Recent progress in preparation of bimodal grain structure and improvement of mechanical and corrosion properties of magnesium alloys

Abstract

Nowadays, magnesium (Mg) alloys are increasingly being considered as a promising lightweight structural material because of their inherent low density and high specific strength. However, the broad application of most Mg alloys is limited by their poor strong-ductility trade-off at room temperature. At present, tailoring the bimodal grain structure (BGS) has the potential to concurrently enhance strength and ductility. Therefore, in order to achieve controllable preparation of BGS, recent research on the formation of Mg alloy with BGS was evaluated from the perspectives of alloy design, initial microstructure of billets, and process parameter. In addition to summarizing the role of BGS in synergistically improving strength and ductility, the influence of BGS on other properties of Mg alloys was also summarized, such as tension-compression yield asymmetry, corrosion resistance, fatigue performance, and fracture toughness. The research progress in controlling BGS provides valuable insights into the design and production of high-performance Mg alloys.

1. Introduction

Mg alloys have garnered significant attention due to their inherent characteristics, including low density, high specific strength, and high stiffness, particularly in the automotive and aerospace fields [1,2]. However, Mg and its alloys often exhibit poor strength, low ductility, and strong anisotropy at room temperature, which limits their application. Hitherto, researchers have made great efforts to improve the mechanical properties through various strategies, such as alloying [3] and grain refinement [[4], [5], [6]]. However, alloying is difficult to simultaneously improve strength and ductility, and the addition of alloying elements further increases the cost of Mg alloys. Similarly, enhancing the strength of Mg alloys can be achieved by refining grains to build structures with ultrafine grains (UFG) or nano grains (NG). Nevertheless, this is usually accompanied by a decrease in ductility [7]. A major reason is the low dislocation storage efficiency within UFG or NG. Therefore, UFG or NG structures are susceptible to plastic instability due to the loss of work-hardenability [8].

A novel approach involving heterogeneous microstructure has recently been adopted to improve the strength-ductility trade-off dilemma. The excellent mechanical properties are mainly related to heterogeneous deformation, which leads to the accumulation of geometrically necessary dislocations (GNDs) at the interface of soft/hard domains, resulting in hetero-deformation induced (HDI) strengthening [9]. The role of heterogeneous microstructure in simultaneously improving the strength and ductility has been discovered in steels [10,11], titanium alloys [12], and Cu alloys [13], etc. The introduction of BGS in Mg alloys, a kind of heterogeneous microstructure, has been proven to be an effective method for obtaining excellent mechanical properties, including strength, ductility, and yield asymmetry [[14], [15], [16]].

In general, there are two main types of BGS in Mg alloys, one consisting of coarse dynamic recrystallized (DRXed) grains and fine DRXed grains (Fig. 1a). This type of BGS can be attributed to the inhomogeneous distribution of dynamic precipitates [14,17] or abnormal grain growth [18]. Another type is composed of coarse undynamic recrystallized (unDRXed) grains with strong texture and DRXed grains with weakened texture (Fig. 1b, c). For this type, partial dynamic recrystallization (DRX) and grain growth are the key mechanisms to form this type of BGS [[19], [20], [21], [22]]. Overall, BGS in Mg alloys typically exhibit a combination of fine/ultrafine and coarse grains. Currently, there exist various techniques for fabricating BGS in Mg alloys, such as extrusion [[23], [24], [25]], rolling [[26], [27], [28]], powder metallurgy [29], and surface mechanical grinding treatment [30]. Process parameters can be utilized to adjust the degree of DRX and DRXed grain size, hence influencing the formation of BGS and ultimately achieving improvement in the mechanical properties of Mg alloys.

Fig. 1.

Microstructural characteristics of BGS: (a) as-extruded AM80 alloy [14], (b) as-extruded Mg-11.5Gd-4.5Y-1.5Zn-0.3Zr (wt.%) alloy [22], (c) as-rolled AZ91 alloy [28].

This paper provides a comprehensive overview of the factors influencing the formation of BGS, such as alloy composition, the initial structure prior to processing, and processing parameters. The primary focus is on the impact of BGS on the mechanical properties of Mg alloys and the associated underlying mechanisms, aiming to offer valuable insights for the manufacture of high-performance Mg alloys.

2. The formation of bimodal grain structure

2.1. The effect of alloying on bimodal grain structure

Recent studies have demonstrated that the formation of BGS is closely related to the addition of alloying elements, such as Al [31,32], Mn [33,34], Ce [35], and Sm [36]. On the one hand, the inclusion of alloying elements, especially rare earth elements (REs), greatly reduces the stacking fault energy (SFE), thereby enhancing the driving force for recrystallization and facilitating DRX. As a result, a wider area is covered by the DRXed grains [37]. On the other hand, the addition of alloying elements can lead to the formation of second phases in Mg alloys and promote DRX through the particle stimulated nucleation (PSN) mechanism induced by large particles (>1 µm) [38,39]. For instance, Yu et al. [32] found that the PSN effect of Al₂Nd and Al₁₁Nd₃ phases in the hot-extruded Mg-3Nd alloys increased with higher Al content, resulting in the transformation of the microstructure from BGS to a nearly fully recrystallized structure, and consequently, an increased volume fraction of the DRXed grains. This situation was also found in as-extruded AZ31-xSm alloy, as shown in Fig. 2 [36]. However, in the Mg-Sn-Ca-Mn alloy, the volume fraction of the DRXed grains decreased with the addition of Al [31]. This was attributed to the hindered precipitation of large CaMgSn phases by the addition of Al, leading to a reduction in nucleation sites for DRXed grains and a subsequent decrease in the proportion of DRXed grains. Therefore, it should be considered whether the formed second phase can act as a nucleation site when incorporating alloying elements to regulate BGS.

Fig. 2.

Schematic illustration showing microstructure evolution of AZ31-xSm alloys under as-cast and as-extruded conditions [36].

2.2. The effect of initial microstructure on bimodal grain structure before deformation

2.2.1. Initial grain size

The effect of the initial grain size on BGS is related to the deformation temperature, which is known to determine the DRX mechanism. In general, at low temperatures (< 473 K), DRX is associated with twinning, while continuous DRX (CDRX) occurs at intermediate temperatures (473–523 K) and discontinuous DRX (DDRX) dominates at high temperatures (> 573 K) [33].

In the case of the DDRX mechanism, grain boundaries are preferred sites for nucleation. Therefore, a finer initial grain size provides more nucleation sites and faster recrystallization kinetics, resulting in a significant increase in the size and proportion of DRXed grains [40,41]. For example, large numbers of fine grains produced by equal-channel angular pressing (ECAP) processing provided an abundance of grain boundaries, offering abundant nucleation sites for DRX during subsequent rolling, thus forming the high proportion of fine grains [42]. It is well known that stored energy is the basis for recrystallization. Grain boundaries effectively impede the movement of dislocations and facilitate the formation of dislocation pile-ups, making them favorable sites for DRX [43]. Therefore, the initial grain size can significantly influence the microstructure of wrought Mg alloys under DDRX-dominant deformation conditions. A critical grain size (d_c) of DDRX exists under specific deformation conditions. If the initial grain size exceeds the critical grain size, achieving full DRX becomes challenging, as depicted in Fig. 3 [44]. Conversely, if the initial grain size in the as-cast conditions is smaller than the critical size, the critical strain required for the initiation and completion of DRX is reduced, increasing the opportunity to obtain a complete DRX microstructure under the same processing conditions [45].

Fig. 3.

A model for the grain refinement of different initial structures of Mg alloys processed by ECAP. The left column shows the initial condition, the second column shows the structure after one pass, and the third column shows the structure after multiple passes [44].

While CDRX occurs through the continuous absorption of dislocations at low-angle grain boundaries (LAGBs), ultimately leading to the formation of high-angle grain boundaries (HAGBs) and nucleation [46]. Therefore, the initial grain size has no influence on the microstructure of wrought Mg alloys under the dominant deformation conditions of CDRX.

2.2.2. Texture

Azeem et al. [47] found that in addition to fine DRXed and large elongated unDRXed grains, there were also abnormally grown grains in as-extruded AZ21 alloy, and believed that this was related to the texture of the pre-extruded material. Therefore, according to the initial grain orientation and extrusion direction (ED), the grains of the pre-extruded materials are divided into three types: the grain basal plane is aligned within 5 along the ED (AGs); the grain basal plane slightly deviates from the ED (MGs); the grain basal plane is in random direction (RGs). During extrusion, all the grains are subjected to strain which tends to elongate them in the ED. Depending on the grain orientation and Schmid factor, the strain hardening and stored plastic energy of each grain are different [48]. Since the AGs grain orientation is aligned with the ED, it is easy to slip in the basal slip system have the least stored plastic energy, while RGs are forced to have dislocation activity in multiple slip systems to accommodate strains, so they have greater strain hardening and greater stored plastic energy. The AGs grain orientation is between the AGs and the RGs, and the stored plastic energy is also between the two grains. The stored plastic energy can trigger recrystallization under high temperature extrusion, and the nucleation rate during recrystallization is proportional to the stored strain energy density. Thus RGs with high storage energy undergo a high rate of recrystallization. MGs have lower storage energy and therefore a lower nucleation rate, resulting in larger DRX grains. AGs have the least stored plastic energy and cannot trigger recrystallization, so they become unDRXed grains. This shows that the pre-extruded texture affects the recrystallization behavior during thermal deformation and is illustrated in Fig. 4.

Fig. 4.

Schematic representation of (a) pre-extruded microstructure; (b) intermediate microstructure during extrusion; (c) actual binary micrograph of the as-extruded materials [47].

In addition, in the static annealing process of wrought magnesium alloys, due to the different stored energy or orientation of each grain [49], specific grains preferentially grow at the expense of smaller grains, resulting in a bimodal grain size distribution [18,47,50]. In general, the area fraction of large grains increases with annealing temperature and that of small grains decreases. The decrease in the number density of small grains can be attributed to two independent mechanisms: the first mechanism is the consumption of small grains by large ones due to surface-driven phenomenon and Ostwald ripening. The second mechanism is the small grains with non-basal orientation grow abnormally at the expense of small grains with basal orientation [18].

2.2.3. Twinning

Moreover, the presence of artificially formed twins within the billet can also significantly influence the DRX behavior during hot deformation, resulting in a great variation in the final microstructure. Park et al. [51] reported that twin-induced DRX (TDRX) occurred at pre-existing twins during hot deformation of as-rolled AZ31 alloy, resulting in the formation of BGS. Similarly, the cold pre-forging (CPF) process induced a large number of {10$\bar{1}$2} twins in the billet, which in turn provided nucleation sites for DRX during the extrusion process, leading to an increase in the proportion of DRXed grain in the as-extruded alloy [52]. Furthermore, pre-compression prior to ECAP can also induce twinning, and AZ31 alloy with a higher degree of induced twinning exhibited the highest volume fraction of DRXed grains under the same shear deformation condition [53]. Fig. 5 illustrates the mechanism of TDRX [54]. In the deformation process of Mg-6Al-3Sn-1Zn (wt.%) alloy, the movement of the dislocation was limited by the precipitated phase and the twin boundary resulting in the formation of dislocation cells. These tangled dislocation cells bulged outside the twin boundaries, causing the twins to acquire a bamboo-like shape. A substructure is formed in the twin. These are quadrilateral structures with low density dislocations surrounded by entangled dislocations. And tangled dislocations can then absorb adjacent dislocations to form LAGBs, leading to the formation of subgrains. Then the increase of misorientation transformed LAGBs into HAGBs which eventually leads to the formation of BGS.

Fig. 5.

Schematic illustration showing the TDRX. The formation of (a) dislocations and precipitates; (b) tangled dislocations; (c) subgrains; and (d) DRXed grains [54].

2.2.4. Precipitates

The as-cast microstructure will affect the microstructure of subsequent hot-deformed samples. Michiaki et al. [55] found that the average grain size of the fine DRXed grain in the as-extruded Mg-Zn-Y alloy remained constant regardless of the secondary dendrite arm spacing (SDAS) in the as-cast alloy, but the volume fraction of the DRXed region increased with the decreasing SDAS (Fig. 6a). This is because the dispersion degree of long-period stacking ordered (LPSO) phase increased with decreasing SDAS in the as-cast Mg-Zn-Y alloy (Fig. 6b), and DRX occurred adjacent to the LPSO phase due to high strain concentration. This suggests that the SDAS is a significant factor in determining the morphology of the LPSO phase in the as-cast state, as well as the volume fraction of the DRXed grains region in the as-extruded alloy.

Fig. 6.

(a) Relationship between the as-extruded Mg₉₇Zn₁Y₂ alloys and the SDAS with respect to the as-solidified state in LPSO phase dispersion, (b) Change in the volume fractions of two types of matrices (the DRXed α-Mg region and the hot-worked α-Mg region) as a function of SDAS [55].

In addition, the increase in dislocation density caused by the interaction between dislocations and pre-existing precipitates during hot extrusion may enhance the occurrence of DRX. Therefore, pre-existing precipitates can also affect the DRX behavior during hot deformation, leading to significant alterations in the final microstructure. The formation of BGS can be achieved by adding alloying elements to obtain undissolved precipitates. For example, the addition of Y to the Mg-Zn-Zr alloy can lead to the precipitation of Mg₃Zn₃Y₂ (W) phase. The undissolved coarse W phase in the as-homogenized alloy can be crushed into micron-sized particles (>1 µm) during the extrusion process, and then the W phase particles act as nucleation sites of the DRXed grains by the PSN mechanism [56].

Besides, various precipitates can be obtained and the density of precipitates can be controlled by different heat treatments (such as pre-solution, pre-aging, etc.). Compared with AZ80 alloy without pre-aging, the fine Mg₁₇Al₁₂ precipitates formed by aging prior to extrusion (APE), which can reduce the average size of DRXed grains and increase the proportion of fine DRXed grains in AZ80 alloy after extrusion [17]. Furthermore, as shown in Fig. 7, the density of precipitates increased with the increase of aging time, and the proportion of fine grain layers gradually increased due to the pinning effect of a large number of nanoscale precipitates dispersed throughout it, forming various volume fraction of BGS [57]. Therefore, controllable preparation of BGS can be achieved by adjusting the density of precipitates.

Fig. 7.

Scanning electron microscope (SEM) micrographs and inverse pole figure (IPF) maps of with different pre-aging times [57].

2.3. The influence of dynamic precipitation and solute segregation behaviors on bimodal grain structure

2.3.1. Dynamic precipitation

The effect of dynamic precipitation on BGS has been studied extensively. In many wrought magnesium alloys, such as Mg-Al-based alloys [58], Mg-Gd-based alloys [[59], [60], [61]], Mg-Sn-based alloys [62], dynamic precipitation is widely observed in the DRXed region and mainly occurs on the DRXed grain boundaries, but few precipitates are found within the interiors of deformed grains. The dynamic precipitation can hinder the growth of DRXed grain via the Zener pinning effect, resulting in DRXed grain refinement. The pinning effect can be expressed by Zener pressure, and its formula is as follows [58]:

$$P_Z=\pi r_P\gamma n_S$$ (1)

where P_Z is the Zener pressure of particles with radius of r_P and the quantity density per unit boundary area of n_S; γ is the grain boundary tension. It is evident from Eq. 1 that the pinning effect depends on both the size and the quantity density of the precipitates. For the former, smaller particles (<1 µm) tend to reduce the DRXed grain size and inhibit the DRX process [63,64], while larger particles (>1 µm) can facilitate DRX through PSN. For the latter, the proportion of DRX tends to decrease significantly with the increase of dynamic precipitation [59].

The occurrence of dynamic precipitation is related to deformation temperature or strain rate. He et al. [58] established and verified a phenomenological model describing the quantity density n_P and average size d_P of precipitates at different deformation conditions, and the function forms are as follows:

$$\begin{array}{c}{\mathrm{n}}_{\mathrm{P}}=5.760\times 10^{-10}{\dot{\epsilon }}^{0.326}\text{exp}\left({\displaystyle \frac{1.031\times 10^5}{\text{RT}}}\right)\\ {\mathrm{d}}_{\mathrm{p}}=1.643\times 10^4{\left[\dot{\epsilon }\text{exp}\left({\displaystyle \frac{1.093\times 10^5}{\text{RT}}}\right)\right]}^{-0.202}\end{array}$$ (2)

where $\dot{\epsilon }$ is strain rate; T is deformation temperature; R is the molar gas constant. It is evident from Eq. 2 that an increase in temperature or a decrease in strain rate results in a decreased number density and an increased average size of the precipitates. By integrating this model with the Zener pressure formula, it becomes feasible to predict the ultimate microstructure after thermal deformation under various deformation conditions.

In addition, dynamic precipitation is mainly concentrated at DRXed grain boundaries, which is closely related to the obvious localized strain between unDRXed and DRXed regions. Strain-induced defects may lead to non-equilibrium segregation of solute atoms and promoting nucleation of precipitates [65], which leads to dynamic precipitation near the grain boundaries. With the development of DRXed grains, the newly formed grain boundaries became a new barrier, constantly hindering dislocation movement. The reduction of interfacial and misfit energy promotes nucleation of dynamic precipitates [62]. Thus, the dynamic precipitation is non-uniformly distributed in front of the DRXed/unDRXed boundary, and the area fraction of dynamic precipitation increases with the expansion of the DRXed region into the deformed grain, as shown in Fig. 8.

Fig. 8.

Schematic diagram of the dynamic precipitation influenced by DRX [62].

2.3.2. Solute segregation

Solute segregation along DRXed grain boundaries has been widely observed in many magnesium alloys, such as Mg-Al-based alloys [[66], [67], [68]], Mg-RE-based alloys [61,69,70], Mg-Sm-based alloys [71]. Similar to dynamic precipitation, solute segregation along grain boundaries can effectively pin grain boundaries and strongly inhibit grain boundary migration. Solute segregation along grain boundaries has two effects on recrystallization behavior. On the one hand, it strongly inhibits the nucleation and growth of DRXed grains [66]. On the other hand, it inhibits DDRX behavior by hindering grain boundary migration, thus promoting the occurrence of CDRX [70]. Therefore, grain boundary segregation is advantageous for the development of BGS.

2.4. The effect of processing parameters on bimodal grain structure

2.4.1. Extrusion

In recent years, there has been a growing interest in as-extruded Mg alloys due to their superior mechanical properties compared to cast Mg alloys. Additionally, as-extruded Mg alloys are easier to fabricate using a one-pass process compared to as-rolled products. The microstructure of as-extruded Mg alloy can be affected by not only the initial billet conditions and alloy composition, but also the extrusion process parameters (such as temperature, speed, and ratio, etc.) and extrusion die [[23], [24], [25]]. The hot extrusion process can impact the DRX degree and DRXed grain size, allowing for controlled preparation of BGS.

2.4.2. Extrusion temperature

The temperature has an impact on the dislocation density [72], the number and distribution of second phases [73,74], as well as the diffusion rate of grain boundaries, thereby affecting recrystallization nucleation and growth. In general, the deformed Mg alloys at relatively low temperatures usually exhibit a partially DRXed microstructure, consisting of both fine equiaxed DRXed grains and coarse elongated unDRXed grains [[75], [76], [77]]. As the temperature increases, the proportion of DRXed grains gradually increases. The existence of unDRXed grains can be attributed to solute segregation or second phases leading to grain boundary pinning, which hinders grain boundary migration and retards DRX [31,78]. And the effect of temperature on DRXed grain size is generally expressed in terms of the Zener-Hollomon parameter (Z), as follows:

$$\mathrm{Z}=\dot{\epsilon }exp\left(\frac{Q}{\text{RT}}\right)$$ (3)

where $\dot{\epsilon }$ is strain rate, Q is activation energy similar to that of self-diffusion and R is gas constant.

A relation between Z and grain size is given by Eq. 4 [79], as follows:

$$\mathrm{Z}{\mathrm{D}}^{\mathrm{m}}=\text{constant}$$ (4)

where D is the grain size and m is a positive constant. Therefore, a lower temperature leads to a higher Z and a finer grain size.

2.4.3. Extrusion speed

The microstructure of as-extruded Mg alloy is influenced not only by the extrusion temperature, but also by the extrusion speed. As the extrusion speed increased from 0.3 mm/s to 3.5 mm/s, the microstructure characteristics transformed from BGS to the near complete DRX structure in Mg-2Sn-1.95Ca-0.5Ce (wt.%) alloy, and the grain size of DRXed grains gradually increased [80]. Similar observations were also observed in Mg-11Gd-3Y-0.5Nd-Zr (wt.%) alloys [81] and Mg-8Al-0.5Zn (wt.%) alloys [82]. This is attributed to the generation of excess heat during extrusion, resulting from plastic deformation and friction. In particular, both grain growth and DRX during extrusion necessitate thermal activation, thus the driving force for initiating grain growth and DRX will increase as the temperature of the billet increases [83]. Besides, with the extrusion speed increases, the volume fraction of the second phase gradually decreases while the size gradually increases, leading to a weakened pinning force on the grain boundary [80]. In combination with these reasons, the grain size and fraction of DRXed grains will increase with the increase of extrusion speed. Evidently, the extrusion speed is a crucial parameter influencing the DRX process. Therefore, controlling the extrusion speed to limit the temperature rise is essential for managing recrystallization and achieving partial recrystallization in Mg alloy.

2.4.4. Extrusion ratio

The extrusion ratio has a direct impact on the die exit temperature and strain rate, thus indirectly affecting the microstructure of as-extruded Mg alloys. The surface temperature rises by approximately 50 °C when the extrusion ratio is increased from 10 to 25 or from 25 to 50 while maintaining a constant initial billet temperature of 250 °C [84]. A higher extrusion ratio corresponds to a higher extrusion speed for the same ram speed [79]. Consequently, the increase in die exit temperature and speed leads to a higher proportion of DRXed grains and a relatively coarse DRXed grain size. However, some studies have found that with the increase of extrusion ratio, the proportion of DRXed grains increased while the average grain size decreased [85,86].

The variation of grain size is related to Z parameter, which is derived from Eq. 3. However, since the strain rate cannot be calculated simultaneously, Eq. 5 can be used to estimate the average strain rate [85], as follow:

$$\dot{\overline{\epsilon }}=\frac{{6\mathrm{D}}_{\mathrm{B}}^2\text{Vln}\left(\text{ER}\right)}{{\mathrm{D}}_{\mathrm{B}}^3{-\mathrm{D}}_{\mathrm{E}}^3}$$ (5)

where D_B and D_E indicate the inlet and outlet diameters of the extrusion die, respectively, V defines the punch velocity, ER is defined as extrusion ratio.

It is generally understood that a critical value of strain is required to induce recrystallization, and the size of recrystallized grains reduces as the amount of strain increases up to a second critical strain value. Beyond this point, the amount of strain has little further effect on the grain size [79]. At low extrusion ratios, the Z parameter increases with the increasing extrusion ratio, resulting in a finer grain size distribution in the microstructure [85]. As the extrusion ratio increases, the extrusion strain also increases and may exceed this critical value, which explains why finer grain sizes are not observed for the higher extrusion ratio.

The three extrusion parameters mentioned are closely related to Z parameters, which has a fundamental influence on the DRX kinetics. Li and Wu et al. [87,88] studied the effect of Z parameter on DRX of Mg-Zn-Zr-RE alloy and found that the alloy tends to achieve a higher DRX fraction when deformed at lower Z values at the same deformation degree, and the DRX fraction gradually approaches the constant value of 1 when the strain increases, indicating that complete recrystallization can be attained under sufficient deformation degree (Figs. 9(a-g)). Moreover, in it can be seen from Fig. 9(h) that the reduction in Z parameter accelerates the recrystallization process. The higher temperatures significantly enhance the migration of grain boundaries, which the rapid growth of new nuclei. Concurrently, the lower strain rates can provide sufficient time and space for the DRX nuclei to grow up prior to the emergence of the next generation of recrystallized nuclei [89]. Consequently, by considering the profound effect of the Z parameter on DRX, the formation of BGS can be predicted by adjusting the deformation temperature and strain. This provides a strategic approach to tailoring the microstructure and properties of alloys through precisely controlling extrusion parameters.

Fig. 9.

The influence of the Z parameters on (a-f) microstructure of specimens deformed at the strain of 0.6, (g) the DRX fraction, and (h) the DRX rate [87].

2.4.5. Extrusion die angle

Considering the impact of die geometry, it stands to reason that the die angle plays an important role in determining the microstructure and quality of the extruded products. Under a certain extrusion ratio, an increase in die angle leads to an increase in grain size, that is, the grain size increases in direct proportion to die angle. This relationship between die angle and grain size is particularly pronounced at lower extrusion ratios [85]. Therefore, it can be concluded that die angle is an effective parameter on grain size but that is not as significant as extrusion ratio. Besides, the die angle can influence the microstructural homogeneity of the as-extruded Mg alloy owing to variation in the flow rate, deformation temperature, and effective strain near the work zone during extrusion. As shown in Fig. 10, the microstructural inhomogeneity (i.e., unDRX fraction) of AZ31 alloy after indirect extrusion decreased in cross section with an increase in the die angle [90]. A faster metal flow, higher temperature, and larger effective strain are generated in an alloy extruded with a 90° die angle, thereby enhancing DRX behavior during extrusion and ultimately resulting in a homogeneously DRXed microstructure.

Fig. 10.

Inverse pole figure (IPF) maps of the sub-surface (S), quarter (Q), and center (C) regions of longitudinally sectioned samples of AZ31 extruded with different die angles. d_DRX indicates the average size of the recrystallized grains [90].

2.4.6. Rolling

At present, various rolling methods such as high strain rate rolling (HSRR) [91,92], asynchronous rolling (AR) [27,93], hard-plate rolling (HPR) [94,95] have been identified, leading to the formation of BGS during the deformation process. Furthermore, the DRX process in Mg alloys is significantly influenced by rolling parameters, resulting in noticeable changes in microstructure.

2.4.7. Rolling temperature

The effect of deformation temperature on the microstructure of as-rolled Mg alloys is similar to that of as-extruded Mg alloys, wherein a decrease in temperature leads to a simultaneous reduction in DRXed grain size and proportion. For example, due to the partial DRX caused by HPR at 350 °C, a BGS composed of coarse unDRXed grains (>∼70 µm) and fine DRXed grains (<∼5 µm) was obtained in AZ91 alloy [28]. In addition, reducing the deformation temperature to ∼300 °C results in coarse unDRXed microstructure containing substantial substructure, while increasing the deformation temperature to ∼400 °C results in almost uniform recrystallized grain structure. This phenomenon is also observed in high-speed rolling of AZ31 alloy from 300 °C to 400 °C, which is attributed to the promoted TDRX behavior [96].

2.4.8. Rolling cumulative reduction

The cumulative reduction is one of the key parameters in the rolling, which has a significant impact on the formation of the BGS. The rolling cumulative reduction can be controlled by changing thickness reduction, the number of rolling passes or HPR rolling. It is observed that as the cumulative reduction increases during hot rolling, there is an apparent increase in the fraction and refinement of the average size of fine grains, leading to the formation of an optimized BGS characterized by coarse grains uniformly embedded in fine grain regions [94]. Fig. 11 shows the relationship between the strength, average grain size of the coarse and fine grains, and total reduction ratio. It is noted that as the total reduction ratio increases, the average grain size gradually decreases while the fraction of fine grains gradually increases [26]. This is because DRX not only occurs around the α-Mg grain boundaries and the interphase boundaries but also nucleates at the twins.

Fig. 11.

Effect of the total reduction ratio on the strength (yield strength (YS) and ultimate tensile strength (UTS)) and mean grain size of the test specimen. The mean grain sizes of the fine and coarse grain regions are shown [26].

When the alloy is rolled with lower rolling cumulative reduction of 20% and 40%, insufficient driving force for recrystallization results in preferential twinning, as illustrated in Fig. 12b, c. When the rolling cumulative reduction was increased to 60% and 70%, the large strain is difficult to evenly disperse among the original grains, leading to more strain concentration near the grain boundary and accumulation of high-density dislocation nearby. This accumulation provides the driving force for the occurrence of DRX, resulting in the formation of a BGS where coarse grains are encased in fine ones, as shown in Fig. 12d [97]. Further increase in rolling cumulative reduction causes dislocations to reconcentrate at the newly formed DRXed grain boundaries, increasing the position of grain boundary induced nucleation, which helps promote the occurrence of DRX and further form finer DRXed grains [98,99].

Fig. 12.

Schematic diagram of the microstructure evolution of AZ91 alloys with varied rolling cumulative reductions. (a) 0%, (b) 20%, (c) 40%, (d) 60% & 70% [97].

HPR is also an effective method to improve rolling accumulation reduction, and the reduction of HPR is larger than that of conventional rolling (CR) under the same pass [100]. Moreover, the HPRed region contains a higher volume fraction of finer grains, where the volume fractions of fine grains in the HPRed and CRed regions are∼19% and∼6%, respectively (Fig. 13) [94]. This difference is attributed to the generally higher dislocation density in the HPRed region compared to the CRed region. The accumulation of high misorientation angle in coarse grains also indicates that coarse grains may transform into fine grains through CDRX with further deformation [101]. Thus, this suggests that fine grains are more likely to form through CDRX in the HPRed region due to faster accumulation of dislocations.

Fig. 13.

(a) IPF map of the intermediate-rolled sample in the RD-ND plane after rolling at 350 °C with a reduction of 55%, (b) and (c) details of the HPRed and CRed areas, respectively, (d) and (e) corresponding misorientation profiles measured along lines L1 and L2 [94].

2.4.9. Rolling speed

Currently, research has found that the initiation of DRX during hot rolling is related to both rolling speed and thickness reduction, rather than being solely influenced by a single factor [102]. In fact, rolling speed and thickness reduction can affect strain rate, and the corresponding average strain rate is calculated by Eq. 6 as follows:

$$\bar{\dot{\epsilon }}=\frac{\mathrm{H}-\mathrm{h}}{\mathrm{H}}·\frac{\mathrm{v}}{\sqrt{\mathrm{R}(\mathrm{H}-\mathrm{h})}}$$ (6)

where H is the initial thickness, h is the final thickness, v is the roll circumferential speed, and R is the roll radius.

Research has found that with a smaller thickness reduction (20% rolling reduction), the microstructures of samples rolled at different rolling speeds exhibited nearly identical characteristics. When the thickness reduction increased to 40%, there were differences in microstructure evolution in the samples rolled at different rolling speeds, that is, with the increase of rolling speed, the proportion of DRXed grains slightly increased and surrounded the large original grains [102]. Consequently, it can be inferred from the above discussion that DRX can only be operated on samples rolled at relatively high speeds and have large thickness reduction, which means that deformation at high strain rates can achieve controllable preparation of the BGS of the as-rolled Mg alloys.

At present, there is significant interest in an innovative HSRR technique. According to previous reports, the high strain rate was increased to over 5 s^-1 and even 9.6 s^-1 [91,92]. The Mg-Zn-Mn alloy was hot rolled by the single-pass HSRR at a strain rate ($\dot{\epsilon }$) of 9.1 s^-1, forming a BGS [91]. HSRR has the capability to generate numerous high strain microstructures, such as twinning and shear bands, which can be used as favorable recrystallization nucleation sites for non-basal orientation [92]. This is because the rapid rolling speed and delayed slip occurrence, resulting in greater stress concentration and potential rapid twinning deformation. Kim et al. It is found that during HSRR, many {10$\overline{1}$2} extension twins that were difficult to recrystallize form first [103]. Then, with the increase of the rolling reduction, {10$\overline{1}$1} contraction twins and {10$\overline{1}$1}-{10$\overline{1}$2} double twins were formed, which were beneficial for the crystal orientation of basal slip. These twins aggregated more dislocations, making them suitable for nucleation of new grains and promoting TDRX.

In general, the speed of the upper and lower rollers is consistent during the hot rolling process. However, a new machining technique known as differential speed rolling has been discovered, which introduces AR. This process involves the use of differential circumferential velocities of the upper and lower rolls to induce shear strain across the thickness of the sheet [27,93]. As a result, the samples produced through AR display a typical BGS, which is attributed to the shear force generated by the asynchronous rolling. This, in turn, leads to a higher driving force for recrystallization during the rolling process [104,105]. Therefore, the larger asynchronous ratio results in an increased driving force for recrystallization, leading to a higher recrystallization fraction, as shown in Fig. 14.

Fig. 14.

IPF map of different asynchronous ratio samples: (a) 1, (b) 1.1, (c) 1.2, (d) 1.3, recrystallization grain: (e) 1, (f) 1.1, (g) 1.2, (h) 1.3. (i) frequency of recrystallization grain [93].

In fact, the degree of microstructure evolution (i.e., the average size and the fraction of DRXed grains) is the result of the comprehensive effect of hot deformation controlling parameters, namely, deformation temperature, deformation rate and deformation amount. The operational influence of these processing parameters governs the nucleation process and significantly affects the rate of DRX, the size of DRXed grains, and the distribution of DRXed grains. Therefore, predicting the microstructural evolution throughout entirety of the hot deformation process requires simultaneous consideration of the influence of the above-mentioned microstructure control parameters.

2.4.10. Other deformation processes

In addition to the most commonly used processing techniques mentioned above that can prepare BGS in Mg alloys, the surface mechanical grinding treatment [30], powder metallurgy (PM) [29,106] and severe plastic deformation (SPD) [53,107,108] have also successfully prepared BGS of Mg alloys. For instance, the Mg­-0.85Zn-­2.05Y­-0.35Al (at%) alloy, prepared from rapidly solidified (RS) ribbons and heat-treated below 400 °C, exhibited a BGS consisting of coarse unDRXed grains (∼2.8 µm) with high kernel average misorientation (KAM) values (∼1.8°) and ultrafine DRXed grains (∼0.68 µm) with intermediate KAM values (∼1.1°) [109]. Additionally, Fu et al. [110] successfully prepared a bimodal Mg-7Y (wt.%) alloy by a novel route of high pressure boriding and semi-solid extrusion. Zou et al. [111] also successfully prepared ZK60 alloy with BGS using radial forging (RF) process. In the early deformation stage of forging, the material underwent significant grain refinement, but an inhomogeneous grain structure was formed due to the strain gradient along the radial direction. By increasing the RF pass, the grains in different radial parts were gradually refined, ultimately forming a BGS comprising coarse grains (∼14.1 µm) and fine (∼2.3 µm) grains. Moreover, other techniques such as multi-directional forging (MDF) [112] and hard plate hot forging (HPHF) [113] also exhibited BGS in Mg alloys, with the average grain size of the BGS gradually refining as the number of forging passes increased.

Furthermore, after reciprocating upsetting extrusion (RUE), DRXed grains with random textures were formed under the combined action of CDRX and DDRX mechanisms [114]. And as the strain increased, the deformed grains were gradually consumed by DRXed grains. In addition, Mg alloys treated with strong shear strain, such as ECAP [44] and high-pressure torsion (HPT) [115,116], have been observed to exhibit BGS with large initial grain sizes, attributed to the formation of shear bands within the alloys. When the strain reached a certain level, low angle boundaries were formed within the grains, releasing accumulated strain and forming shear bands. Further strain led to larger crystal rotations within the shear band, evolving into regions favorable for DRXed grains through low angle boundaries and ultimately developing into high angle boundaries [117].

3. The effect of bimodal grain structure on the properties of Mg alloys

3.1. Good strength-ductility trade-off

Magnesium alloy, as the lightest metal structural material, has great application potential in the field of lightweight. In order to achieve the wide application of magnesium alloys in load-bearing parts, high strength of magnesium alloys is required, in addition to high ductility in manufacturing processes involving deformation, such as stamping, forging and bending. Therefore, the strength-ductility trade-off is important for magnesium alloys because it directly affects their formability and suitability for various applications [118,119]. However, in certain cases, the formation of BGS has been found to significantly enhance the combination of strength and ductility in Mg alloys. For example, the as-extruded Mg-Gd-Y-Zn alloy exhibited a high UTS of 352 MPa and an elongation of approximately 14% [120]. Similarly, the Mg-8Gd-3Y-0.5Zr (wt%) alloy demonstrated a YS of 371 MPa, a UTS of 419 MPa, and an elongation of 15.8% [121]. The ZK60 alloy with BGS was obtained by conventional extrusion method with high TYS of ∼ 345 MPa, UTS of ∼ 370 MPa and high tensile strain of ∼ 20.5%, superior to most homogeneous ZK60 alloys reported so far (Fig. 15) [122]. Furthermore, the AZ91 alloy with BGS prepared by the hard plate rolling (HPR) method displayed a TYS of 374 MPa and an elongation of 19.6%, which was significantly better than the counterpart with a uniform recrystallized grain structure [28]. The mechanical properties of representative Mg alloys with BGS are detailed in Table 1. These findings suggest that the incorporation of BGS in Mg alloys represents a promising approach for achieving a favorable balance between strength and ductility.

Fig. 15.

YS vs. total elongation of ZK60 alloy with BGS and other reported homogeneous Mg-Zn based alloys [122].

Table 1.

Mechanical properties of representative Mg alloys with good strength-ductility synergy.

Alloy composition (wt%)	Processing condition	TYS (MPa)	UTS (MPa)	Elongation (%)	Ref.
Mg-9.32Gd-3.72Y-1.68Zn-0.72Zr	Extrusion	∼302.3	∼382.5	∼8.7	[123]
Mg-9Gd-4Y-1Zn-0.5Zr	Extrusion	∼323	∼381	∼15.1	[124]
Mg-8.2Gd-3.8Y-1.0Zn-0.4Zr	Rolling	∼385	420	∼19	[125]
Mg-8.2Gd-3.8Y-1Zn-0.4Zr	Rolling	∼318	∼403	∼13.7	[126]
Mg-6.2Gd-3.7Y-0.9Zn-0.3Zr-0.3Ag	Forging	∼292	∼384	∼9.0	[107]
Mg-0.6Mn-0.5Al-0.5Zn-0.4Ca	Rotary swaging	∼371	∼400	∼6.7	[108]
Mg-3.5Al-3.3Ca-0.4Mn	Extrusion	∼410	∼420	∼5.6	[77]
Mg-5Al-3Ca-0.3Mn	Extrusion	∼420	∼451	∼4.1	[21]
Mg-8Gd-4Y-0.5Mn-0.2Sc	Extrusion	∼315	∼380	∼11	[127]
Mg-10Gd-3Y-0.5Nd-0.5Zr	Extrusion	∼326	∼389	∼13.1	[75]
Mg-13Gd-4Y-2Zn-0.6Zr	Rolling	∼367	∼446	∼10.3	[128]
Mg-10Gd-3Y-1Zn-0.4Zr	Extrusion	∼420	∼500	∼4.6	[129]

In contrast to the homogeneous structure, Mg alloys with the BGS exhibit a higher tensile strength due to the presence of fine DRXed grains and coarse unDRXed grains with strong basal texture [130]. The fine DRXed grains contribute to strengthening through gain refinement hardening according to Hall-Petch relation for Mg alloys [5], while the strong fiber texture in the unDRXed grains leads to greater strength by reducing the Schmid factor of the basal slip system [131]. Meanwhile, severe plastic deformation introduces a mass of dislocations into the unDRXed grains, which are retained due to the low temperature, leading to numerous residual stresses. The combination of basal texture and residual stress in the unDRXed grains contributes to the high strength of the alloy [132]. However, the BGS composed of coarse and fine grains with similar textures does not always have a positive impact on strength, as it depends on the average grain size and the proportion of coarse and fine grains. The BGS only enhances the YS when satisfies Eq. 7 [133], as follow:

$${\mathrm{A}}_{\mathrm{f}}>\left(\frac{1}{\sqrt{{\mathrm{d}}_{\mathrm{u}}}}-\frac{1}{\sqrt{{\mathrm{d}}_{\mathrm{c}}}}\right)/\left(\frac{1}{\sqrt{{\mathrm{d}}_{\mathrm{f}}}}-\frac{1}{\sqrt{{\mathrm{d}}_{\mathrm{c}}}}\right)$$ (7)

where A_f is the proportion of fine grains in the BGS; d_f and d_c are the average grain size of fine and coarse grains in BGS, respectively; d_u is the average grain size of the uniform structure.

In addition, the randomization of basal texture in the fine DRXed grains can increase ductility by activating more deformation modes. Meanwhile, the role of unDRXed grains in enhancing ductility is currently mainly focused on two aspects. On the one hand, tensile twinning is formed in the unDRXed grains, which plays an important role in coordinating deformation and enhancing ductility in hexagonal compact packing (H.C.P.) systems by reorienting the crystal direction within the twinning to favorable slip directions [134]. However, the elongation to failure is greatly reduced in comparison with the alloy composed entirely of DRXed grains due to the generation of {10$\overline{1}$1}-{10$\overline{1}$2} double twinning within the unDRXed grains [25]. And it should be noted that the number of twin appearing in the unDRXed regions is very limited, making their contribution to ductility negligible. On the other hand, when the applied resolved shear stress exceeds the critical resolved shear stress (CRSS) value of the pyramidal slip in the Mg alloy, the coarse grains begin to play an important role in the storage of dislocation, thus promoting work hardening and resulting in a simultaneous higher strength and ductility compared to the fine-grained Mg alloy [135].

In fact, the synergistic improvement in both strength and ductility of BGS materials is mainly due to HDI strengthening and HDI strain hardening [136,137]. HDI strengthening increases YS, whereas HDI strain hardening plays a crucial role in enhancing the ductility. The HDI strengthening and HDI strain hardening arise from the combined effects of interactions between back stress and forward stress. Fig. 16 illustrates the emission of GNDs from a Frank-Read (F-R) dislocation source located within the soft zone. It is assumed that the F-R source becomes active at the critical shear stress τ₀, when τ_a> τ₀, the soft zone undergoes plastic deformation while the hard zone remains elastic, resulting in the accumulation of GNDs at the interface between the two zones to accommodate the strain gradient in the soft zone. The GND generated near the interface has the same Burgers vector, resulting in the generation of long-range internal stress that acts in an opposing direction to the applied shear stress. This long-range internal stress serves to inhibit the emission of more dislocations from the F-R source, thereby impeding plastic deformation within the soft zone, so it is called the back stress (τ_b). The higher applied stress is required to overcome both τ_b and τ₀ for the F-R source to release more dislocations, consequently strengthening the soft zone. Accumulation of GNDs at the boundary leads to stress concentration at the tip of the pile. This stress concentration, originating from the soft zone side, necessitates a balancing stress on the hard zone side, known as forward stress [136], which facilitates deformation within the hard zone. At the boundary, the back stress and forward stress exhibit same magnitudes but opposite directions, so the two cancel each other out. Nevertheless, these stresses have different profiles depending on the distance from the boundary, contributing to the observed HDI strengthening and HDI strain hardening.

Fig. 16.

Schematics of a GND pile-up, inducing back stress in the soft domain, which in turn induces forward stress in the hard domain [137].

Typically, unDRXed/deformed grains and fine DRXed grains are identified as hard zones, while coarse DRXed grains are considered as soft zones. The two zones are mechanically incompatible during the elastic-plastic deformation stage and plastic deformation stage. The mechanical incompatibility leads to strain distribution during deformation, where the softer zones are subjected to higher plastic strain [9,136,138]. The strain distribution has been identified through in-situ digital image correlation (DIC) technology in Mg-Gd-Y-Zn-Zr alloys with BGS, and it was observed that the multi-layer structure mitigates strain localization by transferring strain from DRXed to unDRXed regions, thereby enhancing the ductility of the alloy (Fig. 17) [139].

Fig. 17.

Strain distribution maps measured by DIC at different tensile strains of the Mg-Gd-Y-Zn-Zr alloys along ED: (a, b) low-strain range of 0.25∼1 pct and (c, d) high-strain range of 4∼14 pct. (a, c) normal strain ε_xx and (b, d) shear strain ε_xy. The grains with (0001) <11$\overline{2}$0> Schmid factor > 0.4 are marked by yellow arrows. The DRXed grains and unDRXed grains are schematically highlighted by white lines [139].

Furthermore, strain hardening is key to retaining the ductility of metals, especially those of high strength [140]. For the conventional homogeneous metals, an increase in dislocation density plays a dominant role in strain hardening. Strain hardening can be derived from the Taylor's equation [141], as follow:

$$\frac{\mathrm{d}\sigma }{\mathrm{d}\epsilon }=\alpha \text{MGb}\frac{\mathrm{d}\sqrt{{\rho }_{\mathrm{G}}+{\rho }_{\mathrm{S}}}}{\mathrm{d}\epsilon }$$ (8)

where ε is the applied strain; ${\rho }_{\mathrm{G}}$ is the density of GNDs; ${\rho }_{\mathrm{S}}$ is the density of statistically stored dislocations (SSDs); α is a constant; M is the Taylor factor; G is the shear modulus; b is the magnitude of Burgers vector.

According to Eq. 8, the key to improve the strain hardening rate is to increase the dislocation accumulation rate. However, for metals with heterogeneous microstructure, the contribution of back stress to work hardening much more that of dislocation accumulation [12]. Back stress hardening is caused by the pile-ups of GNDs, which are necessary to accommodate strain gradient near the interfaces during strain partitioning occurs between the hard and soft domains. The development of the long-range back stress can hinder dislocation slip in the tensile direction, while facilitating dislocation slip in the opposite direction, effectively increasing the ductility [141].

Based on the above theory, in order to achieve the highest HDI strengthening and HDI strain hardening, the soft zone should be fully constrained by the matrix of the hard zone, so it is crucial to control the proportion of the soft and hard zone. At present, it has been confirmed by the computer simulation method that the plastic flow of magnesium alloys with BGS distribution depends on the volume concentration ratio of fine grains to coarse grains [142]. Currently, experimental research on the effect of BGS ratio on mechanical properties is gradually increasing. The Mg-1.2Zn-0.1Ca (wt.%) alloy with different volume fractions of BGS was obtained by low temperature extrusion and annealing at 250 °C for different times. It was found that a short 10-min annealing sample showed superior YS of 305 MPa and good elongation of 20% [138]. Similarly, the AZ91 alloy with different volume fractions of BGS was obtained by controlling the APE time, and it was found that the sample with the proportion of 26.8% in the fine-grained layer exhibited the excellent combination of strength and ductility [57]. He et al. [143] controlled the volume fractions of fine grains and coarse grain through combining process of hot forging, extrusion and annealing, and found that the bimodal grain size distribution had little effect on the strength, and as the coarse grain volume fraction increased from 0.65% to 36.40%, the elongation of the alloy increased from 12.8% to 19.5% and then decreased to 10.4%, that is, the elongation reached the maximum value when the coarse grain volume fraction was 8%. However, in general, the effect of BGS volume fractions on the ductility of magnesium alloys needs to be further studied. Literature [137] suggests that the volume fraction of the hard zone should be high enough to adequately constrain the soft zone, so the optimal volume fraction of the soft zone needs to be well below 50%. On the other hand, when the volume fraction of the soft zone is greater than 50%, the soft zone will form a network, making the plastic strain path around the hard zone, resulting in the overall yield. In other words, the ability to embed hard zone heterostructures in the soft matrix to improve the overall yield strength is limited.

3.2. Tension-compression yield asymmetry

The strong tension-compression (T-C) yield asymmetry is one of the most important problems limiting the widespread application of deformed Mg alloys [144]. Therefore, it is imperative to mitigate or eliminate this asymmetry for the development of Mg alloys. It is generally believed that homogeneous structures with weaker textures are usually associated with low yield asymmetry [145]. However, recent studies have found that BGS has a significant positive effect on reducing the yield asymmetry. The as-extruded Mg-Gd-Zn-Zr alloy with BGS exhibited considerably lower T-C yield asymmetry, compared to the equiaxed Mg-Gd-Zn-Zr alloy (Fig. 18a) [130], and lower T-C yield asymmetry compared to most of the commercial ZK60 and AZ31 alloys [146,147].

Fig. 18.

(a) Engineering tensile and compressive stress-strain curves obtained in the BGS and equiaxed samples, (b) BSE-SEM micrographs of the sample with BGS compressed to plastic strains of 0.25 % and fracture [130].

The primary explanation for the lower T-C asymmetry in the BGS is attributed to different effects of the coarse and fine grains in the BGS on the T-C yield asymmetry [20,148]. Specifically, the basal slip activated in fine grains with weakened texture under tension before the yield points and the inhibition of extension twinning by fine grains under compression [133]. However, when the fine DRXed grains are unable to accommodate the plastic strain, the unDRXed grains will be forced to deform and the twining will occur (Fig. 18b) [130]. Furthermore, it has been observed in the literature [3,149] that alloys with BGS exhibited low but different yield asymmetries, suggesting that the influence of the BGS yield asymmetry may be associated with the ratio and size of unDRXed and DRXed grains. This relationship needs further investigation in future researches.

3.3. Superplasticity

The importance of BGS in promoting superplasticity has been identified. Lapovoke et al. [150] reported that a AZ31 alloy with BGS exhibited a higher total elongation in comparison with a material with uniformly fine grains. Additionally, a remarkable superplasticity was also observed in a ZK60 alloy with BGS that the tensile ductility at 220 °C was 2040% and 1400% for the strain rates of 3 × 10^–4 s^-1 and 3 × 10^–3 s^-1, respectively [151]. The improved superplastic deformation behavior was also noted in the AZ91 alloy with BGS, demonstrating high superplasticity with an elongation of approximately 580% during tensile tests at 300 °C and 1 × 10⁻³ s⁻¹ [152]. In addition to achieving superplasticity at low strain, high strain rate superplasticity of the AZ91 alloy with BGS was successfully obtained by rapidly solidified flaky powder metallurgy (RS-FP/M), resulting in fracture elongation of 465% at temperature of 350 °C and initial strain rate of 10^–2 s^-1 [29].

On the one hand, the prior occurrence of CDRX in coarse grains during hot tension leads to a dynamic balance between hardening from dislocation migration and softening from CDRX, accommodating significant deformation strain, and reducing the possibility of microcrack formation. On the other hand, the progressive transformation of coarse grains into nearly strain-free fine grains via CDRX enhances grain boundary sliding (GBS) activation. In addition, the prior consuming of local strain via CDRX in the coarse grains could decrease coarsening rate of fine equiaxed grains. Consequently, CDRX in coarse-grained regions and GBS in fine-grained regions collectively control superplastic deformation at early tension stage (Fig. 19) [152].

Fig. 19.

Microstructure evolution and deformation mechanisms of AZ91 alloy with BGS during superplastic deformation [152].

3.4. Corrosion resistance

The formation of Mg oxide on the surface of Mg alloys is facilitated by the lower chemical potential of MgO compared to Mg. Upon exposure to a corrosive environment, the cubic Mg oxide transforms hexagonal Mg hydroxide. The compressive stress arising between the corrosion product layer and the oxide layer can lead to the fracture of the former and a subsequent decrease in corrosion resistance [153]. Current studies indicate that the corrosion resistance of Mg alloys can be influenced by factors such as grain size, texture, and dislocation density [[154], [155], [156]]. The fine grain can stabilize the corrosion product layer by increasing the grain boundary and releasing the compressive stress between the corrosion product layer and oxide layer, thus improving the corrosion resistance. Texture also plays a role in the corrosion resistance of Mg alloys, with the closely packed basal plane exhibiting greater corrosion resistance due to its higher atomic density, resulting in higher binding energy, lower surface energy, and higher activation energy for removal of atoms from the metal lattice [157]. Consequently, the corrosion resistance should be reduced as follows: basal plane, pyramidal plane and prismatic plane. Therefore, the DRXed region with weak basal texture is expected to be more easily to corrosion than the unDRXed region with strong texture (Fig. 20) [156]. It is worth noting that dislocation density can also play a role in the corrosion behavior of deformed Mg alloys, as regions with high dislocation density typically exhibit higher anodic dissolution rates, rendering them more electrochemical active in the corrosion process [155]. In this sense, the unDRXed grains are expected to be more susceptible to corrosion than the DRXed grains. Overall, the combined influence of grain size, texture and dislocation density should be taken into account when considering the corrosion behavior of Mg alloys with BGS.

Fig. 20.

Optical micrographs of the ECAPed alloy (a) before and (b) after the local potentiodynamic polarization test in 3.5 wt.% NaCl using the electrochemical microcell technique. (c) Local potentiodynamic polarization curves of the DRXed and unDRXed regions in the ECAPed alloy. SEM micrographs of the (d) unDRXed region and (e) the DRXed region after the local electrochemical test and after removal of corrosion products [156].

3.5. Other properties

The above provides an overview of the impact of BGS on mechanical and corrosion properties. In fact, BGS also has a certain impact on other properties, such as fatigue properties, fracture toughness, damping capacity, etc.

The BGS can improve the fatigue properties of Mg alloy compared to the fully recrystallized structure alloy, particularly in low or high cycle fatigue regions, due to its higher YS [158,159]. The fully recrystallized structure leads to cyclic hardening-induced stress concentration until failure, while the BGS can relieve the stress concentration during cyclic loading, ensuring stable cyclic deformation and persistent cyclic softening at low and high strain amplitudes, respectively. In addition, the BGS effectively suppresses the generation of cracks induced by residual twinning in the equiaxed structure.

Ji et al. [124] designed a Mg-9Gd-4Y-1Zn-0.5Zr (wt.%) alloy with a BGS, with a YS of 430 MPa and a plane strain fracture toughness K_IC of 23.1 MPa m^1/2. The activation of non-basal 〈a〉 dislocations and 〈c + a〉 dislocations in DRXed grain dissipates large irreversible plastic energy, reducing the driving force for local crack initiation and hindering aggregation of microvoids. The curved crack propagation path generated by unDRXed grains promotes extrinsic toughening. In addition, the strong strain hardening ahead of crack tip, caused by the plastic deformation incompatibility between DRXed and unDRXed grains, can slow down the microvoid growth rate and result in a high K_IC [160]. The schematic diagram of the toughening mechanisms for the BGS during the fracture toughness test is shown in Fig. 21.

Fig. 21.

Schematic diagram of the toughening mechanisms for the BGS during the fracture toughness test [161].

Besides, the presence of unDRXed grains in BGS is most likely to balance the damping capacity and mechanical properties. The Mg-Zn-Y-Zr alloys containing unDRXed grains has been observed to a high damping capacity above 0.01 in the strain-amplitude independent damping region, which is indicative of a high damping material. Conversely, alloys with fully recrystallized grain exhibits low damping capacity [162]. This discrepancy is attributed to the facilitation of dislocation movement by coarse grains, with the damping capacity showing an increase as the proportion of unDRXed grains increases, ultimately leading to higher damping capacities [163].

4. Summary and outlook

This review discusses the recent developments in BGS Mg alloys from two perspectives: The variables influencing BGS and the influence of BGS on the mechanical properties of Mg alloys, as well as insights into the relevant mechanisms for performance enhancement. The construction of a BGS can be achieved by adjusting the alloy composition, the grain size, second phase distribution, and twinning in the initial material. Combined with the adjustment of processing parameters, that is, deformation temperature, deformation speed, cumulative deformation, etc., control over the BGS can be achieved. Furthermore, recent studies have demonstrated that BGS can improve the strength-ductility synergy of Mg alloys due to the respective roles of coarse/fine grains in BGS and their roles in adjusting local strain. In addition, bimodal structure also plays a certain role in T-C yield asymmetry, fatigue property, fracture toughness and corrosion resistance. In conclusion, BGS is expected to pave the way for the future development of high-performance Mg alloys.

Declaration of competing interest

The authors declare that they have no conflicts of interest in this work.

Biographies

Jing Zuo obtained her Bachelor's degree in materials science and engineering from Harbin Institute of Technology in 2021. She is currently pursuing a Ph.D. in Materials Science and Engineering at Harbin Institute of Technology. Her research interest focuses on the deformation mechanism in magnesium alloys.

Chao Xu(BRID-06650.00.11082) graduated from Harbin Institute of Technology in 2012 with a Ph.D. degree in materials science. He then became a postdoctoral fellow at Nagaoka University of Technology in Japan and served as an assistant professor in 2016. He joined Harbin Institute of Technology and served as a professor in 2018. His current research interests include the design, preparation, processing technology and deformation mechanism of high-performance magnesium alloys and their composites.

Lin Geng(BRID-08879.00.10338) graduated from Harbin Institute of Technology in 1990 with a Ph.D. degree in materials science. He started his research career at Harbin Institute of Technology and obtained the position of professor at the school of material science and engineering, Harbin Institute of Technology in 1995. His current research interests include intermetallic compound matrix composites, interface research of metal matrix composites, heat treatment process research of metal matrix composites, high temperature deformation of metal matrix composites, etc.