Tailoring strength-ductility balance of caliber-rolled AZ31 Mg alloy through subsequent annealing

Tin Kong, Byung J Kwk,, Jonghyun Kim, Jong Hun L, Sung Hyuk Prk,Ji Hoon Kim, Young Hoon Moon, Hyun Sik Yoon, Tkyung L,?

a School of Mechanical Engineering, Pusan National University, Busan 46241, Republic of Korea

b Advanced Forming Process R&D Group, Korea Institute of Industrial Technology, Ulsan 44413, Republic of Korea

c College of Materials Science and Engineering, Chongqing University, Chongqing 400044, China

d School of Materials Science and Engineering, Kyungpook National University, Daegu 41566, Republic of Korea

e Department of Naval Architecture and Ocean Engineering, Pusan National University, Busan 46241, Republic of Korea

Abstract Recently,multi-pass caliber rolling has been shown to be effective for Mg alloys.This study investigated the effect of subsequent annealing on the mechanical properties of a caliber-rolled AZ31 Mg alloy to modulate the strength-ductility relationship. This annealing gave rise to different trends in mechanical properties depending on the temperature regime. Low-temperature annealing (T≤473K) exhibited a typical trade-off relationship, where an increase in annealing temperature resulted in increased ductility but decreased strength and hardness. Such a heat treatment did not degrade the high strength-ductility balance of the caliber-rolled alloy, suggesting that the mechanical properties could be tailored for different potential applications. In contrast, high-temperature annealing (T > 473K) caused a simultaneous deterioration in strength, hardness, and ductility with increasing annealing temperature. These differences are discussed in terms of the varying microstructural features under the different investigated annealing regimes.

Keywords: AZ31 Mg alloy; Caliber rolling; Annealing; Grain growth; Mechanical improvement; Twinning.

1. Introduction

Increasing attention towards the environment and energy saving has given rise to active studies on weight reduction and lightweight materials. In this trend, Mg alloys are currently attracting significant attention from the automotive,aerospace,and electronic industries[1].For example,62%weight saving was achieved by adopting Mg sheet metal in body panel applications [2]. However, the low strength of Mg alloys is one of the issues hindering their wider applications. Researchers have suggested inducing an ultrafine-grained (UFG) structure composed of submicrocrystalline grains to increase the mechanical strength of Mg alloys [3]. This led to the development of various severe plastic deformation (SPD) processes in recent decades, such as high-pressure torsion (HPT)[4], equal-channel angular pressing (ECAP) [5], accumulative back extrusion (ABE) [6], high-ratio differential speed rolling(HRDSR) [7], and accumulative roll-bonding (ARB) [8].

Despite effective strengthening, HPT, ECAP, and ABE are limited to lab-scale fabrication owing to their sample dimensions being of centimeter-order as well as difficulty in mass production. HRDSR and ARB are sheet metal forming processes; the latter also has the issue of non-continuous manufacturing. In other words, there is strong demand for a rod manufacturing process that enables (i) UFG structure, (ii)bulk dimension, and (iii) mass production. Multi-pass caliber rolling has attracted attention from this viewpoint due to its capabilities of effective grain refinement and continuous mass production of steels [9,10], Ti alloys [11–13], and even high-entropy alloys [14].

Fig. 1. Graphical scheme of caliber rolling employed in the present study. Red line indicates the basal pole aligned with TD before the rolling. Circular figures indicate the cross-section of caliber-rolled alloy before (dotted line) and after (solid line) the deformation for each pass. Numbers in the circular figures indicate the number of applied caliber-rolling pass.

In contrast to those materials, a relatively limited number of studies have investigated caliber-rolled Mg alloys. Inoue et al.[15]and Mukai et al.[16]first reported the potential of caliber rolling for Mg alloys in enhancing mechanical strength and hardness. Doiphode et al. investigated the deformation behavior of a caliber-rolled AZ31 Mg alloy at room temperature [17] and at elevated temperatures [18]. Our latest studies have verified that the advantages of caliber rolling apply to Mg alloys as well. One study [19] reported an enhanced strength-ductility balance and disappearance of yield asymmetry in a caliber-rolled ZK60 Mg alloy for the first time,and another study [20] investigated microstructural evolution and mechanical improvement with increasing caliber-rolling passes.

Despite a few previous studies, a considerable amount of essential information about caliber-rolled Mg alloys is still unknown because this research area is relatively new. One missing piece of information is the effect of subsequent annealing.Although a previous study [21] reported a grain growth tendency with respect to annealing,it investigated a narrow range of temperatures (i.e., 573–723K) and provided no data of tensile properties. Thus, the present study conducted a wider(373–673K and 10–360min) and more systematic investigation of the effects of subsequent annealing on the mechanical properties of caliber-rolled AZ31 Mg alloy. Interestingly, the properties did not show a monotonic tendency as expected from the previous work [21]. The mechanisms behind this peculiar trend were interpreted in terms of grain refinement and growth, texture hardening, and mechanical twinning.

2. Experimental procedures

A commercial hot-rolled AZ31 Mg plate with a thickness of 50mm was prepared for this study. The chemical composition of the plate was measured to be Mg-3.6Al-1.0Zn-0.5Mn(numbers represent mass percentage). This plate was homogenized at 673K for 10h and machined into a rod with a diameter of 26mm and total length of 280mm; the length consisted of a 130-mm head section with an inclined angle of 5° and a 150-mm body section. An image of this sample was presented in our previous work [19]. All examinations reported hereafter were performed only on the body section.

The AZ31 rod was soaked in a furnace at 673K for 1h and then subjected to caliber rolling in an ambient atmosphere without interval heating. The basal pole was set to be parallel to the transverse direction (TD) for the first-pass deformation(i.e., the basal plane parallel to the normal direction (ND)).The samples were rotated 90° clockwise around the rolling direction (RD) for each of seven deformation passes, which imposed a total area reduction of 87% (Fig. 1). This alloy is referred to as ‘CR’ hereafter.

The CR rod was cut to pieces of 120mm in length,which were subjected to subsequent annealing using an electric furnace followed by quenching in a water bath. The annealing was conducted at 373K, 423K, 473K, 573K,and 673K, respectively. The applied annealing temperature varied with intervals of 10min, 30min, 1h, 3h, and 6h,respectively. Microstructure samples and tensile specimens were fabricated from these 120-mm long annealed materials.

The microstructural characteristics of the annealed alloys were investigated using scanning electron microscopy (SEM)and electron backscatter diffraction (EBSD) technique. In addition, microstructural evolution after tensile fracture was observed using EBSD technique; the samples each obtained at a point 5mm away from fracture surface in this case. All samples were mechanically polished using#600 to#1200-grit SiC paper, 3-μm to 1-μm alcohol-based diamond suspension, and 0.04-μm colloidal silica in that order. For the SEM observation, they were further processed with chemical etching in a solution of 4g picric acid, 10ml acetic acid, 10ml distilled water, and 70ml ethanol. The obtained EBSD data were analyzed using the TexSEM Laboratories orientation imaging microscopy software. Only the data with confidence indexes higher than 0.1 were employed to provide a reliable analysis.The step size varied from 0.05μm to 0.75μm depending on the grain size.

Fig. 2. SEM micrographs of the investigated alloys. ‘T’ and ‘t’ indicate the annealing temperature and time, respectively. Note that the alloys annealed at 673K use a different scale owing to larger grains.

Mechanical evaluation of the investigated alloys was conducted with a uniaxial tensile test in an ambient atmosphere.Tensile specimens were machined based on the ASTM-E8 standard with a gage length, width, and thickness of 25mm,6mm, and 1.5mm, respectively. The tests were conducted at a strain rate of 5×10?3s?1using a laser extensometer to obtain precise data. The Vickers hardness test was performed under a load of 2 kgf. The tests were repeated five times per annealing condition.

3. Results

The microstructures of the investigated AZ31 Mg alloys varied significantly depending on the annealing condition(Fig. 2). The CR alloy had a UFG structure with an average grain size of 0.9μm. This effective grain refinement by caliber rolling resulted from (i) significant redundant strain generated in the complex three-dimensional stress states and (ii)strain accumulation at the center of the material cross-section through the oval/circular-shaped calibers. The microstructural and mechanical characteristics of this CR alloy have recently been discussed in more detail elsewhere [20].

The subsequent annealing induced a grain growth, the extent of which was proportional to the annealing temperature and time (Fig. 3). The deviation in grain size at a given temperature ranged from 0.2μm (at 373K) to 5.6μm (at 673K).In contrast, the deviation radically increased to 4.2–9.6μm when the annealing time was fixed. These results indicate a primary contribution of annealing temperature to the microstructural evolution of annealed alloys, which is consistent with the literature [22]. As a result, microstructures are interpreted in terms of annealing temperature rather than annealing time hereafter.

Fig. 3. Variation in grain size for the investigated alloys with respect to annealing temperature and time.

Fig. 4. Inverse pole figure maps and grain-size distributions of the alloys representing Group I and Group II, respectively: (a) annealed at 373K for 10min and (b) 673K for 10min.

Fig. 5. (0002) pole figures of the investigated alloys: (a) CR alloy and alloys annealed at (b) 373K, (c) 473K, (d) 573K, and (e) 673K for 10min.

The annealed AZ31 Mg alloys were classified into two groups for further discussion. Group I contain the samples annealed at low temperatures (i.e., 373–473K), excluding the sample annealed at 473K for 6h. Group II includes the samples annealed at high temperatures (i.e., 573–673K) as well as the aforementioned sample. The grains were uniform in size for Group I samples, while the microstructures became inhomogeneous for Group II samples, as shown in Fig. 2.Such a difference gave rise to different shapes in a grain-size distribution (Fig. 4). For example, the alloy annealed at 373K exhibited the single peak around a grain size of 1μm in the diagram. This is contrasted by the two peaks around 5μm and 12μm for the alloy annealed at 673K. In other words,Group II samples had a bimodal grain structure composed of fine-grained (FG) and coarse-grained (CG) regions. The similar phenomenon was reported in AZ31 Mg alloys subjected to FSP [23] and severe cold drawing [24].

The values of grain orientation spread (GOS) of the CR and annealed alloys were calculated through EBSD analysis.In the CR alloy,most of the investigated area(88.2%)showed a low GOS value of less than 1°. The subsequent annealing rarely changed this trend; the area fractions with GOS values less than 1° were 85.6%, 87.5%, 91.2%, and 86.3% for the samples annealed at 373K, 473K, 573K, and 673K, respectively. GOS value is defined as the deviation in orientation for a given location with respect to an average orientation value.In other words, the high frequency of low GOS values for the investigated alloys indicates a high fraction of recrystallized grains possessing a low strain energy [25].

The CR alloy showed a basal texture where the basal poles were tilted from ND to TD, while the basal plane remained parallel to RD (Fig. 5). The ND and TD in the pole figure map correspond to those presented in Fig. 1. The current data are consistent with the texture development reported in other caliber-rolled Mg alloys [19,20], implying a primary contribution of the process rather than the alloying system. The subsequent annealing scarcely induced a meaningful change from the initial texture of the CR alloy, except for an increase in texture intensity with increasing annealing temperature.

The CR alloy exhibited a yield strength (YS) of 313MPa,ultimate tensile strength (UTS) of 336MPa, uniform elongation (UE) of 13%, and elongation to failure (EL) of 20%.The hardness of the CR alloy was measured to be 83 HV.These results indicate only a small deviation compared with the data of a caliber-rolled AZ31 Mg alloy measured elsewhere (YS of 298MPa, UTS of 378MPa, UE of 13%,and EL of 21% [20]), which strengthens the data’s reliability.The tensile properties indicate simultaneous improvement of strength and ductility throughout caliber rolling in comparison to a commercial AZ31 alloy (UTS of 188MPa and EL of 17% [17]). This improvement was attributed to the caliber rolling rather than the material because similar results were recently reported with a caliber-rolled ZK60 Mg alloy[19].

The subsequent annealing gave rise to distinguished differences in mechanical properties depending on the heattreatment condition. As similar to the discussion with the microstructure data above, annealing temperature also had a higher contribution to tensile properties than annealing time.The deviations of tensile properties between the minimum and maximum annealing times (i.e., 10min and 6h) were limited to 35MPa for YS, 63MPa for UTS, 3% for UE, and 6% for EL for all samples. These ranges are notably narrower than the deviation depending on annealing temperature: 50MPa,125MPa, 4%, and 10%, respectively. Consequently, the mechanical properties were presented as a function of temperature to clarify the trends (Fig. 6). An increase in annealing temperature gave rise to a monotonic decrease in mechanical strength (i.e., YS and UTS) as well as hardness. In contrast,ductility did not follow such a monotonic trend. Both UE and EL increased up to an annealing temperature of 473K and then decreased at higher temperatures.

Fig. 6. Variation in mechanical properties depending on annealing temperature: (a) strength, (b) hardness, and (c) ductility. The horizontal lines indicate the mechanical properties of the CR alloy.

Fig. 7. Strength-ductility balance represented by a YS×EL diagram. The dotted lines indicate the strength-ductility balance corresponding to the given YS×EL values.

The comprehensive mechanical performances of the investigated alloys were evaluated in terms of YS×EL (Fig. 7).The diagram clearly shows a tendency of tensile properties changing with annealing temperature. Increasing the annealing temperature improved the ductility but reduced strength for Group I samples. It is worth noting that the Group I data did not markedly deviate from the red dotted line despite the trade-off relationship between strength and ductility, suggesting that the enhanced YS×EL value of CR alloy (i.e.,6310 MPa%) was preserved after subsequent annealing. On the other hand, Group II samples presented a simultaneous reduction of strength and ductility with an increase in annealing temperature, resulting in the YS×EL values decreasing. Therefore, it is possible to tailor the tensile properties of a caliber-rolled AZ31 Mg alloy through low-temperature annealing at ≤473K (i.e., Group I). Higher annealing temperatures (i.e., Group II) induce loss of the strength-ductility balance, and thus they are not recommended.

Both strength and hardness showed similar trends with respect to grain size (Fig. 8). Group I and Group II showed an obvious domain in these diagrams as well. The YS results of the investigated alloys followed the Hall-Petch relationship, expressed as YS=σ0+kd?0.5, where σ0(= 133MPa)is a friction stress, k (= 153 MPa μm0.5) is a Hall-Petch coefficient, and d is grain size. The friction stress of the investigated alloys is higher than those of AZ31 Mg alloys subjected to ECAP (85MPa [26]) and friction stir processing (FSP) (24MPa [27]). Such results are further discussed in Section 4.2 in terms of texture development. The hardness data had a similar form relationship: HV=H0+KHd?0.5[21],where H0(= 47 HV) and KH(= 30 HV μm0.5) correspond to friction stress and Hall-Petch coefficient of the Hall-Petch relationship, respectively.

Fig. 8. Variation in (a) YS and (b) hardness with respect to the inverse root of grain size for the investigated alloys.

Fig. 9. Inverse pole figure maps and image quality maps in the alloy annealed at (a) 573K and (b) 673K for 1h after tensile fracture; (c) enlarged images presenting the types of generated twins.

Microstructural characterization of Group II samples after tensile fracture provided deeper insight into the annealing effect with the bimodal grain structure (Fig. 9). The alloy annealed at 573K for 1h consisted of 76% FG region(d=4.6μm) and 24% CG region (d=14.5μm). Mechanical twinning mainly occurred in the CG region (i.e., the proportion of twin boundaries ftb=6.5%), whereas it was significantly suppressed in the FG region (ftb=1.5%). The increase in annealing temperature from 573K to 673K did not change the bimodal ratio (i.e., 76% for FG and 24% for CG), but it did increase the overall grain sizes: 12.9μm and 40.8μm for FG and CG regions, respectively. The frequency of mechanical twinning considerably increased in this alloy.In particular, mechanical twinning occurred in both the CG region (ftb=12%) and FG region (ftb=4.4%). Analysis of misorientation angles revealed the type of observed twins as {10?11} contraction twins, {10?12} extension twins in the contraction twin band, and {10?11}-{10?12} double twins. Such twin types are typical in a wrought Mg alloy whose basal plane is parallel to the tensile direction[28,29].

Fig. 10. Relationship between number fraction of twin boundaries and average grain size of fractured tensile specimens.

The proportion of twin boundaries was arranged as a function of their grain sizes (Fig. 10). The homogeneous microstructure yielded one datum per sample in Group I,whereas the bimodal structure of samples in Group II provided two data, one for each region. It is interesting to note that the data indicates a linear relationship between the twin fraction and grain size. This proposes a primary role of grain refinement in the suppression of mechanical twinning in the investigated AZ31 Mg alloys.

4. Discussion

4.1. Effects of grain refinement and growth

Group I samples had a trade-off between strength and ductility without loss of their balance (i.e., the YS×EL value).Such a strength-ductility relationship is typical for FG/UFG metals [30]. Grain growth reduces the proportion of grain boundaries acting as barriers against dislocation movement,weakening the grain-boundary strengthening.This mechanism is applicable to the present annealed alloys by considering the high consistency with the Hall-Petch relationship, as shown in Fig. 8. For example, the alloys annealed at the lowest temperature (i.e., 373K) exhibited the similar strength, hardness,and ductility as compared with CR alloy.This is understood in terms of the similar grain sizes among these samples. Meanwhile, dislocations were nearly saturated in the UFG structure, thereby decreasing strain hardening [31]. The small gap between YS and UTS (i.e., 23MPa) of the CR alloy supports this deduction. The grain growth through subsequent annealing recovered a strain-hardening capability, of which the extent increased with increasing annealing temperature.These factors rationalize the variation in strength and ductility maintaining the comparable YS×EL values for Group I samples.

Doiphode et al. [17] also applied caliber rolling to AZ31 Mg alloys at 523–723K, which showed YS×EL values ranging from 2730 MPa% to 4262 MPa%; the data are presented in Fig.7.However,the YS×EL value of Group I was remarkably improved (i.e., 1.5–2.5 times higher) compared to values reported in the literature. The difference in tensile properties between the two categories of caliber-rolled AZ31 Mg alloys stemmed from the efficiency of grain refinement. The CR alloy fabricated in this work had a UFG structure(d=0.91μm),as mentioned above. In contrast, the previous work [17] developed relatively coarse grains with grain size ranging from 3.6μm to 12.5μm depending on the rolling temperature.Oval/circular-shaped calibers employed in this work have a higher efficiency of strain accumulation compared to the typical shape of calibers (i.e., diamond cross-section), leading to more effective grain refinement in this study [32].

Regarding Group II samples, it is worthwhile noting the high consistency of them with the Hall-Petch relationship in Fig. 8. This implies that the strength reduction by grain growth confirmed with Group I samples is also available for the Group II samples. Such a deduction explains the monotonic decrease in YS, UTS, and hardness with increasing annealing temperature, as shown in Fig. 6a and b. Consequently, the deviation of Group II from the red dotted line in Fig. 7 must have resulted from degraded ductility rather than strength and hardness.This is further discussed in Section 4.3.

Hardness showed a negative correlation with annealing temperature, as shown in Fig. 6b. The caliber rolling gave rise to a strong dynamic recrystallization to accommodate high rolling temperature and large deformation [19], which released the strain energy in grains. Thermal energy provided by the subsequent annealing was insufficient to cause an additional static recrystallization, because most grains had already been recrystallized during caliber rolling. However,the supplied thermal energy increased the mobility of lattice defects, thereby leading to the recovery and resultant release in stored energy. This is further supported by kernel average misorientation (KAM) values of selected samples(Fig. 11). The KAM values decreased with an increase in either annealing temperature (Fig. 11a and b) or annealing time (Fig. 11b–d). The reduced KAM values imply the decreasing density of geometrically necessary dislocations[33], suggestive of active recovery under these annealing conditions. The residual strain energy of CR alloy is also reduced by the atomic diffusion at grain boundaries towards the increasing radius of grain curvatures during grain growth[34]. The reduced energy resulted in the decrease in hardness with increasing annealing temperature.

4.2. Effects of texture hardening

The entire AZ31 Mg alloys investigated in this work presented the strong basal texture with the basal plane being parallel to the tensile direction. Such a texture development is similar to those observed in extruded Mg alloys without rare-earth elements [35]. The strong basal texture inhibited the activation of basal slip, thereby increasing friction stress of an alloy. This phenomenon, referred to as texture hardening,led to similarly high friction stresses for the present AZ31 Mg alloys (133MPa) and extruded alloy (130MPa [36]). In contrast, the basal plane is no longer parallel to the tensile direction in Mg alloys subjected to conventional ECAP processing routes (e.g., Bcor C routes) and FSP [27]. For example,shear deformation induced by ECAP tilted the basal pole toward the extrusion direction [37]. The activation of non-basal slips alleviates the extent of grain-boundary strengthening under this condition. Consequently, texture hardening due to the strong basal texture resulted in the higher friction stress of the present AZ31 Mg alloys as compared with those subjected to the conventional ECAP and FSP.

Fig. 11. KAM maps in the alloys annealed at (a) 473K for 1h, (b) 473K for 6h, (c) 573K for 6h, and (d) 673K for 6h.

Table 1 Average SF values for basal slip under tension along the RD for the alloys annealed for 10min as well as CR alloy. Note that these annealing conditions correspond to those used for Fig. 5.

The intensity of basal texture increased with increasing annealing temperature, as shown in Fig. 5. Such a trend is in good agreement with the literature. Roostaei et al. [38] confirmed a positive correlation between the basal-texture intensity and annealing temperature(from 523K to 723K)in ARBprocessed AZ31 Mg alloys. Pérez-Prado and Ruano [39] also reported an increment in basal-texture intensity with increasing annealing temperature (from 723K to 793K) and annealing time (from 30min to 3h) in AZ61 alloy.

Nevertheless, texture hardening through the subsequent annealing was insufficient to induce further mechanical strengthening for the annealed alloys. This is supported by EBSD calculation of Schmid factors (SF) for basal slip under tension along the RD (Table 1). The calculation yielded similar SF values regardless of annealing temperatures, suggesting a considerably lower contribution of texture hardening in the annealed alloys as compared with the effect of grain growth.Recalling Fig. 3, a grain size rose a maximum 680% after 10-min annealing and 1150% after 6-h annealing. Such an increment is significantly higher than the increase in texture intensity through the annealing. Su et al. [37] investigated the effect of annealing conditions on tensile properties of AZ31 Mg alloys subjected to ECAP. The variation in tensile properties in the previous study is highly consistent with the trends in Fig.6;YS monotonically decreased with increasing annealing temperature, whereas EL exhibited the highest value at 473K.The caliber rolling and ECAP give rise to different texture developments, as aforementioned. The consistent trends in tensile properties, despite the different textures, strengthen our argument that the texture hardening had a negligible influence on tensile properties when annealing FG/UFG AZ31 Mg alloys.

4.3. Effects of mechanical twinning

The non-monotonic trends in ductility arose from the exclusive activation of mechanical twinning in Group II.Both mechanical twinning and dislocation slip follow the Hall-Petch relationship, of which the k value of twinning is significantly higher than that for slip [40]. This suggests that the resistance to mechanical twinning increases via grain refinement, and its rate is considerably higher than that of the dislocation slip. Such a difference in k value leads to the suppression of twinning in FG/UFG metals with grain sizes less than a critical value [41]. Li et al. [42] reported the reduction of twin frequency under compression of pure Mg with decreasing grain size to 4μm, which is consistent with the present data in Fig. 10. More importantly, grain refinement to UFG (d=0.8μm) and FG (d=2μm) structures resulted in an absence of twins in pure Mg. Those authors thus suggested the critical grain size for the activation of twinning to be ～2.7μm. Interestingly, this suggestion almost matches the results of this study; the critical grain size for the investigated alloys was measured to be 2.3μm in this work.Such a consistency supports our conclusion that the excessive annealing temperature allowed mechanical twinning in the Group II samples.

The basal planes parallel to the tensile direction gave rise to activation of {10?11} contraction twins, {10?12} extension twins in the contraction twin band, and {10?11}-{10?12} double twins. These types of twins are unfavorable for ductility because twin boundaries induce highly localized shear deformation and strain incompatibility [43]. Therefore,the formation of harmful twins degraded ductility, thereby reducing the YS×EL values of the Group II samples. Annealing at 673K induced more active formation of these twins than that at 573K, as shown in Figs. 9 and 10, resulting in the most significant deterioration in the strength-ductility balance shown in Fig. 7.

5. Conclusions

The present study revealed the effects of subsequent annealing on the mechanical properties of caliber-rolled AZ31 Mg alloy with a homogeneous UFG structure. The annealed alloys were classified into two groups that exhibited different microstructural and mechanical properties. In Group I (i.e.,alloys annealed at 373–473K), strength and hardness monotonically decreased with increasing annealing temperature,whereas ductility showed a monotonic increase. Accordingly,Group I samples preserved the enhanced strength-ductility balance of the initial CR alloy. In Group II (i.e., alloys annealed at 573–673K), all mechanical properties deteriorated with increasing annealing temperature, thereby degrading the strength-ductility balance. The reduced strength and hardness were interpreted in light of the Hall-Petch relation. In other words, grain refinement and growth mainly contributed to these mechanical properties. The decrease in ductility of Group II samples stemmed from the activation of mechanical twinning that induced a highly localized shear deformation and strain incompatibility. Texture hardening had a negligible influence on mechanical properties among the investigated alloys as compared with the other two factors.

Acknowledgment

This work was supported by the National Research Foundation of Korea(NRF)grant funded by the Korea government(MSIT) through GCRC-SOP (No. 2011-0030013).