Heterogeneous nucleation of β-type precipitates on nanoscale Zr-rich particles in a Mg-6Zn-0.5Cu-0.6Zr alloy

Zirconium (Zr) is an important alloying element to Mg-Zn-based alloy system. In this paper, we report the formation of the β-type precipitates on the nanoscale Zr-rich particles in a Mg-6Zn-0.5Cu-0.6Zr alloy during ageing at 180°C. Scanning transmission electron microscopy examinations revealed that the nanoscale Zr-rich [0001]α rods/laths are dominant in the Zr-rich core regions of the as-quenched sample after a solution treatment at 430°C. More significantly, these Zr-rich particles served as favourable sites for heterogeneous nucleation of the Zn-rich β-type phase during subsequent isothermal ageing at 180°C. This research provides a potential route to engineer precipitate microstructure for better strengthening effect in the Zr-containing Mg alloys.

The key strengthening precipitates in this alloy system have been considered as two types of Zn-rich precipitates, the rod-like β 1 ′ precipitates perpendicular to the (0001) α plane and the plate-like β 2 ′ precipitates parallel to the (0001) α plane [1][2][3][4][5]. Hardening by precipitation of β-type precipitates is believed to be the main strengthening mechanism of Mg-Zn-based alloys [1].
Recently, a peak-aged Mg-6Zn-0.5Cu-0.6Zr cast alloy has been reported to possess excellent mechanical properties with an ultimate tensile strength of 266.3 MPa, a 0.2% proof yield strength of 185. 6 MPa and an elongation of 16.7% [5]. Both the strength and ductility of the newly designed Mg-6Zn-0.5Cu-0.6Zr alloy are superior to those of the traditional Mg-6Zn-xCu-0.5Mn alloys [5,6]. Since Zr-rich particles may form after a solution treatment in Zr-containing Mg alloys [2,7,8], the present research aims to unveil the effect of these pre-existing nanoscale Zr-rich particles on the formation of the subsequent β-type precipitates of the Mg-6Zn-0.5Cu-0.6Zr alloy during age hardening.

Methods
The alloy with a nominal composition of Mg-6Zn-0.5Cu-0.6Zr (wt.%) for this study was prepared by melting high-purity Mg and Zn with Mg-28.78 wt.% Cu and Mg-31.63 wt.% master alloys, in a steel crucible and by casting into a permanent mould under an Ar atmosphere. Samples sectioned from the ingot were solution-treated for 24 h at 430°C. To investigate the microstructural evolution of the Zr-rich and Zn-rich precipitates, the water-quenched samples were subsequently aged in an oil bath for 20 and 120 h at 180°C. Thin foil specimens for scanning transmission electron microscopy (STEM) and transmission electron microscopy (TEM) were prepared by a twin-jet electropolisher using a solution of 10.6 g LiCl, 22.32 g Mg(ClO 4 ) 2 , 200 ml 2-butoxi-ethanol and 1,000 ml methanol at about −45°C and 70 V. The STEM study was conducted using a JEOL 2200FS microscope (JEOL Ltd., Tokyo, Japan) equipped with a high-angle annular   Figure 1a are enriched with Zn and Zr. This is in good agreement with the previous reports showing that Zrrich phases exist in various Zr-containing Mg-Zn-based alloys after a solution treatment [2,7,8]. EDXS analysis detected no enrichment of Cu in the Zr-rich particles.

Results and discussion
In order to investigate the effect of these pre-existing Zrrich particles on the formation of Zn-rich strengthening precipitates during subsequent isothermal ageing, HAADF imaging and EDXS mapping were conducted on samples aged at 180°C for different time. The 1 210 ½ α HAADF image of the 20-h-aged sample, as shown in Figure 2a, reveals that a dispersion of particles was mostly elongated along the [0001] α direction, with only one marked β 2 ′ perpendicular to the [0001] α direction. After tilting a large angle of approximately 51°to the 0 111 ½ α zone axis (Figure 2e), all particles observed in Figure 2a were found to be separate without overlapping with each other. The β 2 ′ precipitate, marked in Figure 2a, is a plate containing a brighter core, which corresponds to an enrichment of Zr (Figure 2b). The Zr map, Zn map and a combined Zr and Zn map, as shown in Figure 2b,c, and d, demonstrate that most of the elongated particles were composites containing a Zn-rich part and a Zr-rich segment. Careful examinations of the EDXS maps and the HAADF image confirmed that each Zr-rich segment was located either at the end or in the middle of an individual elongated precipitate. Therefore, we conclude that those Zr-rich segments of the precipitates are, in fact, the remains of the Zr-rich particles initially present in the as-quenched condition. We further deduce that these Zr-rich particles served as a precursor phase for the heterogeneous nucleation of Zn-rich β 1 ′ precipitates ([0001] α rods) and β 2 ′ precipitates ((0001) α plates) in the Zr-rich core regions of the Mg alloy during subsequent ageing. Figure 3 shows the HAADF image and the corresponding EDXS mapping result of the 120-h-aged sample. Both the length of [0001] α β 1 ′ rods and the thickness of (0001) α β 2 ′ plates grew significantly with the ageing time. The Zr map, Zn map and a combined Zr and Zn map, as shown in Figure 3b,c and d, indicate that many β 1 ′ rods and the β 2 ′ plate contain a Zr-rich segment. The sizes of Zr-rich segments observed in the 120-h-aged sample are smaller than those observed in the 20-h-aged sample. It appears that the size of the Znrich segments gradually increased at the expense of the Zr-rich segments during the isothermal ageing. After tilting approximately 36°from the 1 210 ½ α beam direction, a 1 543 ½ α HAADF image (Figure 3e) further confirms the existence of the Zr-rich segments in the Zn-rich precipitates. All experimental evidences above indicate that the heterogeneous nucleation on the pre-existing Zrrich particles is significantly important for the formation of Zn-rich precipitates (β 1 ′ and β 2 ′) in the Zr-rich core regions of the Mg alloy during ageing at 180°C.
To explore the crystallographic characteristics of these Zr-rich [0001] α rods, we examined the as-quenched microstructure using TEM with the beam parallel to the [0001] α direction, as shown in Figure 4. Most of the Zr-rich particles (>80%) of the as-quenched sample in Figure 4a have a low aspect ratio in the range of 1:1 to approximately 1:3 and a thickness in the range of 6 to approximately 12 nm with their long side, which is less than 25 nm, parallel to the < 11 20> α directions. They are Zr-rich [0001] α rod/ lath particles observed previously by STEM examinations (Figure 1a). The rest of the Zr-rich particles (<20%), marked with black arrows in Figure 4a, are thin rods with aspect ratios of 1:3 to approximately 1:20 and a thickness of 2 to approximately 5 nm, with their long axis approximately 23°away from the < 11 20> α directions. They are similar to the type C Zr-rich rods reported by Gao et al [8]. In contrast, the size and aspect ratio of the dominant Zrrich [0001] α rods/laths in the end-on view are significantly different from the Zr-rich < 11 20> α rods reported by Gao et al [8]. This difference is possibly due to the different alloy systems and the heat treatment techniques. Chemical microanalysis of these [0001] α rods using EDXS indicated that the atomic ratio of Mg:Zn:Zr was about 51:19:30 (inset, Figure 4a), suggesting that these [0001] α rods were Zr-rich precipitates with a Zn:Zr ratio close to 2:3. The corresponding micro-beam diffraction patterns (Figure 4b) confirm that these Zr-rich [0001] α  rods have a tetragonal structure similar to that of Zn 2 Zr 3 δ phase (a = b = 7.633 Å, c = 6.965 Å, α = β = γ = 90 [8,9]). The orientation relationship (OR) implied by the superimposed precipitate and matrix patterns was such that 1 10 ½ δ == 0001 ½ α , 110 ð Þ δ == 1 100 ð Þ α and 001 ð Þ δ == 1 120 ð Þ α . By combing the commonly reported OR between β 1 ′-MgZn 2 [3,10] /β 1 ′-Mg 4 Zn 7 [11,12] and α-Mg matrix with the OR of the δ-Zn 2 Zr 3 phase determined in this work, the possible ORs and the crystallographic disregistries between δ phase and β 1 ′ phase were determined and listed in Table 1. The inter-planar misfits between the matching planes 001 the directional misfits along the matching directions 1 10 ½ δ ÀZn 2 Zr 3 == 11 20 ½ β 1 ′ ÀMgZn 2 , 1 10 ½ δ ÀZn 2 Zr 3 == 001 ½ β 1 ′ ÀMg 4 Zn 7 were calculated as 2.5%, 5.4%, 5.1% and 3.2%, 1.8%, which are less than the critical values of 6% and 10% given in the edge-to-edge matching model [13]. The low lattice mismatch between these two phases explains why β 1 ′ rods form directly on the end plane (001) δ of the Zr-rich rods, as shown in Figures 2 and 3. The presence of the initial Zr-rich phases can provide much lower activation energy barrier and a favourable crystallographic correlation for the nucleation of the subsequent Zn-rich precipitates according to the classical nucleation theory [14]. It is a significant finding that the Zr-rich phases can act as the precursor phase for the heterogeneous nucleation of Zn-rich β-type strengthening phases in the Mg alloy, given that the Zr-rich core region is a major microstructural feature of Zr-containing Mg alloys [7,8]. By effectively engineering Zr-rich [0001] α rods in the Zrrich cores of Mg alloys using a solution treatment, the formation of [0001] α β 1 ′ rods could be promoted according to the heterogeneous nucleation mechanism revealed by this research.

Conclusions
In summary, we have demonstrated that the nanoscale Zrrich [0001] α rods/laths were predominant in Zr-rich core regions of the Mg-6Zn-0.5Cu-0.6Zr (wt.%) alloy after a solution treatment at 430°C. The nanoscale Zr-rich particles served as a precursor phase for the heterogeneous nucleation of the Zn-rich β-type strengthening precipitates during subsequent isothermal ageing at 180°C. These results are important for controlling Zr-rich particles in the Zrrich core regions for enhancing the overall strength of the Mg alloy.