Microstructure evolution and mechanical properties of as-extruded Mg-Gd-Y-Zr alloy with Zn and Nd additions

Author’s Accepted Manuscript Microstructure evolution and mechanical properties of as-extruded Mg-Gd-Y-Zr alloy with Zn and Nd additions Zijian Yu, Chao Xu, Jian Meng, Xuhu Zhang, Shigeharu Kamado www.elsevier.com/locate/msea

PII: DOI: Reference:

S0921-5093(17)31668-4 https://doi.org/10.1016/j.msea.2017.12.070 MSA35912

To appear in: Materials Science & Engineering A Received date: 26 April 2017 Revised date: 19 November 2017 Accepted date: 16 December 2017 Cite this article as: Zijian Yu, Chao Xu, Jian Meng, Xuhu Zhang and Shigeharu Kamado, Microstructure evolution and mechanical properties of as-extruded MgGd-Y-Zr alloy with Zn and Nd additions, Materials Science & Engineering A, https://doi.org/10.1016/j.msea.2017.12.070 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Microstructure evolution and mechanical properties of as-extruded Mg-Gd-Y-Zr alloy with Zn and Nd additions Zijian Yua*, Chao Xua, Jian Mengb*, Xuhu Zhangc, Shigeharu Kamadoa a. Department of Mechanical Engineering, Nagaoka University of Technology, Nagaoka 940-2188, Japan b. State Key Laboratory of Rare Earth Resources Utilization, Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Changchun 130022, P. R. China c. Aerospace Research Institute of Materials & Processing Technology, Beijing 100076, P. R. China

: [email protected] (Zijian Yu) , [email protected] (Jian Meng)

*Corresponding author. Tel: +81-258-47-7304; Fax: +81-258-47-9770.

Abstract The

microstructure

evolution

and

mechanical 1

properties

of

as-extruded

Mg-11.5Gd-4.5Y-0.3Zr (wt.%) alloy with Zn and Nd additions were investigated. The addition of Zn inhibits the dynamic recrystallization (DRX) due to the presence of the long-period stacking ordered (LPSO) phase. The addition of Nd promotes the precipitation of the Mg5RE (RE: rare earth) phase. The existence of the densely distributed Mg5RE phase before hot extrusion promotes the DRX in subsequent hot extrusion process and leads to grain refinement. The increase in the number of Mg5RE phase particles degrades the mechanical properties of the resultant alloy. After hot extrusion, the studied alloys exhibit a bimodal microstructure consisting of fine dynamic recrystallized (DRXed) grains of several microns and

strongly

textured

course

un-DRXed

grains.

The

as-extruded

Mg-11.5Gd-4.5Y-1.5Zn-0.3Zr alloy exhibits an excellent balance of strength and ductility (tensile yield strength of 371 ± 3.0 MPa and elongation of 7.2 ± 0.8 %). The alloy strengthening is attributed to the bimodal microstructure, the Mg5RE and LPSO phases, and the basal texture. The tensile yield strength of the as-extruded Mg-11.5Gd-4.5Y-1.5Zn-0.3Zr alloy can be further increased to 425 ± 2.5 MPa by precipitation hardening with the T5 treatment. Keywords: Magnesium alloy; Rare earth; Mechanical properties; Microstructure

1. Introduction 2

Mg alloys have attracted considerable attention in the automobile and aerospace industries due to their competitive features such as low density, high specific strength, good damping capacity and excellent castability [1]. Compared with commercial Mg alloys, such as Mg-Al-Zn (AZ) and Mg-Y-RE (WE) systems, RE containing Mg alloys, such as Mg-Gd/Y, Mg-Gd/Y-Nd, Mg-Gd/Y-Zn systems, exhibit better mechanical performances at both room temperature and elevated temperatures [2-5]. In addition, these alloys exhibit a remarkable ageing-hardening response, since the high solubility of Gd and Y in Mg (23.5 wt % for Gd and 12.5 wt % for Y at their eutectic temperatures) declines sharply with decreasing temperature [5, 6]. The addition of Nd in Mg-Gd/Y alloys further improves the ageing-hardening response and mechanical properties by increasing the strengthening phase   and  precipitates [7-9]. Mg-Gd-Nd alloys show an even better creep resistance in comparison to the commercial magnesium alloy WE54 at 250 °C [10]. The addition of Zn in Mg-Gd/Y alloys has different effects on the ageing-hardening response. Gao et al [11] reported that additions of a small amount of Zn (~ 1 wt %) could significantly enhance the ageing-hardening responses of Mg-Gd alloys with small concentrations of Gd (~ 6 wt %). Honma et al [12] reported that Zn additions (~1 wt %) degrade the ageing-hardening response of Mg-Gd-Y alloys containing high concentrations of Gd (~11 wt%) and Y (~4 wt %), but lead to the formation of LPSO phase. This LPSO phase could significantly increase the mechanical properties via short-fiber strengthening [13, 14]. To date, many investigations have been performed on the Mg-Gd-Y based alloys with the aim of gaining desired mechanical performance by microstructural modification, including grain refinement, precipitation, and texture control by using various thermo-mechanical 3

treatments such as hot extrusion and post-deformation ageing [14-19]. After hot extrusion, the refined DRXed grains and homogeneously dispersed particles can significantly improve the mechanical properties of Mg alloys via grain boundary strengthening and dispersion strengthening

[15,

16].

The

metastable

phase



precipitates,

formed

during

post-deformation ageing, further improves the mechanical properties by suppressing the basal slip via precipitation strengthening [9]. Texture significantly influences the mechanical performance of wrought Mg alloys. The strong basal texture of the conventional Mg extrusions, such as AZ31 or AZ61, provides a poor orientation for basal slip and twinning in tension along the extrusion direction (ED), leading to high strength but low ductility [20-22]. Gd and/or Y additions lead to the formation of <  > RE texture and weaken the basal texture [23, 24]. The <  > oriented grains are beneficial for basal slip, consequently resulting in improved ductility. Homma et al [14] reported a high-strength and ductile Mg-10Gd-5.7Y-1.6Zn-0.5Zr (wt %) alloy prepared by hot extrusion and ageing. This alloy exhibited a tensile yield strength (TYS) of 473 MPa and an elongation (EL) of 8.0 %. Hou et al [25] developed an as-extruded Mg-8Gd-2Y-1Nd-0.3Zn-0.6Zr (wt %) alloy with a high dispersion of second-phase precipitates exhibiting a high TYS of 270 MPa and an EL of 14.2

%.

Yu

et

al

[17,

18]

successfully

fabricated

a

high-strength

Mg-11Gd-4.5Y-1Nd-1.5Zn-0.5Zr (wt %) alloy having a TYS of 502 MPa and an EL of 2.6 % at room temperature, and a TYS of 316 MPa and an EL of 6.3 % at 250 °C. The excellent mechanical performance is attributed to the presence of the LPSO and Mg5RE phases. The previous investigations have demonstrated that Mg-Gd-Y based alloys with Nd and/or Zn additions could be a promising candidate for producing high-strength and heat resistance Mg 4

alloys. Recently, it was reported that the existence of extensive second-phase particles before hot extrusion can significantly improve the DRX in subsequent hot extrusion process, and give rise to the grain refinement [26-29]. As a result, a considerable improvement in strength and ductility can be achieved. The influence of second-phase particles on the DRX largely depends on their size, spacing and fraction [29]. Robson et al [27] revealed that the second-phase particles (> 1 μm) contribute to the DRX as they lead to the formation of new grains by particle simulated nucleation (PSN) during hot deformation. Xu et al [28] observed that the decrease in spacing between second-phase particles assists the DRX and increases the DRX fraction. Note that this influence also depends on the features of the second-phase particles. For instance, in the Mg-Gd-Y based alloys the Mg5RE phase assists the DRX, whereas the LPSO phase resists it [16, 30]. After hot extrusion the different microstructures can be obtained in these alloys, accordingly, their mechanical performances would be very different. However, the relevant investigations have been rarely reported. Although it has been well known that both the Mg5RE and LPSO phases contribute to the alloy strengthening, their effects on the DRX, post-deformation microstructure, and subsequent mechanical performance still need to be further investigated. In the present work, we carried out a comparative study on the high-strength Mg-11.5Gd-4.5Y-0.3Zr (wt.%) alloy with and without a small amount of Zn and Nd additions. Mg-11.5Gd-4.5Y-0.3Zr (wt.%) alloy was selected as a benchmark in this comparative study with the Mg-11.5Gd-4.5Y-1Nd-0.3Zr (wt.%) and Mg-11.5Gd-4.5Y-1.5Zn-0.3Zr (wt.%) alloys, with the aim of gaining an in-depth understanding of the effects of the LPSO and Mg5RE 5

phases on the microstructure and mechanical properties of this alloy. The purpose of Zn and Nd additions in the Mg-11.5Gd-4.5Y-0.3Zr (wt.%) alloy is to obtain extensive LPSO and Mg5RE phases.

2. Experimental procedure The

alloys

with

designed

compositions

of

Mg-11.5Gd-4.5Y-0.3Zr

(wt.%),

Mg-11.5Gd-4.5Y-1Nd-0.3Zr (wt.%) and Mg-11.5Gd-4.5Y-1.5Zn-0.3Zr (wt.%) were melted in an electrical furnace at 760 ± 3 °C under a SF6/CO2 (1:99) atmosphere. After the melts were cooled to 720 ± 3 °C, they were cast into a steel mould with a diameter of 92 mm. The mould was preheated to 200 ± 3°C before casting. The chemical compositions of these alloys were analyzed using the inductively coupled plasma (ICP) spectrometer (Thermo Scientific iCAP6300, China). The results were summarized in Table 1. The cylindrical billets were machined into round bars with diameters of 43 mm and lengths of 38 mm. The round bars were then homogenized at 520 ± 3 °C for 48 h (T4 treatment) followed by quenching in water at 25 °C. The homogenized bars were subject to preheat treatments at 450 ± 3 °C for 8 min to reach the extrusion temperature. The hot extrusion was performed at 450 ± 3°C with an extrusion ratio of 20:1 at a ram speed of 0.1 mm s-1. After hot extrusion was complete, ageing treatment (T5 treatment) was conducted at 200 ± 3 °. Tensile specimens (gauge dimensions: Ф 4 mm×20 mm) were tested at room temperature along the ED with a speed of 0.001 s-1 using a Shimadzu Autograph AG-1 (50 kN) testing machine. The microstructural observations were conducted by optical microscope (OM) using a Olympus BX51M light optical microscope, field-emission scanning electron microscopy (FE-SEM) using a JEOL JSM-7000F SEM equipped with TSL electron back scattered diffraction (EBSD) operating at 6

15kV, and transmission electron microscopy (TEM) using a JEOL JEM-2100F TEM operating at 200 kV. The specimens for the OM and SEM observations were ground using silicon carbide emery paper up to 4000 grit, and then polished with 3-micron diamond and finally with 0.05-micron silica solution (Struers OP-S suspension). The TEM foils were prepared by mechanical grinding to a thickness of approximately 60 µm and then ion-polished to perforation using a Gatan Precision Ion Polishing System (GATAN691). X-ray diffraction (XRD) measurements were obtained for the phase analysis using a diffractometer (Rigaku MiniFlex600, 40 kV and 15 mA) with Cu Kα radiation and a secondary monochromator (step rate of 0.03 ° s-1, dwell time 3 s). The SEM-EBSD measurements (15 keV, step size 0.3 μm) were used to analyze the local microstructure and texture. To ensure statistical rigor, more than 1000 grains were examined for each condition. In this study, the Mg-11.5Gd-4.5Y-0.3Zr (wt.%), Mg-11.5Gd-4.5Y-1Nd-0.3Zr (wt.%) and Mg-11.5Gd-4.5Y-1.5Zn-0.3Zr (wt.%) alloys were designated as GW, GWN and GWZ, respectively. The as-cast and homogenized samples were designated as as-cast and T4, respectively. The extruded samples without or with ageing treatment were designated as EX and T5, respectively. The average grain size of the DRXed grains was measured using linear intercept method.

3. Results 3.1. Microstructural characterizations 3.1.1. Microstructure before hot extrusion

The microstructures of as-cast GWZ, GWN and GW alloys were observed using OM 7

(Fig. 1). The OM micrographs show that the as-cast alloys consisted mainly of dendritic grains and eutectic phase distributing along the dendrite arms. In addition to the eutectic phase, the addition of Zn led to the formation of lamellar LPSO phase in the as-cast GWZ alloy (Fig. 1a and d). After T4 treatment, the dendritic grains were replaced by the equiaxed grains (Fig 2a-f). The eutectic phase was resolved into the α-Mg matrix, whereas the lamellar LPSO phase was retained at the grain boundaries (Fig. 2g). A few Zr-rich phase particles and cuboid phase particles can be observed in these alloys (Fig. 2g-i). The Zr-rich phase particles appeared in a clustered matter as indicated by circles in Fig. 2d-f. Moreover, a few eutectic phase particles were still retained in the GWN alloy (Fig. 2h). After T4 treatment, the average grain sizes were 48.7f2.9 μm, 73.2f5.0 μm and 112.8f7.3 μm for the GWZ, GWN and GW alloys, respectively. To analyze the chemical compositions of the second-phases, energy-dispersive X-ray spectroscopy (EDS) elemental point analysis was conducted with these T4-treated alloys. The typical second-phase particles (indicated by arrows in Fig. 2g-i) were designated as P1-P8, respectively. The SEM-EDS results show that the Zr-rich phase (P1, P6 and P8) had a Zr content in the range of 15 - 17 at % (Table 2). The cuboid phase (P2, P5 and P7) was enriched of Gd and Y, resulting in a stoichiometric ratio of Mg to RE (Gd, Y) of 1:1.5, which is close to that of MgRE2 phase. This cuboid phase was previously reported in the similar Mg-Gd-Y based alloys [31, 32], and its formation was a result of the dissolution of Mg5RE phase [33]. In this dissolution procedure, the network Mg5RE phase was transformed to a spheroidized Mg5RE phase and a cuboid Mg5RE phase. Finally, the former was completely dissolved, while the later was remained due to its excellent thermostability. The chemical compositions 8

of

the

lamellar

LPSO

phase

(P3)

and

the

eutectic

phase

(P4)

were

Mg84.66Gd7.98Y3.17Zn4.19Zr0.08 (at %) and Mg77.99Gd14.85Y2.39Nd4.77Zr0 (at %)ˈ respectively. Phase identification was further conducted by XRD, revealing that the identical phase constitution included Mg and Mg5RE phases in the as-cast GW and GWN alloys (Fig. 3a). The GWZ alloy contained these phases in addition to the Mg 12REZn and Mg3RE phases. After T4 treatment, the Mg12REZn phase can still be identified through the XRD pattern, whereas the Mg5RE and Mg3RE phases were not evident in the XRD pattern (Fig. 3b). This result is consistent with the microstructure observations. It indicated that the eutectic phases were resolved into the α-Mg matrix while the lamellar LPSO phase still remained at the grain boundaries. The XRD patterns of GW and GWN alloys demonstrated the dissolution of eutectic Mg5RE phase during the T4 treatment. Considering the SEM-EDS result, the residual eutectic phase (P4) in the T4-treated GWN alloy can be identified as Mg5RE phase. The microstructures of these alloys after preheating to the extrusion temperature were observed using OM and SEM (Fig. 4). The OM micrographs show that a few particles formed at the grain boundaries during the short-term preheat treatment. The GWN alloy exhibited enhanced second-phase precipitation in comparison to the GWZ and GW alloys as a large number of plate-like phase particles precipitated inside the grains (Fig. 4b). The plate-like phase particles densely distributed in the region containing the clustered Zr-rich phase particles. Fig. 4d-f show the SEM micrographs of these newly formed particles at a higher magnification. The typical particles were designated as P9-P12, respectively, and detected by SEM-EDS. The SEM-EDS results show that the particles at grain boundaries had a similar chemical composition, which was (Mg, Zn, Zr)89.26Gd8.82Y1.92 (at %) for P9, (Mg, 9

Zr)88.53Gd8.19Y1.59Nd1.69 (at %) for P10 and (Mg, Zr)89.36Gd8.10Y2.54 (at %) for P12. These particles at grain boundaries can be identified as Mg5RE phase since its characteristic peaks reappeared in the XRD patterns after the preheat treatments (Fig. 3c). The plate-like particle (P11) had a chemical composition of (Mg, Zr)91.14Gd6.41Y0.92Nd1.53 (at %). Our previous investigation revealed that these plate-like phase particles were the Mg5RE phase [16]. Due to their small size, the SEM-EDS results were influenced by the α-Mg matrix, thus the plate-like particles had a higher Mg content over 91 at %. As for the GWN alloy, the addition of Nd effectively promoted the precipitation of Mg5RE phase. Accordingly, the characteristic peaks of Mg5RE phase were more obvious in the XRD pattern of GWN alloy (Fig. 3c).

3.1.2. Microstructure after hot extrusion

The microstructures of the studied alloys after hot extrusion were observed using SEM and EBSD inverse pole figure (IPF) maps (Fig. 5). These alloys exhibited a bimodal microstructure consisting of fine DRXed grains with several microns and coarse un-DRXed regions (Fig. 5a-c). Several fine particles were formed at the DRXed grain boundaries by dynamic precipitation (upper right corners of Fig. 5a-c). The typical fine particles are shown in the TEM micrographs (Fig. 6a-c). According to the SAED patterns recorded from the particle P13, P15 and P16, these fine particles can be identified as the Mg5RE phase (F P , a = 2.2 nm) [9]. In the GWZ alloy, in addition to the Mg5RE phase, the lamellar LPSO phase with a large aspect ratio was aligned parallel to the ED (Fig. 5a). Moreover, a few thin plate-shaped LPSO phase were observed in the DRXed grains. In the GWN alloy, the number

10

of Mg5RE phase was significantly increased. Several bulk particles appeared in a form of “strings” (Fig. 5b). The SAED pattern recorded from a typical bulk particle (P14) revealed that they are also the Mg5RE phase (Fig. 6b). The EBSD IPF maps of these alloys are shown in Fig. 5d-f. Note that the high angle grain boundaries (HAGBs, > 15°) and the low angle grain boundaries (LAGBs, <15°) were outlined by black and white lines, respectively. A comparison between Fig. 5a and d suggested that the large black regions are lamellar LPSO phase particles. In the un-DRXed regions, several LAGBs were formed and gave rise to the subgrain structures (Fig. 5g-i). In the GWZ alloy, the average DRXed grain size was calculated as 1.6 ± 0.5 µm. Compared with the GWZ alloy, the GWN alloy had a reduction in the un-DRXed regions. The average DRXed grain size was 1.8 ± 0.6 µm, which is comparable to that in the GWZ alloy. In the GW alloy, the average DRXed grain size was 2.5 ± 0.8 µm, which is slightly larger than that in the GWN and GWZ alloys. The un-DRXed regions in the GW alloy was also reduced, but the decrement was still less than that in the GWN alloy. The DRX fraction was 89.4 %, 93.0% and 84.0% for the GW, GWN and GWZ alloys, respectively. The textures were presented using IPFs obtained by the EBSD measurements (Fig. 7). These as-extruded alloys exhibited a basal texture, i.e., most grains were oriented with the [  ] axis parallel to the ED. The decrease in the maximum texture intensity for the GWN alloy could be related to the increase in the DRX fraction since the un-DRXed regions exhibited a very strong basal texture.

3.1.3. Recrystallization behaviors

To seek the dominate recrystallization mechanism during hot extrusion, a comparative

11

microstructural analysis, at the same location: 4 mm (partially deformed area) away from die exit, was performed using SEM and EBSD for the GWZ, GWN, and GW alloys (Fig. 8). In the GWZ alloy, the lamellar LPSO phase was deformed to accommodate the deformation of parent grains (Fig. 8a). The EBSD IPF map shows that the dynamic recrystallization (DRX) primarily took place at the prior grain boundaries (P-GBs) and gradually propagated into the grain mantle regions (Fig. 8b). The formation of new grains at the P-GBs was indicated by the black square. In contrast to the P-GBs, the DRX in the vicinity of the lamellar LPSO phase was insufficient. In the GWN alloy, several particle strings can be observed were in the matrix in addition to the fine Mg5RE particles (Fig. 8c). These particle strings were formed by cracking of plate-like Mg5RE particles during the hot extrusion. As a result, they consisted of several fragments (bulk Mg5RE particles) with similar sizes that were aligned one after another. Compared with the GWZ alloy, the GWN alloy had a raised DRX fraction. At the P-GBs a few new grains can still be observed as indicated by the black square in Fig. 8d. The DRX gradually propagated into the grain interiors, and the extensive DRX occurred in the regions with densely distributed bulk Mg5RE particles. It should be noted that the uneven distribution of bulk Mg5RE particles gave rise to a bimodal microstructure consisting of fine DRXed grains and coarse un-DRXed regions. In the region free of bulk Mg5RE particles the incomplete DRX resulted in a subgrain structure as indicated by the circles in Fig. 8c and d. In the GW alloy, the DRX was less pronounced in comparison to the GWN alloy (Fig. 8e). The new grains mainly formed at the P-GBs as indicated by the black square, while a few grains formed in the grain mantle regions (Fig. 8f).

3.2. Mechanical properties 12

Fig. 9 shows the age-hardening curves for the as-extruded GWZ, GWN and GW alloys. The GW alloy exhibited a strong ageing-hardening response. The hardness reached a peak, 130.7 HV, at 48h. The GWZ and GWN alloys reached their peak hardness of 128.2 HV and 125.6 HV at 48 h and 24 h, respectively. The addition of Zn to the GW alloy slightly decreased the peak hardness but had a no influence on the peak ageing time. The decrease in the peak hardness is due to the lowered number density of the b ¢ phase by the addition of Zn [12]. With the addition of Nd to the GW alloy, although the peak ageing time became shorter, the peak hardness was further decreased. This result is in contradiction with the previous findings, such that the ageing-hardening response can be enhanced significantly by the addition of Nd, in the similar Mg-RE based alloys [7, 34]. Fig. 10 shows the typical tensile stress-strain curves for the as-extruded alloys with and without T5 treatment. The corresponding tensile properties are summarized in Table 3. In tension, the GWZ alloy exhibited the highest ultimate tensile strength (UTS) and TYS. However, its EL was slightly lower than that of the GW alloy. Compared with the GWZ and GW alloys, the GWN alloy had a poor mechanical performance. Specifically, the UTS, TYS and EL were 424 ± 2.4 MPa, 371 ± 3.0 MPa, and 7.2 ± 0.8 %, respectively, for the GWZ alloy, and 379 ± 2.6 MPa, 347 ± 3.9 MPa, and 3.3 ± 0.5 %, respectively, for the GWN alloy, and 402 ± 10 MPa, 336 ± 6.1 MPa, and 7.6 ± 1.7 %, respectively, for the GW alloy. The T5 treatment effectively strengthened these alloys but decreased their ductility. The UTS and TYS were increased to 480 ± 0.4 MPa and 425 ± 2.5 MPa for the GWZ alloy, to 416 ± 5.1 MPa and 402 ± 7.2 MPa for the GWN alloy and to 483 ± 5.5 MPa and 423 ± 9.0 MPa for the GW alloy, respectively. Their EL was reduced to 3.7 ± 0.3 %, 1.4 ± 0.1 % and 2.2 ± 0.1 %, respectively. In these 13

alloys, the GWZ alloy exhibited an excellent balance of strength and ductility.

4. Discussion 4.1 Recrystallization mechanism The dominate recrystallization mechanisms for the studied alloys are discussed on the basis of the microstructure observation of the partially deformed samples, Fig. 8. In the GWZ alloy, new grains are formed at the P-GBs by the grain boundary bulging as indicated by the black square, Fig. 8b. It is evident that discontinuous DRX (D-DRX) is involved in this microstructure evolution and forms a necklace structure consisting of fine DRXed grains around the parent grains [29]. In contrast to the P-GBs, the DRX in the vicinity of the lamellar LPSO phase is insufficient. It is indicating that the lamellar LPSO phase hinders the DRX. During the hot extrusion, the stress concentration may occur at Mg/LPSO interfaces [30]. The LPSO phase can be deformed by kink bands to accommodate the deformation of parent grains and release the stress concentration [35]. The stable Mg/LPSO interfaces are unfavourable for the nucleation of new grains [35]. In the GWN alloy, a large number of bulk Mg5RE particles are observed in the matrix in addition to the fine Mg5RE particles (Fig. 8c). Compared with GW and GWZ alloys, the DRX behaviour of GWN alloy is enhanced. Although a few new grains form via the D-DRX mechanism at the P-GBs and the particles simulated nucleation at the bulk Mg5RE particles, most grains were formed by continuous dynamic recrystallization (C-DRX) [16, 27]. The bulk Mg5RE particles promotes the subdivision of parent grains during the hot extrusion, Fig. 8c and d. The resultant sub-structures provide favourite nucleation sites for the DRXed grains, and the rapid transformation of LAGBs to HAGBs leads to the formation of DRXed grains by the continuous consumption of sub-grains [36, 37]. 14

The DRX is extensive in the regions with densely distributed bulk Mg5RE particles, whereas it is insufficient in the regions lack of bulk Mg5RE particles. It is indicating that bulk Mg5RE particles facilitate the C-DRX by promoting the grain subdivision during the hot extrusion. In the GW alloy, the DRX was less pronounced in comparison to the GWN alloy due to the lack of bulk Mg5RE particles during hot extrusion (Fig. 8e). The new grains mainly form via the D-DRX mechanism at the P-GBs, while a few new grains form via the C-DRX mechanism in the grain mantle regions (Fig. 8f).

4.2 Microstructures and strengthening The strengthening of the Mg alloys is primarily attributed to grain refinement, texture, and second phase particles [14, 38, 39]. The studied alloys exhibited a bimodal structure consisting of fine DRXed grains of several microns (1.6 – 2.5 μm) and coarse un-DRXed grains. According to Hall-Petch relation, these fine DRXed grains improve the TYS to some extent via grain boundaries strengthening [40]. The texture should be another contributor to the strengthening of the studied alloys. A basal texture with < 1010 > components parallel to the ED was observed after hot extrusion, revealing that most grains had the c-axis perpendicular to the ED (Fig. 7). These textured grains, especially for the un-DRXed grains having a strong basal texture, effectively raise the TYS by providing a hard crystallographic orientation for the basal slip and tensile twinning under tensile loading along the ED [41, 42]. In addition, the sub-grain structures in the un-DRXed grains also serve as effective barriers to the dislocations motion during tensile test [43]. The existence of un-DRXed grains raises the TYS, but reduces the EL. In contrast to the un-DRXed grains, the DRXed grains with relatively random orientations facilitate the basal 15

slip in tension. They can release the strain localization transferred from the un-DRXed grains and hence promote the homogeneous plastic deformation during tensile test [19, 44]. The Mg5RE and LPSO phase particles also contributes to the strengthening of the studied alloys. The densely distributed Mg5RE particles at grain boundaries could improve the mechanical properties via dispersion strengthening [18, 39]. The lamellar LPSO phase particles with a high aspect ratio are aligned parallel to the ED (Fig. 5a). Hagihara et al. [13, 45] reported that the (0001)<11 2 0> basal slip is the dominant deformation mode in the LPSO phase. The lamellar LPSO phase have the layered interface parallel to the (0001)LPSO and to the ED, where is in hard orientation for the basal slip under tensile loading. The aligned lamellar LPSO phase, acts as a reinforcement, strengthens the alloy via short-fibre strengthening [13]. In addition, the stable coherent interface between the LPSO phase and Mg grains are not favourable for the nucleation of defects during deformation [35]. Thus, the lamellar LPSO phase can effectively improve the strength and ductility. In addition to the lamellar LPSO phase, a few thin plate-shaped LPSO phase were observed in the DRXed grains (Fig. 6a). It was reported that the thin plate-shaped LPSO phase precipitates on the basal plane of α-Mg matrix, and it exhibits a slight impediment to basal slip due to the fact that the plates on the basal plane provide insufficient barriers to the movement of basal dislocations [9, 12, 19]. The thin plate-shaped LPSO phase can effectively impede the movement of non-basal dislocations and further improve the strength. Solid solution strengthening also contributes to the alloy strengthening. Alloying Mg with solutes is an effective approach to improve its mechanical properties, because the solute atoms could bring hardening effects to deformation systems [46-48]. These effects are largely 16

determined by the size of solid solutes and misfits with respect to atomic radius. The atomic radius is 0.180 nm for Gd, 0.212 nm for Y, 0.206 nm for Nd, 0.133 nm for Zn and 0.160 nm for Mg, respectively. The substitution of Mg atom with a RE atom leads to a positive misfit and a compression strain, while the substitution with a Zn atom leads to negative misfit and a tensile strain [46, 48]. In the α-Mg matrix, these misfits cause the lattice distortion and act as barriers to the dislocation movement. Thus, the RE and Zn additions strengthen the alloys by the solid solution strengthening. After T5 treatment, the strength was significantly increased,

whereas the ductility was reduced (Table 3). The strengthening of the T5-treated alloys is attributed to the nano-scale b ¢ phase (see the SAED pattern at upper right corner of Fig. 6a), which precipitates on the prismatic plane of Mg matrix making a sufficient resistance to the basal slip [9]. The orientation relationship between b ¢ phase and Mg matrix is (001) ! // &&&&0]"#$% . These alloys are strengthened via precipitation (0001)"#$% and [001] ! //[211 strengthening mechanism [9]. In the present study, the GWZ, GWN and GW alloys exhibit different mechanical performances. Compared with the GW alloy, the GWZ alloy exhibited a raised UTS and TYS but a lowered EL. The GWN alloy exhibited the worst mechanical performance among the studied alloys. The different mechanical behaviours are attributed to the differences in their microstructures. The hot extrusion gives rise to various DRX fractions, which are 89.4 %, 93.0% and 84.0%, respectively, for the GW, GWN and GWZ alloys. The effect of the coarse un-DRXed grains on the mechanical properties are apparently different, such that it is weakened as the DRX fraction increases in the GW alloy, but enhanced as the DRX fraction decreases in the GWZ alloy. Since the un-DRXed grains improve the TYS but degrade the EL, 17

the GWZ alloy exhibits a raised TYS and a lowered EL in comparison to the GW alloy. Although GWN alloy has the highest DRX fraction, it exhibits the lowest EL in the studied alloys. The poor ductility should be attributed to another influence factor, i.e. the extensive bulk Mg5RE particles. The additions of Nd and Zn in GW alloy lead to differences in the microstructure with respect to the second-phase particles. In the GWN alloy, the addition of Nd significantly promoted the precipitation of plate-like Mg5RE particles, Fig. 4b. These plate-like Mg5RE particles were cracked into the bulk Mg5RE particles during hot extrusion, Fig. 8c. The extensive bulk Mg5RE particles not only assist the DRX through the PSN mechanism, but also bring about effective pinning on the grain boundaries and strengthen the studied alloy via the dispersion strengthening mechanism [18, 39, 40]. However, the existence of bulk Mg5RE particle causes the crack initiation and provides a path for crack propagations at grain boundaries, resulting in a decrease in ductility [16]. Compared with the GW and GWZ alloys, the GWN alloy has a raised number of bulk Mg5RE particles. The extensive bulk Mg5RE particles degrade the ductility of the GWN alloy. In the GWZ alloy, the number of bulk Mg5RE particles was less than that in the GWN alloy, resulting in a less degradation of the ductility. In addition, the formation of LPSO phase can hinder the propagation of the cracks, so that an increase in the ductility can be expected [35, 49]. Although the GWZ alloy has a comparable grain size to the GWN alloy, it exhibits a superior ductility to the GWN alloy due to the existence of LPSO phase and the lack of bulk Mg5RE particles. After hot extrusion the ageing-hardening responses are different in these alloys. Compared with the GWZ and GW alloys, the GWN alloy had a weaker ageing-hardening response, resulting in a relatively low 18

TYS. In the contrast, the GWZ and GW alloys are comparable in the TYS due to the similar ageing-hardening response. It is noteworthy that the GWZ alloy exhibits an excellent balance of strength and ductility (TYS of 425 ± 2.5 MPa and EL of 3.7±0.3 %) after T5 treatment.

5. Conclusions This study investigates on the microstructure evolution and mechanical properties of as-extruded Mg-11.5Gd-4.5Y-0.3Zr (wt.%) alloy with Zn and Nd additions. The results show that the addition of Zn inhibits the dynamic recrystallization due to the presence of the LPSO phase. The addition of Nd promotes the precipitation of the Mg5RE phase. The existence of the densely distributed Mg5RE phase before hot extrusion promotes the dynamic recrystallization in subsequent hot extrusion process and leads to grain refinement. The increase in the number of Mg5RE phase particles degrades the mechanical properties of the resultant alloy. Following extrusion, the studied alloys exhibit a bimodal microstructure consisting of fine DRXed grains of several microns and strongly textured course un-DRXed grains. The as-extruded Mg-11.5Gd-4.5Y-1.5Zn-0.3Zr alloy exhibits an excellent balance of strength and ductility (TYS of 371±3.0 MPa and EL of 7.2±0.8 %). The alloy strengthening is attributed to the bimodal microstructure, the Mg5RE and LPSO phases, and the basal texture. The TYS of the as-extruded Mg-11.5Gd-4.5Y-1.5Zn-0.3Zr alloy can be further increased to 425±2.5 MPa by precipitation hardening with the T5 treatment.

Acknowledgements Zijian Yu acknowledges financial support from the Japan Society for the Promotion of Science (JSPS) in the form of a JSPS post-doctoral fellowship. This work is partially supported

by

the

National

Key

Technologies 19

R&D

Program

(2012BAE01B04,

2012DFH50100, KGFZD-125-13-021, 201001C0104669453).

References [1] B.L. Mordike, T. Ebert, Mater. Sci. Eng.: A 302 (2001) 37-45. [2] B.L. Mordike, Mater. Sci. Eng.: A 324 (2002) 103-112. [3] L. Lin, L. Chen, Z. Liu, J. Mater. Sci. 43 (2008) 4493. [4] J.W. Lu, D.D. Yin, L.B. Ren, G.F. Quan, J. Mater. Sci. 51 (2016) 10464-10477. [5] L.L.Rokhlin, London and New York :Taylor & Francis Ltd, (2003) 1-256. [6] T.B. Massalski, H. Okamoto, Binary alloy phase diagrams, ASM International, Materials Park, Ohio, 1990. [7] Q. Peng, Y. Wu, D. Fang, J. Meng, L. Wang, J. Mater. Sci. 42 (2007) 3908-3913. [8] F. Penghuai, P. Liming, J. Haiyan, M. Lan, Z. Chunquan, Mater. Sci. Eng.: A 496 (2008) 177-188. [9] J.F. Nie, MMTA, 43 (2012) 3891-3939. [10] Y. Negishi, T. Nishimura, M. Kiryuu, S. Kamado, Y. Kojima, R. Ninomiya, Journal of Japan Institute of Light Metals, 45 (1995) 57-63. [11] X. Gao, J.F. Nie, Scr. Mater. 58 (2008) 619-622. [12] T. Honma, T. Ohkubo, S. Kamado, K. Hono, Acta Mater. 55 (2007) 4137-4150. [13] K. Hagihara, A. Kinoshita, Y. Sugino, M. Yamasaki, Y. Kawamura, H.Y. Yasuda, Y. Umakoshi, Acta Mater. 58 (2010) 6282-6293. [14] T. Homma, N. Kunito, S. Kamado, Scr. Mater. 61 (2009) 644-647. [15] A. Sadeghi, M. Pekguleryuz, J. Mater. Sci. 47 (2012) 5374-5384. [16] Z. Yu, Y. Huang, C.L. Mendis, N. Hort, J. Meng, Mater. Sci. Eng.: A 624 (2015) 23-31. [17] Z. Yu, Y. Huang, X. Qiu, G. Wang, F. Meng, N. Hort, J. Meng, Mater. Sci. Eng.: A 622 (2015) 121-130. [18] Z.J. Yu, Y. Huang, X. Qiu, Q. Yang, W. Sun, Z. Tian, D.P. Zhang, J. Meng, Mater. Sci. Eng.: A 578 (2013) 346-353. [19] C. Xu, T. Nakata, X. Qiao, M. Zheng, K. Wu, S. Kamado, Sci. Rep. 7 (2017) 40846. [20] M.T. Pérez-Prado, O.A. Ruano, Scr. Mater. 48 (2003) 59-64. [21] M.T. Pérez-Prado, O.A. Ruano, Scr. Mater. 46 (2002) 149-155. [22] J. Bohlen, S.B. Yi, J. Swiostek, D. Letzig, H.G. Brokmeier, K.U. Kainer, Scr. Mater. 53 (2005) 259-264. [23] N. Stanford, D. Atwell, A. Beer, C. Davies, M.R. Barnett, Scr. Mater. 59 (2008) 772-775. [24] N. Stanford, Mater. Sci. Eng.: A 527 (2010) 2669-2677. [25] X.L. Hou, Q.M. Peng, Z.Y. Cao, S.W. Xu, S. Kamado, L.D. Wang, Y.M. Wu, L.M. Wang, Mater. Sci. Eng.: A 520 (2009) 162-167. [26] D. Yu, D. Zhang, J. Sun, Y. Luo, J. Xu, H. Zhang, F. Pan, J. Alloy. Compd. 690 (2017) 553-560. [27] J.D. Robson, D.T. Henry, B. Davis, Acta Mater. 57 (2009) 2739-2747. [28] S.W. Xu, K. Oh-ishi, S. Kamado, T. Homma, Scr. Mater. 65 (2011) 875-878. [29] F.J. Humphreys, M. Hatherly, Pergramon Press, Oxford, (1995). [30] J. Yu, Z. Zhang, Q. Wang, X. Yin, J. Cui, H. Qi, J. Alloy. Compd. 704 (2017) 382-389. [31] R. Alizadeh, R. Mahmudi, A.H.W. Ngan, T.G. Langdon, Adv. Eng. Mater. 18 (2016) 1044-1049. [32] Z. Yang, Z.H. Wang, H.B. Duan, Y.C. Guo, P.H. Gao, J.P. Li, Mater. Sci. Eng.: A 631 (2015) 160-165. [33] Y. Gao, Q. Wang, J. Gu, Y. Zhao, Y. Tong, Mater. Sci. Eng.: A 459 (2007) 117-123. 20

[34] X. Zhang, L. Meng, C. Fang, P. Peng, F. Ja, H. Hao, Mater. Sci. Eng.: A 586 (2013) 19-24. [35] X.H. Shao, Z.Q. Yang, X.L. Ma, Acta Mater. 58 (2010) 4760-4771. [36] Z. Yu, C. Xu, J. Meng, S. Kamado, Mater. Sci. Eng.: A 703 (2017) 348-358. [37] Z. Yu, C. Xu, J. Meng, X. Zhang, S. Kamado, J. Alloy. Compd. 729 (2017) 627-637. [38] J. Bohlen, M.R. Nürnberg, J.W. Senn, D. Letzig, S.R. Agnew, Acta Mater. 55 (2007) 2101-2112. [39] S.W. Xu, K. Oh-ishi, S. Kamado, F. Uchida, T. Homma, K. Hono, Scr. Mater. 65 (2011) 269-272. [40] K. Oh-ishi, C.L. Mendis, T. Homma, S. Kamado, T. Ohkubo, K. Hono, Acta Mater. 57 (2009) 5593-5604. [41] M.R. Barnett, Mater. Sci. Eng.: A 464 (2007) 1-7. [42] S.H. Park, S.-G. Hong, C.S. Lee, Scr. Mater. 62 (2010) 202-205. [43] R. Sedláček, W. Blum, J. Kratochvíl, S. Forest, MMTA, 33 (2002) 319-327. [44] H. Wu, G. Fan, M. Huang, L. Geng, X. Cui, H. Xie, IJP, 89 (2017) 96-109. [45] K. Hagihara, N. Yokotani, Y. Umakoshi, Intermetallics, 18 (2010) 267-276. [46] Z.R. Liu, D.Y. Li, Acta Mater. 89 (2015) 225-233. [47] N. Stanford, R. Cottam, B. Davis, J. Robson, Acta Mater. 78 (2014) 1-13. [48] S. Sandlöbes, Z. Pei, M. Friák, L.F. Zhu, F. Wang, S. Zaefferer, D. Raabe, J. Neugebauer, Acta Mater. 70 (2014) 92-104. [49] K. Wen, K. Liu, Z. Wang, S. Li, W. Du, Mater. Sci. Eng.: A 674 (2016) 33-39.

21

Figure Captions Fig. 1. OM images of as-cast alloys: (a and d) GWZ alloy, (b and e) GWN alloy, and (c and f) GW alloy. Fig. 2. OM and SEM images of as-cast alloys after the T4 treatment: (a, d and g) GWZ alloy, (b, e and h) GWN alloy, and (c, f and i) GW alloy. Fig. 3. X-ray diffraction patterns of the studied alloys under different states: (b) T4-treated state and (c) preheated state.

(a) as-cast state,

Fig. 4. OM and SEM images of the studied alloys after the preheat treatment: (a and d) GWZ alloy, (b and e) GWN alloy, and (c and f) GW alloy.

Fig. 5. SEM images and corresponding EBSD IPF maps of the studied alloys after the hot extrusion (a, d and g) GWZ alloy, (b, e and h) GWN alloy, and (c, f and i) GW alloy.

Fig. 6. TEM images and corresponding SEAD patterns of Mg5RE phase in the studied alloys: after the hot extrusion and the T5 treatment: (a) GWZ alloy, (b) GWN alloy, and (c) GW alloy.

Fig. 7. IPF of the studied alloys after the hot extrusion: (a and d) GWZ alloy, (b and e) GWN alloy, and (c and f) GW alloy.

Fig. 8. SEM images and corresponding EBSD IPF maps of the studied alloys after the hot extrusion: (a and b) GWZ alloy, (c and d) GWN alloy, and (e and f) GW alloy. Note that the observed areas were selected at the same location: 4 mm away from die exit. Fig. 9. Ageing-hardening curves of the as-extruded alloys. Fig. 10. Tensile stress-strain curves of the as-extruded alloys with or without the T5 treatment.

22

Table 1 Summary of chemical compositions of the studied alloys

Alloys GWZ GWN GW

Nominal compositions

Chemical compositions (wt %) Gd Y Nd Zn Mg-11.5Gd-4.5Y-1.5Zn-0.3Zr 11.49 4.31 1.53 Mg-11.5Gd-4.5Y-1Nd-0.3Zr 11.43 4.22 1.16 Mg-11.5Gd-4.5Y-0.3Zr 11.71 4.90 -

Zr 0.38 0.33 0.31

Mg Bal. Bal. Bal.

Table 2 Summary of chemical compositions of various phases Alloys GWZ

GWN

GW

Particle P1 P2 P3 P4 P5 P6 P7 P8

Chemical compositions (at %) Gd Y Nd Zn 2.51 1.48 4.88 31.97 35.77 1.58 7.98 3.17 4.19 14.85 2.39 4.77 29.80 27.73 2.53 2.38 0.42 0 19.52 31.28 3.89 0.76 -

Phase Zr 15.08 0 0.08 0 1.47 16.11 3.46 17.08

Mg 76.05 30.68 84.66 77.99 38.47 81.09 45.74 78.27

Zr-rich phase Cuiboid phase LPSO phase Mg5RE phase Cuboid phase Zr-rich phase Cuboid phase Zr-rich phase

Table 3 Mechanical properties and hardness with average values and standard deviations. The as-extruded and T5-treated alloys were tested along ED at room temperature (25°C). State

Samples

TYS[MPa]

UTS [MPa]

EL [%]

Hardness [HV]

EX

GWZ GWN GW GWZ GWN

371±3.0 347±3.9 336±6.1 425±2.5

424±2.4 379±2.6 402±10 480±0.4

7.2±0.8 3.3±0.5 7.6±1.7 3.7±0.3

402±7.2 423±9.0

416±5.1 483±5.5

1.4±0.1 2.2±0.1

105±1.2 106±0.5 98±2.0 128±1.8 126±2.4

T5

GW

23

131±1.5

Highlights:

1. The Nd additions enhance the precipitation of Mg5RE phase and promote the DRX. 2. Extensive Mg5RE phase degrades the mechanical properties of the GWN alloy 3. The LPSO phase inhibits the dynamic recrystallization (DRX) of the GWZ alloy. 4. A GWZ alloy with UTS of 480±0.4 MPa and EL of 3.7±0.3% was produced.

24

Figure(s)

Figure(s)

Figure(s)

Figure(s)

Figure(s)

Figure(s)

Figure(s)

Figure(s)

Figure(s)

Figure(s)