Chromate-Free Corrosion Protection Strategies for Magnesium Alloys—A Review: Part II—PEO and Anodizing

Although hexavalent chromium-based protection systems are effective and their long-term performance is well understood, they can no longer be used due to their proven Cr(VI) toxicity and carcinogenic effect. The search for alternative protection technologies for Mg alloys has been going on for at least a couple of decades. However, surface treatment systems with equivalent efficacies to that of Cr(VI)-based ones have only begun to emerge much more recently. It is still proving challenging to find sufficiently protective replacements for Cr(VI) that do not give rise to safety concerns related to corrosion, especially in terms of fulfilling the requirements of the transportation industry. Additionally, in overcoming these obstacles, the advantages of newly introduced technologies have to include not only health safety but also need to be balanced against their added cost, as well as being environmentally friendly and simple to implement and maintain. Anodizing, especially when carried out above the breakdown potential (technology known as Plasma Electrolytic Oxidation (PEO)) is an electrochemical oxidation process which has been recognized as one of the most effective methods to significantly improve the corrosion resistance of Mg and its alloys by forming a protective ceramic-like layer on their surface that isolates the base material from aggressive environmental agents. Part II of this review summarizes developments in and future outlooks for Mg anodizing, including traditional chromium-based processes and newly developed chromium-free alternatives, such as PEO technology and the use of organic electrolytes. This work provides an overview of processing parameters such as electrolyte composition and additives, voltage/current regimes, and post-treatment sealing strategies that influence the corrosion performance of the coatings. This large variability of the fabrication conditions makes it possible to obtain Cr-free products that meet the industrial requirements for performance, as expected from traditional Cr-based technologies.


Foreword
Magnesium is a lightweight and versatile material that has become increasingly important as a structural material in light of global warming and the urgent need for weight reduction. It is able to provide sustainable and affective green corrosion protection solutions in transport, airspace, electronics, and consumer goods sectors. Due to its highly negative standard electrode potential (E • = −2.37 V SHE ), Mg is susceptible to atmospheric corrosion if left unprotected [1,2]. This review is motivated by the search for an effective alternative to Cr-based surface treatments for Mg alloys that has been going on for the last two decades, spurred by constantly tightening environment-concerned regulations worldwide. Voltage is a crucial factor that determines the final characteristics of the anodic film There is a certain potential threshold, i.e., a so-called dielectric breakdown potentia above which the mechanism of the electrochemical oxidation process changes (Error! Ref erence source not found. 2). Below the breakdown potential, the growth of the oxide coating is relatively slow and the obtained structure is mostly amorphous. In the case of Mg and its alloys, it i  [11] with permission from Elsevier. Reprinted from [11] with permission from Elsevier.
Voltage is a crucial factor that determines the final characteristics of the anodic film. There is a certain potential threshold, i.e., a so-called dielectric breakdown potential, above which the mechanism of the electrochemical oxidation process changes (Error! Reference source not found. 2). Below the breakdown potential, the growth of the oxide coating is relatively slow, and the obtained structure is mostly amorphous. In the case of Mg and its alloys, it is typically composed of hydrated magnesium oxide with a thin barrier layer adjacent to the   [49] and three-layer (c-e) [53] PEO coatings on Mg alloys. Schematic diagrams of two-layer and three-layer PEO coatings (f). (a,b) Reprinted from [49], (c-e) reprinted from [53] with permission from Elsevier.
The anti-corrosion properties of PEO coatings are mainly attributed to their submicrometric barrier layer (Error! Reference source not found. 5e) [54,55]. Nevertheless, the middle layer, which often develops under AC regimes, also contributes to the corrosion resistance to a certain extent, since its pores are not penetrating and mostly not interconnected, limiting species diffusion.  [49] and three-layer (c-e) [53] PEO coatings on Mg alloys. Schematic diagrams of two-layer and three-layer PEO coatings (f). (a,b) Reprinted from [49], (c-e) reprinted from [53] with permission from Elsevier. usually exhibits a crystalline structure, a thicker barrier layer, and a composition that is strongly dependent on the electrolyte constituents, while the latter is easily incorporated during the treatment through short-circuit paths [56]. In alkaline environments, PEOformed MgO undergoes partial conversion to Mg(OH)2, which has a Pilling-Bedworth ratio of 1.77 [57], resulting in greatly improved surface coverage. The ability of PEO processes to incorporate the electrolytes species opens new possibilities for coating design and obtaining the desired properties of surface-treated Mg alloys ( Figure 6). For instance, in the presence of aluminum compounds, a very stable MgAl2O4 spinel can be formed in the coating in addition to magnesium oxide [58]. This spinel is characterized by excellent chemical, thermal, dielectric, and mechanical features and also has a higher Pilling-Bedworth ratio than MgO (1.3 [59]), which is crucial for providing good corrosion protection. Additionally, PEO coatings on Mg alloys offer advantages like high thickness, fast growth rate, and good paintability, enabled by the relatively rough and porous top part of the coating. Mechanical properties like hardness and wear resistance can also be improved by PEO treatment [6,60]. It is worth noting that, unlike in pre-breakdown anodizing, PEO processing does not require special surface pre-treatment, Due to the much more desirable properties of the PEO coatings compared with prebreakdown anodic films on Mg, the last two decades have witnessed a surge of publications on PEO research in China, UK, Russia, Germany, France, Spain, and Turkey [39,49,56,58,[62][63][64][65][66][67], associated with the strategic interest of lightweight materials and the need of their protection and functionalization.

Cr-Based Anodic Treatments
As mentioned above, chromium-based protective coatings have been used for a long time. The first satisfactory method of Mg anodic treatment was developed in 1937, and it was Cr-based. The electrolyte was composed of 10% sodium dichromate (Na 2 Cr 2 O 7 ) and 2-5% monosodium phosphate (NaH 2 PO 4 ). The process was carried out at a constant current of 0.5-1 A m −2 , 50 • C, for 30-60 min. After painting, the anodic coating exhibited better anti-corrosion properties than a traditional CC obtained by pickling in highly concentrated Cr-based solution (180 g/L of Na 2 Cr 2 O 7 ·2H 2 O with 190 mL/L of HNO 3 ) and painting [68]. One of the oldest commercial anodizing processes for Mg, DOW17 (1942), which is still commonly used in industrial treatments, also owes its anti-corrosion properties to the presence of Na 2 Cr 2 O 7 in the electrolyte, with the other components of the concentrated electrolyte being NH 4 HF 2 and H 3 PO 4 (pH~5) [69,70]. The DOW17 procedure produces 5-75 µm-thick crystalline coatings containing MgF 2 , NaMgF 3 , Mg x+y/2 O x (OH) y , with some small amounts of Cr 2 O 3 (Figure 7a), by applying AC or DC at a voltage below 100 V and at a temperature above 70 • C [60,69,71].
which is still commonly used in industrial treatments, also owes its anti-corrosion properties to the presence of Na2Cr2O7 in the electrolyte, with the other components of the concentrated electrolyte being NH4HF2 and H3PO4 (pH~5) [69,70]. The DOW17 procedure produces 5-75 µm-thick crystalline coatings containing MgF2, NaMgF3, Mgx+y/2Ox(OH)y, with some small amounts of Cr2O3 (Figure 7a), by applying AC or DC at a voltage below 100 V and at a temperature above 70 °C [60,69,71].
An example of the DOW17 coating structure is presented in Figure 7b. Note that its morphology is strikingly similar to that of PEO coatings. As can be seen, the obtained coating is quite porous and irregular and has significant cracks; nevertheless, the process has been widely used because, for a very long time, there was no better alternative. Figure 7. (a) X-ray diffraction pattern [71] and (b) backscattered electron cross-sectional micrograph of DOW17 coating on AZ91 alloy [72], the latter produced by applying constant current with the voltage rising to 200 V.  [71] and (b) backscattered electron cross-sectional micrograph of DOW17 coating on AZ91 alloy [72], the latter produced by applying constant current with the voltage rising to 200 V.
An example of the DOW17 coating structure is presented in Figure 7b. Note that its morphology is strikingly similar to that of PEO coatings. As can be seen, the obtained coating is quite porous and irregular and has significant cracks; nevertheless, the process has been widely used because, for a very long time, there was no better alternative.
Surprisingly, despite the search for environmentally friendly alternatives, Cr (VI)containing processes are still being developed. For instance, recently, DC anodizing of a Mg-Li alloy was reported in a mixture of K 2 Cr 2 O 7 and (NH 4 ) 2 SO 4 (pH 5.5, 24 • C, 60 min) [70]. The corrosion-resistant coating was obtained due to reactions (1)- (2): In the presence of chromate, the protective coating is formed by a chemical reaction between the electrochemically oxidized Mg and hexavalent chromium, with the latter being partially reduced to the trivalent state. The excellent corrosion performance of anodic and conversion chromium-based coatings is related to several factors. Studies on CCC have shown that after drying samples treated in Cr(VI) solution, dehydration of chromium hydroxide occurs, leading to the formation of amphoteric Cr 2 O 3 [73]. Cr 2 O 3 is almost insoluble in water, slightly soluble in acids and alkalis [74], and is characterized by a corundum structure and high hardness (almost 9 on the Mohs scale). Moreover, the presence of trapped hexavalent chromium in the form of either salts or oxides is responsible for the self-healing ability of the coating during corrosion [75]. Additionally, some studies have suggested that the presence of polar oxo-Cr(VI) anions prevents the adsorption of de-passivating anions like chlorides [76].

Cr-Free Anodic Treatments: State-of-the-Art Technologies and Patents
Since Cr(VI) is highly toxic and carcinogenic, Cr-based industrial processes must be replaced by new, environmentally-friendly technologies. Further, there is an unceasing need to improve the functional properties of coatings and to reduce fabrication costs, which both affect the economic sustainability of Mg-based components. The majority of the alternative anodizing technologies adopt safe, alkaline-based and chromate-free electrolytes. The optimal electrolyte compositions and process parameters that provide the best coating properties are being intensively researched.
One of the first commercial treatments which departed from the use of chromates was HAE (1955) [77]. The highly alkaline electrolyte (pH~14), composed of KOH, Al(OH) 3 , K 2 F 2 , Na 3 PO 4 , and K 2 MnO 4 , operates at 20-30 • C. The treatment is carried out under an AC regime at 1.5-2.5 A dm −2 and, in most cases, the final voltage does not exceed 125 V. The coating thickness ranges between 5-75 µm and is determined by the cut-off voltage limit (60 V for thinnest coatings). The processing time varies between 7 and 60 min for thin and thick coatings, respectively [60]. The quality of the obtained coatings is reported to surpass that of DOW17 [78].
Modern commercial developments in Mg anodizing technology offer anti-corrosion properties that are at least comparable to or better than older, acidic chromate-based processes like DOW17 or fluoride-based HAE ( Figure 8). The commercial anodization Anomag (Magnesium Technology) process, as well as PEO processes like Keronite (Keronite), and Magoxid-Coat (Aalberts Surface Technologies), use Cr-free dilute alkaline solutions. The treatment conditions and anticorrosion characteristics of popular commercial procedures, as well as selected recent related studies, are presented in Table 1. The commercial anodization Anomag (Magnesium Technology) process, as well as PEO processes like Keronite (Keronite), and Magoxid-Coat (Aalberts Surface Technologies), use Cr-free dilute alkaline solutions. The treatment conditions and anti-corrosion characteristics of popular commercial procedures, as well as selected recent related studies, are presented in Table 1. It can be seen that commercial processes are based on silicate, phosphate, and sometimes, fluoride compounds. The coating corrosion resistance properties are often evaluated in terms of their performance in neutral salt spray tests (NSST); unfortunately, different coating thicknesses make the comparison of different commercial processes difficult. However, it is evident that newer commercial treatments, which abandon the use of toxic compounds, show much greater corrosion resistance than older methods like the DOW17 or HAE. A cross-section view of a Keronite coating [86] is shown in Figure 9.
Materials 2022, 15, x FOR PEER REVIEW 13 of 58 evaluated in terms of their performance in neutral salt spray tests (NSST); unfortunately, different coating thicknesses make the comparison of different commercial processes difficult. However, it is evident that newer commercial treatments, which abandon the use of toxic compounds, show much greater corrosion resistance than older methods like the DOW17 or HAE. A cross-section view of a Keronite coating [86] is shown in Figure 9. The last two rows in Table 2 present selected research results. They demonstrate that the electrolytes containing compounds like vanadates [84] or borates [76] can provide better anti-corrosion properties compared with HAE and DOW17 treatments, which are the most commonly used commercial techniques. Nevertheless, it must be considered that both chemicals are hazardous, toxic, and suspected of damaging fertility, similarly to chromates [87,88].
It is worth noting that PEO processes typically work in a high voltage range and therefore are characterized by relatively high energy consumption. Expensive power supplies with high energy output and specially designed output waveforms are often required. These drawbacks limit the application of PEO technology for the production of large parts in the aircraft or automotive industries; this will be discussed further below.
The number of patents of anodizing techniques for Mg alloys has been steadily increasing over the last 20 years ( Table 2). As can be seen in the observations, many of these patents focus on the chemistry of the electrolyte and the effect of additives such as particles and organic/inorganic compounds. It is also common to find patents where twostep processes are used to confer additional functionality to the coating.  The last two rows in Table 2 present selected research results. They demonstrate that the electrolytes containing compounds like vanadates [84] or borates [76] can provide better anti-corrosion properties compared with HAE and DOW17 treatments, which are the most commonly used commercial techniques. Nevertheless, it must be considered that both chemicals are hazardous, toxic, and suspected of damaging fertility, similarly to chromates [87,88].
It is worth noting that PEO processes typically work in a high voltage range and therefore are characterized by relatively high energy consumption. Expensive power supplies with high energy output and specially designed output waveforms are often required. These drawbacks limit the application of PEO technology for the production of large parts in the aircraft or automotive industries; this will be discussed further below.
The number of patents of anodizing techniques for Mg alloys has been steadily increasing over the last 20 years ( Table 2). As can be seen in the observations, many of these patents focus on the chemistry of the electrolyte and the effect of additives such as particles and organic/inorganic compounds. It is also common to find patents where two-step processes are used to confer additional functionality to the coating.

Method
Anodization process in a phosphate alkaline electrolyte containing a sequestering agent to suppress plasma during anodization and a tertiary amine. Outcome -Controlled coating thickness and porosity by choosing various combinations of both current density and time (e.g., high current density for a short time produce a less porous layer).
-The addition of a small amount of a phosphonate such as "Dequest" 2066 or 2041 to the anodizing bath allows the anodizing process to proceed with both pulsed waveforms and also DC. [96] WO/2003/083181 Process and device for forming ceramic coatings on metals and alloys, and coatings produced by this process Method PEO (at high frequency pulses) supported by sonic acoustic vibrations for the use of stable hydrosols as electrolytes (for the introduction of fine-disperse particles). Outcome -Controlled micro-discharges during PEO process.

Method
Anodization process in a phosphate alkaline electrolyte containing a sequestering agent and a tertiary amine. Outcome Process to create oxide layers with sequestering agents in the form of ethylene diamine tetramethylene phosphonic acid and DEQUEST. [98] CN106119846 a Method for preparing corrosion resistant and abrasion-resistant coating on surface of magnesium alloy Method Two-step treatment process: micro-arc oxidation treatment followed by microwave plasma vapor deposition. Outcome Improved corrosion and wear properties.

Corrosion of Anodized Mg Alloys
Understanding the influence of the anodic coatings on the mechanisms of corrosion of Mg alloys is crucial for the identification and design of coating characteristics that ensure good protection. The stages leading to corrosion initiation in the presence of anodic coatings are discussed below.
A protective anodic layer separates the alloy from the environment; its main role is to delay the ingress of aggressive species to the substrate [1]. As a consequence, features like big cracks and pores in the coating that enable easy access of the electrolyte to the metal and quick initiation of the corrosion process are undesirable. This is the case for the top porous layer of the coating, which is non-protective. The thick, intermediate layer with discontinuous porosity can considerably delay the corrosion process because the solution cannot easily infiltrate the coating [54,115]. The compact pore-free barrier layer, adjacent to the substrate, is the most protective part of the coating. Defects and partial dissolution of the outer layers, especially when they are amorphous, eventually lead to the penetration of corrosive species down to the barrier layer. The stability of this layer is greatly influenced by the local pH. Low values lead to the fast dissolution of the barrier layer and the corrosion of the substrate, whereas high pH values promote the formation of Mg(OH) 2 , which is voluminous and causes a blocking effect that limits the corrosion rate. This alkalization is particularly facilitated in the limited volume of the pore band that is sometimes present between the barrier layer and the intermediate porous layer [116]. A schematic of the corrosion process is shown in Figure 10. Typical corrosion morphologies of PEO-coated Mg alloys include general undercoating corrosion [49,117], localized corrosion [118], and coating dissolution [116,119] (Figure 11). Typical corrosion morphologies of PEO-coated Mg alloys include general undercoating corrosion [49,117], localized corrosion [118], and coating dissolution [116,119] (Figure 11).
Another aspect to bear in mind is that the influence of the type of Mg alloy only becomes relevant when the PEO coating has failed [54]. Grain boundaries or impurities in the Mg alloy, despite being active corrosion points in the bare substrate, are not so active when covered by a dielectric passive film. Indeed, the probability that continuous pores will be located directly above the active point is small; therefore, the corrosion process is slowed [55]. However, recent research has shown that the second phase segregates in the substrate may also contribute to more extensive discharges and increase the probability of through-going pores there [120]. The impurity level determines how long the coating survives under the same test conditions; this is particularly true for flash-PEO of alloys with strong micro-galvanic coupling [121]. Typical corrosion morphologies of PEO-coated Mg alloys include general undercoating corrosion [49,117], localized corrosion [118], and coating dissolution [116,119] (Figure 11).  [49,117], (b) localized corrosion [118], and (c) coating dissolution [116]. Reprinted (a) from [49], (b) from [118], (c) from [116] with permission from Elsevier.
Considering the described mechanism, the corrosion of PEO-coated Mg alloys can be minimized by enhancing of the chemical stability of the coatings and reducing the pore number and size (especially open porosity), which can be realized in several ways during and/or after anodizing. Fine-tuning the electrical input characteristics, the in situ incorporation of corrosion inhibiting species into the coating from the electrolyte, and sealing the pores during post-treatment are examples of current, successful strategies. Further, PEO coatings exhibit very good paintability due to their high roughness; therefore, the application of polymer top-coats significantly improves the corrosion resistance.

Energy Input
Anodic treatments can be performed using either constant voltage or constant current (DC/AC) regimes or in more advanced unipolar and bipolar pulsed modes. The energy driven into the system feeds several processes: the electrochemical oxidation and Joule heat in the case of conventional anodizing and plasma chemical reactions, thermal oxidation, and electrolyte vaporization in the case of the PEO [122][123][124][125]. As a result, the way the energy is supplied and distributed has a significant impact on the final structure of the coating [126][127][128][129]. For instance, DC constant current conditions sustain intensive, longlasting, and less mobile micro-discharges which promote the formation of large channels and coating material destruction [130,131], greater pore size, and lower pore population density. Eventually, DC micro-discharges transform into microarcs, accompanied by more intensive gas evolution, the formation of very large size pores, and the thermal cracking of the coating [10].
Typically, anodic treatments of Mg and its alloys use current densities between 0.01-0.1 A cm −2 . The ranges of final voltages are quite varied. The lowest values can be found below 50 V [132], whereas the highest can be up to 500-600 V [66,133]. In most cases, applied voltages are within the 150-350 V range [60]. Compared to aluminum, the durations of anodic treatments of Mg-based alloys are relatively short. Typically, within 3-15 min, the coating thickness can reach about 4-40 µm, with a growth rate of 0.5-13 µm min −1 . These thicknesses are not maximal for Mg-based PEO coatings but are optimal for practical applications, especially from an economical and industrial point of view. Optimizing the PEO parameters makes it possible to achieve relatively low energy consumption, for instance, 0.632 kWh m −2 µm −1 [134]. It was recently shown that so-called flash-plasma electrolytic oxidation coatings (FPEO), with a process duration <90 s and corrosion performance better than that of a commercial Cr(VI)-based coating, consume a very low amount of energy (~1 kWh m −2 µm −1 ) [135,136]. In concentrated alkaline electrolytes (>150 g/L dissolved solids) under a DC regime, the energy consumption can be as low as~0.07 kWh m −2 µm −1 [137]; however, there is a lack of information on the corrosion behavior of such coatings.
Nowadays, unipolar or bipolar current pulse modes are frequently used in PEO of Mg alloys, as they can decrease the coating porosity and improve homogeneity, and thus, improve the corrosion resistance. Combinations of anodic and cathodic pulses and modifications of the frequency and duty cycle provide a wide range of energy input options that can help to avoid destructive, long-lasting micro-discharges, increase their population density [138], and obtain denser coatings with better corrosion resistance. Some examples of different energy input-related processing parameters and their effect on coating properties are summarized in Table 3.  10 Hz → best EIS performance Z 10 mHz-0.5 h 2.5 × 10 5 Ω cm 2 Z 10 mHz-50 h 1.1 × 10 5 Ω cm 2 [143] Unipolar pulse(S1):  Non-sealed 3 min coating-no corrosion in 80 h NSST, pitting at 100 h. [145] As can be seen in Table 3, there are many possibilities in terms of modifying the energy parameters, but the final effect on the anti-corrosion properties of the coatings is always a result of all system factors, such as voltage, current density, frequency, and duty cycle, and it is hard to determine the effect of an individual variable. For instance, the influence of the application of uni-or bipolar pulses modes on corrosion resistance is ambiguous, as shown in the contradictory results presented in [144][145][146]148]. Current frequencies tested in the range of 10-2000 Hz exhibit a tendency of corrosion resistance increase with higher frequency values [139,140,142,149]; however, in one case, the opposite trend was observed [143]. Based on the summarized results, it can be concluded that duty cycles should typically be below 40% for both unipolar [139,140,149] and bipolar treatments [141,146], and that higher values are not suitable for the improvement of the anti-corrosion properties of PEO coatings [139]. The coating growth rate in these conditions can be slower than under DC; nevertheless, in well-chosen conditions, it is comparatively fast, usually between 0.77-12.44 µm min −1 ( Table 3). The influence of system parameters on coating thickness can be arranged from the most to least relevant as follows: final voltage > current density > duty cycle > frequency. The significance of individual parameters on the anti-corrosion properties may be arranged in the following order: final voltage > frequency > duty cycle > current density [140]. Importantly, a literature analysis confirmed that high thickness is not always a guarantee of the best anti-corrosion properties of PEO coatings on Mg alloys.

Electrolyte Composition
The type and concentration of the electrolyte play a crucial role in the morphology, composition, and ultimately, in the corrosion resistance [8,60]. The first and most basic distinction is the pH of the solution. Acidic electrolytes are not preferred for PEO treatment because they completely suppress micro-discharges below pH 3 [150], enhancing the oxide dissolution rate and leading to the formation of porous-type structures with poor corrosion resistance [151,152]. Electrolytes with higher pH values promote the earlier onset of microdischarges, and the oxide film develops typical PEO three-layered structures.
To ensure an alkaline environment, KOH or NaOH are often chosen as primary components. This also prevents excessive anodic dissolution. With regard to the specific role of K + or Na + ions, it has been shown that the former (as a hydroxide, phosphate, or both) promotes more compact coatings, whereas the latter reduces the pore size in the outer part of the coating [153]. The effects of KOH or NaOH concentrations on coating properties have been intensively investigated. For instance, a three-fold increase of KOH concentration (up to 0.27 M) increased the pH (by~1 unit) and conductivity (by more than 2.5 times), decreased the breakdown and final voltage values, and increased the compact layer thickness and the MgO/Mg(OH) 2 ratio, resulting in higher corrosion resistance of coating in salt spray [154]. Similar results were found for a fluoride-and phosphate-based electrolyte containing 1.5 M and 3 M KOH, i.e., better corrosion performance was observed with a more concentrated solution due to the reduced coating porosity [155].
Comparing the results for 5 g/L and 100 g/L NaOH-based electrolytes, the breakdown voltages were 282 V and 82 V, respectively. Furthermore, the coating obtained from the concentrated electrolyte exhibited higher corrosion resistance (two orders of magnitude lower corrosion current and one order of magnitude higher total impedance, |Z| 10mHz ) due to the increased coating thickness [156]. It is evident that due to the effect of decreasing breakdown voltage, the concentration of the electrolytes is a significant factor in minimizing the energy consumption. Sustainable recycling and circular economy routes need to be developed for such electrolytes in order to avoid the increase of the environmental footprint when they need to be disposed of.
Other additives, like phosphates, silicates, fluorides, aluminates, zirconates, permanganates, etc., are used to modify the coating composition and properties. Under the electric field, anions like PO 4 3− , SiO 3 2− and F − migrate inward and get incorporated into the coating [157]. The following reactions (3)- (7) are examples describing the different phase formation in the presence of some of these additives: The X-ray diffraction patterns of PEO-treated Mg alloys confirmed the presence of the aforementioned compounds [143,158,159]. These phases can have a great effect on the physico-chemical properties of the coatings; for instance, the following order of corrosion resistance to Cl − attack was identified: amorphous material < MgO < Mg 3 (PO 4 ) 2 < Mg 2 SiO 4 [116].
The beneficial effect of fluoride-based electrolytes on the corrosion resistance of PEO coatings on Mg alloys is often related to the formation of passivating MgF 2 phase [160,161], associated with more compact layers, nobler corrosion potential, and reduced susceptibility to pitting [133,160,162,163]. It was also reported that the addition of fluoride to various electrolytes (based on NaAlO 2 , Na 2 SiO 3 , NaAlO 2 with Na 3 PO 4 , and Na 2 SiO 3 with Na 3 PO 4 ) always caused much quicker growth of the PEO layer, which led to a significant reduction of specific energy consumption [135]. On the other hand, it was shown that similar corrosion resistance can be achieved in more benign, fluoride-free electrolytes, such as aluminatephosphate [164].
The influence of phosphates, common additives in commercial electrolytes (Anomag CR1 and CR2, Henkel Japan Co., Ltd.), on corrosion properties has been shown [49] and seemed to depend on the various types of phosphates (orthophosphate, pyrophosphate, polyphosphate, hydrophosphates, etc.) and the treatment conditions [165][166][167]. Phosphorus-containing coatings typically favor the formation of amorphous phases, although Mg 3 (PO 4 ) 2 can also be detected [165]. It has been suggested that amorphous phosphate participates in a self-repairing mechanism consisting of its dissolution and subsequent re-deposition as an insoluble crystalline hydrated magnesium phosphate at the locations where the PEO coating fails [168].
The presence of sodium orthophosphate in the electrolyte has been shown to promote the coating growth rate through a more uniform distribution of micro-discharges, avoiding local coating destruction [165]. However, the phosphate concentration must be carefully chosen. According to [164], too low concentrations (<0.21 M) result in non-uniform coatings, whereas concentrations higher than 0.25 M lead to recurrent micro-discharges at specific locations on the surface.
The addition of aluminates, like NaAlO 2 , to the electrolyte has been shown to decrease the number and size of discharge channels [58] and increase the coating thickness [159]. More importantly, the formation of MgAl 2 O 4 results in enhanced corrosion resistance, which is associated with its chemical stability and a higher Pilling-Bedworth ratio compared with MgO [48,169,170]. It has been proven that for a long immersion time (9 days), the corrosion resistance of the aluminate-containing coating was higher than that of silicate-and phosphate-based electrolytes [171]. Less frequently used aluminum nitrate, Al(NO 3 ) 3 , also helps to form a thick and uniform passive layer when used in the 0.15-0.20 M range [164].
Perhaps the most common component of PEO electrolytes is sodium silicate, Na 2 SiO 3 . Compared with equivalent amounts of phosphates and aluminates, it has a greater influence on the electrolyte conductivity, final voltage, and therefore, coating thickness. A corrosion study comparing the effect of phosphate, aluminate, molybdate, and silicate sodium salts revealed that the latter provided the biggest improvement. This was probably due to a higher coating thickness and less open porosity [159].
Frequently, mixtures of the aforementioned chemicals work much better than simple composition electrolytes. The study of the FPEO (>90 s treatment,~5 µm coatings thickness) obtained from varied electrolytes based on mixtures of KOH with one, two, or three components from NaAlO 2 , Na 2 SiO 3 , Na 3 PO 4 , KF showed the highest performance with the most complex (four elements) compositions (either aluminate or silicate mixtures with potassium hydroxide, phosphate, and fluoride). The best coating achieved superb corrosion protection, significantly outperforming commercial Cr(VI)-based reference samples [135,136].
Sodium borate, Na 2 B 4 O 7 , is usually used in combination with silicate, phosphate, or aluminate salts. Borate ions promote coating growth, since the decomposing B 4 O 7 2− is an additional source of oxygen. The presence of sodium tetraborate in a solution of Na 2 SiO 3 , KOH, NaH 2 PO 4 , and, optionally, KF, improved the coating resistivity by ten fold, although it tended to cause a "coral reef" surface morphology [172]. A Mott-Schottky examination of the electronic properties of the borate-and fluoride-based coatings indicated their lower donor concentration, much more negative flat band potential, and, therefore, higher corrosion resistance. Some other works have also shown the beneficial effect of Na 2 B 4 O 7 on corrosion resistance and even suggested that it assists in a self-repairing mechanism [158,173].

Organic and Inorganic Soluble Additives
The compounds discussed above are the most frequently used constituents of electrolytes in PEO of Mg alloys. However, a lot of attention is currently being paid to the further improvement of the coating properties, which can be achieved by using various types of organic and inorganic additives in the basic electrolyte mixtures ( Figure 12).
It should be noted that while the introduction of some soluble inorganic compounds strongly modifies the chemical composition of the coating, it also changes the pH and conductivity of the electrolyte, which influence the breakdown and final voltages and discharge density, resulting in structural modifications to the coatings [173][174][175]. The main criterion behind the choice of an additive salt is its capacity to form highly stable compounds that improve the passivity of the coating [84,132]. For example, elements like Mo (VI) [176] or rare-earth Ce(III) and La(III) [174] have the potential to provide a self-healing capacity of the coating under corrosion.
Titanium [177][178][179] and zirconium compounds [150,[180][181][182][183][184] are being actively researched as additives due to their very high stability and pore-sealing effect [177,178,184]. While titanium is typically used in a form of K 2 TiF 6 , various types of zirconium salts have been tested. The best results (denser internal film and a more uniform surface) were yielded by K 2 ZrF 6 compared with ZrOCl 2 and Zr(NO 3 ) 4 [150]. The inconvenience of K 2 ZrF 6 is the hydrolysis of Zr 4+ ions at pH > 4.0, which leads to the precipitation of Zr(OH) 4 in the electrolyte and, consequently, non-uniform discharge distribution and low coating quality. This limits the choice of basic components of the electrolytes to those that support relatively low pH, such as NaH 2 PO 4 [182,183], which, in turn, limits the range of coating phase compositions. Some authors have circumvented this by dual-or two-step anodizing (Figure 13), where PEO of magnesium alloy in the acidic electrolytes [180] is preceded by a short anodic pre-treatment in alkaline electrolyte; this prevents the excessive anodic dissolution of the magnesium alloy and ensures good corrosion protection [180,181].
Organic additives usually do not change the coating composition but can provide some unique effects compared with inorganic compounds. Firstly, they can either decrease or increase the breakdown potential [185,186], with the latter imposing an additional economic burden. Further, they adsorb on the electrode and significantly reduce the anode/electrolyte interface tension [66,187], modulating gas evolution and the discharge population density [185,187]. All these effects can affect the mechanism of the discharge and change the structure of PEO coatings, reduce the surface porosity, inhibit defect development and improve the corrosion resistance [188]. The detailed effects of organic and inorganic additives are presented in Table 4. [158,173].

Organic and Inorganic Soluble Additives
The compounds discussed above are the most frequently used constituents of electrolytes in PEO of Mg alloys. However, a lot of attention is currently being paid to the further improvement of the coating properties, which can be achieved by using various types of organic and inorganic additives in the basic electrolyte mixtures ( Figure 12). It should be noted that while the introduction of some soluble inorganic compounds strongly modifies the chemical composition of the coating, it also changes the pH and conductivity of the electrolyte, which influence the breakdown and final voltages and discharge density, resulting in structural modifications to the coatings [173][174][175]. The main criterion behind the choice of an additive salt is its capacity to form highly stable compounds that improve the passivity of the coating [84,132]. For example, elements like Mo (VI) [176] or rare-earth Ce(III) and La(III) [174] have the potential to provide a selfhealing capacity of the coating under corrosion.
Titanium [177][178][179] and zirconium compounds [150,[180][181][182][183][184] are being actively researched as additives due to their very high stability and pore-sealing effect [177,178,184]. While titanium is typically used in a form of K2TiF6, various types of The analysis in Table 4 indicates that additives in a mixture of 2-3 basic components (e.g., silicates, fluorides, phosphates) can provide significantly improved corrosion protection; however, the final result is not a direct sum of the effects of the individual constituents. Initial studies of the effect of electrolyte additives carried out by McNeil in 1957 demonstrated a positive influence of tungstate, vanadate, and stannate on corrosion resistance (although it is worth noting that the substrates were pickled in dichromate solution before anodic treatment and, optionally, after treatment). The best results before and after the post-treatment were obtained in the presence of vanadate.
In most of the cases presented in Table 4, the additives led to enhanced corrosion protection compared to additive-free PEO coatings (higher corrosion potential, lower corrosion current, lower tribocorrosion rate) [84,[174][175][176][184][185][186][187]191,192,196,197,201]; however, negative results were also reported [132,189]. Often, the additives demonstrated positive effects within a certain concentration limit, beyond which the properties deteriorated [126,176,179,185,192,201,202]. It is also evident that additives in the electrolyte often reduce the coating thickness, without reducing the coating protective properties, which is associated with improved morphological parameters of the coatings, such as compactness, roughness, or pores size [174,175,192,196,203,204].
The reduction of corrosion current by two orders of magnitude compared with the untreated substrate (e.g., down to~10 −7 A/cm 2 ) is frequently achieved in the presence of an additive. One of the most notable improvements was obtained by the addition of KMnO 4 (10 −9 A/cm 2 ) ( Figure 14) [175] or EDTA (10 −8 A/cm 2 ) into the electrolyte [196].
uniform surface) were yielded by K2ZrF6 compared with ZrOCl2 and Zr(NO3)4 [150]. The inconvenience of K2ZrF6 is the hydrolysis of Zr 4+ ions at pH > 4.0, which leads to the precipitation of Zr(OH)4 in the electrolyte and, consequently, non-uniform discharge distribution and low coating quality. This limits the choice of basic components of the electrolytes to those that support relatively low pH, such as NaH2PO4 [182,183], which, in turn, limits the range of coating phase compositions. Some authors have circumvented this by dual-or two-step anodizing (Figure 13), where PEO of magnesium alloy in the acidic electrolytes [180] is preceded by a short anodic pre-treatment in alkaline electrolyte; this prevents the excessive anodic dissolution of the magnesium alloy and ensures good corrosion protection [180,181].  (d-i) X-ray elemental mapping of the MAO coating deposited on Mg in two stages, sequentially using the alkaline silicate electrolyte in the first stage, followed by the acidic zirconate electrolyte in the second stage, both at 300 V and for 3 min each stage [180]. Reprinted from [180] with permission from Elsevier. PDP in 3.5 wt.% NaCl (i corr ) 0 g/L: 7.46 × 10 −9 A/cm 2 0.4 g/L: 6.82 × 10 −9 A/cm 2 0.6 g/L: 7.09 × 10 −9 A/cm 2 0.8 g/L: 7.17 × 10 −9 A/cm 2 [190] Ce(NO 3 ) 3 0.    A/cm 2 NaOH, Na 2 SiO 3 : −1.14 V; 6.125 × 10 −7 A/cm 2 NaOH, 4 g/L: −1.513 V; 1.226 × 10 −5 A/cm 2 NaOH,4 g/L, Na 2 SiO 3 : −1.349 V; 1.385 × 10 −7 A/cm 2 [197] benzotriazole (BTA) 3, 5, 10 g/L AZ31 KOH, Na 2 SiO 3 , Na 2 B 4 O 7 , Na 2 CO 3 , Na 2 SiO 3 DC 1.5 A dm −2 23 [198] ethylene-glycol 10 PDP in 3.5 wt.% NaCl (i corr ) PENC + PEO: 1.92 × 10 −8 A/cm 2 , two and one orders of magnitude greater than for PENC and PEO. EIS in 3.5 wt.% NaCl (|Z| 10mHz ), 30 min/72 h PENC + PEO: 10 7 Ω cm 2 /10 6 Ω cm 2 PEO: 3 × 10 6 Ω cm 2 /3 × 10 3 Ω cm 2 PENC: 1.7 × 10 4 Ω cm 2 /1.7 × 10 2 Ω cm 2 [200] Recently, a number of works have applied orthogonal experiment designs in order to evaluate the interaction of several electrolyte components, such as K 2 ZrF 6 and EDTA-Na, with the electrical parameters. Both of these additives are being increasingly used in new PEO coating developments [205][206][207][208]; K 2 ZrF 6 as a source of Zr, with the aim of forming ZrO 2 in the coating from particle-free electrolytes, and EDTA-Na as a complexing agent to stabilize cations of interest in the electrolyte [193]. K 2 ZrF 6 is problematic on its own in terms of the surface passivation and coating growth that occur due to its acidity; however, it can successfully form self-sealing coatings when combined with other additives, such as phytic acid and ammonium fluoride [206]. an additive. One of the most notable improvements was obtained by the addition of KMnO4 (10 −9 A/cm 2 ) ( Figure 14) [175] or EDTA (10 −8 A/cm 2 ) into the electrolyte [196]. Recently, a number of works have applied orthogonal experiment designs in order to evaluate the interaction of several electrolyte components, such as K2ZrF6 and EDTA-Na, with the electrical parameters. Both of these additives are being increasingly used in new PEO coating developments [205][206][207][208]; K2ZrF6 as a source of Zr, with the aim of forming ZrO2 in the coating from particle-free electrolytes, and EDTA-Na as a complexing Organic additives in the electrolyte typically decompose during PEO micro-discharge, leading to excessive gas generation and, as a result, a looser coating microstructure with compromised corrosion resistance. However, additives such as silanes, e.g., 3aminopropyltrimethoxysilane (APTMS), have been shown to increase the coating thickness and corrosion resistance due to the densification of the coating and the reduction of pore size associated with the formation of Mg-O-Si chemical bonds [199].
The introduction of N and C into the coating via plasma electrolytic nitrocarburizing (PENC) from formamide/NaOH aqueous electrolytes, where Mg acts as a cathode, followed by PEO in an alkaline silicate electrolyte is a new strategy that offers sustainable, high corrosion resistance of a Mg alloy due to the formation of a thick and extremely compact coating [200].

Particle Addition
New PEO coatings with added functionality are being investigated through the in situ incorporation of insoluble particles in the electrolyte (Figure 15). This approach modifies both the electrochemistry of the PEO process (e.g., conductivity, pH, breakdown voltage, etc.) and the coating characteristics (e.g., porosity, thickness, phase composition, compactness of the layer, etc.). The added particles bring new functionalities to the coating and can be introduced into the coating in an inert manner, without reaction or the formation of new phases, or by reactive or partially reactive incorporation, when a reaction between the particles and the coating matrix takes place [209]. agent to stabilize cations of interest in the electrolyte [193]. K2ZrF6 is problematic on its own in terms of the surface passivation and coating growth that occur due to its acidity; however, it can successfully form self-sealing coatings when combined with other additives, such as phytic acid and ammonium fluoride [206].
Organic additives in the electrolyte typically decompose during PEO microdischarge, leading to excessive gas generation and, as a result, a looser coating microstructure with compromised corrosion resistance. However, additives such as silanes, e.g., 3-aminopropyltrimethoxysilane (APTMS), have been shown to increase the coating thickness and corrosion resistance due to the densification of the coating and the reduction of pore size associated with the formation of Mg-O-Si chemical bonds [199].
The introduction of N and C into the coating via plasma electrolytic nitrocarburizing (PENC) from formamide/NaOH aqueous electrolytes, where Mg acts as a cathode, followed by PEO in an alkaline silicate electrolyte is a new strategy that offers sustainable, high corrosion resistance of a Mg alloy due to the formation of a thick and extremely compact coating [200].

Particle Addition
New PEO coatings with added functionality are being investigated through the in situ incorporation of insoluble particles in the electrolyte (Figure 15). This approach modifies both the electrochemistry of the PEO process (e.g., conductivity, pH, breakdown voltage, etc.) and the coating characteristics (e.g., porosity, thickness, phase composition, compactness of the layer, etc.). The added particles bring new functionalities to the coating and can be introduced into the coating in an inert manner, without reaction or the formation of new phases, or by reactive or partially reactive incorporation, when a reaction between the particles and the coating matrix takes place [209]. Reactive incorporation includes one of the following: melting, phase transformation, and reaction with electrolyte constituents, substrate, or other phases in the coating. Crucially, the extent of these processes is dependent on many factors, such as the size, melting point, concentration, and the zeta potential of the added particles, the composition of the electrolyte, type of substrate, and energy input from the micro- Reactive incorporation includes one of the following: melting, phase transformation, and reaction with electrolyte constituents, substrate, or other phases in the coating. Crucially, the extent of these processes is dependent on many factors, such as the size, melting point, concentration, and the zeta potential of the added particles, the composition of the electrolyte, type of substrate, and energy input from the micro-discharges. Examples of particles that reactively incorporate into PEO coatings on Mg are Al 2 O 3 , ZrO 2 , SiO 2 , and clay particles [209].
The size of the suspended particles should be small enough, i.e., less than 500 nm, to prevent their sedimentation and to enable their easy transport into discharge channels. The zeta potential is an important parameter that characterizes the degree of electrostatic repulsion between the particles in the electrolyte; the more negative its value, the more uniform the dispersion of particles and the more homogeneous their distribution in the coating. Typically, in alkaline solutions, particles are negatively charged, which facilitates their electric field-assisted incorporation into the coating matrix. More information on the stabilization of particles in the electrolytes using organic additives (e.g., urea, sodium dodecyl sulfate) and the mechanisms of particle incorporation into PEO coatings on Mg alloys can be found in [210,211].
Recently, a new range of particles came into focus. For instance, graphene oxide (GO) was successfully introduced into a PEO coating [242][243][244], reducing the number of micropores and improving the corrosion resistance due to the increased tortuosity of the electrolyte species diffusion pathway. Similarly, the addition of graphite [237,245,246], multi-walled carbon nanotubes (CNT) [247,248], or carbon spheres (CS) [249] into the electrolyte produced a coating densifying effect, thereby increasing the corrosion resistance while the CNT oxidized during PEO (new). However, the primary benefits of the incorporation of carbon-based particles are enhanced hardness, wear resistance, and heat dissipation [247][248][249]. Another innovative idea is the in situ incorporation of nanocontainers loaded with corrosion inhibitors (e.g., halloysite or aluminosilicate nanotubes loaded with benzotriazole, molybdate, or vanadate salts or 8-hydroxyquinoline) [250][251][252]. The inhibitor is released when the pH changes during the corrosion process, which confers self-healing properties upon the PEO coating. The effects of particle addition on the coating properties and corrosion resistance are summarized in Table 5.  [212] 5 g/L n-SiO 2~1 2 nm µ-SiO 2 1-5 µm        Based on the results in Table 5, it can be concluded that the incorporation of in situ particles is a very promising way to improve the anti-corrosion properties of PEO coatings. In almost all studies, the corrosion current density decreased compared with the particle-free PEO coatings fabricated under the same conditions. The lowest values were obtained for treatments with the addition of zirconia sol (1.4 × 10 −8 A/cm 2 ) [212], clay (2.3 × 10 −8 A/cm 2 ) [116], and alumina sol (2.6 × 10 −8 A/cm 2 ) [228]. Comparing these values with those shown in Table 4, it is evident that the addition of particles can be more effective than the use of soluble electrolyte additives.
However, the testing period appears to be important. Recently, it was reported that some particle-incorporated coatings provide only a short-term increase in corrosion resistance; longer testing times often reveal the opposite effect, with faster relative degradation [209,234,253]. Additionally, as observed with other electrolyte constituents, particles are only effective up to a certain concentration, usually below 20 g/L, beyond which their positive effect on corrosion resistance declines [239,242]. Their effect on coating thickness is variable and not correlated with corrosion resistance. This notwithstanding, particle-modified coatings are good for additional functionalities, such as, for instance, wear resistance [254].

Organic Electrolytes
Anodizing Mg alloys in non-aqueous electrolytes is an alternative route that has been actively explored in the last ten years [255][256][257][258][259][260][261]. Compared to aqueous anodizing, it avoids water decomposition, decreases Mg dissolution during anodizing, and opens a window for the development of defect-free anodic films. Ethylene glycol, triethanolamine, and glycerol are often used as base electrolytes which may also contain fluoride anions and a small percentage of water. Depending on the source of fluoride ions (HF, NH 4 F, etc.) and the presence of water, barrier type [261] or self-organized nanoporous/nanotubular [260] films can be formed. For instance, ordered oxy-fluoride nanostructures (nanopores and nanotubes) with 70-100 nm pore diameter were formed in a WE43 Mg alloy by anodizing in ethyleneglycol and 0.2 M HF in the potential window between 70 and 120 V [260]. In another study, the effect of water content was explored regarding the transformation from a porous to a compact film morphology. In a solution of ethylene glycol and trimethylamine with 1% water, a porous structure (100 nm pore diameter) was formed, whereas a water content between 10% and 40% led to the formation of a barrier-type film. The latter showed better anticorrosion properties than the nanoporous films in a salt-spray test [259]. FTIR and GDOES results showed the incorporation of C, N, and O in the films and that the amount of incorporated species increased with decreasing water content.
It is worth mentioning that anodizing in organic electrolytes in the presence of fluoride ions helps to increase the efficiency of the process and increases the Pilling-Bedworth ratio up to~1.7 [261] due to F-enrichment (~0.2 O/F ratio [258]), which is expected to be beneficial for corrosion protection. Efficiencies close to 100% are commonly reported for anodizing using these types of electrolytes (the lower the water content, the higher the current efficiency) [255].
Anodizing voltages can be as high as 440 V without dielectric breakdown [257], and the thickness of the coatings ranges from several hundreds of nanometers, in the case of barrier-type films (Figure 16), to several tens of micrometers in nanoporous/nanotubular films (Figure 16b) [256]. nanoporous/nanotubular [260] films can be formed. For instance, ordered oxy-fluoride nanostructures (nanopores and nanotubes) with 70-100 nm pore diameter were formed in a WE43 Mg alloy by anodizing in ethyleneglycol and 0.2 M HF in the potential window between 70 and 120 V [260]. In another study, the effect of water content was explored regarding the transformation from a porous to a compact film morphology. In a solution of ethylene glycol and trimethylamine with 1% water, a porous structure (100 nm pore diameter) was formed, whereas a water content between 10% and 40% led to the formation of a barrier-type film. The latter showed better anticorrosion properties than the nanoporous films in a salt-spray test [259]. FTIR and GDOES results showed the incorporation of C, N, and O in the films and that the amount of incorporated species increased with decreasing water content. It is worth mentioning that anodizing in organic electrolytes in the presence of fluoride ions helps to increase the efficiency of the process and increases the Pilling-Bedworth ratio up to ~1.7 [261] due to F-enrichment (~0.2 O/F ratio [258]), which is expected to be beneficial for corrosion protection. Efficiencies close to 100% are commonly reported for anodizing using these types of electrolytes (the lower the water content, the higher the current efficiency) [255].
Anodizing voltages can be as high as 440 V without dielectric breakdown [257], and the thickness of the coatings ranges from several hundreds of nanometers, in the case of barrier-type films (Figure 16), to several tens of micrometers in nanoporous/nanotubular films (Figure 16b) [256].  [261]; (b) nanotubular type [260]. Reprinted (a) from [261] and (b) from [260] with permission from Elsevier.
Qi et al. recently obtained a~15 µm-thick PEO coating on AZ31B alloy in ethylene glycol-based electrolyte with 8 wt.% NH 4 F under pulsed bipolar conditions ( Figure 17). The coating comprised pure MgF 2 and had a gradient porosity reducing inward from 24.93% to 2.87%. The coating reduced the corrosion current density of the alloy in neutral and acidic (pH3.0) 3.5%NaCl by approximately two and one orders of magnitude, to~5 × 10 −7 A/cm 2 and to~3 × 10 −6 A/cm 2 , respectively [262].
So far, studies on anodizing in organic electrolytes have been carried out using a range of commercial magnesium alloys (e.g., ZE41, WE43, AZ91D, Mg-Zn-RE, etc.). It has been observed that the second phases can be anodized and incorporated into the anodic films, locally affecting the film thickness and composition. However, there is a severe lack of applied and systematic studies of the film properties, energy efficiency, and corrosion behavior. So far, studies on anodizing in organic electrolytes have been carried out using a range of commercial magnesium alloys (e.g., ZE41, WE43, AZ91D, Mg-Zn-RE, etc.). It has been observed that the second phases can be anodized and incorporated into the anodic films, locally affecting the film thickness and composition. However, there is a severe lack of applied and systematic studies of the film properties, energy efficiency, and corrosion behavior.
Another successful approach is the incorporation of corrosion inhibitors into the PEO pores by sample immersion in the inhibitor solution under low-pressure conditions. It has been proven that PEO pores, as a high capacity reservoir for corrosion inhibitors, can provide long-term, active protection for Mg substrates [136]. For a deeper analysis of the corrosion inhibition mechanism using organic inhibitors added as a post-treatment, the reader is referred to Part III of this review [3]. The application of stable compounds such as SiO 2 [86], TiO 2 [285], and SiO 2 -ZrO 2 [286] via sol-gel has also proved to be successful. The idea is to form a top barrier layer by covering the PEO surface and filling the micropores. Sol-gel coatings are not incorporated into the inner parts of the coating and rarely interact with the PEO coating material in an active way. Morphologically, sol-gel coatings are more uniform than conversion coatings produced by simple immersion, although cracking is likely to occur on their surface when thick multi-layers are produced. A typical practice in sol-gel treatments is to repeat the immersion/heat-treatment/drying sequence several times in order to obtain a coating that is thick enough to sufficiently cover the pores of the PEO layer [86,[285][286][287].
Another novel approach consists of hydrothermal post-treatment, resulting in the formation of layered double hydroxides (LDH) [288]. LDHs, with a chemical formula [M 2+ 1−x M 3+ x (OH) 2 ][A n− ] x/n ·zH 2 O, are composed of positively charged brucite-like layers, containing both Mg 2+ and Al 3+ , separated by regions containing anions and solvation molecules. LDH flakes can easily seal the pores of PEO coatings, but their greatest advantage is the possibility of loading their interlayer spaces with inhibitors which are released when triggered by chemical changes in the environment. This was first demonstrated on PEO-treated aluminum alloy, bringing about a remarkable increase in the corrosion resistance achieved through vanadate-intercalated LDH post-treatment [289].
The first directly grown LDH on PEO of Mg alloys without autoclave was achieved by Petrova et al. [290]. Currently, LDH post-treatment is being successfully exploited for the fabrication of composite PEO/inhibitor-loaded LDH coatings on Mg alloys in an increasing number of works [291][292][293][294][295][296][297]. Chen et al. investigated PEO/LDH systems with a variety of cations in the LDH layer, such as Ni-LDH, Zn-LDH, Al-LDH, and MgFe-LDH [294,295], of which Ni-LDH was found to form a continuous LDH layer over the PEO surface and produce an order of magnitude increase in total impedance (up to 5 × 10 5 Ωcm 2 ) compared with a stand-alone PEO coating. It was also shown that PEO coatings modified with composite Zn-Al LDH with reduced graphene oxide (rGO) nanosheets could effectively enhance corrosion resistance and reduce i Corr to 5 nA/cm 2 [298]. Alternatively, in contrast to the direct synthesis method, LDH nanocontainers can be in situ incorporated into PEO coatings [299] or can be formed as a secondary reaction product of the primary deposited MnOOH film, spontaneously reacting with the corrosion-produced Mg 2+ and OH − [300]. Apart from the sealing effect of the LDH layer that improves the barrier properties of the PEO coating, the abilities to exchange the anionic inhibitors between its hydroxide layers and release them "on demand" offer active corrosion protection. More details on the active inhibition properties of LDH are provided in Part III of this review [3].
Superhydrophobic composite coatings have been fabricated on magnesium alloys by combining micro-arc oxidation (MAO) with post-treatments such as cyclic assembly in phytic acid and Ce(NO 3 ) 3 solution [301], stearic acid ethanol solution [302], alkynol inhibitor loading followed by hydrophobic wax application [303], hydrothermal treatment in boiling water followed by self-assembled monolayer formation in perfluorodecyltrichlorosilane (FDTS) [304], LDH treatment followed by perfluorodecyltriethoxysilane (PFDS) and perfluoropolyether (PFPE) lubricant infusion to form so-called slippery liquid-infused porous surface (SLIPS) [305], with some of these approaches maintaining remarkably high corrosion resistance for up to 21 days ( Table 6). The application of fluorocarbons, such as polytetrafluoroethylene and fluoroparaffines [306][307][308][309], produced changes in impedance modulus and corrosion current density of up to two orders of magnitude (increase and decrease, respectively) and contact angles of 130 • -152 • .
The application of polymeric compounds, which infiltrate the coating pores, is a common practice for anodized Mg components [310]. Research on this topic has gone in several directions, including silanes [274,311] and E-coatings [312][313][314]. For instance, silane treatment KH550 has been shown to enhance the corrosion resistance of a magnesium alloy with a silicate-based PEO coating [315]. E-coatings are widely applied in the automotive industry; they consist of the deposition of epoxy, epoxy polyester, or polyester resins by electrostatic powder spraying. For instance, AM60B alloy anodized at 70 V for 15 min in a strongly alkaline bath (KOH, Na 3 PO 4 , KF, Al(NO 3 ) 3 ), subjected to electrostatic powder spraying and curing (190 • C, 20 min), displayed no evidence of corrosion after salt spray testing for up to 595 h for a polyester paint [316]. Wierzbicka et al. reported that varied FPEO coatings covered with a three-component epoxy primer after a week of NSST were rated at up to 7 or 8 (out of a maximum of 10, according to the ASTM D 1654 standard) ( Figure 18). These results are comparable with the rating of Cr(VI)-based commercial conversion coatings used in the aircraft industry [135]. The same authors demonstrated that FPEO coating loaded with 4-MSA inhibitor and coated with epoxy-based chromate-free primer with an active inhibition system based on lithium leaching technology outperformed a commercial Cr(VI)-based conversion coating with chromate-based primer after 1000 h of exposure in NSST ( Figure 19) [136].
E-coating baths have also been shown to be suitable for the sealing of PEO coatings by a simple immersion process in an E-coating bath consisting of an epoxy resin and titanium dioxide [317]. changes in impedance modulus and corrosion current density of up to two orders of magnitude (increase and decrease, respectively) and contact angles of 130°-152°.
The application of polymeric compounds, which infiltrate the coating pores, is a common practice for anodized Mg components [310]. Research on this topic has gone in several directions, including silanes [274,311] and E-coatings [312][313][314]. For instance, silane treatment KH550 has been shown to enhance the corrosion resistance of a magnesium alloy with a silicate-based PEO coating [315]. E-coatings are widely applied in the automotive industry; they consist of the deposition of epoxy, epoxy polyester, or polyester resins by electrostatic powder spraying. For instance, AM60B alloy anodized at 70 V for 15 min in a strongly alkaline bath (KOH, Na3PO4, KF, Al(NO3)3), subjected to electrostatic powder spraying and curing (190 °C, 20 min), displayed no evidence of corrosion after salt spray testing for up to 595 h for a polyester paint [316]. Wierzbicka et al. reported that varied FPEO coatings covered with a three-component epoxy primer after a week of NSST were rated at up to 7 or 8 (out of a maximum of 10, according to the ASTM D 1654 standard) (Figure 18). These results are comparable with the rating of Cr(VI)-based commercial conversion coatings used in the aircraft industry [135]. The same authors demonstrated that FPEO coating loaded with 4-MSA inhibitor and coated with epoxy-based chromatefree primer with an active inhibition system based on lithium leaching technology outperformed a commercial Cr(VI)-based conversion coating with chromate-based primer after 1000 h of exposure in NSST ( Figure 19) [136].
E-coating baths have also been shown to be suitable for the sealing of PEO coatings by a simple immersion process in an E-coating bath consisting of an epoxy resin and titanium dioxide [317].  In [63], the mechanical and corrosion performance of a polymer-coated AZ31 magnesium alloy pre-treated by plasma electrolytic oxidation (PEO) was compared with that of a polymer coated fluorotitanate-zirconate conversion coating (Gardobond X4707) on the same alloy. The electrostatically sprayed polymer was composed of polyester resin (50 wt.%), triglycidyl isocyanurate, polypropylene wax (≤1 wt.%), inorganic pigments, and fillers (barium carbonate and barium sulfate). Tests showed that both coatings (PEO and conversion coating) passed the adhesion test (rating 0); however, the results of impact and impact + adhesion tests revealed better performance of the coating with Ti/Zr treatment ( Figure 20). Surface appearance of (a) Ti/Zr + polymer and (b) PEO + polymer coatings after impact + adhesion tests [63]. Reprinted from [63] with permission from Elsevier.
Atmospheric corrosion tests, as per ASTM B117, and cyclic exposure to salt fog, as per VDA 621-415, revealed no damage either in the PEO or Ti/Zr treatments (Figure 21c,d).
Greater differences were observed on scribed (ASTM D1654, procedure A) specimens, with the PEO + polymer coating showing a higher rating number, i.e., superior performance, than the Ti/Zr + polymer coating (Figure 21e,f). The superior anticorrosion In [63], the mechanical and corrosion performance of a polymer-coated AZ31 magnesium alloy pre-treated by plasma electrolytic oxidation (PEO) was compared with that of a polymer coated fluorotitanate-zirconate conversion coating (Gardobond X4707) on the same alloy. The electrostatically sprayed polymer was composed of polyester resin (50 wt.%), triglycidyl isocyanurate, polypropylene wax (≤1 wt.%), inorganic pigments, and fillers (barium carbonate and barium sulfate). Tests showed that both coatings (PEO and conversion coating) passed the adhesion test (rating 0); however, the results of impact and impact + adhesion tests revealed better performance of the coating with Ti/Zr treatment ( Figure 20).  In [63], the mechanical and corrosion performance of a polymer-coated AZ31 magnesium alloy pre-treated by plasma electrolytic oxidation (PEO) was compared with that of a polymer coated fluorotitanate-zirconate conversion coating (Gardobond X4707) on the same alloy. The electrostatically sprayed polymer was composed of polyester resin (50 wt.%), triglycidyl isocyanurate, polypropylene wax (≤1 wt.%), inorganic pigments, and fillers (barium carbonate and barium sulfate). Tests showed that both coatings (PEO and conversion coating) passed the adhesion test (rating 0); however, the results of impact and impact + adhesion tests revealed better performance of the coating with Ti/Zr treatment ( Figure 20). Surface appearance of (a) Ti/Zr + polymer and (b) PEO + polymer coatings after impact + adhesion tests [63]. Reprinted from [63] with permission from Elsevier.
Atmospheric corrosion tests, as per ASTM B117, and cyclic exposure to salt fog, as per VDA 621-415, revealed no damage either in the PEO or Ti/Zr treatments (Figure 21c,d).
Greater differences were observed on scribed (ASTM D1654, procedure A) specimens, with the PEO + polymer coating showing a higher rating number, i.e., superior performance, than the Ti/Zr + polymer coating (Figure 21e,f). The superior anticorrosion Figure 20. Surface appearance of (a) Ti/Zr + polymer and (b) PEO + polymer coatings after impact + adhesion tests [63]. Reprinted from [63] with permission from Elsevier.
Atmospheric corrosion tests, as per ASTM B117, and cyclic exposure to salt fog, as per VDA 621-415, revealed no damage either in the PEO or Ti/Zr treatments (Figure 21c,d).
Greater differences were observed on scribed (ASTM D1654, procedure A) specimens, with the PEO + polymer coating showing a higher rating number, i.e., superior performance, than the Ti/Zr + polymer coating (Figure 21e,f). The superior anticorrosion properties of the PEO coating can be easily observed in Figure 22, which presents cross-sectional views of the studied samples after scribing and corrosion tests. properties of the PEO coating can be easily observed in Figure 22, which presents crosssectional views of the studied samples after scribing and corrosion tests.  Other examples of polymer top coats enhancing the corrosion resistance of PEO coatings are multiple immersion in a titanium-based organic polymer [318] and sealing with hybrid epoxysilane [319], low-molecular weight polymers (e.g., MALPB, maleic anhydride-g-liquid polybutadiene) [320], and epoxy resins [321], which can be filled with inhibitor loaded nano-or micro-carriers, such as Ce 3+ loaded zeolite microparticles [314] to ensure self-healing.
Post-treatment procedures and their effect on the PEO coating corrosion performance are presented in Table 6.
Based on the corrosion data, post-treatments are the most effective way to improve the corrosion resistance of anodic films on Mg alloys. Corrosion current density values in the order of 10 −9 -10 −10 A cm −2 can be achieved [286,318]. The top coat not only seals the pores but also forms an additional layer that isolates the substrate from external aggressive factors. The thickness of top coats can be up to 130 µm [321]. Methods providing a self-healing effect, such as inhibitor loaded LDH nanocontainers, deserve special attention and should be explored in more detail, since they can actively protect Mg substrates during service by repairing small defects. Some disadvantages of post-treatments are the longer time required to obtain the final product and added cost of the high temperature drying/curing steps which are common in these processes. Other examples of polymer top coats enhancing the corrosion resistance of PEO coatings are multiple immersion in a titanium-based organic polymer [318] and sealing with hybrid epoxysilane [319], low-molecular weight polymers (e.g., MALPB, maleic anhydride-g-liquid polybutadiene) [320], and epoxy resins [321], which can be filled with inhibitor loaded nano-or micro-carriers, such as Ce 3+ loaded zeolite microparticles [314] to ensure self-healing.

Perspectives
In summarizing the recent advances, it is evident that the combined use of protective strategies, i.e., the enrichment of porous PEO layers with corrosion inhibitors, followed by sealing with hybrid or organic coatings loaded with an inhibitor, achieves enhanced long-term corrosion protection for magnesium alloys, providing a strong barrier (when the coating in intact) and active corrosion protection in case the coating is damaged. The protection level of such a system is comparable to that of the gold standard for many decades-a chromate-based conversion layer coated with an organic chromate-based primer. The fault tolerance of the composite coating is the result of the cumulative performance of all its components, i.e.,: i) the anodic film, which provides barrier protection and serves as the repository for ii) corrosion inhibitors that ensure active corrosion suppression; and iii) an organic coating that enhances the protective properties of the barrier, seals the porous anodic layer with inhibitors, and enhances the active corrosion inhibition due to its ability to inhibit the action of corrosive agents [136,326].
Further development can be expected of each element of such combined protection systems, which may also focus on coating multifunctionalization, such as abrasion, wear, thermal resistance, biocidal effect, superhydrophobicity, self-cleaning, etc. For anodic layers in particular, there is a large scope of future work to further minimize energy consumption. In these regards, organic electrolytes, as well as more concentrated alkaline electrolytes, should be explored to bring the coating formation voltage to below 150-200 V and limit the treatment times to under 60 s. Appropriate recycling strategies will have to be considered for such electrolytes in order to achieve sustainability and successful integration into the circular economy. For sealing strategies, trends such as PEO/LDH and PEO/sol-gel combinations are showing great potential. LDH scaffolds can be easily loaded with smart inhibitors, but two major challenges of this strategy remain: (i) optimization of the PEO layer thickness and composition to facilitate the growth of LDH; and (ii) the search for suitable, environmentally friendly inhibitors that can be loaded into LDH. Sol-gel layers, on the other hand, can be loaded with smart nanocontainers, such as clay particles, multiwalled CNT, and mesoporous silica [327]. With regard to the latter in particular, there is a wealth of knowledge in the field of smart drug release for biomedical applications [9,328], but this appears to be almost completely overlooked when it comes to corrosion-resistant engineering applications. The adaptation of mesoporous silica to smart inhibitor release and making it part of a hybrid coating system, e.g., within the LDH layer [329] or organic top-coat [330], seem to be the next logical next steps. The paintability of such hybrid coatings is largely unknown and may present challenges of its own, for instance, due to the hydrophobic nature of some of the loaded inhibitors.
Last but not least, the rapid development of additive manufacturing technologies for magnesium alloys [331][332][333] (which, until recently, has been lagging behind those for other light alloy systems, such as aluminum and titanium, although this is a topic for a separate discussion) will inevitably spur research in the field of corrosion protection by anodic coatings, among other approaches. Such research is still in an embryonic state and so far has been limited exclusively to biomedical applications [334,335].