Understanding microstructural evolution during three-axial thermo-mechanical processing involving severe plastic deformation of magnesium alloys

ABSTRACT

Magnesium-rare earth alloy was selected as a model alloy system to explore and fundamentally understand the evolution of microstructure during three-axial thermo-mechanical processing involving severe plastic deformation. The microstructural evolution was studied via electron backscatter diffraction technique in a scanning electron microscope. Dynamic recrystallization mechanism during severe plastic deformation led to significant grain refinement of the alloy to ultrafine/fine-grained structure with random orientation of grains.

KEYWORDS:

• Magnesium alloy
• rare-earth element
• electron back-scattered diffraction

Introduction

Polycrystalline metals and alloys consist of microstructural constituents including grains, grain boundaries and precipitates. Electron backscatter diffraction (EBSD) combined with scanning electron microscope (SEM) enables orientation of individual grains, grain size distribution, local texture, local strain, point-to-point orientation correlation, and phase identification and distributions to be determined on the surface of polycrystalline materials [1]. EBSD has been widely used to study structure–property relationship. This is important because of the need to develop high strength-high ductility combination in metals and alloys [2–14]. However, the application of EBSD to understand the microstructural evolution during severe plastic deformation processing of alloys is limited.

Thus, the objective of the study described here is to develop an understanding of the microstructural evolution during different stages of three-axial (x-, y-, and z-axis) processing of a rare-earth (RE) element containing magnesium alloy via EBSD. This study enabled us to predict the underlying phenomenon responsible for the ultra-fine/fine-grained structure of the alloy. Magnesium-RE bearing alloy was selected as a model alloy system because of the significant interest in light-weight materials for automotive, aerospace, and biomedical applications.

Experimental procedure

Magnesium alloy of nominal composition, Mg-0.77Zn-0.32Gd (at%) was as-cast using traditional melting and casting methods. Prior to the three-axial (x-, y-, and z-axis) thermo-mechanical processing involving severe plastic deformation, the alloy was annealed to relieve internal stresses. The details of melting, casting, and processing are given elsewhere [3].

The steps in the EBSD sample preparation procedure and analysis were the following:

1. The specimens were cut into appropriate sizes, with length and width not exceeding 5 mm each, and a thickness not less than 0.4 mm.

2. SiC paper of different grits was used to grind the thickness of the specimen to 0.25-0.3 mm. The sequence of SiC used was: 600 - 800 -1500 - 2500.

3. The EBSD specimens were prepared by twinjet electropolishing (Fischione Instruments, Model 110 Automatic Twin-Jet Electropolisher) using a solution consisting of 5% perchloric acid and 95% alcohol at a applied potential of 20 V for 20–30 s, and the temperature was maintained at −30°C.

4. Following electropolishing, ion polishing was carried out by milling (Fischione Instruments, Model 1061) process at a potential of 6 kV for 15 min, with the temperature maintained at 0°C.

5. EBSD analysis was carried out using a Zeiss Gemini 300 scanning electron microscope equipped with a Symmetry S2 EBSD detector and Aztec Crystal software. An acceleration voltage of 20 kV was used, and the aperture size was set to 120 μm-diameter in a high current mode with a probe current of 12 nA. Variable step size was used from 0.1 to 2.5 micrometres, depending on the sample.

Results and discussion

The microstructure of the as-cast alloy in terms of information obtained by EBSD, band contrast (BC), inverse pole figure (IPF), and Kernel average misorientation (KAM) images is presented in Figure 1. In the EBSD figures, the square examination area in the as-cast samples was ~1000 × 800 µm² (collected with a step size of 2.5 µm). The grain size of the as-cast sample exhibits a normal distribution with a peak about 275 µm (as shown later in Figure 5(a)). In Figure 1(a), the IPF mapping shows the elongated grain morphology, some of which are over 200 micrometres in length. Moreover, the IPF and BC images (Figure 1(a,b)) reveal a fraction of twinning structures inside and cross elongated grains. In the KAM image (Figure 1(c)), the twinning structures across grains exhibit a higher lattice strain field. It can be suggested that these twinning structures accompanying the lattice strain concentration hinder the dislocation movement. Alternatively, some twin structures grown within grains presumably divide the grain size, providing a twin boundary strengthening effect. Additionally, a high density of precipitates, exhibiting black dot contrast in Figure 1(b), is observed, located within the matrix or adjacent to the twin boundaries, growing inside grains. The twin boundaries can be recognized as the nucleation sites for precipitates, thereby bringing about a lower lattice strain field.

Figure 1. (a) EBSD IPF, (b) EBSD BC, and (c) EBSD KAM mappings of the as-cast alloy.

Figure 2(a-f) present the IPF, BC, and KAM images of annealed alloy. In the EBSD figures, the square examination areas in the annealed samples were ~1000 × 800 µm² and ~120 × 180 µm² (collected with a step size of ~2.5 µm and ~0.5 µm), respectively. The sample contains the normal grain size distribution similar to that of the as-cast sample (as shown later in Figure 5(b)). The average grain size of the annealed sample is ~520 µm which is larger than that of the as-cast sample. In Figure 2(a,d), a small fraction of twin structures within grains is observed. It can be suggested that the annealing treatment enhances the growth of twins, resulting in a thicker twin structure. Moreover, Figure 2(b,e) reveals that coarsening and dissolution of precipitates followed by Ostwald ripening mechanism occurred during the annealing process. Alternatively, the KAM images (Figure 2(c,f)) indicate the presence of twin structures, which grow across the grains, and exhibit a lower lattice strain field compared to twins grown within grains.

Figure 2. (a, d) EBSD IPF, (b, e) EBSD BC, and (c, f) EBSD KAM mappings of the annealing alloy.

Similarly, low and high magnification, BC, IPF, KAM images of annealed alloy after x-, y-, and z three-axial processing are presented in Figure 3(a-f), respectively. After the first-round of three-axial severe plastic deformation, the grain size distribution of magnesium alloy is strikingly different from the annealed alloy. The average grain size of the first-round deformed samples is ~ 6.8 $\pm$ 6.8 µm. The large deviation in the average grain size is presumably associated with bimodal microstructure, which are composed of fine and coarse grain structures [15]. Figure 5(c) displays two peaks in the grain size distribution chart of the first-round deformed samples, respectively. The larger grains possess a peak distribution at about ~25.1 µm and the smaller grains, at about ~4.4 µm. It can be suggested that the stored lattice strain energy, which is caused by the three-axial severe plastic deformation, is non-uniformly released, thereby creating the bimodal structure after annealing. Moreover, in Figure 3(a,d), there is evidence of continuous dynamic recrystallization mechanism with a necklace-type structure and there are large elongated grains in the vicinity of extremely fine equiaxed grains. Many of these many elongated grains with near <$\bar{1}2\bar{1}0$> orientations are fragmented. The high magnification IPF image, Figure 3(d), better reveals the orientation distribution of grains. Different regions within the large grains show changes in orientation, indicating that different rotations occurred in different areas of the large grains during the deformation process. After three-axial severe plastic deformation, large grains can form slightly differently oriented compared to the small grains, which can be observed by comparing the BC and IPF image. In Figure 3(c,f), the KAM images indicate that when different regions within the large grains undergo rotations in different directions, higher strains are formed at the interfaces of different regions. The regions corresponding to fine grains have lower strain and are related to the fact that recrystallization releases the strain.

Figure 3. (a, d) EBSD IPF, (b, e) EBSD BC, and (c, f) EBSD KAM mappings of the first-round deformed samples.

As shown in Figure 4, the BC, IPF, and KAM, images after the second round of three-axial thermo-mechanical processing are presented. The effective grain size is further refined to ultrafine/fine-grained structure and the grains are equiaxed with a single peak of grain sizes around 3.3 µm, as shown in Figure 5(d). In Figure 4(a,d), the IPF images show that fine grains are randomly oriented and there is no preferential orientation of grains. However, some large grains appear to be oriented along the <$01\bar{1}0$> orientations. Alternatively, in Figure 4(c,f), the KAM images that show the strain distribution and the degree of strain suggest that the strain at grain boundaries is slightly greater than the regions in the grain interior. Grain boundaries are interfaces between individual grains in a polycrystalline material. Due to their different crystallographic orientation, they act as barrier to transmission of dislocations and impede the deformation process. As a result, strain can accumulate at grain boundaries, leading to a slightly higher strain compared to the regions in the grain interior.

Figure 4. (a, d) EBSD IPF, (b, e) EBSD BC, and (c, f) EBSD KAM mappings of the second round of deformed samples.

Figure 5. The grain size distribution charts show the grain structures of (a) the as-cast, (b) the annealed, (c) the first-round deformed, and (d) the second-round deformed samples.

The number fraction versus misorientation angle for the four different conditions is presented in Figure 6. It is apparent that after the first and second round of thermo-mechanical processing involving severe plastic deformation, there is a wide distribution of misorientation angles as compared to the as-cast and annealed alloy. This implies random orientation of grains in the ultrafine/fine-grained structure of the magnesium-rare earth element containing alloy. We attribute the random distribution of grains to the cumulative effect of rare-earth element (gadolinium) and three-axial thermo-mechanical processing involving severe plastic deformation, both of which activate slip on systems other than basal slip. While the significant refinement of grain size via dynamic recrystallization and fragmentation of grains occurred during severe plastic deformation process.

Figure 6. The misorientation angle distribution charts of (a) as-cast, (b) annealed, (c) first-round deformed, and (d) second-round deformed samples.

It is pertinent to indicate here that the final grain size after two-rounds of thermo-mechanical processing involving severe plastic deformation processing depends on the as-cast structure (grain size of columnar grains), which can vary depending on the casting conditions, despite similar processing conditions. The important aspect that is underscored here is that three three-axial thermo-mechanical processing involving severe plastic deformation is a promising process to obtain ultrafine/fine-grained structure.

Conclusions

The study of microstructural evolution as a function of different stages of thermo-mechanical processing involving severe plastic deformation by EBSD clearly suggested that significant grain refinement occurred via dynamic recrystallization such that ultrafine/fine-grained structure is obtained after two rounds of three-axial severe plastic deformation processing. Furthermore, the grains are randomly oriented.

Acknowledgments

RDKM sincerely thanks Professor Y.R. Yang, former colleague at the University of Cambridge for guidance and provided training to TFC.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.