Journal of Materials Science & Technology  2020 , 43 (0): 104-118 https://doi.org/10.1016/j.jmst.2020.01.018

Research Article

Unveiling annealing texture formation and static recrystallization kinetics of hot-rolled Mg-Al-Zn-Mn-Ca alloy

Qinghang Wanga, Bin Jiangab*, Aitao Tanga, Jie Fua, Zhongtao Jiangc, Haoran Shengd, Dingfei Zhanga, Guangsheng Huangab, Fusheng Panab

a State Key Laboratory of Mechanical Transmissions, College of Materials Science and Engineering, Chongqing University, Chongqing 400044, China
b Chongqing Academy of Science and Technology, Chongqing 401123, China
c Research Institute for New Materials Technology, Chongqing University of Arts and Sciences, Chongqing 402160, China
d Shanghai Aerospace Equipment Manufactory, Shanghai 200245, China

Corresponding authors:   ∗Corresponding author at: State Key Laboratory of Mechanical Transmissions, College of Materials Science and Engineering, Chongqing University, Chongqing 400044, China. E-mail address: jiangbinrong@cqu.edu.cn (B. Jiang).

Received: 2019-07-17

Revised:  2019-09-4

Accepted:  2019-09-24

Online:  2020-04-15

Copyright:  2020 Editorial board of Journal of Materials Science & Technology Copyright reserved, Editorial board of Journal of Materials Science & Technology

More

Abstract

The development of Mg-Al-Zn-Mn-Ca series alloys provides a potential prospect to achieve high strength and formability at room temperature (RT). The formation of elliptical annular texture is treated as a crucial factor for the enhanced RT formability. However, the origin of such an elliptical annular texture formation has been rarely reported. Herein, we unveiled the formation and evolution of elliptical annular texture in the hot-rolled Mg-1.6Al-0.8Zn-0.4Mn-0.5Ca (AZMX1100, wt.%) alloy after annealing at different temperatures for 1 h, and its static recrystallization (SRX) kinetics in given annealing temperature for different time. The results revealed that the formation of elliptical annular texture in the hot-rolled AZMX1100 alloy after annealing was derived from nucleation-oriented SRX mechanism, which took place in 200-300 °C, induced by cracked chain-shaped Al2Ca phases, contraction twins, intersections of double twins, intersections of double twins and grain boundaries and non-basal slips. On further annealing from 300-450 °C, the grains with 45°-70° transverse direction (TD) preferentially grew, which made elliptical annular texture extended along the TD. Based on the Johnson-Mehl-Avrami-Kolmogorov (JMAK) model, Avrami exponent n value was estimated to be 0.68-1.02, attributed to non-random SRX nucleation, giving rise to the lower activation energy QR of nucleation of ∼74.24 kJ/mol. Since the co-segregation of Al, Zn and Ca atoms in grain boundaries created a strong interaction of solutes and grain boundaries, the hot-rolled AZMX1100 alloy exhibited the higher activation energy Qg (∼115.48 kJ/mol) of grain growth.

Keywords: Mg-Al-Zn-Mn-Ca alloy ; Elliptical annular texture ; Nucleation ; Grain growth ; Static recrystallization kinetics

0

PDF (7546KB) Metadata Metrics Related articles

Cite this article Export EndNote Ris Bibtex

Qinghang Wang, Bin Jiang, Aitao Tang, Jie Fu, Zhongtao Jiang, Haoran Sheng, Dingfei Zhang, Guangsheng Huang, Fusheng Pan. Unveiling annealing texture formation and static recrystallization kinetics of hot-rolled Mg-Al-Zn-Mn-Ca alloy[J]. Journal of Materials Science & Technology, 2020, 43(0): 104-118 https://doi.org/10.1016/j.jmst.2020.01.018

1. Introduction

One of the lightest structural materials, magnesium (Mg) alloys, have attracted considerable attention for potential applications in the automotive industries [1]. Comparable to aluminum (Al) alloys and steels, however, commercial wrought Mg alloys such as AZ or ZK series alloys are significantly difficult to achieve a perfect matching with high strength and excellent formability at room temperature (RT). It is considered as the “strength-formability trade-off dilemma” [[1], [2], [3]], which to large extent impedes the extensive application of Mg alloys.

Formability of Mg alloys is strongly susceptible to texture characteristic, and modifying and weakening basal texture is in favor of the enhanced formability at RT [[1], [2], [3], [4], [5], [6]]. The addition of RE (Ce [7], Y [8] and Gd [9]) elements into Mg-Zn alloy extremely improved the Index Erichsen (I.E.) value to ∼9.0 mm, an indicator of the stretch formability at RT. Nevertheless, Mg-Zn-RE alloys are still not satisfied with the demands of industrial applications because of their low strength. For instance, Mg-1.5Zn-0.2Y (wt.%) [8], Mg-1.0Zn-2.0Gd (wt.%) and Mg-1.0Zn-3.0Gd (wt.%) [9] alloy sheets demonstrated the low average tensile yield strength of below 130 MPa. And even Mg-1.5Zn-0.2Ce (wt.%) alloy sheet was only ∼100 MPa [7]. To facilitate the use of Mg alloys in the automotive industry, it is essential to exploit the RT high-strength and formable wrought Mg alloys.

Recently, the development of Mg-Al-Zn-Mn-Ca series alloys provides a potential prospect to achieve high RT strength and formability. Mg-1.2Al-0.8Zn-0.4Mn-0.5Ca (AZMX1100, wt.%) alloy exhibited a large I.E. value of ∼7.7 mm in a solution treated (T4) condition, and the tensile yield strength remarkably increased to ∼204 MPa after aging at 200 °C for 1 h (T6) [10]. By bake treatment (T4 at 450 °C for 1 h + tension of 2 % + aging at 170 °C for 20 min), Mg-1.3Al-0.8Zn-0.7Mn-0.5Ca (AZMX1110, wt.%) alloy presented a higher tensile strength of ∼238 MPa, accompanied with a high I.E. value of ∼7.8 mm in T4 condition [3]. Such high strength was mainly attributed to the co-segregation of Al, Zn and Ca atoms to basal <a> dislocations hindering dislocation motions, along with the co-clustering of these atoms [3]. Moreover, a weak and elliptical annular texture characteristic contributed to the excellent RT formability of AZMX1100 [10] and AZMX1110 [3] alloys. It is important to reveal the origin of the formation of this elliptical annular texture for effective texture modification. However, detailed mechanisms still remain unknown.

For the formation of non-basal texture, researchers believe that it is mainly related to the recrystallization process. Two classical theories known as nucleation-oriented and growth-oriented recrystallization mechanisms are used to explain the origin of recrystallized texture. Several nucleation mechanisms have been proposed to explain the formation of non-basal texture: for instance, particle stimulated nucleation (PSN) [2,11], activation of non-basal slip systems [2,12,13], existence of compression and double twinning [2,14,15], shear band nucleation (SBN) [12,16]. In general, the growth-oriented recrystallization mechanism is also deemed as the selective growth of grains, which is closely related to the difference of storage energy [17,18] and the co-segregation of solute atoms [19,20]. However, especially for the formation of elliptical annular texture, the detailed microstructural and textural evolution during annealing remains unclear and the proposed mechanisms are still debatable.

In this work, the as-cast Mg-1.6Al-0.8Zn-0.4Mn-0.5Ca (AZMX1100, wt.%) alloy was multi-pass hot-rolled and subsequently annealed at different temperatures. Analysis of electron backscattered diffraction (EBSD) was carried out to reveal the origin of elliptical annular texture formation during hot-rolling and subsequent static recrystallization (SRX). In addition, annealing at given various temperatures for different time was performed to investigate SRX (grain nucleation and growth) kinetics of hot-rolled AZMX1100 alloy.

2. Experimental procedures

The as-cast Mg-1.6Al-0.8Zn-0.4Mn-0.5Ca (AZMX1100, wt.%) alloy (Ф 80 mm × 200 mm) was prepared in an electric resistance furnace under the atmosphere of SF6 and CO2 mixed gas (mixing volume ratio was 1:99). The materials used in this study were commercial pure Mg ingot (≥ 99.99 wt.%), pure Al ingot (≥ 99.8 wt.%), pure Zn granules (≥ 99.8 wt.%), Mg-10 wt.% Mn and Mg-25 wt.% Ca master alloy. The melt was held at 720 °C for 20 min after alloying elements were dissolved into Mg matrix, and then it was poured into a steel mold of 200 mm × 50 mm × 100 mm in length, width and height, which was preheated to 350 °C. The chemical composition of as-cast AZMX1100 alloy was measured by inductively coupled plasma-atomic emission spectroscopy (ICP-AES) and the result is shown in Table 1.

Table 1   Detailed chemical composition of as-cast AZMX1100 alloy (wt.%).

AlloyMgAlZnMnCa
AZMX1100Bal.1.640.840.380.49

New window

As-cast AZMX1100 alloy strips with an initial thickness of 8.0 mm machined by wire-electrode cutting were subjected to homogenize at 450 °C for 12 h, and then were hot-rolled at 400 °C. The AZMX1100 alloy strips were rolled down from 8.0 mm to a final thickness of 2.7 mm by 5 passes with a reduction per pass of 15 %. The samples were reheated at 400 °C for 10 min between each pass to keep the rolling temperature constant. It was observed that no any crack occurred on the surfaces and edges of hot-rolled AZMX1100 alloy strips. In order to demarcate the texture characteristic of hot-rolled AZMX1100 alloy after annealing at different temperatures, the samples with sizes of 6.0 mm × 5.0 mm × 2.7 mm were machined to carry out annealing at 200, 250, 300, 350, 400 and 450 °C for 1 h. For investigating the SRX kinetics of hot-rolled AZMX1100 alloy, the samples with sizes of 6.0 mm × 5.0 mm × 2.7 mm were machined to perform annealing at given temperatures for different time.

Microstructural observations of hot-rolled and annealed samples were analyzed using an Optical Microscopy (OM, ZEISS Axiovert 40 MAT) and an electron backscattered diffraction (EBSD, JEOL JSM-7800 F) technique. EBSD specimen preparation consisted of grinding on SiC papers of grit sizes 280, 400, 600, 800, 1000, 1200 and 2000, washing, blow-drying as well as electro-polishing at a voltage of 20 V and an electric current of 0.03 A for 120 s at a temperature of -20 °C with a special electrolyte named as AC2. A step size of EBSD scan was set as 0.5 μm. All EBSD data were analyzed using the Channel 5 software. Average grain sizes of samples were determined by the measurement of diameter using Image-pro plus software. The observation of second phases in the hot-rolled and annealed samples was carried out by a scanning electron microscopy (SEM, TESCAN VEGA 3 LMH SEM) equipped with energy dispersive spectrometer (EDS) and an X-ray Diffraction device (XRD) using Cu radiation at a wavelength of 0.15406 nm by Rigaku D/Max 2500. The volume fraction of second phases was measured by Image-pro plus software. Vickers hardness testing was conducted on the aging treated specimens under a load of 50 g and the holding time of 60 s. Transmission electron microscopy (TEM, FEI TECNAI G2 F20) was used to characterize the crystal defects and second phases. Thin foils for TEM were prepared by mechanical polishing (∼40 μm) and then ion beam thinning (GATAN, PIPS 691). TEM observation was then conducted with an accelerating voltage of 200 kV.

3. Results and discussion

3.1. Microstructures and texture evolution of hot-rolled and annealed samples

Fig. 1 reveals the OM images taken in the longitudinal section (RD-ND plane, RD and ND represent rolling direction and normal direction, respectively) and the (0002) pole figures obtained by X-ray texture analysis of samples in the hot-rolled condition and fully annealed one at 450 °C for 1 h. It can be seen that, in the hot-rolled condition, the microstructure consists of deformed grains containing massive twins and a few black chain-shaped bands marked by red arrows (these bands will be described in SEM images of Fig. 6) without the apparent occurrence of dynamic recrystallization. Moreover, shear bands are not obvious. This may be attributed to the multi-pass high temperature rolling at 400 °C and the intermediate annealing, which is in favor of weakening strain localization in form of shear bands [21]. On the X-ray texture, it exhibits a double-peak texture characteristic with maximum pole intensity of 6.186 tilted ± 15° from the ND towards the RD and slightly broadens along the transverse direction (TD). Such a texture feature has been reported in rolled Mg-Zn-RE alloys [12,21,22] and extruded Mg-RE alloys [23]. Numerous studies have revealed that the formation of such a double-peak texture was closely related to the activation of pyramidal <c+a> slip and the occurrence of double twins [12,24]. From the observation of TEM of hot-rolled sample (see Fig. 2), based on the gb = 0 (b, Burger vector) criterion [25], pyramidal <c+a> dislocations (marked by red arrows) are simultaneously visible under two-beam diffractions with g = 0002 and g = 01 $\bar{1}$ 0. Obviously, the high rolling temperature of 400 °C provides a high energy enough to take over the critical resolved shear stress (CRSS) of pyramidal <c+a> slip activity. Of course, basal <a> dislocations (see Fig. S1 in Supplementary Material) and prismatic <a> traces also can be observed in the hot-rolled sample. In addition, it can be observed that there are a large number of “black bands” existed in the hot-rolled sample by EBSD data (see Fig. 3(a)), and the result from line AB (point-to-point) exhibits that the misorientation angles between “black bands” and matrix contain 38°±5° and 56°±5° consistent with the definitions of {10 $\bar{1}$ 1}-{10 $\bar{1}$ 2} double twin and {10 $\bar{1}$ 1} contraction twin (see Fig. 3(b)). Therefore, we can identify that multiple slip mechanisms and twin types are simultaneously activated during multi-pass high rolling, which contributes to the formation of double-peak texture characteristic. These results indicate that RE-free AZMX1100 alloy demonstrates the same microstructural properties as RE-containing Mg alloys in the hot-rolled condition.

Fig. 1.   OM images and (0002) pole figures obtained by X-ray texture analysis of (a) hot-rolled AZMX1100 alloy and (b) fully annealed AZMX1100 alloy at 450 °C for 1 h.

Fig. 2.   Two-beam bright-field TEM images showing the microstructure of hot-rolled sample, where these images are taken under two beam conditions using diffraction vectors of (a) g = 0002, (b) g = 01 $\bar{1}$ 0.

Fig. 3.   (a) EBSD map of hot-rolled sample; (b) line profiles of point-to-point along white arrow AB in (a).

In the fully annealed condition, the microstructure is comprised of the homogeneous complete SRXed grains with an average grain size of ∼ 8.0 μm and a few black chain-shaped bands marked by red arrows (these bands also will be described in SEM images of Fig. 7). Regarding the recrystallization texture characteristic, we can observe that the maximum pole intensity of 4.342 focuses on the location tilted ±60° from the ND towards the TD. Moreover, the texture contour lines expand to the RD, forming a weak RD texture component of 1.895 with basal poles tilted from the ND towards the RD. Such a new texture characteristic is named by the elliptical annular texture. Except for the AZMX1100 and AZMX1110 alloys reported by Bian et al. [3,10], the hot-rolled and annealed Mg-Zn-Ce alloy [21] and the cold-rolled Mg-Zn-Gd alloy [26] also showed the similar elliptical annular texture characteristic. Such a large difference in texture characteristic between the hot-rolled and the annealed conditions is intimately associated with the SRX process. In other words, Grain nucleation and growth of SRX process determine the formation of elliptical annular texture in the annealed sample.

In order to unveil the effect of SRX process on the origin of elliptical annular texture and its evolution, the hot-rolled samples were subjected to annealing at 200, 250, 300, 350, 400 and 450 °C for 1 h. The microstructures and texture evolution of hot-rolled samples suffered from annealing at different temperatures are displayed in Fig. 4. Regarding the microstructure evolution, there are extremely few locations, where SRX behavior occurs at low temperature of 200 °C. The volume fraction of SRXed grains accounts for ∼3.4 % (detailed statistic method of SRX is shown in Fig. S2 in Supplementary Material). Moreover, mass locations are still not demarcated by EBSD, since the amount of internal stress especially located insides contraction twins and double twins have not been removed yet at low temperatures. With increasing the annealing temperature to 450 °C, the volume fraction of SRXed grains in the annealed samples gradually enhance to approach complete SRX (for ∼91.7 %). Moreover, unidentified areas by EBSD also sharply decrease, which suggests that these areas become the nucleated sits of SRXed grains facilitating the SRX and resulting in the release of mass internal stress. On the average grain size, it also gradually raises from ∼1.9 μm to ∼7.9 μm with annealing temperature. As the mentioned above, the SRX characteristic of hot-rolled AZMX1100 alloy can be considered as the co-existence of grain nucleation and growth until the alloy happens to the complete SRX. In terms of texture evolution, the maximum pole intensity of annealed sample lies in the location tilted ∼20° from the ND towards the RD at low temperature of 200 °C, and a new weak texture component form in the place tilted ∼25° from the ND towards the TD. With the increase of annealing temperature, the double-peak texture of hot-rolled sample along the RD gradually transforms into the elliptical annular texture, accompanied with the weakening of RD texture component and the enhancement of TD one. Moreover, the maximum pole intensity has a gradually downward tendency with the development of SRX process.

Fig. 4.   Microstructures, grain size distributions of SRXed grains and (0002) pole figures of hot-rolled samples subjected to annealing at different temperatures: (a-f) 200, 250, 300, 350, 400 and 450 °C for 1 h, respectively.

For better describing the formation of elliptical annular texture and its evolution with annealing temperature, the elliptical annular texture is simplified as the one consisted of RD and TD texture components. Fig. 5 shows the relationship among annealing temperature, texture intensity and angle of RD and TD texture components, and the schematic diagram of elliptical annular texture evolution. As shown in Fig. 5(a), the texture intensity of RD component has a rapid descending trend from 200 °C to 300 °C, and then reaches a platform. Nevertheless, the texture intensity of TD component exhibits a distinct fluctuation. In this temperature range from 200 °C to 300 °C, the grain nucleation dominates the whole SRX process, since the volume fraction of SRXed grains sharply increases from ∼ 3.4 % to ∼ 86.6 %. It suggests that the orientations of SRXed grains are primary TD texture components. Over 300 °C, it remarkably surpasses that of RD component. Note that it gradually decreases from 300 °C to 400 °C, and then rises from 400 °C to 450 °C. The texture intensities of RD and TD components are close and lowest at 400 °C, which indicates that the hot-rolled sample demonstrates a uniform and weak elliptical annular texture after annealing at 400 °C for 1 h. Over 400 °C, we can assume that the apparent grain growth gives rise to the enhancement in texture intensity of TD component, thereby making the annealed sample show the remarkable preferred TD texture component. Regarding the tilted angles of RD and TD texture components, it can be found that with raising the annealing temperature, both the tilted angles of RD and TD texture components gradually increase, as shown in Fig. 5(b). Note that the increasing level of tilted angle for TD texture component is apparently higher than that for RD texture component. Such a tendency may be closely related to grain growth. In recent study, Zhao et al. [27] also found a similar TD texture component shift phenomenon in cold-rolled Mg-2Zn-1Gd (wt.%) alloy during SRX process. They attributed this result to the preferred growth of grains with TD texture component [27]. Therefore, in this work, the TD texture component shift may result from the preferred growth of grains with TD orientation. Based on the analysis of texture intensities and tilted angles of RD and TD components with annealing temperature, the elliptical annular texture evolution is divided into two stages (see Fig. 5(c)). The annealing at 200-300 °C is marked by the first stage. In this stage, the nucleation of SRXed grains with TD orientation dominates the whole SRX process, along with the weakening of RD texture component and the enhancement of TD one. The annealing at 300-450 °C is labeled by the second stage. In this stage, the preferred growth of grains with TD orientation is main SRX behavior, which gives rise to TD texture component away from the ND.

Fig. 5.   (a) Texture intensity values of RD and TD texture components as a function of annealing temperature; (b) Tilted angles of RD and TD texture components as a function of annealing temperature; (c) Schematic diagram of elliptical annular texture evolution.

Fig. 6.   SRX nucleation of hot-rolled sample during annealing at 200 °C: (a) image quality including multiple twin types; (b) SEM image, where chain-shaped phase is marked by yellow dotted lines; (c) extracted SRXed grains from EBSD map in Fig. 3(a); (d) (0002) pole figure of extracted SRXed grains corresponding to (c).

Fig. 7.   (a, b) SEM images of as-cast sample via homogenization at 450 °C for 12 h; (c) local magnification view and EDS results corresponding to (b); (d, e) SEM images of hot-rolled sample; (f) local magnification view and EDS results corresponding to (e).

3.2. Nucleation mechanisms and preferred growth of SRX process

3.2.1. Nucleation-oriented SRX texture

As we analyze above, the elliptical annular texture characteristic has begun to take shape at low temperature annealing of 200-300 °C, along with the nucleation of grains with TD orientation. It indicates that the SRX texture of hot-rolled AZMX1100 alloy after annealing becomes the nucleation-oriented texture type. In general, SRX nucleation sites may locate at the around of second phases [11,28], insides twins (especially contraction twins [29,30] and double twins [14, 15), at twins and grain boundaries [13], at the intersections of twins and grain boundaries [27], and insides deformed grain [13,31]. In order to unveil the detailed nucleation mechanisms during annealing at 200-300 °C, we make use of EBSD technique to inspect the nucleation sites of SRXed grains. Fig. 6 shows the SRX behavior of hot-rolled sample during annealing at 200 °C. As we see, there still keep massive twins, especially contraction twins (blue lines) and double twins (green lines) after annealing (see Fig. 6(a)). These contraction twins and double twins tend to be unidentified areas due to high storage energy [15]. Insides these twins or at twin boundaries, we can hardly find any SRXed grains. It infers that twins do not become effective nucleation sites at low temperature annealing. Under the careful observation combined with SEM image (Fig. 6(b)), a few fine SRXed grains extracted from Fig. 6(a) distribute along black chain-shaped bands marked by yellow dotted lines in SEM image, as shown in Fig. 6(c). In Fig. 6(d), these SRXed grains are extracted to measure their orientation distribution in the (0002) pole figure. We are amazed to find that most SRXed grains mainly distribute at the locations tilted from the ND to the TD, and a small part situate at the site tilted from the ND to the RD. Such a texture characteristic is approach to an elliptical annular texture. Therefore, we conclude that the main nucleation mechanism is the chain-shaped band-induced SRX to promote the formation of elliptical annular texture at low temperature annealing of 200 °C.

Herein, there is a crucial question: what are chain-shaped bands? In order to answer this question, we carry out SEM observations, as shown in Fig. 7, Fig. 8. Fig. 7 displays the SEM images and EDS results of homogenized and hot-rolled samples. It can be seen that vast chain-shaped phases and a few granule-shaped phases randomly distributes on the whole matrix of homogenized sample (see Fig. 7(a)). Based on the EDS results in Fig. 7(c), Chain-shaped and granule-shaped phases may be Al-Ca (point A) and Al-Mn (point B) phases, respectively. According to the XRD result and phase diagram of AZMX alloy (not shown), chain-shaped and granule-shaped phases are identified by Al2Ca and Al8Mn5 phases, respectively. After multi-pass hot-rolling, there still exists massive chain-shaped Al2Ca and a small number of granule-shaped Al8Mn5 phases. However, these chain-shaped Al2Ca phases happen to broken along the RD, as seen in Fig. 7(f). Fig. 8 shows the SEM images of hot-rolled samples after annealing at different temperatures. At 200-450 °C, Al2Ca and Al8Mn5 phases can hardly dissolve into the matrix since the melting points of these phases take over 450 °C. Therefore, the volume fraction of second phases in all the annealed samples is almost identical (∼28 %). In addition, it is fairly apparent that these cracked chain-shaped Al2Ca phases induce the nucleation of SRXed grains, especially at 200-250 °C. We can clearly observe that SRXed grains prior to nucleate at the around of these cracked chain-shaped Al2Ca phases (marked by yellow dotted line area in Fig. 8(b)) at 200 °C. Therefore, the main SRX nucleation mechanism is the cracked chain-shaped Al2Ca phase-induced one to facilitate the formation of elliptical annular texture at low temperature annealing of 200 °C, combined with the analysis of Fig. 6. With increasing the annealing temperature, the nucleation sites of SRXed grains increase. Unrecognized high storage energy areas may become effective nucleation sites. At 250 °C, besides the cracked chain-shaped Al2Ca phases, there are also many SRXed grains insides twins (green dotted line areas) and deformed grains (red dotted line areas), as shown in Fig. 8(d). It suggests that multiple nucleation mechanisms jointly contribute to the SRX behavior at 250 °C.

Fig. 8.   (a) SEM image of hot-rolled sample after annealing at 200 °C; (b) local magnification view corresponding to (a). Yellow dotted line area is SRX region at the around of cracked chain-shaped Al2Ca phases; (c) SEM image of hot-rolled sample after annealing at 250 °C; (d) local magnification view corresponding to (c). Yellow dotted line area is SRX region at the around of cracked chain-shaped Al2Ca phases. Green dotted line area is SRX region insides twins. Red area is SRX region insides deformed grain; (e-h) SEM images of hot-rolled samples after annealing at 300, 350, 400 and 450 °C, respectively.

In order to further understand the detailed nucleation mechanisms at 250 °C, the EBSD analysis of different areas from Fig. 4(b) is shown in Fig. 9, Fig. 10, Fig. 11. At 250 °C, massive cracked chain-shaped Al2Ca phases still can act the effective SRX nucleation sites to form a large number of SRXed grains. This nucleation mechanism induced by cracked chain-shaped Al2Ca phases is identical with that at 200 °C. Herein, we will do not elaborate in detail. Fig. 9 describes the SRX nucleation induced by contraction twins in the annealed sample. SRXed grains and deformed grains can be distinguished by kernel average misorientation (KAM) map. KAM within a grain is constructed by calculating the average misorientation between every pixel and its surrounding pixels. Blue color regions represent low-strain, even free-strain ones, and green color areas are regarded as high-strain ones in the KAM map. The grains corresponded to blue color regions in the KAM map are considered as SRXed grains (Fig. 9(b)). These SRXed grains derived from twins are labeled by G1-G4 (see Fig. 9(a)). The detailed twin type can be recognized by the relative misorientation angle distribution corresponding to lines A and B from Fig. 9(a). We can find that the misorientations between grains G1 and G2 and matrix are close to 56°±5° (see Fig. 9(c)), which is consistent with the one between contraction twin and matrix. Thus, this nucleation mechanism is affirmed by the contraction twin-induced SRX. For revealing the relationship between this nucleation mechanism and SRX texture, the orientation distribution of SRXed grains G1-G4 is shown in the (0002) pole figure. As seen in Fig. 9(d), Grains G1 and G3 locate at the place tilted from the ND towards the RD, and grains G2 and G4 situate at the locations tilted from the ND towards the TD. Such an orientation distribution is also close to the elliptical annular texture orientation. Therefore, we verify that the SRX nucleation induced by insides contraction twins is also in favor of the formation of elliptical annular texture.

Fig. 9.   SRX nucleation induced compression twins in hot-rolled sample during annealing at 250 °C: (a) EBSD map extracted from Fig. 3(b); (b) KAM map corresponding to (a); (c) line profiles of point-to-point along the white line AB in (a); (d) orientation distributions of SRXed grains G1-G4 and matrix in (0002) pole figure.

Fig. 10.   SRX nucleation induced double twin intersection in hot-rolled sample during annealing at 250 °C: (a) EBSD map extracted from Fig. 3(b); (b) KAM map corresponding to (a); (c) image quality including multiple twin types; (d) orientation distributions of SRXed grains G1-G14, double twins D1-D5 and matrix in (0002) pole figure.

Fig. 11.   SRX nucleation induced intersections of double twins and grain boundaries in hot-rolled sample during annealing at 250 °C: (a) EBSD map extracted from Fig. 3(b); (b) KAM map corresponding to (a); (c) image quality including multiple twin types; (d) orientation distributions of SRXed grains G1- G2, double twin D1 and matrix in (0002) pole figure.

It is strange that no any sign is found to nucleate insides double twins. However, massive previous studies have proved that double twins as significant nucleation sites gave rise to texture weakening in RE-containing Mg alloys [15]. So, what a role do double twins play in the SRX process of hot-rolled AZMX1100 alloy?

Fig. 10 shows the SRX nucleation induced by double twin intersection in the annealed sample. According to the KAM map (Fig. 10(b)), SRXed grains are highlighted and marked by G1-G14 from Fig. 10(a). Based on the image quality map (Fig. 10(c)), there are still many {10 $\bar{1}$ 1}- {10 $\bar{1}$ 2} double twins (a misorientation of ∼38° between double twin and matrix) reserved at 250 °C. These SRXed grains G1-G14 just right nucleate at midland of double twins. We seem to be able to find that G1, G2, G3, G6, G7, G8, G9 grains exhibit a chain-shaped distribution. Similarly, G4, G5, G10, G11, G12, G13, G14 grains also have a chain-shaped distributed characteristic. Therefore, an assumption can be proposed to state that G1, G2, G3, G6, G7, G8, G9 grains may result from the SRX nucleation of intersection of double twins D1, D2, D3 and a vanished twin D6 labeled by yellow dot-line region, and G4, G5, G10, G11, G12, G13, G14 grains may result from the SRX nucleation of intersection of double twins D1, D2, D3, D4, D5 and a vanished twin D7 marked by while dot-line region. Here, it is difficult to find the traces of twins D6 and D7. However, the misorientation values between SRXed grains G1-G14 and matrix (∼38° in Fig. 10(d)) can provide an evidence to confirm that vanished dot-line regions D6 and D7 both are double twins. This SRX nucleation mechanism can be considered as one induced by double twin intersection, which also has been reported in the hot-rolled and annealed Mg-4Zn-1Ce (wt.%) alloy [13]. The orientations of double twin intersection-induced SRXed grains are different from double twins and parent grain in Mg-4Zn-1Ce alloy [13]. In this work, SRXed grains G1-G14, as demonstrated in Fig. 10(d), distribute at the locations tilted from the ND to the RD and TD in the (0002) pole figure, which are also distinguished with the orientations of double twins (D1-D5) and matrix, especially like G1, G2, G10 and G11. Such a scattered orientation distribution is also beneficial for the formation of elliptical annular texture. Besides double twin intersection, the intersection of double twin and grain boundary also acts a significant role on the SRX nucleation. In the hot-rolled and annealed Mg-1.5Zn-0.2Ce (wt.%) alloy, Huang et al. [21] revealed that SRXed grains mainly nucleated at the intersections of double twins and pre-existing grain boundaries. In the present study, we also observe this similar SRX nucleation phenomenon. Fig. 11 uncovers the SRX nucleation induced by intersections of double twins and grain boundaries. As seen in Fig. 11(a), SRXed grains G1 and G2 situate at the intersections of double twin and grain boundaries (SRXed grains and double twin can be identified by KAM map (see Fig. 11(b)) and image quality map (see Fig. 11(c)), respectively). In addition, the orientations of grains G1 and G2, located at the places tilted from the ND to the TD, can be reflected in the (0002) pole figure, as shown in Fig. 11(d). It indicates that the SRX nucleation mechanism induced by intersection of double twin and grain boundary can remarkably promote texture scattering. Therefore, the mentioned role resulted from double twins activating the SRX nucleation and is a crucial origin of elliptical annular texture formation.

Of course, we cannot neglect the role of dislocation slips on the SRX nucleation. It has been revealed that a high activity of non-basal slip (i.e., prismatic <a> slip) was closely relative to the formation of a weak texture with basal pole spread in the TD of rolled-annealed Mg-Zn-Nd and extruded Mg-Sn-Zn-Y sheets [2,12]. During rolling deformation, there exists frequently an obvious orientation gradient inside a deformed grain. The orientation gradient derived from dislocation slips can result in sub-grains and, on further annealing, lead to the formation of high-angle grain boundaries [13]. The activation of more non-basal slips tends to facilitate dislocation climb and/or cross-glide, which is necessary for dislocations to rearrange themselves into cell and sub-grain boundaries [32]. In this work, an amount of non-basal slips are activated in the hot-rolled sample giving rise to the formation of dislocation wall (see Fig. 12), which is beneficial for the formation of sub-grain boundaries (or low-angle grain boundaries) during low temperature annealing of 200 °C, and on further annealing at 250 °C, even 300 °C, these sub-grain boundaries gradually transforms into high-angle grain boundaries. Sanjari et al. [13] pointed out that the orientations of SRXed grains induced by prismatic <a> slip were completely different from parent grains. Therefore, the SRX nucleation mechanism induced by non-basal slips can effectively result in the formation of new grains with wider orientation distribution.

Fig. 12.   Two-beam dark-field TEM images showing microstructure of hot-rolled sample, where these images were taken under two beam conditions using diffraction vectors of (a) g = 0002, (b) g = 01 $\bar{1}$ 0, (c) view of high magnification of red rectangular frame in (b).

3.2.2. Preferred growth of grains with TD texture component

As we observe in Fig. 4, the volume fractions of SRXed grains for the hot-rolled samples after annealing at 300-450 °C are up to beyond 90 %, and with increasing the annealing temperature, their average grain sizes gradually raise. Based on the analysis in Fig. 5, the elliptical annular texture has formed at 300 °C, and on further annealing, this texture characteristic has not an apparent alteration, but its TD texture component shift is found away from the ND, which means that there may be a preferred growth of grains with a high-angle TD texture component. These grains with high-angle TD texture component prior to grow making the intensity of low-angle TD texture component reduce and that of high-angle TD one strengthens. Until 400 °C, the intensity of total TD texture component intensity is lowest. Further annealing to 450 °C, the apparent growth of grains with high-angle TD texture component leads to the enhancement of intensity of total TD texture component.

In order to quantitatively describe the preferred growth of grains with high-angle TD texture component, the elliptical annular texture is divided into four groups with 0°-20°, 20°-45°, 45°-70° and 70°-90° texture components, respectively [2]. These four groups are labeled by TCA, TCB, TCC and TCD texture components, respectively. Fig. 13 shows the EBSD maps and (0002) pole figures of these four groups in the hot-rolled samples after annealing at 400 °C and 450 °C, and corresponding to their volume fractions and average grain sizes. The detailed data are listed in Table 2. It can be seen that the volume fractions of TCB and TCC texture components are remarkably higher than these of others at 400 °C and 450 °C. However, with raising the annealing temperature from 400 °C to 450 °C, TCA and TCD texture components almost keep stable, but TCB and TCC texture components have an apparent fluctuation. The volume fraction of TCB texture component decreases by 14.2 %, whereas the TCC texture component increases by 14.3 %. On the average grain size, we can also obviously observe that with the increase of annealing temperature, the increasing level of the average grain size for TCC texture component is highest (∼49.1 %) among four texture components. Therefore, it is concluded that the preferred growth of grains with 45°-70° TD texture component results in, with increasing the annealing temperature, TD texture component shift in the elliptical annular texture.

Fig. 13.   EBSD maps and (0002) pole figures of four groups of grain orientation components 0°-20° (TCA), 20°-45° (TCB), 45°-70° (TCC) and 70°-90° (TCD) tilted away from the ND in the hot-rolled sample after annealing at 400 °C (a-d) and hot-rolled sample after annealing 450 °C (e-h); (i, j) volume fraction and average grain size of each group of grain orientation component in hot-rolled samples after annealing at 400 °C and 450 °C.

Table 2   Summary of volume fractions and average grain sizes of TCA, TCB, TCC and TCD texture components in hot-rolled samples after annealing at 400 °C and 450 °C.

Texture component400 °C450 °C
Volume fraction (%)Ave. grain size (μm)Volume fraction (%)Ave. grain size (μm)
TCA8.54.87.86.8
TCB52.25.138.07.1
TCC30.15.344.37.9
TCD9.85.39.97.0

New window

Only realizing that the elliptical annular texture is closely relative to the preferred growth of grains with high-angle TD texture component is insufficient. The reason for preferred growth of grains with non-basal orientations is necessary to be understood. Based on the previous researches, this preferred growth of grains with non-basal orientations may be mainly attributed to the difference in storage energy [17,18] and the segregation of solute atoms [19,20].

Some researchers found with increasing the annealing temperature, the <2 $\bar{11}$ 1 > RE texture component was strengthened in extruded Mg-1Gd (wt.%) alloy gradually at the expense of the <10 $\bar{1}$ 0> texture component, which could be ascribed to the difference in storage energy between the <2 $\bar{11}$ 1> grains with low KAM values and the <10 $\bar{1}$ 0> grains with high KAM values [18]. However, in our work, the annealed samples happen to the apparent SRX behavior at 300-450 °C. The fairly high SRX level (∼90 %) gives rise to no distinction in storage energy between grains with basal and non-basal orientations, since all grains have low KAM values. Thus, it is reasonable to say that the storage energy has no obvious effect on the preferred growth in our study.

Massive studies have been reported that RE, Al, Zn, and Ca solute atoms could segregate in grain boundaries or twin ones to hinter the growth of grains or twins [3,10,19,20,33]. Basu et al. [19] clearly pointed out that the enhanced solute drags in Mg-1Gd by Gd segregation significantly amplified the growth advantage of RE orientations over basal nuclei during SRX process. For RE-free cold-rolled Mg-0.3Zn-0.1Ca (at.%) alloy, the co-segregation of Zn and Ca atoms effectively impeded the preferred growth of grains with strong basal orientation during SRX process, which resulted in a weak and scattered texture [33]. In previous studies, Bian et al. [3,10] and Trang et al. [34] have proved that Al, Zn and Ca atoms could co-segregate in grain boundaries in AZMX1110 and AZMX3110 alloys, respectively. Therefore, we argue that the growth of grains with basal orientation may be inhibited because of this co-segregation which hinders the migration of special grain boundaries, so that the grains with high-angle TD orientation appear relatively preferred growth.

Based on the reports of Wang et al. [35], the eight possible low-energy grain boundaries in hexagonal close packed (HCP) metals, shown in Table 3, have been found by simulation. In order to measure the special grain boundaries induced by co-segregation of solute atoms, the rotation axis of these 8 special grain boundaries in the hot-rolled samples after annealing at 400 °C and 450 °C are shown in Fig. 14. In this case, the rotation axis distribution of 26°-30°, 30°-34°, 37°-41° and 41°-45° grain boundaries are random for two annealed samples. However, the rotation axis of 56°-60° and 60°-64° grain boundaries happen to the preferred distribution around [1 $\bar{2}$ 10], and the rotation axis of 71°-75° and 73°-77° grain boundaries exhibit the preferred distribution around [1-210] and [01-10]. It indicates that 58° [1 $\bar{2}$ 10], 62° [1 $\bar{2}$ 10], 73° [1 $\bar{2}$ 10], 75° [1 $\bar{2}$ 10], 73° [01 $\bar{1}$ 0], and 75° [01 $\bar{1}$ 0] grain boundaries are reserved. Note that although both two annealed samples have a similar rotation axis distribution, it is remarkable that the intensities of these rotation axis distribution for the hot-rolled sample after annealing at 450 °C are higher than these for the hot-rolled sample after annealing at 400 °C. According to the results in Table 3, 62° [1 $\bar{2}$ 10], 75° [1 $\bar{2}$ 10] and 73° [01 $\bar{1}$ 0] grain boundaries are pre-existing low-energy ones, and 58° [1 $\bar{2}$ 10], 73° [1 $\bar{2}$ 10] and 75° [01 $\bar{1}$ 0] grain boundaries may be new low-energy grain boundaries induced by co-segregation of solute atoms, since the solute segregation in grain boundaries could apparently reduce the energy of grain boundaries [36]. With increasing the annealing temperature from 400 °C to 450 °C, these solute segregation-induced low-energy grain boundaries intensities obviously enhance, which is rather beneficial for the growth of grains with high-angle TD orientation.

Table 3   Summary of eight possible low-energy grain boundaries in HCP metals and corresponding angle range [35].

Low-energy grain boundaryAngle range (°)
31.99° [1$\bar{2}$10]32±2
43.11° [1$\bar{2}$10]45±2
61.91° [1$\bar{2}$10]62±2
75.06° [1$\bar{2}$10]75±2
28.41° [01$\bar{1}$0]28±2
39.06° [01$\bar{1}$0]39±2
58.36° [01$\bar{1}$0]58±2
72.88° [01$\bar{1}$0]73±2

New window

Fig. 14.   Rotation axis distribution of 8 special grain boundaries shown in Table 3 in hot-rolled samples after annealing at 400 °C and 450 °C.

3.2.3. Schematic diagram of SRX process

In the Sections 3.2.1 and 3.2.2, we have unveiled the SRX nucleation and growth processes of hot-rolled sample during annealing at different temperatures in detail. For making readers better understand this SRX process, it is fairly necessary to draw a simple schematic diagram of SRX process, as shown in Fig. 15. Stage ① represents the initial homogenized state at 450 °C. In this stage, the microstructure is mainly comprised of chain-shaped Al2Ca phases and granule-shaped Al8Mn5 phases distributed uniformly insides grains and at grain boundaries. After multi-pass hot-rolling, there is an apparent characteristic that chain-shaped Al2Ca phases take place breaking, along with the formation of a large number of twins (including extension twins, contraction twins and double twins) and dislocations (containing basal <a> dis., prismatic <a> dis. and pyramidal <c+a> dis.). At low temperature annealing of 200 °C, SRXed grains prior to nucleate at the around of cracked chain-shaped Al2Ca phases, as shown in stage ③. Besides the cracked chain-shaped Al2Ca phases, with the increase of annealing temperature to 300 °C, contraction twins, intersections of double twins, intersections of double twins and grain boundaries, as well as orientation gradients induced by non-basal slips also become effective nucleation sites to promote the formation of SRXed grains, coupled with nucleation-oriented elliptical annular texture formation. On further annealing from 300-450 °C, the grains with 45°-70° TD orientation happen to the preferred growth, since the co-segregation of Al, Zn and Ca atoms in grain boundaries gives rise to the occurrence of new low-energy grain boundaries hindering the growth of grains with basal orientation. This process occurs in stages ⑤-⑥.

Fig. 15.   Schematic diagram of SRX process: stage ①① represents the initial homogenized state at 450 °C; stages ②-④ show the SRX nucleation process at low temperature annealing (200-250 °C); stages ⑤-⑥ demonstrates the grain growth of SRXed grains (300-450 °C).

3.3. SRX kinetics

3.3.1. Nucleation kinetics of SRX process

Vickers hardness tests were carried out to track the SRX process during annealing. The dependences of the hardness on annealing time after annealing at 200, 300 and 450 °C are compared in Fig. 16(a), hardness data are summarized in Table 4, Table 5, Table 6. Under these three annealing temperatures, there is a similar tendency that the hardness firstly drops slowly, and then it decreases rapidly. Finally, the rate of softening decreases until a steady state is reached. With increasing the annealing temperature, the softening time is increasingly shorter. Even at 450 °C, the hardness drops rapidly in the first 10 min. For better comparison of the SRX progress in the hot-rolled samples during annealing at 200, 300 and 450 °C, the hardness data are converted into the fractional softening (XH) form given as follows [37]:

XH=(H0-Hi)/(H0-Hr) (1)

where H0 (∼71 HV) is the hardness of hot-rolled sample, Hi is the measured hardness during annealing, and Hr is the hardness when full recrystallization has been achieved. To avoid the misunderstanding, the state with finest SRXed grain size corresponding to each annealing temperature is regarded as full recrystallization in each annealed sample. Fig. 16(b) shows the fractional softening behaviors of hot-rolled samples during annealing at 200, 300 and 450 °C for different time. It can be seen that XH-t curves exhibits a similar sigmoid shape, which is accorded with Johnson-Mehl-Avrami-Kolmogorov (JMAK) model given as follows [38]:

XH=1-exp(-Btn) (2)

where B is a factor related to grain shape, t is the annealing time, and n is Avrami exponent, which can be calculated by linear relationship from the conversion of Eq. (2). This conversion is given as follows:

lnln(1/(1-XH)=lnB+nlnt (3)

Fig. 16.   (a) Hardness of annealing samples at 200, 300 and 450 °C as a function of annealing time; (b) fractional softening (XH) of annealing samples at 200, 300 and 450 °C as a function of annealing time; (c) lnln(1/(1-XH) of annealing samples at 200, 300 and 450 °C vs. ln t; (d) linear fitting of Eq. (5): ln(1/tR) versus 1/T.

Table 4   Hardness of hot-rolled sample after annealing at 200 °C for different time.

Annealing time (min)
1030601206009602280
Hardness (HV)70.168.064.960.858.257.157.0

New window

Table 5   Hardness of hot-rolled sample after annealing at 300 °C for different time.

Annealing time (min)
0.515102030405060120300
Hardness (HV)69.068.164.361.559.658.157.657.156.955.954.3

New window

Table 6   Hardness of hot-rolled sample after annealing at 450 °C for different time.

Annealing time (min)
0.10.5135102030405060120
Hardness (HV)69.968.165.162.261.355.153.753.753.253.153.152.3

New window

The calculated results are shown in Fig. 16(c). The n values are 0.68, 0.70 and 1.02 for samples annealed at 200, 300 and 450 °C, respectively. Su et al. [39] revealed that the n value was found to be a function of annealing temperature and deformation strain, and the n value raised with the increase of temperature and strain. Our result is consistent with literature. In addition, some researches pointed out that the low n value was likely to be relative to the non-random recrystallization [40]. It has been reported that the range of n value in as-deformed AZ31 alloy after annealing was 1.2-3.4 [38,39], which is higher than that in the present study. It indicates that there may be a small number of effective nucleation sites or high storage energy regions in the as-deformed AZ31 alloy leading to the relatively random recrystallization during annealing. Nevertheless, in this work, we have proved that massive cracked chain-shaped Al2Ca phases, contraction twins, intersections of double twins, intersections of double twins and grain boundaries as well as orientation gradients insides deformed grains in the hot-rolled sample are effective nucleation sites. SRXed grains nucleate preferentially in these high storage energy regions, giving rise to the relatively lower n value than that of AZ31 alloy. Recently, Su et al. [37] found that the AZ31 alloy subjected to high speed rolling also exhibited a relatively lower n value (0.7-2.0) after annealing, which was attributed to the formation of a large number of shear bands and twins preferentially acted as nucleation sites. This literature directly verifies our statement.

According to the Arrhenius expression [38], the rate of SRX can be described as follows:

VR=Aexp(-QR/RT) (4)

where VR is the rate of recrystallization and considered to be directly proportional to (1/tR), tR is the annealing time required for the complete SRX, A is a constant, R is the universal gas constant, QR is the activation energy of SRX nucleation and T is the annealing temperature. In order to obtain the QR value, Eq. (4) is transformed into a new formalism as follows:

ln(1/tR)=lnA-QR/RT (5)

According to the fitting ln(1/tR) - (1/T) linear curve shown in Fig. 16(d), the slope of linear curve, -8.94, is calculated. Therefore, the activation energy QR value is ∼74.24 kJ/mol, which is still lower than that of a heavily cold drawn AZ31 alloy (∼85.9 kJ/mol) during annealing, meaning that the hot-rolled sample easily takes place SRX nucleation.

3.3.2. Grain growth kinetics of SRX process

The grain growth of hot-rolled sample upon annealing is further quantitatively analyzed in this section. We select two annealing temperatures, 300 °C and 450 °C, to carry out annealing for different time and observe grain growth behavior. Fig. S3 in Supplementary Material compares the microstructures of hot-rolled samples after annealing at 300 °C and 450 °C for 1, 5 and 10 h, respectively. It can be seen that, combined with the average grain size as a function of the annealing time of 1-10 h (see Fig. 17(a)), the average grain size for both samples increase with the increase of annealing time. In order to calculate the grain growth kinetics of SRX process, we use a model proposed by Burke and Turnbull [41]. This model was based on a hypothesis that SRXed grains grew by grain boundary migration. They pointed out that the growth of grains was primarily driven by a pressure on grain boundaries, and the rate of migration was inversely proportional to the radius of the curvature of grain boundaries [41]. According to the theory of Burke and Turnbull, the relationship between the average grain size (D) of sample and the annealing time (t) can be given as follows [41]:

$D_{ m } -D_{0}^{ m } =kt $ (6)

where D0 is the initial average grain size (at t = 0), m is the grain growth exponent, and k is constant depending on the material composition and temperature. In general, D0 is roughly considered as zero by assuming D>> D0 [38,42]. After differential treatment, Eq. (6) is conversed into Eq. (7) as follows:

dDm/dt=k (7)

Fig. 17.   (a) Average grain size, D, of hot-rolled samples after annealing at 300 °C and 450 °C as a function of annealing time; (b) linear fitting of Eq. (11): ln(dD/dt) as a function of lnD; (c) linear fitting of Eq. (13): lnk versus (1/T).

Eq. (7) can be further written as follows:

$\frac{dD^{m}}{dD} \cdot \frac{dD}{dt}=k$ (8)

and then

m∙Dm-1∙$\frac{dD}{ dt }$ =k (9)

and then taking the logarithm of both sides as follows:

lnm+(m-1)lnD+$ln\frac{dD}{ dt } $=lnk (10)

and then rearranging as follows:

$ln\frac{dD}{ dt } $ =-(m-1)lnD+ (11)

Fig. 17(b) shows the results of linear fitting of Eq. (11) for the samples annealed at 300 °C and 450 °C. The slope of linear fitting curve is -(m-1), and the intercept is $ln\frac{k}{ m }$. According to the calculated results, the grain growth exponent m values are 1.97 and 2.27, respectively, and the k values are 7.90×10-14 mm/s and 5.33×10-16 mm/s, respectively. From the atomic point of view, the grain boundary migration actually is an atomic diffusion process by thermal activation. Based on the Arrhenius-type equation [43], the relationship between the constant k and the activation energy of grain growth Qg can be written as follows:

k=k0exp(-$\frac{Q_{ g } }{ RT }$) (12)

where k0 is the constant, and then taking the logarithm of both sides as follows:

lnk-lnk0=(-$\frac{Q_{ g } }{ RT }$) (13)

Fig. 17(c) shows the result of linear fitting for Eq. (13). The slope of linear fitting curve is regarded as -$\frac{Q_{g}}{ R }$, and the calculated Qg value is ∼115.48 kJ/mol, which is remarkably higher than that of AZ31 alloy (80.8-92.5 kJ/mol). In other words, the diffusivity of Mg atom for the hot-rolled sample in this work is apparently lower than that for AZ31 alloy during grain growth. It has been reported that the addition of solutes was beneficial for the enhanced activation energy of grain boundary migration [41,44,45], because grain growth is necessary to conquer solute drag effect, especially grain boundary pinning. When the interaction energy of solutes and grain boundary is higher than the driving force of grain boundary migration, the activation energy is mainly depended on solutes diffusion. If this driving force takes over the solutes-grain boundary interaction energy, the activation energy is mainly controlled by boundary self-diffusion [44]. Obviously, in this work, the co-segregation of Al, Zn and Ca atoms in grain boundaries for the hot-rolled sample after annealing strongly induces the increase of solutes-grain boundary interaction energy giving rise to the increased activation energy, compared with AZ31 alloy.

4. Conclusions

In this work, we investigated the evolution of elliptical annular texture in hot-rolled AZMX1100 alloy after annealing at different temperatures for 1 h, and unveiled the origin of such an elliptical annular texture formation. In addition, the static recrystallization kinetics of hot-rolled AZMX1100 alloy was also systematically analyzed in given annealing temperature for different time. The primary conclusions were listed as follows:

(1)AZMX1100 alloy subjected to multi-pass hot-rolling exhibited a double-peak texture characteristic oriented with c-axis tilted ∼15° towards the RD. With increasing the annealing temperature, such a double-peak texture gradually transformed into an elliptical annular texture extended along the TD. When annealing temperature was up to 400 °C, it had a uniform weak elliptical annular texture.

(2)When annealing temperature was ranged from 200-300 °C, grain nucleation dominated the SRX behavior of hot-rolled AZMX1100 alloy. Nucleation mechanisms induced by cracked chain-shaped Al2Ca phases, contraction twins, intersections of double twins, intersections of double twins and grain boundaries and non-basal slips jointly contributed to the formation of elliptical annular texture.

(3)When annealing temperature was ranged from 300-450 °C, grain growth played a significant role on the SRX behavior of hot-rolled AZMX1100 alloy. The grains with 45°-70° TD orientation happened to the preferred growth, since the co-segregation of Al, Zn and Ca atoms in grain boundaries gave rise to the occurrence of new low-energy grain boundaries (58° [1 $\bar{2}$ 10], 73° [1 $\bar{2}$ 10] and 75° [01 $\bar{1}$ 0] grain boundaries) hindering the growth of grains with basal orientation.

(4)The nucleation kinetics of static recrystallization of hot-rolled AZMX1100 alloy could be well described by JMAK model in 200-450 °C. The Avrami exponent n value was ranged from 0.68 to 1.02 due to the non-random nucleation mechanisms mentioned above of recrystallization nuclei, which led to the lower activation energy QR of nucleation of ∼74.24 KJ/mol.

(5)Activation energy Qg for grain growth was determined by annealing ranged from 300 °C to 500 °C. The Qg value, fitted to an Arrhenius relationship, was estimated to be ∼115.48 KJ/mol. It was attributed to the strong pinning of co-segregated Al, Zn and Ca atoms in grain boundaries.

Acknowledgements

The work was financially supported by the National Natural Science Foundation of China (Nos. 51531002 and U1764253), the National Key Research and Development Program of China (Nos. 2016YFB0301104 and 2016YFB0101700), the Chongqing Scientific & Technological Talents Program (No. KJXX2017002), the Chongqing Science and Technology Commission (No.cstc2018jcyjAX0472), and the Science and Technology Research Program of Chongqing Municipal Education Commission (No. KJQN201801306).


/