PEO samples for 24h at room temperature. Finally, the samples were removed, washed with tert-butyl methyl ether, and then dried under ambient conditions.

LDH formation (PEO-X-LDH)
The PEO samples with the applied xerogel layer were subjected to hydrothermal LDH conversion method using the same procedure described in previous work [24][25]28]. Aside from the PEO samples covered with xerogel, another set of reference PEO samples without xerogel treatment underwent a direct LDH growth following the same procedure.

Surface characterization and phase analysis
A Tescan Vega3 SB scanning electron microscope (SEM, Brno, Czech Republic) equipped with an eumeX energy dispersive X-ray (EDS, Heidenrod, Germany) spectrometer was used to obtain images of the specimen surface and their elemental composition. The specimens were studied in both top and cross-sections views.
In addition to the SEM-EDX, tomographic data was obtained at the imaging beamline (IBL) P05 at PETRA III storage ring (DESY, Hamburg, Germany). The indirect detector system consists of a CdWO4-scintillator converting the X-rays into visible light which is magnified by a factor of 9.9 with microscope optics mounted on a back-illuminated camera system [31][32]. For measurements, an X-ray energy of 28.5keV was used with a field of view of 3.7 x 3.7 mm. The detector used is a CMOS camera with 5120 x 3840 pixels and a linear pixel size of 6.4 µm. For tomographic reconstruction, a filtered back projection using ASTRA was employed. Surface area analysis was done with the programs ImageJ and Avizo.
The phase composition and structure was investigated using a Bruker D8 Advance diffractometer (Karlsruhe, Germany). The measurement (with a Cu K α radiation, step size 0.02°, dwell time 1s) was carried at room temperature with an incident angle set to 3°.
The glow discharge optical emission spectroscopy (GDOES) depth profile analysis of the coatings was performed with a HORIBA GD-Profiler 2 (Longjumeau, France) using a 4 mm anode, an operating pressure of 650 Pa and a power of 30W.

Electrochemical and corrosion assessment
The corrosion performance of the different specimens was examined using a Gamry

3.1.Structure and morphology
SEM cross-section micrograph of the PEO coating is shown in Fig 1.a. The coating contains a number of pores, cracks and defects, which were developed due to discharges during the PEO process. However according to the literature and the present results, the choice of the electrolyte in terms of the composition and concentration of phosphate and silicate played a role on the final morphology. Studies reported that a higher concentration of phosphate in the electrolyte generally favors the production of a more compact and stable PEO coating but leads to the creation of more cracks and larger pores whereas silicates induces higher numbers of pores but smaller in size. A combination of phosphate with silicates in equal amounts resulted in the production of a compact and thicker coating with a lower number of pores [33][34][35], see Fig. 1.a. The XRD pattern ( Fig. 1.b) shows a broad signal at ~15 to 35° corresponding to a potential amorphous phase and some characteristic peaks related to the aluminum substrate. SRµCT experiments were performed to estimate the ratio between the surface area of the compact sample (bare) and porous PEO surface of the formed layer. Furthermore, the pore size distribution was calculated. The tomography studies (Fig. 2) performed on the PEO covered AA2024 allow a clear differentiation between the AA2024 alloy (light grey area with bright inclusions) and the PEO layer (dark grey region) grown on the surface. The SRµCT measurements show a thickness of 50µm of the PEO layer with irregular shaped pores. For a qualitative analysis, the data was segmented into background, PEO layer and bulk material.
Afterwards the surface area of the AA2024 and the pores was determined by Avizo using the connected components and label analysis plugins. This analysis revealed a surface area of 1.33 mm 2 for the compact bare surface and area of 6.8 mm 2 for the outer porous layer and these values include the calculation from both closed and open pores. Furthermore, a broad pore size distribution of (28 ± 23) µm 3 was determined using tomography. (Fig. 2. After characterization of the PEO coated samples, an attempt to grow LDH on the latter without any prior pre-treatment was performed. The results are depicted in Fig. 3.   Fig. 3.a and b show the SEM micrographs of the PEO coated AA2024 before and after an attempt to grow LDH-NO 3 on the surface. The planar view SEM images of the PEO coated sample (Fig. 3.a) support the previous statement regarding the porosity and nature of the obtained PEO coating. Pores which developed due to the discharges or sparking during the treatment can be observed on the surface of the PEO-treated aluminum alloy. Following the LDH treatment no changes on the PEO surface are noticed (Fig. 3.

b)
The additional XRD measurements (Fig. 3.c) did not reveal any distinctive peaks coming from LDH, and this together with SEM images support the hypothesis that the characteristics of the chosen PEO coating in terms of tortuosity and thickness is indeed not suitable for LDH formation.  After the LDH treatment, small flake-like structures appear and are on average homogeneously distributed on the PEO-X sample surface. It can be observed that the LDH grows all along the pores (Fig.4.b). The surface appearance remains the same after the LDH ion exchange with vanadate ( Fig. 4.c). Simultaneously after the SEM images ( Fig. 3 and 4), EDS elemental distribution data were collected (Fig. 5), for the four samples of interest; PEO, PEO-X, PEO-X-LDH-NO 3 and PEO-X-LDH-VO x .
Al (from the substrate, PEO, xerogel and LDH), P and Si from the PEO coating are present in all specimens. No difference in the elemental distribution between the PEO sample with and without xerogel can be observed with EDS analysis since the additional xerogel layers is mostly composed of Al oxide. Zn was revealed on both PEO-X-LDH-NO 3 (Fig. 5.c) and PEO-X-LDH-VO x (Fig. 5.d) specimens and located in regions with pores and defects of PEO coating. The same fact was described in a previous work [24]. However, in this case, it can be justified by the xerogel settling preferentially in these defected zones during the posttreatment, which consequently is followed by the LDH formation in the same zones.
Vanadium was found only on the PEO-X-LDH-VO x (Fig. 5.d) specimens, and is overlapping with regions where Zn is present.  The signal of vanadate can be observed only for PEO-X-LDH-VO x compared to a vanadate signal close to background level for both PEO and PEO-X-LDH-NO 3 . In this case, as well, an increase of the vanadate signal is observed at the PEO coating interface. The vanadate profile displays the same tendency as for the zinc in the presence of LDH (Fig. 7.d), which corroborates the assumption of the presence of LDH in the PEO pores.
In view of the gathered results, the mechanism involved in the formation of LDH films in this particular system (schematically illustrated in Fig. 8), can be explained gradually according to the following steps ; a) Xerogel formation covers two different processes; the preparation of the wet gel and its drying. There are various methods to prepare aluminum oxide wet gel [36][37]. In this work, the convenient and mild epoxide-based procedure proposed by Simpson et al. [37] was used for wet gel preparation. This procedure itself can be divided in several stages; The first stage concerns the dissolution of the metal salt. Since, in aqueous media metal atoms  However, it should be noted that the acid formed during the hydrolysis process does not allow the reactions to proceed completely.
Al oxide gel formation occurs due to the formation of chemical bonds between sol particles (condensation process with elimination of a water molecule). Therefore, for gelation, it is necessary to promote the further condensation process.
The epoxide-assisted method is based on the ability of alkene oxides to scavenge acids irreversibly, facilitating slow and controlled aluminum salt hydrolysis and sol particles' condensation.
The above explanation are resumed in form of chemical equations (1), (2) and (3): Propylene oxide scavenges protons (and counterions) and promotes further hydrolysis reaction as following, Nu H b) Subsequently to the covering of the PEO specimens with xerogel, they are immersed in a solution containing zinc and nitrate precursors for LDH formation [24,25,28].
-The LDH-NO 3 can be formed accordingly; It is important to mention that the described mechanisms are simplistic. The actual processes taking place in the aqueous media are much more complex and are highly dependent on a number of factors namely the pH, the solution concentration, solvent, temperature, etc.
(3) Therefore, it can be more challenging to identify the exact reaction mechanisms and that is outside the current work scope.

3.2.Electrochemical characterization
The overall corrosion resistance of the four systems was evaluated using electrochemical impedance spectroscopy (EIS). The Bode plots of the systems were recorded after 1h, 48h and 168h immersion in 0.5 % NaCl and are depicted in Fig. 9.
At first glance into the low frequencies (10 -2 Hz) of the different spectra, it can be clearly seen that the impedance modulus of the different systems continuously diminishes with time.
Despite that, the corrosion resistance of the PEO-X system presents a better performance than the other three systems at the beginning (1h). This can be related to the additional barrier effect offered by the xerogel layer. After 168h, the PEO-X-LDH-VO x sample shows the highest EIS modulus.
During the first hour of immersion, the PEO coated AA2024 exhibits three time constants.
The first time constant at high frequencies (Freq > 10 4 Hz) can be ascribed to a contribution of the barrier response of the outer layer and the pores that are initially not entirely filled up with solution, thus explaining the relatively high impedance at high frequencies. The second constant at mid-frequencies (10 0 -10 1 Hz) represents the response from the inner barrier layer that appears to be stable during time. The time constant at lower frequencies (Freq < 10 -1 ) can be ascribed to the response from the double layer capacitance and the charge transfer resistance due to the initiated corrosion process.
The PEO-X sample shows the highest barrier protection at the start, this is probably due to the xerogel sealing properties as mentioned above. With time, the additional protection offered by the xerogel layer decreases and after 168h, it is no longer observed.
In the case of PEO-X-LDH-NO 3 , no significant improvement of the impedance modulus can be noticed during the whole immersion period. Although, the LDH-NO 3 is expected to show some increase of the corrosion protection due to a sealing effect [24][25] but it is not apparent in this case. This behavior can be explained by the decrease/consumption of the xerogel layer during the synthesis of LDH, hence reducing the sealing effect that could be provided by the latter.
The PEO-X-LDH-VO x exhibits higher impedance modulus compared to PEO and PEO-X-LDH-NO 3 samples but still lower then PEO-X at the first hour of immersion. However, after 168h it shows a better performance, while the PEO-X impedance decreases. The PEO-X-LDH-VO x system shows fours time constant, the additional relaxation process appearing between 10 1 -10 0 Hz can be assigned to the inhibition effect of vanadate from the LDH upon release and formation of a thin protective layer [38]. Localized corrosion processes was examined with the scanning vibrating electrode technique (SVET) after 4h, 12h and 22h immersion in 0.05M NaCl. In order to observe the self-healing effect on the different samples, two artificial defects of 100µm depth were drilled into the samples at a distance of 1mm from each other. The measuring probe was placed at approximately 100 µm above the sample. The SVET maps are shown in Fig. 10.
A clear improvement of the corrosion resistance can be noticed for the sample with LDH-VO x . For the reference samples with only PEO, corrosion starts mainly at the defects after the first hours of immersion and spreads in other areas of the sample with time (22h). Significant corrosion activities were observed for the PEO-X samples and increase with time as it can be seen by the different cathodic and anodic regions on the map.
In the case of PEO-X-LDH-NO 3 sample and according to previous reported results [24], it was expected that the system would provide an additional protection due to the nano-trapping properties of LDH-NO 3 since it was previously disclosed that it could trap corrosive species such as Clinto the galleries and contribute to the reduction of corrosion reactions taking place at the interface [39][40]. Yet, this response is masked and no remarkable improvement was observed neither in the EIS nor in the SVET results. Indeed, after longer immersion times until 22h, the corrosion activity also increased for the PEO-X-LDH-NO 3 sample. Slight corrosion activity is also detected for PEO-X-LDH-VO x in the beginning. However, it does not increase greatly after longer immersion time, which backs up the results of the preceding EIS and the idea of the functionality played by vanadate in the formation of a protective layer. Considering all results, there is a clear potential for the use of xerogel as an intermediate posttreatment to enable the growth of nano-structured LDH-container layers on various range of PEO coatings, especially for the ones that are not suitable for direct LDH growth [28].
Although, we demonstrated that there is an improvement in the corrosion performance, there is still a need to develop the process further and achieve higher efficiency. This can be achieved by first, carrying out a systematic study on PEO coatings with different morphologies to have a better grasp on how to associate an ideal barrier protection from the PEO coating itself and the sealing plus active corrosion protection from the "smart" LDH nanocontainers. Secondly, the parameters for both xerogel layer application pre-treatment on PEO and the LDH growth can be optimized (concentration, temperature, time, etc.).
Moreover, one can think of other types of corrosion inhibitors with better corrosion performance and more importantly environmentally friendly which is not the case for vanadate that was used in this work as a model.