MATERIALS TRANSACTIONS
Online ISSN : 1347-5320
Print ISSN : 1345-9678
ISSN-L : 1345-9678
Microstructure of Materials
Microstructure and Mechanical Properties of Ultrafine-Grained Mg-9Al-1Zn Alloy Fabricated by Multi-Directional Forging at Room Temperature
Yojiro ObaChihiro WatanabeMasakazu KobayashiHiromi Miura
Author information
JOURNAL FREE ACCESS FULL-TEXT HTML

2025 Volume 66 Issue 9 Pages 1138-1142

Details
Abstract

Mg-9Al-1Zn (AZ91Mg) alloy was multi-directionally forged at room temperature (rMDFed) employing by small pass strains of Δε = 0.1 up to a cumulative strain of ΣΔε = 2.0 at maximum. Ultrafine-grained structure with an average grain size of 0.2 µm was obtained via subdivision mainly by multiple mechanical twinning. The dense mechanical twinning as well as small pass strains suppressed sharp basal texture evolution, which enabled rMDFing up to high cumulative strain regions, and hence evolution of homogeneous ultrafine-grained structure. Well-balanced mechanical properties of very high yield strength of 520 MPa, ultimate tensile strength of 550 MPa, hardness of 1.32 GPa with reasonable plastic strain to fracture of 5% were achieved. These superior mechanical properties were results of extremely high forging stress over 400 MPa with changing forging direction to induce multiple slips and multiple twinning, which brought extremely large strain hardening and grain orientation randomization.

1. Introduction

Magnesium (Mg) alloys are promising structural materials because of their lightest weight and the specific strength among the practical metals and alloys. For industrial applications, further improvement of the mechanical properties is one of the most important issues. To date, researchers have intensively carried out various new approaches, such as adding rare-earth elements [14] and grain refinement by severe plastic deformation (SPD) [517]. The former gives notable strengthening because of evolution of long-period stacking-ordered (LPSO) structure. For instance, a yield strength (YS) and an ultimate tensile strength (UTS) of 460 and 526 MPa, respectively, were reported in Mg-Ni-Y alloy [3]. In the latter case, ultrafine grains (UFGs) smaller than 1 µm produced by SPD bring an excellent balance of strength and ductility without any addition of costly and strategic rare-earth elements [517]. Miura et al. reported that multi-directional forging at room temperature (rMDFing) of AZ80Mg alloy achieved an extremely high YS of 530 MPa, UTS of 650 MPa, and plastic strain to fracture of 9% [15]. Compared to the multi-directional forging under decreasing temperature conditions (dMDFing) employing large pass strains of Δε = 0.8 and a grain-refinement mechanism of dynamic recrystallization, rMDFing is accomplished by adoption of small pass strains of Δε = 0.1. This is effective to suppresses the evolution of sharp basal texture, in which basal {0001} poles are aligned parallel to the forging direction [14, 15, 18]. This dedicated procedure prevents cracking during repeated forging processes of Mg alloy and hence allows rMDFing up to high cumulative strains. Furthermore, rMDFing at ambient temperature involves extravagant work hardening and grain subdivisions by mechanical twinning till high cumulative strain regions, which are completely different behavior and operative mechanisms during dMDFing and conventional thermomechanical processing techniques. As the results, rMDFing can provide excellent mechanical properties because of various effects of grain refinement by mechanical twinning, grain orientation randomization by multiple twinning, extremely high work hardening, and activation of multiple slips due to notably high tensile stress [15, 18].

In the present study, rMDF was applied to a Mg-9Al-1Zn (AZ91Mg) alloy, which is one of the most common Mg alloys in a variety of fields, such as the automotive industry [19, 20]. In Mg alloys, the addition of Al generally promotes mechanical twinning by reducing stacking fault energy (SFE) [21, 22] because of decrease in the twinning stress [23]. Zn changes not only SFE but also the segregation state of Al [19, 22]. The changes in the microstructure and mechanical properties of AZ91Mg alloy rMDFed are compared with the previous results of rMDFed AZ80Mg alloy and dMDFed Mg alloys [10, 1315, 18].

2. Experimental Procedure

A commercial hot-extruded square bar of AZ91Mg alloy (Mg-9.2Al-0.86Zn-0.13Mn in mass%) with an initial grain size of 29 µm and the length of about 1 m along the extrusion axis (EA) was cut into a rectangular shape to have aspect ratio of 1.11:1.05:1.00 for rMDFing. Details of MDFing process are reported elsewhere [1315]. The combination of the aspect ratio and the small pass strain of Δε = 0.1 enables MDFing continuously without any reshaping. The small pass strain, which is much smaller than that of dMDFing, contributes to suppress the evolution of sharp basal texture. rMDFing was carried out on an Amsler universal mechanical testing machine at an initial strain rate of 3 × 10−3 s−1. The initial forging axis (FA) was parallel to the hot extrusion axis. The specimens were rotated 90° pass by pass. rMDFing was conducted up to a maximum cumulative strain of ΣΔε = 2.0 (i.e., 20 passes).

The evolved microstructure was observed on the planes parallel to EA or final FA using transmission electron microscopy (TEM) and orientation imaging microscope (OIM). The color decoding of the OIM maps was normal to the planes (crystal direction perpendicular to EA or final FA). On the same plane, change in the hardness during rMDFing was investigated using a micro-Vickers hardness tester. Tensile tests were conducted using specimens with gauge dimensions of 0.7 mm (thickness), 2.5 mm (width), and 5 mm (length) on an Instron-type mechanical testing machine at 298 K at an initial strain rate of 1.0 × 10−3 s−1. Tensile direction was perpendicular to the final FA.

3. Results and Discussions

Figure 1 shows true stress–cumulative strain curves during rMDFing of the AZ91Mg alloy. Each curve shows yielding followed by a large work hardening region. The maximum flow stress increased with increasing cumulative strain and appeared to approach a constant value at a high cumulative strain region. These features are similar to those of the other rMDFed Mg alloys [14, 15, 18]. The results in Fig. 1 confirmed that rMDFing of AZ91Mg alloy was successfully performed up to high cumulative strain of ΣΔε = 2.0. The almost constant value of the maximum flow stresses in the high cumulative strain region implies saturation of microstructural evolution, including grain refinement and dislocation accumulation. It is interesting to note that the maximum flow stresses at the 1st, 4th, 7th, 10th, 13th, 16th, and 19th forging passes were slightly higher than those at other passes (indicated by arrows in Fig. 1). This can be explained by the basal texture evolved in the as-hot-extruded bar, which persistently remained to cause so-called texture hardening until the high cumulative strain region [14, 15, 18]. The slight increment of the maximum flow stress at each three passes decreased gradually as the cumulative strain increased, which would suggest weakening of the basal texture by rMDFing.

Fig. 1

True stress versus cumulative strain curves during rMDFing of AZ91Mg alloy. Arrows indicate the passes having the slightly higher maximum flow stress compared to that of the other passes. (online color)

Change in the hardness of the AZ91Mg alloy during rMDFing is displayed in Fig. 2. The hardness increased rapidly at the low cumulative strain region, and then little by little in the medium and high cumulative strain regions. The obviously large strain hardening rate at the low cumulative strains should be affected by the dynamic Hall-Petch hardening effect induced by mechanical twinning and grain refinement during rMDFing [24]. This tendency is similar to the cumulative strain dependence of the maximum flow stress shown in Fig. 1. The highest value of 1.32 GPa at ΣΔε = 2.0 is remarkably higher compared with that of the commercial AZ91Mg alloy before rMDFing, and even slightly higher than those of the high-strength Mg alloys; age-hardened Mg-Zn-Gd alloy (1.25 GPa) [2] and rMDFed AZ80Mg alloy (1.26 GPa) [15]. The marvelously high hardness should be induced by extremely high dislocation density as well as development of UFGed structure. Distribution of precipitates can also contribute to the hardness via pinning of grain boundaries, accumulation of dislocation, stress concentration, and suppression of twinning [15, 18, 19, 2527] and will be described below with Figs. 3, 4. The high forging stress over 400 MPa and change in the forging direction pass by pass, where the latter contributes to activate even various non-basal systems, caused multiple slips to derive notable work hardening [15, 18].

Fig. 2

Chang in the Vickers hardness as a function of cumulative strain during rMDFing of AZ91Mg alloy.

Fig. 3

OIM maps of rMDFed AZ91Mg alloy with (a) ΣΔε = 0 (as-hot-extruded material), (b) ΣΔε = 0.3, and (c) ΣΔε = 0.5, respectively. (d), (e), (f) and (g), (h), (i) are the corresponding misorientation distributions and inverse pole figures of rMDFed AZ91Mg alloy with (a) ΣΔε = 0 (as-hot-extruded material), (b) ΣΔε = 0.3, and (c) ΣΔε = 0.5, respectively. White, black and purple lines in (a), (b), and (c) indicate boundaries with misorientation angles θ < 3°, θ ≥ 3°, and twin boundaries, respectively. The black-contrasted areas observed in (a) and (b) are mainly of β phase particles. The arrow outside of the photographs denotes either extrusion axis (EA) or forging axis (FA). (online color)

Fig. 4

Microstructures of AZ91Mg alloy (a) before rMDFing and (b) after rMDFing to a cumulative strain of 2.0. The arrow outside of the photographs denotes either extrusion axis (EA) or forging axis (FA). Inset in (b) is a selected-area diffraction pattern obtained from the area indicated by a circle in (b). The pattern indicates the presence of three grains (M, T, and N). While the misorientation angle between the grains M and T corresponds to the relationship of $\{ 10\bar{1}1\} $ contraction twin, the grain N has no typical primary twin relationships with respect to the grains M and T.

Microstructural evolution during rMDFing was examined by OIM. Figure 3 shows the OIM maps, corresponding misorientation distributions, and inverse pole figures. The as-hot-extruded sample is comprising coarse equi-axed grains with coarse twins, which can be indicated by the peaks at approximately 86° in the misorientation distributions and coincided with the primary $\{ 10\bar{1}2\} $ twins [28, 29]. It is apparent from the evolution of fine acicular grains in Fig. 3 that grain subdivision was proceeded mainly by mechanical twinning during rMDFing. The small peaks at around the misorientation angle of 60° are the higher order twins [13]. Furthermore, the peak at a low misorientation angle appeared during rMDFing is related with kinks, subboundaries, and low-angle boundaries [2831]. Accumulation of dislocations on the boundaries during SPD also induces change in the grain-boundary misorientation angles in addition to grain fragmentation by low-temperature dynamic recrystallization [32]. Grain subdivision was, therefore, carried out by combined mechanisms mainly of multiple twinning, kinking and low-temperature dynamic recrystallization. These combined mechanisms also promoted the misorientation randomization as seen in Fig. 3(f) similar to the Mackenzie distribution [33]. In this way, the grain size decreased gradually with increasing cumulative strain. Even while the area observed by OIM was quite small, destruction of the sharp basal texture in the as-hot-extruded sample during rMDFing was distinct by the decrease of the intensity at {0001} in the inverse pole figures. These results were qualitatively matched to the previous studies which reported the high intensity at {0001} in the as-hot-extruded alloy [18, 28] and reduction by MDFing [15, 18]. This was one of the reasons that MDFing to high cumulative strain regions at room temperature could be accomplished. That is, the employment of small pass strains of Δε = 0.1 effectively suppressed the evolution of sharp basal texture and contributed to orientation randomization, which enabled rMDFing to high cumulative regions.

The microstructures evolved before and after rMDFing to ΣΔε = 2.0 were investigated by TEM (Fig. 4). The initial microstructure contained fine precipitates smaller than a few tens of nanometers. In the AZ series of Mg alloys (Mg-Al-Zn-Mn systems), it is known that Mg17Al12 (β phase) and Al8Mn5 are finely precipitated [19, 25, 34]. Although previous studies reported that the fine precipitates hinder mechanical twinning [26, 27], acicular-UFGed microstructure mainly composed of mechanical twins was developed homogeneously by rMDFing to ΣΔε = 2.0 (Fig. 4(b)). These microstructural characteristics were same as those in the other rMDFed Mg alloys [14, 15, 18]. The average grain size decreased to 0.2 µm, which was also comparable to the other rMDFed Mg alloys [14, 15, 18]. The dense twinning must be possible under extraordinary high forging stresses over 400 MPa (Fig. 1), which is much higher than the various twinning stresses [35]. Changing FA at each pass also facilitates mechanical twinning by changing the shear direction to generate various types and variants of twins [15, 18]. The misorientation angles with respect to the straight boundaries did not always match the relationships for typical $\{ 10\bar{1}1\} $, $\{ 11\bar{2}2\} $, and $\{ 10\bar{1}2\} $ mechanical twins. This confirmed the change in the grain-boundary misorientation angles by the accumulation of dislocations [32]. Multiple twinning can also take place and weaken the crystallographical relationships of primary twins [18].

Tensile tests of the rMDFed samples were carried out at room temperature. The attained flow curves were exhibited in Fig. 5(a). For comparison, the result of the as-hot-extruded one tensile tested parallel to EA is also shown. It is evident in Fig. 5(a) that both YS and UTS drastically increased with increasing cumulative strain. The YS of 110 MPa and UTS of 260 MPa of the as-hot-extruded sample increased remarkably up to 520 and 550 MPa at ΣΔε = 2.0. It is also interesting to note that the increase in the YS and UTS appears more gradual at the high cumulative strain region similar to the changes in the maximum flow stress and hardness during rMDFing (see Figs. 1 and 2), suggesting upper limit of modification of the mechanical properties. Unclear tendency of the plastic strain to fracture εf as a function of cumulative strain can be caused by the distribution of β phase particles (see Fig. 3), which promote cracking to reduce ductility [36]. According to the previous studies [19, 36], the increase in cumulative strains brings high dislocation density and stress concentration at the interface between the matrix and the β phase during deformation. This probably causes easier initiation of cracks in the rMDFed AZ91Mg alloys than in the as-hot-extruded alloy. On the other hand, the β phase particles can be fragmented [34, 36] and thus the distribution of the β phase particles can be changed during rMDFing. The unclear behavior of εf should result from these complex factors. Nevertheless, well-balanced mechanical properties of YS = 520 MPa, UTS = 550 MPa and εf of 5% could be achieved at ΣΔε = 2.0. It is remarkable in Fig. 5 that plastic strain to fracture εf of 5% is still preserved even after straining by rMDFing to ΣΔε = 2.0 and it tended to increase with cumulative strain. This should be induced by the occurrence of multiple slips due to extraordinarily high tensile strength over 500 MPa [15]. These mechanical properties were compared with several representative rare-earth added Mg alloys (Mg-RE) [14] and UFGed Mg alloys [517] in Fig. 5(b). The values attained in the present rMDFed AZ91Mg alloys are comparable to those of the previous high-strength Mg alloys; Mg-Ni-Y alloy (YS = 460 MPa, UTS = 526 MPa, and εf = 8%) [3] and rMDFed AZ80Mg alloy (YS = 530 MPa, UTS = 650 MPa, and εf = 9%) [15]. Figure 5(b) also indicates that rMDFing provides significantly higher UTS than the other SPD methods. The UFGed microstructures developed by rMDFing effectively strengthened Mg alloys. The slightly lowered mechanical properties of the rMDFed AZ91Mg alloy compared with those of the rMDFed AZ80Mg alloy should be influenced by the distribution of β phase particles, as already mentioned above. Even so, much higher YS and UTS was attained at much lower cumulative strain region by rMDFing compared with those necessary for MDFing under decreasing temperature conditions (ΣΔε ≥ 6.4) [10, 13].

Fig. 5

Results of tensile test of the MDFed AZ91Mg alloys; (a) true stress vs. nominal strain curves, and (b) UTS vs. elongation εf of the rMDFed AZ91Mg alloy at ΣΔε = 2.0 as well as representative rare-earth added Mg alloys (Mg-RE) [14] and UFGed Mg alloys [517]. Filled circles, grey diamonds, and open triangles denote the values of rMDFed Mg alloys, Mg-RE, and the other UFGed Mg alloys, respectively.

4. Conclusion

Hot-extruded AZ91Mg alloy was successfully multi-directionally forged at room temperature (rMDFed) up to a cumulative strain of 2.0 at maximum. Mechanical twinning mainly contributed to subdividing the coarse initial grains into ultrafine-grained structure with an average grain size of 0.2 µm. Employment of small pass strains of Δε = 0.1 enabled rMDFing to high cumulative regions by the suppression of basal texture evolution and grain-orientation randomization. Notably high forging stress during rMDFing induced large work hardening by multiple slips and multiple twinning even under the presence of dense precipitation. The rMDFed AZ91Mg alloy exhibited quite high hardness of 1.32 GPa as well as high yield strength and ultimate tensile strength of 520 MPa and 550 MPa. Even while after rMDFing (i.e., one of the SPD methods) to high cumulative strain region, reasonable plastic strain to fracture of 5% could be attained. Hence, well-balanced mechanical properties could be achieved.

Acknowledgments

The authors thank Mr. T. Shibazaki for the help with the experiment.

REFERENCES
 
© 2025 The Japan Institute of Metals and Materials
feedback
Top