Influence of sealing post-treatments on the corrosion resistance of PEO coated AZ91 magnesium alloy

Applied Surface Science 433 (2018) 653–667

Contents lists available at ScienceDirect

Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc

Full Length Article

Influence of sealing post-treatments on the corrosion resistance of PEO coated AZ91 magnesium alloy B. Mingo a,b,∗ , R. Arrabal b , M. Mohedano b , Y. Llamazares b , E. Matykina b , A. Yerokhin a , A. Pardo b a b

School of Materials, University of Manchester, Oxford Road, Manchester, M13 9PL, UK Departamento de Ciencia de Materiales, Facultad de Ciencias Químicas, Universidad Complutense, 28040, Madrid, Spain

a r t i c l e

i n f o

Article history: Received 25 July 2017 Received in revised form 10 October 2017 Accepted 11 October 2017 Available online 13 October 2017 Keywords: PEO Magnesium Corrosion Coatings Sealing Contact angle

a b s t r a c t The effect of three different post-treatments carried out on Plasma Electrolytic Oxidation (PEO) coated magnesium alloys are evaluated in terms of characterisation and corrosion resistance. Special interest is given to the role of a common additive (NaF) to the coating properties. The post–treatments are based on immersion sealing processes in aqueous solutions of inorganic salts (cerium and stannate based salts) and alcoholic solution of an organic acid (octodecylphosphate acid, ODP). Sealing mechanisms for each post-treatment are proposed. Cerium and stannate sealings are based on filling of the pores with the products of dissolution/precipitation reactions, while the ODP acid sealing is based on the formation of a thin layer of ODP over the coating through specific interactions between the polar part of the organic acid and the coating surface. All coatings are evaluated by salt fog test and analysed by electrochemical impedance spectroscopy. All sealings show a slight increase in the corrosion resistance of the coatings formed in the NaF-free electrolyte, but their positive influence is boosted in case of the coatings obtained in the NaF-containing electrolyte. This is related to the chemical and morphological changes at the coating surface induced by the presence of NaF in the electrolyte. © 2017 Elsevier B.V. All rights reserved.

1. Introduction The growing use of magnesium in weight-sensitive applications has triggered the development of surface modification treatments capable of increasing its performance, especially its corrosion resistance. Amongst them, Plasma Electrolytic Oxidation (PEO) stands out as an eco-friendly plasma-assisted electrolytic technique capable of obtaining highly stable ceramic coatings with improved hardness, adhesion, corrosion and wear resistance compared to the coatings obtained by other electrolytic methods, such as anodising [1]. PEO operates at direct, alternating, unipolar or bipolar current modes in the voltage range of 400–1000 V [2] using, generally, alkaline aqueous solutions, although acidic electrolytes can also be used [3]. During the treatment, plasma micro-discharges are generated at the substrate-electrolyte interface due to the dielectric breakdown of the surface oxide film formed at the early stages of the process. At the micro-discharge sites, plasma-assisted ther-

∗ Corresponding author at: School of Materials, University of Manchester, Oxford Road, Manchester, M13 9PL, UK. E-mail address: [email protected] (B. Mingo). https://doi.org/10.1016/j.apsusc.2017.10.083 0169-4332/© 2017 Elsevier B.V. All rights reserved.

mochemical reactions take place whose products solidify within the discharge channels at extremely high cooling rates (108 K s−1 ) [4], leading to the formation of high-temperature phases and metastable compounds containing elements from both the substrate and the electrolyte. PEO coatings usually possess a layered structure comprising a nanometre scale (100–600 nm) dense inner barrier layer and a thick (5–100 ␮m) outer layer with variable porosity [1,5]. Although the morphology of the barrier layer is difficult to modify, the thickness, porosity and composition of the outer layer can be tailored as required for specific applications. Such versatility is one of the major advantages of this technique since it allows it to be used in a wide variety of applications. For instance, porous coatings are usually preferred in biomaterials, since they promote cell adhesion, while dense, compact and impermeable coatings are desired to prevent corrosion. To achieve required coating properties, the following three approaches are usually pursued: (i) optimisation of the electrical parameters, (ii) selection of the electrolyte composition and (iii) application of post-treatments. Regarding the first approach, the applied current mode, voltage, current density, pulse frequency and duty cycle have been

654

B. Mingo et al. / Applied Surface Science 433 (2018) 653–667

previously reported to influence the nature and characteristics of micro-discharges (e.g. ignition voltage and plasma temperature) [6], which affected drastically the coating morphology [1]. Generally, AC and bipolar current modes result in denser and more homogeneous coatings compared to those produced by DC and unipolar modes, as the former allow the discharge lifetime to be controlled. High voltages and current densities increase the oxidation rate of magnesium due to the higher energy input, which leads to higher growth rates and, therefore, coatings thickness [7]. Increasing the duty cycle causes a greater energy intensity and lower density of micro-discharges [8,9] which results in higher coating porosity and lower thickness [1]. However, in case of PEO coatings formed on magnesium alloys, contradictory results have been observed. Gao et al. [10] evaluated the influence of the current mode (pulsed unipolar and pulsed biopolar) formed on cp-Mg and observed that the passivation of the metal surface depends on the applied negative pulse. They observed that passivation was hindered under more energetic negative pulses, which results in lower growth rates. The coatings formed under bipolar conditions presented more defects compared to the ones formed under unipolar conditions, which lead to coatings with lower corrosion resistance. Gnedenkov et al. [11] also studied the effect of current mode on the properties of PEO coatings formed on a Mg-Mn alloy and observed the opposite behaviour, coatings formed under biopolar mode had a higher growth rate and improved mechanical and corrosion properties compared to the coatings formed under unipolar conditions. This suggests that there are other factors that influence the coatings growth mechanism and therefore the coatings morphology. The selection of the electrolyte is one of these fundamental factors that affect both the composition and morphology of the coatings [4]. NaF is a common additive to PEO electrolytes and it is able to modify greatly the coating morphology. It has been demonstrated that F− ions promote Mg passivation [12] which leads to lower breakdown potentials and results in materials with increased hardness, wear [13] and corrosion resistance [14]. The composition of the coating can also be tailored modifying the composition of the electrolyte, since cataphoretic effects occur during the treatments which allows the incorporation of insoluble particles added to the electrolyte. Previous works by the authors demonstrated successful incorporation of various nanoparticles, e.g. ZrO2 [15] and CeO2 [16] to enhance the corrosion resistance, Al2 O3 [17] and SiC [18] to improve the tribological properties and Ca- and P-rich [19] compounds to promote the biocompatibility of PEO coatings. However, these strategies do not always confer sufficient protection against corrosion in aggressive environments since the inherent porosity of PEO coatings allows the penetration of aggressive species, which would eventually reach the metal substrate. Therefore, the application of post-treatments adapted from the sealing processes used in anodising is currently becoming a wide spread strategy to improve the long-term corrosion resistance of PEO coatings. The simplest post-treatment consists in immersion of PEOcoated specimens in boiling water. Following this procedure, Chu et al. [20] observed an improvement in the corrosion resistance of PEO coatings on a magnesium alloy due to partial blocking of the coating porosity by hydroxides and oxides formed in the pores. Based on the same principle, different types of immersion posttreatments have been carried out using different solutions. For instance, immersion in aqueous solutions of various salts provides a simple and economic treatment based on the formation of compounds with low solubility, which would precipitate into pores and cracks of the coating, thus obstructing the paths for corrosive agents [21]. Different types of solutions have been successfully used, including phosphates [22], silicates [23], hydroxyapatite [24] or rare earth based compounds (Ce [22], La [25]). However, these

processes are often reversible so the deposited salts can be redissolved with time, losing the sealing capacity. An alternative approach to the immersion in salt solutions is the application of organic top-coats. Organic compounds penetrate easily into the pores and cracks of PEO coatings, sealing them completely and resulting in a smooth surface layer very effective against corrosion. Organic top-coats are based on either organic acids or polymeric compounds synthesised mainly by sol–gel processes. Octadecylphosphonicacid (ODP), on the other hand, is an organic type of sealant which has the ability to interact metal oxide surfaces during simple immersion, forming a hydrophobic monolayer (or a few monolayers) film that improves the corrosion resistance. There are a few examples in the literature where the sealing properties of ODP acid are evaluated on aluminium alloys, i.e. in organic coatings and on a PEO coating [26,27]. However, the results are rather limited and further studies are needed. On the contrary, there are numerous works where the improvement in the corrosion resistance of PEO coatings was promoted by the application of polymeric top-coats [23,28–30], but their main limitation is their low mechanical properties. Scratches or external impacts can easily damage the organic film and the corrosion attack can spread relatively easily under the undamaged areas [31], therefore it is important to have an effective corrosion inhibiting sealant under the polymer top-coats. In this work, three different types of post-treatments performed on PEO coated AZ91 magnesium alloy are studied and compared from a characterisation and corrosion protection points of view. The applied post-treatments are based on immersion sealings in different media (aqueous solutions of cerium salts and sodium stannate as well as an alcoholic solution of phosphonic acid). Although the use of post-treatments is a common practice in Mg based coatings, there is little information available related to these three types of sealing post-treatments. The main objective of this work is to elucidate the mechanisms of different sealing post-treatments on PEO coatings formed on AZ91 magnesium alloy and their effect on the corrosion resistance. Special interest is given to the role of NaF, a common additive of PEO electrolytes, on the resulting coating morphology and composition. 2. Materials and methods 2.1. Material A commercial alloy AZ91D with nominal composition of (wt.%) 8.80 Al, 0.68 Zn, 0.30 Mn, 0.01 Si, < 0.008 Ni, 0.004 Fe, < 0.001 Cu, Mg – balance supplied by Magnesium Elektron Ltd. was used to carry out the investigation. The metallic substrates (34 × 24 × 2 mm3 ) were ground using successive grades of SiC paper (up to P1200), rinsed in water, cleaned with isopropyl alcohol and dried in warm air. Then the working area was delimited to ∼4 cm2 using a commercial resin (Lacquer 45, MacDermid plc.). 2.2. PEO coatings PEO coatings were obtained using an AC voltage-controlled power supply (EAC-S2000 ET Systems electronic) and an electrochemical cell equipped with a thermostatic jacket (20 ± 1)◦ C under continuous electrolyte agitation. An AISI 316 steel plate of 7.5 × 15 cm2 size was used as a counter electrode. A square waveform voltage signal was applied with a positive-to-negative pulse ratio of 420 V/60 V, using a 60 s initial ramp to reach the voltage amplitude, at 500 Hz frequency. The voltage peak-to-peak value was kept constant and the root mean square (rms) current density was limited to 200 mA cm−2 . To evaluate the influence of NaF on the composition and morphological features of PEO coatings, two dif-

B. Mingo et al. / Applied Surface Science 433 (2018) 653–667

ferent electrolytes were used (electrolyte 1: Na2 SiO3 ·5H2 O, 8 mL/L, KOH 6 g/L and electrolyte 2: Na2 SiO3 ·5H2 O 8 mL/L, KOH 6 g/L, NaF 2 g/L). The process duration was 400 and 200 s for the NaF-free (PEO−1) and NaF-containing coating (PEO−2), respectively. 2.3. Sealing post-treatments The cerium salt sealing was carried out using an aqueous solution composed of 13.3 g L−1 Ce(NO3 )3 ·6H2 O, 3 g L−1 H2 O2 and 1 g L−1 H3 BO3 . Specimens were immersed into the solution under constant agitation during 180 min at 30 ◦ C. After that they were rinsed with water and dried with warm air. The stannate sealing was performed using an aqueous solution of 160 g L−1 Na2 SnO3 3H2 O at 95% during 30 min at 80 ◦ C. The sealed samples were rinsed in water and dried with warm air. The ODP acid sealing was carried out using a solution of 0.1672 g L−1 ODP acid in ethanol prepared 24 h before the immersion. The specimens were immersed without agitation for 24 h at 23 ◦ C, then ultrasonically cleaned in ethanol for 10 min and dried at 40 ◦ C during 30 min. Finally, they were rinsed again in ethanol and dried with warm air. 2.4. Surface characterisation The specimens were analysed in surface plane and cross section by scanning electron microscopy (SEM) using a JEOL JSM-6400 microscope equipped with an Oxford Link attachment for energy dispersive X-ray (EDS) microanalysis. Phase identification was carried out by X-ray diffraction (XRD) using a Philips X’Pert instrument (Cu K␣ = 1.54056 Å). The XRD patterns were taken using Bragg-Brentano geometry in the 2␪ range from 10◦ to 90◦ with a step size of 0.04◦ and a dwell time of 5 s per step. The coating thickness was measured using a Fischer ISOSCOPE FMP10 portable gauge. The results presented were averaged over ten measurements taken at arbitrary locations. These values were also confirmed by the cross-sectional SEM analysis. Image analysis of the coating surface was carried out using three SEM micrographs taken at arbitrarily locations, at 500 × magnification. The micrographs were analysed using ImageJ software and the results presented provide the mean average values of surface porosity area fraction, pore diameter and pore density. The surface roughness was evaluated in five different locations following ISO 4287 standard [32], using a Surtronic 25 tester (Taylor Hobson Precision) and TalyProfile software, with a Gaussian filter of 0.25 mm applied. To evaluate coating wettability, the contact angle measurements were carried out using an FTA 1000/FTA instrument. Four drops of deionized water were analysed separately using FTA32 software. Fifty photographs taken at a frame rate of 0.5 s were analysed per each drop and the results presented results provide the mean average value of all measurements. 2.5. Corrosion evaluation 2.5.1. Salt fog tests The salt fog test of each surface treatment was carried during 7 days using a salt spray cabinet (CCI/CCM-MX) according to the standard ASTM B117 [33] by duplicate. The salt fog atmosphere was created at neutral pH using an aqueous solution of 5 wt.% NaCl, which was atomised keeping the pressure in the range 70–170 kNm−2 and the temperature at 35 ◦ C. After the tests, the specimens were rinsed with water at ∼ 37 ◦ C in order to remove solid salt deposits and dried in air.

655

Table 1 Affected area fraction by corrosion attack with their corresponding score as per ASTM D1654. Affected area (%)

Score

No corrosion 0–1 2–3 4–6 7–10 11–20 21–30 31–40 41–55 56–75 >75

10 9 8 7 6 5 4 3 2 1 0

In un-coated metals, corrosion rates are usually calculated by weight loss after removal of the corrosion products, but in the case of coated specimens this parameter is not possible to calculate since the removal of the corrosion products located at the substrate/coating interface is not possible without destroying the coating. Therefore, in order to evaluate the extent of corrosion attack, the surface appearance of the specimens after exposure to salt fog tests was analysed according to the standard ASTM D1654 [34]. Depending on the fraction of surface area affected by corrosion a score was given according to Table 1, wherein a higher score corresponds to a material exhibiting slighter corrosion attack, therefore, having a higher corrosion resistance. 2.5.2. Electrochemical impedance spectroscopy Electrochemical Impedance Spectroscopy (EIS) was used to evaluate the corrosion resistance of PEO coatings in an aqueous saline solution (NaCl 0.5 wt.%) at 23 ◦ C. For that a GillAC (ACM Instruments) computer-controlled potentiostat and a threeelectrode cell were used. The specimen was connected as a working electrode, with a graphite electrode and a silver–silver chloride (Ag/AgCl) electrode used as the counter and the reference electrode, respectively. The solution inside the reference electrode was KCl 3 M, which provides a potential of 0.210 V with respect to the standard hydrogen electrode. The tests applying a sinusoidal perturbation of 10 mV RMS amplitude in the frequency range of 30 kHz–0.01 Hz were carried out for up to 7 days of immersion. All measurements were triplicated to ensure reproducibility. The impedance data were fitted to an appropriate equivalent circuit using ZView software (Solartron Analytical) with a <10% error for each element and ␩2 values of <0.009. 3. Results and discussion 3.1. Characteristics of surface morphology and composition Fig. 1 shows surface plane and cross-sectional views of PEO coatings formed on AZ91 magnesium alloy in the two silicate based electrolytes, without (PEO–1) and with NaF additions (PEO–2). The surface microstructure formed as a consequence of gas evolution during the coating growth, and microcracks caused by residual stresses formed at the locations of the discharge channels, where the high temperature, pressure and partial evaporation of the electrolyte lead to a high stress state within the coating [35]. Both coatings exhibit a thin (∼0.6 ␮m) compact barrier layer and a porous outer layer, 10–13 ␮m in thickness. The main difference between the two coatings is the porosity (Table 2). PEO−2 shows a greater pore diameter and area fraction and a lower pore density compared to PEO–1. Kazanski et al. [36] observed a similar effect of fluoride ions on the coating porosity. They evaluated different concentration of KF in a silicate based electrolyte on the coating properties and observed

656

B. Mingo et al. / Applied Surface Science 433 (2018) 653–667

Fig. 1. Plan view and cross section SEM micrographs of (a, c) PEO−1 and (b, d) PEO−2.

Table 2 Porosity values of PEO−1 and PEO−2 coatings calculated by image analysis.

PEO−1 PEO−2

Surface porosity (area fraction)

Pore diameter (␮m)

Pore densityPores/cm2

6.4 ± 0.7 8.9 ± 0.9

1.05 ± 0.05 1.80 ± 0.10

(6.2 ± 0.4) × 106 (4.9 ± 0.3) × 106

that increasing concentrations of KF result in a lower number of larger pores. However, the addition of fluoride affects not only the coating porosity, but also the growth rate. It increases electrolyte conductivity and decreases the breakdown voltage resulting in a more efficient coating growth [14,37]. Some authors attribute this to a transition in the sparking regime, for instance Liang [13] stated that F− promotes the transition from a low density population of large micro-discharges to a higher number of smaller ones. In the present study, the coating thickness is only slightly affected by introduction of NaF in the electrolyte, increasing from 10.8 ± 0.5 ␮m to 13.4 ± 0.4 ␮m, although the treatment time was halved, indicating the higher growth rate PEO−2. EDS analysis (Table 3) showed that both coatings present a similar composition formed by elements from the substrate and the electrolyte, mainly O, Mg and Si in the form of MgO and Mg2 SiO4 , as deduced from XDR analysis (Fig. 2) and minor quantities of Al, Na and K. The NaF–containing coating (PEO−2), additionally presents in average ∼5 wt.% of F on its composition. EDS analysis also revealed that fluoride is preferentially incorporated to the inner part of the coatings, reaching concentrations of ∼10 wt.%. This has already been observed by other authors and it is related to the small size and high mobility of fluoride ions [36,38]. These are likely to be incorporated in the form of MgF2 [13,39], however this compound was not identified by XRD, probably due to its low relative amount [11].

Fig. 3 shows cross-sectional and surface plane views of PEO−1 and PEO–2 coatings after sealing post-treatments. The cerium based sealing results in the formation of a flake-shaped superficial layer which covers the whole PEO coating. The cracks observed in this film are probably due to the dehydration of the cerium conversion coating. Content of Ce within this layer varies depending on the location as it can be easily observed in the brighter areas in the micrograph taken in the backscattered electron mode (Fig. 3b). Cerium is mainly located in central regions of the flakes where it is accumulated in the form of a powder-like compound, probably consisting of CeO2 as identified by XRD analysis. At these locations, Ce concentration reaches ∼20 wt.%. These compounds were also formed inside the open pores blocking them partially (Fig. 3a), however they did not penetrate any deeper, so the barrier layer was kept intact as deduced from the EDS analysis. The XRD analysis also revealed formation of Mg(OH)2 during the sealing post-treatment, probably due to hydration reaction of MgO [40]. It is worth mentioning that the coating PEO−2 contains a higher amount of Ce in this superficial layer. This could be related to the presence of a greater porosity in the coating. After the Ce-based post-treatment, a change in the coating colour from white-greyish to orange-yellowish was observed. This is characteristic of Ce rich compounds, with the colour shade depending on the oxidation state of the element. In the case of Ce3+ , the colour resembles lightyellow and in the case of Ce4+ it is shifted towards orange [41].

B. Mingo et al. / Applied Surface Science 433 (2018) 653–667

657

Table 3 EDS analysis of the cross sections of PEO−1 and PEO−2. Elements (wt. %) O

Na

Mg

Al

Si

K

F

PEO−1

Point 1 Point 2 Point 3 Point 4

43.43 42.92 41.46 30.39

1.07 – – –

29.15 30.29 30.86 49.93

1.40 1.47 1.85 4.51

21.05 22.07 23.98 15.17

3.89 3.25 1.85 –

– – – –

PEO−2

Point 1 Point 2 Point 3 Point 4 Point 5

37.29 33.79 30.36 26.05 18.77

1.21 0.66 0.85 – –

34.39 36.60 38.62 41.99 57.07

3.65 4.02 3.97 3.92 9.61

19.14 21,21 20.04 19.12 6.27

1.65 1.29 1.25 1.34 0.29

2.67 2.42 4.91 7.58 8.00

In the case of sodium stannate sealing, a similar response was found in both PEO−1 and PEO−2 coatings. Stannate precipitated on the coating in the form of spherical granules with uniform dimensions of 0.63 ± 0.04 ␮m in diameter. Bagalà et al. [42] studied the influence of immersion time and temperature in a similar post-treatment on a PEO coated magnesium alloy and found that increasing these two parameters, both the granule size and the surface coverage increase. Additionally, Elsentriecy et al. [43] observed that this effect was promoted with increasing solution pH. EDS analysis of the granules showed that they mainly contain Mg, Sn and O, forming MgSn(OH)6 with minor amounts of SnO, as deduced from the XRD analysis (Fig. 2). However, further analysis using other characterization techniques such as XPS would be necessary to confirm the presence of these species. Similar to the cerium based sealing, Mg(OH)2 compounds were also identified on the surface of stannate sealed coatings. In the present study, the spherical granules do not cover the coating surface completely since the features of original PEO coatings are still visible following the post-treatment. However, they penetrate deeper into the coating, partially blocking pores and cracks in it, as can be observed at the cross-sectional micrograph (Fig. 3c). EDS analysis revealed that the maximum of Sn was ∼8 wt.% for both coatings, with little influence of NaF. The ODP acid sealing differs considerably from the cerium and stannate based post-treatments where Ce- and Sn-rich compound precipitate heterogeneously at the coating surface. In this case a thin layer, probably a monolayer [44], of ODP acid is deposited on the coating surface. The phosphonic acid, as the functional group of ODP, interacts with the coating surface, with the remaining hydrocarbon chain arranged towards the solution. However, low atomic weights, non-cristallinity and small amounts of the forming elements (P, O, C and H) make them difficult to be identified by the analytical techniques employed in this study. It is not therefore surprising that the XRD analysis has revealed a similar pattern compared to the untreated coating. Unlike the previous sealing post-treatments, Mg(OH)2 was not formed since the process was carried out in the alcohol based solution. Although the identification of the ODP acid layer was not possible by the used characterisation techniques, its presence was evident as deduced from the change on the coating’s properties (hydrophobic properties and corrosion resistance) as discussed in Sections 3.3 and 3.4.

3.2. Sealing mechanisms

Fig. 2. X-ray diffraction patterns of PEO−1, PEO−2, PEO−1−Ce, PEO−1−Sn and PEO−1−ODP.

3.2.1. Cerium based sealing The sealing mechanism of PEO coatings using cerium salt based solutions is based on a dissolution-precipitation process, where compounds from the coating (MgO and Mg2 SiO4 ) are dissolved in the first stage, leading to a local pH increase which promotes the precipitation of cerium oxide/hydroxide compounds. Considering

658

B. Mingo et al. / Applied Surface Science 433 (2018) 653–667

presence of H2 O2 in the sealing medium, the following mechanism is proposed (Fig. 4a): Step 1, Ce3+ ions from the sealing solution are oxidised by H2 O2 (refer Ref. [41] for Eq. (1)): 3+

2Ce

−

(aq) + H2 O2 (aq) + 2OH (aq) →

2Ce(OH)2+ 2

(aq)

(1)

Step 2, MgO and MgSiO4 are partially dissolved in acidic media (refer Ref. [45] for Eqs. (2) and (3): MgO (s) + 2H+ (aq) → Mg2+ (aq) + H2 O (aq) +

Mg2 SiO4 (s) + 4H (aq) → 2Mg

2+

(2)

(aq) + SiO2 (aq) + 2H2 O (aq) (3)

Both reactions (2) and (3) consume protons, which lead to a local alkalinisation at the proximity of the coating surface. Step 3, Ce(OH)2 2+ ions precipitate in the form of CeO2 due to the local pH increase (refer Ref. [41] for Eq. (4): − Ce(OH)2+ 2 (aq) + 2OH (aq) → Ce(OH)4 (s) → CeO2 (s) + 2H2 O

(4)

Probably remaining Ce3+ ions from the solution also precipitate in the form of Ce(OH)3 and Ce2 O3 . The SiO2 (aq) formed in step (3), partially precipitates back into the coating as it can be deduced from the slight increase in the Si/Mg ratio (calculated from EDS analysis) in the coatings after the

Ce-based post-treatment, namely 0.75 and 0.61 for PEO-1-Ce and PEO-2-Ce and 0.62 and 0.45 for PEO-1 and PEO-2, respectively. 3.2.2. Stannate based sealing The sealing mechanism of the stannate based post-treatment has been adapted from previous works based on stannate conversion coatings [46], since the studies employing this type of solution as a sealant are rather limited. This mechanism is also based on a dissolution-precipitation sequence of chemical reactions, where partial dissolution of the coating in the first step is followed by the precipitation of Sn-rich compounds at the surface (Fig. 4b): Step1, MgO and Mg2 SiO4 compounds from the PEO coating are partially dissolved in the sealing solution resulting in the formation of Mg2+ ions according to reactions (5) and (3) (refer Ref. [40] for Eq. (5), refer Ref. [45] for Eq. (3)). MgO (s) + H2 O (aq) → Mg2+ (aq) + 2OH− (aq) +

Mg2 SiO4 (s) + 4H (aq) → 2Mg

2+

(aq) + SiO2 (aq) + 2H2 O

(5) (3)

It is worth mentioning that the dissolution rate of Mg2 SiO4 decreases considerably in alkaline conditions [47], which is the media (pH = 12.5) at which this process was carried out. In order to ensure the formation of Mg2+ , the sealing post-treatment was

Fig. 3. Plan view and cross section SEM micrographs of PEO coatings after sealing post-treatment: (a, b) PEO−1−Ce, (c, d) PEO−2−Ce, (e, f) PEO−1−Sn, (g, h) PEO−2−Sn, (i, j) PEO−1−ODP and (k, l) PEO−2−ODP.

B. Mingo et al. / Applied Surface Science 433 (2018) 653–667

659

Fig. 3. (Continued)

carried out at 80 ◦ C since temperature promotes the dissolution kinetics. Step 2, Mg2+ from reactions (3) and (5) react with Sn(OH)6 2− ions from the hydrolysis of Na2 SnO3 which leads to the precipitation of MgSn(OH)6 (refer Ref. [48] for Eq. (6), refer Ref. [46] for Eq. (7)): Na2 SnO3 (s) + 3H2 O (aq) → Na2 Sn(OH)6 (aq) → 2Na+ (aq) +Sn(OH)2− 6 (aq)

Mg2+ (aq) + Sn(OH)2− 6 (aq) → MgSn(OH)6 (s)

(6)

(7)

Probably reaction (7) takes place at the vicinity of the coating surface, where there is a higher concentration of Mg2+ ions both in dissolution and adsorbed at the coating surface coming from reaction (5).

XRD analysis revealed the presence of SnO, which might be the result of secondary reactions, where the Sn(OH)6 2− is partially reduced.

3.2.3. ODP based sealing The ODP based sealing mechanism differs considerably from the previous two based on precipitation of solid compounds at the coating surface, partially blocking the porosity. In this case, the organic compound ODP acid is chemisorbed at the coating surface. Guerrero [49] suggested that phosphonic acid group of the ODP is capable of creating P O Metal bonds through condensation reactions between the P OH groups from the phosphonic acid with HO Metal groups from the coating surface and/or coordination covalent bonds P = O → Metal. Additionally, hydrogen bonds can also be formed between P = O and P OH groups from the phosphonic acid and hydroxyl groups from the coating surface. Therefore, the phosphonic functional head of the ODP acid interacts with the

660

B. Mingo et al. / Applied Surface Science 433 (2018) 653–667

Fig. 4. Proposed mechanisms for (a) Ce-based sealing, (b) Sn-based sealing and (c) ODP-based sealing.

Table 4 Roughness values of PEO coatings before and after sealing post-treatment. PEO−1 Post-treatment – Cerium-based sealing (Ce) Stannate-based sealing (Sn) ODP acid sealing (ODP)

Ra (␮m) 0.83 ± 0.15 0.78 ± 0.06 0.80 ± 0.04 0.84 ± 0.07

Table 5 Contact angle values of PEO coatings before and after sealing post-treatment.

PEO–2 Rz (␮m) 6.2 ± 0.9 5.9 ± 0.1 5.4 ± 0.7 5.7 ± 0.4

Ra (␮m) 0.99 ± 0.07 0.85 ± 0.05 0.92 ± 0.11 0.95 ± 0.09

PEO–1 Rz (␮m) 6.7 ± 0.6 5.2 ± 0.2 6.7 ± 0.9 6.3 ± 0.2

coating and the hydrophobic hydrocarbon chain of the organic acid is arranged towards the solution (Fig. 4c). 3.3. Roughness and contact angle measurements Table 4 shows the roughness values of the coatings. All coatings present Ra in the range of 0.78–0.99 ␮m. It can be observed that the addition of NaF (PEO−2) increases the roughness values compared to PEO−1. Regarding the effect of post-treatments, none of the sealing modify greatly these characteristics, although they tend to slightly decrease the surface roughness since in all cases a superficial layer has been formed which covers to a greater or lesser extent the crater-like morphology of PEO coatings. Contact angle measurements are presented in Table 5. Both PEO–1 and PEO–2 coatings are highly hydrophilic, showing con-

Post-treatment – Cerium-based sealing (Ce) Stannate-based sealing (Sn) ODP acid sealing (ODP)

PEO–2 ◦

Contact angle ( ) 10 ± 2 82 ± 4 6±2 46 ± 3

Contact angle (◦ ) 15 ± 3 68 ± 6 5±1 79 ± 4

tact angles <10◦ . This is related to the presence of hydroxyl groups formed as a consequence of hydration reactions of the coatings components, which promotes specific hydrophilic interactions. Also, the high surface porosity might promote the penetration of water through the coating. The application of cerium based and ODP acid post-treatments increase significantly the contact angle, although the values are still below 90◦ , which is considered to be a threshold for hydrophobic surfaces. In the case of the Ce-based sealing (PEO−1−Ce), this increase is probably related to the chemical modification of the coating surface. In this case the surface is mainly covered by CeO2 (4) whose absence of hydroxyl groups hinders water interactions and, therefore, prevents water permeability. It is worth mentioning that PEO−2−Ce, shows a slightly lower contact angle compared to PEO−1−Ce and this could be explained by two factors. First of all, it could be related to the increased porosity

B. Mingo et al. / Applied Surface Science 433 (2018) 653–667 Table 6 Fraction of surface area affected by corrosion after 7 d of exposure to the salt fog test and the score of corrosion resistance for the studied materials according to ASTM D1654. Specimen

Affected surface area (%)

Score

PEO-1 PEO-2 PEO-1-Ce PEO-2-Ce PEO-1-Sn PEO-2-Sn PEO-1-ODP PEO-2-ODP

13.2 ± 2.6 9.2 ± 2.4 12.6 ± 2.0 1.9 ± 1.3 7.1 ± 1.8 5.2 ± 0.5 11.2 ± 2.9 1.1 ± 0.1

5 6 5 8 6 7 5 8

in the cracked layer in case of PEO−2−Ce. Secondly, reaction (4) kinetics will depend on reaction (1), i.e. on availability of OH− . As it will be shown in the corrosion test section, the NaF-based coating is more corrosion resistant, then there is less OH− available to form CeO2 , which may result in a CeO2 layer not as uniform as the one on PEO−1, or most of it forms inside the pores and not on the surface, so it is not as hydrophobic as expected. The Sn-based sealing does not improve the hydrophobic properties of the coating. Probably, the hydrated nature of the Sn-based sealing which is mainly composed by MgSn(OH)6 (7) facilitates water interaction thought the hydroxyl groups. In the case of ODP acid post-treatments, there is an increase in the contact angle and it is related to the presence of a monolayer of a hydrophobic chain at the surface of the coatings which impedes water interaction and not to the sealing of the coating’s porosity. Probably the PEO coating surface was not completely covered by the ODP acid that is why a further improvement is not achieved. It is worth mentioning that in this case, the presence of NaF resulted in a significant increase in the contact angle values, which evidenced that the sealing mechanism is not related to the sealing of the pores. Probably the increased surface roughness of PEO−2−ODP promotes the formation of a denser ODP layer over the coating which prevents the water access. 3.4. Corrosion behaviour of coatings 3.4.1. Coatings resistance to salt fog tests Fig. 5 shows surface plane views of the studied coatings after 7 days of exposure to the salt fog conditions and Table 6 collects the surface area fraction affected by corrosion attack and their corresponding scores. All specimens suffer corrosion degradation to a greater or lesser extent, with PEO−2−Ce and PEO−2−ODP coatings exhibiting the least corrosion damage. The addition of NaF to the coating electrolyte (PEO−2) results in a slight increase of the corrosion resistance score compared to PEO−1 from 5 to 6. This is consistent with previous reports [50,51]. For instance, Apelfeldet al. [50] observed up to two orders of magnitude decrease in corrosion current density of PEO coatings formed on AZ41 magnesium alloy formed in electrolytes with NaF and KF additions. White and co-authors [51] also observed an improvement in the corrosion properties of PEO coatings formed on AZ31 alloy when KF was added to the electrolyte and they attributed such improvement to the formation of a passive layer enriched with MgF2 . All the sealing post-treatments improve the corrosion resistance compared to the as-received PEO coatings. According to this classification, the sealings that provide a better corrosion resistance are those predicted by visual analysis, such as PEO−2−Ce and PEO−2−ODP coatings that have the highest score 8. However, the same sealings applied to the NaF−free coatings (PEO−1−Ce and PEO−1−ODP) do not show significant effects on the corrosion resistance if compared to the unsealed PEO coating (PEO−1). In the case of the stannate based sealing, both coatings (PEO−1−Sn and

661

PEO−2−Sn) show a slight increase in the corrosion resistance, but it is not as significant as for the other two sealings. In the case of the cerium sealing, the improvement in the corrosion properties is related to three different factors. First of all, precipitation of CeO2 oxides during the post-treatment physically obstructs the pores, impeding the electrolyte penetration and therefore increasing the corrosion resistance. The positive influence of Ce-based treatments is also related to the release of Ce3+ ions that may show inhibiting properties. In particular, they react with OH− originated from the cathodic oxygen reduction reaction (2H2 O + O2 + 2e− → 4OH− ) causing precipitation of Ce(OH)3 /Ce(OH)4 /CeO2 xH2 O compounds at the cathodic sites and therefore inhibiting the electrochemical activity at the surface [52,53]. Some authors have also considered that Ce possesses selfheling properties comparable to those of chromate based coatings [54] based on Ce capacity to undergo redox reactions. With increasing pH in aerated conditions, the trivalent cerium can be easily oxidised to Ce4+ [55,56], which can migrate to corrosion initiation sites precipitating in the form of semi-protective oxide/hydroxide compounds [57–59]. Regarding the stannate based sealing, the slight improvement in the corrosion resistance is associated with just partial blockage of the porosity by MgSn(OH)6 and Mg(OH)2 compounds that hamper the diffusion of aggressive ions from the corrosive medium towards the substrate [60]. For this reason, the improvement in this case was not as significant as with the Ce-based sealing. The positive influence of ODP acid post-treatment is related to the formation of a uniform layer of ODP over the coatings surface. The arrangement of the aliphatic chain of the ODP acid towards the corrosive medium (Fig. 4c) impedes the penetration of the aggressive aqueous electrolyte, which was demonstrated with the increased contact angle values after the treatment (Table 5). Ishizakiet al. [61] formed a monolayer of ODP directly on the surface of AZ31 magnesium alloy and also found a considerable increase in hydrophobic properties, although the contact angles were lower than 90◦ too. They found substantial improvement of the corrosion resistance compared to the untreated alloy and attributed this to the partial passivation of magnesium surface due to precipitation of a semi-protective layer of corrosion products. In light of the results gathered in Table 6, it is evident that the presence of fluorine in the original PEO coatings influences significantly their corrosion properties after sealing post-treatments. In the case of Ce- and ODP acid-based sealings, a considerable improvement is observed in the corrosion score, from 5 to 8. In the case of the Sn-based sealing, the improvement was marginal. However, the reason for the fluoride effect on sealing quality in each post-treatment is different. In the case of the Ce-based post treatment, such improvement is mainly related to chemical modification of the surface composition. Fluoride promotes formation of fewer but larger pores, this facilitates the penetration of the sealing solution through the pores and promotes the precipitation of higher quantities of CeO2 . Higher amounts of Ce-rich compounds within the coating enhance the positive effect of Ce on the corrosion resistance, since more Ce3+ ions would be released into the solution, promoting precipitation of Ce-rich compounds at the cathodic sites on the surface. In the case of ODP acid sealing, the addition of NaF slightly increases the surface roughness of PEO−2−ODP coating compared to the PEO−ODP one; therefore, a larger area would be available for the ODP acid to bond with the coating surface. Additionally, the presence of MgF2 compounds probably promotes the ODP acid chemisorption though hydrogen bonds. These two factors lead to the formation of a denser and homogenous ODP film across the coating surface which protects the material against corrosion attack. The positive influence of NaF on the ODP acid sealing can also be associated with the increase in hydrophobic properties,

662

B. Mingo et al. / Applied Surface Science 433 (2018) 653–667

Fig. 5. Surface appearance of the PEO coatings and sealings before and after salt fog corrosion test.

with contact angle increasing from 46◦ to 79◦ , which is a common indicator of the improvement in corrosion properties [62]. 3.4.2. Electrochemical response of PEO coatings Fig. 6 shows Nyquist and Bode plots for the studied coatings after three hours and seven days of immersion in a 0.5 wt.% aqueous solution of NaCl. The equivalent circuit shown in Fig. 7a was used to fit the experimental data for PEO−1 and PEO−2 coatings after 3 h of immersion. Rs represents the resistance of electrolyte solution, Rout /CPEout and Rin/ CPEin loops account for the resistance and the capacitive behaviour of the outer and inner part of PEO coating, respectively. Constant phase elements (CPE) were used instead of capacitors in order to account the non-ideal behaviour arisen from the heterogeneous interfaces of the surface layers.

Two time constants associated with the outer and inner part of the coating can be identified in the Bode plot. The response at high frequencies (103 -104 Hz) is related to the outer porous layer of the coating while the medium-frequency (10 Hz) response can be attributed to the inner barrier layer. The corrosion resistance of the coatings is mainly provided by the inner barrier layer, therefore, analysing the Rin values of PEO−1 and PEO−2 calculated from EIS data fitting (Table 7), it is possible to state that NaF does not improve the protective properties of the studied PEO coatings, since an increase of Rin values was not observed. This was quite unexpected since the opposite trend was observed in the salt fog test and NaF usually influences positively the corrosion resistance of PEO coatings, since it usually promotes passivation [12]. In this case the addition of NaF affects drastically the porosity of the coatings,

B. Mingo et al. / Applied Surface Science 433 (2018) 653–667

663

Table 7 Fitted electrical parameters of EIS spectra after 3 h of immersion in 0.5 wt.% NaCl. Specimen

CPEout -T (S s−n cm−2 )

CPEout -n

Rout ( cm2 )

CPEin -T (S s−n cm−2 )

CPEin -n

Rin ( cm2 )

PEO−1 PEO−2 PEO−1−Ce PEO−2−Ce PEO−1−Sn PEO−2−Sn PEO−1−ODP PEO−2−ODP

5.41 × 10−7 9.21 × 10−8 2.87 × 10−6 5.29 × 10−7 2.46 × 10−7 1.02 × 10−7 3.03 × 10−7 5.63 × 10−8

0.7196 0.8270 0.8570 0.7327 0.7499 0.7726 0.76216 0.82542

1.25 × 104 1.75 × 105 2.02 × 106 1.33 × 106 3.04 × 105 7.67 × 104 2.96 × 104 6.75 × 103

1.19 × 10−6 3.22 × 10−7 4.54 × 10−6 9.55 × 10−6 3.88 × 10−7 2.24 × 10−7 2.10 × 10−6 7.05 × 10−9

0.8692 1.0000 0.9007 0.9900 0.9170 0.8720 0.6500 0.9900

2.71 × 106 1.85 × 105 3.06 × 105 8.20 × 105 1.42 × 105 2.23 × 105 1.46 × 105 8.58 × 105

resulting in an increased fraction of the surface area covered by pores, which increases the coating permeability. Probably, in the salt fog test, where the amount of electrolyte over the coating surface is considerably lower compared to the EIS measurements, the positive influence of NaF on the passive properties of the barrier layer prevails over the increased porosity. When increasing the immersion time to 7 days a decrease in the corrosion resistance is observed as deduced from the smaller diam-

CPEdl -T (S s−n cm−2 )

CPEdl -n

Rct ( cm2 )

2.18 × 10−6

0.5500

1.96 × 106

1.06 × 10−7

0.6630

4.81 × 106

eter of the capacitive loop in the Nyquist plot and in the lower values of modulus of impedance at low frequencies. In this case both outer and inner part of the coating behave as a single layer (Rcoat /CPEcoat ) and a new time constant (Rct /CPEdl ) appears at low frequencies (0.1–1 Hz), which is attributable to the corrosion activity at the metal surface. Here Rct accounts for the charge transfer resistance and CPEdl to the capacitive behaviour of the double layer at the metal/electrolyte interface. In this case the two time constants are

Fig. 6. Nyquist and Bode plot of (a, b, c) PEO−1 and PEO−2, (d, e, f) PEO−1−Ce and PEO−2−Ce, (g, h, i) PEO−1−Sn and PEO−2−Sn and (j, k, l) PEO−1−ODP and PEO−2−ODP after 3 h and 7 days of exposure to naturally aerated 0.5 wt.% NaCl solution.

664

B. Mingo et al. / Applied Surface Science 433 (2018) 653–667

Fig. 6. (Continued)

Fig. 7. Equivalent circuits used to fit the experimental EIS data: (a) PEO−1, PEO−2, PEO−1−Ce, PEO−2−Ce, PEO−1−Sn, PEO−1−ODP after 3 h of immersion; (b) PEO−1, PEO−2, PEO−1−Ce,PEO−2−Ce, PEO−1−Sn, PEO−1−ODP, PEO−2−ODP after 7 days of immersion, and (c) PEO−2−Sn after 3 and 7 days and PEO−2−ODP after 3 h of immersion.

difficult to distinguish due to overlapping. A similar response was observed for PEO coatings on Mg-Ca alloys in simulated body fluid [39]. Fig. 6b–d show the electrochemical response of PEO coatings after post-treatments and Fig. 8 compares the modulus of the

impedance values at 0.1 Hz of the studied materials which shows a strong relationship between the coating morphology and the efficiency of the post-treatments [27]. In the case of NaF-free coatings (PEO−1−Ce, PEO−1−Sn and PEO−1−ODP), the influence on the corrosion resistance of sealing post-treatments is not very

B. Mingo et al. / Applied Surface Science 433 (2018) 653–667

Fig. 8. Impedance modulus at 0.1 Hz of the studied coatings after 3 h and 7 days of exposure to naturally aerated 0.5 wt.% NaCl solution.

significant; however, the NaF-containing coatings (PEO−2−Ce, PEO−2−Sn and PEO−2−ODP) show a substantial improvement after the sealing post-treatment. For both as-received (PEO−1 and PEO−2) and Ce-based sealed coatings (PEO−1−Ce and PEO−2−Ce), similar equivalent circuits were used comprising the response of the outer and inner part of the coating at short immersion times (3 h, Fig. 7a) and a combined response of both layers and the response of electrochemical corrosion reactions at the metal interface at longer immersion times (7days, Fig. 7b). After 3 h of immersion, the PEO−1−Ce coating shows a similar corrosion resistance to the as-received coating as can be observed in Fig. 8. In the case of un-sealed coating, the immersion time of 3 h was not sufficient for the electrolyte to reach the metallic substrate; that is why the influence of the sealing was negligible. However, when increasing the immersion time up to 7 days a slight improvement could be observed. This is related to the presence of Ce-based compounds inside the pores, which obviously offers a greater resistance than hollow pores. Additionally, Ce-rich compounds have inhibiting properties against corrosion leading to the precipitation of Ce(OH)3 /Ce(OH)4 /CeO2 ·xH2 O at the cathodic sites. On the contrary, the electrochemical behaviour of PEO−2−Ce coating is considerably improved compared to the unsealed coating (PEO−2). This is related to the higher amount of Ce-rich compounds deposited both at the surface and inside the pores, which physically impedes the electrolyte access to the coating and offers a higher concentration of Ce3+ available for corrosion inhibition. When increasing the immersion time to 7 days a decrease of the corrosion resistance can be observed, although it is still higher compared to the unsealed coating. This behaviour is related to the partial dissolution of these Ce-rich compounds deposited in the pores which diminishes their blocking capacity and increases the permeability of the coating. The PEO−1−Sn coating shows similar features compared to the un-sealed (PEO−1) coating and the same equivalent circuits were used to fit the experimental data (Fig. 7a and b for 3 h and 7 d, respectively). Similar to the Ce-based sealing, the corrosion resistance of this coating is comparable to that of the unsealed coating. But the incorporation of NaF, PEO−2−Sn, improves the corrosion response as it can be observed in the higher values of impedance in Fig. 8. In this case the electrochemical response was fitted using the equivalent circuit shown in Fig. 7c since three time constants could be observed in the Bode plot. The response at high frequencies (103 to 104 Hz) is related to the outer porous layer, the one at medium frequencies (10 Hz) to the inner barrier layer and that at low frequencies (10−2 to 10−1 Hz) is associated with the electrochemical processes at the metal/electrolyte interface.

665

This equivalent circuit was also used by Mohedano et al. [59] to fit the corrosion response of a PEO coated AM50 magnesium alloy enriched with CeO2 nanoparticles. This improvement is related the precipitation of MgSn(OH)6 blocking the porosity. The total amount of Sn (∼8 wt.%) found in the coating by EDS analysis was considerably lower compared to the concentration of Ce (∼20 wt.%) in the Ce-based sealing. This could explain a slightly lower corrosion resistance provided by the Sn-based sealing. When increasing the immersion time from 3 h to 7 d, the similar equivalent circuit was used but the corrosion resistance was considerably lower, but in this case, the response was better than that of the Ce-based sealing. This is probably related to the lower solubility of MgSn(OH)6 compared to CeO2 compounds. The impedance response of ODP sealed coatings was very similar to those with Sn-based post-treatment. The NaF-free coating showed two time constants associated to the outer and inner part of the coating, Rout /CPEout and Rin/ CPEin , respectively (Fig. 7a), but when the immersion time increased to 7 days, these two show a combined response (Rcoat /CPEcoat ) (Fig. 7b) at medium frequencies (102 Hz) and a new time constant that accounts for electrochemical processes at the metal/electrolyte interface (Rct /CPEdl ) appears at low frequencies (10−2 –10−1 Hz). For both immersion times, the ODP post-treatment (PEO−1−ODP) does not improve the corrosion resistance compared to the unsealed coating (PEO−1); although at least some improvement was expected since increased contact angles suggested lower coating wettability and impeded electrolyte access to the metal surface. It may be that the surface was not completely covered by the ODP acid layer. A significant improvement in the corrosion resistance was observed for the coating formed in NaF-containing electrolyte (PEO−2−ODP). Similar to the PEO−2−Sn coating, the electrochemical response after 3 h of immersion could be fitted to the equivalent circuit containing three time constants (Fig. 7c). This coating possessed the best corrosion resistance after 3 h of immersion, which could be related to the formation of a denser ODP layer over the coating surface as a result of the increase in the surface roughness. This response is consistent with considerable increase in contact angle observed for such coatings following the ODP acid post-treatment. However at longer immersion times the corrosion resistance decreased dramatically. The partial dissolution of the most superficial layer of the PEO coating leads to the desorption of ODP monolayer [44] which, combined with the increased porosity caused by fluoride additions, results in an aggressive corrosion attack via a greater number of access points for the electrolyte to the metal-oxide interface. Thus, the three studied sealing post-treatments have little influence on the corrosion behaviour of NaF-free coatings (PEO−1). However, in the presence of NaF (PEO−2), the effect of the sealing post-treatments is boosted. In the case of the inorganic sealing post-treatments (Ce and Sn), such improvement is related to the physical blockage of the porosity which impedes the penetration of aggressive species into the coating. In the case of Ce, it is also associated with Ce inhibition capacity. In the case of organic ODP acid sealing, an improvement is only achieved at short immersion times where the ODP aliphatic chains hinder the access of the aqueous electrolyte to the surface. At longer immersion times, it shows even a lower resistance compared to the unsealed coating and this can be explained by the partial desorption of the ODP layer and increased surface porosity induced by the presence of NaF in the electrolyte, which increases the coating permeability. Therefore, the ODP acid sealing seems to be the most promising post-treatment amongst the studied ones, however it is necessary to improve its long term resistance. A possible approach to achieve this would be the chemical modification of the PEO coatings in order to promote the formation of a greater number of more

666

B. Mingo et al. / Applied Surface Science 433 (2018) 653–667

intense chemical interactions between the coating and the ODP which would ensure its chemisorption and therefore, its durability. 4. Conclusions – The addition of NaF to the electrolyte increases the size and area fraction of pores and slightly increases the roughness. NaF also increases the growth rate of the coating. PEO−1 and PEO−2 are mainly formed of MgO and Mg2 SiO4 and possess hydrophilic properties (contact angle <90◦ ). – The Ce- and Sn-based post-treatments consist on dissolution/precipitation reactions at the PEO surface which results in the formation of Ce- and Sn-rich compounds (CeO2 , MgSn(OH)6 , SnO) which physically obstruct the electrolyte access within the coating. The ODP acid sealing results in the formation of a thin layer of ODP chemisorbed to the coatings surface. – Sealing treatments based on Ce, Sn and ODP acid slightly improve the corrosion performance of NaF-free PEO coatings. However, their corrosion resistance is boosted in the coatings formed on NaF-containing electrolytes, which is mainly related to the chemical modification of the superficial layer and morphological changes induced by fluoride ions. At short immersion times, the ODP-based sealing provides the best corrosion resistance which is attributable to the improved hydrophobic properties provided by the ODP monolayer, although for longer exposure times, Ceand Sn-based post-treatments appear to perform better which is related to the physical blockage of the porosity. Acknowledgements The authors are grateful to the European Research Council under the ERC Advanced Grant (320879 ‘IMPUNEP’) and to MINECO/FEDER (Spain, Project MAT2015-66334-C3-3-R) for financial support. M. Mohedano is also grateful to MINECO for financial support via Young Researchers Challenges Programme (Proyectos Retos Jovenes Investigadores) (MAT2015-73355-JIN).

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21] [22]

[23]

[24]

[25] [26]

[27]

References [28] [1] T.S.N. Sankara Narayanan, I.S. Park, M.H. Lee, Strategies to improve the corrosion resistance of microarc oxidation (MAO) coated magnesium alloys for degradable implants: Prospects and challenges, Prog. Mater. Sci. 60 (2014) 1–71. [2] A. Nominé, J. Martin, G. Henrion, T. Belmonte, Effect of cathodic micro-discharges on oxide growth during plasma electrolytic oxidation (PEO), Surf. Coat. Technol. 269 (2015) 131–137. [3] H. Fakhr Nabavi, M. Aliofkhazraei, A. Sabour Rouhaghdam, Morphology and corrosion resistance of hybrid plasma electrolytic oxidation on CP-Ti, Surf. Coat. Technol. 322 (2017) 59–69. [4] A. Yerokhin, X. Nie, A. Leyland, A. Matthews, S. Dowey, Plasma electrolysis for surface engineering, Surf. Coat. Technol. 122 (1999) 73–93. [5] R. Arrabal, E. Matykina, T. Hashimoto, P. Skeldon, G.E. Thompson, Characterization of AC PEO coatings on magnesium alloys, Surf. Coat. Technol. 203 (2009) 2207–2220. [6] R.O. Hussein, P. Zhang, X. Nie, Y. Xia, D.O. Northwood, The effect of current mode and discharge type on the corrosion resistance of plasma electrolytic oxidation (PEO) coated magnesium alloy AJ62, Surf. Coat. Technol. 206 (2011) 1990–1997. [7] P. Bala Srinivasan, J. Liang, C. Blawert, M. Störmer, W. Dietzel, Effect of current density on the microstructure and corrosion behaviour of plasma electrolytic oxidation treated AM50 magnesium alloy, Appl. Surf. Sci. 255 (2009) 4212–4218. [8] V. Dehnavi, B.L. Luan, D.W. Shoesmith, X.Y. Liu, S. Rohani, Effect of duty cycle and applied current frequency on plasma electrolytic oxidation (PEO) coating growth behavior, Surf. Coat. Technol. 226 (2013) 100–107. [9] Y. Tang, X. Zhao, K. Jiang, J. Chen, Y. Zuo, The influences of duty cycle on the bonding strength of AZ31B magnesium alloy by microarc oxidation treatment, Surf. Coat. Technol. 205 (2010) 1789–1792. [10] Y. Gao, A. Yerokhin, A. Matthews, Effect of current mode on PEO treatment of magnesium in Ca- and P-containing electrolyte and resulting coatings, Appl. Surf. Sci. 316 (2014) 558–567. [11] S.V. Gnedenkov, O.A. Khrisanfova, A.G. Zavidnaya, S.L. Sinebryukhov, V.S. Egorkin, M.V. Nistratova, A. Yerokhin, A. Matthews, PEO coatings obtained on

[29]

[30]

[31]

[32]

[33] [34]

[35]

[36]

[37]

[38]

[39]

an Mg–Mn type alloy under unipolar and bipolar modes in silicate-containing electrolytes, Surf. Coat. Technol. 204 (2010) 2316–2322. E. Gulbrandsen, J. Taftø, A. Olsen, The passive behaviour of Mg in alkaline fluoride solutions. Electrochemical and electron microscopical investigations, Corros. Sci. 34 (1993) 1423–1440. J. Liang, B. Guo, J. Tian, H. Liu, J. Zhou, T. Xu, Effect of potassium fluoride in electrolytic solution on the structure and properties of microarc oxidation coatings on magnesium alloy, Appl. Surf. Sci. 252 (2005) 345–351. G.-H. Lv, H. Chen, X.-Q. Wang, H. Pang, G.-L. Zhang, B. Zou, H.-J. Lee, S.-Z. Yang, Effect of additives on structure and corrosion resistance of plasma electrolytic oxidation coatings on AZ91D magnesium alloy in phosphate based electrolyte, Surf. Coat. Technol. 205 (Suppl. 1) (2010) S36–S40. R. Arrabal, E. Matykina, P. Skeldon, G. Thompson, Incorporation of zirconia particles into coatings formed on magnesium by plasma electrolytic oxidation, J. Mater. Sci. 43 (2008) 1532–1538. M. Mohedano, C. Blawert, M. Zheludkevich, Silicate-based plasma electrolytic oxidation (PEO) coatings with incorporated CeO2 particles on AM50 magnesium alloy, Mater. Des. 86 (2015) 735–744. R. Arrabal, M. Mohedano, E. Matykina, A. Pardo, B. Mingo, M.C. Merino, Characterization and wear behaviour of PEO coatings on 6082-T6 aluminium alloy with incorporated ␣-Al2O3 particles, Surf. Coat. Technol. 269 (2015) 64–73. R. Arrabal, E. Matykina, P. Skeldon, G.E. Thompson, Coating formation by plasma electrolytic oxidation on ZC71/SiC/12p-T6 magnesium metal matrix composite, Appl. Surf. Sci. 255 (2009) 5071–5078. E. Martínez Campos, A. Santos-Coquillat, B. Mingo, R. Arrabal, M. Mohedano, A. Pardo, V. Ramos, J.L. López Lacomba, E. Matykina, Albumin loaded PEO coatings on Ti — potential as drug eluting systems, Surf. Coat. Technol. 283 (2015) 44–51. C.L. Chu, X. Han, F. Xue, J. Bai, P.K. Chu, Effects of sealing treatment on corrosion resistance and degradation behavior of micro-arc oxidized magnesium alloy wires, Appl. Surf. Sci. 271 (2013) 271–275. G.-L. Song, An irreversible dipping sealing technique for anodized ZE41 Mg alloy, Surf. Coat. Technol. 203 (2009) 3618–3625. N.V. Phuong, B.R. Fazal, S. Moon, Cerium- and phosphate-based sealing treatments of PEO coated AZ31 Mg alloy, Surf. Coat. Technol. 309 (2017) 86–95. U. Malayoglu, K.C. Tekin, S. Shrestha, Influence of post-treatment on the corrosion resistance of PEO coated AM50B and AM60B Mg alloys, Surf. Coat. Technol. 205 (2010) 1793–1798. P. Tian, X. Liu, C. Ding, In vitro degradation behavior and cytocompatibility of biodegradable AZ31 alloy with PEO/HT composite coating, Colloids Surf. B: Biointerfaces 128 (2015) 44–54. L. Pezzato, K. Brunelli, R. Babbolin, P. Dolcet, M. Dabal, Sealing of PEO coated AZ91 magnesium alloy using La-Based solutions, Int. J. Corros. 2017 (2017) 13. V. Dalmoro, J.H.Z. dos Santos, E. Armelin, C. Alemán, D. Schermann Azambuja, Phosphonic acid/silica-based films: a potential treatment for corrosion protection, Corros. Sci. 60 (2012) 173–180. M. Mohedano, E. Matykina, R. Arrabal, B. Mingo, A. Pardo, PEO of pre-anodized Al–Si alloys: corrosion properties and influence of sealings, Appl. Surf. Sci. 346 (2015) 57–67. J. Wang, J. Tang, Y. He, Top coating of low-molecular weight polymer MALPB used for enhanced protection on anodized AZ31 B Mg alloys, J. Coat. Technol. Res. 7 (2010) 737–746. M.B. Kannan, D. Gomes, W. Dietzel, V. Abetz, Polyoxadiazole-based coating for corrosion protection of magnesium alloy, Surf. Coat. Technol. 202 (2008) 4598–4601. M. Golabadi, M. Aliofkhazraei, M. Toorani, A.S. Rouhaghdam, Evaluation of La containing PEO pretreatment on protective performance of epoxy coating on magnesium, Prog. Org. Coat. 105 (2017) 258–266. S. Hutchins, P. Shashkov, V. Samsonov, A. Shatrov, New Generation of Keronite Process for Corrosion Protection of Magnesium, In: Magnesium, Wiley-VCH Verlag GmbH & Co KGaA, 2005, pp. 553–558. ISO, BS EN ISO 4287, Geometrical product specification (GPS) Surface texture: Profile method Terms, definitions and surface texture parameters: British Standards Institute, (2000). A. International, ASTM B117, Standard Practice for Operating Salt Spray (fog) Apparatus, 2003. A. International, ASTM D1654–08, Standard test method for evaluation of painted or coated specimens subjected to corrosive environments, West Conshohocken, PA (2008). R.H.U. Khan, A. Yerokhin, X. Li, H. Dong, A. Matthews, Surface characterisation of DC plasma electrolytic oxidation treated 6082 aluminium alloy: effect of current density and electrolyte concentration, Surf. Coat. Technol. 205 (2010) 1679–1688. B. Kazanski, A. Kossenko, M. Zinigrad, A. Lugovskoy, Fluoride ions as modifiers of the oxide layer produced by plasma electrolytic oxidation on AZ91D magnesium alloy, Appl. Surf. Sci. 287 (2013) 461–466. C.E. Barchiche, D. Veys-Renaux, E. Rocca, A better understanding of PEO on Mg alloys by using a simple galvanostatic electrical regime in a KOH–KF–Na3PO4 electrolyte, Surf. Coat. Technol. 205 (2011) 4243–4248. ´ R. Vasilic, ´ J. Radic-Peri ´ ´ M. Peric, ´ Characterization of plasma S. Stojadinovic, c, electrolytic oxidation of magnesium alloy AZ31 in alkaline solution containing fluoride, Surf. Coat. Technol. 273 (2015) 1–11. M. Mohedano, B.J.C. Luthringer, B. Mingo, F. Feyerabend, R. Arrabal, P.J. Sanchez-Egido, C. Blawert, R. Willumeit-Römer, M.L. Zheludkevich, E.

B. Mingo et al. / Applied Surface Science 433 (2018) 653–667

[40] [41]

[42]

[43]

[44]

[45] [46]

[47]

[48]

[49]

[50]

[51]

Matykina, Bioactive plasma electrolytic oxidation coatings on Mg-Ca alloy to control degradation behaviour, Surf. Coat. Technol. 315 (2017) 454–467. L.F. Amaral, I.R. Oliveira, P. Bonadia, R. Salomão, V.C. Pandolfelli, Chelants to inhibit magnesia (MgO) hydration, Ceram. Int. 37 (2011) 1537–1542. F. Scholes, C. Soste, A. Hughes, S. Hardin, P. Curtis, The role of hydrogen peroxide in the deposition of cerium-based conversion coatings, Appl. Surf. Sci. 253 (2006) 1770–1780. P. Bagalà, F.R. Lamastra, S. Kaciulis, A. Mezzi, G. Montesperelli, Ceria/stannate multilayer coatings on AZ91D Mg alloy, Surf. Coat. Technol. 206 (2012) 4855–4863. H.H. Elsentriecy, K. Azumi, H. Konno, Effects of pH and temperature on the deposition properties of stannate chemical conversion coatings formed by the potentiostatic technique on AZ91 D magnesium alloy, Electrochim. Acta 53 (2008) 4267–4275. T. Ishizaki, M. Okido, Y. Masuda, N. Saito, M. Sakamoto, Corrosion resistant performances of alkanoic and phosphonic acids derived self-assembled monolayers on magnesium alloy AZ31 by vapor-phase method, Langmuir 27 (2011) 6009–6017. F. Wang, D.E. Giammar, Forsterite dissolution in saline water at elevated temperature and high CO2 pressure, Environ. Sci. Technol. 47 (2012) 168–173. N. Yang, Q. Li, F. Chen, P. Cai, C. Tan, Z. Xi, A solving-reprecipitation theory for self-healing functionality of stannate coating with a high environmental stability, Electrochim. Acta 174 (2015) 1192–1201. F.K. Crundwell, The mechanism of dissolution of minerals in acidic and alkaline solutions: part II Application of a new theory to silicates, aluminosilicates and quartz, Hydrometallurgy 149 (2014) 265–275. Y.-I. Choi, S. Salman, K. Kuroda, M. Okido, Synergistic corrosion protection for AZ31 Mg alloy by anodizing and stannate post-sealing treatments, Electrochim. Acta 97 (2013) 313–319. G. Guerrero, J.G. Alauzun, M. Granier, D. Laurencin, P.H. Mutin, Phosphonate coupling molecules for the control of surface/interface properties and the synthesis of nanomaterials, Dalton Trans. 42 (2013) 12569–12585. A. Apelfeld, B. Krit, V. Ludin, N. Morozova, B. Vladimirov, R.Z. Wu, The characterization of plasma electrolytic oxidation coatings on AZ41 magnesium alloy, Surf. Coat. Technol. 322 (2017) 127–133. L. White, Y. Koo, S. Neralla, J. Sankar, Y. Yun, Enhanced mechanical properties and increased corrosion resistance of a biodegradable magnesium alloy by plasma electrolytic oxidation (PEO), Mater. Sci. Eng.: B 208 (2016) 39–46.

667

[52] L. Ecco, M. Fedel, F. Deflorian, J. Becker, B.B. Iversen, A. Mamakhel, Waterborne acrylic paint system based on nanoceria for corrosion protection of steel, Prog. Org. Coat. 96 (2016) 19–25. [53] S.A. Hayes, P. Yu, T.J. O’Keefe, M.J. O’Keefe, J.O. Stoffer, The phase stability of cerium species in aqueous systems i. e-ph diagram for the ce hclo 4 h 2 o system, J. Electrochem. Soc. 149 (2002) C623–C630. [54] J. Mosa, N. Rosero-Navarro, M. Aparicio, Active corrosion inhibition of mild steel by environmentally-friendly Ce-doped organic−inorganic sol−gel coatings, RSC Adv. 6 (2016) 39577–39586. [55] A. Aldykiewicz, A. Davenport, H. Isaacs, Studies of the formation of cerium-Rich protective films using X-Ray absorption near-Edge spectroscopy and rotating disk electrode methods, J. Electrochem. Soc. 143 (1996) 147–154. [56] K.A. Yasakau, M.L. Zheludkevich, S.V. Lamaka, M.G. Ferreira, Mechanism of corrosion inhibition of AA2024 by rare-earth compounds, J. Phys. Chem. B 110 (2006) 5515–5528. [57] M. Fedel, A. Ahniyaz, L.G. Ecco, F. Deflorian, Electrochemical investigation of the inhibition effect of CeO2 nanoparticles on the corrosion of mild steel, Electrochim. Acta 131 (2014) 71–78. [58] K. Aramaki, Preparation of chromate-free, self-healing polymer films containing sodium silicate on zinc pretreated in a cerium(III) nitrate solution for preventing zinc corrosion at scratches in 0.5 M NaCl, Corros. Sci. 44 (2002) 1375–1389. [59] M. Mohedano, C. Blawert, M.L. Zheludkevich, Cerium-based sealing of PEO coated AM50 magnesium alloy, Surf. Coat. Technol. 269 (2015) 145–154. [60] H. Huo, Y. Li, F. Wang, Corrosion of AZ91D magnesium alloy with a chemical conversion coating and electroless nickel layer, Corros. Sci. 46 (2004) 1467–1477. [61] T. Ishizaki, K. Teshima, Y. Masuda, M. Sakamoto, Liquid phase formation of alkyl- and perfluoro-phosphonic acid derived monolayers on magnesium alloy AZ31 and their chemical properties, J. Colloid Interface Sci. 360 (2011) 280–288. [62] S.V. Gnedenkov, V.S. Egorkin, S.L. Sinebryukhov, I.E. Vyaliy, A.S. Pashinin, A.M. Emelyanenko, L.B. Boinovich, Formation and electrochemical properties of the superhydrophobic nanocomposite coating on PEO pretreated Mg–Mn–Ce magnesium alloy, Surf. Coat. Technol. 232 (2013) 240–246.