Effects of temperature-induced variation of the second phase on the microstructure, texture and properties of Mg-RE-Zn alloys

In this paper, the Mg-9.32Gd-3.72Y-1.68Zn-0.72Zr (wt%) alloy has undergone three times of repetitive upsetting extrusion deformation. Alloys with different morphologies and distribution patterns of the second phase have been prepared by varying the deformation temperature in each pass. The effects of the second phase on microstructure, texture and mechanical properties are investigated. The results show that (i) the second phase, including Mg5Gd and LPSO phase, has an important effect on the dynamic recrystallization (DRX) behavior of the alloy. (ii) Appropriate lamella distance and block phase size can promote the activation of the slip system and effectively weaken the texture strength. (iii) The DT sample has an effective combination of fine grain strengthening and second phase strengthening due to the reasonable second phase distribution and size, which greatly improves the UTS and YS of the alloy.


Introduction
In recent years, magnesium alloy, one of the lightest metal structural materials at present, has the characteristics of low density, high specific strength, high specific stiffness and weak anisotropy.It has broad application prospects in aerospace, national defense and military, automotive electronics and other fields [1][2][3].However, due to the disadvantages of low strength and poor plasticity of magnesium alloy, its application has been greatly limited [4,5].Among them, Mg-RE-Zn alloy greatly improves the ultimate tensile strength and yield strength of magnesium alloy because it introduces LPSO phase with high hardness and elastic modulus [6][7][8].
In recent decades, several innovative methods of large plastic deformation (SPD) have been developed and widely used to improve material properties [9].And among the many SPDS techniques, the repetitive upsetting extrusion (RUE) technique combining upsetting and extrusion deformation provides a good plasticity processing method for difficult to deform magnesium alloys [10,11].It can be used for preforming of large parts with low equipment requirements.At the same time, compression and extrusion stress states are loaded alternately in the RUE process.The microstructure and texture are regulated by controlling the deformation conditions.The alloy properties are effectively strengthened.The strengthening mode of Mg-RE-Zn alloy is mainly fine crystal strengthening and second phase strengthening [12].Wu et al [13] investigated the texture evolution mechanism of rare-earth magnesium alloys containing LPSO phase during deformation and their texture stability during subsequent heat treatment.The relationship between texture weakening and deformation mechanism during multi-pass deformation was further investigated.Du et al [14] investigated the average grain size of Mg-Gd-Y-Zn-Zr alloy from the initial 64 mm to 3.4 mm at the completion of 6 passes of RUE deformation.Li et al [10] investigated the effect of repetitive upsetting and extrusion (RUE) on the anisotropy of microstructure, texture and mechanical properties of Mg-9Gd-4Y-Zn-0.4Zralloy.Texture weakening and grain refinement play a synergistic role in the weakening of mechanical property anisotropy.In addition, the changes of LPSO phase during RUE deformation and its effect on dynamic recrystallization (DRX) were studied in detail by Meng et al [15].However, the effect on the mechanical properties of the second phase was not mentioned.
Nevertheless, the level of research on the RUE process is still limited.Currently, most of the attention is focused on the influence of the number of deformation passes or deformation paths on the microstructure and mechanical properties.The effects of the distribution and size of the second phase on the material structure and properties have not been mentioned.Therefore, regulating the morphology and distribution of the second phase of alloys based on different RUE deformation temperatures to enhance the mechanical properties is of great significance for the further application of the RUE process.
In this paper, by changing the deformation temperature of each pass of repetitive upsetting extrusion (RUE), the purpose of regulating the morphology and distribution of the second phase in alloy is achieved.The strength of the alloy in three passes is much higher than that of the isothermally deformed alloy with the same extrusion ratio.In addition, the effects of different second-phase morphologies and distributions on the dynamic recrystallization behavior, texture and properties of the alloy are investigated in detail.

Experimental procedure
The material used in this paper is Mg-9.32Gd-3.72Y-1.68Zn-0.72Zralloy.The billet is cylindrical bar with diameter of 79 mm and height of 343 mm cut from the ingot.According to some previous studies on the alloy by our team [11,13,15,16], the most optimal homogenization system was selected.This method can not only achieve the effect of homogenization, but also avoid excessive grain growth.The alloy is homogenized and hold at 510 °C for 16 h prior to RUE deformation.Afterwards, it is cooled in air.The deformation is mainly realized by repetitive upsetting extrusion (RUE).The billet is placed in the mold with diameter D to complete upsetting; then it is extruded in the channel with diameter d.The experimental flow is shown in Figure 1(a).In this experiment, D is 82 mm and d is 79 mm.The RUE was subjected to a total of three passes with an extrusion ratio of 1.96.The extrusion speed was 0.5 mm/s.The experiment was divided into three groups: the first group was decreasing RUE(DT), where the temperature was gradually reduced from 480 °C to 430 °C.The second and third groups were isothermal RUE with temperatures of 430 °C(IT-430) and 480 °C(IT-480), respectively.The experimental procedure is shown in figure 1(b).The cumulative strain is calculated to be approximately 0.45.According to the equation (1) where e is equivalent strain, n is the number of passes.The deformation was carried out in a 4-THP61-630 machine.After that the center position was taken as a sample for observation, followed by room temperature tensile testing.The tensile testing bar size and sample location is shown in the figures 1(c), (d).The microstructure of samples was characterized by optical microscopy (OM, DM2500M, Leica Microsystems, Wetzlar, Germany) and scanning electron microscopy (SEM, SU5000, Hitachi, Tokyo, Japan) at an accelerating voltage of 20 kV.The samples used for OM measurements were mechanically polished and chemically etched using a solution containing 3 g of picric acid, 2 ml of acetic acid, 5 ml of distilled water and 60 ml of ethanol.X-ray diffraction (XRD) was used for the identification of the phases during the deformation process.In a further step, scanning electron microscopy (SEM) equipped with electron backscatter diffraction (EBSD) was used to observe the evolution of the grain orientation.TSL OIM analysis version 7.3 (EDAX Inc., Mahwah, NJ, USA) is used to quantify the details of grain misorientation angle, grain size, microstructure, average grain misorientation and Schmidt factor.The mechanical properties of the alloys are tested using an Instron 3382 tensile tester (Instron, Norwood, MA, USA) at a strain rate of 0.01 mm min −1 at room temperature.To ensure the accuracy of the experimental results, three tensile samples were processed in each condition for testing and the results are averaged.

Microstructure evolution
Figure 2(a) shows the SEM microstructure of alloy after homogenization treatment at 510 °C × 16 h.From the figure, it can be seen that the alloy mainly consists of α-Mg matrix phase, intragranular lamellar LPSO phase and intergranular block LPSO phase.There is a small amount of rare-earth-rich phase randomly distributed inside the grains and at the grain boundaries (the part circled in red).Gao et al [17]  It can be seen that the average grain size of the alloy is 201.72 μm.The grain orientation is relatively random.The maximum polar densities are 7.208 (PF) and 2.292 (IPF), respectively.Furthermore, the average width of the block LPSO phase is 12.05 μm.lamellar LPSO phase were measured by randomly selecting 100 data points from each of the three samples.The average lamellar distance of the lamellar LPSO phase in the deformed grains of each sample has been measured.∼0.70 μm, ∼0.47 μm as well as ∼0.92 μm, respectively.The grains of the DT samples are broken to different degrees.But there still exist some large deformed grains, forming a typical bimodal structure.Dense lamellar LPSO phases are retained inside the large deformed grains.These lamellar phases undergo different degrees of kinking.The IT-430 sample formed a more typical bimodal microstructure than DT sample, including a large number of coarse deformed grains and fine dynamic recrystallization (DRX) grains distributed around the large grains.A more densely lamellar LPSO phase is retained inside the large grains, and only different degrees of bending occur.At the same time, DRX behavior can be clearly observed on the bending bands.The grain refinement of the IT-480 sample is very obvious, and dynamic recrystallization is almost complete.Only a few deformed grains remain.The size is much smaller than the remaining two groups of samples.The retention of lamellar LPSO phases is significantly lower than that of the remaining two groups of samples.
Figures 3(d)-(f) shows the SEM images of the samples under different deformation conditions.Compared with the homogenized samples, the block LPSO phase of the deformed samples changes significantly.The block LPSO phases of the DT sample and the IT-430 sample are slightly elongated and no obvious fragmentation is observed.On the contrary, the block LPSO phase of the IT-480 sample is observed to be significantly elongated as well as fragmented.Like the lamellar phase been, the average widths (w) of the block LPSO phases of the DT sample and the IT-430 sample are measured to be ∼10.08 μm and ∼12.02 μm, respectively.Compared with the homogenized sample, the dimensions do not change significantly.While the width of the block LPSO phase of the IT-480 sample decreased dramatically, and the average width reached 5.20 μm.In addition, a large number of β-Mg5 (Gd, Y, Zn) phase are obviously precipitated compared initial alloy.They are mostly distributed at the grain boundaries of the DRXed grains and inside the block phase.Among them, IT-430 precipitated the most β-Mg5 (Gd, Y, Zn) phase, followed by DT and IT-480 the least.[10-14, 16, 18, 19], it can be seen that during the RUE process, the Mg-9.32Gd-3.72Y-1.68Zn-0.72Zralloy is mainly composed of α-Mg matrix, β-Mg 5 (Gd, Y, Zn) phase and different Mg12 (Gd, Y, Zn) phases (block and lamellar shape).Among them, the content of LPSO phase is much higher, which is the main reinforcing phase in the second phase.
In order to characterize the grain refinement during different RUE processes, each sample is analyzed using EBSD.Figure 4 shows the orientation microstructure and the corresponding images of average grain size distribution of different samples.The grains with Grain Orientation Spread (GOS) < 2°are defined as DRX grains; those with 2° GOS 5°are defined as sub-grains and those with GOS > 5°are defined as deformed grains.The black area in the IPF maps is mainly the LPSO phase.The EBSD system cannot recognize the LPSO phase due to the lack of corresponding grain parameters [20,21]  instead, was significantly larger than that of others.This suggests that the degree of grain refinement gradually increases with the increase of deformation temperature.

Texture of different alloy states
Figure 5 shows the pole figure (PF) and Inverse pole figure (IPF) of the alloy in different states.After deformation, the samples all formed a typical basal texture.The texture intensity was significantly weakened compared to the initial state, and the texture components were all relatively dispersed.As the deformation temperature increases, the texture strength shows a gradual decrease.Moreover, the degree of bimodality of the texture of the samples is significantly weakened with the increase of the deformation temperature.For the (0001) PF map, the IT-430 sample has the largest texture intensity of 6.746, followed by the DT sample with 4.995.And the weakest texture intensity is that of the IT-480 sample, which is 2.773 only.The maximum texture intensities of IPF are 2.798 (DT), 6.386 (IT-430), and 1.397 (IT-480), respectively.In addition, the samples showed completely different texture components.As can be seen from the (0001) PF map in figure 5, the DT sample exhibits a distinct extruded texture.The grain c-axis is distributed in the region of 0 ∼ 30°from the TD direction.As can be seen from the IPF, the DT samples formed a strong texture component of ED∥〈20-21〉, a weak fiber texture component of ED∥〈10-10〉, and a weak rare-earth texture component of ED∥〈4-2-23〉.The (0001) PF of the IT-430 sample (figure 7(b)) exhibits a distinct basal bimodal texture.As shown in the IPF, the IT-430 sample formed a strong abnormal texture component of ED∥[0001] and a secondary rare-earth texture component of ED∥〈10-11〉-〈20-23〉 [22].After three passes of deformation, the (0001) basal texture of the IT-480 sample is maximally weakened, forming an anomalous texture of ED∥〈0001〉 and a weak fiber texture of ED∥〈10-10〉-〈2-1-11〉.

Mechanical properties
Figure 6 shows the uniaxial tensile mechanical properties of different samples.The loading direction is parallel to the ED.The results show that the DT samples have the highest comprehensive performance under the same extrusion ratio and the same cumulative strain.The ultimate tensile strength (UTS) and tensile yield strength (TYS) increase by ∼80 MPa compared to the other samples.While the elongation (EL) increases by ∼0.8% compared to that of the IT-430 sample.The values of UTS, YS, and EL for the DT sample are ∼394.3MPa, ∼301.8MPa, and ∼8.4%, respectively.The UTS and YS of the IT-430 sample are both slightly higher than those of the IT-480 sample.The UTS and YS increased by ∼0.1 MPa and ∼19 MPa respectively than those of the IT-480 sample.EL is much smaller than that of IT-480 with a difference of ∼10.6%.It can be seen that the UTS and TYS of the material first increase and then decrease with the increase of the layer spacing of lamellar LPSO phase and the decrease of the block LPSO phase size, while the EL gradually decreases.It will be clarified in section 4.3.

. Dynamic recrystallization mechanism
From figure 4, it can be seen that each sample exhibits a completely different dynamic recrystallization behavior.In order to analyze the recrystallization mechanism of the samples in more detail, the characteristic regions (black boxed parts) of each sample in figure 4 were selected for analysis in figure 7. Figures 7(a)-(c) show selected grains from the DT sample, G1.It is obvious from figure 7(a) that the grain refinement is dominated by grain breaking.There are more sub grains and a few DRX grains around the parent grain (figure 7(b)).There are a large number of low-angle grain boundaries (LAGBs, 2°and 15°) in the parent grains and sub grains.The fluctuation of the misorientation angle from the point to the origin along the black arrows A-B is quite large, reaching 35°.It indicates that there is a high degree of dislocation activity within the grain.In addition, it can be seen from figure 7(c) that the DRXed grains and sub grains (black parts) effectively weaken the texture strength, but there is an inheritance relationship between them and the parent grain (the color section).The parent grain undergoes lattice rotation during deformation, and the grain c-axis is gradually deflected toward 〈10-10〉, 〈2-1-10〉, and 〈0001〉.The final grain orientation is formed which is clustered in the region labeled in figure 7(c).In addition, a large amount of LAGBs are still retained inside the sub grains.This suggests that these sub grains can be further fragmented to achieve DRX.This also explains why the area fraction of sub grains in the DT samples is much higher than others.It can be inferred that DRX formation is related to the continuous absorption of grain dislocations by sub grains and their gradual conversion into high angle grain boundaries (HAGBs, >15°).This phenomenon is classified as a continuous dynamic recrystallization (CDRX) mechanism.Beyond that, we find DRX grains formed along the block LPSO phase in the deformed grains in figure 4(a).Owing to the high hardness and high modulus vector of the block LPSO phase, it leads to a strong stress concentration at the α-Mg/LPSO phase interface [23].Therefore, it can further promote DRX through the particle-stimulated nucleation (PSN) mechanism.
Figures 7(d)-(f) shows the selected characteristic grains from the IT-430 sample.Its parent grain exhibits a completely different DRX behavior from G1 although the misorientation is not much significant.It can be clearly seen that the grain refinement is dominated by discontinuous dynamic recrystallization (DDRX) promoted by block LPSO phase and grain boundary motion.Meanwhile, a large number of sub grains are formed within the parent grain by the interaction of the pre-formed DRX grains with the parent grain, which initiates the CDRX mechanism.Unlike G1, the fine grains around the parent grain are almost all DRX grains.In addition, the DRX grain orientations are all more random compared to G1.The inheritance behavior in G1 does not occur, as shown in figure 7(f).Apart from that, there are also a plenty of LAGBs among the parent grain.The point-to-origin misorientation of 36.9°forC-D is lower than that in G1.As can be inferred, these DRXed grains consume part of the dislocations and hinder the CDRX process.Beyond that, it can be seen that the DRXed grains gradually expand inward along the LAGBs.Unlike the fragmentation of grains in G1, the sub grains in G2 provide more nucleation sites for DRX.It further promotes the DDRX process, as shown by the red arrow in figure 7(d).In summary, it can be judged that the DRX behavior of the IT-430 sample is dominated by the DDRX behavior.
Figures 7(g)-(i) shows the selected characteristic grain G3 from the IT-480 sample.The dislocation activity in G3 is relatively low compared to G1 and G2.The point-to-origin misorientation of E-F is only 22.8°.As in G2, a necklace DRX grains exists around the parent grain.Meanwhile, the phenomenon of energy absorption by LAGBs transforming into HAGBs, i.e., CDRX behavior, can be clearly seen within the parent grain.According to figure 7(i), it can also be observed that G3 has both the inherited behavior of DRXed grains in G1 and a large number of random orientations of DRXed grains in G2.Therefore, it can be recognized that the refinement of the deformed grains at this stage mainly relies on the synergy of DDRX and CDRX.
Common sense would suggest that the DRX area fraction should gradually increase as the deformation temperature increases.This is different from the pattern presented in this paper.As can be seen from figures 4(a)-(c), IT-480 has almost completed DRX.The DT sample is still in the deformed state.The sub grains and deformed grains still occupy the majority of the fraction.The IT-430 sample is also in the deformed state.But, the fraction of DRX grains is larger than DT sample.While the fraction of sub grains is obviously lower than that of DT sample.This is inevitably related to the distribution and size of the second phase of the alloy.

Effect of second-phase relative dynamic recrystallization behavior
The second phase, as strengthening phase, has an essential influence on the dynamic recrystallization behavior of alloys.On the one hand, it can hinder the dislocation motion and delay the DRX; on the other hand, it can also provide more nucleation sites for DRX grains and promote the occurrence of DRX [24].The pinning effect it produces can effectively inhibit the growth of DRX grains [25].And what kind of role it plays mainly depends on the strain with its own morphology and distribution [26].From the XRD results in figure 3(g), it can be seen that the second phase in this alloy mainly consists of the LPSO phase and β-Mg5(Gd, Y) phase.According to previous studies [15,22,27,28], it is evident that the β-phase can provide nucleation sites for DRX grains.Besides, the pinning effect it produces can effectively restrain the development of DRXed grains.The influence of LPSO phase on the DRX behavior needs to be further explored.It is mentioned in section 3.1 that the grains of DT and IT-430 samples contain dense lamellar LPSO phase and coarse block LPSO phase.However, there is a slight difference in the lamellar phase spacing between the two.In contrast, the lamellar LPSO phase inside the IT-480 sample has almost disappeared, and only a small amount of fine block LPSO phase remains.
To investigate the effect of LPSO phase on the DRX behavior of the alloys, the deformation behavior of LPSO phase in DT and IT-430 samples are analyzed (figure 8).As can be seen, the lamellar distance of the lamellar phase has a very strong influence on the DRX behavior of the grains.The average lamellar distance (d) in the DT sample is 0.70 μm (figure 3(a)).The lamellar LPSO phase in the same deformed grains undergoes kinking to varying extents, and the kinking bands are twisted.The kinking angles ranged from 19.3°to 75°, where all of them formed high-angle grain boundaries.In addition, some of the lamellar LPSO phases are fractured along the lamellae, as shown by the red circles in figure 8(a).This verifies the opinion that the grain refinement of DT samples mainly relies on grain fragmentation.Meanwhile, the twisted kink bands provide a large number of nucleation sites for the DRXed grains.As a result, a large number of DRXed grains are discovered around the kink bands (The part indicated by the blue arrow).This is also the reason why DRXed grains with orientations significantly different from those of the parent grain appear around the HAGBs in figure 8(a).
The average lamellar distance in the IT-430 sample (figure 3(b)) is 0.47 μm, and the lamellar phases are bent only at a small angle (12°), and the bending zone is basically distributed along a straight line (figure 8(b)).There are a large number of DRXed grains on the bending band, accompanied by the precipitation of a large number of β-phases.This indicates that DRX occurs within the deformed grains in the IT-430 sample before the formation of high-angle grain boundaries.In other words, the LAGBs generated by lamellar phase bending provide nucleation sites for DRX nucleation.This confirms the opinion in section 4.1.1.At the same time, the stable lamellar phase structure also prevents DRX from occurring, resulting in the retention of a large number of deformed grains (43.5% area fraction) in sample IT-430.
The average lamellar distance in the IT-480 sample is 0.92 μm ( figure 3(c)).Unlike the previous two samples, as shown in figure 8(c), the lamellar phase is continuously kinked in the same direction, and the cumulative kink angle reaches 103.2°.The lamellar phases inside the deformed grains are bent only by 7°and 11.4°.The dislocation activity is not obvious.Instead, the parts close to the grain boundaries undergo lattice rotations with larger angles (33.5°as well as 51.3°) due to grain boundary motions.This is consistent with the phenomenon observed in section 4.1.1.
Since the intragranular lamellar LPSO phase can limit DRX by dislocation pinning [29], the degree of DRX is closely related to the density of dislocations that can be accommodated by the lamellar phase.The dislocation accommodation capacity of the Un-DRXed grains (deformed grains + sub grains) with different lamellar distance lamellar phases inside was characterized using KAM maps as shown in figures  where ρ(cm −2 ) is dislocation density, KAM ave (°) is the average value of the local misorientation obtained by EBSD, μ(nm) is the scanning step of EBSD, b(nm) is the Burgers vector of the matrix.Equation (2) shows that the dislocation density is directly proportional to the KAM value.In other words, the lower the lamella distance, the higher the dislocation density that the grain can accommodate.This also explains why the deformed grains in the IT-430 sample have the least amount of lamellar phase deformation within them.
In addition to the lamellar LPSO phase, the block LPSO phase also has significant influence on the DRX behavior of the alloy.As mentioned earlier, the block phase can promote the occurrence of DRX of alloys through the PSN mechanism.And it needs to be further explored how the size of the block phase affects the DRX behavior.For this question, an empirical equation was suggested [31]: where θ(°) is the lattice rotation at a distance x(μm) from a particle of width w(μm), θ max (°) is the maximum rotation angle, and c 1 is a constant equal to 1.8 [29].
For the particles of diameter greater than 2.5 μm [32], θ max was found to be a function only related to effective strain ε.The relationship between θ max and ε can be given as: ( ) q e = where c 2 is a constant of the order of unity 1.1, where θ max was 24°at strain of 0.45 [29].From equation (3), the size w of the block LPSO phase is proportional to the distance x from the deformation zone when the grain rotation angle θ is equal.For the block LPSO phase with larger size, it can more effectively hinder the dislocation motion and promote the lattice rotation around the block phase.And when the size of the block phase is fine, the dislocation activities it can hinder are limited, with the accumulated energy being limited, and fewer DRXed grains being produced.
In summary, the second phase has an essential influence on the DRX behavior of the alloy.The β phase can provide nucleation sites for DRX grains.On the other hand, the pinning effect it produces can effectively inhibit the growth of DRXed grains.Alloys with different lamellar distance lamellar phases show a completely different DRX behavior.The degree of grain deformation increases with increasing lamellar distance.Smaller lamella distance inhibits the DRX process, and larger lamellar distance facilitates the initiation of various DRX mechanisms and improves the degree of grain refinement.And the DRX region affected with the size of the block LPSO phase increasing.As a result, the three samples show completely different DRX behaviors.
As shown in the figure 9, the DT sample has moderate lamellar phase distance and block phase size.The portion of the alloy grain away from the block phase undergoes lattice rotation due to the accumulation of a certain amount of dislocations in the lamellar phase.This is manifested by the kinking and breaking of the lamellar phase, i.e., CDRX.At the same time, the kinking and fragmentation of the lamellar phase provided nucleation sites for DRX, which led to the occurrence of DDRX behavior around it.And the part close to the block phase accumulates a large number of dislocations under the influence of the block phase, causing the lattice to rotate.In this case, the DDRX around the block phase due to the PSN mechanism and the kinking of the lamellar phase inside the grain are observed.
The IT-430 sample has the smallest lamellar phase distance with the largest block phase size.The inner portion of the grain away from the block phase is only bent at a small angle due to the overcrowding of the lamellar phase.The bending bands provide nucleation sites for CDRX and DDRX behavior.The part near the block phase has accumulated a large number of dislocations, but only a small amount of DDRX occurs due to the high capacity of the lamellar phase.
The IT-480 sample has the widest lamellar phase distance with the minimal block phase size.The portion of the grain interior away from the block phase shows continuous large-angle kinking of the lamellar phase due to the small dislocation capacity of the sparse lamellar phase, and CDRX occurs.Consistent with the DT sample, the kinked and fragmented lamellar phase provides nucleation sites for DRX, which results in DDRX behavior in its surroundings.The portion near the block phase had a relatively limited area of influence due to the small size of the block phase.

Effect of second phase on alloy texture
As shown in figure 5, the Un-DRX grains of each sample have obvious selective orientation.Both the Un-DRX and DRX grains of the DT sample and the IT sample exhibit obvious rare-earth texture.And the maximum intensity point gradually approaches to ED∥〈10-10〉-〈2-1-10〉 as the layer distance of the lamellar phase increases.For the IT-480 sample with the largest lamellar distance, the rare-earth texture has almost disappeared and transformed into ED∥〈10-10〉-〈2-1-10〉 fiber texture.In addition, anomalous textures are found in IT-430 and IT-480 samples.To explore the reasons for the formation of rare-earth texture and anomalous texture, the Schmidt factor (SF) of the characteristic region of each sample is selected for analysis here, as shown in figure 10. Figure 10(a) shows the comparison of the average Schmid factor of each slip system of the three samples, and figure 10(b) shows the schematic diagram of the rotation axis of each slip system [33].Overall, the DT sample has the most active dislocation activity.Only the mean Schmid factor value 〈0.3 for basal 〈a〉 slip, which is 0.295.The remaining slip coefficient activities are at the highest level of the three samples.The prismatic 〈a〉 slip and pyramidal 〈a〉 slip SF for the IT-480 sample are middle of the range between the DT sample and the IT-430 sample.The SF for pyramidal 〈c+a〉 slip varies little from that of the DT sample.The basal 〈a〉 slip of the IT-430 sample is significantly more active than the remaining two groups.While the prismatic 〈a〉 slip system is less active.Therefore, it can be inferred that the lamella distance and block phase size of DT can promote the initiation of the slip system.
In the DT sample, as can be seen in figure 5, the highly active non-basal plane slip makes the grain gradually change from ED∥〈20-21〉 orientation in Un-DRX grain to ED∥〈10-10〉-〈4-2-23〉, forming the current DRX texture (ED∥〈10-10〉-〈20-21〉-〈4-2-23〉).At the same time, the less active basal 〈a〉 slip makes the deformed grains of ED∥〈4-2-23〉 have limited ability to rotate in the direction of ED∥〈0001〉 and ED∥〈2-1-10〉.Thus a weak rare-earth texture of ED∥〈4-2-23〉 is formed.However, the whole texture shows Un-DRX texture due to the stronger Un-DRX texture.In other words, the formation of the rare-earth texture in the DT sample is caused by the lower activity of the basal 〈a〉slip while the non-basal slip is more active.The IT-430 sample has deflected the ED∥〈0001〉 oriented Un-DRX grains in the ED∥〈11-20〉 direction by an certain angle due to the activation of basal slip.The strong texture at ED∥〈0001〉 is weakened.And the pyramidal slip causes the ED∥〈10-11〉 grain orientation in the Un-DRX grains to be gradually deflected toward ED∥〈0001〉 and ED∥〈10-10〉.Therefore, the anomalous texture in the IT-430 sample is composed of both Un-DRX grains and DRX grains.Furthermore, it has the strongest texture due to the fact that the combined activity of the slip system is smaller than others.Unlike the IT-430 sample, the anomalous texture of the IT-480 sample is provided by the DRX grains.The strong texture of ED∥〈10-10〉 in the Un-DRX grains is effectively weakened by the combined effect of non-basal slip.The grain orientation gradually transforms from ED∥〈10-10〉 to ED∥〈0001〉 and ED∥〈11-20〉 directions.
In summary, it can be concluded that the appropriate lamella distance and block phase size can promote the activation of the slip system and effectively weaken the texture intensity.In addition, the rare-earth texture in the alloy is gradually deflected toward ED∥〈10-10〉-〈2-1-10〉 with the increase of the lamella distance and the decrease of block phase.And the formation of the anomalous texture is also closely related to the lamellar spacing of the alloy.For the IT-430 sample with dense lamellar phase, the anomalous texture is formed by the combined effect of Un-DRX grains and DRX grains in non-prismatic slip, while for the IT-480 sample with large lamellar spacing, the anomalous texture is formed by DRX grains in non-prismatic slip.

The contribution of second phase strengthening mechanisms
The strengthening mechanisms after RUE deformation of Mg-Gd-Y-Zn-Zr alloys are mainly attributed to the following [13]: (i) fine-grain strengthening induced by DRX grains; (ii) (0001) basal texture strengthening of coarsely deformed grains and dislocation stacking strengthening of high-density substructures within the grains; and (iii) second-phase strengthening containing β-Mg 5 RE, lamellar and block LPSO phase.In other words, the strengthening mechanism of alloys can be divided into 4 main parts: fine grain strengthening (σ GS , MPa), dislocation strengthening (σ ρ , MPa), texture strengthening (σ t , MPa), and second phase strengthening (σ sp , MPa).The yield strength (σ ρ , MPa) of the alloy is calculated based on the following sum of products equation [34]: Due to the low content of β-phase and very small grain size, the strengthening effect it produces is limited and is neglected here [35].According to equation (5) it can be obtained that the second phase strengthening can be expressed by the following equation: Where σ 0 (MPa) refers to the yield strength of undeformed pure Mg (20 MPa) [36].
Fine grain strengthening is also known as grain boundary strengthening, which follows the Hall-Petch relationship [37]: where σ 0 refers to the yield strength of pure Mg, d(μm) is the average grain size, and k(MPa•μm −1/2 ) is the stress concentration factor representing grain boundaries (GBs) as an obstacle to slip.Previous investigation shows that the value of σ 0 is approximately 46.5 MPa for Mg-Gd alloy [38].k is approximated to be 188 MPa•μm −1/2 in the as-extruded Mg-Gd-Y alloy [6].σ GS of DT, IT-430 and DT-480 samples are calculated to be ∼84.75MPa, ∼79.79 MPa and ∼110.65 Mpa, respectively.Dislocation strengthening can be described by equation (7): where M is the Taylor factors of grains, α is a constant with a value of 0.24 [39], b is the Burgers vector of the matrix approximately equal to 0.32 nm and G (MPa) is the matrix shear module.For magnesium alloys, G = 1.66 × 10 4 MPa [40].σ ρ of DT, IT-430 and DT-480 are calculated to be ∼3.19MPa, ∼2.94 MPa and ∼2.21 MPa.
Texture strengthening can be described by equation (8): where σ 0 is a constant.Based on the strength value of α-Mg single crystal in Mg-Gd alloy, 46.5 MPa is taken [41], m is the average Schmid factor.σ t of DT, IT-430 and DT-480 are calculated to be ∼29.17MPa, ∼29.65 MPa and ∼29.46 MPa.Thus, σ se of DT, IT-430 and DT-480 are easily calculated to be 164.64MPa, 110.95 MPa and 62.08 MPa.It can be clearly seen that the order of the contribution of each strengthening mechanism is: σ se > σ GS > σ t > σ ρ .Among them, the second phase strengthening and fine grain strengthening dominate.The second phase provides a significant strengthening effect on the yield strength of the alloy, and its pattern is consistent with the distribution pattern of the second phase.The strengthening effect of the second phase gradually diminishes with the increase of the layer distance of lamellar LPSO phase and the decrease of the block LPSO phase size.And the DT sample has an effective combination of fine-crystal strengthening and second-phase strengthening due to the reasonable second-phase distribution and size.The UTS and YS of the material are greatly improved.
Furthermore, the second phase has a significant effect on the ductility of the alloy.Elongation, one of the ductility testing standards, can clearly be seen that of the alloys increased with the increase of the lamellar phase distance and the decrease of block phase size.It is not difficult to find that the second phase is often the crack initiation place of the alloy by consulting the previous literature [16,[42][43][44].Therefore, when the content of the second phase increases, the ductility of the material decreases.Meanwhile, it can be known from section 3.1 that the grain refinement degree of the IT-480 sample is far greater than that of the other two patterns.For magnesium alloys [45], the ductility improvement brought by grain refinement cannot be underestimated.

Conclusion
In current research, Mg-9.32Gd-3.72Y-1.68Zn-0.72Zr(wt%) alloys with different morphologies and distribution patterns of the second phase are prepared by varying the deformation temperature of each pass.The effects of the second phase on microstructure, texture and mechanical properties were investigated.The following conclusions were obtained: (1) DT and IT-430 samples contain dense lamellar LPSO phase and large block LPSO phase inside the grain.
However, there is a slight difference in the lamellar phase spacing between the two.In contrast, the lamellar LPSO phase inside the IT-480 sample is almost dissolved back into the matrix, and only a small amount of finely ground block LPSO phase remains.In addition, a large number of β phase are obviously precipitated compared initial alloy.Among them, IT-430 precipitated the most β phase, followed by DT and IT-480 the least.
(2) The tensile and yield strengths of the material are significantly increased by changing the deformation temperature of the passes.The DT samples reach 394.3 MPa and 301.75 MPa, respectively.
(3) The second phase has an important influence on the DRX behavior of the alloys.β-phase promotes DRX and inhibits its growth by PSN mechanism and pinning grain boundaries.For the intragranular lamellar LPSO phase, alloys with different lamellar spacings show quite different DRX behavior.The degree of grain deformation increases with increasing lamella spacing.Smaller lamella spacing hinders the dynamic recrystallization process.While larger layer spacing facilitates the initiation of various DRX mechanisms and improves the degree of grain refinement.For the block LPSO phase, the larger the size d of the block phase, the stronger its ability to promote dynamic recrystallization.
(4) Appropriate lamella spacing and block phase size can promote the initiation of the slip system and effectively weaken the texture strength.In addition, the rare-earth texture in the alloy is gradually deflected toward ED∥〈10-10〉-〈2-1-10〉 with the increase of lamella spacing and the block phase.And the formation of the anomalous texture is also closely related to the lamellar spacing of the alloy.For the IT-430 sample with dense lamellar phase, the anomalous texture is formed by the combined effect of Un-DRX grains and DRX grains in nonprismatic slip.For the IT-480 sample with large lamellar spacing, the anomalous texture is formed by DRX grains in nonprismatic slip.
(5) The second phase strengthening and fine grain strengthening are dominant in the alloy strengthening mechanism.The yield strength of the second phase alloy provides a significant strengthening effect.With the increase of the interlayer spacing of lamellar LPSO phase and the decrease of the size of block LPSO phase, the strengthening effect of the second phase gradually weakens.Due to the reasonable distribution and size of the second phase, the fine grain strengthening and the second phase strengthening of DT samples are effectively combined, and the UTS and YS of the materials are greatly improved.

Figure 1
Figure 1 (a) The schematic diagram of RUE deformation; (b) the temperature diagram under different RUE deformation; (c) tensile testing bar size; (d) sample location.
have identified them as β-Mg 5 (Gd, Y, Zn) phases with FCC crystal structure.Inverse pole figure (IPF), 0001 pole figure (PF) and Inverse pole figure (IPF) of the initial alloy are shown in Figures 2(b)-(c).

Figure 2 (
d) shows the XRD results of the initial alloy.It can be seen that alloy is mainly composed of α-Mg matrix, β-Mg 5 (Gd, Y, Zn) phase and different Mg 12 (Gd, Y, Zn) phases (block and lamellar shape).Figures 3(a)-(c) shows the microstructure of the samples under different deformation conditions.It can be seen that the microstructures of the three groups of experiments are obviously different.Here the average distance (d) of the lamellar phase was counted using Nano Measurer 1.2 software.The dimensions of the

Figure 3 (
Figure 3(g) is the XRD pattern of the samples under different deformation conditions.According to XRD patterns and other studies[10-14, 16, 18, 19], it can be seen that during the RUE process, the Mg-9.32Gd-3.72Y-1.68Zn-0.72Zralloy is mainly composed of α-Mg matrix, β-Mg 5 (Gd, Y, Zn) phase and different Mg12 (Gd, Y, Zn) phases (block and lamellar shape).Among them, the content of LPSO phase is much higher, which is the main reinforcing phase in the second phase.In order to characterize the grain refinement during different RUE processes, each sample is analyzed using EBSD.Figure4shows the orientation microstructure and the corresponding images of average grain size distribution of different samples.The grains with Grain Orientation Spread (GOS) < 2°are defined as DRX grains; those with 2° GOS 5°are defined as sub-grains and those with GOS > 5°are defined as deformed grains.The black area in the IPF maps is mainly the LPSO phase.The EBSD system cannot recognize the LPSO phase due to the lack of corresponding grain parameters[20,21].After three passes of deformation, the microstructures of the samples in different second phase distribution states are obviously different.the average grain sizes (GS) of DT, IT-430, and IT-480 samples are 24.16,47.47, and 8.59 μm, respectively.The difference in grain sizes of different samples is very huge.The GS of IT-430 samples is significantly larger than that of the remaining two groups.The samples under different deformation parameters exhibited different degrees of DRX (figures 4(d)-(f)).The DRX area fraction was 16.2% for the DT sample, 40.4% for the IT-430 sample, and 81.1% for the IT-480 sample.The average grain size of the DRX grains of the IT-480 sample and the DT do not differ much, at 3.98 μm and 4.86 μm, respectively.The GS ave of the DRX grains of the IT-430 sample (7.57μm), Figure 3(g) is the XRD pattern of the samples under different deformation conditions.According to XRD patterns and other studies[10-14, 16, 18, 19], it can be seen that during the RUE process, the Mg-9.32Gd-3.72Y-1.68Zn-0.72Zralloy is mainly composed of α-Mg matrix, β-Mg 5 (Gd, Y, Zn) phase and different Mg12 (Gd, Y, Zn) phases (block and lamellar shape).Among them, the content of LPSO phase is much higher, which is the main reinforcing phase in the second phase.In order to characterize the grain refinement during different RUE processes, each sample is analyzed using EBSD.Figure4shows the orientation microstructure and the corresponding images of average grain size distribution of different samples.The grains with Grain Orientation Spread (GOS) < 2°are defined as DRX grains; those with 2° GOS 5°are defined as sub-grains and those with GOS > 5°are defined as deformed grains.The black area in the IPF maps is mainly the LPSO phase.The EBSD system cannot recognize the LPSO phase due to the lack of corresponding grain parameters[20,21].After three passes of deformation, the microstructures of the samples in different second phase distribution states are obviously different.the average grain sizes (GS) of DT, IT-430, and IT-480 samples are 24.16,47.47, and 8.59 μm, respectively.The difference in grain sizes of different samples is very huge.The GS of IT-430 samples is significantly larger than that of the remaining two groups.The samples under different deformation parameters exhibited different degrees of DRX (figures 4(d)-(f)).The DRX area fraction was 16.2% for the DT sample, 40.4% for the IT-430 sample, and 81.1% for the IT-480 sample.The average grain size of the DRX grains of the IT-480 sample and the DT do not differ much, at 3.98 μm and 4.86 μm, respectively.The GS ave of the DRX grains of the IT-430 sample (7.57μm),
Figures 4(g)-(i) shows the area fraction of sub grains with their grain size distribution.34.8% for DT sample, while 12.2% for IT-430 sample and 9.9% for IT-480 sample.The sub grain size of DT samples and IT-480 samples are not much different from each other, which are 10.84 μm and 6.44 μm, respectively.And the sub grain size of IT-430 sample is consistent with the DRX GS law, which is obviously larger than the remaining two groups, at 21.46 μm.Figures 4(j)-(l) shows the area fraction of deformed grains with their grain size distribution.Obviously, the area fraction of deformed grains and grain size decrease significantly with the increase of layer spacing of lamellar LPSO phase and the decrease of block LPSO phase size.The area fraction of deformed grains of IT-430 sample is not much different from that of DT sample, which is 43.8% (IT-430) and 40.4% (DT).While only 3.6% remains in the IT-480 sample.The GS ave of the deformed grains is 91.29 μm for IT-430, 27.65 μm for DT, and only 7.61 μm for IT-480.

Figure 5 .
Figure 5. (0001) PF and IPF for different states of the alloy.

4. 1 .
Effects of the second phase on dynamic recrystallization mechanism 4.1.1

Figure 6 .
Figure 6.Uniaxial tensile mechanical properties of different samples.
8(a)-(c).The KAM ave values of the DT sample, the IT-430 sample, and the IT-480 sample are 1.20°, 1.45°, and 0.75°, respectively.The dislocation density of the Un-DRX grains in the respective sample was obtained from equation (1) [30]:

Figure 9 .
Figure 9. Diagram of the mechanism of different samples for DRX.

Figure 10 .
Figure 10.(a) Schmid factor for each sample; (b) The axis of rotation of each slip system.Reprinted from [33], Copyright (2023), with permission from Elsevier.