2025 Volume 65 Issue 11 Pages 1725-1734
Primary recrystallization texture strongly influences the magnetic properties of grain-oriented electrical steel through secondary recrystallization. In this study, the effect of applied tensile stress during annealing on grain growth and primary recrystallization texture formation in Fe-3%Si alloy was investigated. It was revealed that grain growth is promoted under the condition of applied tension, and the intensity of {411}<148> orientation increases, while the intensity of {111}<112> orientation decreases. The changes in grain diameters and textures are explained by the normal grain growth with the size advantage of larger {411}<148> grains and disadvantage of smaller {111}<112> grains than the average sized grains. Moreover, KAM value of {111}<112> grains was confirmed to be larger than that of {411}<148> grains after tension annealing. This suggests that the stored energy in {111}<112> grains is larger than that in {411}<148> grains, which would promote the selective growth of {411}<148> grains with SIBM mechanism by consuming {111}<112> grains with relatively higher stored energy.
Grain-oriented electrical steel is a magnetic material mainly used in transformer cores. It exhibits excellent magnetic properties in one direction owing to secondary recrystallization, which sharpens {110}<001> orientation, known as Goss orientation. For abnormal grain growth—where secondary recrystallization of sharp Goss grains occurs—the inhibitors that suppress the growth of non-Goss grains, as well as the primary recrystallization texture and grain size, must be controlled.1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16) The coincidence site lattice (CSL) boundary model states that grains with a Σ9 CSL relationship to Goss orientation are important because they promote the preferential growth of Goss oriented grains. This model has been proposed as a texture-control technology in secondary recrystallization.2,3,4,5,6,7,8,9,10,11,14,15,16) {411}<148> and near {111}<112> orientations exhibit a Σ9 CSL relationship to Goss orientation, and these orientations must be incorporated into the primary recrystallization texture to achieve sharp secondary recrystallization. Yasuda et al.17) investigated the development mechanism of {411}<148> orientation in Fe-3%Si alloy steel. They proposed that {411}<148> recrystallized grains have a relatively large driving force for grain growth because they are adjacent to recovered grains with high dislocation density, belonging to the α-fiber texture with the RD//<011> direction (where RD denotes the rolling direction). Moreover, the grain size of {411}<148> recrystallized grains increases in the later stages of primary recrystallization. Furthermore, they clarified that during the subsequent normal grain growth, the larger {411}<148> grains tend to grow further owing to their size advantage,18) consuming surrounding smaller grains.17) In addition, Yasuda et al.19) investigated the origin of {411}<148> recrystallized grains and estimated that some grains near {211}<011> orientation—a stable orientation in cold rolling—rotate into {411}<148> orientation and serve as nuclei for {411}<148> grains during primary recrystallization.
However, in the industrial production of grain-oriented electrical steel, various external factors influence the steel sheet due to production efficiency and equipment constraints. One such factor is the tension applied to the steel sheet during annealing. In the annealing processes such as primary recrystallization, continuous annealing, where the coil is continuously uncoiled and annealed, instead of batch annealing, is extremely effective for efficient production. However, during continuous annealing, unavoidable tension is applied to the steel sheet, which may affect texture formation during primary recrystallization and subsequent grain growth. Few studies have investigated the effect of tension during annealing on texture and grain size. Hutchinson20) reviewed various external factors affecting texture formation and noted that no clear relationship was found between stress and recrystallization behavior in studies where elastic stress was applied to steel sheets during annealing. By contrast, Onuki et al.21,22,23) confirmed that, under plastic stress, the area fraction of {100} grains increases and that of {111} grains decreases as uniaxial compression deformation is applied at higher temperatures and lower strain rates, suggesting that {100} grains with a low Taylor factor preferentially grow. Additionally, they estimated that the migration of high-angle grain boundaries between {100} and {111} grains is influenced by the low-angle grain boundaries within each grain.21,22,23) However, to the best of our knowledge, no studies have been conducted on texture changes during high-temperature annealing with tensile deformation in Fe-3%Si alloy. Therefore, this study investigates the effect of applied tension during the grain growth stage on grain growth behavior and texture development in Fe-3%Si alloy.
A hot-rolled sheet with a thickness of 2.6 mm and a composition of Si: 3.4%, C: 0.06%, Mn: 0.1%, S: 0.01%, Al: 0.03%, and N: 0.01% (mass%) was annealed, cold rolled to 0.23 mm, and subjected to primary recrystallization annealing to obtain primary recrystallized grains, where decarburization was also achieved. The primary recrystallization annealing was conducted at 800, 820, and 850°C for 90 s to produce three distinct grain size levels. Under all conditions, the C content of the primary recrystallized sheet was less than 0.005%. Subsequently, the sheets were annealed at 850°C for 120 s under varying tensions (tensile stresses) in a hoop furnace. After air cooling, the microstructure and texture of the steel sheets were analyzed. The load applied to the sheet was measured at the exit side of the annealing furnace, and the tension was controlled by adjusting the torque at the payoff reel on the entrance side while winding the sheet at a constant speed. The applied tension was set at 9.8, 19.6, and 24.5 MPa. A condition of 0 MPa (no tension) was achieved by annealing the sample on a mesh belt. Before annealing, parallel lines perpendicular to the direction of applied tension were drawn on the steel sheet at approximately 250 mm intervals. The elongation of the sample due to tension annealing was determined by measuring the distance between these lines before and after annealing using a caliper.
The microstructures of the steel sheets were investigated by mechanically polishing the normal direction (ND)–RD cross-section, followed by etching with nital and observation using an optical microscope. Vickers hardness measurements were performed with a load of 0.98 N and a holding time of 5 s, measuring five points each. Textures were analyzed by mechanically polishing the transverse direction (TD)–RD cross-section of the samples to a half-thickness layer, followed by electrolytic polishing. Pole figures of {200}, {220}, and {222} planes were measured using X-ray diffraction (XRD; RINT2500HF, Rigaku Corp., Japan) to obtain the orientation distribution function (ODF). Detailed texture and grain structure analyses were conducted using scanning electron microscopy (SEM; JSM-800HL, JEOL Ltd., Japan) and electron backscatter diffraction (EBSD; EDAX OIM Data Collection, AMETEK Inc., America). The measurement area for each field of view was set to 1000 μm × 600 μm, with a step interval of 2 μm, considering the grain size. The observation fields included between 839 and 2579 grains. Although fewer than 1000 grains were present under some conditions, the average texture results obtained from EBSD were considered reliable.24)
The results of the sheet elongation after tension annealing are shown in Fig. 1. Elongation increased with increasing tension, regardless of the primary recrystallization annealing temperature. Although annealing was performed under low tension for a short duration, plastic deformation occurred during annealing. The elongation values were similar for primary recrystallization temperatures of 800 and 820°C. However, at 850°C, the elongation was slightly lower at tensions of 19.6 and 24.5 MPa compared to the other conditions.
The microstructures of the steel sheets observed by optical microscope after primary recrystallization and tension annealing are shown in Fig. 2. Grain size varied depending on the primary recrystallization temperature: the higher the temperature, the larger the grain size. After subsequent tension annealing at 850°C, grain growth occurred under all conditions. Moreover, grain growth increased with higher tension, particularly in the 800°C-annealed material, which had a smaller grain size after primary recrystallization annealing. No cementite was observed in the microstructure under any condition.
An example of the inverse pole figure (IPF) maps measured by EBSD on the RD–TD cross section is shown in Fig. 3. Figure 3(a) shows the IPF map of the sample annealed at 800°C for primary recrystallization. Figure 3(b) presents the IPF map of the same sample after subsequent annealing at 850°C for 120 s without tension. Figure 3(c) shows the IPF map of the sample subsequently annealed at 850°C under a tension of 24.5 MPa. Even without tension, grain growth occurred after annealing at 850°C for 120 s, along with an increase in pink-colored {411} grains and a decrease in blue-colored {111} grains. These tendencies were more pronounced under the application of 24.5 MPa tension, as shown in Fig. 3. Low-angle grain boundaries of 2°–15°, indicated in white in the figure, were observed in both {411} and {111} grains. The quantitative changes in average grain size are shown in Fig. 4. When the primary recrystallization annealing temperature was 800°C, the grain size was approximately 15.0 μm. At 820 and 850°C, the grain size increased to approximately 16.9 and 20.7 μm, respectively, indicating that higher annealing temperatures lead to larger grain sizes. Although some variation was observed, the results confirmed that grain size after tension annealing increased with higher applied tension. When the primary recrystallization annealing temperature was 820°C, the grain size after subsequent annealing at 850°C without tension was approximately 21.2 μm, while it increased to approximately 24.6 μm under a tension of 24.5 MPa. This result is qualitatively consistent with the optical microscopy observations shown in Fig. 2. The changes in average grain size due to annealing with and without tension (non-tension annealing) are listed in Table 1. The changes due to tension annealing were calculated as the differences between the average grain sizes obtained under 19.6 and 24.5 MPa tension and those obtained under non-tension annealing, accounting for variations. The changes in grain size during non-tension annealing were attributed to normal grain growth and were 5.8, 4.3, and 2.2 μm for primary recrystallization temperatures of 800, 820, and 850°C, respectively. By contrast, the changes in average grain size between tension and non-tension annealing—attributed to the effect of tension—were 2.0, 2.7, and 0.8 μm for primary recrystallization temperatures of 800, 820, and 850°C, respectively. These values were approximately half of the grain-size changes caused by normal grain growth.
| Samples | Change in average grain diameters by 850°C annealing without tension | Difference average grain diameters by 850°C annealing between without/with tension |
|---|---|---|
| Recrystallization annealed at 800°C | 5.8 μm | 2.0 μm |
| Recrystallization annealed at 820°C | 4.3 μm | 2.7 μm |
| Recrystallization annealed at 850°C | 2.2 μm | 0.8 μm |
The Vickers hardness values measured at the mid-thickness layers are shown in Fig. 5. After primary recrystallization, the Vickers hardness increased at lower annealing temperatures, which is likely due to smaller grain sizes. However, after tension annealing, the Vickers hardness varied, and no clear trends related to either the primary recrystallization annealing temperature or applied tension were observed. The indentation size during Vickers hardness testing was approximately 30 μm, which is similar to the average grain size of 15–25 μm. This suggests that the variation may have been increased owing to changes in the amount of strain in each grain and the effect of grain boundaries. As shown in Fig. 4, lower primary recrystallization temperatures resulted in more grain growth after tension annealing at 850°C. This resulted in smaller differences in grain size after tension annealing, which likely contributed to the absence of a clear trend. Furthermore, although higher tension generally promoted grain growth, the introduction of strain during tension annealing also influenced the variation in Vickers hardness.
The ODFs (ϕ2 = 45° section) for the mid-thickness layer, measured by XRD after primary recrystallization annealing and tension annealing, are shown in Fig. 6. Under all conditions, high intensities were observed in {h,1,1}<1/h,1,2> fiber texture, centered around {411}<148>–{100}<012> orientations, and in the γ-fiber texture, consisting of ND//<111> centered around {111}<112> orientation. Furthermore, the texture changed with tension annealing; the intensity of {411}<148> orientation increased with increasing tension, while {111}<112> intensity decreased.
To further clarify these textural changes, ODF intensity difference maps before and after tension annealing are shown in Fig. 7. At a primary recrystallization annealing temperature of 800°C, subsequent non-tension annealing at 850°C increased {411}<148> intensity and decreased {111}<112> intensity, as shown in Fig. 7(a1). Furthermore, the greater the tension applied during annealing, the higher the intensity of {411}<148> orientation and the lower the intensity of {111}<112> orientation became. By contrast, at primary recrystallization annealing temperatures higher than 800°C, the texture changes due to subsequent non-tension annealing at 850°C were minimal. For example, for a primary recrystallization annealing temperature of 850°C, the texture changes were insignificant after non-tension annealing, as shown in Fig. 7(c1). However, under tensions of 19.6 MPa or greater, noticeable changes were observed, with an increase in the intensity of {411}<148> orientation and a decrease in the intensity of {111}<112> orientation, as shown in Figs. 7(c3) and 7(c4).
Figure 8(a) shows the changes in the intensity of {411}<148> orientation with tension annealing, and Fig. 8(b) shows the changes in the intensity of {111}<112> orientation. The figures show that the intensity of {411}<148> orientation after tension annealing increased with increasing tension, while the intensity of {111}<112> orientation decreased. Figure 9 shows the changes in the intensities of {411}<148> and {111}<112> orientations before and after tension annealing. As shown in Fig. 9(a), the changes in the intensity of {411}<148> orientation were the largest at a primary recrystallization annealing temperature of 800°C, followed by 820 and 850°C, under all tension conditions. Regardless of the primary recrystallization annealing temperature, the changes in the intensity of {411}<148> orientation were largest under a tension of 24.5 MPa. As shown in Fig. 9(b), the intensity of {111}<112> orientation decreased the most under a tension of 24.5 MPa.
The changes in {411}<148> and {111}<112> orientations due to non-tension and tension annealing are shown in Table 2. Here, the changes in the texture due to tension annealing were evaluated as the difference between the average values of the intensities under the annealing conditions without tension and those with tensions of 19.6 and 24.5 MPa, considering the variation. The changes in {411}<148> orientation due to non-tension annealing, i.e., normal grain growth, were 1.1, 0.4, and 0.3 for primary recrystallization temperatures of 800, 820, and 850°C, respectively. By contrast, the changes in {411}<148> orientation due to tension annealing, that is, the effect of the applied tension, were 0.9, 1.1, and 0.5 for primary recrystallization temperatures of 800, 820, and 850°C, respectively. Notably, these changes were equal to or greater than those caused by normal grain growth. However, the changes in {111}<112> orientation due to normal grain growth were −0.8 to −1.6, and the changes due to applied tension were −0.7 to −1.3. Thus, the effects of normal grain growth and applied tension were similar, with some variations. Considering that the effect of tension on the changes in the average grain size was approximately half of that of normal grain growth, the changes in {411}<148> and {111}<112> orientations due to tension annealing were greater than the changes in the average grain size.
| Samples | {411}<148> orientation | {111}<112> orientation | ||
|---|---|---|---|---|
| Change in intensity by 850°C annealing without tension | Difference in intensity between without/with tension | Change in intensity by 850°C annealing without tension | Difference of intensity between without/with tension | |
| Recrystallization annealed at 800°C | 1.1 | 0.9 | −1.6 | −0.7 |
| Recrystallization annealed at 820°C | 0.4 | 1.1 | −0.8 | −0.8 |
| Recrystallization annealed at 850°C | 0.3 | 0.5 | −0.8 | −1.3 |
The mechanism by which the intensity of {411}<148> orientation increases and that of {111}<112> orientation decreases during tension annealing is discussed. Two hypotheses are considered: (1) Size advantage—grain growth driven by grain boundary energy, where larger {411}<148> grains consume smaller {111}<112> grains; (2) Strain-induced grain boundary migration—during tension annealing, {411}<148> grains with lower strain accumulation consume {111}<112> grains with higher strain accumulation, resulting in orientation-selective grain growth. These hypotheses are discussed in detail below.
4.1. Development of {411}<148> Orientation by the Size AdvantageThe frequency and average grain sizes of {411}<148> and {111}<112> grains after primary recrystallization and tension annealing were calculated from the EBSD measurements. The extraction conditions were set at a tolerance angle of 15° from the ideal orientation, and regions with five or more measurement points (equivalent to a diameter of 4.7 μm) were considered grains. The calculated average grain sizes are presented in Fig. 10. Figures 10(a) and 10(b) show the average grain sizes of {411}<148> and {111}<112> grains, respectively, while Fig. 10(c) presents the ratio of the average grain size of each grain orientation to that of all grains. Both {411}<148> and {111}<112> grains exhibited growth during tension annealing. Similar to the trend observed for the overall average grain size, their sizes increased with increasing tension. To evaluate whether {411}<148> grains could grow due to size advantage, the ratio of each orientation’s average grain size to the average of all the grains (Fig. 10(c)) was examined. Although some variation was observed, {411}<148> grains were equal to or larger than the average grain sizes. Hillert18) reported an equation for the driving force of grain growth based on size advantage, as shown in Eq. (1):
| (1) |
where P is the driving force for grain growth, α is the shape factor, E is the grain boundary energy, RC is the critical grain size of the matrix grains, and R is the grain size of the individual grain. According to Eq. (1), when the size of a grain exceeds the average matrix grain size, it experiences a possible grain growth due to its size advantage. However, while {111}<112> grains grew during tension annealing, as seen in the experimental results in Fig. 10, their average grain size remained smaller than that of all grains, indicating they are more likely to be consumed during normal grain growth. Yasuda et al.17) investigated the grain growth behavior of {411}<148> and {111}<112> grains and found that {411}<148> grains were larger after recrystallization. Their study revealed that {411}<148> grains tend to nucleate earlier in regions of high stored energy and dislocation density. Moreover, they provide a significant driving force for grain growth during the later stages of recrystallization, resulting in relatively large grains upon their completion. Consequently, it was presumed that the coarse {411}<148> grains would grow larger owing to their size advantage, even during the subsequent normal grain growth stage after recrystallization. Although the recrystallization process was not investigated in this study, the results align with the mechanism proposed by Yasuda et al.:17) the average grain size of {411}<148> grains was larger than that of all grains after primary recrystallization annealing, while {111}<112> grains were smaller. Therefore, the growth of {411}<148> grains was inferred to be driven by their size advantage. However, the following section examines whether tension annealing introduces an additional driving force for grain growth other than grain boundary energy, enabling {411}<148> grains to more readily consume {111}<112> grains that would be expected from size advantage alone.
In the previous section, the selective grain growth of {411}<148> grains was discussed in terms of normal grain growth driven by grain boundary energy. However, as shown in Fig. 10(a), the growth of {411}<148> grains progressed with increasing tension during tension annealing, which cannot be explained solely by their size advantage. Murakami et al.25,26) investigated the mechanism of preferential growth of Goss grains after skin-pass rolling and subsequent annealing of primary recrystallization annealed sheets. They proposed that {111}<112> grains accumulate more strain during skin-pass rolling, whereas Goss grains accumulate less, resulting in a low dislocation density. Consequently, Goss grains recover more easily during the annealing after skin-pass rolling, leading to their preferential growth by consuming of the surrounding high-strain regions. To investigate whether a similar phenomenon occurs during dynamic tension annealing, the average misorientation of each grain (the average kernel average misorientation (KAM) value of each grain) was evaluated, as shown in Fig. 11. Here, KAM represents the average orientation difference between adjacent measurement points; a larger value indicates a higher density of geometrically necessary dislocations.27) Figures 11(a) and 11(b) show the average misorientations of {411}<148> and {111}<112> grains, respectively. Although some variation was observed, the average misorientation of {111}<112> grains was higher than that of {411}<148> grains. This suggests that when these two oriented grains are in contact, {411}<148> grains—having relatively lower strain (low dislocation density)—preferentially consume {111}<112> grains with higher strain (high dislocation density), facilitating their growth. Figure 11(c) shows the ratio of the average misorientation of {411}<148> or {111}<112> grains to that of all grains. Despite some variation, {411}<148> grains generally exhibited lower strain compared to all grains, while {111}<112> grains exhibited higher strain. Hillert’s equation, with an added strain-effect term, is given in Eq. (2). Here, i denotes the grain of interest, Δγi is the strain difference between the grain of interest and the matrix, and β is a coefficient. For example, if the strain in the grain of interest is larger than that in the matrix, Δγi is positive, thereby reducing the driving force for grain growth.
| (2) |
Onuki et al.21) analyzed the grain growth behavior during high-temperature compression deformation and reported that the area fraction of {100} grains increased, while that of {111} grains decreased. In this study, the intensity of {411}<148> orientation increased, and that of {111}<112> orientation decreased after tension annealing. Although the deformation modes differ, a trend similar to that observed by Onuki et al. was confirmed. Their study did not include a description of the amount of strain in {100} and {111} grains; however, they discussed the strain behavior using the Taylor factor. As in this study, {100} grains were presumed to experience a lower strain, whereas {111} grains experienced higher strain. Onuki et al. assumed that the migration of high-angle grain boundaries between {100} and {111} grains was influenced by low-angle grain boundaries (subgrain boundaries in the recovered structure) within each grain.21) In this study, detailed analysis was not possible owing to the large EBSD measurement interval of 2 μm. Nevertheless, a macroscopic grain boundary distribution was observed in the IPF map shown in Fig. 3. In Fig. 3(c), low-angle grain boundaries of 2°–15° are seen in both the near-{411} and near-{111} grains. However, in this study, both grains often exhibited grain boundaries, rather than retaining subgrain boundaries within the grains. The strain rate during tension annealing in this study was 5×10−4 s−1 under a maximum deformation condition of 24.5 MPa tension, which is equivalent to the conditions used by Onuki et al. However, the amount of strain was 0.06, which is approximately an order of magnitude smaller than the compression strain (e.g., −0.50) reported by Onuki et al. Therefore, it can be inferred that differences in strain led to differences in local texture and recovery structure.
4.3. Analysis of Strain-Accumulation Behavior during Room-Temperature Tensile DeformationAs the introduction and simultaneous recovery of strain during high-temperature tension annealing are complex processes, a room-temperature tensile test was conducted to analyze the strain-accumulation behavior during deformation in detail. A sample that had undergone primary recrystallization annealing at 850°C was deformed by 6% at a tensile speed of 1.8 mm/min at room temperature, and its texture was analyzed using EBSD. The IPF maps are shown in Fig. 12, and the ODFs (ϕ2 = 45° section) are presented in Fig. 13. In the room-temperature tensile test, strain was applied at a rate of 4×10−4 s−1, which is equivalent to tension annealing at high temperatures. This suggests that the changes in crystal orientation were minor. However, as shown in Table 3, the average misorientation within each oriented grain was smaller for {411}<148> grains than for all grains, and larger for {111}<112> grains. This corresponds to the discussion in the previous section that {411}<148> grains have a relatively low dislocation density compared to {111}<112> grains, supporting the hypothesis of strain-induced grain boundary migration.
| Average Misorientation | |||
|---|---|---|---|
| Samples | All | {411} <148> | {111} <112> |
| Recrystallization annealed at 850°C | 0.27 | 0.26 | 0.29 |
| Stretching by 6% at room temperature after recrystallization annealed at 850°C | 0.70 | 0.66 | 0.81 |
Figure 14 shows the crystal-orientation map and KAM map of the sample that was primary recrystallization annealed at 850°C and then subjected to a room-temperature tensile test. The KAM value of {111}<112> grains was higher than that of {411}<148> grains. Generally, the amount of strain accumulated during processing varies with crystal orientation,21,22,23,25,26,28) and is related to the Taylor factor.21,22,23,25,26) The results of this study suggest that {111}<112> grains have a crystal orientation that tends to accumulate strain more readily than {411}<148> grains during tensile deformation. Additionally, for both grain types, smaller grains exhibited higher KAM values. The distribution of KAM values within grains was nonuniform, with a tendency for higher KAM values near the grain boundaries. This is presumably because dislocations accumulate more readily near grain boundary regions than in the grain interiors, at a deformation level of approximately 6%. The average misorientation value used in the previous discussions corresponds to the average KAM value for each grain. Since smaller grains have a larger proportion of grain boundary region, they are more strongly influenced by the grain boundary. Thus, {111}<112> grains likely experienced greater strain accumulation due to their relatively small grain size, and a similar phenomenon may have occurred at high temperatures. The elongation measured after tension annealing at 850°C showed that the specimen at this temperature before tension annealing had lower elongation compared to specimens annealed at lower temperatures. Although Fig. 14 shows the results of the room-temperature tensile test, it is presumed that a relatively large strain accumulated near the grain boundaries during high-temperature tensile deformation, leading to hardening, and the progression of creep deformation via grain boundary sliding. Generally, larger grain sizes suppress high-temperature creep deformation. Therefore, it is assumed that the larger the grain size before tension annealing, the more creep deformation by grain boundary sliding is suppressed, resulting in reduced elongation at 850°C. Separating the effects of grain size and crystal orientation on the amount of strain accumulation during tensile deformation will be addressed in future work.
Based on the above discussion, the mechanism by which the intensity of {411}<148> orientation increases and that of {111}<112> orientation decreases with tension annealing is considered as follows: The average size of {411}<148> grains was equal to or larger than that of all grains, whereas {111}<112> grains—though they grew during tension annealing—were smaller than that of all grains and were more readily consumed during normal grain growth. Thus, {411}<148> grains were preferentially promoted owing to size advantage. Moreover, since the average misorientation of {111}<112> grains was higher than that of {411}<148> grains after tension annealing, {411}<148> grains preferentially consumed the higher-strain {111}<112> grains, further promoting their own growth.
This study investigated the effect of applied tension during annealing on texture development in Fe-3%Si alloy. Grain size increased after tension annealing, accompanied by an increase in the intensity of {411}<148> orientation and a decrease in the intensity of {111}<112> orientation. Moreover, these changes became more pronounced with increasing tension. A detailed analysis of the texture development revealed that the average grain size of {411}<148> grains was larger than that of all the grains, whereas {111}<112> grains were smaller.
The mechanism behind the enhanced intensity of {411}<148> orientation and the reduced intensity of {111}<112> orientation under tension annealing was examined. The average size of {411}<148> grains was equal to or larger than that of all grains, while the average grain size of {111}<112> grains, which grew during tension annealing, was smaller than that of all grains. Therefore, the preferential growth of {411}<148> grains was promoted owing to their size advantage, which was driven by grain boundary energy, and {111}<112> grains were consumed. Additionally, the average misorientation (dislocation density) of {111}<112> grains was higher than that of {411}<148> grains after tension annealing. This supports the strain-induced grain boundary migration mechanism, wherein {411}<148> grains preferentially consume the higher-strain {111}<112> grains. Together, these mechanisms facilitated the preferential growth of {411}<148> grains.
