Corrosion Behavior of Extruded AM60-AlN Metal Matrix Nanocomposite and AM60 Alloy Exposed to Simulated Acid Rain Environment

The present work compared the initial stages of corrosion process development on the AM60-AlN metal matrix nanocomposite surface and on AM60, during their exposure for 30 days to simulated acid rain solution (SAR). The AlN nanoparticles were observed as “attached” to those of Mn-rich AlMn intermetallic particles, forming clusters. The introduction of 1.0 wt.% AlN (≈ 80 nm) in the AM60 alloy carried a slight grain refinement and favored the formation of a denser and more protective corrosion layer, suggested by the electrochemical impedance spectroscopy (EIS) values of higher charge transfer resistance (R2) and capacitance, characteristic of the double layer in the presence of corrosion products, and also suggested by Rn (EN) values, compared to those of the AM60 alloy. Thus, the concentration of the released Mg-ions from the composite surface was lower. Due to the increase in time of the SAR solution pH, Al de-alloying may occur, as well as Al(OH)3 formation, as confirmed by XPS analysis. Due to the presence of Cl-ions in SAR solution, localized corrosion was observed, suggested as fractional Gaussian noise of a stationary and persistent process in time, according to the PSD of the corrosion current fluctuations (EN).


Introduction
Magnesium is the lightest structural metal (1.74 g cm −3 ) [1] and its rapidly cooled alloys can reach a strength/weight ratio of 490 kN m kg −1 [2]. When the weight reduction is an important factor, the excellent physical and mechanical properties of the Mg-alloys make them ideal materials for automotive and aerospace applications [3,4]. In recent years, air pollution has become a serious problem, causing an increase in atmosphere aggressivity, and metal structures suffer from accelerated environmental damage [5]. The use of fuels in the transport sector has contributed to pollutants, such as SO 2 and CO 2 , causing a decrease in the atmospheric pH and, as a consequence, acid rain known worldwide. Therefore, a necessary trend has been to reduce the fuel consumption, which would lower the environmental impact [6]. One way to help resolve this problem is linked to the development of new lightweight materials for automobile structures [7]. Promising materials are the Mg-alloys and its nanocomposites.
Mg-Al alloys are divided into AZ-series with the addition of Zn, and AM-series with the addition of Mn. Aluminum is considered one of the most important alloying elements, because of its tendency to form a stable passive oxide layer on its surface in neutral environments. It has been reported that up to 4% Al may provide a decrease in the Mg-Al corrosion rate [8,9]. On the other hand, the addition of Mn also favors such as hardness and wear resistance in a similar way as the introduction of B 4 C and SiC particles [56]. It has been reported that the corrosion rate of Mg-matrix may decrease in the presence of AlN-reinforced particles, because of the high AlN electrical resistivity (1 × 10 12 Ω cm) [57]. Moreover, the AlN particles are considered an excellent choice for grain refinement of magnesium alloys [58,59]. An AM60-based matrix nanocomposite (MMNC), containing 1 wt.% of AlN particles (≈80 nm of average size), has been successfully produced using an ultrasound-assisted indirect chill casting process [60]. The results revealed that the AlN nanoparticles caused a grain refinement and improvement in various mechanical properties (yield strength, ultimate tensile strength, and ductility). Elektron21-AlN nanocomposites have also been produced by ultrasound-assisted casting with different concentrations of AlN (1% and 2%) [61]. A reduction in grain size (to 72.8 µm) was reported for Elektron21 with 2% of AlN, as well as increases in thermal and electrical conductivity.
The aim of the present work was to characterize the initial stages of the corrosion process of the AM60-AlN metal matrix composite and to compare them with those of the AM60 alloy, during their 30 days of exposure to simulated acid rain solution (SAR). A combination of immersion test and surface characterization methods were employed to evaluate the mass loss (corrosion rate), attack on the surface, and the formed product layers. The change in time of SAR pH and concentration of the released Mg-ions into the electrolyte was also monitored, in order to correlate with the corrosion rate and values of the corrosion open circuit potential (OCP). The current fluctuations, considered as electrochemical noise (EN), were analyzed in the time and frequency domain to identify the corrosion process. Electrochemical impedance spectroscopy (EIS) measurements were carried out to compare the performance of the studied alloys in the simulated acid rain solution.

Materials and SAR Model Solution
The nominal composition (wt.%) of the extruded AM60, according to the producer, (Magontec, Bottrop, Germany) is: 6.0 Al; 0.2-0.4 Mn; and the balance Mg. For the preparation of the nanocomposite, 15 kg of the alloy was melted and 150 g of AlN nanoparticles with an average diameter of ≈80 nm was added under stirring; this corresponds to an AlN concentration of 1 wt.%. The nanoparticles were prepared at Tomsk State University in Tomsk, Russia, and their properties and manufacturing process have been previously described [60,62,63]. The mixture of melt and nanoparticles were treated with high-shear dispersion-treatment (HSDT) for 10 min and, for this purpose, a rotor-stator device (Zyomax Ltd., London, UK) was used at 1000 rpm, exposing the melt permanently to a protective gas mixture of Ar + 1% SF 6 . The composite melt was poured into cylindrical molds, suspended in a three-zone resistance furnace, and then slowly lowered into a water bath, initiating the solidification from the bottom of the mold. This means that there is always residual melt on top of the solidified phase, with the result that shrinkage is fed during solidification and the microstructure of the cylinder is largely free of pores. An identical process was carried out for the comparative sample of AM60 without nanoparticles. The cylinders were machined to 49 mm and indirectly extruded at 300 • C in a 2.5 MN extrusion press (Müller Engineering, Todtenweis, Germany) to 10 mm diameter round profiles, and they were taken for the investigation. The extrusion speed was 2.2 mm/s and the extrusion ratio was 25:1.
The tested materials were cut into disc-shapes of 1 mm in thickness; some of them were used for the immersion test while others were used as electrodes for electrochemical experiments. All specimens were abraded with SiC paper to 2000 grit, using ethanol as a lubricant, and they were then sonicated in ethanol for 5 min and dried at room temperature.

Immersion Test and Analysis of Surface
Samples of the AM60 alloy and AM60-AlN nanocomposite were immersed in triplicate in 20 mL of SAR solution and in independent containers for 1, 7, 10, 15, and 30 days following ASTMG31-12a [65]. At the end of each period of exposure, the samples were withdrawn, rinsed with deionized water, and dried in air at room temperature. The waste solutions were stored in order to measure their pH and the concentration of the released Mg-ions by means of photometry (HI83200, Hanna Instruments, Woonsocket, RI, USA). With the purpose of evaluating the mass loss and damage on the surface samples after each exposure period, the layer of corrosion products was chemically removed in the solution of 200 g L −1 of CrO 3 , 10 g L −1 of AgNO 3 , and 20 g L −1 of Ba(NO 3 ) 2 , according to ASTM G1-03 [66]. To calculate the mass loss (∆m = W), the initial (w 0 ) and final mass (w 1 ) of the samples were measured by an analytical balance (VE−204, Velab, CDMX, México). The annual corrosion rate (CR in mm yr −1 ) was calculated through Equation (2), specified by ASTM-G1-03 [66]: where K is a constant (8.76 × 10 4 ), W is the mass loss (g), A is the surface area (cm 2 ), T is the time of exposure (h), and D is the metal density (g cm −3 ). The possible crystalline phases were analyzed by X-ray diffraction (Siemens, D−500, Munich, Germany). The changes in morphology and composition of the formed layers on the surface after exposure to SAR were characterized by scanning electron microscopy (SEM-EDS, XL−30 ESEM-JEOL JSM−7600F, JEOL Ltd., Tokyo, Japan) and XPS spectroscopy (K-Alpha Surface Analyzer, Thermo Scientific, Waltham, MA, USA). The binding energies of all XPS spectra were normalized to the C1s peak at 284.8 eV.

Electrochemical Characterization
The study of the electrochemical corrosion behavior was carried out through several electrochemical methods, such as registration of the corrosion potential and current fluctuations (at open circuit potential, OCP) and electrochemical impedance spectroscopy (EIS). An interface−1000E potentiostat/galvanostat/ZRA (Gamry Instruments, Philadelphia, PA, USA) was used with a typical three-electrode cell configuration, inside a Faraday cage: the working electrodes (0.78 cm 2 ) of AM60-AlN and AM60, the Pt mesh (Alfa Aesar, Ward Hill, MA, USA) as an auxiliary electrode, and a saturated calomel electrode (SCE) as a reference electrode (Gamry Instruments, Philadelphia, PA, USA). All experiments were carried out at room temperature of 21 • C. The EIS spectra were acquired for 1 h and 1, 7, 10, and 15 days of immersion in SAR solution, employing a perturbation amplitude of ±10 mV vs. OCP (stabilized after 1 h), and a frequency interval from 100 kHz to 10 mHz. The obtained EIS data were analyzed with Gamry Echem Analyst software (Gamry Instruments, Philadelphia, PA, USA).
The fluctuations in the current (at OCP, recorded over 3 h) were measured employing two identical working electrodes and the SCE reference electrode, connected to the potentiostat in zero resistance ammeter mode (ZRA), according to ASTM G199-09 [67]. The sampling interval was of 0.1 s and at different times: initial, 1, 7, 10, 15, and 30 days. The spontaneous current fluctuations associated with corrosion phenomena were considered as electrochemical noise (EN) and examined with Electrochemical Signal Analyzer ® (Gamry Instruments, V.7.0.1, Philadelphia, PA, USA), transforming the data to the frequency domain by fast Fourier transform (FFT), in order to analyze their power spectral density (PSD) [68,69]. The values of the β-slope, as an exponent of PSD plots, were used to characterize the dynamics of the corrosion process.

Microstructure of AM60-AlN and AM60
The optical images ( Figure 1) compare the grain size of the AM60-AlN nanocomposite (Figure 1a,b) and AM60 alloy (Figure 1c,d), presented in the longitudinal and cross-sections. The introduction of 1.0 wt.% AlN nanoparticles in the AM60 alloy carried a slight grain re-finement (Table 1) with a size reduction from 3.3 ± 1.1 to 2.9 ± 0.1 µm (longitudinal section) and from 3.3 ± 0.2 to 2.9 ± 0.4 µm (cross-section). This reduction in grain size may have a beneficial effect on ductility [70][71][72]. On the other hand, the Vickers hardness (Table 1) of AM60-AlN resulted in a slight improvement (7%), probably due to the grain refinement, as well the reduction in dislocation movements in the presence of AlN nanoparticles [61].
to characterize the dynamics of the corrosion process.

Microstructure of AM60-AlN and AM60
The optical images (Figure 1) compare the grain size of the AM60-AlN nanocomposite (Figure 1a,b) and AM60 alloy (Figure 1c,d), presented in the longitudinal and crosssections. The introduction of 1.0 wt.% AlN nanoparticles in the AM60 alloy carried a slight grain refinement (Table 1) with a size reduction from 3.3 ± 1.1 to 2.9 ± 0.1 µm (longitudinal section) and from 3.3 ± 0.2 to 2.9 ± 0.4 µm (cross-section). This reduction in grain size may have a beneficial effect on ductility [70][71][72]. On the other hand, the Vickers hardness (Table  1) of AM60-AlN resulted in a slight improvement (7%), probably due to the grain refinement, as well the reduction in dislocation movements in the presence of AlN nanoparticles [61].        (Table 2) suggests may be considered as the intermetallic Mn-rich particles of AlMn, in the presence of C and O elements, the contents of which were attributed   (Table 2) suggests may be considered as the intermetallic Mnrich particles of AlMn, in the presence of C and O elements, the contents of which were attributed to the reported effective nucleating particles of Al-C-O (Al 2 CO) [73]. The AlN nanoparticles (labeled as 3, Figure 3a) were observed as "attached" to those of AlMn, forming clusters. Although the content of Mg (64.4 wt.%, Table 2) was slightly higher than that of 42-58 wt.% given in the binary phase diagram for Al-Mg [74], the areas labeled as 4 were assigned to the β-Mg 17 Al 12 secondary phase, present in both alloys.   (Table 2) suggests may be considered as the intermetallic Mn-rich particles of AlMn, in the presence of C and O elements, the contents of which were attributed to the reported effective nucleating particles of Al-C-O (Al2CO) [73]. The AlN nanoparticles (labeled as 3, Figure 3a) were observed as "attached" to those of AlMn, forming clusters. Although the content of Mg (64.4 wt.%, Table 2) was slightly higher than that of 42-58 wt.% given in the binary phase diagram for Al-Mg [74], the areas labeled as 4 were assigned to the β-Mg17Al12 secondary phase, present in both alloys.

Test Solution Monitoring and Mass Loss Measurement
It has been suggested that the corrosion behavior of aluminum nitride (AlN) in a relatively neutral solution (pH = 7 of deionized water) may be described by the following reactions [75]: Figure 4 presents the change in time of SAR pH ( Figure 4a) and concentration of Mg 2+ ions release (Figure 4b) into the SAR model solution, during the immersion of AM60-AlN nanocomposite and AM60 alloy samples. It can be seen (Figure 4a) that, since the first day, there was an abrupt change in pH initial value of 2.73, reaching pH ≈ 10.0 for AM60 and less alkaline values for AM60-AlN (pH ≈ 9.40). These facts indicated the beginning of the corrosion process, related to the H 2 evolution in the acid SAR solution and the excess of OHions (Equation (1)). As the pH solution reached alkaline values (Figure 4a), the higher content of OHions may favor the formation of Al(OH) 3 (Equation (5)). [76]. The concentration of the released Mg-ions (Equation (7)) reached at 30 days was 5.4 ± 0.34 g cm −2 for AM60, 24% higher than the 4.1 ± 0.25 g cm −2 for AM60-AlN (Figure 4b).
MgCl → Mg + 2Cl (7) The change in mass (∆m) and annual corrosion rates (Equation (2)) are presented in Figure 5. After the first day, the mass loss of AM60 was of 2.20 ± 0.19 mg cm −2 , twice greater than that of the AM60-AlN composite. At 7 days, the ∆m of both samples showed an increasing trend of up to 30 days, presenting 9.5 ± 0.21 mg cm −2 for AM60 and a lower value   (6)) into highly soluble MgCl 2 (56.0 g mL −1 ) [76]. The concentration of the released Mg-ions (Equation (7)) reached at 30 days was 5.4 ± 0.34 g cm −2 for AM60, 24% higher than the 4.1 ± 0.25 g cm −2 for AM60-AlN ( Figure 4b). Mg The change in mass (∆m) and annual corrosion rates (Equation (2)) are presented in Figure 5. After the first day, the mass loss of AM60 was of 2.20 ± 0.19 mg cm −2 , twice greater than that of the AM60-AlN composite. At 7 days, the ∆m of both samples showed an increasing trend of up to 30 days, presenting 9.5 ± 0.21 mg cm −2 for AM60 and a lower value of 8.1 ± 0.19 mg cm −2 for AM60-AlN (Figure 5a). The calculated annual corrosion rates, based on ∆m values corresponding to 30 days, were 0.65 ± 0.01 and 0.55 ± 0.01 mm yr −1 , for AM60 and AM60-AlN, respectively (Figure 5b). These values were influenced by the pH change over time and Mg 2+ ions release ( Figure 4).

SEM-EDS Analysis
After exposure to acid rain (SAR) solution for up to 30 days, the layers grown on the AM60-AlN nanocomposite and AM60 alloy surfaces showed cracked layers of corrosion products (Figure 6a,b), which may act as channels connecting the Mg-matrix with the SAR solution, and also improve the diffusion of the released Mg 2+ ions and H2 bubble evolution. The H2 bubble pressure within the pores of those layers may favor their fracture and eventual local detachment of the corrosion layer.

SEM-EDS Analysis
After exposure to acid rain (SAR) solution for up to 30 days, the layers grown on the AM60-AlN nanocomposite and AM60 alloy surfaces showed cracked layers of corrosion products (Figure 6a,b), which may act as channels connecting the Mg-matrix with the SAR solution, and also improve the diffusion of the released Mg 2+ ions and H 2 bubble evolution. The H 2 bubble pressure within the pores of those layers may favor their fracture and eventual local detachment of the corrosion layer.
After exposure to acid rain (SAR) solution for up to 30 days, the layers grown on the AM60-AlN nanocomposite and AM60 alloy surfaces showed cracked layers of corrosion products (Figure 6a,b), which may act as channels connecting the Mg-matrix with the SAR solution, and also improve the diffusion of the released Mg 2+ ions and H2 bubble evolution. The H2 bubble pressure within the pores of those layers may favor their fracture and eventual local detachment of the corrosion layer. (The SEM images of (a,c,e) were performed by the Philips microscope, while those of (b,d,f) were performed by the JEOL microscope.) EDS analysis (Table 3) of the marked areas of interest confirmed the presence of the Mn-rich phase of Al-Mn intermetallic particles (labeled as 1) on the corroded surfaces (Figure 6a,b) even after the removal of the formed corrosion layers (Figure 6c,d). The SEM images confirmed the clusters of AlN-AlMn intermetallic particles, as observed on reference samples (Figure 3a). Their high content of Al (13.3 wt.%) suggests that the Al(OH)3 corrosion product was formed on their surface because of Al de-alloying (Equations (3)-(5)). It was reported that Al(OH)3 may be stable up to pH ≈ 8.5 because of its minimal solubility, and at pH > 9.5, the solubility will increase [77]. During the exposure of each material, the pH of SAR solution reached a value of ≈ 10 ( Figure 4a). Thus, after 30 days of exposure to SAR solution, EDS did not detect the presence of the N element (Table 3), as being part of AlN nanoparticles. On the other hand, hydrolysis of aluminum nitride  (Table 3) of the marked areas of interest confirmed the presence of the Mn-rich phase of Al-Mn intermetallic particles (labeled as 1) on the corroded surfaces (Figure 6a,b) even after the removal of the formed corrosion layers (Figure 6c,d). The SEM images confirmed the clusters of AlN-AlMn intermetallic particles, as observed on reference samples (Figure 3a). Their high content of Al (13.3 wt.%) suggests that the Al(OH) 3 corrosion product was formed on their surface because of Al de-alloying (Equations (3)-(5)). It was reported that Al(OH) 3 may be stable up to pH ≈ 8.5 because of its minimal solubility, and at pH > 9.5, the solubility will increase [77]. During the exposure of each material, the pH of SAR solution reached a value of ≈ 10 ( Figure 4a). Thus, after 30 days of exposure to SAR solution, EDS did not detect the presence of the N element (Table 3), as being part of AlN nanoparticles. On the other hand, hydrolysis of aluminum nitride (Equations (3)-(5)) may start at pH ≈ 3 [78] and, thus, this reaction should also be considered as a possible in SAR solution. Table 3. EDS elemental analysis (wt.%) of different surface zones on AM60-AlN nanocomposite and AM60 alloy (labeled in Figure 6), after exposure to SAR solution up to 30 days. The Al-Mn particles are considered local efficient cathodes [79][80][81][82] and, thus, they subsist on the Mg-matrix after the removal of corrosion layers (Figure 6c,d). However, it has been reported that, in the presence of chloride ions (NaCl), the Al-Mn phase is unstable and may suffer de-alloying, forming Al(OH) 3 with its posterior delamination [82]. Scanning Kelvin probe force macroscopy (KPFM) maps of the Volta potential difference measurements, for Al-Mn alloys immersed in NaCl solution, have shown the cathodic character of Al-Mn particles, as nobler than the adjacent Mg-matrix; thus, induced localized corrosion is observed in their vicinity [80,81]. The cross-sectional morphology of the AM60-AlN (Figure 6e) nanocomposite and that of the AM60 alloy ( Figure 6f) revealed a maximum depth of 141.0 µm of less intensive localized attack in the case of AM60-AlN and 155 µm for the AM60 alloy.

Zone
On the other hand, EDS analysis ( Table 3) indicated clearly that the main corrosion product of the cracked surface of each alloy was Mg(OH) 2 , in addition to MgO (Figure 6a,b).
The study of AM50 and AZ31 exposed to NaCl + MgCl 2 solution reported the formation of Mn 2 O 3 as an additional corrosion product [82]. The Mn standard potential (−1.18 V) was nobler than those of Al and Mg.

XPS Analysis
To complement the EDS elemental composition of the corrosion layers, formed on each alloy surface after 30 days of exposure to the acid rain model solution (SAR), XPS spectra analysis was performed (Figure 7). The binding energies (Table 4) of O1s, Mg2p, and Al2p were identified (Figure 7a,b). The high-resolution peak of O1s consists of two signals, according to the fitting-curves: the lower oxygen signal at 529.9 eV was attributed to MgO (O 2− ) and the high signal at 531.6 eV was associated with (OH − ) present in Mg(OH) 2 , as the main corrosion product, as well as a part of Al(OH) 3 . On the other hand, the signal of Mg2p contains the contributions of MgO and Mg(OH) 2 [83][84][85], and the signal of Al2p contains the contribution of Al(OH) 3 [86].
character of Al-Mn particles, as nobler than the adjacent Mg-matrix; thus, induced localized corrosion is observed in their vicinity [80,81]. The cross-sectional morphology of the AM60-AlN (Figure 6e) nanocomposite and that of the AM60 alloy ( Figure 6f) revealed a maximum depth of 141.0 µm of less intensive localized attack in the case of AM60-AlN and 155 µm for the AM60 alloy.
On the other hand, EDS analysis ( Table 3) indicated clearly that the main corrosion product of the cracked surface of each alloy was Mg(OH)2, in addition to MgO (Figure  6a,b). The study of AM50 and AZ31 exposed to NaCl + MgCl2 solution reported the formation of Mn2O3 as an additional corrosion product [82]. The Mn standard potential (−1.18 V) was nobler than those of Al and Mg.

XPS Analysis
To complement the EDS elemental composition of the corrosion layers, formed on each alloy surface after 30 days of exposure to the acid rain model solution (SAR), XPS spectra analysis was performed (Figure 7). The binding energies (Table 4) of O1s, Mg2p, and Al2p were identified (Figure 7a,b). The high-resolution peak of O1s consists of two signals, according to the fitting-curves: the lower oxygen signal at 529.9 eV was attributed to MgO (O 2− ) and the high signal at 531.6 eV was associated with (OH − ) present in Mg(OH)2, as the main corrosion product, as well as a part of Al(OH)3. On the other hand, the signal of Mg2p contains the contributions of MgO and Mg(OH)2 [83][84][85], and the signal of Al2p contains the contribution of Al(OH)3 [86].    Table 4 compares the change in time of the open circuit potential (OCP) values of the AM60-AlN nanocomposite and AM60 alloy, exposed for 30 days in the SAR (acid rain model solution). After 7 days, the OCP reached less negative values, because of the formed corrosion layers on the alloy surfaces. Until the end of the experiment (30 days), the OCP values were almost constant, with those of the AM60-AlN nanocomposite being less negative. These results are consistent with the mass losses (Figure 5a), as well as with the concentration of the released Mg-ions (Figure 4b). Figure 8 presents the fluctuations in corrosion current density (at OCP) for the AM60-AlN nanocomposite (Figure 8a) and AM60 alloy (Figure 8b), after 30 days of immersion in SAR solution. It is notable that the AM60 current density (≈7.8 µA cm −2 ) and its average amplitude (5.0 µA cm −2 ) were higher than those of the AM60-AlN nanocomposite (≈6.1 µA cm −2 and amplitude of 3.0 µA cm −2 ).
The spontaneous corrosion current fluctuations, associated with corrosion phenomena, were considered as electrochemical noise (EN), and the data were analyzed in the time and frequency domain, to identify the corrosion process. The analysis enabled the EN resistance (R n ) values, related to the standard potential (σ P ) and current (σ i ) deviations, to be determined [67]. The R n magnitude is considered an equivalent of the R p , (polarization resistance) [87][88][89], and it is inversely proportional to the corrosion rate [90], when the corrosion mechanism is controlled by charge transfer [88]. The R n values of the AM60-AlN nanocomposite (Table 5) after the first day and up to 30 days of exposure were 2-3 times higher than those of AM60, an indication of more efficient protective corrosion layers on the composite surface. The decreased R n values at the end of the experiment (30 days) correlated with the observed localized corrosion on the surfaces of both studied alloys (Figure 6e,f).
ative. These results are consistent with the mass losses (Figure 5a), as well as with the concentration of the released Mg-ions (Figure 4b). e 4. Change in time of values of AM60-AlN nanocomposite and AM60 alloy surfaces exposed to SAR (acid rain) ion for up to 30 days.  Figure 8 presents the fluctuations in corrosion current density (at OCP) for the AM60-AlN nanocomposite (Figure 8a) and AM60 alloy (Figure 8b), after 30 days of immersion in SAR solution. It is notable that the AM60 current density (≈7.8 µA cm −2 ) and its average amplitude (5.0 µA cm −2 ) were higher than those of the AM60-AlN nanocomposite (≈6.1 µA cm −2 and amplitude of 3.0 µA cm −2 ). The spontaneous corrosion current fluctuations, associated with corrosion phenomena, were considered as electrochemical noise (EN), and the data were analyzed in the time and frequency domain, to identify the corrosion process. The analysis enabled the EN resistance (R n ) values, related to the standard potential ( ) and current ( ) deviations, to be determined [67]. The Rn magnitude is considered an equivalent of the Rp, (polarization resistance) [87][88][89], and it is inversely proportional to the corrosion rate [90], when the corrosion mechanism is controlled by charge transfer [88]. The Rn values of the AM60-AlN nanocomposite (Table 5) after the first day and up to 30 days of exposure were 2-3 times higher than those of AM60, an indication of more efficient protective corrosion layers on the composite surface. The decreased Rn values at the end of the experiment (30 days) correlated with the observed localized corrosion on the surfaces of both studied alloys (Figure 6e,f).   The R n (R p ) values were used to calculate the j corr (Equation (8)) and the mass loss rate (MR), according to Equation (9), considering the tentative Tafel constant B' of 52.80 mV [91], and the data are summarized in Figure 9.
where K is a constant (8.958 × 10 −3 g cm 2 /µA m 2 d) and EW is the equivalent weight (11.91). Figure 10 presents the PSD (power spectral density) of corrosion current fluctuations vs. frequency for the AM60-AlN nanocomposite and AM60 alloy (on a bi-logarithmic scale), after 30 days of exposure in the SAR model solution. The extracted exponent β values (0.94 and 0.95) of both alloys were similar and corresponded to fractional Gaussian noise (fGn, −1 < β < 1), as a stationary and persistent process [92][93][94]. This process may be considered as the persistent localized corrosion attacks, according to the SEM images (Figure 6e,f). Figure 11 shows the Nyquist and Bode plots for the AM60-AlN nanocomposite and AM60 alloy exposed to SAR solution for different periods of immersion. After the first hour of immersion, the impedance diagram displayed two capacitive loops at higher (HF) and medium frequencies (MF), followed by the beginning of an inductive loop at low frequencies (LF). The HF-capacitive loop diameter, as a measure of the charge transfer resistance (Rt) at the alloy interface, depends on the double-layer capacitance particularities, including the effect of the formed corrosion layers [86,[95][96][97][98][99][100][101][102]. Its value is inversely related to the corrosion rate through the Stern-Geary equation [95,96,98]. The MF-capacitive loop was associated with mass transport processes connected to the diffusion of the released Mg 2+ ion through the layer of the corrosion product [95][96][97][98][99], while the LF-inductive loop may result from the beginning of localized corrosion on the alloy surfaces and/or linked to adsorption/desorption of intermediates on the electrode surface. The inductive loop could disappear when the corrosion protective layer is formed on the surface [95][96][97][98][99]. The increase in immersion time led to a rise in the diameter of both HF and MF capacitive loops, suggesting the formation of corrosion layers on the surfaces, acting as a physical barrier over time against the SAR aggressive media. The Bode plots were found to be in good agreement with those observed in Nyquist diagrams. The Rn (Rp) values were used to calculate the jcorr (Equation (8)) and the mass loss rate (MR), according to Equation (9), considering the tentative Tafel constant B' of 52.80 mV [91], and the data are summarized in Figure 9.

Electrochemical Impedance Spectroscopy (EIS)
where K is a constant (8.958 × 10 -3 g cm 2 /µA m 2 d) and EW is the equivalent weight (11.91).  Figure 10 presents the PSD (power spectral density) of corrosion current fluctuations vs. frequency for the AM60-AlN nanocomposite and AM60 alloy (on a bi-logarithmic scale), after 30 days of exposure in the SAR model solution. The extracted exponent β values (0.94 and 0.95) of both alloys were similar and corresponded to fractional Gaussian noise (fGn, −1 < β < 1), as a stationary and persistent process [92][93][94]. This process may be considered as the persistent localized corrosion attacks, according to the SEM images (Figure 6e,f).  The Rn (Rp) values were used to calculate the jcorr (Equation (8)) and the mass loss rate (MR), according to Equation (9), considering the tentative Tafel constant B' of 52.80 mV [91], and the data are summarized in Figure 9.

Figure 9.
Change in time of jcorr and MR of AM60-AlN nanocomposite and AM60 alloy, exposed to SAR solution for up to 30 days. Figure 10 presents the PSD (power spectral density) of corrosion current fluctuations vs. frequency for the AM60-AlN nanocomposite and AM60 alloy (on a bi-logarithmic scale), after 30 days of exposure in the SAR model solution. The extracted exponent β values (0.94 and 0.95) of both alloys were similar and corresponded to fractional Gaussian noise (fGn, −1 < β < 1), as a stationary and persistent process [92][93][94]. This process may be considered as the persistent localized corrosion attacks, according to the SEM images (Figure 6e,f). The EIS experimental data were fitted with two equivalent circuits, which correspond to the initial immersion period of 1 h (Figure 12a) and the longer period up to 15 days (Figure 12b) [103]. The main components of the suggested circuits were: R1 (Rt) and CPE1 double layer "capacitance" at the substrate/electrolyte interface; R2 (Rct) and CPE2 "capacitance," both characteristic of the charge transfer process in the presence of the corrosion layer, through which the released ions of Mg 2+ diffuse. The constant phase element (CPE) was introduced instead of a capacitor, to account for the dispersion of the time-constant due to different factors such as heterogeneities and surface roughness [95,98]. In addition, the circuit (a) includes R L and L1 as inductive resistance and inductance, respectively, presenting the dependence of the corrosion process from the surface coverage by the corrosion layer and its characteristics (porosity, cavities, and cracks). The fitted values of the circuit elements (Tables 6 and 7) were used to compare the corrosion behavior of the AM60-AlN nanocomposite to that of AM60 (the circuits have a fitting of 10 −4 ). related to the corrosion rate through the Stern-Geary equation [95,96,98]. The MF-capacitive loop was associated with mass transport processes connected to the diffusion of the released Mg 2+ ion through the layer of the corrosion product [95][96][97][98][99], while the LF-inductive loop may result from the beginning of localized corrosion on the alloy surfaces and/or linked to adsorption/desorption of intermediates on the electrode surface. The inductive loop could disappear when the corrosion protective layer is formed on the surface [95][96][97][98][99]. The increase in immersion time led to a rise in the diameter of both HF and MF capacitive loops, suggesting the formation of corrosion layers on the surfaces, acting as a physical barrier over time against the SAR aggressive media. The Bode plots were found to be in good agreement with those observed in Nyquist diagrams. The EIS experimental data were fitted with two equivalent circuits, which correspond to the initial immersion period of 1 h (Figure 12a) and the longer period up to 15 days (Figure 12b) [103]. The main components of the suggested circuits were: R1 (Rt) and CPE1 double layer "capacitance" at the substrate/electrolyte interface; R2 (Rct) and CPE2 "capacitance," both characteristic of the charge transfer process in the presence of the corrosion layer, through which the released ions of Mg 2+ diffuse. The constant phase element (CPE) was introduced instead of a capacitor, to account for the dispersion of the timeconstant due to different factors such as heterogeneities and surface roughness [95,98]. In addition, the circuit (a) includes RL and L1 as inductive resistance and inductance, respectively, presenting the dependence of the corrosion process from the surface coverage by the corrosion layer and its characteristics (porosity, cavities, and cracks). The fitted values of the circuit elements (Tables 6,7) were used to compare the corrosion behavior of the AM60-AlN nanocomposite to that of AM60 (the circuits have a fitting of 10 -4 ).    The value of the polarization resistances (R p ) was calculated to compare the corrosion resistance of the studied materials exposed to SAR acid solution (Tables 6 and 7), using the following equations derived from the equivalent circuits ( Figure 12): Equivalent circuit (a) 2 capacitive loops and 1 inductive : Equivalent circuit (b) 2 capacitive loops : The R p values show that, with the development of the corrosion process, in the SAR aggressive environment, the AM60-AlN nanocomposite presented a R p ≈ 7% higher than that of AM60, as an indication of the more efficient protective corrosion layer formed.  Figure 13a compares the evolution in time of charge transfer resistance (R2) values (Tables 6 and 7) characteristic of the double layer in the presence of corrosion products. The higher R2 of the AM60-AlN nanocomposite suggests that the formed corrosion layer was probably denser, more protective, and less fractured than those of the AM60 alloy surface. The reported difference in R2 values (Figure 13a), based on EIS measurements, coincided with those obtained from EN measurements (Rn, Table 5). The reported difference in R2 values (Figure 13a), based on EIS measurements, coincided with those obtained from EN measurements (Rn, Table 5). Figure 13a compares the evolution in time of charge transfer resistance (R2), corresponding to the AM60-AlN nanocomposite and AM60 alloy. According to the Brug Equation (12), the values of CPE2 (Tables 6,7) were transformed to the capacitance values (C2) of the double layer in the presence of corrosion products, using Rs, R2, and n2 values [104,105] (Figure 13b). The increase in time of AM60 alloy capacitance (Figure 13b) may be attributed to the formation of a thicker, although less protective, corrosion layer (lower R2 values).

Conclusions
This present study compared the initial stages of the corrosion process on a magnesium-aluminum AM60-AlN nanocomposite surface with those of AM60 immersed in model acid rain solution (SAR) for up to 30 days.
• The incorporation of 1.0% wt.% AlN aluminum nitride nanoparticles (≈ 80 nm average diameter) in the AM60 matrix favored a grain size reduction of 12%, as revealed by the optical images, as well as a slight 7% increase in the Vickers hardness. The AlN particles were "attached" to the phase of Mn-rich Al-Mn intermetallic particles, forming a cluster, according to the SEM images. • The immersion test revealed that the increase in alkaline values of SAR solution was lower during the exposure of the AlN nanocomposite than that of the AM60 alloy. This fact correlates with the 24% lower concentration of the released Mg-ions from the AM60-AlN surface. However, the pH increase suggests that Al de-alloying may occur, as well as Al(OH)3 formation as a corrosion product, as suggested by XPS analysis. • The formed corrosion layers showed cracks at the end of the exposure and localized   (Tables 6 and 7) were transformed to the capacitance values (C2) of the double layer in the presence of corrosion products, using Rs, R2, and n 2 values [104,105] (Figure 13b). The increase in time of AM60 alloy capacitance (Figure 13b) may be attributed to the formation of a thicker, although less protective, corrosion layer (lower R2 values).

Conclusions
This present study compared the initial stages of the corrosion process on a magnesiumaluminum AM60-AlN nanocomposite surface with those of AM60 immersed in model acid rain solution (SAR) for up to 30 days.

•
The incorporation of 1.0% wt.% AlN aluminum nitride nanoparticles (≈ 80 nm average diameter) in the AM60 matrix favored a grain size reduction of 12%, as revealed by the optical images, as well as a slight 7% increase in the Vickers hardness. The AlN particles were "attached" to the phase of Mn-rich Al-Mn intermetallic particles, forming a cluster, according to the SEM images.

•
The immersion test revealed that the increase in alkaline values of SAR solution was lower during the exposure of the AlN nanocomposite than that of the AM60 alloy. This fact correlates with the 24% lower concentration of the released Mg-ions from the AM60-AlN surface. However, the pH increase suggests that Al de-alloying may occur, as well as Al(OH) 3 formation as a corrosion product, as suggested by XPS analysis.

•
The formed corrosion layers showed cracks at the end of the exposure and localized attacks near Al-Mn intermetallic particles, acting as efficient cathodic sites. The crosssectional images revealed a higher intensity of attacks on the AM60 alloy surface without reinforcement.

•
The β exponents extracted from the PSD plots of the corrosion current fluctuations classified the corrosion of the studied materials as fractional Gaussian noise (fGn) of the stationary persistent localized process. • XPS analysis suggests that the main corrosion products were MgO and Mg(OH) 2 , as well as a lower content of Al(OH) 3 that may delay the advance of the AM60-AlN corrosion process.

•
The EIS Nyquist plots and the parameters of the adjusted equivalent circuits indicated that the charge transfer resistance (R2) and capacitance values, characteristic of the double layer in the presence of corrosion products, were higher in the presence of AlN nanoparticles. This may favor the formation of a more dense and efficient protective corrosion layer because of a slight grain refinement. The tendency of the increase in R2 values for the AM60-AlN nanocomposite coincided with that of Rn values obtained from EN measurements. Data Availability Statement: Data presented in this study are available on request from the corresponding author. The data are not public, due to privacy issues.