Effect of Ce Addition on Modifying the Microstructure and Achieving a High Elongation with a Relatively High Strength of As-Extruded AZ80 Magnesium Alloy

Forming magnesium alloys with rare earth elements (La, Gd, Nd, Y, Ce) is a routine method for modifying their microstructure and properties. In the present work, the effect of Ce addition on the microstructure evolution and the mechanical properties of as-extruded Mg-8Al-0.5Zn (AZ80) alloy was investigated. All of the extruded AZ80-xCe (x = 0, 0.2, 0.8 and 1.4 wt %) alloys exhibited equiaxed grains formed by fully dynamic recrystallization, and the grain size of the extruded AZ80 alloy was remarkably reduced by ~56.7% with the addition of 1.4 wt % Ce. Furthermore, the bulk-shaped Al4Ce phase formed when Ce was first added, with the Ce content rising to 0.8 wt % or higher, and Al4Ce particles in both the nano- and micron sizees were well distributed in the primary α-Mg matrix. The area fraction of the Al4Ce particles expanded with increasing Ce content, providing more nuclei for dynamic recrystallization, which could contribute to the grain refinement. The results of the tensile tests in this study showed that Ce addition effectively improved the room temperature formability of the as-extruded AZ80 alloy, without sacrificing strength. The significantly improved mechanical properties were ascribed to excellent grain refinement, weakened texture strength, an increased Schmid factor, and a reduced area fraction of low-angle grain boundaries, all resulting from Ce addition to the as-extruded AZ80 alloy. The contribution of the nano-Al4Ce precipitates on improving the mechanical properties was also discussed in this paper.


Introduction
Reducing vehicle weight is an important strategy for boosting fuel economy and depressing carbon dioxide emissions. Magnesium and its alloys, as the lightest metallic structural materials, have been of great interest for automotive industries [1]. However, the widespread adoption of Mg alloys has been conspicuously limited by its unsatisfied room temperature formability. This is due to a lack of active slip systems, resulting from the hexagonal close-packed (HCP) crystal structure of Mg alloys [2]. Furthermore, most of the commercial products fabricated from magnesium alloys are nowadays produced with die casting, whereas the use of wrought magnesium alloys is very uncommon, because of its limited formability, despite wrought Mg alloys performing better in terms of mechanical properties than the casting Mg alloys [3]. Therefore, in order to expand the industrial and Mg-28 wt % Ce master alloys. The melting procedure was carried out in an electric resistance furnace under the protection of an inert atmosphere with a mixture of CO 2 and SF 6 at around 700 • C, and then the melts were poured into a steel die preheated to 200 • C. The dimensions of each cylindrical billet were 94 mm in diameter and 200 mm in length. The chemical compositions of each billet were measured by an optical spectrum analyzer (ARL 4460, Ecublens, Switzerland); each condition has been tested over three times, and for each time, the date result was obtained by measuring the values of 3-4 points on the tested specimens and calculating their average number. The results are presented in Table 1, which shows that the measured compositions were very close to the nominal ones. The cast billets were then machined into cylinders of 90 mm in diameter and 100 mm in length. In order to improve the alloys' hot deformation performance, a homogenized treatment is necessary to make the non-equilibrium structure of the cast billets uniform in composition by atom diffusion at a high temperature. According to previous experience in our laboratory [16], the billets were homogenized at 415 • C for 20 h, followed by water-quenching. Before extrusion, the billets and the extrusion container were preheated to 430 • C, and then the extrusion was performed at a speed of 0.4-0.6 m/min, and an extrusion ratio of about 32:1. Figure 1 illustrates the schematic diagram of the extrusion process in this work. furnace under the protection of an inert atmosphere with a mixture of CO2 and SF6 at around 700 °C, and then the melts were poured into a steel die preheated to 200 °C. The dimensions of each cylindrical billet were 94 mm in diameter and 200 mm in length. The chemical compositions of each billet were measured by an optical spectrum analyzer (ARL 4460, Ecublens, Switzerland); each condition has been tested over three times, and for each time, the date result was obtained by measuring the values of 3-4 points on the tested specimens and calculating their average number.
The results are presented in Table 1, which shows that the measured compositions were very close to the nominal ones. The cast billets were then machined into cylinders of 90 mm in diameter and 100 mm in length. In order to improve the alloys' hot deformation performance, a homogenized treatment is necessary to make the non-equilibrium structure of the cast billets uniform in composition by atom diffusion at a high temperature. According to previous experience in our laboratory [16], the billets were homogenized at 415 °C for 20 h, followed by water-quenching. Before extrusion, the billets and the extrusion container were preheated to 430 °C, and then the extrusion was performed at a speed of 0.4-0.6 m/min, and an extrusion ratio of about 32:1. Figure 1 illustrates the schematic diagram of the extrusion process in this work. Bal. Phase analyses were conducted by an X-ray diffractometer (XRD; 18KW/D/Max 2500PC, Rigaku, Tokyo, Japan) with Cu Kα radiation at a voltage of 40 kV at a scanning rate of 4° min -1 , with the sample tilt angle ranging from 20° to 80°. An optical microscope (OM; Carl ZEISS Axio-Imager-A2m, Göttingen, Germany), a scanning electron microscope (SEM; Carl ZEISS EVO-18, Göttingen, Germany) equipped with an energy dispersive spectrometer (EDS; Oxford INCA-X-Max, England, UK), and a field-emission transmission electron microscope (TEM; JEOL 200KV/JEM-2100F, Akishima, Japan) were used to characterize the microstructures of the extruded samples. In addition, at least eight OM or TEM micrographs for each condition were collected, to obtain the average grain size of the extruded samples, as well as the average size of the second phases by using the Nano Measurer System 1.2 software. The grain structures and microtextures were analyzed by using a SEM (VEGA3 XMU, TESCAN, Brno, Czech) equipped with an electron backscattered diffraction (EBSD) detector (Oxford Aztec-Nordlys-Nano, England, UK). The textures were measured with a step size determined by the grain size, and the Aztec and Channel 5.0 software (Aztec Software Private Limited, Bengaluru, Karnataka) were utilized to manipulate and to collect the date. The metallographic samples for the OM and SEM observations were firstly ground with sandpapers, and mechanical polished with 0.5 μm diamond pastes, and then chemically etched in acetic picric solution for about 30 s. The picric solution was composed of 5 g picric acid, 5 mL acetic acid, 10 mL distilled Phase analyses were conducted by an X-ray diffractometer (XRD; 18KW/D/Max 2500PC, Rigaku, Tokyo, Japan) with Cu Kα radiation at a voltage of 40 kV at a scanning rate of 4 • min −1 , with the sample tilt angle ranging from 20 • to 80 • . An optical microscope (OM; Carl ZEISS Axio-Imager-A2m, Göttingen, Germany), a scanning electron microscope (SEM; Carl ZEISS EVO-18, Göttingen, Germany) equipped with an energy dispersive spectrometer (EDS; Oxford INCA-X-Max, England, UK), and a field-emission transmission electron microscope (TEM; JEOL 200KV/JEM-2100F, Akishima, Japan) were used to characterize the microstructures of the extruded samples. In addition, at least eight OM or TEM micrographs for each condition were collected, to obtain the average grain size of the extruded samples, as well as the average size of the second phases by using the Nano Measurer System 1.2 software. The grain structures and microtextures were analyzed by using a SEM (VEGA3 XMU, TESCAN, Brno, Czech) equipped with an electron backscattered diffraction (EBSD) detector (Oxford Aztec-Nordlys-Nano, England, UK). The textures were measured with a step size determined by the grain size, and the Aztec and Channel 5.0 software (Aztec Software Private Limited, Bengaluru, Karnataka) were utilized to manipulate and to collect the date. The metallographic samples for the OM and SEM observations were firstly ground with sandpapers, and mechanical polished with 0.5 µm diamond pastes, and then chemically etched in acetic picric solution for about 30 s. The picric solution was composed of 5 g picric acid, 5 mL acetic acid, 10 mL distilled water, and 80 mL ethanol, of which the picric acid was manufactured by China Taishan Yueqiao Reagent Plastics Company Limited, and the acetic acid and ethanol were manufactured by Beijing Chemical Works. Samples for TEM observations were firstly ground to about 70 µm in thickness, and this was followed by ion thinning for about 2 h via ion beam milling (Leica EM-RES101, Solms, Germany). Specimens for EBSD analyses were electro-polished with a specific commercial AC2 electrolyte solution at 20 V for 25-30 s after mechanical polishing.
Tensile tests were carried out via a tensile testing machine (INSTRON 5869, Norwood, MA, USA) at room temperature under a strain rate of 1.0 × 10 −3 s −1 . Tensile specimens with a gauge size of 10 × 3.5 × 1.8 mm were machined, with the load axis parallel to the extrusion direction (ED). Moreover, the values that appeared in this work were the average of at least 3-5 sets of measured dates, and the stress-strain curves presented were good representatives. Further, the grains of Ce-containing alloys are much finer than those of the alloy with no Ce addition, When the Ce content is gradually added from 0 to 1.4 wt %, the grain size of the as-extruded AZ80 alloy decreases obviously from~34.2 µm to~14.8 µm. Tensile tests were carried out via a tensile testing machine (INSTRON 5869, Massachusetts, USA) at room temperature under a strain rate of 1.0 × 10 −3 s −1 . Tensile specimens with a gauge size of 10 × 3.5 × 1.8 mm were machined, with the load axis parallel to the extrusion direction (ED). Moreover, the values that appeared in this work were the average of at least 3-5 sets of measured dates, and the stress-strain curves presented were good representatives. Further, the grains of Ce-containing alloys are much finer than those of the alloy with no Ce addition, When the Ce content is gradually added from 0 to 1.4 wt %, the grain size of the as-extruded AZ80 alloy decreases obviously from ~34.2 μm to ~14.8 μm. By analyzing the XRD results of the as-extruded AZ80-xCe (x = 0, 0.2, 0.8, and 1.4 wt %) alloys ( Figure 3), it can be indicated that there were two phases in the as-extruded AZ80 alloy: primary α-Mg and β-Mg17Al12; a new phase, Al4Ce, emerged with the addition of Ce. Due to the low content of Zn in these alloys, there were no Zn-containing phases detected. Probably, most of the Zn dissolved into the α-Mg matrix uniformly. It is reported that in the AZ series alloys, if the Zn/Al ratio exceeds 1/3, Mg-Al-Zn ternary phases may appear [17]. In this work, Zn/Al ratio was ~1/16, far less than 1/3, By analyzing the XRD results of the as-extruded AZ80-xCe (x = 0, 0.2, 0.8, and 1.4 wt %) alloys ( Figure 3), it can be indicated that there were two phases in the as-extruded AZ80 alloy: primary α-Mg and β-Mg 17 Al 12 ; a new phase, Al 4 Ce, emerged with the addition of Ce. Due to the low content of Zn in these alloys, there were no Zn-containing phases detected. Probably, most of the Zn dissolved into the α-Mg matrix uniformly. It is reported that in the AZ series alloys, if the Zn/Al ratio exceeds 1/3, Mg-Al-Zn ternary phases may appear [17]. In this work, Zn/Al ratio was~1/16, far less than 1/3, and hence showing no Mg-Al-Zn phase. The results also showed that with the increase of Ce addition, the intensity of diffraction peaks corresponding to the Al 4 Ce phase rose, while the intensities of the diffraction peaks of β-Mg 17 Al 12 phase declined. This was due to the consumption of Al during the formation of the Al 4 Ce phase, and thus causing the reduction of the β-Mg 17 Al 12 phase. From the diffraction peaks at the angle of~33 degrees,~58 degrees, and~71 degrees, it can be seen that the intensity of the diffraction peaks corresponding to the α-Mg matrix declined, while the full width at half maxima (FWHM) became broader with the increase of Ce content. This was mainly because the grains of the α-Mg matrix were greatly refined with the addition of Ce (Figure 2), and a smaller grain size often results in more serious X-ray dispersion and increases the FWHM. The decline of the intensity of the α-Mg matrix's diffraction peaks can also indicate that the texture of α-Mg matrix was weakened. Furthermore, it is worth mentioning that the phase peaks of the XRD spectrum shifted, which might be attributed to the changes of the crystal lattice constant caused by the extrusion processes. and hence showing no Mg-Al-Zn phase. The results also showed that with the increase of Ce addition, the intensity of diffraction peaks corresponding to the Al4Ce phase rose, while the intensities of the diffraction peaks of β-Mg17Al12 phase declined. This was due to the consumption of Al during the formation of the Al4Ce phase, and thus causing the reduction of the β-Mg17Al12 phase. From the diffraction peaks at the angle of ~33 degrees, ~58 degrees, and ~71 degrees, it can be seen that the intensity of the diffraction peaks corresponding to the α-Mg matrix declined, while the full width at half maxima (FWHM) became broader with the increase of Ce content. This was mainly because the grains of the α-Mg matrix were greatly refined with the addition of Ce (Figure 2), and a smaller grain size often results in more serious X-ray dispersion and increases the FWHM. The decline of the intensity of the α-Mg matrix's diffraction peaks can also indicate that the texture of α-Mg matrix was weakened. Furthermore, it is worth mentioning that the phase peaks of the XRD spectrum shifted, which might be attributed to the changes of the crystal lattice constant caused by the extrusion processes.   Figure 3b-d, it can be deduced that these second phase particles may be Al4Ce. From SEM micrographs, it can be clearly seen that Mg17Al12 particles often segregated near the serrated grain boundaries, while they rarely appeared inside the grains, as well as being adjacent to the flat grain boundaries. The reason can be that the serrated GBs will provide more nucleation for the Mg17Al12 particles [12]. The corresponding backscattered electron (BSE) images of Figure 4a-d are shown in Figure 4m-p, respectively. The Al4Ce particles in these images are revealed as bright white, and their area fractions are estimated by the software Image-Pro Plus. One can see that when Ce is added, Al4Ce particles in bulk shape with a size of below 10 μm show up. Small amounts of Al4Ce particles distributed along the ED in streamline can be seen in the alloy containing 0.2 wt % Ce. When the Ce content is 0.8 wt % or higher, the Al4Ce particles are well-dispersed in the Mg matrix. Meanwhile, a gradual rise in the area fraction of the Al4Ce phases can be observed. During hot extrusion, Al4Ce particles can act as nuclei for the DRX, and so it is verified that the increase in the amount of the thermodynamically stable Al4Ce particles with an increasing content of Ce addition can provide more nucleation sites for the DRX grains, and thus achieve grain refinement [18]. The excellent grainrefining effect of the Ce addition can also be concluded in the results depicted in Figure 2a-d.   Figure 3b-d, it can be deduced that these second phase particles may be Al 4 Ce. From SEM micrographs, it can be clearly seen that Mg 17 Al 12 particles often segregated near the serrated grain boundaries, while they rarely appeared inside the grains, as well as being adjacent to the flat grain boundaries. The reason can be that the serrated GBs will provide more nucleation for the Mg 17 Al 12 particles [12]. The corresponding backscattered electron (BSE) images of Figure 4a-d are shown in Figure 4m-p, respectively. The Al 4 Ce particles in these images are revealed as bright white, and their area fractions are estimated by the software Image-Pro Plus. One can see that when Ce is added, Al 4 Ce particles in bulk shape with a size of below 10 µm show up. Small amounts of Al 4 Ce particles distributed along the ED in streamline can be seen in the alloy containing 0.2 wt % Ce. When the Ce content is 0.8 wt % or higher, the Al 4 Ce particles are well-dispersed in the Mg matrix. Meanwhile, a gradual rise in the area fraction of the Al 4 Ce phases can be observed. During hot extrusion, Al 4 Ce particles can act as nuclei for the DRX, and so it is verified that the increase in the amount of the thermodynamically stable Al 4 Ce particles with an increasing content of Ce addition can provide more nucleation sites for the DRX grains, and thus achieve grain refinement [18]. The excellent grain-refining effect of the Ce addition can also be concluded in the results depicted in Figure 2a Engineering tensile stress-strain curves of the as-extruded AZ80-xCe (x = 0, 0.2, 0.8 and 1.4 wt %) alloys at room temperature are presented in Figure 5a. The detailed tensile properties are summarized in Table 2. With the addition of Ce, a slight increase for both the yield strength (YS) and ultimate tensile strength (UTS) was obtained, i.e., from ~156.4 MPa and ~305.3 MPa for the AZ80 alloy, to ~180.4 MPa and ~326.2 MPa for the AZ80-0.8Ce alloy, respectively. When the Ce content climbed to 1.4 wt %, both YS and UTS reduced slightly. This phenomenon can be attributed to the reason that when Ce addition rises from 0.8 wt % to 1.4 wt %, the volume fraction of the Al4Ce particles increases greatly (Figure 4o-p). These high-melting-point hard phase particles break up during hot extrusion, and their edges and corners become sharp, which can lead to stress concentration and become the source of microcracks, thus reducing strength. However, the grain size of the alloy with the addition of 1.4 wt % Ce was much finer (Figure 2c-d); the strength still remained at a relatively high level, due to the effect of grain refinement strengthening. In contrast, there appeared an obvious and persistent increase for the elongation value from 22.1% for the AZ80 alloy, to 33.9% for the AZ80-1.4Ce alloy. The significant enhancement in the elongation of the Ce-containing extruded alloys is mainly a result of grain refinement, as shown in Figure 2a-d. The reason for such a shift lies in the fact that grain refinement can suppress the nucleation and propagation of cracks by forming more serrated grain boundaries, and reducing the size of possible flaws that may cause stress concentration [19]. Besides, grain refinement can also make the deformation more uniform. Therefore, a AZ80-1.4Ce alloy with the finest grain size can endure the greatest amount of deformation before fracture. Engineering tensile stress-strain curves of the as-extruded AZ80-xCe (x = 0, 0.2, 0.8 and 1.4 wt %) alloys at room temperature are presented in Figure 5a. The detailed tensile properties are summarized in Table 2. With the addition of Ce, a slight increase for both the yield strength (YS) and ultimate tensile strength (UTS) was obtained, i.e., from~156.4 MPa and~305.3 MPa for the AZ80 alloy, to~180.4 MPa and~326.2 MPa for the AZ80-0.8Ce alloy, respectively. When the Ce content climbed to 1.4 wt %, both YS and UTS reduced slightly. This phenomenon can be attributed to the reason that when Ce addition rises from 0.8 wt % to 1.4 wt %, the volume fraction of the Al 4 Ce particles increases greatly (Figure 4o-p). These high-melting-point hard phase particles break up during hot extrusion, and their edges and corners become sharp, which can lead to stress concentration and become the source of microcracks, thus reducing strength. However, the grain size of the alloy with the addition of 1.4 wt % Ce was much finer (Figure 2c-d); the strength still remained at a relatively high level, due to the effect of grain refinement strengthening. In contrast, there appeared an obvious and persistent increase for the elongation value from 22.1% for the AZ80 alloy, to 33.9% for the AZ80-1.4Ce alloy. The significant enhancement in the elongation of the Ce-containing extruded alloys is mainly a result of grain refinement, as shown in Figure 2a-d. The reason for such a shift lies in the fact that grain refinement can suppress the nucleation and propagation of cracks by forming more serrated grain boundaries, and reducing the size of possible flaws that may cause stress concentration [19]. Besides, grain refinement can also make the deformation more uniform. Therefore, a AZ80-1.4Ce alloy with the finest grain size can endure the greatest amount of deformation before fracture.  The work-hardening rate (Θ = ∂σtrue/∂εtrue) vs the true strain curves for the uniform plastic deformation stage in tensile tests of the as-extruded AZ80-xCe (x = 0, 0.2, 0.8 and 1.4 wt %) alloys are shown in Figure 5b. One can see that the value of the work-hardening rate Θ decreases rapidly after yielding. The declining rates obviously slows down obviously at a true strain of ~7%. The Θ value of the AZ80 alloy begins to plunge rapidly again, when the true strain comes to ~17%. In comparison, the critical true strain values where the declining rate of Θ values speed up again gradually increase with increasing Ce content. When the Ce content is 1.4 wt %, the critical true strain rises to ~27%. Furthermore, AZ80-1.4Ce alloy always shows the highest work-hardening rate at the same true strain after yielding among all of these alloys. These results coincide with the phenomenon observed in Figure 4a, that a higher Ce content contributes to a larger elongation of the as-extruded AZ80-xCe (x = 0, 0.2, 0.8 and 1.4 wt %) alloys. The relationship between the true flow stress and the true strain can be described by the Hollomon equation (σtrue = Kεtrue n ), where n is the work-hardening index, and K is the strength coefficient. The value of the work-hardening index n can reflect the uniform deformation ability of metal materials. Generally, higher n values always come with a better formability of alloys at room temperature [20]. Based on the Hollomon equation, the average n values of the as-extruded AZ80-xCe (x = 0, 0.2, 0.8 and 1.4 wt %) alloys are calculated to be 0.34, 0.36, 0.37, and 0.41, respectively. It demonstrates that the addition of Ce makes the AZ80 alloy more capable of deformation. Researchers have reported a dynamic competitive relationship between the dislocation accumulation caused by plastic deformation, and the dislocation annihilation induced by dynamic recovery, which affects the work-hardening capability of as-extruded alloys [21]. As shown in the insets of Figure 4fh, Al and Zn elements are mostly dissolved into the Mg matrix in the as-extruded alloys. The large amount of solute atoms can remarkably reduce the dynamic recovery rate of the dislocations, and is beneficial for work-hardening. On the other hand, when the Ce content is 0.8 wt % or higher, a large amount of well-dispersed fine Al4Ce particles formed in the Ce-containing as-extruded alloys can hinder the recovery of dislocations, which also improves the work-hardening ability. Consequently,  The work-hardening rate (Θ = ∂σ true /∂ε true ) vs. the true strain curves for the uniform plastic deformation stage in tensile tests of the as-extruded AZ80-xCe (x = 0, 0.2, 0.8 and 1.4 wt %) alloys are shown in Figure 5b. One can see that the value of the work-hardening rate Θ decreases rapidly after yielding. The declining rates obviously slows down obviously at a true strain of~7%. The Θ value of the AZ80 alloy begins to plunge rapidly again, when the true strain comes to~17%. In comparison, the critical true strain values where the declining rate of Θ values speed up again gradually increase with increasing Ce content. When the Ce content is 1.4 wt %, the critical true strain rises to~27%. Furthermore, AZ80-1.4Ce alloy always shows the highest work-hardening rate at the same true strain after yielding among all of these alloys. These results coincide with the phenomenon observed in Figure 4a, that a higher Ce content contributes to a larger elongation of the as-extruded AZ80-xCe (x = 0, 0.2, 0.8 and 1.4 wt %) alloys. The relationship between the true flow stress and the true strain can be described by the Hollomon equation (σ true = Kε true n ), where n is the work-hardening index, and K is the strength coefficient. The value of the work-hardening index n can reflect the uniform deformation ability of metal materials. Generally, higher n values always come with a better formability of alloys at room temperature [20]. Based on the Hollomon equation, the average n values of the as-extruded AZ80-xCe (x = 0, 0.2, 0.8 and 1.4 wt %) alloys are calculated to be 0.34, 0.36, 0.37, and 0.41, respectively. It demonstrates that the addition of Ce makes the AZ80 alloy more capable of deformation. Researchers have reported a dynamic competitive relationship between the dislocation accumulation caused by plastic deformation, and the dislocation annihilation induced by dynamic recovery, which affects the work-hardening capability of as-extruded alloys [21]. As shown in the insets of Figure 4f-h, Al and Zn elements are mostly dissolved into the Mg matrix in the as-extruded alloys. The large amount of solute atoms can remarkably reduce the dynamic recovery rate of the dislocations, and is beneficial for work-hardening. On the other hand, when the Ce content is 0.8 wt % or higher, a large amount of well-dispersed fine Al 4 Ce particles formed in the Ce-containing as-extruded alloys can hinder the recovery of dislocations, which also improves the work-hardening ability. Consequently, as-extruded AZ80-0.8/1.4Ce alloys can achieve an extraordinary combination of high ductility and a relatively high strength. Figure 6 shows the results of EBSD analyses of the as-extruded AZ80-xCe (x = 0, 0.2, 0.8 and 1.4 wt %) alloys at a cross-section paralleled to the ED. From the orientation maps shown in Figure 6a-d, the grain size of the as-extruded AZ80 alloy is greatly refined with the addition of Ce, which is in accordance with the results that the OM images express (Figure 2), as well as the results of the SEM micrographs ( Figure 4). A typical sheet texture can be observed in all samples, and the as-extruded AZ80 alloy shows a relatively stronger basal texture. The texture peak intensity at the {0001} pole figure of the as-extruded AZ80 alloy is 18.32 multiples of a random distribution (MRD). It is gradually weakened after adding Ce to the alloy, and it declines to 11.50 MRD when the content of Ce is 1.4 wt %, which corresponds to the results that are indicated in the XRD images ( Figure 3). Besides, the texture slightly spreads out from the {0001} plane with the addition of Ce. This phenomenon is well consistent with the extruded RE-containing Mg alloys that are mentioned in [22][23][24], which have weaker textures than those of conventional Mg alloys. Stanford et al. [25] found that a weaker texture could significantly improve the ductility. The responsible reasons are as followed: the main deformation mechanism of Mg alloys at room temperature is basal slip, and the extrusion textures formed in the Mg alloys can make it difficult for basal slips to operate during tensile deformation in the ED, because the basal poles are aligned with the sheets' normal direction (ND). Consequently, the weakened texture leads to reduced barriers for basal slip, and the deformability of the alloys will also be ameliorated. The distribution histograms of the Schmid factors for basal slips are shown in Figure 6e-h, and the average value of the Schmid factor increases from 0.232 (for AZ80) to 0.293 (for AZ80-1.4Ce), which is consistent with the phenomena that finer grains always come with higher Schmid factors for basal slips [16,26]. Besides, with the addition of Ce, the area fraction of the proportion with a Schmid factor value that is greater than 0.4 significantly surges from 6.6% (for AZ80) to 17.7% (for AZ80-0.2Ce), and then the variation becomes relatively gentle, with area fractions of 20.6% (for AZ80-0.8Ce) and 21.3% (for AZ80-1.4Ce), respectively. So far, many reports have revealed that a higher Schmid factor is usually accompanied with better formability in Mg alloys [26][27][28]. The Schmid factor (m) can be calculated by the following equation [29]: m = cos θ × cos λ, where θ and λ are the angles between the stress direction and the slip direction and the normal direction of the slip plane, respectively. Thus, Schmid factor (m) is bounded by 0 and 0.5, and the value of m approaches 0.5 when θ and λ are close to 45 • , in which condition the slip system is in the orientation of easy slip, also known as the soft orientation. In this work, the orientations with m values ranging between 0.4 and 0.5 are regarded as soft orientations, and the area fractions of the basal slip systems in soft orientations are calculated. As shown in Figure 6e-h, the area fraction rises with increasing Ce content, which is in favor of facilitating dislocation slips and promoting the formability of the as-extruded AZ80 alloy. Figure 6i-l shows the distribution histograms of the misorientation angle. One can find that the fraction of low angle grain boundaries (LAGBs) of the as-extruded AZ80 alloy is 14.4%, while it is 12.3%, 10.0%, and 9.1% in alloys with additions of 0.2 wt.%, 0.8 wt.%, of 1.4 wt.% Ce, respectively. For microcrystalline alloys, plastic deformation is dominated by both intragranular and intergranular dislocation motions [30]. High-angle grain boundaries (HAGBs) with misorientation angles larger than 15 • exhibiting relatively lower thermal and mechanical stabilities, and providing barriers for the transmission of dislocations from one grain to another, then creating dislocation pile-ups at the boundaries and enhancing the macroscopic strength [31]. In addition, when cracks occur in grains, a larger number of HAGBs are responsible for better crack propagation resistance [32]. On the contrary, LAGBs can even be described as an array of dislocations; thus, cracks can be transferred directly through the LAGBs into other grains. In conclusion, the reduction of the fraction of LAGBs can suppress the propagation of cracks, and it is beneficial for mechanical properties.  In order to further identify the phases in the as-extruded alloys, TEM analyses were performed with an as-extruded AZ80-0.8Ce alloy. Bright-field TEM micrographs are presented in Figure 7a, c, e, g, i. Corresponding diffraction patterns acquired from selected regions marked by the arrows A, B, C, and D in Figure 7a, c, e, g are shown in Figure 7b, d, f, h, respectively. One can see that plate-like and block-shaped Mg17Al12 precipitates (region A and B) with sizes of 300-600 nm, disperse in the AZ80-0.8Ce alloy. Coarse tabular-blocky Al4Ce phases (region C) with sizes of 500-700 nm, as well as fine rod-like and spherical Al4Ce phases (region D) with sizes of 50-200 nm can also be observed.   In order to further identify the phases in the as-extruded alloys, TEM analyses were performed with an as-extruded AZ80-0.8Ce alloy. Bright-field TEM micrographs are presented in Figure 7a, c, e, g, i. Corresponding diffraction patterns acquired from selected regions marked by the arrows A, B, C, and D in Figure 7a, c, e, g are shown in Figure 7b, d, f, h, respectively. One can see that plate-like and block-shaped Mg17Al12 precipitates (region A and B) with sizes of 300-600 nm, disperse in the AZ80-0.8Ce alloy. Coarse tabular-blocky Al4Ce phases (region C) with sizes of 500-700 nm, as well as fine rod-like and spherical Al4Ce phases (region D) with sizes of 50-200 nm can also be observed. In order to further identify the phases in the as-extruded alloys, TEM analyses were performed with an as-extruded AZ80-0.8Ce alloy. Bright-field TEM micrographs are presented in Figure 7a,c,e,g,i. Corresponding diffraction patterns acquired from selected regions marked by the arrows A, B, C, and D in Figure 7a,c,e,g are shown in Figure 7b,d,f,h, respectively. One can see that plate-like and block-shaped Mg 17 Al 12 precipitates (region A and B) with sizes of 300-600 nm, disperse in the AZ80-0.8Ce alloy. Coarse tabular-blocky Al 4 Ce phases (region C) with sizes of 500-700 nm, as well as fine rod-like and spherical Al 4 Ce phases (region D) with sizes of 50-200 nm can also be observed. These results are in good agreement with the XRD results shown in Figure 3c. It is well known that Mg 17 Al 12 precipitates are strengthening phases in the extruded AZ80 alloys, which can be formed by dynamic precipitation from a supersaturated solid solution during hot extrusion [33]. However, adding Ce can restrain the precipitation of Mg 17 Al 12 particles, since the formation of a thermodynamically stable Al 4 Ce phase would consume Al. Fine nano-sized Al 4 Ce particles can precipitate during the hot extrusion process, while irregular bulk-like submicron-/micron-sized Al 4 Ce particles may be formed from the primary Al 4 Ce, emerging in a smelting process and breaking in the extrusion process. Figure 7i shows a typical fine rod-like Al 4 Ce particle in the α-Mg matrix. A high-resolution TEM (HRTEM) image of zone F in Figure 7i is shown in Figure 7j. A clear interface between the nano-Al 4 Ce particle and the α-Mg matrix can be observed, illustrating good interfacial bonding, which is beneficial for transferring the load from the Mg matrix to the particles, and thus improving the strength of the alloys. Figure 7k presents the inverse fast-Fourier-transform (IFFT) image of zone E in Figure 7j. Apparently, a large number of dislocations are distributed in the α-Mg matrix adjacent to the nano-Al 4 Ce particle. A high dislocation density is reported to be beneficial to the precipitation of nano-Al 4 Ce particles as heterogeneous nucleation sites [34]. Besides, the nano-Al 4 Ce particles could also act as strong obstacles to the dislocation motion and result in a strengthening effect. However, on account of the incomplete solid solution of Al 4 Ce, the number of nano-Al 4 Ce particles is quite limited, and the effect of precipitation reinforcement is insignificant.

Conclusions
In the present study, the effect of Ce content on the microstructure and tensile properties of as-extruded AZ80-xCe alloys (x = 0, 0.2, 0.8 and 1.4 wt %) were investigated. In the AZ80 alloy, two phases could be observed, including the primary α-Mg phase and the Mg 17 Al 12 phase. With the addition of Ce, a bulk-shaped Al 4 Ce phase below 10 µm appeared, and a small number of Al 4 Ce particles were distributed along the ED for the AZ80-0.2Ce alloy. When the Ce content was 0.8 wt % or higher, Al 4 Ce particles in both nano-and micron sizes were well dispersed in the α-Mg matrix. The area fraction of Al 4 Ce particles rose with increasing Ce content, indicating a larger number of thermodynamically stable Al 4 Ce particles, which could provide more nuclei for DRX during hot extrusion, and promote grain refinement. The grain size of the as-extruded AZ80-1.4Ce alloy was significantly reduced by~56.7%, compared to that of the AZ80 alloy. Adding Ce could lead to significant grain refinement, a weakened texture strength, an increased Schmid factor, and a reduced area-fraction of LAGBs. All of these were beneficial to improving the deformability of the extruded AZ80 alloy. With Ce content increasing to 0.8 wt % or higher, a superior combination of tensile properties could be achieved, with a high elongation of~29-34%, and a relatively high ultimate tensile strength of~320-327 MPa, respectively.

Conflicts of Interest:
There are no conflicts of interest.