Investigations of Electrochemical Characteristics of Mg-Al-Ca Alloys

Abstract

Magnesium alloys possess a high strength-to-density ratio, thereby increasingly being utilized as lightweight structural materials in a range of industrial applications. Nevertheless, to compete with established materials like aluminum alloys, it is essential to understand the corrosion behavior of Mg and its alloys, as their high reactivity hampers industrial application. The addition of Ca to wrought Mg-Al alloys has gained attention for its ability to improve mechanical properties while also enhancing processing behavior. However, the wide range of alloy compositions within the class of Mg-Al-Ca alloys results in a variety of different corrosion properties. Consequently, this study contributes by investigating the corrosion behavior of two Mg-Al-Ca alloys, highlighting the influence of chemical composition and microstructure.

1. Introduction

Products made of lightweight materials are in high demand for modern, low-emission transport applications. Magnesium (Mg) alloys are of particular interest due to their great weight saving potential, mainly realized by their high specific strength. Applications are manifold, and parts made from Mg can be produced using a wide range of manufacturing processes, with casting and forming being the most common [1]. Wrought Mg typically shows higher strength compared to cast parts, as deformation with subsequent recrystallization can produce a fine-grained microstructure, and effects like grain boundary strengthening and precipitation hardening can be utilized more efficiently. Recently, extruded Mg-Al-Ca alloys were extensively researched, as they offer a unique combination of properties, e.g., high mechanical strength [2,3,4]. Calcium (Ca) as an alloying element in Mg-Al reduces the texture intensity, enables age hardening, and increases oxidation as well as flammability resistance [5,6]. While the influence of elements such as manganese (Mn) and impurities like nickel (Ni) and iron (Fe) on corrosion behavior is well investigated in general [7], interactions between composition, microstructure, and corrosion resistance in Mg-Al-Ca alloys are not yet fully understood.

It is known that the addition of Ca to Mg-Al alloys suppresses the discontinuous precipitiation of Mg₁₇Al₁₂, forming various Ca-containing phases instead [8,9]. The increase of the Ca/Al ratio from 0 to 1 leads to a change of the secondary phase from Mg₁₇Al₁₂ (A12) to the Laves phases Al₂Ca (C15), (Mg,Al)₂Ca (C36), and Mg₂Ca (C14) [10]. With increasing amounts, these Ca-containing phases crosslink and form a skeleton-like structure [11,12]. By adjusting the Ca/Al ratio, the nature of the Ca-containing phases as well as their morphology can be changed, which strongly affects the resulting mechanical properties [13]. The addition of Ca to Mg-Al improves the corrosion resistance by (i) reducing the micro galvanic effect of Mg₁₇Al₁₂, (ii) forming a protective oxide film, and (iii) creating continuously precipitated Laves phases (Mg,Al)₂Ca, which can act as sacrificial anodes [14,15]. Yang et al. studied the corrosion behavior of Mg-5Al-0.3Mn-xCa alloys (x = 0.2–4.0 wt. %) and achieved the best corrosion resistance using 2 wt. % Ca [15].

Wu et al. [16] studied the corrosion properties of cast and extruded Mg-2Ca-xAl alloys (x = 0, 2, 3, 5) and achieved the highest corrosion resistance in the extruded Mg-2Ca-2Al material. The alloys without Al were inferior in as-cast and as-extruded state, since only Mg₂Ca was present as an intermetallic phase, which is less noble than the Mg matrix and distributed along the grain boundaries, leading to corrosion channels. By adding Al, the type and morphology of the phases were altered, leading to improved corrosion resistance. Additionally a barrier effect of the continuous distributed (Mg,Al)₂Ca on the grain boundaries of the cast samples was described [16]. Chauldry et al. observed an increase in the corrosion resistance of AZ31 when Ca (0.5 wt. %) was added and attributed this to the formation of (Mg,Al)₂Ca and the altered morphology of Mg₁₇Al₁₂ [17].

Fe is a common impurity in Mg alloys and has a significant effect on their corrosion rate [18]. When its solubility limit is exceeded, primary intermetallic phases can precipitate, forming galvanic cells with the surrounding Mg matrix, therefore increasing the corrosion rate [19,20]. Via additions of Al and Mn, the stoichiometry of the Fe-containing primary phases can be modified, potentially reducing its negative effect on the corrosion behavior.

In this work, the corrosion properties of two Mg-Al-Ca alloys are studied based on thermodynamic calculations, microstructural investigations, and corrosion resistance evaluations using cyclic potentiodynamic polarization (CPP) and electrochemical impedance spectroscopy (EIS) measurements.

2. Materials and Methods

Two Mg-Al-Ca alloys (for composition, see

Table 1. Alloy compositions.

Alloy	Mg [wt. %]	Al [wt. %]	Ca [wt. %]	Mn [wt. %]	Fe [ppm]	Ni [ppm]
AX22	balance	1.88	1.88	0.00	340	30
AX32	balance	3.02	2.06	0.46	150	20

) were prepared using commercially pure Mg (99.8%) and Al (99.85%), Ca granules (99.9%), and MnCl₂ (98%) in a protective gas atmosphere (Ar + 1 vol. % SF₆). Billets with a dimension of Ø 75 mm × 240 mm were cast into a preheated (200 °C) mild steel mold via low pressure die casting. The cast material was machined to extrusion feedstock of Ø 48.5 mm × 180 mm. Forming was performed at a temperature of 400 °C with a ram speed of 0.5 mm/s via direct extrusion, using a 1.5 MN extrusion press (NEHP 1500.01, Müller Engineering, Todtenweis, Germany). Profiles with a rectangular cross section of 25 mm × 12 mm were produced at an extrusion ratio of 1:6.5. The chemical composition was analyzed via optical emission spectroscopy (SPECTROMAXx 6, SPECTRO Analytical Instruments, Kleve, Germany).

Thermodynamic calculations in equilibrium state were performed using the software Thermo-Calc (version 2023b) and magnesium database TCMG6 (both from Thermo-Calc Software AB, Solna, Sweden).

Polishing was performed using diamond polishing suspensions with particle sizes of 3, 1, and ¼ μm, followed by final polishing using OP-S Chem silica suspension and distilled water (ratio 1:1). The microstructure was analyzed via optical light microscope (Olympus BX60) and scanning electron microscopy (SEM) (Brno, Czech Republic), TESCAN MIRA 3. Electron micrographs were taken using a four-quadrant backscattered electron detector (BSE) at 15 kV accelerating voltage and 15 mm working distance. Elemental analysis (EDS) was performed using an Octane Elect Super detector (EDAX) (Mahwah, NJ, USA).

To determine the corrosion resistance of AX22 and AX32 magnesium alloys, cyclic potentiodynamic polarization (CPP) measurements and electrochemical impedance spectroscopy (EIS) (BioLogic, Seyssinet-Pariset, France) were used. All measurements were repeated at least three times to ensure sufficient accuracy of the obtained results.

As preparation for the corrosion tests, the extruded samples were ground using SiC abrasive papers (grit size up to p1200), providing a sample area of 1 cm². CPP measurements were performed for an exposure time of 15 min in 0.01 M NaCl at a temperature of 22 ± 2 °C, using a laboratory potentiostat Biologic SP-300. The potential range was set from −200 mV to +500 mV vs. open circuit potential (OCP) with a potential change of 1 mV·s⁻¹. Measurements were taken using a three-electrode cell system consisting of a working electrode, a platinum electrode (counter electrode), and a saturated calomel electrode (reference electrode). For obtaining electrochemical characteristics such as corrosion potential (E_corr), corrosion current density (i_corr), corrosion coefficients (β_a and β_c), and corrosion rate (r_corr), Tafel analysis (EC-Lab V11.10 software) was performed.

The electrochemical impedance spectroscopy (EIS) method was used to further investigate the corrosion behavior of the samples for 1 h to 168 h immersion in a 0.01 M NaCl solution (22 ± 2 °C). The measurements were carried out in the frequency range from 100 kHz to 10 mHz, and the change in frequency was set to 10 times per decade using a three-electrode cell system. The potential signal was set at 15 mV, and the mean value was identical with the open circuit potential (OCP). Corrosion parameters were obtained from a quantitative analysis of the measured Nyquist plots using the equivalent circuit method through EC-Lab V11.10 software. The corresponding equivalent circuit (Figure 1) was used to analyze and fit the obtained diagrams to provide information about the corrosion process. It can be asserted that the actual electrochemical interface is characterized by a surface that is typically not smooth, uniform, or defect-free, as commonly found in a pure capacitor. Instead, it consists of numerous pores, irregularities, and surface roughness [21,22]. Due to this fact, three types of equivalent circuits were used to study mechanically treated AX22 and AX32 magnesium alloys. The solution resistance (R_s) is the resistance between the reference and the working electrode. The CPE element expresses a constant phase element, replacing the capacitor in the circuit, describing the heterogeneity and volume imperfections of the electrode surface [23,24]. CPE is according to [25] defined as:

CPE ≈ C = [C(jω)ⁿ]⁻¹

(1)

with n described as follows: when n equals to 1, the CPE is a pure capacitance, whereas when the value of n is equal to 0, the CPE is a pure resistance. The equivalent circuit with an inductive loop consists of the inductive resistance (R_L), the charge transfer resistance of the base material (R₁), and the inductive element L, representing the inductance (Figure 1b). The appearance of an inductive loop is associated with the adsorption and desorption of the intermediate species at the low-frequency section [26]. For the analysis of the Nyquist diagrams, the equivalent circuit method was used [23,27,28]. Polarization resistance R_p describes the impedance of the measured system, and it is directly related to the corrosion resistance of the measured surface [24].

The obtained EIS data were fitted and the R_p values calculated based on the equivalent circuit (EC) shown in Figure 1. The R_p value in case of the system with one capacitive loop (Figure 1a) is equivalent to pore resistance. When the Nyquist plot consists of two capacitive loops, the final corrosion resistance is the sum of the partial resistances R₁ and R₂ (Figure 1b), whereas the R_p value in case of the one capacitive and one inductive loop can be estimated using following equation [23,29,30]:

$${\mathrm{R}}_{\mathrm{p}}=\frac{{\mathrm{R}}_1\mathrm{ }\times \mathrm{ }{\mathrm{R}}_{\mathrm{L}}}{{\mathrm{R}}_1+\mathrm{ }{\mathrm{R}}_{\mathrm{L}}}$$

(2)

Further information on the equivalent circuit measurement method can be found in [24,31,32].

3. Results and Discussion

3.1. Microstructure

Figure 2 shows the calculated ternary phase diagram of Mg-Al-Ca at room temperature (RT) and the corresponding phase fractions of AX22 and AX32 given in atomic percent (at. %). The alloy AX22 is located in the α-Mg + Mg₂Ca + Al₂Ca phase region while AX32 is located in the region α-Mg + Al₂Ca. As can be seen in the calculated values, the higher Al content of AX32 leads to a significantly increased amount of Al₂Ca (3.80 at. % in AX32 in comparison to 2.74 at. % in AX22). At the same time, no Mg₂Ca can be found in AX32, while in AX22 the amount of Mg₂Ca is calculated to be 0.71 at. %. The total amount of Ca-containing phases is 3.80 at. % in AX32 (Al₂Ca) and 3.45 at. % in AX22 (Al₂Ca + Mg₂Ca). It should be mentioned that (Mg,Al)₂Ca is not stable at RT; therefore, the equilibrium calculation does not determine this phase. Nevertheless, this phase can also occur at RT in cast samples produced using conventional solidification conditions. Furthermore, Fe-containing phases can be found, which are Al₂Fe in AX22 and Al₅Fe₂ in AX32. The lower Fe-content in AX32 compared to AX22 can be due to settling of Mn-Fe particles in the melt before casting the material [33]. According to the calculation, the Fe-content in Al₈Mn₅ is negligible, while according to the literature, this phase plays an important role in binding Fe [34].

Figure 3a,b show the microstructures of the extruded alloys in transverse direction (TD). Al and Ca containing intermetallic phases are arranged along the extrusion direction (ED) in both alloys, corresponding well with the literature [16,35] and the calculated ternary phase diagram. A distinction between Al₂Ca and (Mg,Al)₂Ca was not possible via EDS. While the total amount of intermetallic phases is comparable in both alloys, no Mg₂Ca was found in AX22. In AX22 (Figure 3a), fine particles (<5 µm) are homogenously dispersed in the matrix. The EDS analysis identified Al and Fe, while in AX32 (Figure 3b), Al-Mn-Fe phases form instead, due to the presence of Mn. While both alloys show similar grain sizes (average grain sizes: AX22 = 9 μm, AX32 = 12 μm), the Fe-containing intermetallic phases are larger, and the Ca-containing phases are smaller in AX32.

In the calculated isothermal sections (Figure 4a,b), the solubility limit of Fe at 660 °C in AX22 and AX32 is shown as a function of the Mn content. In both cases, the solubility of Fe is rather small (<30 ppm). When the Fe content exceeds this solubility limit, primary Fe-phases form in the melt. In Mg-Al-Ca alloys without Mn, as in AX22, the phases are of the type Al_xFe_y. With the addition of Mn, Al-Mn-Fe phases with variable stoichiometry and crystal structure (BCC, CUB_A13 and Al₈Mn₅) can be formed. These Mn-containing phases are known to bind Fe, thereby improving the corrosion resistance of the alloys by reducing the microgalvanic effect between the Mg matrix and Fe-containing intermetallic phases [20,36].

3.2. Corrosion Testing

3.2.1. Potentiodynamic Polarization Tests

The corrosion properties were examined in the transverse direction TD (cf. Figure 3). Figure 5a shows the open circuit potential (OCP) curves of extruded AX22 and AX32 immersed in 0.01 M NaCl solution for 900 s. At the beginning of the test, the OCP value of AX22 is ~40 mV/SCE lower than that of AX32. During the test, the OCP values of both alloys increase. The value for AX22 shows a steeper increase, and after ~250 s, the OCP of AX22 is higher than that of AX32. The small fluctuations in the curves can be a sign of ongoing pitting corrosion [16,37]. The higher OCP values for AX22 indicate a higher thermodynamical stability compared to AX32. The more negative potential value of AX32 is most likely attributed to the differences in the Fe- and Ca-containing phases in the alloys. Based on the calculations (Figure 2), the number of Fe-containing phases in AX32 is lower than in AX22, while at the same time, the proportion of Al₂Ca is higher. According to Mingo et al., the Al₂Ca phase has a higher positive potential compared to the Mg matrix [38]. Due to this, the corrosion of the substrate is accelerated because Al₂Ca acts as a cathode. Furthermore, the calculation shows a presence of Mg₂Ca in AX22. The Mg₂Ca phase has a more negative corrosion potential than the Mg matrix, therefore acting as an anode [39]. During the corrosion process, this phase dissolves first and protects the α-Mg matrix against corrosion.

Figure 5b shows the cyclic potentiodynamic polarization (CPP) curves for both alloys. The curves were analyzed via Tafel extrapolation analysis and were plotted on the semi-logarithmic scale. From the Tafel analysis of the CPP curves, the following corrosion characteristics were obtained: corrosion potential (E_corr), corrosion current density (i_corr), corrosion rate (r_corr), and Tafel constants (β_a and β_c). The values are summarized in

Table 2. Electrochemical characteristics measured in 0.01 M NaCl.

Alloy	Ecorr [mV vs. SCE]	icorr [μA cm−2]	βc [mV dec−1]	βa [mV dec−1]	Corrosion Rate [mm y−1]
AX22	−1 475 ± 12	20.40 ± 2.34	226 ± 7	158 ± 7	0.47 ± 0.05
AX32	−1 482 ± 11	9.81 ± 1.02	195 ± 11	149 ± 9	0.23 ± 0.03

. While the thermodynamic part of the corrosion reaction is described by E_corr, the description of the kinetic part is attributed to i_corr. The value of the corrosion current density (i_corr) is significantly higher for AX22 (20.40 μA cm⁻²) than for AX32 with 9.81 μA cm⁻². This can most likely be attributed to the described formation of Al-Mn-Fe particles in AX32 and Al-Fe particles in AX22 and the differences in Ca-containing phases. As already stated, ferrous particles containing Mn have less negative influence on electrochemical behavior of Mg alloys, because they have a lower microgalvanic effect than particles without Mn, i.e., Al₂Fe and Al₅Fe₂ [34]. Therefore, the lower content of Fe and the higher amount of Mn in AX32 can contribute to the improved corrosion behavior of the alloy [40]. Furthermore, it is suspected that the size of the intermetallic particles has an influence on the corrosion characteristics; nevertheless, this effect cannot be elaborated in detail with the corrosion measurement methods used. However, extruded Mg alloys are characterized by a strong basal texture, which is coupled with the hexagonal crystal structure of the alloys. According to Jeong et al. and Liu et al., the crystal plane orientation and the basal texture of Mg alloys have a great influence on the corrosion behavior [35,41]. The obtained value for the corrosion potential (E_corr) is higher for AX22 (−1475 mV) than for AX32 (−1482 mV), which is assumed to be an effect of the higher Al and Ca content in AX32. For example, Yang et al. found that with increasing Ca content, the value of E_corr decreases [15]. Furthermore, the value of the corrosion rate (r_corr) is lower for AX32 (0.23 mm/y) compared to AX22 (0.47 mm/y), which is attributed to the lower value of i_corr for alloy AX32 as r_corr can be calculated using Faraday’s law according to the following equation [42]:

$${\mathrm{r}}_{\mathrm{corr}}=\frac{{\mathrm{i}}_{\mathrm{c}\mathrm{o}\mathrm{r}\mathrm{r}}·\mathrm{M}}{\mathrm{n}·\mathrm{F}} [\mathrm{mm}/\mathrm{y}]$$

(3)

where M describes the atomic weight [g/mol], n expresses the number of exchanged electrons in the dissolution reaction, and F marks Faraday’s constant [96,485 C/mol].

3.2.2. Electrochemical Impedance Spectroscopy

To provide a more precise evaluation of the electrochemical behavior of AX22 and AX32 alloys during corrosion, EIS measurements were caried out. Figure 6 shows the Nyquist plots of AX22 (Figure 6a) and AX32 (Figure 6b) immersed in 0.01 M NaCl solution. The corresponding results of electrochemical corrosion characteristics are given in

Table 3. Corrosion electrochemical characteristics of AX22 Mg alloy measured in 0.01 M NaCl.

Time	Rs (Ω·cm2)	R1 (Ω·cm2)	RL (Ω·cm2)	Rp (Ω·cm2)	L (H·cm2)	CPE1(F·sn−1·10−6)	n1
1 h	657 ± 12	1120 ± 102	1820 ± 132	693 ± 58	10,501 ± 136	11.3 ± 0.2	0.9
2 h	644 ± 18	726 ± 33	1087 ± 115	435 ± 26	5215 ± 99	24.9 ± 4.3	0.9
4 h	636 ± 15	551 ± 41	735 ± 44	315 ± 21	3383 ± 123	76.3 ± 0.1	0.8
8 h	628 ± 11	532 ± 42	1163 ± 112	365 ± 31	1163 ± 55	87.4 ± 1.5	1
12 h	634 ± 20	493 ± 47	1100 ± 159	340 ± 36	11,649 ± 63	91.5 ± 3.4	1
24 h	584 ± 13	377 ± 38	983 ± 76	273 ± 25	10,755 ± 42	80.9 ± 3.1	1
48 h	522 ± 14	243 ± 24	829 ± 23	188 ± 12	7885 ± 49	76.5 ± 5.2	1
96 h	448 ± 15	155 ± 12	536 ± 41	120 ± 9	5563 ± 21	77.2 ± 1.1	1
168 h	432 ± 23	128 ± 18	659 ± 28	107 ± 11	3728 ± 35	85.8 ± 7.2	0.9

and

Table 4. Corrosion electrochemical characteristics of AX32 Mg alloy measured in 0.01 M NaCl.

Time	Rs (Ω·cm2)	R1 (Ω·cm2)	R2 (Ω·cm2)	RL (Ω·cm2)	Rp (Ω·cm2)	L (H·cm2)	CPE1 (F·sn−1·10−6)	CPE2 (F·sn−1·10−6)	n1	n2
1 h	731 ± 20	2283 ± 125	1120 ± 182	-	3403 ± 307	-	-	950.9 ± 10.2	0.9	0.9
2 h	727 ± 19	2946 ± 150	1552 ± 175	-	4498 ± 325	-	-	804.2 ± 15.3	0.9	0.9
4 h	714 ± 24	5326 ± 77	3434 ± 163	-	8760 ± 240	-	9.5 ± 0.1	495.7 ± 7.8	0.9	0.7
8 h	707 ± 20	5012 ± 111	4215 ± 192	-	9227 ± 303	-	11.3 ± 1.2	514.1 ± 6.1	0.9	0.8
12 h	703 ± 18	4541 ± 114	3064 ± 120	-	7605 ± 234	-	12.1 ± 0.7	666.2 ± 7.3	0.9	0.8
24 h	674 ± 22	5835 ± 102	-	-	5835 ± 102	-	17.1 ± 0.1	-	0.9	-
48 h	638 ± 15	5449 ± 67	-	6 548 ± 62	2974 ± 32	38,460 ± 420	23.4 ± 0.4	-	0.9	-
96 h	566 ± 13	5962 ± 78	-	9 602 ± 85	3678 ± 41	77,624 ± 231	42.2 ± 2.3	-	0.8	-
168 h	532 ± 14	5455 ± 82	-	-	5455 ± 82	-	37.4 ± 1.8	-	0.9	-

. It can be seen in Figure 6a,b that the Nyquist plots of both alloys consist of two types of electrochemical processes: a high-frequency capacitance loop representing the charge transfer reaction of surface oxide or hydroxide film and a low-frequency inductance loop expressing the adsorption and desorption of the intermediate species [43]. According to Yang et al., the appearance of an inductance loop is caused by pitting corrosion [15]. However, the appearance of the inductive loop can also be attributed to the dissolution caused by chlorides [29]. The charge transfer reaction represents the exchange of electrons between molecules or atoms; thus, this describes the oxidation and reduction process [44]. Corrosion resistance is determined based on the size of the loop radii of the single lines in the Nyquist plots [45].

The value of R_p is used to evaluate the overall corrosion resistance where the better corrosion resistance is attributed to the Mg alloy with the higher R_p value [30]. The highest value of polarization resistance R_p (9227 Ω·cm²) was measured for alloy AX32 after 8 h of exposure (Table 4), whereas, for the alloy AX22, the highest value of R_p was obtained after 1 h of exposure with 445 Ω·cm² (Table 3). After 168 h of exposure, the values of R_p were 5455 Ω·cm² for AX32 and 107 Ω·cm² for AX22 (Table 4).

Alloy AX22 exhibits its highest values for R₁ and R_L during short exposure times, and these values decline as the exposure time increases. This can potentially be attributed to the occurrence of Al₂Ca precipitates, leading to a reduction in the protective behavior of the surface films and a simultaneous deterioration during the electrochemical corrosion reaction [16,30]. In contrast, for AX32, the values for the partial resistances R₁ and R₂ show an increasing trend with longer exposure times, indicating the formation of a protective film on the alloy surface, improving the corrosion behavior during the electrochemical reaction.

In our investigations, the AX32 alloy shows significantly higher CPE₁ values compared to AX22, which is assumed to be a result of more severe local corrosion in AX32 due to the dissolution of α-Mg. This is confirmed by Wu et al., who showed larger CPE₁ values in an AC32 alloy (Mg-2Ca-3Al) compared to AC22 (Mg-2Ca-2Al) [16].

3.2.3. Corrosion Morphology Characterization

Figure 7 shows OM and SEM micrographs of the corroded areas of AX22 (Figure 7a,c,e) and AX32 (Figure 7b,d,f) after 168 h EIS measurement in 0.1 M NaCl solution. The corrosion is more pronounced in AX22 (Figure 7a) than in AX32 (Figure 7b). This can potentially be attributed to the lower amount of Al in AX22, as a higher content of Al atoms promotes the formation and compactness of protective corrosion films (Al(OH)₃ or Al₂O₃) during the corrosion process [30] and the reduced microgalvanic effect of Fe in AX32 due to the presence of Mn. In addition, the alloying element Ca can also form an adherent oxide film on the surface of magnesium alloys. This surface film can act as a protection against corrosion caused by Cl⁻ ions. The reaction of Ca with atmospheric CO₂ thereby ensures the formation of hydroxides and carbonates [46]. Common corrosion products of Mg-Al-Ca alloys can include the following compounds: MgO, Mg(OH)₂, Al₂O₃, Al(OH)₃, CaO, and Ca(OH)₂. The EDX analysis of the corrosion products formed on AX alloys (

Table 5. Results of EDX analyses (in wt. %) of the corrosion products on the surfaces of AX22 and AX32 alloys after 168 h in 0.1 M NaCl solution.

Alloy	Mg [wt. %]	O [wt. %]	Al [wt. %]	Ca [wt. %]	C [wt. %]
AX22	22.2	59.8	2.1	0.9	15.1
AX32	76.0	1.3	0.5	0.4	21.8

) showed primarily Mg-containing oxides, i.e., MgO and/or Mg(OH)₂. These products are porous; therefore, the aggressive solution can continuously penetrate though this protection and react with the magnesium substrate [47].

4. Conclusions

In this work, the microstructure and corrosion characteristics of two Mg-Al-Ca alloys with a Ca content of ~2 wt. % was examined, and the following conclusions can be drawn:

➢

The microstructure of as-extruded AX22 and AX32 alloys is characterized by a fine-grain structure with Mg-Al-Ca phases arranged along the extrusion direction.

➢

Intermetallic phases of the types Al-Mg-Ca, Al-Fe, and Al-Mn-Fe were found, while Mg₁₇Al₁₂ and Mg₂Ca were not detected.

➢

In AX32, the Mn transforms the Al-Fe phases to Al-Mn-Fe phases.

➢

AX32 showed better corrosion resistance with values of i_corr (9.81 μA cm⁻²) and r_corr (0.23 mm/y) in comparison with the AX22 alloy. This can be mainly attributed to the addition of Mn, binding the Fe in Al-Mn-Fe phases, and the higher content of Al and the slightly higher content of Ca.

➢

R_p values measured via EIS confirmed that the AX32 alloy has better corrosion resistance compared to AX22.