2.1. Alloy Preparation and Casting

The alloys were prepared using a high-frequency induction furnace under a controlled SF₆ + CO₂ protective atmosphere to prevent Li oxidation. Commercial high-purity Mg, Li, Al, Zn and Bi ingots (99.9% purity) were used as starting materials. The chemical composition and calculated densities of the investigated alloys are listed in Table 1. The Mg ingot was melted first, followed by the sequential addition of Al, Zn, and Bi. To minimize oxidative loss, Li was introduced into the melt using a controlled charging method designed to limit its exposure to the ambient atmosphere. After alloying, the melt was stabilized and thoroughly stirred to ensure chemical homogeneity. During the pouring process, a continuous flow of SF₆ + CO₂ gas was maintained over the crucible and the mold entrance to establish a localized protective curtain, thereby preventing direct contact between the molten metal and the surrounding air. The melt was subsequently gravity-cast into a preheated steel mold (ø75 × 280 mm, 200 °C). Although Mg-Li alloys are typically processed under vacuum or inert atmospheres due to the high reactivity of Li, the present alloys were successfully fabricated via gravity casting in air by ensuring a consistent protective gas cover throughout the entire melting and pouring sequence. Under the controlled melting and casting conditions employed in this study, Li oxidation and evaporation were effectively suppressed, as evidenced by the reproducible phase constitution and consistent microstructural features observed across all alloy compositions.

2.2. Hot Extrusion

For hot extrusion, the as-cast billets were machined to ø75 × 100 mm and preheated to 200 °C for 2 h. Extrusion was conducted with a container temperature of 350 °C and a die temperature of 200 °C. The extrusion ratio was 14:1, and the exit speed was maintained at approximately 26 m/min, producing plate-type extruded specimens with a thickness of 4 mm.

2.3. Microstructural Characterization

Microstructural observations were carried out on both as-cast and as-extruded specimens. Samples were prepared through standard metallographic procedures, including grinding up to #4000 grit and final polishing with 1 μm diamond suspension and colloidal silica. The microstructures were examined on the plane parallel to the extrusion direction (ED) using a field-emission scanning electron microscope (FE-SEM, JEOL JSM-7800F) operated at 15 kV. Backscattered electron (BSE) mode was utilized to clearly distinguish the constituent phases based on atomic number contrast. The phase fractions of the α-Mg and β-Li phases were quantitatively determined from multiple representative BSE images using image analysis software (i-Solution DT-L, IMT). In this analysis, the brighter regions were identified as the α-Mg phase, while the darker regions corresponded to the β-Li phases. Phase identification was performed via X-ray diffraction (XRD, Bruker D8 ADVANCE) over a 2θ range of 10-90°.

2.4. Mechanical Testing

Room-temperature tensile tests were performed using a universal testing machine in accordance with ASTM E8/E8M. Tensile specimens were machined along the extrusion direction (ED) with a gauge length of 30 mm. The tests were conducted at a constant crosshead speed of 1.8 mm/min (initial strain rate of 1.0 × 10^-3s^-1). At least three specimens were tested for each alloy to ensure statistical reliability.

3.1 Phase identification and as-cast Microstructure

Fig. 1 shows the X-ray diffraction (XRD) patterns of the as-cast Mg-Li alloys with different Al and Bi additions. All of the alloys exhibited diffraction peaks corresponding to the α-Mg (HCP) and β-Li (BCC) phases, confirming the formation of a dual-phase microstructure. Notably, in Alloys 1-3(Bi-free alloys), a distinct shift of the diffraction peaks toward higher 2θ angles was observed as the Al content increased from 4 to 8 wt.%. This shift is primarily attributed to the lattice contraction of the matrix caused by the substitution of Mg atoms with Al atoms, which possess a smaller atomic radius (1.43 Å for Al vs. 1.60 Å for Mg). In contrast, for the Bi-added alloys (Alloys 4-6), this peak shift was significantly suppressed despite the higher Al concentrations. This suggests that Bi addition effectively stabilizes the matrix lattice. It is inferred that the strong chemical affinity between Mg and Bi promotes the preferential formation of Mg₃Bi₂ intermetallic compounds, which consumes Mg from the matrix and mitigates the lattice distortion typically induced by excessive Al dissolution. The diffraction peaks corresponding to Mg₃Bi₂ can be clearly observed in Alloys 4-6, even at a relatively low Bi concentration of 2 wt.%, due to the large electronegativity difference between Mg and Bi^{[16, 18, 19]}. According to previous thermodynamic studies, the formation enthalpy of Mg₃Bi₂ is significantly lower than that of other potential phases, promoting its preferential precipitation in the Mg matrix^{[18, 19]}.

The result confirms that Bi preferentially reacts with Mg to form Mg₃Bi₂ rather than dissolving into the α-Mg or β-Li matrix. This preferential formation of Mg₃Bi₂ is a key factor in the alloy design, as it effectively consumes Mg from the matrix, thereby potentially influencing the α/β phase equilibrium.

Although the overall phase constitution remained similar among the alloys, the relative intensities of the α-Mg and β-Li diffraction peaks varied with Al and Bi additions, indicating changes in the phase fractions. Specifically, the variation in peak intensities suggests that Al promotes the α-Mg phase, while Bi addition counteracts this effect by altering the effective matrix composition. These tendencies are in good agreement with the quantitative phase analysis results presented in Figs. 4 and 5. For these reasons in this study XRD analysis was primarily employed to identify the constituent phases and to provide supporting evidence for the subsequent microstructural and mechanical property analyses.

Figs. 2 and 3 show the BSE-SEM microstructures of the as-cast Mg-Li alloys with varying Al and Bi additions at low and high magnifications, respectively. All of the alloys exhibited a typical dual-phase microstructure composed of α-Mg and β-Li phases, which is consistent with the XRD results (Fig. 1). In the BSE mode, the α-Mg phase appears as relatively bright regions, while the β-Li phase forms a darker matrix; this contrast occurs because the average atomic number of the α-Mg phase (rich in Mg, Z=12) is higher than that of the β-Li phase (rich in Li, Z=3), leading to higher electron backscattering efficiency in the α-Mg regions. As the Al content increased from Alloy 1 to Alloy 3, a gradual increase in the fraction of the α-Mg phase was observed, accompanied by a corresponding decrease in the β-Li phase fraction.

In addition, Al-containing intermetallic compounds, identified as the AlLi phase, were locally observed at the α/β phase boundaries and within the β-Li matrix. The formation of the AlLi phase becomes more pronounced with increasing Al content, indicating that Al actively participates in phase stabilization and intermetallic formation during solidification. The localized precipitation of AlLi at the phase boundaries suggests that Al reaches its solubility limit in the matrix during the final stages of solidification, triggering the formation of these intermetallics^{[5, 10, 14]}.

High-magnification images in Fig. 3 further reveal the presence of fine Zn-rich phases distributed within the β-Li matrix. Although these Zn-rich phases are not clearly detected in the XRD patterns due to their limited volume fraction and nanoscale size, their preferential formation in the β-Li phase is consistent with the higher solubility of Zn in the BCC β-Li structure compared to the HCP α-Mg structure. These Zn-rich features are expected to contribute to local strengthening of the β-Li matrix through Orowan-type or solid-solution strengthening mechanisms^{[11, 12]}.

With increasing Al content, the morphology of the α-Mg phase evolved from relatively continuous regions to a serrated or fragmented structure, particularly in Alloy 2 and Alloy 3. This serrated α-Mg morphology is likely associated with the increased formation of AlLi intermetallic compounds, which can locally interrupt the continuous growth of the α-Mg phase during solidification by acting as physical barriers or by altering the local interfacial energy. Similar morphological transitions of the α-Mg phase, attributed to changes in phase distribution and interfacial stability induced by Al-containing intermetallic formation, have been reported in previous studies on Mg-Li-Al alloys^{[10-14, 20-22]}.

In contrast, the Bi-containing alloys (Alloy 4-6) exhibited distinct microstructural characteristics. The formation of Mg₃Bi₂ intermetallic particles can be clearly observed, mainly distributed along phase boundaries and within the β-Li regions. The precipitation of Mg₃Bi₂ is strategically significant as it is accompanied by a noticeable reduction in the α-Mg phase fraction and a relative increase in the β-Li phase fraction. This confirms that the Bi addition effectively modulates the phase balance by preferentially consuming Mg from the matrix to form thermodynamically stable Mg₃Bi₂ intermetallics, thereby shifting the alloy’s position toward the Li-rich side of the phase diagram.

To quantitatively assess these microstructural changes, image analysis was conducted on the as-cast microstructures. The spatial distribution of each phase is illustrated in Fig. 4, while the corresponding phase fractions are summarized in Fig. 5 and Table 2. As shown in Fig. 5, the increasing Al content leads to higher fractions of α-Mg and AlLi phases, whereas Bi addition results in a decrease in the α-Mg fraction and an increase in the β-Li and Mg₃Bi₂ fractions.

Fig. 1. X-ray diffraction (XRD) patterns of the as-cast Mg-Li alloys with different Al and Bi contents.

Fig. 2. BSE-SEM microstructures of as-cast Mg-Li alloys with varying Al and Bi additions. (a) Alloy 1, (b) Alloy 2, (c) Alloy 3, (d) Alloy 4, (e) Alloy 5, and (f) Alloy 6.

Fig. 3. High-magnification BSE-SEM images showing second-particles in the as-cast Mg-Li alloys. (a) Alloy 1, (b) Alloy 2, (c) Alloy 3, (d) Alloy 4, (e) Alloy 5, and (f) Alloy 6.

Fig. 4. Image-analyzed phase maps of as-cast Mg-Li alloys with varying Al and Bi additions. (a) Alloy 1, (b) Alloy 2, (c) Alloy 3, (d) Alloy 4, (e) Alloy 5, and (f) Alloy 6.

Fig. 5. Phase fraction of α-Mg, β-Li, and second phases in as-cast Mg-Li alloys with varying Al and Bi additions.

3.2. Microstructural Evolution and Phase Distribution after Extrusion

Fig. 6 shows the BSE-SEM microstructures of the extruded Mg-Li alloys with varying Al and Bi additions. Compared with the as-cast condition, all of the extruded alloys exhibited significantly refined and elongated microstructural features aligned along the extrusion direction (ED), reflecting the severe plastic deformation introduced during hot extrusion. The intensive shear strain during the high-speed extrusion process (26 m/min) promoted the fragmentation of the eutectic-like structures observed in the as-cast state, resulting in a more streamlined phase distribution.

All of the extruded alloys retained a dual-phase microstructure composed of α-Mg and β-Li phases. However, the morphology and distribution of these phases were markedly altered by extrusion. The α-Mg phase appears as an elongated or fragmented regions oriented parallel to the extrusion, while the β-Li phase forms a relatively continuous matrix. This morphological evolution is attributed to the differential flow stress between the HCP α-phase and the BCC β-phase, where the latter undergoes more extensive plastic flow^{[7, 20]}. Second-phase particles, including AlLi and Mg₃Bi₂ particles, are observed as fine and uniformly dispersed features, indicating that extrusion effectively breaks up and redistributes the intermetallic phases formed during solidification^{[10, 20]}.

Figs. 7 and 8 present the phase fraction variations of the extruded alloys as a function of Al and Bi additions, while the corresponding numerical values are summarized in Table 3. With increasing Al content from Alloy 1 to Alloy 3, the fraction of the α-Mg phase increased (from 68.23% to 75.21%), accompanied by a reduction in the β-Li phase fraction (from 28.72% to 20.02%). This trend is consistent with that observed in the as-cast alloys, suggesting that Al addition continues to promote α-Mg stabilization even after thermomechanical processing. In addition, the fraction of the AlLi phase increases with increasing Al content, reflecting enhanced intermetallic formation and retention during extrusion.

In contrast, the Bi-containing extruded alloys (Alloy 4-6) exhibited a distinct phase fraction balance compared with their Bi-free counterparts. As shown in Figs. 7 and 8 and Table 3, Bi addition leads to a decrease in the α-Mg phase fraction and a corresponding increase in the β-Li phase fraction. For instance, the β-Li fraction significantly increased from 23.02% in Alloy 2 (Bi-free) to 31.68% in Alloy 5 (Bi-added). Furthermore, the presence of Mg₃Bi₂ as a stable second phase was confirmed in the Bi-added alloys. These results indicate that Bi addition alters the phase balance by preferentially consuming Mg to form Mg₃Bi₂ thereby stabilizing the β-Li phase during extrusion. This phase modulation is critical as it potentially increases the volume fraction of the more ductile β-Li phase, which can accommodate greater strain during subsequent deformation.

Fig. 6. BSE-SEM microstructures of as-extruded Mg-Li alloys with varying Al and Bi additions. (a) Alloy 1, (b) Alloy 2, (c) Alloy 3, (d) Alloy 4, (e) Alloy 5, and (f) Alloy 6.

Fig. 7. Image-analyzed phase maps of as-extruded Mg-Li alloys with varying Al and Bi additions. (a) Alloy 1, (b) Alloy 2, (c) Alloy 3, (d) Alloy 4, (e) Alloy 5, and (f) Alloy 6.

Fig. 8. Phase fraction of α-Mg, β-Li, and second phases in as-extruded Mg-Li alloys with varying Al and Bi additions.

3.3. Mechanical Properties

Fig. 9 shows the engineering stress-strain curves of the extruded Mg-Li alloys with varying Al and Bi additions, and the corresponding tensile properties are summarized in Table 4. All of the extruded alloys exhibited continuous yielding behavior and relatively large elongation, which is characteristic of dual-phase Mg-Li alloys containing the ductile β-Li phase^{[8, 13]}.

For the Bi-free alloys (Alloy 1-3), increasing Al content led to a clear increase in yield strength (YS) and ultimate tensile strength (UTS), accompanied by a gradual reduction in elongation. For instance, as the Al content increased from 4 wt.% (Alloy 1) to 8 wt.% (Alloy 3), the YS increases from 159.35MPa to 191.08MPa, while the elongation dropped significantly from 22.15% to 7.25%. This strength enhancement can be attributed to the combined effects of solid-solution strengthening by Al, the increased fraction of the hard α-Mg phase, and precipitation strengthening associated with the formation of AlLi intermetallic compounds. Meanwhile, the reduction in elongation is closely related to the increased α-Mg fraction and the serrated or fragmented morphology of the α-Mg phase, which limits the continuity of the softer β-Li matrix and reduces the overall deformation accommodation capability^{[5, 9, 22]}.

In contrast, the Bi-containing alloys (Alloy 4-6) exhibited a different mechanical response. Although increasing Al content within the Bi-added alloys still resulted in higher strength and reduced ductility, a direct comparison between Bi-free and Bi-containing alloys with identical Al contents (Alloy 1 vs. 4, Alloy 2 vs. 5, and Alloy 3 vs. 6) reveals that Bi addition leads to simultaneous improvements in both strength and elongation. Notably, Alloy 5 (with Bi) exhibited a YS of 205.52 MPa and an elongation of 16.75%, which are significantly higher than the YS of 178.85 MPa and elongation of 9.01% observed in Alloy 2 (without Bi). This remarkable strength-ductility synergy is attributed to the combined effects of phase fraction modification and microstructural evolution induced by Bi addition^{[13, 18]}.

Specifically, Bi addition promotes the formation of Mg₃Bi₂ particles, which are uniformly distributed in the matrix. These hard particles contribute to precipitation strengthening by effectively pinning dislocation movement during plastic deformation. At the same time, the reduction in the α-Mg phase fraction and the corresponding increase in the ductile β-Li phase fraction resulting from the preferential consumption of Mg by Bi provides a more continuous and deformable matrix. This restored β-Li fraction plays a crucial role in accommodating greater strain and delaying strain localization, and is thereby responsible for the significant recovery in ductility observed in the Bi-added alloys^{[15, 18, 23]}.

However, it is worth noting when comparing Alloy 5 directly to Alloy 1 that the elongation of Alloy 5 is somewhat reduced, despite having a more favorable proportion of the ductile β-Li phase. This apparent discrepancy is primarily driven by the significant increase in the total volume fraction of secondary phases. In Alloy 5, the higher combined additions of Al and Bi result in the dense co-precipitation of both AlLi and Mg₃Bi₂ particles.

Although quantifying the exact experimental phase ratio of these two individual precipitates is challenging, thermodynamic data clearly support the prominent formation of Mg₃Bi₂. The formation enthalpy of Mg₃Bi₂ (approximately -36kJ/mol of atoms) is highly negative and significantly lower than that of AlLi and other potential Mg-Al phases, making it thermodynamically highly stable. While the increased β-Li phase enhances the intrinsic ductility of the matrix, these abundant hard precipitates act as strong stress concentration sites at the particle-matrix interfaces during plastic deformation. Consequently, the localized embrittlement and early microcrack initiation induced by the dense secondary phases prevent the material from reaching its full elongation potential, overshadowing the ductilizing effect of the matrix.

Overall, the mechanical properties of the extruded Mg-Li alloys are governed by the combined effects of phase fraction, the distribution of second-phase particles, and solid-solution strengthening. Al addition primarily enhances strength at the expense of ductility through α-Mg stabilization and intermetallic formation, whereas Bi addition effectively compensates for the ductility loss by increasing the β-Li fraction via thermodynamic phase control. These results demonstrate that synergistic control of alloy composition and microstructure is a promising strategy for optimizing the mechanical performance of ultralight Mg-Li alloys.

Fig. 9. Engineering stress-strain curves of the as-extruded Mg-Li alloys with varying Al and Bi additions.