Effect of Initial Microstructure Prior to Extrusion on the Microstructure and Mechanical Properties of Extruded AZ80 Alloy with a Low Temperature and a Low Ratio

In structural applications, wrought magnesium alloys are very attractive in terms of weight reduction, but their mechanical strength is relatively low at ambient temperatures [1‐4]. In the recent works, researchers find that there is a synergistic effect between the coarse grains with a strong basal texture and submicron-sized grains with a weak texture, resulting in an increase in strength [5‐7]. The idea of achieving the heterogeneous microstructures with high mechanical properties has received significant attention [8‐11]. For example, a bimodal-grained AZ91 alloy prepared by hard-plate rolling exhibits a high ultimate tensile strength (UTS) of 370 MPa and a high elongation of 24% compared with the fine-grained AZ91 alloys [12]. A Mg-7Y-3Zn (wt.%) alloy with multimodal microstructure developed by extrusion shows a high UTS of 385 MPa and an elongation of 7% [13]. However, the strengthening mechanisms for the high strength of Mg alloys with heterogeneous microstructure still lack systematic investigation. The strengthening effect from grain refinement is suggested to be enhanced with increasing volume fraction of recrystallized grains, but the decreasing volume fraction of deformed grains with strong texture will result in a loss of strength. Besides, the size and distribution of second-phase particles also play vital role on the strength of Mg alloys with heterogeneous microstructure. The precipitates not only directly influence the strengthening effect from second phase, but also affect the formation of heterogeneous microstructure [14‐18]. Liao et al. reported that the precipitates with an average size of 1 μm in a Mg-1Gd-1Y-1Zn alloy affected the heterogeneous microstructure by particle stimulated nucleation (PSN) mechanism [15]. Zou et al. reported that the nano-sized precipitates (50 nm) in a Mg-5Zn-1Mn alloy suppressed dynamic recrystallization (DRX), resulting in a bimodal grain size distribution [16]. The influence of precipitates on DRX behavior depends on the deformation temperature because the interaction of dislocations and particles is temperature-dependent [17]. Thus, it is necessary to clarify the role of grain refinement, texture, and second phase on the strength of Mg alloys with different heterogeneous microstructure.

Extrusion is one of the most important processing methods for Mg and its alloys. Extrusion temperature and ratio play critical roles in the grain refinement of extruded Mg alloys. For example, the extrusion temperatures and ratio of AZ91 alloys are commonly controlled over 240 °C and 11, respectively, to achieve a homogeneous micron-scale microstructure [19‐22]. The decreased extrusion temperature and extrusion ratio are suggested to reduce the recrystallization degree of extruded alloys and lead to a heterogeneous microstructure [23‐26]. However, the effect of such a heterogeneous microstructure achieved at low temperatures on the mechanical properties is lacking investigation. Based on these works, we choose a commercial cast AZ80 alloy as initial material and design an extrusion process with low temperature (220 °C) and low ratio (4) to achieve heterogeneous microstructures. The extrusion ratio is chosen based on our previous works on extruded Mg alloys [19, 27], the selected ratio of 4 is suitable to achieve heterogeneous microstructure and ultrafine recrystallized grains. We compare the microstructure and mechanical properties of homogenized-extruded (HE) and cast-extruded (CE) AZ80 alloys. The role of grain refinement, texture, and second phase on the strength of AZ80 alloys are discussed in detail.

In this work, a commercial AZ80 alloy with the actual composition of Mg-7.6Al-0.45Zn-0.27Mn (wt.%) was used. The composition was detected by inductively coupled plasma atomic emission spectrometry (ICP-AES). The cylindrical-shaped samples with a height of 100 mm and a diameter of 80 mm were cut from the billet. Some cylindrical samples were homogenized at 420 °C for 24 h and then quenched at room temperature. These as-cast and as-homogenized samples were heated to 220 °C and then extruded into the rods with a diameter of 40 mm, followed by a rapid quench into water at room temperature. The chamber temperature and mold temperature were both 240 °C. The extrusion ratio was set to 4. The ram velocity for extrusion was 0.2 mm/s. The details of extrusion processes are listed in Table 1. The cast-extruded and homogenized-extruded samples were marked as CE and HE samples, respectively.

Tensile bars with a gauge length of 25 mm and a diameter of 5 mm were tested at a strain rate of 10⁻³/s at room temperature using a Shimadzu Autograph AG-I (100 kN). The tensile direction was parallel to the extrusion direction (ED), and three samples were tested under the same parameters. The microstructure was characterized with an OLS3000 optical microscope (OM), a FEI NOVA400 scanning electron microscope (SEM) and an Oxford HKL Channel 5 electron backscatter diffraction (EBSD) detector, as well as a JEOL JEM-2100F transmission electron microscope (TEM) with energy dispersive spectroscopy (EDS) detector. A mixed solution (2 g picric acid + 5 mL distilled water + 5 mL acetic acid + 4 mL nitric acid + 85 mL alcohol) was used for etching the specimens for OM and SEM observation. An accelerating voltage of 20 kV was used for the SEM and EBSD characterization. EBSD characterization was carried out using a step size of 0.6 μm. Kernel average misorientation map (KAM) is constructed based on the degree of misorientation between a measurement point (kernel) and all its surrounding neighbours [28]. EBSD samples were prepared by mechanical polishing followed by argon ion milling. The TEM characterization under high-angle annular dark field (HAADF) condition and EDS characterization were carried out at 200 kV. Thin-foil samples for TEM observation were prepared using low-energy ion beam thinning. The phase composition of the sample was examined by X-ray diffraction (XRD, Model D/Max 2500PC Rigaku) at 40 kV and a scanning speed of 4°/min. The sizes of grains and precipitates were obtained by Nano Measurer System 1.2 and derived from at least five images.

Figure 1 shows the microstructures of the HE and CE samples. As can be seen, bimodal-grained microstructure forms in the HE and CE samples. In the HE sample, the un-recrystallized elongated grains show an average length of 79 μm and an average width of 27 μm. The volume fraction of elongated grains in the HE sample is 68%. In comparison, elongated grains in the CE sample exhibit a smaller size (with an average length of 57 μm and an average width of 20 μm) but higher volume fraction (89%). Moreover, the elongated grains in the CE sample still have the dendritic morphology.

Figure 2 shows more details of the microstructures for the HE and CE samples. The images are obtained from the planes perpendicular to ED. Both in the HE and CE samples, the second-phase particles are mainly distributed between the coarse grains. In the HE sample, the average size of second-phase particles is 0.45 μm, and the second-phase particles homogenously distribute in the areas of DRXed grains (Figure 2c). These second-phase particles are the dynamic precipitates that form during the extrusion process because the alloying elements have been dissolved during homogenization treatment. Compared to the particles in the HE sample, the second-phase particles in the CE sample are coarse and concentrated (Figure 2f). It indicates that the second-phase particles in the CE sample form mainly during the solidification process of the as-cast alloy and are broken during the extrusion process. The XRD results are shown in Figure 3, which are used for investigating the phase composition of the extruded samples by varying the incident angle (θ). The results prove that the second phases in the HE and CE samples are mainly the Mg₁₇(Al, Zn)₁₂ phase, and the CE sample has a higher content of Mg₁₇(Al, Zn)₁₂ phase. Due to the Mg₁₇(Al, Zn)₁₂ phase in the HE sample is dynamically precipitated during extrusion, which means that the content of precipitates is restricted by the gap of solid solubility between the homogenizing temperature and extrusion temperature. In comparison, the Mg₁₇(Al, Zn)₁₂ phase in the CE sample forms during solidification to room temperature, thus the concentration of solute atoms in the matrix is much higher and numerous second phases can be formed.

Figures 4 and 5 show the TEM images and EDS mapping results of the precipitates in the HE sample. The submicron-sized block-shaped and rod-shaped precipitates can be observed in the HE sample. The block-shaped phase is enriched with Al and Zn elements. The diffraction pattern in Figure 6 proves that the block-shaped phase is the Mg₁₇(Al, Zn)₁₂ phase, which is consistent with the XRD result. The TEM images in Figure 5 show the details of the diffusely distributed nano precipitates. The spherical precipitates have an average size of 18 nm. The rod-shaped precipitates exhibit an average width of 21 nm and an average length of 113 nm. These precipitates are enriched with Al and Mn elements. According to the works on Mg-Al-Zn system alloys [29], the precipitates should be the Al₈Mn₅ phase. As there are no diffraction peaks of Al₈Mn₅ phase can be detected in the XRD result (Figure 3), the volume fraction of Al₈Mn₅ phase should be quite low in the HE sample. In the CE sample, these second-phase particles are micron-sized, submicron-sized, and nano-sized particles, and almost no Al₈Mn₅ phase can be observed (Figure 7).

EBSD results (Figures 8 and 9) show the details of DRXed grains and texture of the HE and CE samples. The images are obtained from the planes perpendicular to ED. The sizes of the DRXed grains in the HE and CE samples are 0.8 μm and 0.7 μm, respectively. The microtextures of deformed and DRXed grains in the HE and CE samples are shown in Figures 2e and f. The deformed grains exhibit a strong basal texture in the HE and CE samples, and the recrystallized grains exhibit a weak texture. Such a heterogeneous texture is commonly observed in the Mg-Al-Zn alloys with heterogeneous recrystallized grain structures [5, 7‐10, 30]. Due to the higher volume fraction of elongated grains, the texture of the CE sample is stronger than that of the HE sample. The KAM result (Figure 10) shows that the orientation gradient within the deformed region is larger than that in the DRXed grains. This is most likely caused by the high density of residual dislocations to accommodate the local lattice strain in the unDRXed grains. Accordingly, the average KAM value of deformed grains in the HE sample (0.64°) is lower than that in the CE sample (0.83°) because of the low volume fraction of unDRXed regions. The density of residual dislocations in the HE sample is thus identified to be lower than that in the CE sample.

The tensile properties of the HE and CE samples are shown in Figure 11. The corresponding yield strength (YS), ultimate tensile strength (UTS) and elongation are also listed. The YS, UTS, and elongation of the HE sample are 284 MPa, 331 MPa, and 4.8%, respectively. In contrast, the CE sample shows a lower YS (218 MPa) and UTS (315 MPa) but a comparable elongation (5.8%). The YS of the HE sample is significantly higher than that of the CE sample, but their ultimate tensile strength is quite close, which results from the stronger work-hardening ability of the CE sample than that of the HE sample.

As mentioned above, numerous tiny and dispersed precipitates exist in the HE sample, which are mainly formed during the extrusion process. It is suggested that strain-induced defects (such as dislocations and vacancies) can act as heterogeneous nucleation sites to promote dynamic precipitation [31]. A high density of dislocations and vacancies is induced during the extrusion process, which promotes the dynamic precipitation of second-phase particles with high number density. Meanwhile, the low extrusion temperature and the short time of the deformation process restricts the growth of these precipitates. According to the works on the effect of second-phase particles on the DRX behaviour of Mg-Al-Zn alloys [17, 32], the phase boundaries between the second-phase particles and Mg matrix are the preferential sites for the nucleation of DRXed grains. In the present work, the number density of dynamically precipitated second-phase particles in the HE sample is significantly higher than that of the coarse compounds in the CE sample (Figure 2). It means that there are more phase boundaries exist in the HE sample than that in the CE sample, which facilitate the nucleation of DRXed grains. Thus, the HE sample achieves a higher volume fraction of DRX grains than the CE sample. Although the effects of the coarse compounds and the dynamically precipitated second-phase particle on the volume fraction of DRXed grains are quite different, the size and microtexture of the DRXed grains in the HE and CE samples are similar. It indicates that the coarse compounds and the dynamical precipitates have a similar influence on the size and orientation of DRXed grains in the present work. In this sense, the formation of ultrafine grains with weak texture in the HE and CE samples should be mainly attributed to the low extrusion temperature.

As shown in Figures 1 and 2, the HE sample has a higher volume fraction of DRXed grains, which will result in a stronger strengthening effect from grain refinement in the HE sample than that in the CE sample. Moreover, the high number density of the dynamic precipitates will also contribute to the strength of the HE sample. The formation of nano-size Al₈Mn₅ phase also improve the strength of the HE sample. Although the HE and CE samples have different grain size, texture, precipitates and DRX degree, the high density of dislocations that form during the low-temperature extrusion process result in the quite low elongation for both. It should be noted that the basal texture of the CE samples is stronger than that of the HE samples, and the residual dislocation density of the CE samples is higher than that of the HE samples, which will result in higher strengthening effect. However, the YS and UTS of the HE sample are still higher than those of the CE sample, indicating that the strengthening effects from grain refinement and dynamical precipitates are stronger than the strengthening effect from the texture in the extruded AZ80 alloys.

In the present work, the microstructure and mechanical properties of a commercial AZ80 alloy extruded at low temperature and low ratio are investigated. The main conclusions are: