I. INTRODUCTION
Magnesium alloys are the lightest structural metallic materials and they are attractive for applications in aerospace and automobile industries. Reference Antion, Donnadieu, Perrard, Deschampa, Tassin and Pisch1–Reference Li, Fu, Peng, Wang and Jiang9 However, with poor room temperature properties and low high temperature creep resistance and strength, the wider uses of conventional Mg alloys are limited. Improvements in the mechanical properties of pure Mg can be achieved through alloying and heat treatment. Owing to the solid solution and precipitation hardening, addition of rare earth (RE) elements such as Gd, Nd, Dy, Sm, and Y is an effective way to improve the strength of Mg alloys both at room and elevated temperatures. Reference Antion, Donnadieu, Perrard, Deschampa, Tassin and Pisch1–Reference Li, Fu, Peng, Wang and Jiang9 Among the Mg-RE alloys, Mg–Gd and Mg–Nd system alloys are widely studied and used because of their high performance. The Mg–Gd system alloys generally have a high content of heavy RE elements, which significantly increases their costs and limit their application. In contrast, with low content of RE elements, Mg–Nd–Zn alloys present high strength, good ductility, and creep resistance by the formation of β″ and β′ phases. Reference Fu, Peng, Jiang, Chang and Zhai3–Reference Sanaty-Zadeh, Luo and Stone7 In addition, these alloys are potential biodegradable material with suitable mechanical properties and good biocompatibility. Reference Qin, Zhao, An, Cheng, Wang, Cheng, Wang, Wang, Jiang, Zhang and Yuan8
Two traditional casting methods including permanent mold casting (PMC) and low-pressure sand mold casting (LSMC) were used to produce Mg–Nd–Zn alloys. However, the resulting PMC alloys contained a few porosities, and the grain size of the LSMC alloys was larger than that cast by a metal mold. Reference Fu, Peng, Jiang, Ma and Zhai6 In contrast, rheoforming is a promising casting process for forming near-net shaped components of high integrity and performance. Reference Flemings10,Reference Canyook, Wannasin, Wisuthmethangkul and Flemings11 Compared with traditional casting techniques, rheoforming can not only decrease the grain size but also significantly decrease or eliminate voids. For this forming technology, one key procedure is the preparation of semisolid slurries with small and nearly spherical primary particles surrounded by the liquid phase. Many external physical fields such as twin screw shearing, Reference Xia, Huang, Cassinath and Fan12 gas bubble stirring, Reference Wannasin, Canyook, Wisutmethangoon and Flemings13,Reference Zhang, Wu, Liu, Zhang, Pang and Ding14 ultra-sonic vibration, Reference Wan, Yan and Xu15 and electromagnetic stirring (EMS) Reference Lu, Wu, Wan and An16–Reference Chen, Zhang, Liu, Wu and Ding18 have been used to prepare these unique semisolid slurries. Among them, EMS is widely used and studies have demonstrated that EMS treatment is also suitable for preparing Mg-RE alloy semisolid slurries. Reference Wang, Chen, Zhang, Wu and Ding17,Reference Chen, Zhang, Liu, Wu and Ding18 After treated by EMS, both the average particle sizes of an Mg–Gd–Zn alloy Reference Wang, Chen, Zhang, Wu and Ding17 and Mg–Nd–Zn alloy Reference Chen, Zhang, Liu, Wu and Ding18 were effectively refined, and the non-dendritic semisolid slurries could be obtained.
The subsequent shaping of the semisolid slurry is another key procedure for the rheoforming process. High pressure die casting (HPDC) is the dominant manufacturing method adopted for rheoforming, but it is hard to completely eliminate porosities during rheo-diecasting because of the high rate mold filling. Alternatively, squeeze casting can compensate the shortcomings of HPDC owing to the slower mold filling and solidification under high pressure. In the conventional squeeze casting (CSC) process, the alloy melt is solidified under an applied pressure during solidification, resulting in a high cooling rate and temperature gradient. Thus, the final components produced by CSC generally have low density of porosities, heat treatability, and high performance. To further improve the mechanical properties of Mg alloys, many studies have been focused on the rheo-squeeze casting (RSC) technique which combines the advantages of rheoforming and CSC. For instance, Zhang et al. Reference Zhang, Wu, Liu, Zhang, Pang and Ding14 found that RSC combined with the gas bubbling process could refine the grain size and reduce porosity and thus effectively improved the mechanical properties of an AZ91-2 wt% Ca (AZX912) alloy. Fang et al. Reference Fang, Lü, Zhao, Wang, Liu and Wu19 reported that a high performance Mg–3RE–6Zn–1.4Y alloy with a fine and uniform microstructure could be obtained when the RSC process was properly applied by combining with the direct ultrasonic vibration process. Guo et al. Reference Guo, Zhang, Yang, Liu and Jin20 found that after the RSC process, a high integrity Mg–9Al–Zn alloy with a fine and compact microstructure could be obtained and the mechanical properties of this alloy were enhanced. The existing studies focused on the effect of applied pressure on the microstructure and mechanical properties of the Mg alloys prepared by the RSC process. The processing parameter of mold temperature also has large influence on the microstructure and scarce studies focused on the effect of mold temperature on the microstructure (e.g., the dislocation density introduced by plastic deformation, compactness of the microstructure, and the secondary solidification behavior) and mechanical properties of RSC Mg-RE alloys. Furthermore, it is expected that further heat treatment can significantly enhance the mechanical properties of the RSC Mg-RE alloys because of the strengthening of solid solution and precipitation. However, the related investigations about this were also very scarce.
In this work, the semisolid slurry of an NZ30K alloy was prepared by EMS and then formed by direct squeeze casting. The main purpose was to study the effect of mold temperature and subsequent heat treatment on microstructure and mechanical properties of RSC the NZ30K alloy and to investigate different strengthening contributions between the T6-treated CSC-NZ30K alloy and RSC-NZ30K alloy.
II. EXPERIMENTAL PROCEDURE
An alloy of nominal composition Mg–3Nd–0.2Zn–0.11Zr (NZ30K, wt%) was prepared by pure Mg and Zn ingots and Mg–90 wt% Nd and Mg–30 wt% Zr master alloys in an electrical resistance furnace under the protection of a mixed atmosphere of CO2 and SF6 with the ratio of 100:1. The pure Mg ingot was melted in the steel crucible, and then the pure Zn and Mg–Nd master alloys were added into the steel crucible when the melt temperature reached 680 °C and 730 °C, respectively, followed by manual stirring for 3 min. The melt was subsequently held at 780 °C for 30 min to allow the RE to completely dissolve. The preheated Mg–Zr master alloy was introduced to the melt at 760–780 °C. After it was purified by 1.5% JDMJ (refining agent with main compositions of 40∼50% MgCl2, 15∼25% KCl, 3∼5% BaCl2, 15∼20% NaCl, 5∼10% CaF2. 3∼5% BaCl2) flux (refining agent for Mg alloys), the melt was then held at 780 °C for 30 min before casting. The molten metal was poured into a steel mold pre-heated to 200 °C. The actual chemical compositions of the alloy (Mg–3.3Nd–0.21Zn–0.41Zr) were analyzed by Inductive Coupled Plasma Atomic Emission Spectroscopy (ICP-AES).
The semisolid slurry of the NZ30K alloy was prepared by EMS. More details can be found in Ref. Reference Chen, Zhang, Liu, Wu and Ding18. The stirring time, applied voltage, and rotational frequency were selected at 120 s, 350 V, and 20 Hz, respectively. To obtain the semisolid microstructure, samples with sizes of Φ10 × 50 mm3 were obtained from the 1/2 radius of the crucible and then quickly quenched into water at room temperature. After treated by EMS, the semisolid slurry was subsequently formed by the direct squeeze casing process at different mold temperatures (200, 250, 300, and 350 °C) under 80 MPa. For comparison, the NZ30K alloy was also prepared by the CSC technique at a pouring temperature of 660 °C.
Samples cut from the RSC and CSC components were solution-treated at 530 °C for 10 h and quenched, followed by aged at 200 °C for 16 h in an oil bath. Rectangular tensile specimens with dimensions of 10 mm width, 1.5 mm thickness, and 18 mm gauge length were prepared by an electric-sparking wire-cutting machine. Tensile testing was carried out on a Zwick/Roell Z100 (Zwick Roell Group, Ulm, Baden–Württemberg, Germany) material test machine. One cross-section of each sample was ground and polished. The as-cast and T6-treated samples were etched in a 4 vol% nital and the solution treated alloys were etched in a solution of 12 g picric acid +80 mL acetic acid +80 mL water +350 mL ethanol. Subsequently, the microstructure characterization was observed with an optical microscope (OM; ZEISS Axio Observer A1, ZEISS, Oberkochen, Germany), scanning electron microscope (SEM; Phenom XL, Phenom, Shanghai, China), and transmission electron microscope (TEM; JEOL-2100F, JEOL Ltd., Tokyo, Japan). Phase analyses were examined by an X-ray diffractometer (XRD; D/max 2550V, Rigaku Corporation, Tokyo, Japan) using Ni-filtered Cu Kα radiation. The densities of the samples were determined by Archimedes’ principle using water immersion. Quantitative metallographic analysis was carried out by quantitative image analysis software. The grain size was measured from the obtained OM micrographs, and it was determined by the linear intercept method using the equation D = L/(NV), where L is the total length of the test line, N is the number of intercepts, and V is the magnification of the photomicrograph. The shape factor (defined as c 2/4πS, where S is the area of the particle and c is the perimeter) was used to describe the roundness of the solid particles. If the particles are perfectly spherical, the shape factor has a value of 1; it increases for less spheroidal particles. At least three typical images on each sample were measured.
III. RESULTS AND DISCUSSION
A. Microstructure and mechanical properties of as-cast RSC-NZ30K alloy
Figure 1(a) shows the microstructure of quenched NZ30K alloy slurry. The particles are small and spheroidal, and the primary α phase is dispersed in a matrix of finer particles and eutectic structure (solidified from residual liquid phase). These particles are expected to grow up through ripening during stirring, and the solid fraction is about 23.9% under present conditions. Reference Chen, Zhang, Liu, Wu and Ding18 In addition, a uniform distribution of solute and temperature in the alloy melt is obtained by stirring. After RSC [Fig. 1(b)], the nearly spherical primary α-Mg particles are surrounded by the secondarily solidified particles (smaller particles which solidified from the residual liquid phase in the semisolid microstructure). The primary particle size increases from 31 μm [Fig. 1(a)] to 38 μm [Fig. 1(b)] after the RSC process, and the particles with size smaller than 30 μm can be regard as secondarily solidified particles [Fig. 2(a)].
Representative micrographs of the NZ30K alloy (a) semisolid slurry, (b) as-cast microstructure rheoformed at mold temperature of 200 °C.
Microstructures of NZ30K alloys rheoformed at different mold temperatures: (a) 200 °C, (b) 250 °C, (c) 300 °C, and (d) 350 °C.
Figure 2 presents the as-cast microstructures of NZ30K alloys prepared by the RSC process at different mold temperatures. With the mold temperature rise from 200 to 350 °C, the particles gradually transform to irregular ones, and the average shape factor increases continuously [Fig. 3(a)]. It indicates that plastic deformation happened between the neighboring particles. Moreover, the amount of secondarily solidified particles and eutectic structure is decreased with increasing mold temperature (Fig. 2). It reveals that the liquid phase flows toward the edge of the component and the primary particles approach each other along the direction parallel to the pressure Reference Yang, Peng, Fu, Hu, Ding and Yu21 during RSC, which is commonly found in the semisolid forming of Mg and Al alloys. Reference Chen, Huang, Huang, Ma and Hao22,Reference Chen, Chen, Zhang and Li23 It is well known that the deformation process of a semisolid non-dendritic slurry under pressure includes four steps: liquid flow (LF), flow of liquid incorporating solid particles (FLS), sliding between solid particles (SS), and plastic deformation of solid particles (PDS). Reference Chen, Chen, Zhang and Li23 The former two steps are dominant when the solid particles are surrounded by the liquid phase while the latter two operate when the solid particles are in contact with each other. During the initial stage of squeeze casting of the semisolid slurry, the LF and FLS should operate in succession because the primary particles are always surrounded by the liquid phase at this stage. The LF regime can generate constituent segregation, the liquid phase flows toward the edge of the component, and the primary particles approach each other along the direction parallel to the pressure. Simultaneously, solidification of the liquid phase occurs. Due to these two reasons, the liquid phase between the primary particles is decreased. The higher the mold temperature, the slower the solidification rate is, and more liquid phase flows to the edge, resulting in the decrease of secondarily solidified particles and eutectic structure. Furthermore, the average size of the primary particle gradually increases from 38 to 44 μm with the mold temperature rise from 200 °C to 350 °C [Fig. 3(a), at mold temperatures of 250 °C and 300 °C, the average primary sizes are 40 μm and 41 μm, respectively]. The density of this alloy increases from 1.7936 to 1.8038 g/cm3 by rising the mold temperature from 200 to 250 °C, and it is maintained at a value of 1.8075 g/cm3 with the mold temperature rise to 300 and 350 °C.
(a) Variations of average particle size, average shape factor, and density of the as-cast RSC-NZ30K alloy with mold temperature; (b) Variations of the primary particle size and average shape factor of the T6-treated RSC-NZ30K alloy with mold temperature.
In addition, the massive eutectic phase along the boundaries of particles can be observed in the alloy rheoformed at 250 °C [Fig. 2(b)]. According to the EDS results (Table I), the compositions of the eutectic phase are mainly composed of Mg, Nd, and a small amount of Zn element. The eutectic phases become much finer as the mold temperature increases from 250 to 350 °C [Fig. 2(d)]. Moreover, the Nd content in the eutectic phases is increased while the Zn content is decreased, and the α-Mg lattice constants are increased with the rise of mold temperature.
EDS results of the eutectic phases and lattice constants a and c of α-Mg and dislocation density in NZ30K alloy rheoformed at different mold temperatures.
The XRD patterns of RSC-NZ30K samples rheoformed at 250 and 350 °C are shown in Fig. 4(a). It is found that this alloy mainly consisted of α-Mg matrix and eutectic Mg12Nd phase and these phases remain unchanged at different mold temperatures. Moreover, as mention above, plastic deformation occurs between the neighboring particles during RSC of the NZ30K alloy. Plastic deformation would induce lattice strain. According to Williamson and Hall, the XRD diffraction line broadening is due to crystallite size and strain contribution. Reference Bindu and Thomas24 To analyze the strain effect, the instrumental corrected broadening is calculated using the following equation:
$${W_{hkl}} = {\left[ {{{\left( {{W_{hkl{\rm{ ‐ M}}}}} \right)}^2} - {{\left( {{W_{hkl{\rm{ ‐ I}}}}} \right)}^2}} \right]^{1/2}}\quad ,$$
where W hkl-M and W hkl-I are the measured and instrumental broadening, respectively. According to the Williamson–Hall (W–H) equation, the relation between crystallite size (d) and lattice strain (ε) is given by Reference Bindu and Thomas24
$${W_{hkl}}\cos {\theta _{hkl}} = \left( {{{M\lambda } / d}} \right) + 4\varepsilon \sin {\theta _{hkl}}\quad ,$$
where M is the shape factor (0.9), λ is the wave length (0.15406 nm), and θ hkl is the Bragg angle. The corresponding W–H plots of the RSC-NZ30K alloy rheo-formed at different mold temperatures are presented in Fig. 4(b). Based on the plots of W hkl cos θ hkl as a function of 4 sin θ hkl , the crystalline size d and the strain ε can be estimated from the y-intercept and slope of the linear fit. Then the dislocation density (ρ) can be obtained using the following formula:
$$\rho = {{2\sqrt 3 \varepsilon } / {\left( {db} \right)}}\quad ,$$
where b is the Burgers vector (b = 3.21 × 10−10 m). Reference Tang, Liu, Chen, Liu and Deng25 By substituting d and ε, the dislocation densities in the alloys are calculated. For the alloy rheoformed at 250 and 350 °C, the dislocation densities are 5.86 × 1013 m−2 and 2.44 × 1012 m−2 (Table I), respectively, indicating that the plastic deformation degree decreases with the mold temperature.
(a) and (b) are XRD patterns and the corresponding Williamson–Hall plots of the as-cast NZ30K alloy rheoformed at 250 °C and 350 °C, respectively. (c) and (d) are mechanical properties of the as-cast and T6-treated RSC samples at different mold temperatures.
Figure 4(c) shows the mechanical properties of RSC NZ30K alloys at different mold temperatures. The ultimate tensile strength (UTS), yield strength (YS), and elongation initially increase when the mold temperature rises from 200 to 300 °C and then decreases when the mold temperature further rises to 350 °C. When the mold temperature is 300 °C, the YS, UTS, and elongation are 116 MPa, 197 MPa, and 6.7%, respectively, which are 30.5, 12.6, and 91.4% higher than the corresponding values of the samples produced at mold temperature of 200 °C.
The mold temperature of 200 °C is the lowest temperatures used in this work, and the solidification rate of the semisolid slurry during RSC is the fastest. Thus the mold-filling ability and shrinkage-feeding ability are the lowest. Solidification shrinkage and porosities thereby easily generate in the last solidified zones, i.e., the secondarily solidified structures are the week points in the RSC-NZ30K alloys. Therefore, the density of the alloy is the lowest [Fig. 3(a)]. Cracks generally develop along the secondarily solidified structures Reference Chen, Huang, Huang, Ma and Hao22,Reference Chen, Chen, Zhang and Li23 and thus the properties of this alloy (YS of 103 MPa, UTS of 151 MPa, elongation of 3.5%) are the lowest under this condition. As the temperature rises to 250 °C, the solidification process slows down and the feeding ability to shrinkage porosities is improved. The density of the alloy is increased [Fig. 3(a), YS of 105 MPa, UTS of 173 MPa, elongation of 5.4%]. Therefore, the tensile properties are increased due to the improved microstructure compactness. When the temperature is elevated to 300 °C, the density of the alloy is further increased and more compact microstructure is obtained. In addition, the secondarily solidified structures among the particles are decreased. Therefore, though the particle size is slightly larger [Fig. 3(a)], the tensile properties of the alloy are the highest in this work. With the temperature rise to 350 °C, the alloy density is kept unchanged and the particle coarsens to a larger one due to the longer solidification time, thus the YS and UTS of the alloy are decreased (YS of 106 MPa and UTS of 177 MPa). However, the elongation (4.3%) is also decreased in the present work. The reason may be that under higher mold temperature, the Mg slurry is easier to oxidize and tend to introduce some inclusions. The brittle phases are deleterious to the mechanical properties and the decrease in the ductility may be attributed to the brittle Mg12Nd phase and the inclusions in the alloy. It is analogous to the development of strain localization in dual phase steel Reference Yan, Tasan and Raabe26 and nickel superalloy Reference Jiang, Yang, Zhang, Zou, Wang, Dunne and Britton27 with inclusion.
Based on the above analysis, the mold temperature has a large effect on the microstructure and mechanical properties of the RSC-NZ30K alloy. The primary particles in the microstructure gradually coarsen and connect each other with rising temperature, accompanied by decreasing amount of the eutectic Mg12Nd phase. The alloy density is increased and more compact microstructure is obtained with the rise of mold temperature. In addition, the dislocation density introduced by plastic deformation is decreased. The mechanical properties of the as-cast RSC-NZ30K alloy are initially increased and then decreased with rising temperature. The increase in the mechanical properties is mainly attributed to the more compact microstructure obtained. The further decrease of tensile properties may be ascribed to the coarsening of particles. It should be pointed out that the dislocation density measured by XRD provides an average over the probed area. It does not account for the spatial distribution of dislocations in individual phases in the material. The plastic deformation between the pre-solidified particles may result in high dislocation density and therefore affect the mechanical property. Further experiments need to be conducted to obtain the necessary data.
B. Effect of solution treatment on the microstructure of RSC-NZ30K alloy
Figures 5(a)–5(d) present the SEM backscattering micrographs of the NZ30K alloy rheo-formed at 300 °C and solution treated for different durations. The amount of the bright eutectic Mg12Nd phase is decreased significantly after solution-treated for 0.5 h [comparing Figs. 5(a) and 5(b)]. The Zr-containing particles are found and the eutectic phases almost completely dissolved into the α-Mg matrix after solution-treated for 8 h [Fig. 5(c)]. When the solution duration is prolonged to 10 h, the long-rod-like phase is observed [Fig. 5(d)] around the α-Zr particles and it is determined to be the Zn2Zr3 phase according to the TEM image analysis [Figs. 5(e) and 5(f)]. The long-rod-like Zn2Zr3 phase is not found in the as-cast microstructure and the short time solution treated alloy [Figs. 5(a)–5(c)], so its formation may be attributed to the reaction of Zr particles with dissolved Zn atoms.
Micrographs of the NZ30K alloy rheo-formed at 300 °C: (a) as-cast microstructure; (b), (c), and (d) are the solution-treated microstructure for 0.5 h, 8 h, and 10 h, respectively. (e) and (f) are the TEM bright field image of the long-rod-like Zn2Zr3 phase and the HRTEM image with FFT taken from the rectangular region.
Furthermore, the change of average particle sizes (including the primary particle size and secondary particle size) with solution time was analyzed and presented in Fig. 6(a). When the solution time is prolonged from 0 to 12 h, the primary particle size is increased from 23 to 28 μm, and the secondarily primary particle is increased from 43 to 52 μm. The results also show that the growth models of the two type particles obey well the traditional growth model Reference Huang, Chen, Ma and Hao28 in the first stage (0–6 h),
$$d = K{\left( {Dt} \right)^n}\quad ,$$
where d is the average particle size, K is the growth constant, and D is the diffusion coefficient which can be obtained by
$$D = {D_0} \cdot \exp \left( {{{ - Q} / {RT}}} \right)\quad ,$$
where D 0 is the diffusion constant related to temperature T, Q is the activation energy, and its value is 207 kJ/mol for the NZ30K alloy. Reference Zheng, Luo, Dong, Sachdev and Ding29 The calculated n values for primary particles and secondarily solidified particles are 0.058 and 0.054, respectively. Both the values are less than 2, which indicates that phase transformation is the main factor for grain coarsening, and the growth of the two kinds of particles is mainly controlled by the dissolution of eutectic phases in the first stage. Reference Cheng, Chen, Qi, Zhang and Yao30
(a) Variations of the average primary particle size and secondarily solidified particle size of the RSC-NZ30K alloy with solution time. (b) and (c) are XRD patterns of the different solution treated samples and partial enlargement of the α-Mg peak.
The growth in the second stage during solution-treated (6–12 h) is quite different, and the data points may be fitted to the following equation:
$$d_t^n - d_0^n = kt\quad ,$$
where d t and d 0 are the final and initial particle sizes, respectively, k and n are constants that are functions of interfacial energy, atomic diffusivity, and solution concentration, and t is the solution time. The fitting results show that the values of primary particles and secondarily solidified particles are 3.7 and 3.8, respectively. It is believed that if the coarsening is controlled by diffusion across a particle/matrix interface, diffusion through lattice, and diffusion along grain boundaries, n would be 2, 3, and 4, respectively. Reference Wang, Liu and Fan31 The values in this work are larger than 3, but lower than 4, so it can be suggested that the particle coarsening of the RSC alloy in the second stage obeys a mixture model including atom diffusion through crystal lattice and along the grain boundaries. Therefore, the eutectic Mg12Nd phase in the grain boundaries gradually dissolves into the α-Mg matrix, and the Zr-containing particles are observed when the alloy solution-treated for 6 and 10 h [Figs. 5(c) and 5(d)].
To study the phase variation of the RSC-NZ30K alloy during solution treatment, XRD is conducted on each sample and the results are shown in Fig. 6(b). The alloy mainly consists of an α-Mg matrix and Mg12Nd eutectic phase [Fig. 6(b)], and the long-rod-like Zn2Zr3 phase is not observed due to its small amount. Figure 6(c) shows that a little migration of the α-Mg peak appears with different durations. The migration should be attributed to the change of lattice constants. As shown in Table II, the lattice constants decrease with increasing solution time in the initial stage because of the dissolution of the eutectic Mg12Nd phase. However, the lattice parameters are increased when the solution time is prolonged to 8 h. The reason may be that the Mg12Nd phase nearly completely dissolved into the α-Mg matrix at that time, while the reaction of Zr particles with dissolved Zn atoms leads to the increase of lattice constants. As the solution time is extended to 10 h, the lattice constants are slightly decreased, which may be attributed to the formation of the Zn2Zr3 phase.
Lattice constants a and c of α-Mg in NZ30K alloy solution treated for different durations.
In summary, the main phenomena occurred in the RSC-NZ30K alloy during solution treatment is the quick dissolution of the eutectic phase and the resulting rapid coarsening of the particles including primary particles and secondarily solidified particles in the early stage and the normal growth due to the relatively slow coarsening rates in the latter stage. The growth in the early stage was mainly controlled by the dissolution of the eutectic phase and the traditional growth model, while the growth in the latter stage obeyed a mixture model that included atom diffusion along grain boundaries and through the crystal lattice. In addition, the reaction of Zr particles with dissolved Zn atoms led to the formation of the long-rod-like Zn2Zr3 phase and the change of the lattice constants of α-Mg in the alloy.
C. Microstructure and mechanical properties of T6-treated RSC-NZ30K alloy
Figure 7 shows the microstructures of T6-treated NZ30K alloys prepared by the RSC process at different mold temperatures. The primary particle size is continuously increased from 31 to 42 μm with rising mold temperature from 200 to 350 °C [Figs. 7 and 3(b)]. The average shape factor of the primary particle also slightly increases from 1.4 to 2.0 with the rise of mold temperature [Fig. 3(b)]. In addition, the long-rod-like Zn2Zr3 phase is also observed in the T6-treated alloys rheoformed at mold temperatures of 300 and 350 °C [Figs. 7(c) and 7(d)], and it is not found when the mold temperature is lower than 300 °C [Figs. 7(a) and 7(b)] which may be attributed to its small amount and size. The amount and size of the Zn2Zr3 phase are increased with rising mold temperature in the T6-treated alloy [comparing the inset high-magnification images in Figs. 7(c) and 7(d)]. Thus it can be proposed that the formation of the long-rod-like Zn2Zr3 phase during age treatment may be attributed to the reaction of undissolved Zr particles with segregation of Zn atoms. The atom segregation phenomenon is normally found in the early stage during age treatment in the Mg–Nd–Zn alloys, Reference Sanaty-Zadeh, Luo and Stone7 and its degree is more severe with the rise of mold temperature according to the present experimental results.
SEM micrographs of the T6-treated NZ30K alloy rheoformed at different mold temperatures: (a) 200 °C, (b) 250 °C, (c) 300 °C, and (d) 350 °C. The insets in (b) and (d) are the high-magnification showing the long-rod-like phase.
Figures 8(a) and 8(b) present the TEM dark field images and corresponding selected area electron diffraction (SAED) patterns of the T6-treated NZ30K alloy rheoformed at 300 °C. According to the
${\left[ {01\bar{1}0} \right]_\alpha }$
dark field images, two types of precipitates are observed. The rod-like precipitates [Fig. 8(a)] should be the β′ phase and the small plate-like precipitates [Fig. 8(b)] should be the β″ phase according to the SAED patterns and the previous studies.
Reference Fu, Peng, Jiang, Chang and Zhai3,Reference Li, Lv, Liang, Qi, Zhu and Zhang32
The diffraction spots at 1/2{2114}α in
${\left[ {01\bar{1}0} \right]_\alpha }$
, correspond to β″ (Mg3Nd) precipitates with the D019 structure (space group P63/mmc, a = b = 0.64 nm, c = 0.52 nm) and those at 1/2{2114}α in the
${\left[ {01\bar{1}0} \right]_\alpha }$
SAED pattern correspond to β′ (Mg7Nd) precipitates with a base-centered orthorhombic unit cell (a = 0.64 nm, b = 1.14 nm, c = 0.52 nm).
Reference Nie33
It is consistent with the previous studies which indicated that the precipitation sequence in Mg–3 wt% Nd was super saturated solid solution (SSSS) → G. P. Zones → β″ → β′/β1 → β.
Reference Sanaty-Zadeh, Luo and Stone7,Reference Ma, Mishra, Balogh, Peng, Luo, Sachdev and Ding34
Furthermore, for the Mg–3Nd–0.2Zn alloy, the precipitation sequence was identified as SSSS → Clusters of atoms → G. P. Zones (I, II, III) → β′ → β2 → β1/γ → β.
Reference Sanaty-Zadeh, Luo and Stone7
These β-type precipitates are believed to play an important role in the enhancement of alloy’s mechanical properties, and the co-existence of β″ and β′ precipitates can be related to the maximum YS of the alloy.
(a) and (b) are TEM dark field images and corresponding SAED patterns with B//
${\left[ {01\bar{1}0} \right]_\alpha }$
of the T6-treated NZ30K alloy rheoformed at 300 °C. (c) SEM micrograph and (d) TEM dark field image and corresponding SAED pattern with B//
${\left[ {01\bar{1}1} \right]_\alpha }$
of the T6-treated CSC-NZ30K alloy.
Figure 4(b) shows the mechanical properties of T6-treated RSC NZ30K alloys at different mold temperatures. With the mold temperature rise from 200 °C to 300 °C, the UTS is significantly increased from 227 to 296 MPa, the YS is increased from 147 to 160 MPa, and the elongation is increased from 3.3 to 7.7%. However, the tensile properties are decreased as the mold temperature further rises to 350 °C (the YS, UTS, and elongation are 149 MPa, 257 MPa, and 5.7%, respectively). The improvement in mechanical properties may be attributed to the smaller particle size (Fig. 7) and more compact microstructure obtained. The further decreased in mechanical properties should be attributed to the larger particle size observed.
D. Strengthening contributions in the RSC-NZ30K alloy and the CSC-NZ30K alloy
The strengthening contributions in the as-cast Mg-RE alloy include grain boundary strengthening (σgb), solid solution strengthening (ΔσS), dislocation strengthening (Δσdis), and second phase strengthening (Δσ2nd). The grain boundary strengthening (σgb) in the as-cast RSC-NZ30K alloy can be estimated by the Hall–Petch relationship, Reference Tang, Liu, Chen, Liu and Deng25
$${\sigma _{{\rm{gb}}}} = {\sigma _{{\rm{Mg}}}} + k{d^{ - 1{\rm{/}}2}}\quad ,$$
where σMg is the YS of the pure Mg (σMg = 21 MPa), k is a constant related to the material, its value is 280–320 MPa μm−1/2 for Mg. Reference Tang, Liu, Chen, Liu and Deng25 Therefore, the calculated contribution of grain boundary strengthening is 65–71 MPa for the as-cast RSC-NZ30K alloy. Solid solution strengthening occurs when an alloy element is trapped into the metal matrix and interacts with dislocations, resulting in the increase of mechanical properties. The solid solution strengthening contribution from solute element i to the YS can be assumed by Reference Toda-Caraballo, Galindo-Nava and Rivera-Díaz-del-Castillo35
$$\Delta {\sigma _{\rm{S}}} = \Sigma {k_i}{C_i}^n\quad ,$$
where C i is the atomic fraction of solute i and k i is a constant for element i, n = 1/2 or n = 2/3 according to Fleischer’s approach Reference Fleischer36 and Labusch’s approach. Reference Labusch37 It is reported that the strengthening model with n = 2/3 presents a better description for the Mg–Nd–Zn–Zr system alloy than that with n = 1/2. Reference Li, Fu, Peng, Wang and Jiang9 For Nd and Zn in the Mg–Nd–Zn–Zr alloy, the solid solution strengthening of Nd and Zn can be estimated as Reference Li, Fu, Peng, Wang and Jiang9
$$\Delta {\sigma _{\rm{S}}}\left( {{\rm{Nd}}} \right) = 813{C_{{\rm{Nd}}}}^{2/3}\quad ,$$
$$\Delta {\sigma _{\rm{S}}}\left( {{\rm{Zn}}} \right) = 964{C_{{\rm{Zn}}}}^{2/3}\quad .$$
By substituting the content of Nd (0.58 at.%) and Zn (0.084 at.%), the calculated values of ΔσS (Nd) and ΔσS (Zn) are 26 and 9 MPa. In addition, the dislocation in the RSC-NZ30K alloy will influence the mechanical properties of the alloy by interacting with each other and impeding their own movement. According to Bailey and Hirsch’s research, the strengthening contribution induced by dislocation can be estimated by Reference Bailey and Hirsch38
$$\Delta {\sigma _{{\rm{dis}}}} = M\partial Gb{\rho ^{1/2}}\quad ,$$
where M is the mean Taylor factor of random polycrystalline Mg (M = 4.5) Reference Cáceres and Lukáč39 ; α is a constant related to the dislocation–dislocation interaction constant, its value is about 0.7 for random polycrystalline Mg Reference Cáceres and Lukáč39 ; G (G Mg = 16.6 GPa) is the shear modulus of the Mg matrix, b is the magnitude of Burgers vector, and it is about 0.321 nm for Mg Reference Tang, Liu, Chen, Liu and Deng25 ; ρ is the dislocation density and its value is about 9.61 × 1012 m−2 which is calculated by using Eq. (3). The calculated ρ value is larger than that of the Mg–1.1Y–Zn alloy (TEM image showed that the value was 6 × 1012 m−2 after creep deformation). Reference Suzuki, Kimura, Koike and Maruyama40 Based on the above date and Eq. (11), the calculated Δσdis is about 0.07 MPa. The value is small and thus the contribution of dislocation strengthening can be ignored in this work.
During RSC, solidification of the alloy is quite fast under the applied pressure. Solidification is non-equilibrium and the phase composition of the RSC alloy obeys the Scheil equation, leading to the formation of second phases. It is known that when the precipitates are spherical or rod-like morphologies with size below 200 nm, the Orowan strengthening is the main mechanism. However, the second phase in this work is the network-like Mg12Nd intermetallic phase with size larger than 9 μm. Therefore, the second phase strengthening (Δσ2nd) is calculated according to the present experimental results and the different strengthening contributions in the as-cast alloy are listed in Table III. In conclusion, the YS of the as-cast RSC-NZ30K alloy can be expressed as
$${\sigma _{0.2}} = {\sigma _{{\rm{Mg}}}} + k{d^{ - 1{\rm{/}}2}} + {k_{{\rm{Nd}}}}{\left( {{C_{{\rm{Nd}}}}} \right)^n} + {k_{{\rm{Zn}}}}{\left( {{C_{{\rm{Zn}}}}} \right)^n} + {\sigma _{2{\rm{nd}}}}\quad .$$
Strengthening contributions in as-cast, T6-treated RSC-NZ30K alloy, and T6-treated CSC NZ30K alloy.
Figures 8(c) and 8(d) present the SEM micrograph and TEM images of the T6-treated CSC-NZ30K alloy. The grain size of the CSC-NZ30K alloy is about 33 μm [Fig. 8(c)], which is smaller than the RSC alloy [Fig. 5(a), 38 μm]. In addition, the long-rod-like Zn2Zr3 phase is also observed around the Zr-containing particles which are within the α-Mg grains. The TEM dark field image and the corresponding SAED pattern show that the precipitates in the CSC-NZ30K alloy are mainly composed of the β′ phase [Fig. 8(b)]. The small plate-like β″ phase in the RSC sample is not found in the CSC sample.
Generally, the application of pressure during solidification would lead to microstructure refinement, which should be partly attributed to the increase in the heat transfer coefficient by eliminating air gaps between the alloy slurry and mold wall surface. Moreover, the applied pressure during solidification also affects the phase diagram of the alloy by shifting the liquidus and solidus lines and even the eutectic composition. Reference Yang, Peng, Fu, Hu and Ding41 Thus the nucleation rate in the alloy melt is increased and the refined particles are obtained. The particles grow up through ripening during EMS treatment; therefore, the average particle size of the RSC sample is larger than that of CSC sample. However, comparing with the T6-treated RSC sample, the T6-treated CSC sample presents lower tensile properties (YS of 154 MPa, UTS of 290 MPa, elongation of 7.7%) though the particle size is smaller. To study the different strengthening contributions in the T6-treated CSC-NZ30K alloy and RSC-NZ30K alloy, two contributors including the grain boundary strengthening (σgb) and precipitation strengthening (σppt) are discussed. The YS of the T6-treated alloys can be expressed as Reference Li, Lv, Liang, Qi, Zhu and Zhang32
$${\sigma _{{\rm{ys}}}} = {\sigma _{{\rm{Mg}}}} + {\sigma _{{\rm{gb}}}} + {\sigma _{{\rm{ppt}}}}\quad .$$
According to Eq. (7), the calculated contribution of grain boundary strengthening σgb is 70–77 MPa for the T6-treated CSC sample and 65–72 MPa for the T6-treated RSC sample. σppt can be calculated according to the present experimental results and the different strengthening contributions are listed in Table III. σppt of the T6-treated RSC sample offer higher strengthening properties (67–74 MPa for the RSC sample and 56–63 MPa for the CSC sample), which may be attributed to the more precipitates obtained. Therefore, the mechanical properties of the T6-treated RSC sample are higher than those of the T6-treated CSC sample.
IV. CONCLUSIONS
In this study, the effect of mold temperature on microstructure and mechanical properties of NZ30K alloy components prepared by RSC combined with EMS has been investigated. Some conclusions can be drawn as follows:
-
(1) With the rise of mold temperature, the particle size of the RSC-NZ30K alloy is increased, the dislocation density is decreased, the eutectic phases are finer, and the secondarily solidified structures are decreased. In addition, the density of the alloy is increased and a more compact microstructure is obtained.
-
(2) The mechanical properties of the as-cast RSC-NZ30K alloys are initially increased and then decreased as the mold temperature rises. The increase in the mechanical properties is mainly attributed to the more compact microstructure obtained. The further decrease of tensile properties may be ascribed to the coarsening of particles.
-
(3) The microstructure evolution of the RSC-NZ30K alloy during solution-treatment can be divided into two stages: the quick dissolution of the eutectic phase and the resulting rapid coarsening of the particles and the normal growth due to the relatively slow coarsening rates. The growth in the early stage is mainly controlled by the dissolution of the eutectic phase and the traditional growth model, while the growth in the latter stage obeyed a mixture model that included atom diffusion along grain boundaries and through the crystal lattice.
-
(4) After T6-treated, rod-like β′ precipitates are observed in the CSC-NZ30K alloy, while two types of precipitates of the small plate-like β″ phase and β′ phase are found in the RSC-NZ30K alloy, resulting in the larger precipitation strengthening in the RSC-NZ30K sample.
-
(5) With rising mold temperature, the mechanical properties of the T6-treated RSC-NZ30K alloy are also initially increased and then decreased. Compared with the T6-treated CSC-NZ30K sample, the mechanical properties of the RSC sample are higher due to the larger precipitation strengthening contribution and the YS, UTS, and elongation are up to 160 MPa, 296 MPa, and 7.7%, respectively.
ACKNOWLEDGMENTS
This work is supported by National Key Research and Development Program of China (No. 2016YFB0701205), Science Innovation Foundation of Shanghai Academy of Spaceflight Technology (Nos. SAST2015047 and SAST2016048).
