Formation of a Corrosion-Resistant and Anti-Icing Superhydrophobic Surface on Magnesium Alloy via a Single-Step Method

A superhydrophobic surface with chemical and long-time stability was fabricated on AZ31 magnesium alloy in a single-step via hydrothermal synthesis to improve corrosion resistance. The as-prepared surface repelled aqueous solutions with a static water contact angle of 156.7 ◦ and its superhydrophobicity was maintained for more than one year. The superhydrophobic surface showed anti-icing performance in a cold environment. The electrochemical measurement showed the superhydrophobic coating surface improved corrosion resistance for the Mg alloy substrate in 3.5 wt% NaCl solution. In this study, we sought higher efﬁciency and environmental friendliness to enhance the prospects for application of magnesium alloys.

Magnesium alloys are among the lightest and most easily machined metals. Owing to their low density, good electromagnetic screening ability, high strength/weight ratio, low cost, machinability, recyclability and other advantages, Mg alloys are known as the green engineering material of the 21st century. [1][2][3] They are used in many types of applications, such as in the automobile industry, biological materials, aerospace and aircraft fields, and electronic production. [4][5][6] The AZ31 Mg alloy is the most widely used industrial wrought Mg alloy due to its balance of properties and price; it is usually rolled into plate, squeezed into bars or processed by forging. 7 However, Mg alloys have shortcomings that limit their application, including low standard potential, high chemical activity, ease of oxidation, low abrasion performance, and susceptibility to corrosion in humid environments. 8 The development of corrosion-resistant Mg alloys will enable a much wider range of applications based on these materials. So far, many approaches have been used for conferring corrosion resistance on magnesium alloy; for example, the control of metallurgical factors by increasing the purity of the alloy, 9 mixing with rare earth elements, 10 or the use of a rapid solidification processing have been investigated. 11 Alternative approaches for facilitating corrosion-resistance performance include surface modification by laser processing, the deposition of a protective coating layer by a sol-gel route, anodic oxidation, electroplating and spraying oil paint. [12][13][14] Recently, many studies have shown that superhydrophobic surfaces are effective for conferring corrosion resistance. [15][16][17][18] A solid surface is superhydrophobic if the contact angle (CA) of a water droplet on the surface is larger than 150 • . [19][20][21][22] Recent studies have associated superhydrophobic surfaces with self-cleaning by a low sliding angle (SA). [23][24][25] For example, superhydrophobic surfaces of lotus leaves and strider legs found in nature inspire our use of this approach. 26 Owing to the poor wettability of the superhydrophobic surface, there is little contact area between the metal substrate and corrosive medium; therefore, superhydrophobic membranes covering the alloy can be used to achieve high corrosion resistance. [27][28][29] In addition to corrosion resistance, superhydrophobic surfaces are useful for other applications such as self-cleaning, oilwater separation, 30 drag reduction, 31 antifouling, 32 anti-icing 33 and UV resistance. 34 In other words, superhydrophobicity is an important property for materials in the fields of construction, communication, energy, biomedicine, and the environment. However, imperfect fabrication and poor stability of superhydrophobic surfaces limit their practical application. To date, many methods have been used to fabricate superhydrophobic surfaces, such as, template-based techniques, 35 z E-mail: qiliu@hrbeu.edu.cn; zhqw1888@sohu.com chemical vapor deposition, 36,37 electrospinning, 38 chemical or physical etching, 39 sol-gel methods, 40 layer-by-layer (LBL) deposition, 41 hydrothermal methods 42 and electrodeposition. 43 Several approaches have been used to fabricate superhydrophobic surfaces on Mg alloys. For example, electrodeposition was used in some recent studies. 3,24 Cui et al. 44 adopted the micro-arc oxidation (MAO) approach. Several groups used strong acid etching, electrochemical etching, and chemical deposition methods. Xie et al. 45 reported that formation of artificial superhydrophobic surfaces requires two steps: in the first step, a rough structure with a micro/nano pattern is constructed; and in the second step, the structure is modified by a low-surface-energy substance. Although the above technique exhibits many advantages, they require both the construction of a rough structure and immersion in fluoroalkylsilane (CF 3 (CF 2 ) 7 CH 2 CH 2 Si(OCH 3 ) 3 ) or long-chain complex. Therefore, these formation processes are not sufficiently efficient and environmentally friendly. Development of methods for simple fabrication of superhydrophobic surfaces has attracted the attention of researchers in recent years. Single-step fabrication of superhydrophobic surface will be more efficient and effective. This has motivated the effort to develop a greener and more efficient approach for formation of multi-functional superhydrophobic surfaces on Mg alloys.
We previously reported one-step fabrication of a superhydrophobic surface on Mg alloy using a water bath method. 32 In the current study, we used a single-step hydrothermal method to successfully fabricate a more stable superhydrophobic surface with anti-icing and corrosion-resistance properties on the AZ31 Mg alloy; this method exhibits increased efficiency and lower cost and is more environment friendly. The superhydrophobic coating with peony-like micro/nano rough structures makes a high static water contact of 156.7 • , producing AZ31 Mg alloy with corrosion-resistance and anti-icing potential. Furthermore, this coating displays chemical and long-term stability.

Experimental
Materials and sample preparation.-The AZ31 magnesium alloy was used as substrate, with a chemical composition that was mainly Mg, 2.98 wt% Al, 0.88 wt% Zn, 0.38 wt% Mn, 0.0135 wt% Si, 0.0027 wt% Fe, 0.002 wt% Ni, and 0.001 wt% Cu. The magnesium alloy was cut into 30 mm × 20 mm × 1.5 mm strips. To remove the oxide/hydroxide layer and impurities from the alloy surfaces, the substrate was first abraded with 600# to 2000# silicon carbide paper, then ultrasonically cleaned in anhydrous ethanol for 15 minutes, and finally washed with deionized water and dried under atmospheric conditions. One-step fabrication of superhydrophobic surfaces.-All reagents used in this experiment were of analytical grade and were used without further purification. First, nickel sulfate (0.0920 g) was dissolved in 35 mL of deionized water, and stearic acid (0.0996 g) was dissolved in 35 mL of absolute ethanol. The nickel sulfate aqueous solution was then added to the stearic acid ethanol solution. This addition led to the appearance of greenish floc in the mixed solution. The mixture was stirred for 1 min to obtain a homogeneous solution and was then transferred to a 100 mL Teflon-lined autoclave. Then, the single-step fabrication by the hydrothermal method was performed as follows: the dried Mg alloy plate was dipped into the 70 mL mixture of equal volumes of NiSO 4 aqueous solution and the stearic acid absolute ethanol solution in an autoclave. The reaction kettle was heated to 150 • C with a pressure of approximately 1.5 MPa and was maintained at that temperature and pressure for 8 h. The autoclave was then cooled to room temperature. Finally, the obtained surfaces were rinsed with deionized water and dried at 60 • C for 8 h.
Characterization.-The surface wettability was described by the static contact angle measured using a FTA200 drop shape analysis system at room temperature. Morphological analysis was carried out using a scanning electron microscope (SEM, JEOL JSM-6480A). The functional groups on the superhydrophobic film were identified by Fourier-transform infrared (FT-IR AVATAR 360) spectroscopic analysis. The chemical state of the surface was studied by X-ray photoelectron spectroscopy (XPS, PHI5700, Perkin-Elmer, USA). The Al Ka radiation (hv = 1486.6 eV) was used as the excitation source. The corrosion resistance of the Mg alloy with superhydrophobic surfaces was tested at room temperature (25 • C) using an electrochemical workstation (Zennium, IM6, Germany), equipped with a three-electrode system with an Ag/AgCl reference electrode consisting of a saturated KCl solution, a platinum mesh as the counter electrode and, finally, the sample as the working electrode. The performance of the working electrode (1 cm 2 ) was determined in 3.5 wt% NaCl aqueous solution. The potentiodynamic polarization curves were measured at a scanning rate of 1 mV s −1 . Electrochemical impedance spectroscopy (EIS) was conducted in the frequency range from 10 mHz to 100 kHz with a perturbation amplitude sinusoidal signal of 10 mV. All electrochemical tests were repeated at least three times to ensure reproducibility and reliability. The anti-icing experiment was carried out in a cold room at −15 • C and outdoors at about −15 • C (in January in Harbin, China).

Results and Discussion
Wettability of the superhydrophobic surface.-The wettability of the surface was assessed using water static contact angle (CA) measurements. The average CA values were determined by testing the same surface at five different locations. Fig. 1 shows the images of water droplets on different Mg alloy surfaces and the image of 3.5 wt% NaCl solution droplets on the as-prepared Mg alloy surface. The water contact angle of the just abraded Mg alloy is 60 • (Fig. 1a), much lower than that on the as-prepared Mg alloy surface. As shown in Fig. 1b, a high CA of 156.7 • is found on the as-prepared Mg alloy surface; the CA is greater than 150 • , demonstrating that we have fabricated a superhydrophobic surface on Mg alloy by a single-step method. Furthermore, from the photograph of water droplets on the asprepared Mg alloy surface (Fig. 1d), it can be found that the ball-like water droplets roll off easily from the superhydrophobic surface; in other words, there is little contact area between the Mg alloy and water. In addition to the high static CA observed for pure water, droplets with high CA values were observed for other solutions such as the 3.5 wt% NaCl solution on the superhydrophobic surface (Fig. 1c). We can infer that the as-prepared superhydrophobic surface offers corrosion resistance and chemical stability.
Moreover, the as-prepared superhydrophobic surface maintains its superhydrophobicity under acidic and basic conditions, showing a degree of chemical stability. To prove that the as-prepared Mg alloy with superhydrophobic surface is endowed with long-time stability, the coated alloy was stored in air for more than one year. The CA values for deionized water and other solutions with differing pH values on the superhydrophobic Mg alloy surfaces were then measured. Fig. 2 shows the relationship between the CA of different Mg alloy surfaces and the pH value of the liquid droplets. Examination of Fig. 2 shows that irrespective of the amount of time that the surface was exposed to the air, high CAs are obtained for the droplets with different pH values on the superhydrophobic surface. The CAs are approximately 150 o and show little variation with pH. There was no obvious change in the superhydrophobicity of the as-prepared Mg alloy when the surface was exposed to air for more than one year. The wettability of the untreated Mg alloy is very different from that of the as-prepared Mg alloy. The CAs for the untreated surface are less than 70 • and vary according to acidic and basic conditions; furthermore, the image in the inset of Fig. 2a show that bubbles are present in the droplet of pH 2; however, there is no bubble in the droplets of the as-prepared Mg alloy and the superhydrophobic AZ31 stored for one year (Fig. 2b).
In summary, a single-step method was used to fabricate a superhydrophobic surface on an Mg alloy; the superhydrophobicity of the surface was found to exhibit chemical and long-time stability.
Surface morphology.-The SEM images of the morphologies of the bare AZ31 surface and the as-prepared superhydrophobic AZ31 at different magnifications are presented in Fig. 3. Examination of Fig.  3a cannot show any outstanding morphology features except for some abrasion lines on the surface of the bare AZ31 alloy. However, the asprepared superhydrophobic surface is covered with micron peony-like rough structures (Fig. 3c). At higher magnification (Fig. 3d), these can be seen to consist of hexagonal flake-like microstructures, with the flake thickness of approximately sixty nanometres. The SEM images reveal that the surface coating created by the single-step hydrothermal method shows a hierarchical micro-nano-flower-like structure, engendering superhydrophobicity. 46 To determine the thickness of the as-prepared superhydrophobic film, a cross-sectional measurement of the hierarchical structure was performed. As shown in Fig. 3b, the superhydrophobic film is tightly bonded to the Mg alloy substrate and is nearly 50 μm thick.
Composition of the superhydrophobic surface.-The chemical composition of the as-prepared surfaces was studied by FT-IR and XPS. Fig. 4 shows the FT-IR spectrum of the superhydrophobic surface. In the high-frequency region, the absorption peaks at approximately 2851 cm −1 and 2919 cm −1 correspond to the asymmetric and symmetric stretching C-H vibrations, respectively. 47 In the lowfrequency region, the peak at approximately 1465 cm −1 is attributable to the C-H bending. These results show that long-chain aliphatic groups exist on the superhydrophobic surface. 3 The as-prepared sur- face shows additional adsorption peaks at approximately 1572 cm −1 and 1635 cm −1 , indicating that carboxylate is produced following the preparation reaction. It is known that the IR peak for the carboxyl group (-COO) from stearic acid is found at 1701 cm −1 ; 43,48 however this peak does not appear in Fig. 4, proving that no free carboxyl groups exist on the surface. The two adsorption peaks have been attributed to asymmetric and symmetric carboxyl group stretches. 49 Therefore, we conclude that CH 3 (CH 2 ) 16 COO − is present on the surface.
XPS analysis was carried out to further ascertain the type of the chemical bonding and the chemical composition on the as-prepared superhydrophobic surface (Fig. 5). Examination of the full survey spectrum (Fig. 5a) shows that C 1s, O 1s, Ni 2p, Ni 2s and Mg 1s peaks are present, indicating the elemental composition obtained on the superhydrophobic surface. We infer that stearic acid and nickel sulfate have reacted on the Mg alloy surface. Fig. 5b shows the two C 1s peaks of the as-prepared surface. The main peak at 284.49 eV represents hydrocarbon groups (C-H/C-C). The small peak at 288.2 eV is attributable to the O=C-O species. [50][51][52] These results imply that the C-H/C-C content is much greater than that for the O=C-O groups. We therefore infer the presence of long chain carboxylate reaction products on the surface that may contribute to hydrophobicity.   The high-resolution XPS spectrum of the Ni 2p (Fig. 5d) indicates the presence of four peaks of B.E. for nickel ion. Typical signals of Ni 2+ 2p 1/2 at approximately 873.5 eV and of Ni 2+ 2p 3/2 at approximately 855.8 eV are observed with respective satellite peaks, demonstrating the existence of Ni 2+ on the surface. [56][57][58] Based on the XPS and FT-IR analysis, we can deduce that the superhydrophobic surface consists mostly of Ni(CH 3 (CH 2 ) 16 COO) 2 with low surface energy.
Anti-icing potential of superhydrophobic surface.-In light of the unique non-wettability of the as-prepared superhydrophobic surface, we surmise that this surface may be promising for anti-icing. A series of simple tests were performed to prove this hypothesis. 59 First, the abraded AZ31 alloy and the superhydrophobic AZ31 alloy were placed horizontally on a watch-glass, with the abraded alloy on the left and the coated alloy on the right. The watch-glass was placed in the freezer for 3 minutes to make the simples cold from the beginning of the test. During the following observations, the samples remained in the freezer to keep the temperature at −15 • C. Then, 0.04 mL deionized water was poured onto the two surfaces marking the beginning of the test. To clearly show the freezing process, methyl orange was added to deionized water. The color of the liquid would change from dark orange to light orange with the freezing process. Inspection of Fig. 6a shows that the nature of the water on the two sam-ples did not change for approximately 5 seconds. At approximately 120 s, the sample on the left began freezing while the sample on the right did not freeze (Fig. 6b). Water on the abraded Mg alloy surface was completely frozen at 150 s, whereas the water on the as-prepared  superhydrophobic surface had not begun to freeze at this time (Fig. 6c). As shown in Fig. 6d, the two samples were both frozen at 600 s. In the second test, these two samples were placed at the edge of the watch-glass to make them incline. Methyl orange was added to 0.04 mL of deionized water that was then poured onto the samples. After 5 seconds, the water in both samples had not frozen; however, in the sample on the left, all of the water stayed on the untreated Mg alloy, whereas in the sample on the right all of the water had flowed off of superhydrophobic surface (Fig. 6e). At 300 s, the water on the Mg alloy surface of the left sample and flowed off of the superhydrophobic surface of the right sample (Fig. 6f) was frozen. Thus, the results presented in Fig. 6 show that the superhydrophobic surface delays the icing time, demonstrating that the superhydrophobic Mg alloy is more effective for anti-icing than the untreated Mg alloy.
To further explain the nature of the as-prepared superhydrophobic surface with anti-icing potential, the above two types of samples were placed horizontally onto the watch-glass with half volume of water. The abraded Mg alloy was placed on the left, and as-prepared superhydrophobic alloy was placed on the right (Fig. 7). We observe that the abraded magnesium alloy sinks in the water at room temperature, whereas the superhydrophobic magnesium alloy floats, as seen from the top-and side-view images in Figs. 7a and 7b, respectively. The samples were then placed in the open air at a temperature of −15 • C to −17 • C for 5 hours. As shown in Fig. 7c (top view) and Fig. 7d (side view), the water froze completely; the untreated alloy lay under the ice, and the superhydrophobic alloy lay on top of the ice. Moreover, the superhydrophobic alloy was removed easily using tweezers with no damage (Fig. 7e). To demonstrate this effect more clearly, a single superhydrophobic sample was placed on the watch-glass with half volume of water in the open air at −15 • C for 5 hours. As shown in Fig. 7f, the superhydrophobic Mg alloy froze on the ice and the ice under the Mg alloy sample was higher than other locations. Based on these tests, we infer that the as-prepared superhydrophobic surface decreases icing to a certain extent because of its low surface energy and superhydrophobicity. We also speculate that the superhydrophobic surface is conducive to drag reduction. Corrosion resistance performance of the superhydrophobic surface.-The corrosion resistance of the superhydrophobic surface was investigated using an electrochemical workstation from potentiodynamic polarization and EIS. Fig. 8 shows potentiodynamic polarization curves of superhydrophobic AZ31 after immersion in 3.5 wt% NaCl aqueous solution for 3 h, 6 h and 9 h, and of untreated AZ31 immersed for 3 h. Intersections of extrapolation of linear Tafel segments of the cathodic curve from the Tafel region and the open circuit potential lines were obtained based on the Tafel extrapolation method, [60][61][62] to determine the values of corrosion potential (E corr ) (vs. Ag/AgCl), corrosion current density (i corr ) and cathodic Tafel slope (b c ). The relevant electrochemical parameters derived from the polarization curves in Fig. 8 for these two samples immersed for different times are summarized in Table I. According to Fig. 8 and Table I, the i corr of the superhydrophobic surface immersed for 3 h (3.60 × 10 −6 A cm −2 ) decreased by approximately more than one order of magnitude compared to that of the untreated AZ31 immersed for the same time (1.45 × 10 −4 A cm −2 ). The i corr value of the superhydrophobic surface formed on AZ31 after immersion in 3.5 wt% NaCl aqueous solution for 6 h was only slightly shifted to approximately 5.50 × 10 −6 A cm −2 . After immersion for 9 h, the i corr of the surface was approximately 8.62 × 10 −6 A cm −2 , a little higher than that of as-prepared surface immersed for 3 h but still much lower than that of the untreated AZ31. Generally, lower corrosion current densities correspond to lower corrosion rates and improved corrosion resistance. 63,64 The above results therefore indicate that the superhydrophobic surface endows magnesium alloy with relatively high corrosion resistance. The enhanced corrosion resistance may be due to the air layer between the superhydrophobic surface and the NaCl aqueous solution. Moreover, the thickness of the air layer between the superhydrophobic surface and the NaCl aqueous solution remains relatively stable in 9 hours. To prove the existence of the air layer, Photographs of abraded Mg alloy (Fig. 9a) and superhydrophobic Mg alloy (Fig. 9b) immersed in 3.5 wt% NaCl aqueous solution captured with oblique angle were taken. It is obvious that the  superhydrophobic surface immersed in the solution is much brighter than it exposed in the air, while the abraded Mg alloy looks the same no matter whether immersed in the solution or exposed in the air. This phenomenon indicates that the air layer can still be trapped in the micro/nano hierarchical structure according to total reflection theory in physics. 65 Furthermore, the air layer could make larger electrolyte resistance and less contact area between the NaCl solution and the superhydrophobic surface to reduce corrosion rate.
The corrosion behavior of the bare AZ31 and the as-prepared superhydrophobic surface was further investigated using EIS measurements, with the results shown in Figs. 10 and 11. Fig. 10 shows the Nyquist plots of superhydrophobic and untreated AZ31 Mg alloy surface immersed in 3.5 wt% NaCl aqueous solution for 3 h. The AZ31 coated with superhydrophobic surface immersed for 3 h is characterized by two capacitive loops in the high and low frequency ranges (Fig. 10). The enlarged Nyquist plot of the bare AZ31 Mg alloy in Fig. 10 contains a capacitive loop in the high frequency range and an inductive loop in the low frequency range. The inductive loop is attributed to the dissolution of Mg and is indicative of pitting corrosion of the substrate. 44 It is important to note that no low-frequency inductive loop appeared in the experiment on the superhydrophobic surface immersed for 3 h. This means that the superhydrophobic surface can effectively protect the Mg alloy from corrosion. Previous work has found that the diameter of the capacitive loop in the Nyquist plots is a measure of the polarization resistance of the work electrode. 24,66 A larger capacitive loop means a lower corrosion rate. 44 Fig. 10 shows that the impedance value of the bare AZ31 is much lower than that of the as-prepared superhydrophobic surface. This indicates that the superhydrophobic surface can decrease the corrosion rate.   11 shows the EIS results for superhydrophobic surface immersed in 3.5 wt% NaCl aqueous solution for different times. The diameters of the Nyquist plots decrease with increasing immersion time (Fig. 11a). These results are due to the decreased thickness of the air layer between the superhydrophobic surface and the NaCl solution. During the first 9 hours of immersion, the thickness of the air layer remained relatively stable, in good agreement with the polarization curves tests results. After immersed for 24 h, most of the air layer disappeared and the NaCl solution contacted with the organic coating directly. Thus, there is obvious decrease for the diameters of the Nyquist plots. As shown in Fig. 11b, the Nyquist plots for the superhydrophobic samples immersed for more than 24 h contain a high frequency capacitive loop, a medium frequency capacitive loop and a low frequency inductive loop, indicating pitting corrosion occurred on the surface. 44 However, the diameter of the Nyquist plots for the superhydrophobic surface immersed for 120 h is still larger than that for the bare AZ31 surface immersed for 3 h. As shown in Fig.  12, there is obvious difference between the surfaces of two samples immersed for 3 h, 24 h and 120 h respectively. After immersion in 3.5 wt% NaCl aqueous solution for 3 h, the bare AZ31 sample has been corroded with black corrosion mark (Fig. 12a), while there is no obvious corrosion mark on the superhydrophobic surface (Fig. 12d). After immersion for 24 h, there is much corrosion product on the bare AZ31 surface (Fig. 12b) and little pitting corrosion on the superhydrophobic sample (Fig. 12e). Additionally, corrosion perforation is observed on the bare AZ31 sample and little pitting corrosion on the  superhydrophobic sample after immersion for 120 h (Figs. 12c and 12f). These results confirm the enhanced corrosion resistance due to the superhydrophobic surface.
Based on the results of the characterizations and the tests, the corrosion resistance ability of the superhydrophobic surface can be explained as follows: first, the micro/nano flower-like structures roughen the surface to allow the passage of air, keeping the surface relatively dry and preventing the corrosive media from reaching the surface; 3 secondly, the surface energy of the superhydrophobic surface is much lower because of the chemical reaction giving rise to a smaller contact area between the corrosive media and the surface. These features facilitate the resistance to corrosion of the as-prepared superhydrophobic coating of the AZ31 alloy.

Conclusions
In summary, this study has developed a simple, environmentfriendly, low-cost, and single-step method for the fabrication of superhydrophobic surfaces with peony-like micro/nano rough structures on the AZ31 Mg alloy. The CA of the as-prepared surface is as high as 156.7 • and the superhydrophobic surface exhibits chemical and long-time stability. According to the results of XPS and FT-IR analyses, Ni(CH 3 (CH 2 ) 16 COO) 2 forms the basis of the superhydrophobic surface. When immersed in 3.5 wt% NaCl aqueous solution, the as-prepared membrane exhibits an improved anti-icing function and higher corrosion resistance with lower corrosion current density compared to the untreated magnesium alloys. This single-step approach provides surface protection on magnesium alloy, and is promising for further development.