1. Introduction
The energy crisis seriously affects the development of modern society and reducing energy consumption is one of the effective ways to deal with the problem. Today, the concept of lightweight design, which aims to reduce energy demand, improve energy efficiency, and reduce carbon dioxide emissions, has been accepted in modern industries [
1,
2]. Magnesium is one of the lightweight metals with the most potential since the density of magnesium is only a quarter of that of iron and two-thirds of that of aluminum [
3]. Moreover, magnesium and its alloys have high specific strength and stiffness, good damping ability, excellent machinability and dimensional stability, superior electromagnetic shielding performance, good biocompatibility, etc. [
4,
5,
6], attracting considerable attention in aerospace, automotive, biomedical and other fields [
7,
8,
9]. However, their poor corrosion resistance in aqueous solutions still hinders their wider application since magnesium has a negative potential of −2.37 V [
10] and the natural oxide film or hydroxide film formed on magnesium has no protective effect [
4]. Even though magnesium alloys can be used as biodegradable metals due to their intriguing biodegradation and biosorption, their rapid biodegradation rate in a physiological environment is not acceptable [
11].
Usually, the corrosion resistance of magnesium alloys can be improved by three ways, i.e., alloying [
12], microstructure regulation [
13,
14], and surface modification [
15,
16]. Among these strategies, it is relatively economical to introduce corrosion-resistant coatings on magnesium alloys as a physical barrier, isolating the substrates from the aggressive environments. Chemical conversion coatings including chromate coatings [
17,
18], phosphate coatings [
11,
19], carbonate coatings [
20,
21], stannate coatings [
22,
23], permanganate coatings [
24] and phytic acid coatings [
25] have been commonly used as protective barriers. Among them, chromate coating is commonly used as an anti-corrosion coating for magnesium alloy because of its excellent corrosion resistance, self-healing performance, and relatively mature process. However, Cr
6+ is highly toxic and poses a significant threat to humans and the environment [
17,
26]. Therefore, its industrial applications have been restricted. At present, it is still urgent to find a more non-toxic and environmentally friendly coating for corrosion protection of magnesium alloys.
Zhang et al. [
27] proposed a concept involving a pre-corrosion treatment to produce a double-barrier structure on AZ80 Mg alloys to resist aqueous corrosion. They first used a one-step hydrothermal process in water to obtain a hierarchical coating composed of an inner compact hydroxide layer and top uniform Mg-Al LDH micro-sheets. After subsequent modification with fluorinated silanes, a Cassie-state superhydrophobic surface can be achieved with enhanced long-term corrosion resistance and potential self-cleaning ability. Obviously, it is an effective way to prepare a corrosion-protection- and surface-functionalization-integrated coating on the AZ80 magnesium alloy. Besides the self-cleaning ability, superhydrophobic surfaces have other functions, such as anti-contamination and anti/de-icing [
28]. Therefore, there is great potential in developing this coating process based on pre-corrosion.
Deepa B. Prabhu et al. [
29] prepared MgCO
3∙3H
2O coatings by immersing a Mg-4Zn alloy in a saturated NaHCO
3 solution at room temperature, which exhibited a biomineralization effect. Yao et al. [
30] immersed a magnesium–neodymium alloy in potassium dihydrogen phosphate solution at 50 °C for 4 h and obtained a layer of uniform MgHPO
4·3H
2O coating with a thickness of about 20 µm, which improves the corrosion resistance of the Mg alloy in a Cl
−-containing simulated concrete pore solution. Wang et al. [
31] immersed an AZ80 alloy in a simulated concrete pore solution made of saturated Ca(OH)
2 solution for 30 days and found that it has good corrosion resistance due to a spontaneously formed protective Mg(OH)
2 film. Therefore, this indicates that the pre-corrosion strategy is feasible to improve the corrosion resistance of magnesium alloys. However, the application of this concept is still in the initial stage, and it is necessary to continue to expand the scope of exploration and deeply understand the relevant mechanisms behind the immersion process, which can lead us to finally form a complete knowledge system for designing corrosion-protection- and surface-functionalization-integrated coatings in the future. In this study, we attempt to treat a magnesium–neodymium alloy with a mixed solution created by a saturated Ca(OH)
2 solution and 0.1 mol/L NaHCO
3 solution. In addition, the corrosion resistance of treated samples is evaluated in simulated physiological solutions, simulated seawater solutions, and simulated concrete pore solutions. Moreover, the associated mechanisms in this study are discussed.
2. Materials and Methods
As-cast Mg-3.3 Nd (wt.%) alloys were used as substrate materials in this study and cut into cubic samples with sizes of 10 mm × 10 mm × 5 mm. They were first mechanically ground with SiC sandpaper up to #2000, and then ultrasonically cleaned in ethanol for 3 min at room temperature. Equipment that can provide a constant temperature water bath was used for coating preparation. Then, 20 mL of saturated Ca(OH)2 solution and 40 mL of 0.1 mol/L NaHCO3 solution were prepared, and both were preheated at 80 °C for 5 min. The samples were first placed in the saturated Ca(OH)2 solution, and then 40 mL of 0.1 mol/L NaHCO3 solution was slowly poured. Next, the beaker containing the mixed solution and samples was heated at 80 °C for 6 h. After this process, the samples were removed from the solution and rinsed with ethanol and dried in air. After coating, surface modification with fluorinated silane was conducted as follows. First, a mixed solution was prepared by perfluorododecyltriethoxysilane (C16F17H19O3Si, PFDS), deionized water, and ethanol with a mass ratio of 1:4:45. Then, the uncoated and coated samples were soaked in the solution for 2 h, respectively. After being heated in an oven at 140 °C for 1 h, they were dried in air. In this paper, the as-received sample was denoted as uncoated and the coated sample was named as coated. After surface modification with PFDS, these two samples were denoted as PFDS-uncoated and PFDS-coated, respectively.
The coating morphology and chemical composition of the coating sample were observed by scanning electron microscopy (SEM, Hitachi Regulus 8100, Tokyo, Japan) equipped with X-ray energy-dispersive spectroscopy (EDS). Here, the samples were coated by sputtered gold before SEM observation. The phase composition of the substrate and coated samples were studied by X-ray diffraction (XRD, Rigaku SmartLab SE, Tokyo, Japan) with Cu Ka radiation at a scan rate of 2°/min. The chemical composition and chemical state of the coating were analyzed by X-ray photoelectron spectroscopy (XPS, Thermo Scientific K-Alpha Waltham, MA, USA) with an Al Kα X-ray radiation source. In addition, transmission electron microscopy (TEM, FEI Talos F200X G2, Waltham, MA, USA) combined with X-ray energy-dispersive spectroscopy (EDS) was applied to investigate the phase composition and chemical composition of the powder scratched from the coating. The powder was first sonicated in ethanol for 5 min to avoid agglomeration and then used for TEM observation.
In this study, three kinds of solutions were used to evaluate the corrosion resistance of the coated samples. A 0.9 wt.% NaCl solution heated at 37 °C was utilized to simulate the physiological environment. A simulated seawater solution used in this study was composed of 24.53 g/L NaCl, 4.09 g/L Na
2SO
4, 11.11 g/L MgCl
2, 1.16 g/L CaCl
2, 0.04 g/L SrCl
2, 0.70 g/L KCl, 0.20 g/L NaHCO
3 and 0.1 g/L KBr [
32]. In addition, a saturated Ca(OH)
2 solution was proposed to simulate a concrete pore solution and, in order to study the effect of Cl
−, sodium chloride (NaCl) was added into the saturated Ca(OH)
2 solution to increase the concentration of NaCl up to 2.5 wt.%. Electrochemical measurements involving potentiodynamic polarization and electrochemical impedance spectrum (EIS) measurements were conducted using an electrochemical workstation (Gamry, Reference 600+). A saturated calomel electrode (SCE) was used as the reference electrode, and the counter electrode was made of platinum sheets. The test sample with an exposed area of 10 mm × 10 mm acted as the working electrode. After being soaked in a 100 mL solution for 1 h, the potential was scanned from the cathodic region to the anodic region at a rate of 1 mV/s to obtain the polarization curves. The corrosion potential and corrosion current density were obtained using the cathode Tafel extrapolation method. The data of EIS measurement were recorded from 100 kHz to 0.1 Hz with a 5 mV sinusoidal perturbing signal at the open-circuit potential. In addition, the corrosion resistance was also estimated from hydrogen evolution measurements in the aforementioned solutions. The samples were sealed with acrylic-based resin with an exposed area of 10 mm × 10 mm and the method of hydrogen collection was carried out as referenced in the literature [
33]. The static water contact angle (CA) and sliding angle (SA) of the samples were measured with a 4 μL ultrapure water droplet on a contact angle measuring instrument (POWEREACH, JC2000D1,Shanghai, China).
3. Results and Discussion
Figure 1 shows SEM micrographs of the surface and cross-section morphologies of the coated sample. As shown in
Figure 1a, the surface of the coated sample is uniformly covered by plush micro-spheres.
Figure 1b shows the magnified view of the spheres, which reveals that the flower-like sphere consists of many micro-sheets. In addition, it is found that many cubic particles are scattered between micro-sheets.
Figure 1c,d shows the cross-sectional morphology of the coating with different magnifications. Here, the thickness of the coating can be calculated to be 53.62 ± 3.87 µm. Interestingly, the coating is composed of two layers: the top layer is related to the sphere-like array and the inner layer is relatively compact. It can be further observed that the inner layer is much thinner than the top layer.
Figure 2 shows the XRD results of the coated and uncoated samples. By comparison, it is determined that the obtained coating is mainly composed of Mg(OH)
2, CaCO
3, and Mg
5(OH)
2(CO
3)
4∙4H
2O and the peaks of Mg and Mg
12Nd originate from the substrate. EDS results were further utilized to analyze the chemical composition of the coating in this study. The results shown in
Figure 3 and
Figure 4 reveal that the coating mainly consists of Mg, Ca, C, and O. It can be found in
Figure 3 that the concentration of Ca in point 1 is much higher than that in point 2. As shown in
Figure 4, it is corroborated by EDS mapping that the coating consists of two layers. By combining the elemental distribution with the XRD results, it is speculated that the bottom layer marked with a dash line is possible to be Mg(OH)
2.
Figure 5 presents the results of the XPS test performed on the coated sample, which can further determine the chemical binding state of coating elements. As shown in
Figure 5a, the XPS survey spectrum reveals the existence of elements Mg, Ca, C, and O on the surface of the coating. This is also consistent with the aforementioned EDS result. High-resolution XPS spectra are deconvoluted and shown in
Figure 5b–e. The three components with binding energies of 284.8 eV, 286.56 eV, and 290.01 eV are obtained by deconvolution of C 1s in
Figure 5b, which belong to C-C, C-O, and C=O [
34,
35], respectively.
Figure 5c shows that the peak of the high-resolution spectrum of Mg 1s appears at 1304.51 eV, which can be attributed to Mg
2+ rather than metallic magnesium [
36]. The deconvolution of Ca 2p shown in
Figure 5d indicates that the actual peak stems from Ca
2+ and the Auger peak of Mg. Similarly, the deconvolution of O 1s shown in
Figure 5e reveals the presence of CO
32− and OH
−.
In order to further analyze the detailed structural information and composition of the coating, the powder scratched from the coating was characterized by TEM with high-angle annular dark field (HAADF) mode and selected area electron diffraction (SAED) in this study. Two forms of powders are shown in
Figure 6 and
Figure 7. As shown in
Figure 6a, EDS mappings of the flaky powder exhibit the elemental distribution of Mg, O, Ca, and C, which shows that the concentration of Ca is very low.
Figure 6b shows the SAED diffraction pattern measured in the yellow circle area of the bright field TEM image, which matches the crystal data of Mg(OH)
2 and Mg
5(OH)
2(CO
3)
4∙4H
2O.
Figure 7a shows the HAADF image and the corresponding EDS mappings of the blocky powder scratched from the coating. Here, the respective content of Ca, O, and C is much higher than that of Mg.
Figure 7b shows the SAED pattern collected at the circle area in the bright field TEM image, which matches well with the crystal data of CaCO
3.
Based on the aforementioned results, it can be concluded that a self-layered composite coating can be successfully achieved on the magnesium–neodymium alloy via a simple immersion process with a mixed solution of saturated Ca(OH)2 solution and 0.1 mol/L NaHCO3 solution. The top layer is mainly composed of CaCO3, Mg(OH)2 and Mg5(OH)2(CO3)4∙4H2O and the inner layer mainly consists of Mg(OH)2. The formation mechanism of the self-layered structure is speculated as follows: once the magnesium alloy contacts the solution, magnesium will react with the water according to the following formula: Mg + 2H2O → Mg(OH)2↓ + H2↑. Thus, the inner compact Mg(OH)2 layer is formed. Meanwhile, when the saturated Ca(OH)2 meets NaHCO3, several chemical reactions happen. First, HCO3− ionized from NaHCO3 reacts with Ca(OH)2 based on the following formula: 2HCO3− + Ca(OH)2 → CaCO3↓ + CO32− + 2H2O. Thus, CaCO3 precipitation is formed and CO32− is produced as well. Next, CO32− will react with free Mg2+ from the aqueous corrosion of magnesium substrate to produce Mg5(OH)2(CO3)4∙4H2O according to the following formula: 5Mg2+ + 4CO32− + 2OH− + 4H2O → Mg5(OH)2(CO3)4∙4H2O↓. Therefore, the top layer is formed by the Mg5(OH)2(CO3)4∙4H2O micro-sheets embedded with microscale CaCO3.
Figure 8a shows the polarization curves of the coated and uncoated samples in the 0.9 wt.% NaCl solution at 37 °C. The calculated
Ecorr and
Icorr are shown in
Figure 8b and
Table 1. By comparison, the corrosion potential of the coated sample shifts positively and the corrosion current density is decreased by two orders of magnitude, ranging from (1.330 ± 0.072) × 10
−4 A∙cm
−2 to (7.548 ± 1.490) × 10
−6 A∙cm
−2. It is clearly indicated that the coating improves the corrosion resistance of the magnesium alloy in the 0.9 wt.% NaCl solution.
Figure 9a presents the polarization curves of the coated and uncoated samples after being soaked in simulated seawater solution for 1 h.
Figure 9b and
Table 2 show the
Ecorr and
Icorr corresponding to the polarization curves. It can be seen that the corrosion potential of the coated sample has shifted positively compared to that of the uncoated sample. The corrosion current density of the sample after coating decreases from (5.310 ± 0.019) × 10
−4 A∙cm
−2 to (8.650 ± 0.615) × 10
−6 A∙cm
−2. This implies that the coating can improve the corrosion resistance of the magnesium alloy in a simulated seawater solution.
Figure 10 shows the polarization curves of the uncoated and coated samples after immersion in a simulated concrete pore solution for 1 h, as well as the calculated values of
Ecorr and
Icorr. The details of the calculated values are shown in
Table 3. The simulated concrete pore solution used in this test is a saturated Ca(OH)
2 solution and the effect of Cl
− on corrosion behavior is investigated by adding NaCl. It can be seen that in the simulated concrete pore solution without Cl
−, the corrosion potential of the coated sample is closed to that of the uncoated sample and the corrosion current densities of these two samples are of the same order of magnitude. However, in the simulated concrete pore solution containing Cl
−, the corrosion potential of the coated sample shifts to the positive direction significantly and the corrosion current density decreases from (3.147 ± 0.816) × 10
−5 A∙cm
−2 in the uncoated sample to (5.414 ± 0.664) × 10
−6 A∙cm
−2 in the coated one. Therefore, this indicates that the coating can improve the corrosion resistance of the magnesium–neodymium alloy in a simulated concrete pore solution with Cl
−.
The electrochemical impedance spectra of the uncoated and coated samples after being soaked in three kinds of solutions for 1 h are shown in
Figure 11,
Figure 12 and
Figure 13. The experimental data and fitted data are represented by scatter plots and solid lines, respectively. Here, the fitting curves are calculated based on equivalent circuit models referring to studies [
37,
38].
Figure 11 shows the electrochemical impedance spectra of the uncoated and coated samples after being soaked in a 0.9 wt.% NaCl solution for 1 h at 37 °C.
Figure 11a exhibits the Bode plots by impedance versus frequency and
Figure 11b shows the Bode plots by phase angle versus frequency,
Figure 11c gives the Nyquist plots. The equivalent circuit models for fitting are shown in
Figure 11d. Roughly speaking, the larger the |Z| value at the lowest investigated frequency, the stronger the corrosion resistance [
26,
34]. It can be seen from
Figure 11a that when the frequency is close to 0.1 Hz, the |Z| value of the coated sample is much larger than that of the uncoated sample. As shown in
Figure 11c, both samples contain capacitive loop at a higher frequency, but the uncoated sample has an extra inductive loop in the low-frequency region. By comparison, the capacitive loop size of the coated sample is larger than that of the uncoated one. It is generally assumed that capacitance loops at higher frequencies can be attributed to thin film effects and charge transfer processes and the occurrence of induction loops at lower frequencies may stem from mass transfer processes [
39,
40].
Figure 11d shows two equivalent circuit models for the uncoated and coated samples.
Rs(
CPEdlRct(
LRL)) is used to fit the EIS data of the uncoated sample and
Rs(
CPEf(
Rpore(
CPEdlRct))) is applied to fit the EIS data of the coated sample. Here, R
s denotes the solution resistance,
CPEf represents the capacitance of the thin film effect,
Rpore indicates the total resistance of pores in the thin film,
CPEdl represents the electric double-layer capacitance, and
Rct represents the charge transfer resistance. In addition, L and
RL are used to describe the low-frequency inductance and related resistance, respectively. The fitted data are shown in
Table 4. Usually, the polarization resistance R
p is reliable to evaluate corrosion resistance. Its reciprocal is directly proportional to the corrosion rate [
41], i.e., the larger R
p, the stronger the corrosion resistance of the material. According to the data in
Table 4,
and
are proposed to calculate the R
p value of the uncoated sample and the coated sample, respectively. The coated sample has a higher average R
p value of 4493.270 Ω·cm
2 than the uncoated one, with an average R
p value of 91.824 Ω·cm
2, indicating that it has better corrosion resistance.
Figure 12 shows the electrochemical impedance spectra of the uncoated and coated samples after being soaked in a simulated seawater solution for 1 h. The |Z|
f=0.1 Hz of the coated sample is much larger than the |Z|
f=0.1 Hz of the uncoated one, which implies that the coated sample has better corrosion resistance. As shown in
Figure 12c, the Nyquist plots of the two samples consist of a high-frequency capacitive loop and an intermediate-frequency capacitive loop. As mentioned above, the capacitance loop at high frequencies can be attributed to the thin film effect and charge transfer process, while the intermediate-frequency capacitance loop may be related to diffusion. Obviously, the coated sample has larger capacitance loops, indicating stronger corrosion resistance. The two electric circuits in
Figure 12d are used to fit the electrochemical impedance spectra of the uncoated and coated samples.
Rs is the solution resistance,
CPEf represents the capacitance of the thin film effect,
Rpore represents the total resistance of pores in the thin film,
CPEdl represents the electric double-layer capacitance, and
Rct is the charge transfer resistance. In addition,
CPEdiff and
Rdiff represent capacitance and resistance related to diffusion, respectively. Based on the fitted data in
Table 5,
and
are further used to calculate the R
p value of the uncoated sample and the coated sample, respectively. The average R
p value (9793.747 Ω·cm
2) of the coated sample is much larger than that (1434.000 Ω·cm
2) of the uncoated sample, suggesting that the coated sample has better corrosion resistance.
Figure 13 presents the electrochemical impedance spectra of the uncoated and coated samples after being soaked in simulated concrete pore solutions with and without Cl
− for 1 h. The low-frequency impedance of the coated sample in both simulated concrete pore solutions is higher than that of the uncoated sample, indicating that the coated sample is more corrosion-resistant than the uncoated one. In addition, the low-frequency impedance of the sample in simulated concrete pore solution with Cl
− is smaller than that in simulated concrete pore solution without Cl
−, suggesting that Cl
− can accelerate the corrosion of the magnesium–neodymium alloy. It can be seen in
Figure 13c that the Nyquist plots of the samples all contain a high-frequency capacitive loop, while the Nyquist plots of the uncoated samples in a simulated concrete pore solution with Cl
− contain an extra low-frequency inductive loop. As for the size of the capacitive loops, all samples in the simulated concrete pore solution without Cl
− have bigger sizes that the ones in the simulated concrete pore solution with Cl
−, and each coated sample has bigger sizes that the corresponding one.
Figure 13d shows the equivalent circuit models for fitting the EIS data and the meaning of each electronic component is referred to the aforementioned introduction.
Table 6 gives the results of the fitting and, based on the data, the polarization resistance is further calculated for evaluating corrosion resistance. Here,
and
are suggested to calculate the R
p value of the uncoated sample and the coated sample in the simulated concrete pore solution without Cl
−, respectively. The uncoated sample has an average R
p value of (1.431 ± 0.187) × 10
5 Ω·cm
2, which is similar to the coated sample’s average R
p value of (1.636 ± 0.270) × 10
5 Ω·cm
2.
and
are used to calculate the R
p value of the uncoated sample and the coated sample in the simulated concrete pore solution with Cl
−, respectively. As expected, the average R
p value (2498.923 Ω·cm
2) of the coated sample is larger than that (1111.527 Ω·cm
2) of the uncoated one.
The hydrogen evolution of the coated and uncoated samples in three different solutions is shown in
Figure 14. As shown in
Figure 14a,b, the hydrogen evolution volumes of the coated samples in both the 0.9 wt.% NaCl solution and the simulated seawater solution are significantly smaller than those of the uncoated samples, indicating an enhanced corrosion resistance of the coated samples in those solutions.
Figure 14c shows the hydrogen evolution of the investigated samples in the simulated concrete pore solution without Cl
−. During the 120 h investigated, the hydrogen evolution volumes of both kinds of samples were very small, indicating that the sample has long-term corrosion resistance in the simulated concrete pore solution without Cl
−.
Figure 14d shows the hydrogen evolution of the investigated samples in the simulated concrete pore solution with Cl
−. Comparing
Figure 14c with
Figure 14d, the hydrogen evolution volumes obtained in the simulated concrete pore solution containing Cl
− are much higher than those obtained in the simulated concrete pore solution without Cl
−. This indicates that Cl
− can aggravate the corrosion of magnesium alloys in the simulated concrete pore solution. Fortunately, the coating has a certain protective effect against the attack from Cl
−, which is validated by the results of the hydrogen evolution test and electrochemical evaluation.
The static water contact angle and sliding angle are applied to evaluate the wettability of the samples.
Figure 15 gives the static water contact angles of the uncoated and coated samples before and after surface fluorination. Before chemical surface modification, the static water contact angles of the uncoated and coated samples are 32.29 ± 7.33° and 0°, respectively, indicating the hydrophilic nature of their surfaces. After surface chemical modification, the static water contact angles of the uncoated and coated samples significantly increase, reaching 127.65 ± 1.00° and 151.74 ± 0.76°, respectively.
Figure 16 shows the captured view of the movement of water droplets on the uncoated and coated samples after surface fluorination. As for the uncoated sample, the water droplet on it does not start moving until the tilt angle becomes 20°. But for the coated sample, the droplet starts rolling when the tilt angle is only about 4°. Combined with the results of the static water contact angle, it is concluded that, after surface modification, the coated sample achieved a superhydrophobic effect, which is believed to endow the surface with water-repellent and self-cleaning abilities [
42].