Corrosion Behavior and Mechanism of Mg-1Bi and Mg-1Sn Extruded Alloys

Abstract

Improving the corrosion resistance of magnesium (Mg) alloys is a long-term challenge, especially when cost-effectiveness is taken into account. In this work, Mg-1Bi and Mg-1Sn extruded alloys are prepared, and the effects of cost-effective Bi and Sn on the corrosion behavior of Mg alloys are comparatively studied. The corrosion resistance of the Mg-1Sn alloy (P_H: 2.83 ± 0.19 mm y⁻¹) is better than that of the Mg-1Bi alloy (P_H: 13.75 ± 1.12 mm y⁻¹), being about five times greater. In addition to the relatively low dislocation density in Mg-1Sn alloy, the difference in corrosion resistance is mainly attributed to two aspects of influence brought about by the addition of Sn and Bi. The Mg₂Sn phase introduced by the addition of Sn has a potential difference (PD) of ~30 mV, which is significantly lower than that (~90 mV) of the Mg₃Bi₂ phase introduced by adding Bi, thereby weakening the micro-couple corrosion tendency of the Mg-1Sn alloy. The addition of Bi has little effect on the corrosion film, while the addition of Sn makes the corrosion film on the Mg-1Sn alloy contain SnO₂, improving the compactness of the corrosion film and thereby enhancing the corrosion protection effect.

1. Introduction

Magnesium (Mg) alloys are highly regarded as “green engineering materials of the 21st century” mainly due to their lightweight and high specific strength, abundant resources, and easy recyclability [1,2,3]. They have significant application potential in aerospace, military hardware, automobiles, and 3C (computer, communication, consumer electronics) industries [4,5]. As engineering structural materials, Mg alloys have been relatively well studied in terms of their mechanical properties (strength and plasticity). However, in actual production and application processes, due to the high chemical activity and non-dense corrosion product film, the corrosion resistance of Mg alloys is usually poor [6,7,8]. Even though there have been relevant studies, poor corrosion resistance remains a major bottleneck restricting their practical application.

It is well known that appropriate alloying not only enhances the mechanical properties of Mg alloys but also improves their corrosion resistance. The improvement in corrosion resistance mainly results from alloying elements regulating the potential difference (PD) of the alloy matrix and the density of the corrosion film [9,10]. For instance, numerous studies have indicated that appropriate rare earth (RE) alloying can reduce the PD between the secondary phase and Mg matrix, avoiding severe micro-galvanic corrosion, and can also change the structure and composition of corrosion product film, enhancing the protective effect on the alloy matrix [11,12,13]. Nevertheless, considering that REs are a national strategic resource and have a relatively high price, finding cost-effective alternative elements would be beneficial for expanding the application of Mg alloys.

The common low-cost alloying elements currently available include Al [14], Ca [15], Zn [16], Zr [17], Bi [18], and Sn [19]. Among them, Bi and Sn, as beneficial elements for Mg alloys, have attracted the attention of researchers in the last few years. Some Bi and Sn containing Mg alloys with high mechanical properties or heat resistance have been developed, demonstrating significant application potential. Majhi et al. [18] found that the continuous and thermally stable Mg₃Bi₂ phase enhances the dislocation pile-ups along with dislocation entanglement within α-Mg grains, enhancing creep resistance in AZ91-0.55Bi alloy. Li et al. [20] developed an Mg-5Bi-3Al-5Sn (wt%) alloy with a yield strength of 384 MPa and ultimate tensile strength of 400 MPa by taking advantage of the precipitation-strengthening effect of nano-scale Mg₃Bi₂ and Mg₂Sn (approximately 80 MPa). Sun et al. [19] found that after peak aging, the Mg₂Sn precipitates could enable the high-temperature (230 °C) yield strength of the Mg-4Sn-3Al-1Zn alloy to reach 97 MPa, which is about 15% higher than that before aging. At present, only a few studies have investigated the corrosion resistance of alloys containing Sn and Bi. Li et al. [21] found that as the Sn content increases, the corrosion resistance of the Mg-2Zn-xSn-0.5Ca (wt%, x = 0.5, 1, 2) alloys first increases and then decreases, and when the Sn content is 1 wt%, the alloy has the lowest corrosion rate of ~2.1 mm y⁻¹. Xu et al. [22] studied the influence of Mg₃Bi₂ in Mg-xBi (wt%, x = 0.2, 0.8) alloys on the mechanical and anti-corrosion properties, and they calculated the corrosion rate only through the corrosion current density. Overall, the research on the influence of these two elements on the corrosion resistance of Mg alloys is still scarce and lacks an in-depth understanding.

Given that Bi and Sn-containing Mg alloys have exhibited excellent mechanical properties, their corrosion behavior/mechanism is not clear enough. In this study, in order to develop new low-cost, high-performance Mg alloys, Mg-1Bi and Mg-1Sn extruded alloys were prepared, and the focus was on exploring the effects of Bi and Sn on the corrosion behavior/mechanism of Mg alloys. Fortunately, it was found that Sn can play a similar role in the corrosion behavior of Mg alloys to that of RE. This research is conducive to assisting in the in-depth understanding of these two types of potential Mg alloys.

2. Materials and Methods

2.1. Materials Preparation

The raw materials used for preparing the two-component alloy ingots included purity Mg (99.96 wt%), Sn (99.95 wt%), and Bi (99.95 wt%). The melting process was conducted under a mixed protective atmosphere of 99.5% CO₂ and 0.5% SF₆. The raw materials were fully melted at 750 °C. Subsequently, the melt was stirred for 10 min at 740 °C and then held at this temperature for 20 min to purify and eliminate slags/impurities. After this period, the floating impurities were carefully removed using a ladle. Finally, the alloy melt was poured into an iron mold with a water-cooling system, which had a diameter of 90 mm and a length of 800 mm. The actual chemical composition was analyzed by inductively coupled plasma atomic emission spectrometry (ICP-AES, Thermo iCAP 7400, Waltham, MA, USA), as shown in

Table 1. Chemical compositions of the alloys were determined by ICP-AES.

Alloys	Actual Compositions (wt%)
	Mg	Bi	Sn	Fe
Mg-1Bi	Bal.	0.97 ± 0.07	-	0.015 ± 0.002
Mg-1Sn	Bal.	-	1.05 ± 0.06	0.015 ± 0.002

. The ingots were extruded under lubricant-free conditions at 150 °C with the same extrusion ratio of 25:1 and ram speed of 0.3 mm s⁻¹, yielding an extruded bar with a diameter of 15 mm (Figure 1). Before extrusion, turning was required to reduce the diameter of all ingots from 90 mm to 80 mm. The two alloys were labeled as Mg-1Bi alloy and Mg-1Sn alloy.

2.2. Microstructural Characterization

The microstructure of the studied alloys was characterized by a transmission electron microscope (TEM, FEI Talos F200X, Thermo Fisher Scientific, Hillsboro, OR, USA) and a scanning electron microscope (SEM, Thermo Scientific Apreo S, Thermo Fisher Scientific, Hillsboro, OR, USA) equipped with electron backscatter diffraction (EBSD, Oxford Symmetry, Oxford Instruments, Oxford, Oxfordshire, UK). All test samples (except TEM samples) underwent sequential grinding with SiC abrasive papers (grit sizes 320–3000). The grinding process continued until the surfaces achieved planarity without visible scratches, followed by mechanical polishing using a metallographic polishing machine (Shenyang Kejing Auto-Instrument Co., Shenyang, China). The EBSD samples were mechanically polished and then subsequently ion-milled on a Leica RES101 (Leica, Heerbrugg, Switzerland). The EBSD data were collected with a scanning step size of 0.3 μm at 20 kV and were analyzed using AztecCrystal 2.0 software (Version 2.0, Oxford Instruments, Abingdon, Oxfordshire, UK). TEM samples (diameter 3 mm) were prepared by ion beam thinning on a Gatan 695 machine (Gatan, Pleasanton, CA, USA). The X-ray diffraction (XRD, Cu K_α radiation, Panalytical Empyrean, Almelo, Netherlands) was used to analyze the phase composition of the two alloys. Bulk samples with dimensions of 8 mm × 6 mm × 5 mm were tested, and 2θ was scanned in the range of 20° to 80° at a scanning speed of 1° min⁻¹.

2.3. Immersion Experiments

This study employed hydrogen evolution tests and weight loss measurements as the primary immersion experimental methods. Samples with dimensions of 25 mm × 10 mm × 3 mm were immersed at 25 °C for 168 h in 3.5 wt% NaCl solution (artificial sea water). In the hydrogen evolution test, samples were mounted within an inverted glass funnel via nylon ropes and immersed in a 3.5 wt% NaCl solution, while the evolved H₂ gas was captured using an acid burette. The solution was refreshed at 24 h intervals to ensure concentration stability, and the H₂ gas volume was measured every 12 h. The corrosion products of the samples were treated with a chromic acid solution (200 g L⁻¹ CrO₃, 10 g L⁻¹ AgNO₃), subsequently rinsed in deionized water, and finally desiccated under cold air. Weight loss was determined by comparing sample masses before and after immersion testing using an analytical balance with a precision of 0.001 g.

The average corrosion rates via hydrogen evolution (P_H) and weight loss (P_W) were determined using the following equations [23]:

$$P_{\mathrm{H}}={\displaystyle \frac{54.696V_{\mathrm{H}}}{t}}$$

(1)

$$P_{\mathrm{W}}={\displaystyle \frac{87.6W}{t\rho A}}$$

(2)

where the hydrogen evolution rate is symbolized by V_H (mL cm⁻²), the immersion time is indicated by t (h), the mass loss during immersion is signified by W (mg), the surface area of the sample is expressed by A (cm²), and the density of the samples is shown by ρ (g cm⁻³).

The surface morphology and corrosion products formed after 168 h of immersion were characterized via SEM. The cross-sectional elemental surface scans were acquired via the energy dispersive spectrometer (EDS) of the SEM. The composition of corrosion products was characterized via X-ray photoelectron spectroscopy (XPS; ESCALAB 250Xi, Thermo Fisher, Waltham, MA, USA) employing Al K_α radiation (1486.6 eV) at a step resolution of 0.05 eV.

2.4. Electrochemical Measurements

Electrochemical testing was performed with a CS350H electrochemical workstation (Corrtest Instrument Co., Wuhan, China) at 25 °C in 3.5 wt% NaCl solution. A three-electrode cell with an exposed working electrode area of 1 cm² was employed. The saturated calomel electrode (SCE) served as the reference electrode, while platinum mesh acted as the counter electrode. The electrochemical impedance spectroscopy (EIS) measurements were performed after immersion in 3.5 wt% NaCl solution for 2 h. The frequency range spanned from 100 kHz to 0.01 Hz, with an applied sinusoidal perturbation amplitude of ±10 mV. The resulting EIS data were analyzed and fitted using Zsimpwin software. The potentiodynamic polarization was scanned from −500 to +500 mV at 1 mV·s⁻¹ after immersion in 3.5 wt% NaCl solution for 12 h. The tests were performed on the extruded longitudinal surface of specimens. A minimum of three replicates per test ensured experimental reproducibility.

2.5. Localized Potential Distribution

A scanning Kelvin probe force microscope (SKPFM, Dimension Icon, Bruker Corporation, Billerica, MA, USA) was employed to quantify the local Volta potential difference at the interface of the α-Mg matrix and secondary phases. The specimen surface underwent sequential mechanical polishing and electropolishing in a 10% perchloric acid alcohol electrolyte maintained at −20 °C. The SKPFM measurements were performed in dual-scan mode utilizing a magnetically etched silicon probe (MESP-V2, k = 3 N m⁻¹, Bruker Corporation, Fremont, CA, USA). For each alloy, a minimum of three samples were characterized at room temperature, with subsequent data analysis performed using Nanoscope Analysis 1.90 software (Version 1.90, Bruker Corporation, Billerica, MA, USA).

3. Results and Discussion

3.1. Microstructure Characteristics

Figure 2 shows the XRD patterns of these two alloys. It can be inferred that the secondary phase in the Mg-1Bi alloy is mainly the Mg₃Bi₂ phase, while it is mainly the Mg₂Sn phase in the Mg-1Sn alloy, which is consistent with previous reports [24,25,26]. It is noted that the intensity of the Mg-Sn alloy is 2000 counts higher than the actual value. The secondary phases in both alloys were further analyzed via TEM-EDS, and the corresponding results are presented in Figure 3. TEM-EDS elemental mapping quantitatively confirmed the secondary phases as Mg₃Bi₂ in the Mg-1Bi alloy and Mg₂Sn in the Mg-1Sn alloy, complementing XRD phase identification. In addition, STEM morphology analysis further reveals these second-phase particles as extrusion-induced dynamic precipitates. Statistical measurements indicate an average size of 67 nm for Mg₃Bi₂ particles versus 115 nm for Mg₂Sn particles. These fine and randomly distributed precipitates are dynamically precipitated during the extrusion process and cannot serve as nucleation sites to refine grains.

Figure 4 shows the EBSD analysis of the two alloys. Figure 4a,b are the inverse pole figure (IPF) maps, showing the grain morphology/size and distribution of the two alloys. Based on the analysis of the average grain size distribution maps (Figure 4a-3,b-3), both extruded alloys exhibit a significant difference between large and small grain sizes, forming a heterogeneous structure, which is bound to benefit the comprehensive mechanical properties of these alloys [27]. The difference in the average grain sizes of the two alloys (Figure 4a-3,b-3) is slight, at 4.96 μm and 5.25 μm, respectively. In addition, it seems that the dynamic recrystallization degree of the Mg-1Bi alloy is lower than that of the Mg-1Sn alloy, and the kernel average misorientation (KAM) maps also support this (Figure 4a-2,b-2). The KAM map can reflect the dislocation density distribution within the grains. By comparison, obviously, the Mg-1Bi alloy has a higher dislocation density and an uneven dislocation distribution, attributed to its insufficient recrystallization degree, while the Mg-1Sn alloy has a relatively low and uniform dislocation density due to its higher recrystallization degree. Therefore, compared with the addition of Bi, the addition of Sn may promote dynamic recrystallization in Mg alloys during deformation. Figure 4a-1,b-1 show the pole figures of the two alloys. Both alloys exhibit a certain basal texture, and the texture intensity differs just slightly.

3.2. Corrosion Behavior

3.2.1. Corrosion Rate in Immersion Testing

Figure 5 shows the hydrogen evolution and the average corrosion rates calculated after 168 h of immersion in a 3.5 wt% NaCl solution. As shown in Figure 5a, the hydrogen evolution rate of the Mg-1Bi alloy is significantly higher than that of the Mg-1Sn alloy throughout the whole immersion process. The hydrogen evolution and weight loss show similar trends in corrosion rate (Figure 5b); that is, the average corrosion rate of the Mg-1Bi alloy (13.75 ± 1.12 mm y⁻¹ from P_H) is significantly higher than that of the Mg-1Sn alloy (2.83 ± 0.19 mm y⁻¹ from P_H). This indicates that the addition of Bi and Sn has different effects on the corrosion resistance of Mg alloy. Compared with Mg-1Bi alloy, Mg-1Sn alloy exhibits significantly better corrosion resistance.

3.2.2. Corrosion Morphology and Product

As shown in Figure 6, Mg-1Bi and Mg-1Sn alloys exhibit distinct surface morphologies following 168 h of immersion in 3.5 wt% NaCl solution. Figure 6a–c show that the corrosion layer on the Mg-1Bi alloy exhibits a porous and non-uniform structure, which significantly compromises its protective capability for the underlying Mg matrix. In contrast, the corrosion product film on the Mg-1Sn alloy is relatively dense and smooth, demonstrating enhanced protective capability (Figure 6g–i). Furthermore, by comparing the surface morphology of both alloys after removing the corrosion product (Figure 6d–f vs. Figure 6j–l), it is evident that more severe corrosion pits appear on the Mg-1Bi alloy surface, while only relatively slight corrosion occurs on the Mg-1Sn alloy. The observations of corrosion products and corrosion morphology are consistent with the results of the corrosion rate.

Figure 7 presents the cross-sectional morphologies and corresponding EDS mappings of the corrosion films of Mg-1Bi and Mg-1Sn alloys after 168 h immersion in 3.5 wt% NaCl solution. The corrosion film on the surface of the Mg-1Bi alloy is relatively thick, with an average thickness of approximately 56.7 μm, while the corrosion film on the Mg-1Sn alloy is relatively thin, with an average thickness of about 20.9 μm. In addition, it is noted that after the same drying treatment during sample preparation, the corrosion film on the Mg-1Sn alloy has almost no cracks, while the corrosion film on the Mg-1Bi alloy has severe cracks, also suggesting that the corrosion film properties of the Mg-1Sn alloy are better. Cross-sectional EDS mappings show strong Mg and O element signals on the surfaces of both alloys, indicating that the primary corrosion products are MgO/Mg(OH)₂. The key distinction is that a relatively obvious Sn element signal is detected in the corrosion film on the Mg-1Sn alloy, while a very weak Bi element signal is detected in the corrosion film on the Mg-1Bi alloy. It can be speculated that a certain amount of SnO₂ is formed in the corrosion film on the Mg-1Sn alloy [28].

Figure 8 shows the XPS analysis of corrosion product film for Mg-1Bi and Mg-1Sn alloys after 168 h immersion in 3.5 wt% NaCl solution. The Mg 1s spectra of both alloys contain two characteristic peaks corresponding to MgO and Mg(OH)₂. In the O 1s spectrum, the significant difference between the two alloys lies in the presence of a characteristic peak for SnO₂ in the corrosion film on the Mg-1Sn alloy (Figure 8a-2 vs. Figure 8b-2). Notably, no Bi element signal can be detected in the Mg-1Bi alloy surface film (Figure 8a-3), while the Sn 3d spectrum clearly indicates the existence of SnO₂ in the corrosion product layer of the Mg-1Sn alloy (Figure 8b-3), consistent with the cross-sectional EDS mapping results after 168 h of immersion (Figure 7). Unlike Bi-alloying, Sn-alloying enables SnO₂ formation within corrosion films (Figure 8), which densifies the microstructure and enhances barrier properties, ultimately improving corrosion resistance.

3.2.3. Local Potential Distribution

The role of the Mg₃Bi₂ and Mg₂Sn phases in micro-galvanic corrosion was investigated via SKPFM surface potential profiling (Figure 9), including the topographic images, Volta potential maps, and potential line profiles [29]. In the surface potential mapping (Figure 9a-2,b-2), the bright regions correspond to the more positive potential, while dark areas exhibit more negative potentials [30]. The results indicate that both secondary phases, Mg₂Sn and Mg₃Bi₂, in the alloys display a higher potential than the Mg matrix, acting as cathodes during micro-galvanic corrosion. The PD between Mg₃Bi₂ and the Mg matrix measures ~90 mV (Figure 9a-3), while the PD between Mg₂Sn and the Mg matrix is only ~30 mV (Figure 9b-3).

3.2.4. Corrosion Behavior in Electrochemical Testing

Figure 10a shows the polarization curves of two alloys after being immersed in 3.5 wt% NaCl solution at 25 °C for 12 h. The values of corrosion potential (E_corr), corrosion current density (i_corr), cathodic slope (β_c), and breakdown potential of the passive film (E_b), obtained from the polarization curves through the cathodic Tafel extrapolation method, are summarized in

Table 2. Fitting results of polarization curves for the Mg-1Bi alloy and Mg-1Sn alloy.

Alloys	Ecorr (V vs. SCE)	icorr (µA cm−2)	βc (−mV decade−1)	Eb (V vs. SCE)	Eb − Ecorr (mV)
Mg-1Bi	−1.44	105.90	207.51	-	-
Mg-1Sn	−1.48	44.42	289.16	−1.40	76

[31]. Generally, the cathodic polarization branch is dominated by hydrogen evolution kinetics, while the anodic branch corresponds to metal dissolution processes [30,32]. After 12 h of immersion, the anodic branch of the Mg-1Sn alloy shows an E_b of −1.40 V_SCE, while it does not appear on that of the Mg-1Bi alloy. Although the passivation of the Mg-1Sn alloy is relatively weak at this time, compared with the Mg-1Bi alloy, the corrosion film formed on the Mg-1Sn alloy after immersion for 12 h already provides a certain level of protection for the Mg matrix. Through the analysis of cathodic branches, i_corr follows the arrangement of the Mg-1Sn alloy < Mg-1Bi alloy. By comprehensively analyzing the anodic and cathodic branches of the polarization curves of these two alloys, it can be concluded that the corrosion resistance of the Mg-1Sn alloy is superior to that of the Mg-1Bi alloy. Furthermore, the i_corr of the Mg-1Sn alloy is lower than that reported for Sn-containing Mg alloys alloyed with multiple elements [21].

Figure 10b,c show the EIS plots of two alloys after immersion in 3.5 wt% NaCl solution at 25 °C for 2 h. The capacitive loop of the Mg-1Sn alloy on the Nyquist plot is obviously greater than that of the Mg-1Bi alloy, and the differences in peak value in Bode phase plots and impedance modulus at low frequencies in Bode modulus plots are consistent with the Nyquist plots. To further analyze the EIS, the EIS curves were modeled with the equivalent circuit in Figure 10d, and the fitted parameters are summarized in

Table 3. Fitting results of EIS curves for the Mg-1Bi alloy and Mg-1Sn alloy.

Alloys	Rs (ohm cm2)	Rf (ohm cm2)	CPEf		Rct (ohm cm2)	CPEdl		L (H cm−2)	RL (ohm cm2)
			Y0, f* $\left(\text{µ}\mathbf{F}{\mathbf{cm}}^{-2 }{\mathbf{s}}^{\mathbf{n}-1}\right)$	nf		Y0, dl* (µF cm−2 sn−1)	ndl
Mg-1Bi	22.15	11.55	8.93	0.9993	123.3	38.56	0.9819	544.5	21.91
Mg-1Sn	22.57	1373.00	13.06	0.9421	1500.0	1129.00	0.5172	21430.0	714.61

_f^* and _dl^* are the fitting correction factor.

. R_s, R_ct, CPE_dl, R_f, and CPE_f represent solution resistance, charge transfer resistance, double layer capacitance, surface film resistance, and surface film capacitance, respectively. The inductance L in series with the resistance R_L indicates the incompletion of the corrosion product film during the corrosion process [33,34]. Both alloys exhibit significant inductive loops (Figure 10b), indicating that their corrosion product film remains incompletely covered. Here, it should be noted that R_f can express the ability of the surface corrosion product film to resist corrosion, while R_ct can reflect the corrosion tendency on the alloy surface. The much larger R_f value listed in Table 3 indicates that the corrosion product film on the Mg-1Sn alloy provides better protection for the Mg matrix as compared to that of the Mg-1Bi alloy. Furthermore, the much larger R_ct value listed in Table 3 indicates that the Mg-1Sn alloy has a lower corrosion tendency on its surface compared to the Mg-1Bi alloy. Meanwhile, the R_f and R_ct values of the Mg-1Sn alloy are both better than those reported in studies on Sn-containing Mg alloys [21].

In this work, the effects of two promising alloying elements, Bi and Sn, proposed in recent years, on the corrosion resistance of Mg alloys are comparatively studied. Significant differences were found between them; that is, both immersion tests and electrochemical tests indicate that the corrosion resistance of the Mg-1Sn alloy in this study is significantly superior to that of the Mg-1Bi alloy (P_H: 13.75 mm y⁻¹ for Mg-1Bi alloy vs. P_H: 2.83 mm y⁻¹ for Mg-1Sn alloy). We consider this to be mainly correlated with the following aspects.

Influence of the second-phase PD. Due to the high reactivity of Mg itself, the secondary phase in Mg alloys reported so far is mostly the cathode phase relative to the Mg matrix [14,35,36], with only a few being the anode phase [32]. The smaller the PD between the cathode phase and Mg matrix, the weaker the micro-galvanic tendency between them, thereby inhibiting corrosion behavior from the thermodynamic perspective. The secondary phase in the Mg-1Bi alloy is mainly the Mg₃Bi₂ phase, and it is mainly the Mg₂Sn phase in the Mg-1Sn alloy (Figure 2 and Figure 3). Based on the results of SKPFM (Figure 9), both Mg₃Bi₂ and Mg₂Sn are the cathodic phases, whereas the PD of Mg₃Bi₂ in Mg-1Bi alloy is approximately 90 mV, while the PD of Mg₂Sn in Mg-1Sn alloy is just ~30 mV. Therefore, the micro-galvanic corrosion tendency of the Mg-1Sn alloy derived from the secondary phase is significantly lower than that of the Mg-1Bi alloy; this is one of the reasons for the lower corrosion rate of the Mg-1Sn alloy.

Formation of a relatively more protective corrosion product film: It has been reported that the corrosion film of pure Mg generally has a poor protective effect on the Mg matrix [7]. Therefore, improving the compactness of the corrosion film to enhance the protectiveness so as to inhibit the corrosion behavior from the kinetic perspective seems to be even more important, since the secondary phase (i.e., the micro-galvanic corrosion source) is generally inevitable as a strengthening source of Mg alloys [28]. In this study, the appearance of a passivation platform in the polarization curve and a larger R_f in the EIS plot prove that the corrosion film formed on the Mg-1Sn alloy is more stable and protective than that on the Mg-1Bi alloy (Figure 10). Based on SEM observations of the samples in Figure 6 and Figure 7, the corrosion film on the Mg-1Bi alloy is thicker, looser, and uneven, while the corrosion film on the Mg-1Sn alloy is thinner, compact, and relatively uniform, directly showing the higher compactness of the corrosion film formed on the Mg-1Sn alloy. Usually, the Pilling–Bedworth ratio (PBR) serves as a criterion for assessing oxide film protectiveness, with values ranging from 1 to 2 typically yielding adherent and continuous barrier layers [33,37]. In this study, MgO and Mg(OH)₂ are detected in the corrosion films of both alloys, with the associated formation mechanisms described as follows:

Mg → Mg²⁺ + 2e⁻ → (anodic part reaction)

(Reaction 1)

2H₂O + 2e⁻ → 2OH⁻ + H₂ → (cathodic part reaction)

(Reaction 2)

Mg + 2H₂O → Mg(OH)₂ + H₂ → (total reaction)

(Reaction 3)

Subsequently, part of the Mg(OH)₂ dehydrates to form MgO. The PBR of MgO is 0.81, and although that of Mg(OH)₂ is 1.77, the Mg(OH)₂ film tends to form a porous structure, with both suggesting poor protectiveness [38]. The corrosion film on the Mg-1Bi alloy predominantly comprises MgO and Mg(OH)₂ (Figure 7 and Figure 8), which leads to its poor compactness and is responsible for weak protectiveness for the Mg-1Bi alloy matrix. For Mg-1Sn alloy, SnO₂ is also detected in its corrosion film (Figure 7 and Figure 8). Thus, the following reactions are also proposed:

Sn → Sn⁴⁺ + 4e⁻

(Reaction 4)

Sn⁴⁺ + 4OH⁻ → Sn(OH)₄

(Reaction 5)

Then, Sn(OH)₄ dehydrates and undergoes the following reaction:

Sn(OH)₄ → SnO₂ + 2H₂O

(Reaction 6)

It is noted that the PBR of SnO₂ is approximately 1.35 [39], and thus, the presence of SnO₂ is considered to enhance the compactness of the corrosion film on the Mg-1Sn alloy. In addition, the high-valent cations in the corrosion film are expected to elevate the local positive charge density, thereby reducing harmful anions by capturing these negative ions.

Other factors: In the microstructure, the potential nature and the magnitude of PD for the secondary phase, serving as the strengthening phase, have a notable impact on the corrosion resistance of Mg alloys, and in addition, other microstructure factors, such as grain size, dislocation density, and texture, also affect the corrosion resistance to varying degrees. In this study, the average grain size and texture intensity of these two extruded alloys are close (Figure 4), and it is considered that their differences in influencing corrosion could be ignored. Generally, crystal defects such as dislocations have relatively high activity and can accelerate corrosion [40,41]. Mg-1Bi alloy exhibits elevated dislocation density compared to Mg-1Sn alloy (Figure 4), which should also be one of the reasons for its poor corrosion resistance. Subsequently, we will anneal the alloy to reduce dislocation density and make the grain boundaries more stable, thereby further improving its corrosion resistance.

4. Conclusions

In this work, Mg-1Bi and Mg-1Sn extruded alloys were prepared, and their corrosion behavior was comparatively studied. By comparing the influences of Bi and Sn alloying on the corrosion behavior of Mg alloys, the following main conclusions were obtained:

(1)

In immersion and electrochemical experiments, the Mg-1Sn alloy exhibits better corrosion resistance than the Mg-1Bi alloy. The corrosion rate of the Mg-1Sn alloy (P_H: 2.83 ± 0.19 mm y⁻¹) is significantly lower than that of the Mg-1Bi alloy (P_H: 13.75 ± 1.12 mm y⁻¹).

(2)

The main secondary phase of the Mg-1Bi alloy is Mg₃Bi₂, and that of the Mg-1Sn alloy is Mg₂Sn. Their PD values are ~90 mV and ~30 mV, respectively. The lower PD in the Mg-1Sn alloy indicates a weaker tendency of micro-galvanic corrosion, which is one of the reasons for its higher corrosion resistance.

(3)

The addition of Bi has little effect on the corrosion film, while the addition of Sn makes the corrosion film on the Mg-1Sn alloy contain SnO₂. The corrosion film containing SnO₂ is more compact, confirmed by the appearance of a passivation platform and larger Rf, thus improving the protective effect, which is also responsible for the lower corrosion rate of the Mg-1Sn alloy.

(4)

Compared with Bi-containing alloys, Sn alloying demonstrates greater potential for improving the corrosion resistance of Mg alloys, offering a promising pathway to develop high-strength and high-corrosion-resistant Mg alloys.