Microstructure Evolution of Mg_96.9Gd_2.7Zn_0.4 Alloy Containing Multiple Phases Prepared by Spark Plasma Sintering Method

Abstract

The synergic strengthening of multiple phases is an essential way to achieve high-performance Mg alloys. Herein, Mg-Gd-Zn alloy containing four phases was prepared by rapid solidification (RS) ribbons and spark plasma sintering (SPS). The microstructure of the alloy consisted of α-Mg, nanosized β₁ phase particles, lamellar long period stacking ordered (LPSO) phase, and β′ phase precipitates. The microstructural evolution was also investigated. The results show that the metastable β₁ phase was formed in the as-cast solidification through rapid solidification, because both Zn atoms and the short holding-time at molten liquid facilitated the formation of the β₁ phase. The β₁ phase grew from 35.6 to 154 nm during the sintering process. Meanwhile, the fine lamellar LPSO phase was simultaneously formed after the Zn-Gd clusters were generated from the supersaturated solid solution, and the width of the LPSO phase was only in the range of 2–30 nm. The third strengthening phase, the metastable β′ phase, was obtained by aging treatment. The results of hardness testing implied that the hardness of the alloy containing the aforementioned three nanosized strengthening phases significantly improved about 47% to 126 HV compared with that of the as-cast ingot.

1. Introduction

As a vital lightweight metallic structure-material, increasing attention is being paid to magnesium (Mg) alloys due to their high specific strength and easy machinability [1,2]. However, Mg suffers from inferior absolute strength and poor plastic deformation at room temperature [3]. In the various strengthening methods for Mg alloys, second-phase (including aging precipitates) strengthening is one of the most important approaches [4,5,6].

Multiple phases introduced into the metals have been proven to be effective in forming satisfactory stable phase boundaries [2,7]. Through tailoring the microstructure according to strengthening phase assembly, the synergistic strengthening effect could be achieved. Generally, severe plastic deformation (SPD) and thermal-mechanical treatment are utilized to refine the grain size, which occasionally leads to two or three strengthening phases formed simultaneously. It was reported that an extruded Mg-Y-Zn alloy was strengthened by both grain boundaries and the second phase, showing a yield strength of ~380 MPa and an elongation of ~7% [8]. It is worth noting that realizing the synergic enhancement of different strengthening phases is the key point for improving the mechanical properties of Mg alloys. Nevertheless, avoiding both the inhomogeneous microstructure and anisotropic mechanical properties for the conventional SPD manufacturing methods still remains a challenge. Besides, the selection of the second phase is also significantly limited based on the conventional chemical composition design of the alloy. Therefore, utilizing a new manufacturing process to develop multiple phases with the synergic strengthening effect is a potential way to achieve higher performance Mg alloys. Jian Lv et al. used magnetron sputtering to fabricate a dual-phase alloy, which consisted of amorphous Mg₄₉Cu₄₂Y₉ and nanocrystal MgCu₂ phase, with a strength up to 3.3 GPa and a favorable fracture-strain of 4% [2]. On the other hand, spark plasma sintering (SPS) has been applied in the preparation of Mg alloys [9,10] as a relatively low-temperature sintering method. It is easy to obtain the specific phases and homogeneous microstructure via adjusting the sintering temperature/time and loading pressure, which could contribute to the realization and regulation of the synergic strengthening effect with multiple phases.

In the Mg-RE (RE denotes rare earth elements) series alloys, β series phase, including the vital strengthening metastable β′ phase (Mg₇RE) with an orthorhombic structure and the stable equilibrium β phase with a cubic structure (Mg₅RE, a = 2.23 nm), metastable β₁ phase with an face centered cubic (fcc) structure (Mg₃RE, a = 0.73), as well as metastable β′′ phase precipitates with a D0₁₉ structure (Mg₃RE, a = 0.64 nm) [5], can significantly strengthen the Mg matrix. Generally, β′ has been the most expected precipitated phase in these phases. β₁ phase has rarely been studied because the size of this phase is hard to control for forming a net and bulk shape during its solidification process.

Herein, an Mg-Gd-Zn alloy containing three phases, including two metastable β series phases (β₁ and β′ precipitates) and a long period stacking ordered (LPSO) phase, is prepared by rapidly solidified (RS) ribbons and SPS method, and the microstructure evolution of this alloy is particularly investigated. The study shows that the β₁ phase also has a remarkable strengthening effect on Mg alloys.

2. Experimental

2.1. Preparation of Samples

2.1.1. Preparation of the Precursors

The precursors of the sintering Mg-Gd-Zn bulk are the as-cast ingot (or the master alloy) and the RS ribbons in this paper. Firstly, Mg (99.95 wt.%), Zn (99.95 wt.%), and Mg-30Gd (wt.%) were used to prepare the Mg_96.9Gd_2.7Zn_0.4 (at.%) master alloy (low-rate solidification). The alloy was melted in a stainless steel crucible in an electric resistance furnace under the protection of the CO₂ + SF₆ (99:1) atmosphere. The melt was then kept at 993 K for 5–10 min before casting into the ingot. Then, the ingot was re-melted by induction heating and subsequently sprayed to obtain the RS ribbons, and the planar flow casting method was conducted with a single copper roller with a speed of about 20 m·s⁻¹ in N₂ atmosphere (∼500 Pa) [11]. The prepared RS ribbons were several hundred millimeters in length, ∼10 mm in width, and the thickness of the ribbons was in the range of 50–75 μm.

2.1.2. Preparation of Mg-Gd-Zn Alloy Containing Multiple Phases

The RS ribbons were subsequently cut and put in a columned graphite die (Φ 40 mm × 40 mm in the outer side, and Φ 15 mm × 40 mm in the inner hole), layer by layer. Consolidation was carried out with a spark plasma sintering (SPS) system (Fuji Electronic Industrial Co., Ltd., Saitama, Japan). The ribbons were compacted under a load of 10 MPa for 2 min, followed by degassing before the sintering. The sintering process was carried out at 450 °C, with a pressure of 45 MPa for 5 min in a vacuum (<0.001 Pa). When the sintering process was finished, the samples continued to be kept under a pressure of 20 MPa as the temperature cooled to ∼200 °C, and then they were cooled to room temperature without any loading. The size of the obtained bulk alloy was 15 mm in diameter and 6.5 mm height. Aging treatment for the SPS bulk was conducted at 200 °C, with a holding time of 20 h in an isothermal muffle furnace.

2.2. Characterization and Tests

2.2.1. Density Measurement

The density of the SPS bulk was determined by an electronic balancer with an accuracy of ±0.0001 g according to the Archimedes’ principle. The result showed that the measured relative density of the bulk alloy was about 99.95% of the as-cast alloy.

2.2.2. Hardness Tests and Analysis

The samples for the hardness test were machined to the size of Φ 4 mm × 6 mm via electron discharge machining (EDM). The Vickers hardness was measured by a VMT-7S, with a 200 g load, a holding time of 10 s, and at least 10 points for each tested sample were measured. Before conducting these hardness tests, we found three phases (β₁ phase, α phase, LPSO phase) that could be contained in SPS samples when the size of the hardness impress was larger than 10 μm × 10 μm; similarly, in the as-cast samples, the size of the hardness impress with an area of 20 μm × 20 μm could cover the β₁ and α phase. The plentiful β′ phase precipitates were obtained and uniformly distributed on the matrix after the aging treatment of the SPS samples. Therefore, the hardness impresses were larger than 20 μm × 20 μm in this paper, and it could be definitely determined that each hardness impress contained the studied phases.

2.2.3. Microstructure Characterization

The phases were identified by X-ray diffraction (XRD, Rigaku Ultima IV, Tokyo, Japan) with 3 kW Cu-Kα radiation at 40 kV and 300 mA, and 2θ was from 20° to 80° with a scanning rate of 0.02° s⁻¹. A ZEISS Auriga-EVO 18 field-emission scanning electron microscope (FE-SEM, 0–20 kV, ZEISS, Jena, Germany), equipped with an energy dispersive X-ray spectrometer (EDS) system, was used to analyze the microstructure of the alloy. A Tecnai G2 F30 transmission electron microscope (TEM, Thermofisher Scientific Co., Waltham, MA, USA) equipped at 300 kV was also used to investigate the microstructural characterization in detail. Samples for SEM observation were etched in a solution of 4 vol.% nitrate alcohol. The TEM foils were firstly mechanically polished to ∼50 μm, punched into discs of 3 mm in diameter, and then ion milled using a Gatan plasma ion polisher. ImageJ software was used to calculate and analyze the phase sizes and size distribution, and ~200 particles or grains were measured for calculating these sizes.

3. Analysis of Results

3.1. The Microstructure of Precursors

Figure 1 shows the XRD patterns of Mg_96.9Gd_2.7Zn_0.4 (at.%) alloy under different conditions. For the lower solidified rate sample (the as-cast ingot), the main peaks correspond to the α-Mg phase, and another suit of peaks is related to the β₁ phase (the ternary compound (Mg, Zn)₃Gd phase). With increasing solidification rate under RS, the XRD pattern changes dramatically: the number of peaks decrease, and the width of all peaks sharply reduce, which indicates the grain size significantly decreases.

For the SPS bulk alloy, not only α-Mg phase and β₁ phase, but also an LPSO phase were detected under elevated temperatures, which is consistent with the reported LPSO form condition of heat treatment or the thermo-deformation process [12]. During rapid solidification, the main crystal plane of the α-Mg phase increases from (101)_Mg to (002)_Mg. Moreover, the relative intensities of some planes also reduce, like the planes of (100)_Mg, (102)_Mg, and (021)_Mg, as shown in Figure 1, which is mainly because the extremely solidified rate restrains the stacking of atoms for its grow-up along the non-closed stacking plane a axis.

The SEM microstructure of the as-cast ingot is shown in Figure 2a,b, and the microstructure of RS ribbon is shown in Figure 2c,d. The microstructures of both as-cast ingot and RS ribbon consist of dendritic grains and the β₁ phase with a continuous net shape, which is mainly located on the grain boundaries. The rate of the phase is estimated at ~10 vol.% in the as-cast alloy based on the quantitative metallographic technique. The microstructure is shown to be significantly refined through the RS process, with the grain size decreasing to 200 nm.

The TEM microstructure of the β₁ phase of the RS ribbons is shown in Figure 3. The results suggest that the particles are uniformly distributed in the grains with an ellipsoid shape. The size distribution of β₁ phase particles are shown in Figure 3b, and suggests that the sizes are in the range of 10–60 nm and the average size is about 35.6 nm.

3.2. The Microstructure of the Bulk Containing Multiple Phases

Figure 4a,b shows the SEM microstructure of the SPS bulk. The grain size significantly increases due to the elevated temperature during the sintering process. Quantitative analysis of the distribution of the grain size based on TEM images is summarized in Figure 5a. The results suggest that the grain sizes of the sintering alloy are mainly in the range of 3–10 μm, and the mean grain size is about 6.3 μm. Moreover, the β₁ phase also grows a little, which is uniformly located in the grains and their boundaries. The LPSO phase was also observed to form in this process. These two second phases (β₁ and LPSO phase) have been characterized in our previous study [11]. The size distributions of the β₁ phase in the SPS bulk are shown in Figure 5b, and the results show that the particle sizes are in a range of 20–650 nm, in which most are less than 300 nm (about 95 %), and the mean particle size is 154 nm.

The TEM microstructure of the sintering alloy is shown in Figure 4c,d. Fine equiaxed grains with a size of about 2 μm can be observed, as well as β₁ phase particles and lamellar LPSO phases crossing the whole grains. The rate of β₁ phase particles is about 8~11 vol.%, which is approximately equal to that of the as-cast ingot.

4. Discussions

4.1. The β₁ and β′ Phase Transformation

The β₁ phase is one of the families of intermetallic compounds Mg₃X, which is an impressive strengthening phase in Mg-RE-Zn alloys [13]. It usually forms via solid-phase transformation during the aging treatment process, with a long holding-time at 200–250 °C [14]. In specific Mg-RE (Gd, Y)-Zn alloys, the β₁ phase can be formed from the molten liquid at a certain solidification process, and the size is usually smaller than that of the equilibrium β phase. It is well known that the equilibrium β phase is more stable than β₁ phase [12]. Generally, in as-cast Mg-RE (-Zn) alloys with high a content of RE (≥1 at.%), β phase is the predominant second phase due to the relatively low driving force of the β phase nuclear Gibbs free energy of formation (Δo) [12], which can easily lead to aggressive growth of the β phase during solidification. Therefore, β phase is easily observed in as-cast Mg-15wt.%Gd alloy, Mg-8Gd-3Yb-1.2Zn-0.5Zr (wt.%) alloy [15,16], and other Mg-RE alloys.

In this study, the β₁ phase grows a little during the sintering process. From the XRD pattern of the β₁ phase, the intensity of (200)_β1 and (220)_β1 increase significantly, which corresponds to the changing of peaks of the α-Mg matrix. In addition, the interplane distance of β₁ phase with an fcc structure and α-Mg phase with an hexagonal close-packed (hcp) structure, d_(220)β1 (2.59 Å), is near equal to d_(002)α (2.60 Å). The XRD results show the same tendency of the relative intensity of the planes of both (220)_β1 and (002)_α. The orientation relationship between β₁ phase and α-Mg has been determined by many researchers as (−112)_β1//(210)_α, [110]_β1//[1]_α [5,17], which is a nearly perfectly coherent relationship [13]. For Mg_96.9Gd_2.7Zn_0.4 alloy, the solidification favors the formation of the β₁ phase rather than the β phase. The following reasons are considered. Firstly, the Gibbs energy changing of a phase nuclear (ΔG_m = G_F − G_E) needs to be mentioned based on the thermodynamics theory. As is well-known, the ΔG_m of a phase has a negative and smaller value; therefore, the β phase shows more stability during its formation process. The equations for G_F of the β phase and β₁ phase formation are shown in

Table 1. The Gibbs energy of phase formation for β phase and β₁ phase.

GF	Free Zn Atoms [12]	With Zn Addition [12]
β	$G_{Gd:Zn}^{0,GdMg5}=-\mathrm{150,000}+12T+G_{Gd}^{0,hcp}+5G_{Zn}^{0,hcp}$	$G_{Gd,Mg,Zn}^{0,GdMg5}=-\mathrm{107,700}+24T$
β1	$G_{Gd:Zn}^{0,GdMg3}=-\mathrm{140,000}+12T+G_{Gd}^{0,hcp}+3G_{Zn}^{0,hcp}$	$G_{Gd,Mg,Zn}^{0,GdMg3}=-\mathrm{69,500}+12T$

. During the solidification process, though the values G_F of both phases are significantly negative, Zn atoms have a stronger effect on the ΔG_m of the β₁ phase than that of the β phase. Secondly, the lattice parameter of β₁ phase (a = 7.4 Å) is much smaller than that of β phase (a = 22 Å), which results in the nucleation of β₁ phase being more convenient due to the shorter moving distance of Zn atoms. Thirdly, the holding time of this alloy at molten liquid is less than 10 min, restraining the nucleation of the stable β phase, which suggests the β₁ phase prefers forms during the solidification. Notably, the lattice parameter of β₁ phase is slightly smaller, a = 7.17 Å, (a = 7.3–7.4 Å [12]). This is mainly because the Zn addition in β₁ phase can dissolve more Zn atoms than that of β phase, and the Zn content in β₁ phase is about 2 at.% in this work.

It is well-known that the Zn element can effectively improve the precipitation response of Mg-Gd-Zn alloys [18]. The β′ phase forms during the aging treatment with mediate holding time in Mg-Gd (-Y)-Zn series alloys [14,19,20]. The formation of the β′ phase is also discussed in detail in the aging treatment process. The strengthening effect of β′ phase works according to the Orowan by-passing mechanism. Both the distribution of the β’ phase and the sizes can be tailored through controlling the aging temperature and the holding time.

4.2. The LPSO Phase Formation

The LPSO phase, as a long-range ordered intermetallic phase, is a significantly strengthening second phase in Mg alloys, which can obviously impede the dislocations and then improve the strength remarkably. Kinking formation also occurs in the LPSO phase to assist the deformation in maintaining favorable plasticity [4,21].

It should be noted that the differences of electronegativity between Zn and Gd are greater than that between Zn and Mg (1.6 for Zn, 1.3 for Mg, and 1.1 for Gd), and Zn atoms are easily fixed with Gd atoms to form Zn-Gd clusters, such as Zn₆RE₉ clusters [22], which play a critical role in the formation of the LPSO phase.

For the SPSed Mg_96.9Gd_2.7Zn_0.4 alloy, the content of the lamellar LPSO phase was estimated to be about 18 vol.%, which was obtained according to analysis of the strengthening mechanism in our previous study [11]. The width was in the range of 2–30 nm, and the corresponding precipitate diameter (d_p) was in the range of 33–68 nm. The smaller the size of the LPSO phase distributed, the more enhancement of strength is obtained [8,23]. In this study, both the content and the morphology of the LPSO phase could be easily controlled through the parameters of the spark plasma sintering process. The following reasons are given. Firstly, the LPSO phase can be steadily formed in a wide temperature range of 200–520 °C [18,24], which is favorable for the sintering of most Mg-RE-Zn alloys because the sintering is usually conducted in the temperature range of 400–500 °C. Secondly, the precursor of sintering bulk, RS ribbons, consists of supersaturated solid solution (SSS) and a small amount of the β₁ phase particles. SSS provides more possibilities for the nucleation of the LPSO phase. Finally, both short holding time and retaining loading state prevent excessive growing-up to a bulk shape.

The microstructure with multiplex strengthening phases is successfully prepared in this study. The results show that the SPS process is a more effective way to control the component and size of the LPSO phase through the sintering temperature and holding time. High temperature (450 °C) facilitates the activation of atom motion to form the LPSO phase during the sintering process. Besides, the short holding time (5 min) maintains the small size with a fine lamellar shape. Therefore, the LPSO phase is uniformly distributed in the matrix, and the width of the phase is only approximately regulated to tens of nanometers.

4.3. Hardness Response

Through controlling the SPS parameters, the microstructure containing multiple strengthening phases is obtained, which consists of α-Mg, β₁ phase, LPSO phase, and the β′ phase. Although the size of a little β₁ phase particle is larger than 100 nm, most of the LPSO phase and β′ phase is controlled to less than 50 nm. For the strengthening mechanism, as discussed in our previous study [11], the SPS Mg_96.9Gd_2.7Zn_0.4 bulk is mainly strengthened by grain-boundaries, SSS, β₁ phase, and LPSO phase.

The hardness of Mg_96.9Gd_2.7Zn_0.4 alloy at different states (as-cast ingot, SPS bulk, and age-treated SPS bulk) is shown in Figure 6. It can be seen that the hardness increases from 86 HV for the as-cast ingot to 108 HV for the sintering bulk. Furthermore, as the aging treatment proceeds, the hardness increases by 47% to 126 HV from its as-cast ingot.

5. Conclusions

In this work, Mg-Gd-Zn alloy with a synergistic strengthening microstructure consisting of α-Mg, β₁ phase, LPSO phase, and β′ phase, was reported, which was achieved by aging treatment after spark plasma sintering of rapidly solidified Mg_96.9Gd_2.7Zn_0.4 (RS) ribbons. The microstructure evolution of the SPS alloy, as well as its mechanical properties (Vickers hardness), were studied, and the following conclusions were obtained as follows:

(1) The synergistic strengthening effect of Mg alloys through multiple phases could be easily realized by using the SPS method. The content and the sizes of nanosized β₁ phase particles and the lamellar LPSO phase could be controlled by both the sintering temperature and the holding time. Furthermore, β′ phase precipitates could be tailored by aging treatment;

(2) The average grain size of the sintering Mg-Gd-Zn alloy was 6.3 μm. β₁ phase ((Mg, Zn)₃Gd) formed in both the as-cast ingot and RS ribbons, and the size grew from 35.6 nm in RS ribbons to 154 nm in SPS alloy. Moreover, a fine lamellar LPSO phase with the width of 2–30 nm formed from the supersaturated solid solution;

(3) The hardness of the alloy was significantly improved from 86 HV of the as-cast ingot containing both α-Mg and bulk β₁ phase to 126 HV of the sintering bulk, including nanosized multiple strengthening phases.