Materials Science and Engineering A 528 (2011) 4055–4067

Contents lists available at ScienceDirect

Materials Science and Engineering A journal homepage: www.elsevier.com/locate/msea

Dynamic microstructural changes during hot extrusion and mechanical properties of a Mg–5.0 Zn–0.9 Y–0.16 Zr (wt.%) alloy S.W. Xu a,b , M.Y. Zheng a,∗ , S. Kamado b , K. Wu a , G.J. Wang c , X.Y. Lv c a b c

School of Materials Science and Engineering, Harbin Institute of Technology, Harbin 150001, PR China Department of Mechanical Engineering, Nagaoka University of Technology, Nagaoka 940-2188, Japan Northeast Light Alloy Company Limited, Harbin 150060, PR China

a r t i c l e

i n f o

Article history: Received 16 September 2010 Received in revised form 4 January 2011 Accepted 25 January 2011 Available online 1 February 2011 Keywords: Mg–Zn–Y–Zr alloy Quasicrystalline phase Hot extrusion Continuous dynamic recrystallization Twin Mechanical properties

a b s t r a c t In this study, firstly, dynamic microstructural changes of an as-cast Mg–5.0 Zn–0.9 Y–0.16 Zr (wt.%) alloy (designated ZWK510) during hot extrusion at 350 ◦ C and a ram speed of 3.33 mm s−1 was systematically investigated by electron backscattering diffraction (EBSD) analysis. The dynamic recrystallization (DRX) mechanism during hot extrusion was discussed. Then, the effect of microstructure and texture on the mechanical properties of the as-extruded alloy specimens at room temperature was discussed. The ascast ZWK510 alloy consists of a-Mg and quasicrystalline I-phase. During hot extrusion at 350 ◦ C, the main DRX mechanism is the continuous DRX near the original grain boundaries. The I-phase can accelerate the DRX behavior near these areas by obstructing the slip of dislocations. The deformation twins and massive blocky substructures formed in original grains can coordinate the DRX process near the original grain boundaries, however the DRX seldom occurs inside of these area. After further deformation, these deformation twins and massive blocky substructures are elongated along the material flow and become so-called unDRXed area, then a bimodal “necklace structure” composed of fine DRXed grains of about 2.1 ␮m and unrecrystallized coarse area is formed. The  extrudedZWK510 alloy shows a DRX ratio of about 58% and a typical basal fiber texture of (0 0 0 1) 1 0 1¯ 0 // extrusion direction (ED). In the matrix

DRXed area around the crushed eutectic I-phase a large number of fine I-phase precipitates are observed pinning at the newly formed DRXed grain boundaries. The 0.2% proof strength and the ultimate tensile strength of the extruded ZWK510 alloy specimen are 317 and 363 MPa, respectively, with an elongation to failure of 12%, which have been attributed to strong basal fiber texture, refined grain size as well as the existence of fine precipitates formed during the hot extrusion. © 2011 Elsevier B.V. All rights reserved.

1. Introduction The development of high strength wrought magnesium (Mg) alloys is strongly desired to reduce the weight of transportation vehicles due to the increasing demand to reduce carbon dioxide emissions in the transportation sector. On the other hand, few applications of wrought magnesium alloys have been explored because of a number of drawbacks, such as lack of formability at room temperature and the 0.2% proof strength of Mg alloys is relatively low compared to that of aluminum (Al) alloys [1]. In order to compete with Al alloy, the 0.2% proof strength of wrought Mg alloy needs to be improved to over 300 MPa with a tensile elongation of over 10%. The strength of Mg alloy is proved to be determined by the combined contributions of grain size refinement strengthening [2], solid solution strengthening plus dispersion and precipita-

∗ Corresponding author. Tel.: +86 451 86402291; fax: +86 451 86413922. E-mail address: [email protected] (M.Y. Zheng). 0921-5093/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.msea.2011.01.103

tion strengthening [3–7] and control of texture [8]. The Hall–Petch coefficient for Mg is ∼0.7 MPa m−1/2 [9], which is significantly higher than that of Al alloy [10]. Thus, grain refinement based on the understanding of dynamic recrystallization (DRX) mechanisms during thermomechanical processes makes a significant contribution to improving the strength of wrought Mg alloys. Recently, high strength wrought Mg alloys having 0.2% proof strength and ultimate tensile strengths of >300 MPa have been developed in Mg–Ca–Zn [11], Mg–Zn–Mn–Al [12], Mg–Gd–Y–Zn–Zr [13], Mg–Zn–Ag–Ca–Zr [14] and Mg–Zn–Ca–Zr [15] alloy systems using grain size refinement by hot extrusion. In these alloy systems stable secondary-phase particles are contained, so grain boundaries could be pinned with these dispersed or precipitated secondary-phase particles to inhibit grain growth, i.e., the mechanical properties of wrought Mg alloys with stable secondary-phase particles can be improved by microstructural control during thermomechanical processes. Wrought Mg–Zn–Y–Zr alloys, which consist of a thermally stable icosahedral quasicrystalline phase (Mg3 Zn6 Y, I phase) formed in situ as a second phase in the a-Mg matrix during solidifica-

4056

S.W. Xu et al. / Materials Science and Engineering A 528 (2011) 4055–4067

Fig. 1. Typical microstructures and XRD pattern of the as-cast ZWK510 alloy specimen. (a) SEM image, (b) TEM micrograph of eutectic I-phase (bright field image), (c) SADP from I-phase exhibiting 2-fold, mirror fold, 3-fold and 5 fold symmetry and (d) XRD pattern.

tion, have attracted much attention. Quasicrystals are isotropic and possess a specially ordered lattice structure called the quasiperiodic lattice structure [16]. Quasicrystal I-phase has high hardness, thermal stability, high corrosion resistance, low coefficient of friction, low interfacial energy, etc. [17]. Therefore, during hot deformation of Mg–Zn–Y–Zr alloy that contains I-phase, hard I-phase particles are stable against coarsening and can effectively obstruct the slip of dislocations [16–18]. Up to now, fine-grained Mg–Zn–Y–Zr alloys reinforced by quasicrystalline Iparticles have been successfully developed by thermomechanical processes such as hot extrusion [16,18,19], hot compression [21] and equal channel angular pressing (ECAP) [22–24], etc. In these studies, mechanical properties at room temperature [18–20,22], superplastic behaviors at high temperatures [16,22–24], precipitation behavior during thermo-mechanical treatment [18], effects of yttrium on the microstructure and mechanical properties of Mg–Zn–Y–Zr alloys [19], etc., have been investigated. However, up to now there have been limited studies addressing the dynamic microstructural changes during thermomechanical processes, which make it difficult to further improve the mechanical properties of Mg–Zn–Y–Zr alloys through microstructural control. In this study, firstly dynamic microstructural changes of an ascast Mg–5.0 Zn–0.9 Y–0.16 Zr (wt.%) alloy during hot extrusion at 350 ◦ C was systematically investigated by electron backscattering diffraction (EBSD) analysis. The DRX behavior during hot extrusion was discussed. Then, the effect of microstructure and texture on the mechanical properties of the as-extruded Mg–Zn–Y–Zr alloy specimens at room temperature was discussed.

2. Experimental procedures The studied alloy was prepared by electric melting of high purity Mg, Zn and the Mg–Zr, Mg–Y master alloys under a cover gas mixture of CO2 and SF6 in a steel crucible and casting them into a steel mold. The composition of the alloy was Mg–5.00 Zn–0. 92 Y–0.16 Zr (wt%). Hereafter, the alloy is designated as ZWK510. The existence of icosahedral quasicrystalline I-phase in the a-Mg matrix of as-cast alloy specimen was confirmed by transmission electron microscopy (TEM, Philips CM 12) operated at 120 kV. TEM specimens were prepared by twin jet electro-polishing. Then, the as-cast ingot was directly extruded at 350 ◦ C with an extrusion ratio of 9 and a ram speed of 3.33 mm s−1 . In order to examine microstructure evolution during hot extrusion, extrusion was interrupted when the material had emerged ∼6 mm from the die and the alloy then quenched in water. Samples for optical microscopy (OM) and scanning electron microscope (SEM) observation were etched in a solution of acetic picral after mechanical polishing. The average DRXed grain sizes and volume fraction of DRXed grain (designated as the DRX ratio) were measured using an Image-Pro Plus 5.0 software with more than 1000 grains and by more than five continuous OM images at a magnification of 200×. Dynamic microstructure and texture analyses by EBSD were conducted using a JEOL JSM-7000F scanning electron microscope (FESEM) operating at 25 kV equipped with TSL MSC-2200 software. Specimens for tensile tests were machined from the as-cast and extruded alloy specimens. The specimens for tensile tests have a gauge length of 15 mm and cross sectional areas of 6 mm × 2 mm.

S.W. Xu et al. / Materials Science and Engineering A 528 (2011) 4055–4067

4057

Fig. 2. EBSD results showing (a) IQ map, (b) IPF map and (c) pole figures of the as-cast ZWK510 alloy specimen. The EBSD data were obtained on the central section parallel to the casting direction.

Tensile tests were performed at room temperature using an Instron 5569 universal test machine at a crosshead speed of 1 mm/min. The tensile axis was selected to be parallel to the extrusion direction for the extruded alloy specimen.

3. Results and discussion 3.1. Microstructures of the as-cast ZWK510 alloy specimen Fig. 1 shows the typical microstructures and the XRD pattern of the as-cast ZWK510 alloy specimen. The as-cast alloy specimen exhibits dendritic solidification structure, in which the eutectic phases locate in the interdendritic region, as shown in Fig. 1(a). To confirm the existence of the I-phase, TEM observations were performed. Fig. 1(b) and (c) shows the TEM morphology and its selected area diffraction patterns (SADP). The SADPs taken from the eutectic lamellar phase show 2-, mirror, 3- and 5-fold symmetries, respectively, which is a distinct characteristic of the icosahedral quasicrystalline I- phase. The XRD pattern (Fig. 1(d)) also reveals that the as-cast alloy consists of two phases of a-Mg and I-phase. Fig. 2 shows the EBSD results of the as-cast ZWK510 alloy specimen. The as-cast alloy specimen has a uniform grain structure (Fig. 2(a)) with an average grain size of about 60 ␮m, which is due to the grain refining effect of Zr element [14] and a rapid cooling in the steel mold during casting. Their inverse pole figure (IPF) maps (Fig. 2(b)) and corresponding pole figures (Fig. 2(c)) do not exhibit any special orientation distribution, i.e., the texture of ascast ZWK510 alloy specimen is random. Therefore, no special effect from initial texture can be shown on the DRX behavior during the following hot extrusion deformation in this study.

3.2. Microstructural evolution during hot extrusion In order to examine microstructure evolution during hot extrusion, ZWK510 specimens from interrupted hot extrusion was prepared. The optical microstructures of each region marked in the schematic map of the interrupted extrusion sample are shown in Fig. 3. Fig. 3(a)–(d) corresponds to the microstructures before and after passage through the die. During the initiation of hot extrusion, coarse twins are formed within some grains (Fig. 3(a)), which is due to limited slip systems of magnesium [7,9]. As further deformation is carried out (Fig. 3(b) and (c)), the original microstructures are elongated along the material flow by the extrusion process. In these microstructures fine grains with a size of several microns are observed around the original grain boundaries. This indicates that fine grains formed around the grain boundaries during hot extrusion by DRX. In the microstructure after passage through the die (Fig. 3(d)) the mixed grain structures, consisting of coarse and fine grains, twins and original grains, are further elongated along the extrusion direction (compare Fig. 3(d) with Fig. 3(c)). The grain sizes in both the coarse and fine grain bands are further refined by extrusion. Since the above-mentioned dynamic microstructural changes during hot extrusion are complicated, these processes cannot be analyzed only by optical microscope. In the following sections, the twin type and the potential “Twin DRX” process, the microstructural changes at original grain boundaries and around the I-phase during hot compression at 350 ◦ C are investigated in detail by EBSD analysis with high resolution. Figs. 4–9 show the IPF maps of the squared areas indicated in Fig. 3. In the IPF maps grain boundaries are indicated by various lines depending upon the grain-to-grain misorientation angles:

4058

S.W. Xu et al. / Materials Science and Engineering A 528 (2011) 4055–4067

Fig. 3. Schematic map of the interrupted extrusion sample and the optical micrographs following interrupted extrusion marked in the schematic map. (a–d) Correspond to microstructures before and after passage through the die.

white or green (depend on the contrast) for 2◦ <  < 15◦ and black for 15◦ <  < 100◦ . Firstly, the twin type should be confirmed because there are various types of twins in Mg. The type of twin can be deduced from the misorientation angle between the twin and the matrix [25,26]. Namely, the {1 0 1¯ 2} twin, which is the c axis tensile twin, produces a basal-plane tilt of 86.3◦ and the {1 0 1¯ 1} twin, which is c axis compression twin, produces a basal-plane tilt of 56.2◦ . Also, there is a special {1 0 1¯ 1} − {1 0 1¯ 2} double twin, which is generated when {1 0 1¯ 2} re-twinning occurs inside the {1 0 1¯ 1} twin, the net result is areorientation  of the original caxis by 30.1◦ or 37.5◦ around the 1 0 1¯ 2 axis. Fig. 4 shows the EBSD results of the area squared in Fig. 3(a). It can be seen in Fig. 4(e) that the distribution of misorientation angle shows local maxima in the ranges of 4–9◦ and 84–89◦ . The peak in the range of 84–89◦ is attributable to a significant proportion of {1 0 1¯ 2}   twin, since the plotting of 86◦ 1 0 1¯ 2 boundaries (±5◦ on both axis and angle) on IPF Fig. 4(a) shows that they are associated with boundaries having morphology and habit planes consistent with {1 0 1¯ 2} twin (confirmed by author through an TSL OIM5.2 software). This suggests that at the beginning of deformation by extrusion the {1 0 1¯ 2} twin is formed. The thickness of these tensile twin lamellas falls into the range of 15–35 ␮m and their volume faction does not exceed 20%. Furthermore, in some area the {1 0 1¯ 2} re-twinning occurring inside the {1 0 1¯ 2} twin is observed (as shown in Fig. 4(b) that A has the {1 0 1¯ 2} twin relationship with B, and B has the {1 0 1¯ 2} twin relationship with C but A does not have any twin relationship with C), by which

the original grains are further divided up. However, the volume faction of this kind of re-twinning does not exceed 5%. A recent study on microstructure evolution during hot extrusion of a Mg–6 Zn–0.4 Ag–0.2 Ca–0.6 Zr (all in mass%) at 350 ◦ C by Oh-ishi et al. [14] reported that deformation twinning was not detected by the EBSD and TEM analyses in the interrupted extrusion sample, which is inconsistent with our results. This is because for the Mg–6 Zn–0.4 Ag–0.2 Ca–0.6 Zr (all in mass%) alloy, in the interior of the original grains a large amount of finely dispersed precipitates are formed [14], by which the necessary shuffling for twinning is prevented. While for the present alloy, the precipitate is not as many as that in [14] in the twinned area (the SEM image for the extruded alloy specimen will be shown in the following section). Furthermore, in the study on deformation behavior of Mg–6.0 Zn–1.5Y–0.5 Zr (all in mass%) during hot compression test from 200 ◦ C to 400 ◦ C by Zhang et al. [21], the appearance of deformation twinning was not mentioned too. This may be due to the limits of observation equipments, as in [21] the microstructures were observed by optical microscope and TEM, the formed twins engulfed by the following DRX process may be easily neglected. For example as shown in Fig. 4(f), which is an enlargement of the area indicated in Fig. 4(c), many fine areas with different orientations from the original grains were formed along the original grain boundaries. By further EBSD analysis, we found that these fine areas have the twin relationship with the original grains ({1 0 1¯ 2} twin 86.1◦ , {1 0 1¯ 1} twin 56◦ and {1 0 1¯ 1} − {1 0 1¯ 2} twin 37.5◦ ). These relationships have been confirmed by EBSD analysis, using the plotting of misorientation angle/axis boundaries (±5◦ on both axis and angle) on

S.W. Xu et al. / Materials Science and Engineering A 528 (2011) 4055–4067

Fig. 4. EBSD results showing: (a) the IPF map of the area squared in Fig. 3(a); (b) an enlargement of the squared area

4059

in Fig. 4(a), showing the formation of tensile twins;

(c) an enlargement of the squared area in Fig. 4(a), showing the bulging out of the original grain boundaries and the formation of subgrains, substructures, and also some DRXed grains; (d) misorientation profiles along the indicated directions marked in Fig. 4(c); (e) distribution of misorientation angles for the area shown in Fig. 3(a) and (f) an enlargement of the indicated area in Fig. 4(c).

IPF maps. Here we called them “fine areas” because the EBSD map is a planar map, we cannot confirm these “fine areas” are resulted from the twin DRX or they are only the vertical sections of twins. Therefore, further study on this is needed. In the present study, around the area shown in Fig. 3(a) eight EBSD maps including 42 twins were taken and all the twins are investigated. The results are nearly the same as shown in Fig. 4, indicating that at the beginning

of deformation by extrusion the twins such as {1 0 1¯ 2} twins are indeed formed. The other local maxima in the range of 4–9◦ (Fig. 4(e)) is mainly corresponding to the low-angle boundaries formed near the original grain boundaries as shown in Fig. 4(c). At the beginning of deformation by extrusion, the original grain boundaries become serrated and some subgrains are formed at their rims (indicated by

4060

S.W. Xu et al. / Materials Science and Engineering A 528 (2011) 4055–4067

Fig. 5. EBSD results showing: (a) the IPF map of the squared area

red arrows in Fig. 4(c)). Subgrain nucleation in these rims is characterized by original grain boundary bulging going with progressive lattice misorientation. Firstly, a section of original grain boundary locally bows out at a very short distance due to the strain-induced boundary migration caused by interaction between moving lattice dislocations and the original grain boundaries with massive eutectic I-phases. This results in the generation of serrated original grain boundaries. Then with increasing strain, the original grains are further deformed, leading to a progressive increase in dislocation density in these bulging rims, which can be reflected by the increase in misorientation angle (shown in Fig. 4 (c) and (d) along the indicated directions). At last, the nucleation of subgrain occurs

in Fig. 3(b) and (b) the distribution of twins in (a).

by the progressive misorientation of these bulging dislocation cells (see the orientation gradient indicated by color gradient between subgrains and original grain body in Fig. 4(c). Different colors in Fig. 4 indicate different crystal orientation). The microstructure at this stage can be called as “infancy of necklace structure”, which is optically invisible. Furthermore, in Fig. 4(c) some DRXed grains have already been formed near the original boundaries with Iphase, due to high density of dislocations accumulating near these areas, indicating the I-phase can accelerate the DRX behavior by obstructing the slip of dislocations. It also can be seen in Fig. 4(c) that even if inside the same original grain the deformation or dislocations accumulation is not uniform (see the orientation gradient

S.W. Xu et al. / Materials Science and Engineering A 528 (2011) 4055–4067

4061

Fig. 6. IPF map of the squared area in Fig. 3(b), showing the formation of subgrains and DRXed grains near the original grain boundaries and the repetitive formation of subgrains near the newly formed DRXed grain boundaries.

indicated by color gradient inside the same original grain), as a result, the original grains are gradually divided into some blocky substructures. When subjected to further deformation, which corresponds to the OM microstructures shown in Fig. 3(b) and (c) and IPF maps shown in Figs. 5–8, much more subgrains are nucleated near the original grain boundaries (Figs. 5(a) and 6), then a real “necklace structure” is formed with the low-angle grain boundaries of subgrains growing to high-angle grain boundaries as shown in Fig. 7. The blocky substructures can also grow to some massive “new grains” in the same way as shown in Fig. 7. All of these DRX processes belong to continuous DRX mechanisms. Continuous DRX is essentially a one-step phenomenon, i.e., new grains are nucleated homogenously at the original grain boundaries or inside the original grains by forming the massive blocky substructures in the present study and their size can hardly grow. Furthermore, dislocations can also accumulate at the newly formed DRXed grain boundaries then some subgrains or substructures followed by DRXed grains occur repetitively (Figs. 4(c) and 6). Consequently, as further deformation is carried out the width of “necklace structure” increases and the original microstructures are refined, as shown in Fig. 9, which corresponds to the microstructures after passing through the die. The microstructural evolution for the twinning area during hot extrusion is different from that near the original grain boundaries introduced above. As further deformation is carried out (corresponding to microstructural development from Fig. 3(a) to Fig. 3(b) and (c)), much more twins are formed, including {1 0 1¯ 1} twin and {1 0 1¯ 1} − {1 0 1¯ 2} double twin, as shown in Fig. 5(b). However, it seems that most of these twins are only elongated along the material flow, the DRX seldom occurs in these twins. When sub-

Fig. 7. IPF map of the squared area in Fig. 3(c), showing the formation of “necklace microstructure” near the original grain boundaries.

jected to further deformation, which corresponds to the IPF maps shown in Fig. 8, most of the {1 0 1¯ 1} twins are developed to {1 0 1¯ 1} − {1 0 1¯ 2} double twins. Although some subgrains are observed in the {1 0 1¯ 2} twins and {1 0 1¯ 1} − {1 0 1¯ 2} double twins, as indicated in Fig. 8 by dark arrows, the amount is small. The DRX process in {1 0 1¯ 1} − {1 0 1¯ 2} double twin may follow the manner which has been reported in our previous work [27] and it is also a kind of continuous DRX. Based on the above-mentioned analysis, it can be concluded that during the hot extrusion of the ZWK510 alloy at 350 ◦ C, the main DRX mechanism is the continuous DRX near the original grain boundaries. The I-phase can accelerate the DRX behavior near these areas by obstructing the slip of dislocations. The deformation twins and massive blocky substructures formed in original grains can coordinate the DRX process near the original grain boundaries, however the DRX seldom occurs inside these area. As further deformation is carried out, these deformation twins and massive blocky substructures are elongated along the material flow and become so-called unDRXed area, then a bimodal grain structure composed of fine grains of several microns and unrecrystallized coarse area is formed as shown in Fig. 9. 3.3. Microstructures of the extruded ZWK510 alloy specimen Fig. 10 shows the microstructures of ZWK510 alloy hotextruded at 350 ◦ C. The extruded ZWK510 alloy shows a bimodal grain structure consisting of ultrafine grains and coarse unDRXed grains (Fig. 10 (a) and (b)). The DRX ratio is about 58%. In the

4062

S.W. Xu et al. / Materials Science and Engineering A 528 (2011) 4055–4067

Fig. 8. IPF map of the squared area

in Fig. 3(c), showing the microstructure evolution of the twins.

DRXed area around the crushed eutectic I-phase a large number of fine I-phase precipitates with an average size of about 0.2 ␮m are observed pinning at the newly formed DRXed grain boundaries (indicated by arrows in Fig. 10(c)), which can inhibit the DRXed grain growth, therefore an average grain size of about 2.1 ␮m is obtained in these DRXed area (Fig. 10 (c) and (d)). Most of these fine I-phases were precipitated during hot extrusion process; of course some of them are probably the destroyed eutectic I-phase [18]. Also, there are many precipitate-free zones (PFZ), high magnification SEM observation shows that most of the PFZ are located in the above-mentioned coarse unDRXed areas. This may be the main reason why the deformation twin can be formed in the present study, as the necessary shuffling for twinning is not prevented in these zones. Fig. 11 shows the pole figures of the extruded ZWK510 alloy specimen at a central section parallel  to the extrusion direction  (ED) and the (0 0 0 1) 1 1 2¯ 0 basal slip Schmid factor when there is a tensile force along ED. As the DRX ratio is only 58%, the pole figures for the DRXed and unDRXed areas are investigated, respectively. The pole figures of both areas exhibit high intensity along the equatorial circumference in the (0 0 0 1) pole figure (Fig. 11 (a) and (b)), while in the (1 0 1¯ 0) pole figures intensities are high at the top and bottom poles (Fig. 11 (a) and (b)) and two fixed latitude lines (Fig. 11(b)). This is a typical basal fiber texture, plane of the Mg is aligned along the ED with  i.e., the basal  its 1 0 1¯ 0 direction parallel to ED. The unDRXed area has a stronger fiber texture than the DRXed area (compare Fig. 11(b) with Fig. 11(a)), therefore, the actual pole figures shape of the extruded ZWK510 alloy specimen is mainly determined  by the unDRXed area, as shown in Fig. 11(c). The (0 0 0 1) 1 1 2¯ 0 basal slip

Schmid factor is 0.12 and 0.22 for the unDRXed and DRXed area, respectively. This indicates that during the tensile test along the ED the unDRXed area will give a strong contribution to the strength and the DRXed area could also offer some contribution to the duc tility except for the strength. The average (0 0 0 1) 1 1 2¯ 0 basal slip Schmid factor of the extruded ZWK510 alloy specimen is 0.17, as shown in Fig. 11(d). 3.4. Mechanical properties of the as-cast and extruded alloy specimens Fig. 12(a) shows the tensile properties of the as-cast and extruded ZWK510 alloy specimens. The 0.2% proof strength (Pf) and the ultimate tensile strength (UTS) of the extruded ZWK510 alloy specimen are 317 and 363 MPa, respectively, with an elongation to failure (ε%) of 12, which are much higher than that of the as-cast alloy specimen (Pf = 105 MPa, UTS = 168 MPa, ε% = 1.8). The enhanced strengths have been attributed to strong basal fiber texture, refined grain size as well as the existence of fine precipitates formed by the hot extrusion. In general, the strength  of metallic materials can be estimated from the grain size, i.e., expressed by  =  0 + kd−1/2 , the so-called Hall–Petch relationship, where  0 is the frictional stress, k is the slope and d is the grain size. The values of  0 and k for Mg–5.0 wt.% Zn were reported to be about 70 MPa and 217 MPa ␮m−1/2 , respectively [14,28]. For example, for the as-cast ZWK510 alloy in this study the grain size is 60 ␮m, and so the value of  is calculated to be 98 MPa. The actual Pf of the ascast ZWK510 alloy is 105 MPa, nearly consistent with the calculated value. Of course, if the strengthening effect of the eutectic I-phase locating in the interdendritic region is considered, the calculated

S.W. Xu et al. / Materials Science and Engineering A 528 (2011) 4055–4067

4063

Fig. 9. IPF map for the area after passage through the die, corresponding to the microstructure shown in Fig. 3(d).

value will be more accurate. In the case of extruded ZWK510 alloy specimen, as mentioned above, the microstructure consists of a bimodal grain structure of fine DRXed grains of 2.1 ␮m and coarse unrecrystallized zones. If the average DRXed grain size is used for calculation, the value of  is 220MPa, which is much lower than the experimentally obtained Pf of 317 Mpa. This is because that the value of 220 MPa was calculated in an ideal condition. We assumed that the microstructure only consists of DRXed grains with random orientation. However, in fact, there are some coarse unDRXed zones with a strong fiber texture in the (0 0 0 1) plane, so the experimental Pf is higher. This indicates that for the ZWK510 alloy the strengthening effect resulted from a strong fiber texture in the (0 0 0 1) plane is more significant than that from the grain refinement. Furthermore, the presence of fine I-precipitate during hot extrusion could also lead to an additional strengthening. The important effects of fine precipitates on improving the strength and anisotropy of the 0.2%

proof strength during tensile and compression tests have been definitely confirmed by Oh-ishi et al. [14] in the Mg–6 Zn–0.4 Ag–0.2 Ca–0.6 Zr (all in mass%) alloy hot-extruded at 350 ◦ C. The enhanced ductility in extruded ZWK510 alloy has been attributed to the DRX as well as breaking up of eutectic I-phase during hot extrusion. As shown in Fig. 12(b), in the as-cast ZWK510 alloy specimen the fracture is formed in the massive eutectic Iphase during tensile test. However, in the extruded alloy specimen, the microstructures are refined by DRX and the coarse eutectic Iphase is broken and dispersed in the matrix, therefore, the fractures are delayed. As mentioned in Section 3.3, the basal planes in the unDRXed regions of extruded alloy sample are orientated parallel to the ED, namely, (0 0 0 1)matrix //ED, therefore, when tensile tests are performed along the direction parallel to ED, a contraction strain component occurs parallel to the c-axis in these unDRXed regions, then {1 0 1¯ 1} compression twins and subsequent {1 0 1¯ 1} −

4064

S.W. Xu et al. / Materials Science and Engineering A 528 (2011) 4055–4067

Fig. 10. Microstructures of ZWK510 alloy hot-extruded at 350 ◦ C. (a) Low magnification OM microstructure, (b) magnified image for the area squared in Fig. 10(a), (c) SEM image, showing the dispersed I-phase and precipitates and (d) distribution of the DRXed grain size.





Fig. 11. EBSD results showing the (a–c) pole figures and (d) (0 0 0 1) 1 1 2¯ 0 basal slip Schmid factor (when there is a tensile force along the extrusion direction) of ZWK510 alloy hot-extruded at 350 ◦ C. (a) Pole figures of the DRXed area, (b) pole figures of the unDRXed area and (c) pole figures for all the area.

S.W. Xu et al. / Materials Science and Engineering A 528 (2011) 4055–4067

4065

Fig. 12. (a) Tensile stress–strain curves of as-cast and extruded ZWK510 alloy specimens, (b) OM image near the fractured surface of the as-cast specimen after tensile test, (c) OM images for the extruded specimen deformed to near the 0.2% proof strength point during tensile test and (d) OM images near the fractured surfaces of the extruded specimen after tensile test. In Fig. 12(b–d) the horizontal direction is the tensile test direction. The tensile axis was parallel to the extrusion direction for the extruded alloy specimens.

Fig. 13. (a) and (b) IPF maps showing the twinning type near the fractured surfaces of the extruded specimen after tensile test, (c) a schematic showing the crack formation at the interface between the {1 0 1¯ 1} − {1 0 1¯ 2} double twin and the matrix. In Fig. 13(a) and (b) the horizontal direction is the tensile test direction. The tensile axis was parallel to the extrusion direction for the extruded alloy specimen.

4066

S.W. Xu et al. / Materials Science and Engineering A 528 (2011) 4055–4067

{1 0 1¯ 2} double twins are activated, as shown in Fig. 13. Thereby, after tensile test on the fractured surface of the extruded ZWK510 alloy specimen, many twins are observed as shown in Fig. 12(c) and (d). However, after the tensile test many cracks are also observed nucleating at the twin boundaries (Fig. 12(d)). Recently, Ando and Koike [25] reported that during the tensile test of the rolled AZ31 alloy at room temperature, {1 0 1¯ 1} − {1 0 1¯ 2} double twins involving {1 0 1¯ 2} tensile twins were propagated within these original {1 0 1¯ 1} compression twins. These double twins could cause surface undulation and their boundaries could become starting points for the cracks [25]. The crack shown in Fig. 12(d) is resulted from the same mechanism, which has been confirmed by EBSD analysis as shown in Fig. 13. Based on the above-mentioned results and analysis, it can be assumed that a further increase in 0.2% proof strength of ZWK510 alloy with a higher ductility could be expected in the fully DRXed microstructure with a strong fiber texture in the (0 0 0 1) plane and with a large number of fine I-phase uniformly pinning at the DRXed grain boundaries. In order to achieve the above-mentioned microstructure, heat treatment (homogenization or solid solution + pre-aging) before hot extrusion and aging treatment after hot extrusion is considered. The precipitation of a large number of fine I-phase before and during hot extrusion may firstly inhibit the formation of deformation twin, therefore decrease the unDRXed area developed by {1 0 1¯ 2} tensile twin; secondly accelerate the DRX process or accelerate the formation of substructures by obstructing the slip of dislocations; thirdly inhibit the DRXed grain growth through pinning at the grain boundaries. Therefore, the next step of the research will be focused on realizing the microstructural control to further enhance the mechanical properties of the ZWK510 alloy.

(4) The extruded ZWK510 alloy shows a DRX ratio of about 58% and an average DRXed grain size of 2.1 ␮m. In the DRXed area around the crushed eutectic I-phase a large number of fine Iphase precipitates are observed pinning at the newly formed DRXed grain boundaries. While most of the coarse unDRXed areas are precipitate-free. The extrudedZWK510 alloy  shows a //ED. typical basal fiber texture of (0 0 0 1) 1 0 1¯ 0 matrix The unDRXed area has a stronger fiber texture than the DRXed area, therefore, the actual pole figures shape of the extruded ZWK510 alloy specimen  is mainlydetermined by the unDRXed area. The (0 0 0 1) 1 1 2¯ 0 basal slip Schmid factor is 0.12 and 0.22 for the unDRXed and DRXed area, respectively. (5) The 0.2% proof strength (Pf) and the ultimate tensile strength (UTS) of the extruded ZWK510 alloy specimen are 317 and 363 MPa, respectively, with an elongation to failure (ε%) of 12, which are much higher than that of the as-cast alloy specimen (Pf = 105 Mpa, UTS = 168 Mpa, ε% = 1.8). The enhanced strengths have been attributed to strong basal fiber texture, refined grain size as well as the existence of fine precipitates formed by the hot extrusion. The enhanced ductility in extruded ZWK510 alloy has been attributed to the DRX as well as breaking up of eutectic I-phase during hot extrusion. In the as-cast ZWK510 alloy specimen the fracture is formed in the massive eutectic Iphase during tensile test. In the case of extruded alloy specimen, the microstructures are refined by DRX and the eutectic I-phase is broken and dispersed in the matrix, therefore, the factures are delayed, however, the {1 0 1¯ 1} − {1 0 1¯ 2} double twins formed in the unDRXed area could cause surface undulation and their boundaries could become starting points for the cracks during tensile test.

Acknowledgement 4. Summary In this study, firstly dynamic microstructural changes of an ascast Mg–5.0 Zn–0.9 Y–0.16 Zr (wt.%) (designated ZWK510) alloy during hot extrusion at 350 ◦ C and a ram speed of 3.33 mm s−1 with an extrusion ratio of 9 was systematically investigated by electron backscattering diffraction (EBSD) analysis. The DRX behavior during hot extrusion was discussed. Then, the effect of microstructure and texture on the mechanical properties of the extruded ZWK510 alloy specimens at room temperature was discussed. The findings can be summarized as follows:

(1) As-cast ZWK510 alloy consisting of a-Mg and quasicrystalline I-phase was prepared. The as-cast alloy specimen has a uniform grain structure with an average grain size of about 60 ␮m and a random orientation distribution. (2) At the beginning of deformation by extrusion the deformation twins such as {1 0 1¯ 2} twins are formed. As further deformation is carried out, much more twins are formed, including {1 0 1¯ 1} twin and {1 0 1¯ 1} − {1 0 1¯ 2} double twin. (3) During the hot extrusion process at 350 ◦ C, the main DRX mechanism of ZWK510 alloy is the continuous DRX near the original grain boundaries. The I-phase can accelerate the DRX behavior near these areas by obstructing the slip of dislocations. The deformation twins and massive blocky substructures formed in original grains can coordinate the DRX process near the original grain boundaries, however the DRX seldom occurs inside these area. As further deformation is carried out, these deformation twins and massive blocky substructures are elongated along the material flow and become so-called unDRXed area, then a bimodal “necklace structure” composed of fine grains of several microns and unrecrystallized coarse area is formed.

This study was supported by the National Nature Science Foundation of China (no. 50201005) and Project of scientific researchers serving for enterprises supported by Ministry of Science and Technology, China (2009GJB20003).

References [1] K. Hono, C.L. Mendis, T.T. Sasaki, K. Oh-ishi, Scripta Mater. (2010), doi:10.1016/j.scriptamat.2010.01.038. [2] S.X. Ding, W.T. Lee, C.P. Chang, L.W. Chang, P.W. Kao, Scripta Mater. 59 (2008) 1006–1009. [3] K. Oh-ishi, K. Hono, K.S. Shin, Mater. Sci. Eng. A 496 (2008) 425–433. [4] D.G. Leo Prakash, D. Regener, J. Alloys Compd. 467 (2009) 271–277. [5] S.W. Xu, N. Matsumoto, S. Kamado, T. Honma, Y. Kojima, Mater. Sci. Eng. A 517 (2009) 354–360. [6] J. Zhang, R.l. Zuo, Y.X. Chen, F.S. Pan, X.D. Luo, J. Alloys Compd. 448 (2008) 316–320. [7] Y. Kojima, S. Kamado, Mater. Sci. Forum 488–489 (2005) 9–16. [8] D.G. Kim, H.T. Son, J.S. Lee, S.K. Park, Y.M. Kim, M.Y. Hu, J. Alloys Compd. (2010), doi:10.1016/j.jallcom.2010.06.082. [9] J. Koike, Met. Mater. Trans. A 36A (2005) 1689–1696. [10] T. Mukai, K. Higashi, Scripta Mater. 44 (2001) 1493–1496. [11] H. Somekawa, T. Mukai, Mater. Sci. Eng. A 459 (2007) 366–370. [12] S.S. Park, G.T. Bae, D.H. Kang, I.H. Jung, K.S. Shin, N.J. Kim, Scripta Mater. 57 (2007) 793–796. [13] T. Homma, N. Kunito, S. Kamado, Scripta Mater. 61 (2009) 644–647. [14] K. Oh-ishi, C.L. Mendis, T. Homma, S. Kamado, T. Ohkubo, K. Hono, Acta Mater. 57 (2009) 5593–5604. [15] T. Homma, C.L. Mendis, K. Hono, S. Kamado, Mater. Sci. Eng. A 527 (2010) 2356–2362. [16] D.H. Baea, M.H. Lee, K.T. Kim, W.T. Kim, D.H. Kim, J. Alloys Compd. 342 (2002) 445–450. [17] Z.P. Luo, S.Q. Zhang, Y.L. Tang, D.S. Zhao, Scripta Metall. 28 (1993) 1513–1518. [18] X.Q. Zeng, Y. Zhang, C. Lu, W.J. Ding, Y.X. Wang, Y.P. Zhu, J. Alloys Compd. 395 (2005) 213–219. [19] D.K. Xu, L. Liu, Y.B. Xu, E.H. Han, Mater. Sci. Eng. A 443 (2007) 248–256. [20] Y. Zhang, X.Q. Zeng, L.F. Liu, C. Lu, H.T. Zhou, Q. Li, Y.P. Zhu, Mater. Sci. Eng. A 373 (2004) 320–327. [21] Y. Zhang, X.Q. Zeng, C. Lu, W.J. Ding, Mater. Sci. Eng. A 428 (2006) 91–97.

S.W. Xu et al. / Materials Science and Engineering A 528 (2011) 4055–4067 [22] S.W. Xu, M.Y. Zheng, X.G. Qiao, W.M. Gan, K. Wu, S. Kamado, Y. Kojima, Key Eng. Mater. 350–358 (2007) 595–598. [23] M.Y. Zheng, S.W. Xu, X.G. Qiao, K. Wu, S. Kamado, Y. Kojima, Mater. Sci. Eng. A 483–484 (2008) 564–567. [24] M.Y. Zheng, S.W. Xu, K. Wu, S. Kamado, Y. Kojima, Mater. Lett. 61 (2007) 4406–4408.

4067

[25] D. Ando, J. Koike, J. Jpn. Inst. Metals 71 (2007) 684–687. [26] M.R. Barnett, Mater. Sci. Eng. A 464 (2007) 1–7. [27] S.W. Xu, N. Matsumoto, S. Kamado, T. Honma, Y. Kojima, Scripta Mater. 61 (2009) 249–252. [28] G. Mima, Y. Tanaka, Trans. JIM 12 (1971) 317–328.