LIU Yang, ZHAO Yiquan, WANG Ling, , JIN Xiuying, SUN Chao,WANG Xiaomin, WANG Gang, DAI Shiyu, WANG Yinong

(1. School of Materials Science and Engineering, Yingkou Institute of Technology, Yingkou 115014, China; 2. College of Chemistry and Materials Engineering, Hainan Vocational University of Science and Technology, Haikou 330044, China; 3. CATARC Automotive Test Center(Tianjin) Co., Ltd, Tianjin 300300, China; 4. School of Electrical Engineering, Dalian University of Technology, Dalian 116024, China;5. School of Materials Science and Engineering, Dalian University of Technology, Dalian 116024, China)

Abstract: To produce high strength and ductility Mg alloys with high productivity and low energy consumption, the residual heat rolling (RHR) process was initially proposed. The microstucture and mechanical properties of AZ31 processed by RHR were investigated by optical micrscopy (OM), electron backscatter diffraction (EBSD), and electron universal testing machine. The yield strength (YS), ultimate tensile strength(UTS), and elongation to failure of RHRed AZ31 sheet were 194 MPa, 311 MPa, and 22%, respectively. The RHRed AZ31 alloys after annealing have very fine and homogeneous grains. The symmetrical rolled (SR)and RHRed AZ31 exhibit typical {0002} basal textures. The RHRed AZ31 has double-peak basal texture distribution. The basal poles of RHRed AZ31 split from normal direction (ND) to rolling direction (RD).There are few hard orientation distributions on the basal ＜a＞ slip and more soft orientation distributions on the prismatic ＜a＞ slip in the RHRed AZ31 sheets than those in the SRed AZ31sheets.

Key words: magnesium alloys; microstructure; texture; grain refinement; mechanical properties; residual heat rolling

1 Introduction

Magnesium and its alloys are widely applied in the industrial fields due to many advantages, such as low density, high specific strength and excellent damping capacity, thus attracting considerable attentions as potential candidates to replace steel and aluminum alloys in automobile and aerospace industries where lightweight components are necessary[1-4]. However,Mg alloys suffer a lot from low ductility at room temperature, which is due to the hexagonal closed-packed(HCP) crystal structure with limited slip systems.Meanwhile, as is well-known, a strong basal texture with c-axis parallel to ND is usually introduced in Mg alloy sheets normally fabricated by rolling or extrusion process[5]. In this condition, the operative ＜a＞ slip systems (either on basal or prismatic plane) can not accommodate the plastic deformation along c-axis at ambient temperature[6-9]. Hence, extensive application of the Mg alloys is greatly restricted.

An effective method to improve the ductility of Mg alloys is texture modification. Texture modification can be divided into two special approaches: one is to decrease the maximum intensity of basal texture with c-axis still parallel to ND; the other is to tilt the basal pole from ND to other direction in some extent. Sometimes the intensity of basal texture decreasing and basal pole tilting can be obtained simultaneously[10-12]. Methods to weaken the textures have been extensively investigated such as equal channel angular pressing (ECAP),different speed rolling (DSR), single roller drive rolling(SRDR) and rare earth element alloying[13-17]. However,the drawbacks of these processing routes are evidently the forming limitations, process complexity, high cost and low productivity. Therefore, it is necessary to put forward a new processing to get out of the current dilemma. In this study, a cost effective and high-productivity process named by RHR was put forward to get better ductility and higher strength in AZ31 alloy sheets, through a comparison with symmetric rolling(SR). The microstructure, texture and mechanical properties were investigated in as-cast AZ31 alloy sheets processed by RHR and SR.

2 Experimental

2.1 Raw materials

The as-cast AZ31 alloys (Mg-2.6wt%Al-0.7wt%Zn-0.39wt%Mn) were prepared in an electrical furnace by the mild steel crucible. After held for 10 mins at 740 ℃, the melt was poured into a preheated mild steel mold with the dimension of 190 mm×130 mm×30 mm and then cooled in air. The as-cast AZ31 alloys were homogenized for 10h at 350 ℃.

2.2 Rolling process

Rolling was carried out in two AZ31 sheets with the size of 180 mm (rolling direction, RD) ×10 mm(width) ×10 mm (t) cut from the homogenized ingots by wire-electrode cutting. Schematic diagrams of the rolling processes are presented in Fig.1. Both of AZ31 sheets were firstly rolled from 10 mm to 5 mm in thickness after heated for 10 min at 500 ℃. In SR process,the AZ31 alloy sheet annealed at 500 ℃ for 5mins underwent the second rolling pass with 10% reduction from 5 mm to 4.5 mm. While in RHR process, the AZ31 alloy sheet was directly rolled with the reduction of 10%, without annealing. Finally, both of AZ31 sheets were annealed for 10 min at 500 ℃, then cooled in the air.

Fig.1 Schematic diagram of rolling process: (a) Symmetric rolling (SR); (b) Residual heat rolling (RHR)

2.3 Characterization

The microstructure evolution of the RHRed and SRed AZ31 sheet were observed by optical microscope(OM) on the cross sections of the normal direction(ND) in as-rolled sheets. The line-intercept method was applied to measure the grain sizes. The microstructures of RHRed and SRed AZ31 alloy sheets were analyzed by electron backscatter diffraction (EBSD) on a Zeiss Supra 55 scanning electron microscope (SEM). EBSD measurements were implemented at accelerating voltage of 20 kV and a step size of 1.5 μm, with a tilt angle of 70o. Samples for EBSD scanning were prepared by wire-electrode cutting, mechanical polishing and automatic electrochemical polishing in a weak acid solution of 10% perchloric acid and 90% absolute ethyl alcohol at a potential of 10 V for 30 s at -25 ℃.

2.4 Tensile test

Tensile specimens with 15mm in gage length,3 mm in width and 3 mm in thickness were machined from as-annealed sheets with the longitudinal axis along RD. In order to ensure the repeatability of tensile results, the triplicate universal tensile tests were conducted using an electron universal testing machine(CMT6305-300KN) with a strain rate of 1×10-3s-1at ambient temperature.

3 Results and discussion

3.1 Microstructures and texture evolution

The microstructures of starting AZ31 sheets were observed by OM are shown in Fig.2(a). It is composed of coarse grains (with average grain size of about 420 μm) and lots of eutecticβphase (Mg17Al12) precipitated from the matrix, as shown by arrow A and B. In order to improve the rollability of initial as cast AZ31 sheets,the homogenization treatment was carried out at 350 ℃for 10 h[18]. Theβphases were removed after homogenization, as shown in Fig.2(b). There is no appreciable difference between grain sizes of as-cast AZ31 sheet and as-homogenized one.

Fig.2 The optical microstructures of (a) as-cast and (b) as-homogenized AZ31 alloy sheet (arrows A and B in Fig.2(a)indicate β phases)

The microstructure evolution of homogenized AZ31 sheets processed by SR is shown in Figs.3(a)-3(d). The microstuctural characteristics of homogenized AZ31 sheets processed by the 1st pass of SR are fine DRXed grains, twins, intersection of twins and DRXed grains within twins, which is indicated by ellipses A and B. The microstructure of AZ31 processed by the 1st pass SR is inhomogeneous after annealing.Some coarse grains with size of 200 μm could be observed in Fig.3(b). The twins are substituted by coarsening static recrystallized (SRX) grains after annealing as shown in Fig.3(b). The grains of AZ31 processed by the 2nd pass of SR become much finer than those of annealed AZ31 processed by the 1st pass of SR.However, coarse grains are rarely observed in AZ31 processed by the 2nd pass of SR, as shown in Fig.3(c).The microstructure become more homogeneous after annealing for AZ31 processed by the 2nd pass of SR, as shown in Fig.3(d). Only a few coarse grains are observed in Fig.3(d). Fig.3(e) shows that there are abundant deformation twins appearing in the interior of grains and some small grains along pre-exisiting large grains boundaries in RHRed AZ31. It is obvious that all the coarse grains and twins were substituted by homogeneous SRXed grains for RHRed AZ31 sheet after annealing. It is well known that dislocation slipping and twinning are two major deformation mechanism. Due to higher dislocation density and more recrystallization nucleation sites provided by twins, the SRXed grains of as-annealed RHRed AZ31 is more homogeneous than those of as-annealed the 2nd SRed AZ31[19-21].

Fig.3 The optical microstructures of AZ31 alloy sheets processed by SR and RHR: (a) As-rolled by the the 1st pass of SR; (b)As-annealed by the the 1st pass of SR; (c) As-rolled by the the 2nd pass of SR; (d) As-annealed by the the 2nd pass of SR; (e) As-rolled by the RHR; (f)As-annealed by the RHR(The ellipses A and B in Fig3(a) indicate fine DRXed grains,twins, intersection of twins and DRXed grains within twins;the arrows in Fig.3(e) indicate deformation twins and small grains along pre-existing large grains, respectively)

The EBSD IPF maps of AZ31 sheets processed by SR and RHR are shown in Fig.4. Each color in the IPF maps present a special crystal orientation relative to the sample coordinate system. There are a high volume fraction of grains in red color shown in Figs.4(a)and 4(b), which indicates that the sheets possess a strong basal texture with their c-axis parallel to ND.As-annealed AZ31 sheet by the 1st pass of SR exhibited stronger basal texture than the as-rolled one by the 1st pass of SR, which can be seen in Figs.5(a) and 5(b).Additionally, the grain size of as-annealed AZ31 processed by RHR was much finer than that of as-annealed AZ31 sheet processed by the 2nd pass of SR, as shown in Figs.4(c) and 4(d). The (0002) basal pole spreads away from ND to RD, as shown in Fig.5(c). There are considerable blue, green, and even white grains appearing in the as-annealed AZ31 sheet processed by RHR,which indicates that there are also other crystal orientations parallel to ND, not just c-axis. It is observed in Fig.5(d) that c-axis tilts from ND to RD, forming the symmetrical two-peak distribution of basal texture. The maximum values of basal pole intensity of RHRed and SRed AZ31 sheets are almost the same, which are 12.2 and 12.6, respectively. The analogous phenomenon that c-axis tilts from ND to RD and forms the symmetrical texture has been reported in related literatures[22,23]. For instance, by a novel texture modification method including in-plane precompression, prestretching and annealing, HE Junjieet alintroduced a symmetrical fourpeak distribution of basal texture in an extruded AZ31 alloy sheet, which exhibited an excellent Erichsen value of 5.6 mm[24].

Fig.4 EBSD IPF maps of AZ31 sheets processed by SR and RHR: (a) As-rolled by the the 1st pass of SR; (b) As-annealed by the the 1st pass of SR; (c) As-annealed by the the 2nd pass of SR; (d) As-annealed by the RHR

Fig.5 (0002) pole figures of AZ31 sheets processed by SR and RHR: (a) As-rolled by the the 1st pass of SR; (b) As-annealed by the the 1st of pass SR; (c) As-annealed by the the 2nd pass of SR; (d) As-annealed by the RHR

It is noted that the coarse grains in the 1st pass of SRed sheet have basal texture according to the inverse pole figure(IPF) in Fig.4(a). The grains with basal texture were favorable for extension twinning, thus some twins are observed in Fig.3(e). However, the limited twins can not accommodate 10% rolling reduction in RHR process at low rolling temperature. The dislocation slipping in fine DRXed grains inherited from pre-rolling pass is absolutely necessary. The dislocation slipping could induce the fine DRXed grains rotate[25].The angle of resultant force of press and friction could be determined by the friction coefficient. Compared the the 2nd pass of SR and RHR, the rolling temperature was the only factor to change the friction coefficient,and the friction coefficient increases with the decrease of temperature, thus the resultant force tilted for the black arrow to red arrow as shown in Fig.6. The angle of fine DRXed grains rotated was related to the resultant force in forward slip zone and backward slip zone.The probable DRXed rotation situations are schematically shown in Fig.7. Due to the difference of direction of resultant force, the grains rotations are different in backward slip zone and forward slip zone. It is predicted that there are two situations for DRXed grains rotation. The first one is nearly all the fine DRXed grains rotated with the angle ofαin backward slip zone, and then some of the rotated grains rotated back with the angle ofα+βin forward slip zone, as shown in Figs.7(a) and 7(b). The second one is some of the grains rotated with the angle ofαin backward slip zone and the others rotated angle ofβin forward slip zone,as shown in Figs.7(a) and 7(c). According to this prediction, it is understandable to obtain the double peak of basal texture in RHR process.

Fig.6 Schematic diagram of rolling force mechanics

Fig.7 DRXed grains with basal texture rotation mode: (a) Rotation in backward slip zone; (b) Rotated grains rotation in forward slip zone; (c)Rotation in forward slip zone

3.2 Mechanical property at room temperature

The mechanical properties of AZ31 alloy sheets processed by different processing methods at room temperature are listed in Table 1[26-28]. It is obvious that the RHRed sheets have the highest yield strength(YS) and ultimate tensile strength (UTS) among all the processed AZ31 sheets in Table 1. Fig.8 shows the uniaxial tensile stress-strain curves of SRed and RHRed AZ31 alloy sheets. The specific values of mechanical properties of the RHRed sheet are 194 MPa (YS), 311 MPa (UTS), and 22% (El), and the specific values of the SRed one are 168 MPa (YS), 267 MPa (UTS), and 18% (El), respectively. It is obviosly shown that the RHRed sheet possesses not only higher YS and UTS,but also a better ductility than that of the SRed one.The higher strength of the RHRed sheet is attributed tosmaller and more homogeneous grains than that of the SRed one after annealing. According to the Hall-Petch relation, the smaller the grains are, the higher strength the material can get, which is proven by the results in this study. Most conclusions of the early investigations pointed that the basal ＜a＞ slip was the dominant deformation mechanism of Mg and its alloys at room temperature, since it had the lowest critical resolved shear stress (CRSS). However, Chino Yet alsuggested that by the introduction of the special texture in the cross rolling of AZ31 alloy, the prismatic ＜a＞ slip was activated to domain the deformation process during the tensile test near ambient temperature[29]. Therefore, it is necessary to apply the basal ＜a＞ slip and prismatic ＜a＞slip as the principal deformation mechanism to explain the actual results. Thus the Schmid factor distribution maps of both slip systems in tensile direction parallel to RD are shown in Fig.9. In the case of SRed sheet, the average values of the calculated Schmid factors on the basal ＜a＞ slip and prismatic ＜a＞ slip for RD are 0.26 and 0.41, respectively. In the case of RHRed sheet, the corresponding average values are 0.31 and 0.38, respectively. It is easily observed that in the distribution of Schmid factors for hard orientation components on the basal ＜a＞ slip (those crystal orientations with their Schmid factors values are close to 0, which are marked with a blue elliptical lines in Figs.9(a) and 9(c)), the SRed sheet has a higher fraction than the RHRed one,which indicates that the basal ＜a＞ slip is more difficult to activate in SRed sheet. Although the the average value of the calculated Schmid factors of the SRed sheet on the prismatic ＜a＞ slip is higher than that of the RHRed one, there is a lower fraction of soft orientation components on the prismatic ＜a＞ slip in the SRed sheet. The crystal orientations with Schmid factors values close to 0.5 are marked with an yellow elliptical lines in Figs.9(b) and 9(d)). Under the assumption that the basal ＜a＞ slip and prismatic ＜a＞ slip are the principal deformation mechanism, the RHRed sheet can exhibit a better ductility than that of the SRed one at room temperature. From the above results and discussion, it can be deduced that the as-annealed AZ31 alloy sheet processed by RHR exhibited the combination of fine-grained microstructure and excellent mechanical properties compared with as-annealed AZ31 sheet processed by SR, due to the smaller grains and easily activated slip systems.

Table 1 Comparisons of ambient temperature mechanical properties of AZ31 alloy sheets using different rolling modes

Fig.8 The uniaxial tensile stress-strain curves of AZ31 alloy sheets at room temperature:(a) As-annealed by the RHR; (b)As-annealed by the 2st pass of SR(tensile axis parallel to RD)

Fig.9 Schmid factor distribution maps of AZ31 alloy sheets: (a) Basal slip of as-annealed by the the 2nd pass of SR; (b) prismatic slip of as-annealed by the the 2nd pass of SR; (c) Basal slip of as-annealed by RHR pass; (d) Prismatic slip of as-annealed by RHR pass

4 Conclusions

a) The RHR process has a higher efficiency to refine grains than that of SR process. The higher dislocation density and more recrystallization nucleation sites provided by twins in RHR process were considered as the major reason for the good effect of grain refining.

b) Both of as-annealed AZ31 alloy sheets exhibited typical {0002} basal textures.The maximum values of basal pole intensity of RHRed and SRed AZ31 sheets are almost the same, which are 12.2 and 12.6,respectively. The introduction of the symmetrical twopeak texture along RD is likely to be responsible for a better ductility in as-annealed RHRed AZ31 alloy sheet.

c) There are less hard orientation distributions on the basal ＜a＞ slip and more soft orientation distributions on the prismatic ＜a＞ slip in the RHRed sheet than those in the SRed sheet.

d) The RHR process is an effective way to obtain good combination of high strength and ductility of AZ31. The YS and UTS of RHRed AZ31 were increased 26 MPa and 44 MPa, respectively, compared to the SRed AZ31, from 168 MPa to 194 MPa and from 267 MPa to 311 MPa. Simultaneously, the elongation to failure increased by 22.22% in the RHRed sheet(22%) when compared to that of the SRed sheet(18%).