Investigation and Modelling of Magnesium Alloy Grain Size during Hot Strip Rolling with Inter-Pass Annealing

Abstract

Hot rolling combined with inter-pass annealing is an important method for increasing plastic deformation and refining the grain size of magnesium alloys. In this study, we statistically analyzed the influence of the number of annealing treatments and the annealing holding time (5–20 min) between hot rolling passes on the microstructure of AZ31 alloy after four rolling passes and inter-pass annealing. In addition, an exponential model was proposed to predict the functional relationship between the average grain size and the number of annealing treatments and holding time. After a single rolling pass, the average grain size increased exponentially with increasing holding time. However, increasing the annealing holding time to more than 15 min resulted in a nonhomogeneous microstructure owing to grain coarsening and secondary recrystallization. The grain refinement effect weakened with the increasing number of rolling passes, whereas the microstructural uniformity was significantly improved by multi-pass rolling deformation under the same annealing conditions. The higher amount of grain boundary energy accumulated during multi-pass rolling clearly increased the grains size of the dynamic recrystallized in the early stages with increasing holding time in the subsequent annealing processes. The average grain size was determined to be an exponential function of both the number of annealing treatments and holding time. The predicted average grain diameters were in good agreement with the measured results, thus validating the proposed method for establishing the average grain size in hot deformation processes.

Fig. 5 Grain size after the first rolling pass and annealing treatment with different holding times.

1. Introduction

Magnesium alloys are advanced structural materials with low densities, high strength-to-weight ratios, high thermal conductivities, superior damping capacities, and good electromagnetic shielding characteristics.¹⁾ Therefore, magnesium alloy strips show good prospects for wide use in the electronic, communication, transportation, and aerospace industries, as well as other fields.^1,2) The past decade has witnessed increasing interest in the new rolling processes for developing fine or ultrafine grain microstructures to increase the strength and ductility of magnesium alloys such as cross shear rolling (CSR),^3–5) ARB,^6,7) large strain rolling (LSR),^8,9) high-speed rolling (HSR),¹⁰⁾ hot pack rolling (HPR),¹¹⁾ equal-channel angular rolling (ECAR),¹²⁾ normal and cross-roll rolling (NCR),¹³⁾ wave-shaped die rolling (WSDR),¹⁴⁾ and electro-plastic rolling (EPR).^15,16) However, magnesium alloys have limited slip systems owing to their hexagonal close-packed structure, which considerably limits their plastic deformation at room temperature.¹⁷⁾

Hot rolling is a key process in the production of magnesium alloy strips, because of the high-temperature guarantees large deformation, grain refinement, and formability. To reach the required deformation temperature, these strips are heated between hot rolling passes both in practical production (Fig. 1¹⁸⁾) and in experimental research on magnesium alloys strips. The strip temperature decreases during multi-pass rolling, which makes intermediate heating an important factor governing the rolling efficiency, strength, softening, and microstructure, especially with regard to grain refinement.^19,20)

Fig. 1

Production of magnesium alloy strips.

During the rolling of magnesium alloys, cracking often occurs when the sheets are thinner than 1 mm. However, AZ31 alloys have been successfully thinned from 6 mm to 0.3 mm by 14 rolling passes and intermediate annealing for 5–10 min between passes.²¹⁾ Owing to dynamic recrystallization, the grains were refined from approximately 100 µm to less than 4 µm after two rolling passes. Both static recrystallization and grain growth occurred in the microstructure during the intermediate annealing between passes, leading to a final grain size of approximately 7.5 µm after 12 rolling passes. After 12 passes, grain refinement owing to static recrystallization did not continue, but the recrystallization grain growth during annealing resulted in a final grain size of approximately 7.5 µm.

Perez-Prado et al.²²⁾ used the accumulative roll-bonding method proposed by Saito²³⁾ to investigate the effect of four rolling passes on grain refinement. After a considerable decrease in grain size from 38 to 4.2 µm during the first rolling step, the grain size remained at approximately 3 µm in subsequent rolling passes owing to grain growth during the intermediate heating between passes. Valle et al.²⁴⁾ used a large-strain rolling method to refine the microstructure of the AZ61 alloy. The grain size was reduced from 54 to 15.2 µm after the first rolling pass but then increased to 20.8 µm after heating for 5 min before the second rolling pass. Then, the grain size was refined to 8.6 µm after the second rolling pass, but it increased again to 17.9 µm during the subsequent heating.

Many researchers use intermediate heating in their study on the hot rolling of magnesium alloys, and there have been some reports on the influence of intermediate heating on the microstructure. However, few studies have focused on the influence of the number of annealing treatments and holding times on the microstructure or on modelling the grain size evolution. Thus, this study aims to investigate the influence of these two process parameters on the grain size of magnesium alloys. In addition, we propose a new function to describe the relationship among the grain size, number of annealing treatments, and holding time. This study will be beneficial for the research, control, and optimization of inter-pass annealing and rolling parameters for magnesium alloys during hot strip rolling.

2. Experimental Procedure

The raw material used was the commercial rolled AZ31 alloy (3.1% Al, 0.9% Zn, 0.9% Mn, 0.01% Cu, 0.003% Fe, 0.001% Ni, balance Mg) provided in a casting condition. For the purpose of hot rolling, a homogenization treatment was performed on the slabs at 400°C for 12 h, and plates with dimensions of 100 mm (length) × 60 mm (width) × 4 mm (thickness) were cut from the slab. These samples were then rolled into the final sheets with a thickness of 1.6 mm through four rolling passes, with each pass and total rolling reductions of about 20% and 60%, respectively.

Considering the temperature decrease of the strips owing to air cooling during rolling, the first annealing and inter-pass annealing temperatures were set to 350°C with heating speed of 30°C/min, and the measured rolling temperature remained in the range of 320–340°C. The roller keeps ambient temperature and the rolling speed was set to 0.5 m/s. The samples have been put into holding furnace immediately after rolling, and the time is recorded subsequent to furnace temperature rises to 350°C automatically. The holding time during the inter-pass annealing was varied to 5, 10, 15, and 20 min. The experimental procedure of the rolling and inter-pass annealing is shown in Table 1, where the prefix R or RA is labeled with the number of inter-pass annealing treatments, and X = 1, 2, 3, 4 indicate holding times of 5, 10, 15, and 20 min, respectively. For example, sample RA2-2 was prepared by two rolling passes and two inter-pass annealing treatments with a holding time of 10 min, and R2-2 was prepared by two rolling passes but only one inter-pass annealing treatment for 10 min. A total of 25 samples were tested for each set of conditions. The samples were immediately quenched with water after rolling or inter-pass annealing to obtain a deformed or annealed microstructure, respectively.

Table 1 Experimental procedure and sample labels for rolling and inter-pass annealing.

The microstructures of the alloys were examined using a DMI5000M optical microscope (OM). The metallographic samples were ground, polished, and chemically etched in an acetic–picric solution (5 g picric acid, 10 mL acetic acid, 10 mL distilled water, and 80 mL ethanol) for approximately 15 s. The grain size and its distribution were statistically analyzed according to the method for estimating the average grain size of metals described in GBT6394-2017.²⁵⁾

3. Results and Discussion

3.1 Influence of the holding time on grain size

To study the homogeneity of the microstructure, the average grain size D_a and the variance of grain size $S_{D}^{2}$ of the samples were calculated according to the following equation:   

\begin{equation} \left\{ \begin{array}{l} D_{a} = \dfrac{1}{n}\displaystyle\sum_{i = 1}^{n}D_{i}\\ S_{D}^{2} = \dfrac{1}{n}\displaystyle\sum_{i = 1}^{n}(D_{i}-D_{a})^{2} \end{array} \right. \end{equation} (1)

where n is the number of statistically analyzed grains, and D_i is the diameter of each individual grain (µm).

OM images of the microstructures of the AZ31 alloy, after homogenization treatment for 3 h at 430°C (Fig. 2), showed that the microstructure is mainly composed of equiaxed grains with a few grains coarsened by the homogenization treatment. The maximum grain size is approximately 80–90 µm, and the minimum grain size is less than 5 µm. Approximately 27% of the grains ranged between 20 and 30 µm in size, and the statistical analysis of the grain size distribution indicated a D_a value of approximately 33 µm.

Fig. 2

Microstructure (a) and grain size distribution (b) of the initial magnesium alloy.

The microstructure and grain size distribution after the first rolling pass are shown in Fig. 3. After a single rolling pass, the grain boundary exhibits stripes (Fig. 3(a)) owing to the shear effect produced by grain boundary slip. In addition, some lath twins and deformation zones appear in the grains, and the twinning is the main mechanism of plastic deformation under these conditions. The rapid temperature reduction, small plastic deformation, and coarse original grains led to an insufficient driving force for dynamic recrystallization and nucleation; thus, only a small amount of dynamic recrystallization occurred during the first rolling pass. The larger dislocation slip in the coarser grains resulted in a severe stress concentration near the grain boundary; thus, twinning occurred mainly in these grains. The plastic deformation stress is attributed to the compressive stress occurring during the rolling process, and the main twins in the thickness direction are on the $\{ 10\bar{1}1\} $ plane. When the shear and compressive stresses were large, a secondary twin on the $\{ 10\bar{1}2\} $ plane was produced in the $\{ 10\bar{1}1\} $ compression twin.^18,26) Furthermore, some grains in the twin zone dynamically recrystallized owing to the distortion energy produced by interactions and cutting among the primary and secondary twins. More than 40% of the grains changed size to be in the range of 20–30 µm, and D_a was approximately 22.8 µm. Thus, the dynamic recrystallization was not sufficient during a single rolling pass, resulting in limited grain refinement.

Fig. 3

Microstructure (a) and grain size distribution (b) after a single rolling pass.

The microstructures of the magnesium alloy after the first rolling pass and annealing treatments for different holding times are shown in Fig. 4. Sub-dynamic and static recrystallization clearly occurred after heat preservation annealing. In addition, significant grain refinement is noted. On the one hand, a small amount of sub-dynamic recrystallization occurred because of dynamic recrystallization nucleation at high temperatures. On the other hand, the distortion energy produced during rolling and annealing promoted the occurrence of static recrystallization. With increasing holding time, the number of twins decreased, whereas the number and size of the recrystallized grains increased. After the first rolling pass and annealing for 5 min, because of less distortion energy release and insufficient recrystallization within shorter time, the microstructure consisted of some twins, coarse grains and statically recrystallized grains, and the coarse grains were surrounded by these fine, recrystallized grains. Therefore, static recrystallization after plastic deformation involved the nucleation and growth of refined grains at the boundaries of larger grains, thus consuming the original coarse grains and twins. Compared with 5 min, when the holding time was increased to 10 min, there was enough time to release distortion energy and complete the static recrystallization occurred leads to the twins disappeared and relative homogeneous microstructure. Despite the growth of some fine recrystallized grains, the microstructural uniformity and grain refinement significantly improved. When the holding time was further increased to 15 min or more, secondary recrystallized grains appeared in the microstructure, and the grains that recrystallized earlier decreased the microstructural homogeneity.

Fig. 4

Microstructure of the magnesium alloy for different annealing holding times after the first rolling pass: (a) 5 min; (b) 10 min; (c) 15 min; (d) 20 min.

Figure 5 shows the grain size distribution for different holding times after a single rolling pass. Evidently, the grains were significantly refined with respect to the original structure after a single rolling pass and annealing. In addition, the microstructure consisted of fewer dynamically recrystallized grains and a large number of twins. Static recrystallization was clearly induced by these twins during annealing, and D_a was approximately 10.1 µm. However, a few coarse grains remained (>50 µm), and the microstructure was nonhomogeneous with an $S_{D}^{2}$ of 97.65. D_a increased exponentially with increasing holding time. Specifically, when the holding time was 10 min, D_a and $S_{D}^{2}$ were 15.2 µm and 43.67, respectively, thus demonstrating a more homogenous microstructure. Grain coarsening and secondary recrystallization occurred when the holding time was 15 min or greater owing to the decreased homogeneity. Therefore, considering the heating efficiency, microstructural homogeneity, and grain refinement, a 10 min inter-pass annealing treatment is optimal after first rolling for the multi-pass hot rolling of the magnesium alloy.

Fig. 5

Grain size after the first rolling pass and annealing treatment with different holding times.

3.2 Influence of the number of inter-pass annealing treatments on grain size

Figure 6 shows the microstructure obtained after multi-pass rolling and inter-pass annealing for 10 min. After rolling and quenching, the microstructure of the magnesium alloy strip clearly consisted of numerous twins, a small number of dynamically recrystallized grains, and a slice of the original grains. In addition, several square and necklace-shaped grains existed in the crossing twin zones, which were produced by the initial and secondary twins, and the twin grains were significantly refined. The distortion energy and twins caused by plastic deformation during rolling refined a multitude of equiaxed grains in the subsequent annealing treatment. With the increase in the number of rolling passes and the accumulation of plastic deformation, the density of twins in the microstructure after rolling and the volume fraction of static recrystallization increased significantly during the subsequent annealing processes. On the one hand, the twins produced by plastic deformation and the accumulated total strain considerably promoted static recrystallization and grain refinement. On the other hand, after annealing, dynamic recrystallization easily occurred during the rolling when the sample contained refined grains. Furthermore, the higher grain boundary energy that accumulated owing to the refined grains and deformation clearly facilitated the growth of the recrystallized grains during the re-annealing process. Therefore, when the holding time of the inter-pass was constant, the grain refinement effect was weakened with the increase of the rolling passes.

Fig. 6

Microstructure after multi-pass rolling and inter-pass annealing for 10 min: (a) R2; (b) RA2; (c) R3; (d) RA3; (e) R4; (f) RA4.

Figure 7 shows the grain size distribution obtained after multi-pass rolling with inter-pass annealing for 10 min. The grains were significantly refined with increasing number of rolling passes. In particular, more than 30% of the grains had diameters between 10 and 20 µm. In addition, the value of D_a decreased in the order of RA1-2, RA2-2, RA3-2, and RA4-2, from 33.1 µm for the original grains to 15.7, 14.3, 13.1, and 13.3 µm, respectively. The coarse grains, which were similar in size to the original grains, almost disappeared for the RA3-2 and RA4-2 samples. In addition, the value of D_a for RA4-2 is very similar and even coarser to that for RA3-2 (relative value not more than 1%) results probably from more annealing times, longer holding time or higher temperature, but $S_{D}^{2}$ decreased continuously with increasing number of rolling passes. Therefore, under the same annealing conditions, grain refinement became less effective with increasing number of rolling passes, whereas the microstructure homogeneity continued to improve significantly. Hence, grain refinement and microstructural homogeneity can be obtained only by reasonably adjusting and optimizing the number of rolling passes and the inter-pass annealing holding time.

Fig. 7

Grain size distribution (a) and (b) D_a and “$S_{D}^{2}$” after multi-pass rolling with inter-pass annealing for 10 min.

4. Modelling of the Grain Size

4.1 Grain size analysis

Figure 8 shows the microstructure of the magnesium alloy strips after multi-pass rolling with different inter-pass annealing holding times. With increasing number of rolling passes, the distortion energy increases owing to the accumulated plastic strain, which leads to significantly more grain refinement after annealing for the same holding time. Furthermore, with the same number of rolling passes, increasing the annealing holding time clearly increases the size of recrystallized grains. After one rolling pass with subsequent annealing for 5 min, some original coarse grains and twins appear, whereas twins no longer appear in the microstructure with increasing number of rolling passes or annealing holding time. After the second rolling pass and subsequent inter-pass annealing for 5 min, the twins no longer appear. Meanwhile, equiaxed grains obtained are significantly finer because the annealing time is too short to promote the growth of recrystallized grain. With increasing number of rolling passes, more accumulated plastic strain results in further grain refinement, but the effect of the refinement is considerably weakened. After the first rolling pass and subsequent annealing for 10 min, the twins no longer appear, and equiaxed recrystallized grains are obtained. However, these recrystallized grains were larger than those that were annealed for 5 min. When the holding time increases to 15 min or longer, there is sufficient time for grains to grow, and the effect of the increased plastic deformation owing to multi-pass rolling on the grain refinement is evidently weakened. Therefore, increasing the holding time and number of annealing treatments is not conducive to further refining the grains of magnesium alloy strips during multi-pass rolling.

Fig. 8

Microstructure after multi-pass rolling with different inter-pass annealing holding times: (a) 5 min; (b) 10 min; (c) 15 min; (d) 20 min.

Figure 9 shows the changes in the average grain size with different rolling and annealing conditions. Evidently the annealing holding time of 5 min is insufficient for the recrystallized grains to grow and several grains smaller than 10 µm appear in the microstructure. However, a few original coarse grains larger than 80 µm remain (Fig. 9(a) and (b)) because of the lower plastic deformation and incomplete recrystallization. With increasing number of rolling passes, the number of the larger grains decreases significantly, thereby decreasing the average grain size from 33.1 µm for the initial sample to 7.7 µm for RA3-1. A continuously increasing the number of annealing times leads to grain refinement efficiency decreasing and even microstructural coarsening, as similar as the condition of 10 min holding time. With holding times of 10 and 20 min, approximately 60% and 50% of the grains have sizes in the range of 10–20 µm, respectively. Thus, grains obviously grow during the annealing process, thus weakening the grain refinement effect of the subsequent rolling passes and annealing treatments. When the annealing holding time is extended to 20 min, the recrystallized grains grow drastically, and the microstructure is severely coarsened. For example, for RA1-4, approximately 40% of the grains range from 20 to 30 µm in size, and a few grains are even larger than 60 µm. Therefore, increasing the annealing holding time leads to grain growth and secondary recrystallization, thereby degrading the uniformity of the grain size during the multi-pass rolling. Furthermore, even with the weakening effect of the rolling passes, the grain refinement and homogeneity of the microstructure significantly improved. Therefore, the inter-pass annealing holding time for AZ31 magnesium alloy during hot rolling should be less than 15 min. Further, with increasing number of rolling passes, the inter-pass annealing and holding times should be gradually reduced to refine significantly the grain size and attain a homogeneous microstructure. As an example, the microstructure obtained with 10 min holding time after first rolling pass but then 5 min after the second to fourth rolling passes is beneficial certainly for the grain refinement and homogeneous, compared with these of fixed annealing holding time 5 min or 10 min.

Fig. 9

Grain size after different rolling passes and annealing holding times: (a), (b) 5 min; (c), (d) 15 min; (e), (f) 20 min.

4.2 Prediction model

Figure 10 shows the average grain size of the magnesium alloy with different inter-pass annealing conditions. Clearly, the average grain size decreases exponentially with increasing number of inter-pass annealing treatments and the decreasing holding time. Compared with the initial grains, which had a size of 33.1 µm, the grains are significantly more refined after rolling deformation. Figure 10(a) demonstrates that the grains are significantly refined by the initial rolling passes when the holding time is shorter than 10 min, indicating that the lower activation energy of the fine grains is beneficial to the subsequent dynamic recrystallization and plastic deformation processes. With increasing number of rolling passes and inter-pass annealing holding times, the average grain size continues to decrease, but the microstructural refinement effect is weakened owing to grain growth. The average grain size decreases from 33.1 to 10.1 µm for RA1-1, which further decreases to 9.4 and 7.8 µm for RA2-1 and RA3-1, respectively. However, finer grains with higher grain boundary energies facilitate grain growth, resulting in an average grain size of 8.1 µm for RA4-1. With increasing annealing holding time, the recrystallized grains clearly continue to grow. Secondary dynamic recrystallization during subsequent rolling passes and grain growth during subsequent annealing treatments occurred more easily in these refined grains than in those during earlier rolling passes. Therefore, inter-pass annealing is of great importance for subsequent rolling processes in terms of obtaining finer recrystallized grains, reaching the required deformation temperature, and eliminating the deformation stress of the previous pass. However, repeated inter-pass annealing treatments limit the grain refinement and lower the rolling efficiency. Therefore, a descending approach to the number of annealing treatments and holding times should be employed to improve the grain refinement, rolling efficiency, and plastic deformation.

Fig. 10

Average grain size as a function of the (a) number and (b) holding time for inter-pass annealing.

According to the analysis and fitting results in Fig. 10, the average grain size D_a and original grain size D₀ are expressed as a function of the number of annealing treatments and holding times as follows:   

\begin{equation} \frac{D_{a}}{D_{0}} = A \times (R_{p})^{B} \times (t_{a})^{C} \end{equation} (2)

where A is a material constant, B and C are material exponential coefficients, R_p is the number of annealing treatments, and t_a is the annealing holding time.

Taking the natural logarithm to both sides of eq. (2) yields:   

\begin{equation} \ln (D_{a}/D_{0}) = \ln A + B\ln (R_{p}) + C\ln (t_{a}) \end{equation} (3)

where B = ∂ ln(D_a/D₀)/∂ ln(R_p) and C = ∂ ln(D_a/D₀)/∂ ln(t_a). The exponential coefficients B and C can be calculated as −0.18779 and 0.644258, respectively, according to the average values of the slope of the linear fitting lines in Fig. 11. In addition, A can be determined to be 0.11682 from the values of B and C and the slope of the linear fitting line (Fig. 11(c)).

Fig. 11

Relational curves among the different variables.

Based on our findings and the phenomenological representation of the shapes of the fitting curves, functional models were investigated to describe the average grain size during multi-pass rolling and annealing:   

\begin{equation} D_{a} = 0.117D_{0} \cdot (R_{p})^{-0.188} \cdot (t_{a})^{0.644} \end{equation} (4)

Figure 12 compares the average grain size predicted by the constructed model and the measured values. The predicted results are clearly in good agreement with the actual measured values. The maximum relative error is 21.86%, and the average predicted relative error is approximately 8.54%, which validates the reliability of established model in predicting the average grain size and optimizing the process parameters for the hot rolling of magnesium strips. Furthermore, the proposed model for predicting the evolution of average size and obtaining material constants is confirmed to be reliable for hot deformation processes.

Fig. 12

Comparison between the predicted and measured average grain size.

5. Conclusions

This study investigated the influence of the number of annealing treatments and holding time on the grain size of magnesium alloys during hot rolling. Further, a new function was proposed to describe the relationship among the grain size, number of annealing treatments, and holding time. The conclusions are as follows:

1. (1)    The fast temperature decrement, low plastic deformation and coarser original grains lead to the deficiency of dynamic recrystallization nucleation ability. In addition, some lath twins and a few deformation zones existed in the microstructure after a single rolling pass. In the annealing process, the grains were significantly refined by sub-dynamic and static recrystallization. With increasing holding time, the number of twins decreased, while the number of recrystallized grains increased. Further, D_a increased exponentially with increasing holding time. With a holding time of 10 min, D_a and $S_{D}^{2}$ were 15.2 and 43.67 µm, respectively, demonstrating the improved microstructural homogeneity of the sample. In addition, secondary recrystallized grains appeared, and the finer grains grew remarkably when the holding time was increased to more than 15 min.
2. (2)    With increasing number of rolling passes and the accumulation of plastic deformation, the density of twins in the microstructure and the volume fraction of the statically recrystallized grains significantly increased during the annealing processes. The optimal inter-pass annealing holding time for the hot strip rolling of the magnesium alloy is less than 15 min. Furthermore, with increasing number of rolling passes, the number of inter-pass annealing treatments and holding time should be gradually decreased to refine significantly the grain size and homogenize the microstructure.
3. (3)    The average grain size decreased exponentially with increasing number of the inter-pass annealing treatments and the decreasing holding time. The reliability and accuracy of the established model for predicting the average grain size and optimizing processes parameters were thus validated. Further, the proposed methods for modelling the evolution of the average size and obtaining the material constants for hot deformation were found to be reliable.

Acknowledgement

The authors gratefully acknowledge financial support from the Natural Science Foundation-Steel and Iron Foundation of Hebei Province (No. E2018501016), Fundamental Research Funds for the Central Universities (No. N172304042), as well as the Doctoral Scientific Research Foundation of Liaoning Province (No. 20170520314).

REFERENCES

• 1)   S.  Agarwal,  J.  Curtin,  B.  Duffy and  S.  Jaiswal: Mater. Sci. Eng. C 68 (2016) 948–963. doi:10.1016/j.msec.2016.06.020
• 2)   Y.  Estrin and  A.  Vinogradov: Acta Mater. 61 (2013) 782–817. doi:10.1016/j.actamat.2012.10.038
• 3)   H.  Watanabe,  T.  Mukai and  K.  Ishikawa: J. Mater. Sci. 39 (2004) 1477–1480. doi:10.1023/B:JMSC.0000013922.16079.d3
• 4)   S.H.  Lee and  D.N.  Lee: Int. J. Mech. Sci. 43 (2001) 1997–2015. doi:10.1016/S0020-7403(01)00025-X
• 5)   W.J.  Kim,  J.B.  Lee,  W.Y.  Kim,  H.T.  Jeong and  H.G.  Jeong: Scr. Mater. 56 (2007) 309–312. doi:10.1016/j.scriptamat.2006.09.034
• 6)   F.  Schwarz,  C.  Eilers and  L.  Krüger: Mater. Charact. 105 (2015) 144–153. doi:10.1016/j.matchar.2015.03.032
• 7)   L.S.  Huang,  X.B.  Pei,  L.  Luo,  X.X.  Rao,  Y.H.  Jing and  Y.  Liu: Mater. Trans. 61 (2020) 893–902. doi:10.2320/matertrans.MT-MF2019051
• 8)   M.V.  Markushev,  D.R.  Nugmanov,  O.  Sitdikov and  A.  Vinogradov: Mater. Sci. Eng. A 709 (2018) 330–338. doi:10.1016/j.msea.2017.10.008
• 9)   F.  Guo,  D.F.  Zhang,  X.S.  Yang,  L.Y.  Jiang,  S.S.  Chai and  F.S.  Pan: Mater. Sci. Eng. A 607 (2014) 383–389. doi:10.1016/j.msea.2014.04.024
• 10)   J.  Su,  M.  Sanjari,  A.S.H.  Kabir,  I.H.  Jung and  S.  Yue: Scr. Mater. 113 (2016) 198–201. doi:10.1016/j.scriptamat.2015.10.040
• 11)   T.  Wu,  L.  Jin,  W.X.  Wu,  L.  Gao,  J.  Wang,  Z.Y.  Zhang and  J.  Dong: Mater. Sci. Eng. A 584 (2013) 97–102. doi:10.1016/j.msea.2013.07.011
• 12)   W.J.  Kim,  S.J.  Yoo,  Z.H.  Chen and  H.T.  Jeong: Scr. Mater. 60 (2009) 897–900. doi:10.1016/j.scriptamat.2009.02.005
• 13)   D.G.  Kim,  K.M.  Lee,  J.S.  Lee,  Y.O.  Yoon and  H.T.  Son: Mater. Lett. 75 (2012) 122–125. doi:10.1016/j.matlet.2012.01.141
• 14)   H.Y.  Wang,  T.T.  Feng,  L.  Zhang,  C.G.  Liu,  Y.  Pan,  M.  Zha,  X.L.  Nan,  C.  Wang and  Q.C.  Jiang: Mater. Des. 88 (2015) 157–161. doi:10.1016/j.matdes.2015.08.154
• 15)   Z.H.  Xu,  G.Y.  Tang,  S.Q.  Tian,  F.  Ding and  H.Y.  Tian: J. Mater. Process. Technol. 182 (2007) 128–133. doi:10.1016/j.jmatprotec.2006.07.019
• 16)   X.P.  Li,  F.  Wang,  X.H.  Li,  G.Y.  Tang and  J.  Zhu: Mater. Sci. Eng. A 618 (2014) 500–504. doi:10.1016/j.msea.2014.09.060
• 17)   H.  Miyano,  K.  Takemoto,  M.  Tsushida,  H.  Kitahara and  S.  Ando: Mater. Trans. 62 (2021) 1097–1101. doi:10.2320/matertrans.L-M2021826
• 18)  Z.H. Chen: Wrought Magnesium Alloy, (Chemical Industry Press, Beijing, 2005) pp. 239–240 (in Chinese).
• 19)   W.Y.  Kim and  W.J.  Kim: Mater. Sci. Eng. A 594 (2014) 189–192. doi:10.1016/j.msea.2013.11.066
• 20)   W.Z.  Chen,  W.C.  Zhang,  L.X.  Zhang and  E.D.  Wang: J. Alloy. Compd. 646 (2015) 195–203. doi:10.1016/j.jallcom.2015.06.112
• 21)   P.  Yang,  L.  Meng,  W.M.  Mao and  Q.W.  Cai: Trans. Mater. Heat Treat. 26 (2005) 34–39 (in Chinese).
• 22)   M.T.  Pérez-Prado,  D.  Valle and  O.A.  Ruano: Scr. Mater. 51 (2004) 1093–1097. doi:10.1016/j.scriptamat.2004.07.028
• 23)   Y.  Saito,  H.  Utsunomiya,  N.  Tsuji and  T.  Sakai: Acta Mater. 47 (1999) 579–583. doi:10.1016/S1359-6454(98)00365-6
• 24)   J.A.  del Valle,  M.T.  Pérez-Prado and  O.A.  Ruano: Mater. Sci. Eng. A 355 (2003) 68–78. doi:10.1016/S0921-5093(03)00043-1
• 25)  L.J. Chen, Y. Luan, Q. Gu, X.H. Ju and C. Yu: Determination of estimating the average grain size of metal, (China National Standardization Management Committee and General Administration of Quality Supervision, Inspection and Quarantine of the People’s Republic of China, 2017) pp. 5–17 (in Chinese).
• 26)   Y.J.  Wu and  R.  Zhu: J. Mater. Eng. Perform. 28 (2019) 6182–6191. doi:10.1007/s11665-019-04363-x