Improving Corrosion Resistance of Magnesium Alloys via Synergistic Action of TiO₂ Superhydrophobic Coating and Micro-Arc Oxidation

Highlights

What are the main findings?

•

FAS binds via Si-O-Ti bonds, achieving superhydrophobicity (WCA = 160°) and a stable air film.

•

DUV-assisted sol–gel deposits anatase TiO₂ to seal MAO pores and form a dense dual-layer barrier.

What are the implications of the main findings?

•

A three-step strategy was developed to fabricate superhydrophobic FAS-TiO₂-MAO coatings.

•

The composite coating shows ultra-low corrosion current and 6-order-higher corrosion resistance.

•

Synergistic barrier and superhydrophobicity offer a low-cost anti-corrosion route for Mg alloys.

Abstract

To mitigate the intrinsic high corrosion susceptibility of AZ31B magnesium alloy, a three-step synergistic surface modification strategy was developed in this work: initially, a MgO ceramic coating was in situ fabricated on the AZ31B substrate via micro-arc oxidation (MAO); subsequently, a TiO₂ sealing barrier layer was deposited on the MAO coating through a deep ultraviolet (DUV)-assisted sol–gel method; finally, a superhydrophobic top layer was constructed via fluoroalkylsilane (FAS) self-assembly. The microstructural characteristics, chemical compositions and corrosion resistance of the coatings at different modification stages were comprehensively characterized by X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), scanning electron microscopy (SEM), water contact angle (WCA) measurements and electrochemical tests. The results showed that the as-deposited TiO₂ was predominantly anatase phase, and FAS molecules were firmly anchored on the coating surface via Si-O-Ti covalent bonds, endowing the composite coating with a WCA of up to 160°. Electrochemical tests demonstrated that the FAS-TiO₂-MAO composite coating exhibited an ultra-low corrosion current density of 1.31 × 10⁻⁹ A/cm² and a remarkably high charge transfer resistance (R_ct) of 3.46 × 10⁸ Ω·cm². Compared with the bare AZ31B substrate, the corrosion current density was decreased by nearly four orders of magnitude, while the charge transfer resistance was enhanced by approximately six orders of magnitude, indicating a significant improvement in corrosion resistance. Moreover, the composite coating exhibited excellent interfacial adhesion, favorable mechanical durability, and outstanding chemical stability, confirming its reliable long-term corrosion protection and high practical application potential. This work provides a feasible strategy for fabricating high-performance superhydrophobic anticorrosive coatings on magnesium alloys.

1. Introduction

Magnesium (Mg) and its alloy materials have great potential in a wide range of applications, including biodegradable implants due to their excellent mechanical properties, biocompatibility, avirulence and osteogenesis, as well as the electronics, automotive and aeronautical industries due to their strength-weight ratio, light-weight structural characteristics, excellent electromagnetic shielding performance and good castability [1,2,3,4,5]. However, magnesium alloys are not only kinetically corrosion-susceptible and thermodynamically unstable, but also exhibit an excessively high corrosion rate due to a strong dissolution tendency in galvanic coupling systems or chloride-containing environments [6,7]. Therefore, the core of surface modification technologies for enhancing the corrosion resistance of magnesium alloys is to construct a continuous and compact protective coating with excellent interfacial bonding to the substrate, which can effectively isolate the magnesium matrix from corrosive media [8]. In the past few decades, a variety of such surface technologies have been extensively investigated and widely applied, including the preparation technologies of inorganic ceramic coatings (e.g., micro-arc oxidation (MAO) [9,10] and chemical conversion coatings [11,12]), metal plating technologies (e.g., electroless plating and hot-dip plating [13,14]), and organic coating technologies [15] and superhydrophobic surface protection technologies [16].

MAO achieves microarc discharge through the plasma effect and electrochemical reactions in the electrolyte, enabling the in situ formation of an oxide ceramic coating on the magnesium alloy substrate. This coating forms a metallurgical bond with the magnesium alloy substrate, exhibiting excellent bonding strength and spalling resistance. It also features high compactness, along with good wear resistance and oxidation resistance, making it a mainstream technology for fabricating inorganic protective coatings in the field of magnesium alloy surface modification [9]. However, MAO coatings inevitably contain micropores and microcrack defects, which tend to act as permeation pathways for corrosive media. Fortunately, recent studies have demonstrated that the synergistic integration of multiple technologies can significantly enhance the corrosion resistance of magnesium alloys. Representative examples include sol–gel sealing layers to plug MAO pores [17], in situ grown Mg–Al layered double hydroxide (LDH) films followed by hydrophobic modification to construct a dense and water-repellent barrier [18], MAO/epoxy hybrid coatings that seal defects and enable inhibitor-assisted protection/self-healing [19], electrodeposited graphene oxide overlayers that cover porous MAO surfaces and reduce electrolyte ingress [20], and one-step particle-assisted MAO (e.g., hydroxyapatite incorporation) that promotes the “self-sealing” of micropores and densifies the outer layer [21]. Overall, the “MAO + sealing/topcoat” multi-technology synergy is widely recognized as an effective strategy to mitigate the intrinsic porosity of MAO coatings and substantially improve the long-term corrosion resistance of magnesium alloys [22].

Among candidate sealing materials, TiO₂ is particularly attractive due to its high chemical stability and good corrosion resistance. Various TiO₂-containing coatings on magnesium alloys have been developed via different deposition techniques. For instance, Ma et al. employed atomic layer deposition (ALD) to prepare a TiO₂ layer on MAO-coated AZ31B magnesium alloy, which efficiently sealed the micropores and cracks in the MAO coating, thus significantly enhancing the corrosion resistance and biocompatibility [23]; Liu et al. fabricated a hydrophobic TiO₂/MoS₂ nanocomposite coating on AZ31B via electrophoretic deposition, and the silane-modified coating exhibited excellent anti-corrosion performance with a low corrosion current density, supporting the feasibility of TiO₂ as a dense barrier or pore-plugging layer for magnesium alloy protection [24]. Compared with techniques such as ALD and electrophoretic deposition, the sol–gel method is also a mainstream and widely applied technology for fabricating oxide ceramic coatings, boasting distinct advantages including a simple operational process, uniform film formation, and excellent compositional designability. However, the conversion of gel coatings into ceramic functional coatings under high-temperature conditions is accompanied by significant volume shrinkage, which readily induces cracking, delamination, and other failure phenomena of the functional coatings. Fortunately, Kim, Chen et al. have previously developed and demonstrated a modified sol–gel approach, specifically the deep ultraviolet (DUV)-assisted sol–gel technique. Utilizing this method, a series of dense, low-defect oxide coatings have been successfully fabricated at relatively low temperatures, including InGaZnO [25], YBa₂Cu₃O_7−x [26], Zn/Zr/HfO_x [27], (Pb_0.76Ca_0.24)TiO₃ [28]. Their experimental findings also confirmed that photons emitted from DUV irradiation can penetrate the gel films, triggering the formation of free radicals and thereby enabling the conversion of gel films into ceramic films at low temperatures. By promoting the controlled, gradual release of gaseous byproducts during the gel-to-ceramic transformation, this approach mitigates not only the internal stress induced by drastic volume shrinkage but also efficiently minimizes pore defects within the coating, which in turn brings about a remarkable enhancement in the compactness of the as-formed ceramic coatings. Furthermore, considering the relatively low thermal stability of magnesium and its alloys, they are highly susceptible to combustion in oxygen-rich environments at elevated temperatures. Therefore, the DUV-assisted sol–gel technique, characterized by a low thermal budget, represents an ideal strategy for fabricating TiO₂-based anticorrosive coatings on magnesium alloys. Nevertheless, to the best of our knowledge, relevant research has not been reported in the literature to date.

The corrosion behavior of magnesium alloys is closely related to their contact with aqueous corrosive media. Numerous studies have confirmed that constructing a superhydrophobic interface on magnesium alloys via surface modification technologies can form a stable air film on the surface based on the lotus-leaf effect, which effectively blocks direct contact between water, corrosive ions, and the magnesium alloy substrate, and cuts off the mass-transport pathways for corrosion reactions. Meanwhile, superhydrophobicity can significantly reduce the wettability and contact area of corrosive media on the substrate surface, fundamentally inhibiting the initiation and propagation of corrosion [16,29]. This strategy can also be combined synergistically with traditional anti-corrosion technologies to further enhance the corrosion protection performance of magnesium alloys [30,31]. In this work, aiming to improve the corrosion resistance of AZ31B magnesium alloy, an orders-of-magnitude enhancement in corrosion resistance was achieved through a three-step synergistic surface modification strategy: MAO treatment, deposition of a chemically stable TiO₂ barrier layer using a DUV-assisted sol–gel technique, and subsequent superhydrophobic functionalization via fluoroalkylsilane self-assembly. This research provides a feasible route for developing high-efficiency anti-corrosion technologies for magnesium alloys, and may offer a modest contribution to promoting their practical engineering applications in corrosive service environments.

2. Materials and Methods

2.1. Materials

The chemical composition of the AZ31B magnesium alloy sheets is as follows (wt.%): Al 2.5–3.5, Mn 0.2–1.0, Si ≤ 0.08, Cu ≤ 0.01, Zn 0.6–1.4, Ca ≤ 0.04, balance Mg. The alloy was provided by Jiangsu Hengshen Co., Ltd., Zhenjiang, China. Tetra-n-butyl titanate, anhydrous ethanol, acetylacetone, sodium silicate, potassium hydroxide, potassium fluoride, and acetone (all analytical reagent grade, AR) were purchased from Sinopharm Chemical Reagent Co., Ltd., Shanghai, China. 1H,1H,2H,2H-perfluorooctyltrichlorosilane (FAS, AR) was sourced from Shanghai Aladdin Biochemical Technology Co., Ltd., Shanghai, China. Waterproof abrasive paper (280#, 600#, 800#, 1200#) was procured from 3M China Co., Ltd. (Shanghai, China). Deionized water was prepared in the laboratory using a Millipore ultrapure water purification system (Millipore, Burlington, MA, USA).

2.2. Magnesium Alloy Specimen Pre-Treatment

AZ31B magnesium alloy was cut into rectangular specimens with dimensions of 25 mm × 25 mm × 5 mm, followed by sequential grinding and polishing to refine the surface. Grinding was performed using 280#, 600#, 800# and 1200# waterproof abrasive papers in turn, and subsequent mechanical polishing was carried out on a polishing machine. This treatment was intended to smooth the rough and irregular alloy surface, remove surface oxide films and other contaminants, and reduce surface defects. The polished magnesium alloy specimens were ultrasonically cleaned with anhydrous ethanol and acetone, then thoroughly rinsed with deionized water and dried. The dried specimens were subsequently sealed with plastic wrap, which serves to prevent the introduction of detrimental impurities into the electrolyte during subsequent experiments.

2.3. TiO₂ Photosensitive Sol Preparation

The TiO₂ photosensitive sol was prepared via a sol–gel route using a ternary solution system, as reported in our previous work [32]. All solution manipulations were performed in an argon-purged glove box at ambient temperature, with the relative humidity maintained below 30%. Tetra-n-butyl titanate, anhydrous ethanol, and acetylacetone were employed as the precursor, solvent, and chelating agent, respectively. Specifically, 1.60 g of acetylacetone was dissolved in 25.95 g of anhydrous ethanol, and the mixture was stirred magnetically until homogeneous and transparent. Thereafter, 5.45 g of tetra-n-butyl titanate was added dropwise into the solution, followed by continuous magnetic stirring for a further 3 h to yield a stoichiometric precursor solution. The as-prepared mixture was then statically aged for 24 h, affording a pale yellow TiO₂ photosensitive sol with a concentration of 0.4 mol/L.

2.4. MAO Treatment

MAO treatment was performed using a self-developed MAO75–III apparatus manufactured by Xi’an University of Technology (Xi’an, China). AZ31B magnesium alloy specimens were fixed with aluminum wires and subsequently immersed in a prepared electrolyte. The electrolyte was formulated by dissolving 8 g sodium silicate, 10 g potassium hydroxide, and 8 g potassium fluoride in 1 L of deionized water. During the MAO process, the electrical parameters were set as follows: an applied voltage of 420 V, a pulse frequency of 350 Hz, a pulse duty cycle of 12%, and a treatment time of 5 min. The electrolyte temperature was controlled and maintained at 25–40 °C throughout the entire treatment. Temperatures below 25 °C lead to a low coating growth rate and a high density of defects within the coating, thus deteriorating its corrosion resistance. In contrast, temperatures above 40 °C give rise to excessively intense micro-arc discharges, causing surface ablation and crack formation, which significantly reduce the compactness and corrosion resistance of the coating. The as-prepared specimens were denoted as MAO-AZ31B. A comprehensive account of the experimental procedure is shown in Figure 1.

2.5. Deposition of TiO₂ Coatings

TiO₂ gel films were deposited on the MAO layer surfaces using the dip-coating method at a precisely controlled withdrawal speed of 2 mm/s. The resultant gel films were first oven-dried at 80 °C for 15 min. The dried specimens were then transferred onto a hot stage with a preset temperature of 150 °C. Once the temperature of the specimens reached 120 °C, the specimens were exposed to DUV irradiation continuously for 120 min. The DUV light was supplied by a low-pressure ultraviolet mercury lamp, which emitted radiation at wavelengths of 253.7 nm (90%) and 184.9 nm (10%), with an energy density of 25 mW/cm². After irradiation on one side was completed, the specimen was turned over, and the other side of the substrate was subjected to the same irradiation procedure. Ultimately, the specimens were annealed in a muffle furnace at 400 °C for 30 min, and the as-prepared specimens were denoted as TiO₂-MAO-AZ31B. For comparison, two reference samples were prepared using different processes: one via the conventional sol–gel technique (i.e., dipping followed by drying at 80 °C and direct annealing at 400 °C without DUV irradiation), and the other subjected only to DUV irradiation without subsequent thermal annealing.

2.6. Surface Modification for Superhydrophobicity

A 0.5 wt.% FAS/ethanol solution was prepared by dissolving 0.39 g of FAS into 100 mL of absolute ethanol, followed by magnetic stirring for 15 min. The as-prepared TiO₂-MAO-AZ31B specimens were immersed in the aforementioned FAS/ethanol solution and kept statically for 12 h. After immersion, the specimens were removed and blow-dried with nitrogen gas. The specimens were then placed in a preheated oven at 80 °C for a further 12 h of drying. The final specimens obtained after the entire process were denoted as FAS-TiO₂-MAO-AZ31B.

2.7. Material Characterization

(Building 100, Zelenograd, Moscow, Russia). Surface wettability of the specimens was evaluated by means of water contact angle (WCA) measurements using a JC2000A optical contact angle goniometer (Dongguan Shengding Precision Instrument Co., Ltd., Dongguan, China). All characterization tests were performed under ambient light conditions, with the test temperature controlled at 25 ± 1 °C and the relative humidity maintained at 45% ± 5%. During the measurement process, 5 μL deionized water droplets were deposited onto the surface of each specimen. To guarantee the repeatability and reliability of the test results, five independent tests were carried out at randomly selected testing regions on each specimen surface. The average value calculated from these sets of independent data was ultimately adopted as the determined WCA value for each individual specimen. The corrosion resistance of the specimens was systematically investigated utilizing potentiodynamic polarization curves and electrochemical impedance spectroscopy (EIS). All electrochemical measurements were performed in a 3.5 wt.% NaCl solution at room temperature using a conventional three-electrode system: a saturated calomel electrode (SCE) as the reference electrode, uncoated and coated AZ31B samples (exposed area: 1 cm²) as the working electrode, and a platinum electrode as the counter electrode. Before testing, the working electrode was immersed in the electrolyte until a stable open circuit potential (OCP) was achieved. EIS measurements were conducted over a frequency range from 10⁻² Hz to 10⁵ Hz, with 10 points per decade and a sinusoidal voltage amplitude of 10 mV (RMS). For comparison, polarization curves were measured from −0.5 V to +0.5 V vs. OCP at a scan rate of 2 mV/s. All electrochemical tests were performed using a CHI660D electrochemical workstation, and EIS data were fitted and analyzed using ZView (version 2) software.

The adhesion between the FAS-TiO₂-MAO coating and the AZ31B substrate was evaluated via a cross-cut test in accordance with the Chinese national standard GB/T 9286-2021 (Paints and varnishes—Cross-cut test; China Standards Press: Beijing, China, 2021). Six parallel cuts with 1 mm spacing were made on the coating surface, followed by six perpendicular cuts to form a square grid. After removing loose debris, pressure-sensitive adhesive tape was firmly applied to the grid and rapidly peeled off at an angle of approximately 60°. The adhesion grade was determined by visually examining the peeling area according to the standard. For the tribological durability test of FAS-TiO₂-MAO-AZ31B, one side of the sample was fixed to a 500 g weight using double-sided tape. The opposite side was placed on 600# sandpaper and uniformly dragged 30 cm in a single direction. After each cycle, the WCA was measured. This dragging process was repeated ten times to assess the mechanical stability of the FAS-TiO₂-MAO coating. In the chemical stability test, the FAS-TiO₂-MAO-AZ31B specimen was immersed in a 3.5 wt.% NaCl solution. At 24 h intervals, the sample was removed, cleaned with ethanol, and dried prior to WCA measurement.

3. Results and Discussion

When the TiO₂-MAO-AZ31B reference sample was prepared via the conventional sol–gel method, the sample was found to be extremely prone to spontaneous combustion during thermal annealing. This phenomenon can be attributed to the rapid exothermic decomposition of organic components within the TiO₂ gel layer at high temperatures, which triggered the combustion of magnesium. In addition, even when a TiO₂ coating was successfully obtained by chance, it exhibited abundant surface cracks and poor adhesion with severe delamination. This is mainly ascribed to the rapid decomposition of organic species and significant volume shrinkage during direct thermal annealing, together with the relatively slow atomic diffusion rate and the constraint effect from the underlying layer, which generated considerable internal stress within the coating. For the sample treated only by DUV irradiation without thermal annealing, the TiO₂ coating showed very weak adhesion and could be easily detached by slight scratching or abrasion. Both types of coatings are unsuitable for use as anticorrosive protective coatings. Therefore, only the specimens prepared by the DUV-assisted sol–gel method are included in the subsequent discussion.

As depicted in Figure 2, the XRD pattern of the AZ31B substrate is dominated by diffraction signals characteristic of metallic Mg with a hexagonal close-packed structure. The distinct diffraction peaks are mainly situated at around 2θ = 32.2°, 34.4°, 36.6°, 47.8°, 57.4°, and 63.1° which can be sequentially indexed to the (100), (002), (101), (102), (110) and (1120) crystallographic planes of α-Mg (JCPDS No. 35–0812). No diffraction peaks corresponding to Al- or Zn-bearing secondary phases are discernible, implying that the concentrations of alloying elements in the alloy are low and they exist in the Mg matrix mainly in the form of solid solution. Following MAO treatment, fresh diffraction peaks emerge at 2θ values of 42.9° and 62.3°, which are assigned to the (200) and (220) planes of cubic MgO (JCPDS No. 45–0946), respectively. This verifies that a crystalline MgO ceramic layer is in situ constructed on the AZ31B substrate through the MAO process. In the XRD pattern of the TiO₂-MAO-AZ31B specimen, apart from the characteristic peaks of the Mg substrate and the MgO ceramic layer, a set of additional diffraction signals is identified at 2θ = 25.3°, 68.8° and 70.3°. These peaks match well with the (101), (116) and (220) planes of anatase-phase TiO₂, in accordance with the JCPDS standard card No. 21–1272 [33]. These findings confirm that a crystalline TiO₂ coating has been successfully deposited via the DUV-assisted sol–gel method, with anatase as the primary crystalline phase. It is widely acknowledged that the surface of anatase TiO₂ is more prone to hydroxylation relative to the rutile phase, and the resultant hydroxyl groups can covalently bond with fluoroalkylsilane, thereby endowing the surface with high hydrophobicity [34].

Figure 3 depicts the XPS survey and high-resolution spectra for the coatings prior to (group a: TiO₂-MAO-AZ31B) and following FAS functionalization (group b: FAS/TiO₂-MAO-AZ31B). The survey spectrum in Figure 3(a1) reveals that the pristine TiO₂-MAO coating is composed predominantly of Mg, O, Ti, Ca, C, and Si. The elemental inventory of the FAS/TiO₂-MAO specimen is largely analogous, with the only new signature being a distinct fluorine (F) signal (Figure 3(b1)). No discernible shifts in the binding energies of Mg, O, Ti and Ca are observed between the two specimens, a trend consistent with their respective sources: Mg and Ti are derived from the AZ31B substrate and the TiO₂ coating, respectively, while Ca is present as an impurity element introduced from the electrolyte during the MAO process. Relative to the bare TiO₂-MAO surface, the FAS/TiO₂-MAO surface presents a strong characteristic peak of fluorine, which provides direct spectroscopic proof that FAS molecules have been stably anchored onto the oxide coating. Pronounced differences are resolved in the high-resolution C 1s and Si 2p spectra between the two specimen groups. As illustrated in Figure 3(a2), the high-resolution C 1s spectrum of TiO₂-MAO coating can be deconvoluted into two characteristic peaks at 284.8 eV and 288.8 eV. These peaks are assigned to adventitious carbon accumulated during XPS measurement and carbonate species formed from residual carbon upon thermal annealing in the DUV-assisted sol–gel procedure, respectively. The C 1s spectrum of FAS/TiO₂-MAO coating (Figure 3(b2)) displays a more complex multiplet structure. In addition to the two common peaks at 284.8 eV and 288.8 eV, three extra fitted components are detected at 286.2 eV, 291.5 eV, and 293.5 eV, which correspond to C–O moieties, –CF₂ groups and –CF₃ terminal groups in turn. All these functionalities are characteristic structural segments of the FAS molecule [35,36]. The Si 2p peak of the TiO₂-MAO coating is centered at 103.4 eV (Figure 3(a3)), consistent with the binding energy of Si 2p in Na₂SiO₃ [37,38], indicating that this silicon component originates from the electrolyte used in the MAO process. For the FAS/TiO₂-MAO coating (Figure 3(b3)), the Si 2p signal can be curve-fitted into two distinct peaks at 103.4 eV and 102.3 eV. The peak at 103.4 eV is ascribed to residual Na₂SiO₃ from the MAO electrolyte, whereas the peak at 102.3 eV is assigned to Si–O chemical bonds. The appearance of this Si–O feature consolidates the conclusion that FAS is chemically conjugated to the TiO₂ coating via covalent Si–O linkages [38,39].

Figure 4a presents the SEM surface morphology of the mechanically polished AZ31B substrate. Strikingly, the surface is devoid of voids or other visible defects, and features uniformly oriented, continuously distributed shallow parallel polishing scratches. These observations collectively indicate that the AZ31B substrate boasts a dense microstructure and attains excellent overall surface planarity post-polishing. Figure 4b illustrates the SEM surface morphology of the AZ31B substrate following MAO treatment. The surface displays a uniformly distributed honeycomb-like porous architecture, characterized by irregular near-circular or elliptical pores with diameters ranging from 0.5–1.5 μm and subtle size variations. The pore walls are rough and interconnected, thus forming a continuous 3D network framework—a signature morphology resulting from transient high-temperature melting, rapid solidification, and gas evolution during the MAO process. Of note, no discernible cracks or spalling are detected, demonstrating that the MAO ceramic layer exhibits favorable structural integrity and densification. Figure 4c illustrates the SEM surface morphology of the TiO₂-MAO coating. Relative to the as-prepared MAO surface (Figure 4b), the pore count is markedly diminished, with residual pores mostly <1 μm in diameter and no longer interconnected. In addition, the AFM topographic map (Inset of Figure 4c) reveals a smooth surface with an arithmetic mean roughness (Ra) of only 0.18 μm. The TiO₂ coating has efficiently sealed and filled the porous structure of the MAO substrate, thereby yielding a substantial improvement in surface planarity and a relatively dense, continuous film morphology. Nevertheless, some reduced-size depressions and residual pores persist, reflecting incomplete filling of the MAO substrate’s inherent porous architecture. Evidently, no obvious cracks or delamination are observed on the coating surface, verifying robust interfacial adhesion between the TiO₂ coating and MAO substrate, and high structural integrity of the coating itself.

The cross-sectional SEM morphology of the TiO₂-MAO-AZ31B specimen is shown in Figure 4d. A well-defined three-tiered architecture is evident, comprising the AZ31B magnesium alloy substrate (left), MAO ceramic interlayer (middle), and TiO₂ functional coating (right). The AZ31B substrate on the left preserves parallel polishing-induced textures and features a dense, defect-free microstructure, providing a stable foundation for the overlay coatings; the MAO interlayer in the middle possesses a typical porous columnar structure with a uniform thickness of ~8 μm, and this porous configuration not only bonds tightly to the substrate but also enhances mechanical interlocking and offers anchoring sites for TiO₂ deposition; The TiO₂ coating, with a thickness of approximately 3 μm, is generally continuous. The few voids observed in the composite coating can be ascribed to two factors: on the one hand, the entrapped air in the porous MAO layer prevents full infiltration of the TiO₂ sol, so some inherent voids remain even after DUV irradiation and thermal annealing, although the TiO₂ coating still fills and seals most pores and the mechanical interlocking and pinning effect in the filled regions ensure robust interfacial bonding between the TiO₂ coating and the MAO interlayer; on the other hand, these voids are also largely artifacts induced by the brittle fracture during cross-sectional sample preparation, as the brittle and porous MAO ceramic tends to crack under mechanical stress, rather than resulting from the failure of the TiO₂ coating itself. Of supplementary note, FAS molecules are anchored onto the TiO₂ surface through covalent Si–O–Ti bonds, forming a self-assembled monolayer. The as-formed modifier layer possesses a typical thickness of 1–10 nm, which is orders of magnitude smaller than the characteristic characterization scale of scanning electron microscopy (spanning from submicrometer to micrometer levels). Therefore, the surface topography of the FAS-TiO₂-MAO composite coating is highly consistent with that of the unmodified TiO₂-MAO coating, and no distinguishable morphological variations can be identified between these two specimens via SEM characterization.

Figure 5a depicts the WCA micrograph of the AZ31B substrate. The water droplet assumes a flattened, spread-out profile on the surface, yielding a measured WCA of 45°—this confirms the pronounced hydrophilic nature of the AZ31B substrate. The water droplet on the MAO-treated AZ31B retains a flattened, spread-out profile, with a measured contact angle of 26° (Figure 5b)—a marked reduction from the 45° of the bare AZ31B substrate, indicating a substantial enhancement in surface hydrophilicity following MAO treatment. This enhancement is ascribed to the porous architecture of the MAO ceramic layer and abundant hydroxyl (–OH) polar groups derived from magnesium oxide phases, which not only increase surface adsorption sites but also elevate surface energy, thereby facilitating further spreading of water droplets. When a water droplet is deposited onto the surface of the TiO₂-MAO-AZ31B, it spreads rapidly into an extremely thin liquid film (Figure 5c), with a measured WCA of only 3°, demonstrating that the coating exhibits prominent superhydrophilicity. The marked superhydrophilic performance observed here is well elucidated by the Wenzel wetting model (Figure 6a), which correlates the apparent wettability and surface roughness through the equation cos θ_r = r · cos θ_y. In this expression, θ_r is the experimentally measured contact angle on the rough coating surface, r signifies the roughness factor (r > 1 for surfaces with topological roughness), and θ_y refers to the intrinsic contact angle on a smooth counterpart. Owing to the high density of polar hydroxyl (–OH) groups originating from the TiO₂ component, the coating displays strong intrinsic hydrophilicity with θ_y typically lower than 30° [40] corresponding to a large positive cos θ_y. Since the TiO₂-modified MAO structure retains a hierarchically rough morphology with r > 1, the Wenzel mechanism predicts a further enhancement in hydrophilicity, which pushes the apparent WCA down to only 3° and gives rise to the superhydrophilic characteristic. Figure 5d displays the WCA micrograph of the FAS-functionalized TiO₂-MAO coating, where a water droplet rests stably in a near-spherical conformation, registering a WCA of 160°—clear evidence of the coating’s robust superhydrophobic nature. The underlying mechanism is well captured by the Cassie–Baxter wetting model (Figure 6b), which describes wetting behavior on heterogeneous rough surfaces through the core relationship: cos θ_c = f₁cos θ₁ + f₂cos θ₂. In this expression, θ_c denotes the experimentally observed apparent contact angle, f₁ and f₂ represent the fractional areas of the solid–liquid and air-liquid interfaces, respectively (where f₁ + f₂ = 1), θ₁ is the intrinsic contact angle of the solid surface, and θ₂ corresponds to the contact angle of air (≈180°, so cos θ₂ = −1). For the FAS-TiO₂-MAO coating, FAS molecules are anchored to the surface via covalent Si–O–Ti bonds, introducing low-surface–energy fluorocarbon moieties (–CF₃, –CF₂–) that displace the polar hydroxyl (–OH) groups present on the pristine TiO₂-MAO coating. This chemical modification raises the intrinsic contact angle θ₁ to above 120° [39,41], leading to a negative value for cos θ₁. Simultaneously, the coating preserves the micro-nano hierarchical roughness inherited from the MAO substrate, which traps a significant volume of air within its textured crevices. This leads to a high value of f₂, the air–liquid interfacial area fraction. When these parameters are substituted into the Cassie–Baxter equation, the combined influence of the negative cos θ₁ and cos θ₂ values produces a more negative cos θ_c, which corresponds to an apparent contact angle θ_c approaching 180°. The measured 160° WCA is a direct manifestation of this synergistic interplay: chemical functionalization to lower surface energy, coupled with the retained micro-nano roughness, operates through the Cassie–Baxter mechanism to drive the dramatic transition from superhydrophilicity to superhydrophobicity.

Owing to its facile operation and rapid measurement, the Tafel polarization technique has been extensively adopted for the determination of instantaneous corrosion rates. Corrosion resistance of the specimens is generally evaluated on the basis of corrosion current density (i_corr) and corrosion potential (E_corr) [42]. A lower i_corr value coupled with a more positive E_corr is indicative of superior anti-corrosion performance. Figure 7 plots the Tafel polarization curves of the bare AZ31B substrate, MAO-AZ31B, TiO₂-MAO-AZ31B, and FAS/TiO₂-MAO-AZ31B specimens after 1 h immersion in 3.5 wt.% NaCl solution, enabling a comparative analysis of the electrochemical corrosion behaviors among the four specimens. The abscissa represents the logarithm of corrosion current density, while the ordinate denotes the potential relative to the saturated calomel electrode (VSCE). Of particular note, the FAS/TiO₂-MAO-AZ31B specimen exhibits the most positive corrosion potential at approximately −1.34 V, which is considerably higher than those of the TiO₂-MAO-AZ31B specimen (−1.53 V), bare AZ31B substrate (−1.55 V), and MAO-AZ31B specimen (−1.57 V) (

Table 1. Summary of electrochemical corrosion parameters (i_corr, E_corr) for AZ31B-based specimens after 1 h immersion in 3.5 wt.% NaCl solution.

Specimen	Ecorr (V/SCE)	icorr (A/cm2)
AZ31B	−1.55	1.69 × 10−4
MAO-AZ31B	−1.57	5.04 × 10−6
TiO2-MAO-AZ31B	−1.53	1.64 × 10−7
FAS/TiO2-MAO-AZ31B	−1.34	1.31 × 10−9

). From a thermodynamic perspective, the ranking of the four specimens in terms of thermodynamic stability and resistance to spontaneous corrosion is as follows: FAS/TiO₂-MAO-AZ31B > TiO₂-MAO-AZ31B > AZ31B substrate > MAO-AZ31B. Among these, the FAS/TiO₂-MAO-AZ31B specimen possesses the optimal thermodynamic stability and the lowest susceptibility to spontaneous corrosion. Unexpectedly, MAO treatment shifts the corrosion potential of AZ31B substrate negatively by 0.02 V, implying a deterioration in thermodynamic stability and corrosion resistance upon the application of MAO modification. This observation arises primarily because the porous ceramic layer generated via MAO allows electrolyte infiltration to the coating–substrate interface, giving rise to local galvanic couples consisting of the ceramic coating (cathodic phase) and the Mg alloy substrate (anodic phase). Such galvanic coupling drives the overall corrosion potential toward the negative potential regime associated with the anodic substrate. It should be emphasized that thermodynamic parameters merely reflect the intrinsic tendency toward corrosion, whereas the practical corrosion resistance is governed by the kinetic rate of the corrosion process.

The kinetic corrosion rate can be quantitatively described by i_corr. The icorr value of the FAS/TiO₂-MAO-AZ31B specimen is measured at approximately 1.31 × 10⁻⁹ A/cm², which is orders of magnitude lower than those of the TiO₂-MAO-AZ31B specimen (1.64 × 10⁻⁷ A/cm²), MAO-AZ31B specimen (5.04 × 10⁻⁶ A/cm²), and bare AZ31B substrate (1.69 × 10⁻⁴ A/cm²). Accordingly, the practical corrosion resistance of the specimens follows the sequence: FAS/TiO₂-MAO-AZ31B > TiO₂-MAO-AZ31B > MAO-AZ31B > AZ31B substrate, with the FAS/TiO₂-MAO-AZ31B assembly delivering the slowest corrosion rate and the most prominent protective performance. The above trend originates from divergent protective mechanisms inherent to each coating system. The FAS/TiO₂-MAO coating integrates the physical barrier effect of the compact TiO₂-MAO ceramic layer and the hydrophobic shielding imparted by fluorocarbon groups introduced through FAS functionalization. This synergistic interaction significantly impedes electrolyte permeation and access to the underlying metallic surface. For the TiO₂-MAO coating, TiO₂ nanoparticles fill and refine the pore network within the MAO layer, reinforcing the barrier capability. In contrast, the pristine MAO coating consists solely of a porous ceramic structure, imposing only limited restriction on electrolyte penetration. The uncoated AZ31B substrate undergoes direct corrosive attack at the metallic surface, resulting in the highest kinetic corrosion rate among all investigated specimens.

EIS measurements were further performed to evaluate the corrosion resistance of the four prepared specimens. The EIS plots of all specimens after 1 h of immersion in 3.5 wt.% NaCl electrolyte are depicted in Figure 8, and the acquired EIS data were systematically fitted and analyzed using ZView software. The obtained Nyquist plots are shown in Figure 8a–c. Both the FAS/TiO₂-MAO-AZ31B and TiO₂-MAO-AZ31B coatings exhibited the typical impedance response characteristics of magnesium alloy protective coatings. Combined with the Bode phase angle-frequency plot in Figure 8d, it can be confirmed that both specimens presented a distinct double capacitive arc feature, which corresponded to the capacitive response of the coating bulk and the charge transfer process at the coating-substrate interface, respectively. In addition, Figure 8d revealed that, in comparison with the TiO₂-MAO-AZ31B, the FAS/TiO₂-MAO-AZ31B exhibited the broadest phase angle plateau near 80° over the frequency range of 10⁻² to 10⁴ Hz—a typical characteristic indicative of ideal capacitive protective behavior and excellent structural integrity of the coating. Meanwhile, the impedance modulus at 0.01 Hz (|Z|_0.01 Hz) in Figure 8e remains at the 10⁷ Ω·cm² order of magnitude, significantly higher than that of other samples, further verifying the long-term corrosion resistance and reliability of the coating system.

This characteristic directly verified their designed bilayer protective structure, which was composed of a hydrophobic top layer and a TiO₂/MAO ceramic barrier sublayer. Notably, the capacitive arc radius of the FAS/TiO₂-MAO-AZ31B (Figure 8a) was considerably larger than that of the TiO₂-MAO-AZ31B (Figure 8b). This significant performance enhancement was attributed to the fluorocarbon groups introduced by FAS modification on the surface of the TiO₂-MAO coating, which constructed a stable low-surface-energy hydrophobic layer on the coating surface. This hydrophobic layer thus significantly hindered the penetration of electrolyte into the coating top layer, thereby resulting in a remarkable improvement in the corrosion resistance of the coating system. Figure 8c showed that the MAO-AZ31B and bare AZ31B substrates exhibited similar impedance response characteristics, with their Nyquist curves composed of both capacitive and inductive arcs. The radius of the capacitive arc for the MAO coating was significantly larger than that of the AZ31B substrate, which validated the basic protective effect of the MAO coating on the AZ31B substrate [30]. In contrast, the bare AZ31B substrate presented an extremely small capacitive arc, accompanied by a low and drastically fluctuating phase angle (Figure 8d), demonstrating intense corrosion reactions on its surface and extremely poor corrosion resistance. In EIS analysis, the low-frequency impedance modulus |Z|_f→0 from the Bode plot is a crucial parameter for evaluating the corrosion rate of specimens, where a higher |Z|_f→0 value indicates a lower corrosion rate [19]. As shown in Figure 8e, the |Z|_f→0 value of the AZ31B substrate increased from approximately 10² Ω·cm² to about 10³ Ω·cm² after the deposition of the MAO coating. Subsequent deposition of the TiO₂ sealing barrier layer further raised this value to around 3 × 10⁵ Ω·cm², and a dramatic increase to approximately 2 × 10⁷ Ω·cm² was achieved after FAS modification. Collectively, the stepwise modification involving MAO treatment, TiO₂ sealing barrier deposition, and FAS superhydrophobic modification realized a progressive enhancement in the corrosion resistance of the AZ31B magnesium alloy. The synergistic effect of these three coating modification approaches ultimately improved the corrosion resistance of the AZ31B magnesium alloy by nearly six orders of magnitude relative to the bare substrate.

To quantitatively evaluate the corrosion resistance of the FAS-TiO₂-MAO coating, the EIS curves of the specimens were fitted and interpreted using the equivalent circuits illustrated in Figure 9, with the corresponding fitting results summarized in

Table 2. Electrochemical parameters extracted from equivalent circuit fitting of EIS curves.

Specimen	Rcoat	CPEcoat	RMAO	CPEMAO	Rct	CPEdl
	(Ω·cm2)	(F/cm2)	(Ω·cm2)	(F/cm2)	(Ω·cm2)	(F/cm2)
AZ31B	-	-	-	-	1.26 × 102	1.75 × 10−5
MAO–AZ31B	-	-	1.88 × 102	9.42 × 10−7	2.85 × 103	2.93 × 10−6
TiO2–MAO–AZ31B	1.37 × 103	1.88 × 10−9	4.70 × 104	4.52 × 10−7	2.52 × 105	2.86 × 10−7
FAS/TiO2–MAO–AZ31B	2.12 × 105	2.22 × 10−8	2.04 × 105	9.71 × 10−9	3.46 × 108	8.49 × 10−8

. The equivalent circuit in Figure 9a was applicable to the AZ31B substrate, where the resistances R_s, R_ct and R_pit denote the electrolyte resistance, charge transfer resistance and pitting corrosion resistance, respectively. The inductance L is associated with the deposition-dissolution process of corrosion products generated during the corrosion of the magnesium alloy substrate. To compensate for the system inhomogeneity caused by the rough and porous surface of the specimens, all capacitive components in the circuits were replaced by constant phase elements (CPE_coat) [43]. Herein, CPE_dl represents the capacitive component at the magnesium alloy/electrolyte interface. The equivalent circuit in Figure 9b was used for the MAO ceramic coating, with R_MAO and CPE_MAO corresponding to the resistance and capacitance of the MAO coating, respectively. For the TiO₂-MAO composite coating, the equivalent circuit in Figure 9c was adopted, where R_coat and CPE_coat stand for the resistance and capacitance of the outer TiO₂ layer, respectively. It is well established that the capacitive arc in the high or medium frequency region is related to the outer coating layer, while the electrochemical behavior in the low frequency region is associated with the interface layer between the coating and the substrate. A higher R_ct value indicates a greater resistance to the corrosion reaction at the electrode/electrolyte interface, and a larger R_coat value reflects a stronger barrier property of the coating layer. In addition, a smaller CPE value and a CPE exponent n closer to 1 imply that the capacitive behavior of the interface/coating is more similar to that of an ideal capacitor, corresponding to a more uniform surface. Since the FAS surface modification did not alter the serial bulk structure of the coating, the equivalent circuit for the FAS-TiO₂-MAO coating was identical to that for the TiO₂-MAO coating (Figure 9d).

Analysis of the electrochemical data in Table 2 revealed that the AZ31B substrate exhibited an extremely low R_ct value of only 1.26 × 10² Ω·cm², indicating a weak resistance to the interfacial corrosion reaction and thus poor corrosion resistance. After MAO treatment, the R_ct of the coating increased to 2.85 × 10³ Ω·cm², which was approximately 20 times higher than that of the substrate. Meanwhile, the CPE_MAO value decreased to 9.42 × 10⁻⁷ F/cm² compared with the substrate, implying a reduction in the double-layer capacitance. The formation of the MAO coating on the substrate surface constructed a physical barrier that significantly inhibited the corrosion process, thus improving the corrosion resistance to a certain extent. For the TiO₂-MAO coating, the R_ct value reached 2.52 × 10⁵ Ω·cm², an increase of about 80 times relative to the single MAO coating, accompanied by a further decrease in the CPE value. This result demonstrated that the TiO₂-MAO bilayer structure formed a dual protective barrier, which further hindered the penetration of corrosive media and led to a remarkable enhancement in corrosion resistance. As for the FAS/TiO₂-MAO composite coating, it achieved an R_ct value of 3.46 × 10⁸ Ω·cm² and a CPE_coat value of 2.22 × 10⁻⁸ F/cm². Although the CPE_coat value slightly increased in comparison with the TiO₂-MAO coating, the R_ct value was drastically elevated, and the CPE exponent n was closer to 1. These results indicated that the FAS modification effectively improved the hydrophobicity and surface uniformity of the coating, endowing the FAS/TiO₂-MAO coating with the optimal corrosion resistance among all the specimens.

As an anticorrosive coating, the adhesion between the FAS/TiO₂-MAO coating and the AZ31B substrate is critical for the long-term stability and protective performance of FAS/TiO₂-MAO-AZ31B. Tape-peel test results showed no obvious coating delamination at the cross-cut edges or intersections. According to GB/T 9286-2021, the coating achieved a grade 0 adhesion, indicating excellent bonding strength. Such superior adhesion can be attributed to the DUV-assisted sol–gel method, which minimizes defects in the TiO₂ coating, and the pinning effect provided by the TiO₂ filling into the MAO pores. A wear test was conducted to investigate the mechanical durability of the FAS/TiO₂-MAO coating. As shown in Figure 10a, the WCA of the FAS/TiO₂-MAO-AZ31B surface decreases with increasing wear distance; the surface still exhibits a WCA higher than 150° at a wear distance of 180 cm, remaining superhydrophobic, which demonstrates that the FAS/TiO₂-MAO coating possesses reasonable wear resistance. Thereafter, the WCA decreases rapidly due to the significantly increased actual contact area and accelerated wear process.

Figure 10b displays the evolution of the WCA on the surface of the FAS/TiO₂-MAO-AZ31B specimen as a function of immersion time in 3.5 wt.% NaCl solution. Overall, the WCA decreases with increasing immersion time. Within the first 168 h, the WCA declines slowly and remains above 150°, retaining superhydrophobicity, which confirms that the FAS/TiO₂-MAO coating possesses excellent chemical stability. After 168 h, the WCA decreases significantly and drops to 52° at 360 h, indicating a complete loss of hydrophobicity, accompanied by local delamination of the coating. This phenomenon reveals that during the initial immersion stage, the FAS-dominated superhydrophobic coating can trap an air layer to isolate the substrate from corrosive media, thus remarkably reducing the corrosion rate. With prolonged immersion, the air layer gradually disappears, leading to an increased contact area between the corrosive medium and the TiO₂/MAO coating. Meanwhile, the penetration and pitting corrosion of the corrosive solution are continuously intensified, eventually resulting in the failure of the superhydrophobic anticorrosive coating. Therefore, improving the chemical stability and durability of superhydrophobic coatings is of great significance for enhancing their long-term corrosion resistance. In this work, the FAS superhydrophobic layer is a monomolecular self-assembled layer with limited chemical stability and shielding effect against aqueous corrosive media. Future studies focusing on increasing the thickness and stability of the superhydrophobic modifier are expected to achieve superior long-term corrosion resistance.

Overall, the as-prepared FAS/TiO₂-MAO coating exhibits excellent corrosion resistance and durability, with a E_corr of −1.34 V, a i_corr as low as 1.3 × 10⁻⁹ A/cm², and an impedance modulus at 0.01 Hz (|Z|_0.01 Hz) of ~2 × 10⁷ Ω·cm². The coating retains superhydrophobicity and effective corrosion protection for at least 168 h in 3.5 wt.% NaCl solution. As presented in

Table 3. Comparison of corrosion resistance and durability of representative coatings on magnesium alloys.

Coating System	Ecorr (V/SCE)	icorr (A/cm2)	|Z|0.01 Hz (Ω·cm2)	Durability (3.5% NaCl)	Ref.
Superhydrophobic stearic-acid-modified Ni–P/Cu/Ni	−0.37	4.0 × 10−7	~104	NR	[44]
Silica-nanoparticle-reinforced hybrid sol–gel/PANI	NR	NR	~106	≥120 h	[45]
Organically modified silicate hybrid sol–gel	−1.40	~10−7	~106	≥72 h	[46]
Self-healing CeOx	NR	~10−7	~106	≥120 h	[47]
ALD-deposited TiO2 sealing layer on MAO	−1.35	8.1 × 10−7	NR	≥168 h	[23]
Hydrophobic TiO2/MoS2 nanocomposite	−1.38	~10−8	~106	≥120 h	[24]
GPTMS/TEOS-derived organic–inorganic silica	−1.45	3.7 × 10−9	~107	≥96 h	[48]
Superhydrophobic PFOTES-modified PVDF/SiO2	−1.32	2.4 × 10−9	NR	≥168 h	[49]
FAS/TiO2-MAO	−1.34	1.3 × 10−9	~2 × 107	≥168 h	The work

, its comprehensive performance outperforms that of several recently reported protective coatings on AZ31/AZ31B magnesium alloys, including superhydrophobic stearic acid-modified Ni-P/Cu/Ni [44], silica-nanoparticle-reinforced hybrid sol–gel/polyaniline (PANI) [45], organically modified silicate hybrid sol–gel [46], self-healing CeO_x coatings [47], ALD-deposited TiO₂ sealing layer on MAO [23], hydrophobic TiO₂/MoS₂ nanocomposites [24], and 3-glycidoxypropyltrimethoxysilane (GPTMS)/tetraethyl orthosilicate (TEOS)-derived organic-inorganic silica coatings [48]. Meanwhile, it achieves comparable outstanding corrosion resistance to the superhydrophobic 1H,1H,2H,2H-perfluorooctyltriethoxysilane (PFOTES)-modified poly(vinylidene fluoride) (PVDF)/SiO₂ coating [49], as reflected by the similarly ultralow i_corr and high impedance modulus. This work demonstrates that the proposed superhydrophobic multi-technology strategy combining MAO, sealing, and topcoat modification offers great significance for developing high-performance corrosion-resistant coatings on magnesium alloys.

4. Conclusions

In this work, a high-performance FAS/TiO₂-MAO composite coating was successfully fabricated on AZ31B magnesium alloy via a three-step synergistic strategy: MAO pretreatment, DUV-assisted sol–gel deposition of a TiO₂ sealing layer, and subsequent FAS superhydrophobic modification. The structural and functional characterizations confirmed that the DUV-assisted sol–gel technique effectively addressed the cracking and delamination issues of traditional sol–gel coatings, enabling the formation of a dense TiO₂ layer that partly sealed the porous MAO structure. The FAS modification further endowed the coating with superhydrophobicity (WCA = 160°) by introducing low-surface-energy fluorocarbon groups. Electrochemical tests demonstrated exceptional corrosion resistance of the FAS/TiO₂-MAO coating, with a E_corr of −1.34 V vs. SCE, a i_corr as low as 1.3 × 10⁻⁹ A/cm², and an impedance modulus at 0.01 Hz (|Z|_0.01 Hz) of ~2 × 10⁷ Ω·cm². Additionally, the coating exhibited excellent interfacial adhesion and retained stable superhydrophobicity and effective corrosion protection for at least 168 h in 3.5 wt.% NaCl solution. Its overall performance surpasses most and is comparable to the best of recently reported protective coatings on AZ31/AZ31B magnesium alloys.

The proposed “MAO + DUV-assisted sol–gel sealing + FAS superhydrophobic modification” synergistic strategy provides a feasible and efficient route for overcoming the intrinsic porosity of MAO coatings and the poor low-temperature compatibility of traditional sol–gel techniques on heat-sensitive magnesium alloys. This multi-technology integration not only achieves a nearly six-order-of-magnitude improvement in corrosion resistance compared to the bare AZ31B substrate but also balances mechanical durability, chemical stability, and interfacial adhesion. The findings offer valuable insights for the design and fabrication of high-performance protective coatings on magnesium alloys, promoting their practical applications in corrosive environments such as electronics, automotive, and biomedical fields. Future research will focus on enhancing the long-term stability of the superhydrophobic layer by optimizing the modifier thickness and chemical structure, and expanding the strategy to other heat-sensitive light alloys.