Corrosion Behavior and Mechanism of Carbon Ion-Implanted Magnesium Alloy

: Carbon ion implantation was conducted on an AM60 magnesium alloy with ﬂuences between 1 × 10 16 and 6 × 10 16 ions / cm 2 and an energy of 35 keV. The microstructure and electrochemical properties of the samples were systematically characterized by X-ray photoelectron spectroscopy, X-ray di ﬀ raction, Raman scattering, scanning electron microscopy, transmission electron microscopy, and electrochemical methods. These studies reveal that a 250 nm-thick C-rich layer is formed on the surface and the Mg 2 C 3 phase embeds in the ion-implanted region. The crystal structure of the Mg 2 C 3 was constructed, and an electronic density map was calculated by density-functional theory calculation. The large peak in the density of states (DOS) shows two atomic p orbitals for Mg 2 C 3 . The main electron energy is concentrated between − 50 and − 40 eV, and the electron energy mainly comes from Mg ( p ) and Mg ( s ). The electrochemical experiments reveal that the E corr is − 1.35 V and I corr is 20.1 µ A / cm 2 for the sample implanted with the optimal ﬂuence of 6 × 10 16 ions / cm 2 . The sample from C ion implantation gives rise to better corrosion resistance.


Introduction
Magnesium alloys have some obvious advantages such as light weight and recyclability, and they are more and more widely used [1][2][3][4][5]. However, the corrosion resistance of magnesium alloys is not good enough [6,7]. To date, many methods have been carried out to improve the corrosion properties of magnesium alloys such as microarc oxidation [8,9], laser cladding [10,11], and ion implantation [12][13][14][15][16]. Ion implantation causes the substrate and coating have a good binding force compared to other methods. Many ions can be used in the ion implantation. For example, Ti [17,18], Al [19,20], Zr [21,22], and Cr [23,24] ion implantation has been conducted on Mg alloys to improve the corrosion properties. Compared with that of an untreated sample, the corrosion potential of a Ti/N-treated coupon increases by over 600 mV, the corrosion current density decreases by over two orders of magnitude, and the polarization resistance increases by 66.4 times [17]. The Al ion-implanted AZ31 magnesium alloy with an ion implantation dose of 6 × 10 16 ions/cm 2 achieved a high pitting breakdown potential of about −480 mV (saturated calomel electrode (SCE)) [20]. The Zr ion implantation sample shows a smaller I corr due to the formation of ZrO 2 in the implanted layer [21].
In addition to metals, some nonmetals are also used for ion implantation. A Mg-Nd-Zn-Zr alloy was modified with C 2 H 2 gas, and a diamond-like carbon film (DLC) was synthesized. Most of the surface region was protected in the aggressive NaCl solution [25]. The corrosion resistance of the AZ31 Mg alloy formed by N ion implantation was studied [26,27]. The results indicate that a Mg 3 N 2 film is formed. The corrosion potential shift was 100 mV. Besides the corrosion resistance, the cytotoxicity properties of Mg alloys modified by ion implantation have also been investigated [28][29][30].
The corrosion performance of the magnesium alloys can be improved to a certain extent by ion implantation. The C element was also implanted into pure magnesium, and the electrochemical properties were studied by Xu [31], but the microstructure of a C layer is not well known. As a very important element, C has many physical and chemical properties, such as good electrical conductivity and tribological properties. Many new carbon-based materials are being developed, for instance, nanoporous carbon [32] and carbon layers [33].
The application of AM60 materials is limited by poor corrosion properties. The effects of C ion implantation on the AM60 magnesium alloy and related mechanism have not been reported previously. In this work, the AM60 magnesium alloy was implanted with C with different fluences, and the structural and electrochemical properties were studied in detail. The atomic structure is discussed based on both experimental and theoretical results.

Experimental and Theoretical Calculations
Commercially available AM60 magnesium alloy samples were cut into 15 mm × 15 mm × 2 mm pieces and used as substrates in the experiment. Before ion implantation, the samples were finely ground with SiC paper of 400, 800, 1200, 1500, and 2000 grits, sequentially, followed by polishing with fine diamond paste (average grain size, 0.5 µm) to a final roughness of approximately 0.08 ± 0.02 µm, and then they were cleaned in acetone by ultrasonication. Carbon ion implantation was performed on the PI-80A plasma ion implanter in the Plasma Laboratory of City University of Hong Kong, Hong Kong, China. During implantation, the base pressure in the vacuum chamber was 2 × 10 −3 Pa and the acceleration voltage was 35 keV. The ion implantation fluences are shown in Table 1. X-ray photoelectron spectroscopy (XPS, PHI-5000 versaprobe, Kanagawa, Japan) was carried with Al Kα irradiation at a sputtering rate of about 10 nm/min in order to analyze the depth of the implanted elements. The Raman scattering spectra were acquired on the HR LabRAM using a 514.5 nm argon laser, and the structure was confirmed by X-ray diffraction on the D2 Phaser X-ray diffractometer (XRD) (BRUKER, Karlsruhe, Germany) with Cu Kα radiation. The cross-sectional microstructure of the samples was observed under a transmission electron microscope (TEM, JEM-2100, Tokyo, Japan). A TEM sample with a thickness of 3 mm was cut from the center of the ion implantation surface, with a plane size of 3 mm × 0.5 mm. Then, the 0.5 mm particles were further ground and thinned. The hole was thinned near the ion implantation surface, and the observation surface was perpendicular to the ion-implanted surface.
The corrosion tests were performed on an electrochemical workstation (CHI660E, Shanghai, China). A three-electrode configuration with a platinum sheet as the counter electrode, Mg sample as the working electrode (0.785 cm 2 exposed area), and saturated calomel electrode (SCE) as the reference electrode was adopted. The volume of the electrolyte was 1000 mL, the electrolyte was 3.5% NaCl solution, the temperature was 23 ± 1 • C, and the experiment was conducted three times to improve the statistics. The polarization curves were acquired at a scanning rate of 5.0 mV/s from −500 to 400 mV with respect to the open circuit potential. The samples soaked in 3.5% NaCl for 24 h were observed by scanning electron microscopy (SEM, ZEISS SIGMA500, Dresden, Germany).
In order to calculate the electronic structure, the crystal structure of Mg 2 C 3 was constructed by a first-principle calculation with the CASTEP module in the Materials Studio software [34]. The unit cell constant of Mg 2 C 3 and each atomic lattice were completely relaxed until the unit cell energy converged to a fixed value before the calculation. The result of the calculation ignores the rotation effect. Figure 1 displays the XPS depth profile of the ion-implanted AM60 alloy of Sample C3. A carbon-rich layer is formed after C ion implantation with a depth of about 250 nm. The high-resolution XPS spectra of C, O, Al, and Mg acquired after sputtering for different times are displayed in Figure 2. As the sputtering time increases, the C peak intensity decreases ( Figure 2a) and the Mg intensity increases gradually (Figure 2d). The Al intensity also increases (Figure 2c), and O is observed at the sputtering times of 0 and 15 min (Figure 2b). The C binding energy is 284.5 eV (Figure 2a), suggesting a sp 2 hybridization composed of three σ bonds and one π bond perpendicular to the σ bond. This observation is consistent with previous results, and it is generally believed that sp 2 hybridization is formed by low-energy ion implantation [35,36].

Results and Discussion
Coatings 2020, 10, x FOR PEER REVIEW 3 of 10 cell constant of Mg2C3 and each atomic lattice were completely relaxed until the unit cell energy converged to a fixed value before the calculation. The result of the calculation ignores the rotation effect. Figure 1 displays the XPS depth profile of the ion-implanted AM60 alloy of Sample C3. A carbonrich layer is formed after C ion implantation with a depth of about 250 nm. The high-resolution XPS spectra of C, O, Al, and Mg acquired after sputtering for different times are displayed in Figure 2. As the sputtering time increases, the C peak intensity decreases ( Figure 2a) and the Mg intensity increases gradually (Figure 2d). The Al intensity also increases (Figure 2c), and O is observed at the sputtering times of 0 and 15 min (Figure 2b). The C binding energy is 284.5 eV (Figure 2a), suggesting a sp 2 hybridization composed of three σ bonds and one π bond perpendicular to the σ bond. This observation is consistent with previous results, and it is generally believed that sp 2 hybridization is formed by low-energy ion implantation [35,36].    Figure 1 displays the XPS depth profile of the ion-implanted AM60 alloy of Sample C3. A carbonrich layer is formed after C ion implantation with a depth of about 250 nm. The high-resolution XPS spectra of C, O, Al, and Mg acquired after sputtering for different times are displayed in Figure 2. As the sputtering time increases, the C peak intensity decreases (Figure 2a) and the Mg intensity increases gradually (Figure 2d). The Al intensity also increases (Figure 2c), and O is observed at the sputtering times of 0 and 15 min (Figure 2b). The C binding energy is 284.5 eV (Figure 2a), suggesting a sp 2 hybridization composed of three σ bonds and one π bond perpendicular to the σ bond. This observation is consistent with previous results, and it is generally believed that sp 2 hybridization is formed by low-energy ion implantation [35,36].     Figure 3 shows the Raman scattering spectra. Before ion implantation, there is no C peak, but after C ion implantation, a G peak at 1542.14 cm −1 and D peak at 1331.66 cm −1 arising from amorphous carbon with sp 2 and sp 3 hybridization appear, being consistent with the XPS data. Figure 4 shows the XRD patterns, and α-Mg and Mg 17 Al 12 are present. A weak Mg 2 C 3 phase was formed after ion implantation, especially for high fluences such as for Samples C2 and C3. Figure 5 shows the TEM images of Sample C3. Figure 5a shows that the carbon-implanted sample is about 250 nm thick, consistent with the results of XPS. The white line shows the interface between the C layer and substrate. The black objects are the compound formed during ion implantation. A typical compound is revealed in the near surface as shown in Figure 5b. Figure 5c is a higher magnification of part of the white circle area in Figure 5b. The distance of the lattice space of the compound is 0.37 nm. According to the EDS results, the percent of Mg is 34.6%, the percent of C is 61.3%, and the percent of O is 4.1%. The compound is mainly composed of Mg and C. The diffraction patterns of the compound are displayed in Figure 5d. The results of TEM and XRD suggest that the compound is Mg 2 C 3 .

Results and Discussion
Coatings 2020, 10, x FOR PEER REVIEW 4 of 10 Figure 3 shows the Raman scattering spectra. Before ion implantation, there is no C peak, but after C ion implantation, a G peak at 1542.14 cm −1 and D peak at 1331.66 cm −1 arising from amorphous carbon with sp 2 and sp 3 hybridization appear, being consistent with the XPS data. Figure 4 shows the XRD patterns, and α-Mg and Mg17Al12 are present. A weak Mg2C3 phase was formed after ion implantation, especially for high fluences such as for Samples C2 and C3. Figure 5 shows the TEM images of Sample C3. Figure 5a shows that the carbon-implanted sample is about 250 nm thick, consistent with the results of XPS. The white line shows the interface between the C layer and substrate. The black objects are the compound formed during ion implantation. A typical compound is revealed in the near surface as shown in Figure 5b. Figure 5c is a higher magnification of part of the white circle area in Figure 5b. The distance of the lattice space of the compound is 0.37 nm. According to the EDS results, the percent of Mg is 34.6%, the percent of C is 61.3%, and the percent of O is 4.1%. The compound is mainly composed of Mg and C. The diffraction patterns of the compound are displayed in Figure 5d. The results of TEM and XRD suggest that the compound is Mg2C3.    Coatings 2020, 10, x FOR PEER REVIEW 4 of 10 Figure 3 shows the Raman scattering spectra. Before ion implantation, there is no C peak, but after C ion implantation, a G peak at 1542.14 cm −1 and D peak at 1331.66 cm −1 arising from amorphous carbon with sp 2 and sp 3 hybridization appear, being consistent with the XPS data. Figure 4 shows the XRD patterns, and α-Mg and Mg17Al12 are present. A weak Mg2C3 phase was formed after ion implantation, especially for high fluences such as for Samples C2 and C3. Figure 5 shows the TEM images of Sample C3. Figure 5a shows that the carbon-implanted sample is about 250 nm thick, consistent with the results of XPS. The white line shows the interface between the C layer and substrate. The black objects are the compound formed during ion implantation. A typical compound is revealed in the near surface as shown in Figure 5b. Figure 5c is a higher magnification of part of the white circle area in Figure 5b. The distance of the lattice space of the compound is 0.37 nm. According to the EDS results, the percent of Mg is 34.6%, the percent of C is 61.3%, and the percent of O is 4.1%. The compound is mainly composed of Mg and C. The diffraction patterns of the compound are displayed in Figure 5d. The results of TEM and XRD suggest that the compound is Mg2C3.    According to Find It soft, Mg 2 C 3 has a complex cubic structure. The Mg 2 C 3 space group is Pnnm (parameter code 58), and a = 6.4108 Å, b = 5.2786 Å, and c = 3.7283 Å, as shown in Table 2. Based on the data, a unit cell model and the electron density map are shown in Figure 6a,b, respectively. Figure 6c shows the density of states (DOS) of the Mg 2 C 3 unit cell, and the large peak shows two kinds of atomic p orbitals. The main electron energy is concentrated between −50 and −40 eV, and the electron energy mainly comes from Mg (p) and Mg (s). In addition, the valence electron of C (s) and C (p) makes a small contribution. The structure is primarily responsible for the improved corrosion resistance, to be discussed later.  According to Find It soft, Mg2C3 has a complex cubic structure. The Mg2C3 space group is Pnnm (parameter code 58), and a = 6.4108 Å, b = 5.2786 Å, and c = 3.7283 Å, as shown in Table 2. Based on the data, a unit cell model and the electron density map are shown in Figure 6a,b, respectively. Figure  6c shows the density of states (DOS) of the Mg2C3 unit cell, and the large peak shows two kinds of atomic p orbitals. The main electron energy is concentrated between −50 and −40 eV, and the electron energy mainly comes from Mg (p) and Mg (s). In addition, the valence electron of C (s) and C (p) makes a small contribution. The structure is primarily responsible for the improved corrosion resistance, to be discussed later.   Figure 7 presents the potentiodynamic polarization curves of the different samples in the 3.5% NaCl solution. The corrosion potential (Ecorr) and corrosion current density (Icorr) are determined by Tafel extrapolation. If a tangent is made to the cathodic polarization curve and the anodic polarization   Figure 7 presents the potentiodynamic polarization curves of the different samples in the 3.5% NaCl solution. The corrosion potential (E corr ) and corrosion current density (I corr ) are determined by Tafel extrapolation. If a tangent is made to the cathodic polarization curve and the anodic polarization curve, the ordinate of the intersection of the two lines and the E corr potential line is the I corr . Compared to that of the un-implanted AM60 control sample, the E corr of Sample C3 shifts from −1.48 to −1.35 V, indicative of better corrosion resistance in the NaCl solution. The I corr of sample C3 is 20.1 µA/cm 2 , which is much lower than that of the AM60 control sample (50.6 µA/cm 2 ). The more positive E corr and smaller I corr demonstrate that C ion implantation improves the corrosion resistance of the Mg alloy by forming the modified structure as described previously [37,38].  Figure 7 presents the potentiodynamic polarization curves of the different samples in the 3.5% NaCl solution. The corrosion potential (Ecorr) and corrosion current density (Icorr) are determined by Tafel extrapolation. If a tangent is made to the cathodic polarization curve and the anodic polarization curve, the ordinate of the intersection of the two lines and the Ecorr potential line is the Icorr. Compared to that of the un-implanted AM60 control sample, the Ecorr of Sample C3 shifts from −1.48 to −1.35 V, indicative of better corrosion resistance in the NaCl solution. The Icorr of sample C3 is 20.1 μA/cm 2 , which is much lower than that of the AM60 control sample (50.6 μA/cm 2 ). The more positive Ecorr and smaller Icorr demonstrate that C ion implantation improves the corrosion resistance of the Mg alloy by forming the modified structure as described previously [37,38].  The morphology of the different samples after immersion in 3.5% NaCl is shown in Figure 8. Lots of corrosion defects are observed from the surfaces of the control samples (Figure 8a-2). The SEM images were obtained from the marked part in the black box of the digital picture. It can be seen that the corrosion area of sample BM is larger than that of the other samples. The EDS performed on Point 1 as shown in Figure 8d-2 indicates that there are Mg, O, and C elements on the surface after C ion implantation. Figure 9 is the schematic presentation of the corrosion of the different samples. The inner layer is composed of AM60 alloy, and a thin layer of MgO is attached to the outer layer. The MgO layer on the surface is hydrated during exposure to an aqueous environment, which converts the MgO into a larger stable Mg(OH) 2 layer [24], and H 2 is produced in the corrosion process as shown in Figure 9a. The reaction may be expressed as the sum of the following reactions: Mg 2+ + 2(OH) − = Mg(OH) 2 (product formation) Mg The anodic reaction (Equation (1)) probably involves intermediate steps, which may produce the monovalent magnesium ion (Mg + ), which has a short lifetime [39]. The reduction process for hydrogen ions and the hydrogen overvoltage of the cathodic phase play an important role in the corrosion of magnesium [40]. After C ion implantation, there is a C layer on the surface. The sample from C ion implantation gives rise to better corrosion resistance. Mg 2 C 3 scattered in the C layer and C layer can decrease the corrosion and prevent corrosion from entering the substrate as shown in Figure 9b. The morphology of the different samples after immersion in 3.5% NaCl is shown in Figure 8. Lots of corrosion defects are observed from the surfaces of the control samples (Figure 8a-2). The SEM images were obtained from the marked part in the black box of the digital picture. It can be seen that the corrosion area of sample BM is larger than that of the other samples. The EDS performed on Point 1 as shown in Figure 8d-2 indicates that there are Mg, O, and C elements on the surface after C ion implantation.  Figure 9 is the schematic presentation of the corrosion of the different samples. The inner layer is composed of AM60 alloy, and a thin layer of MgO is attached to the outer layer. The MgO layer on the surface is hydrated during exposure to an aqueous environment, which converts the MgO into a larger stable Mg(OH)2 layer [24], and H2 is produced in the corrosion process as shown in Figure 9a. The reaction may be expressed as the sum of the following reactions: 2H2O + 2e = H2 + 2(OH) − (cathodic reaction) Mg 2+ + 2(OH) − = Mg(OH)2 (product formation) Mg + 2H2O = Mg(OH)2 + H2 (overall reaction) The anodic reaction (Equation (1)) probably involves intermediate steps, which may produce the monovalent magnesium ion (Mg + ), which has a short lifetime [39]. The reduction process for  . After C ion implantation, there is a C layer on the surface. The sample from C ion implantation gives rise to better corrosion resistance. Mg2C3 scattered in the C layer and C layer can decrease the corrosion and prevent corrosion from entering the substrate as shown in Figure 9b.

Conclusions
(1) A 250 nm-thick C-rich coating is formed on the surface of AM60 after C implantation with an ion implantation dose of 6 × 10 16 ions/cm 2 , and a Mg2C3 phase is observed from the ion-implanted region.

Conclusions
(1) A 250 nm-thick C-rich coating is formed on the surface of AM60 after C implantation with an ion implantation dose of 6 × 10 16 ions/cm 2 , and a Mg 2 C 3 phase is observed from the ion-implanted region. (2) The large peak in the density of states (DOS) shows the two atomic p orbitals of Mg 2 C 3 . The main electron energy is concentrated between −50 and −40 eV, and the electron energy mainly comes from Mg (p) and Mg (s). (3) The more positive E corr and smaller I corr demonstrate that C ion implantation gives rise to better corrosion resistance in AM60 magnesium alloys.