Enabling manufacturing of multi-axial forging-induced ultrafine-grained strong and ductile magnesium alloys: a perspective of process-structure-property paradigm

ABSTRACT

Magnesium (Mg) alloys are difficult to cold/warm-process due to their hexagonal close-packed (HCP) lattice, which has restricted slip systems and makes plastic deformation at low temperatures a challenging task. Multi-axial forging (MAF) and annealing of as-cast alloys were combined to create a texture-free ultrafine-grained (UFG) structure with a high strength-high ductility combination to address this challenge. The study showed a clear and fundamental change from basal and pyramidal dislocation slip to twinning in the strong and ductile UFG Mg alloy compared to the low strength coarse-grained (CG) Mg alloy counterpart. This implied that the dislocation activity, grain orientation, and particular grain boundary states played an important role in controlling the deformation mechanisms. The potential impact of processing Mg alloys by MAF opens-up a new frontier of strong and ductile low-density materials for light and efficient solutions including energy absorption and formability and provides a perspective in terms of process-structure-property relationship.

KEYWORDS:

• Magnesium alloys
• triaxial forging
• microstructure
• structural and functional materials

1. Introduction and challenge in processing high strength-high ductility Mg-alloys

The element magnesium (Mg) is available in plenty in nature, and its alloys display a variety of desirable characteristics, such as high specific strength and stiffness, excellent damping and shock absorption, high thermal conductivity, and great electromagnetic shielding ability [1] [2–5]. Despite the fact that these characteristics support the use of magnesium alloys in the automotive, aerospace, biomedical, and packaging industries, there are three factors that prevent their broad application. The first is that magnesium alloys are weaker than their aluminium counterparts. Second, it is difficult to plastically deform magnesium alloys at low temperatures due to the hexagonal close-packed (HCP) lattice and limited slip systems that characterize magnesium. Finally, due to the HCP crystal structure, Mg-alloys have poor formability (ductility), and subsequent processing creates crystallographic asymmetry (texture effects).

In regard to the strength, Mg-rare earth (RE) alloys, either binary or with additional alloying elements, are currently of significant interest because of high yield strength that can be achieved through precipitation hardening. In recent years, a wide range of RE elements (La, Ce, Nd, Gd, Y) are being explored to increase the strength of Mg-alloys. Their addition has been promising in increasing the strength through precipitation hardening-induced by phases such as Mg₁₂RE, Mg₃RE or Mg₂RE phases [6].

One significant strategy for boosting engineering material ductility without sacrificing strength is grain refinement. Ultrafine-grained alloys have exceptional cyclic stability during fatigue and a combination of high strength and homogeneous elongation [7,8]. These characteristics have sparked new attention and suggested fresh approaches to alloy processing. When compared to materials that have relatively fewer grain boundaries and interfaces, UFG materials exhibit significantly different structure and properties [9]. This is because UFG materials have a high density of grain boundaries and interphase barriers.

There are some restrictions on both current severe plastic deformation methods and thermomechanical processing (TMP). One of the main techniques for obtaining grain refinement in metals is TMP, which involves strain-induced transformation or dynamic recrystallization. Recent TMP research conducted in the author’s lab [10,11] and at other institutions has shown that extreme plastic deformation processes that result in UFG structure can overcome the grain refinement restrictions imposed by traditional TMP [12]. Equal channel angular pressing (ECAP) [13], accumulative roll bonding (ARB)14,15], high pressure torsion (HPT) [16–19,, repeated compression [20], and upsetting extrusion [21] are the primary SPD techniques. However, as these are laboratory-scale procedures, they cannot be used to produce alloys with uniform and texture-free structures. A bi-modal grain size distribution is typically attained. As a result, it is still difficult to process strong, ductile, texture-free UFG Mg-alloys. We overcame this difficulty by combining the multi-axial forging (MAF) process with the annealing of as-cast magnesium alloy.

The objective here is to provide a perspective and overview of the promise of texture-free Mg-rare earth alloys with high strength-high ductility combination processed by multiaxial forging (MAF). The current understanding in terms of process-structure-property relationships that govern the manufacturing of texture-free ultra-fine grained (UFG) strong and ductile magnesium alloys processed by multi-axial forging with the ability to sustain large plastic deformation, as recently obtained by the author’s group is presented [1–3].

2. Key features and sketch of the multi-axial manufacturing route for Mg-alloys

We processed the as-cast Mg-RE alloy via a combination of annealing and multiaxial forging to obtain ultrafine-grained (UFG) structure with high strength-high ductility combination. In this regard, we considered a relatively dilute and model ternary alloy system, with minimum targeted strength of  ~200 MPa and elongation exceeding ~20% in the multiaxial forged alloy. A schematic illustration of the multiaxial forging process is presented in Figure 1 [1–3].

Figure 1. Schematic of multiaxial forging (MAF) process used to process Mg alloy [1–3]. .

In designing the lean and model Mg-RE alloy, the following aspects were considered after a critical review of the phase diagrams of Mg-RE alloys and fundamental knowledge available in the literature. (a) Gadolinium (Gd) has high solubility of 23.49 wt% at eutectic temperature in Mg [22] and significantly contributes to solid solution strengthening [23]. (b) Alloying Mg with 2 wt% Gd exhibits maximum tensile strength as compared to any other rare earth element for similar percentage of alloying elements. Beyond 2 wt% Gd, there is no significant effect on strength and may reduce the strength because of solute segregation. (c) Gd and Zn together form FCC w-phase with Mg (Mg₃Zn₃Gd) that is suitable for grain refinement. Based on the aforementioned aspects, a lean model alloy system with nominal chemical composition in wt% of Mg-2Zn-2Gd was considered, ensuring the formation of precipitation strengthening w-phase and simultaneously obtaining solid solution strengthening. The combination of annealing and MAF processing route is generic in nature and can be successfully extended to other Mg-alloys based on Nb-, Mo- or Ce-alloy design, in lieu of Gd, and to other grades of non-RE Mg-alloys.

3. Preparation and processing of Mg-2Zn-2Gd alloy

The alloy was melted in a resistance furnace under a protective atmosphere of Ar + 2% SF₆. Pure Mg, Zn, and Gd were used to prepare the designed alloy. First, Mg and Zn were melted in a boron nitride (BN)-coated mild steel crucible, and Gd was added to the liquid melt at ~770°C. This was followed by mechanical stirring to completely dissolve Gd and obtain uniform composition in the melt. Lastly, the liquid melt was poured into a preheated metallic round mould to obtain the as-cast alloy [1–3].

Prior to multiaxial forging (MAF), the as-cast alloy was annealed at 500°C for 2 h to relieve the internal residual stresses. Multiaxial forging (MAF) of as-cast Mg-2Zn-2Gd alloy was carried out at a using a forging press of. Rectangular samples were subjected to MAF, as schematically illustrated in Figure 1 [1–3]. The sample was placed in preheated die and MAF. During MAF, the dimensional ratio of the sample was maintained constant through the MAF process, while the loading direction was changed 90° from pass to pass [1,3].

The MAF device consisted of a punch for external loading and a die, and the punch moved vertically inside the chamber comprising of punch and die. The dimensional ratio of samples was maintained constant throughout the MAF processing, while the loading direction was changed 90° from pass to pass. The samples were MAF processed (two passes), and in each pass the true strain induced was 2.1, such that the total strain after the second pass was 4.2. We can produce different grain structures, from ultra-fine grained (UFG) to sub-micron (SMG) and fine-grained (FG) structure, by controlling the strain in each pass, a key parameter in the MAF process. As a result, we can adjust the structure to have appropriate mechanical properties.

After the first pass, the microstructure was heterogeneous, consisting of coarse deformed grains and fine dynamic recrystallized grains. Some grains were oriented parallel to their (0001) plane, implying the presence of some texture. However, after the second pass, a near homogeneous structure with recrystallized grains was present with randomization of texture (see below).

The multiaxial forging (MAF) was carried out in the temperature range of 350–500°C and working ratio of 0.3–0.4/condition. The preliminary data reported here is for the optimized MAF conditions (temperature of 450°C, working ratio of 0.3/direction) which provided a uniform structure.

4. Microstructure

As-cast and forged samples were mechanically polished using a series of SiC papers. Final polishing was carried out using 0.2 μm alumina. The samples were etched for ~2 minutes in a solution containing 8 g picric acid, 5 ml acetic acid, 10 ml distilled water, and 100 ml ethanol. The microstructure of polished samples was examined by X-ray diffraction (Bruker D8 model) and field emission scanning electron microscopy (Hitachi S4800). To analyse different phases, EVA software available with the XRD equipment was used. Transmission electron microscopy was carried out using 3 mm disks that were electropolished in a solution containing 3% perchloric acid in ethanol (A-C). Figure 2 summarizes the X-ray diffraction data of as-cast and as-cast + annealed + MAF alloy indicating the presence of α-Mg and w-phase (Mg₃Zn₃Gd).

Figure 2. X-ray diffraction data of as-cast, and as-cast + annealed + MAF alloy indicating the presence of α-Mg (matrix) and w-phase (Mg₃Zn₃Gd) [1,2].

Figure 3 [1,2] shows the microstructure of as-cast + annealed and as-cast + annealed + MAF alloy. During annealing, grain growth occurred and the average grain size was increased from~25 µm in the as-cast alloy to 44 ± 2 µm in the annealed alloy. After multiaxial forging, the grain size of the annealed alloy was significantly refined to 416 ± 140 nm. An efficient method to disrupt the grain structure and subdivide the grains during multiaxial forging was to rotate the sample 90° around its axis at each phase, producing an ultrafine-grained (UFG) alloy. We consider here the as-cast and annealed alloy to represent a coarse-grained (CG) structure (average grain size ~44 ± 2 µm) and multiaxial forged alloy to represent UFG structure (average grain size ~416 ± 140 nm). Figures 3c,d illustrate the morphology and size of precipitates observed in the α-Mg matrix. The near-spherical precipitates in the α-Mg matrix were of size range ~50–80 nm and from the selected area electron diffraction pattern were identified to be Mg₃Zn₃Gd with orientation relationship of ($2\bar{2}01$)[$\bar{1}2\bar{1}6$]_α-Mg//(220)[$\bar{1}$11]_Mg3Zn3Gd.

Figure 3. Microstructure of as-cast + annealed alloy referred as coarse-grained (CG) and as-cast + annealed + multiaxial forged referred as ultrafine-grained (UFG) Mg-2zn-2gd alloy. (c) Transmission electron micrograph illustrating size range of the Mg3Zn3Gd precipitates in α–Mg matrix, as confirmed by (d) electron diffraction pattern [1,2].

From the pole figures of CG (as-cast + annealed) Mg alloy (Figure 4) [1], it may be seen that before MAF, in the CG alloy there was a strong texture on all easy slip planes with the maximal value of 52.59 (Figure 4a), implying strong anisotropy in the Mg alloy before MAF. In summary, before forging almost all the planes had maximal texture. However, after MAF, the majority of grains had no obvious texture on the slip planes including prismatic plane <a> $\left\{1\bar{1}00\right\}$, pyramidal plane <a> $\left\{1\bar{1}01\right\}$, pyramidal plane <a+c> $\left\{11\bar{2}2\right\}$, extension twin plane $\left\{10\bar{1}2\right\}$ and compression twin plane $\left\{10\bar{1}1\right\}$. Even on basal plane <a>$\left\{0001\right\}$, the maximal value of texture was only 8.29 (Figure 4b). This clearly implied that the MAF significantly suppressed texture. There was some texture at {0001}. The significant change in texture is attributed to the combined effect of MAF and Mg₃Zn₃Gd particle-simulated nucleation of new grains. Dynamic precipitation of Mg₃Zn₃Gd impacts nucleation of new grains and also hinders grain growth. It may also be noted that FCC precipitates inherited from the as-cast alloy have high deformability during MAF and fragmented during forging. It is also likely that besides dynamic recrystallization, grain fragmentation via kinking mechanism occurred during MAF [1–3].

Figure 4. Pole figures at different slip planes of magnesium alloy before (a) and after (b) multiaxially forging process (MAF) obtained by EBSD [1].

5. Mechanical properties of MAF Mg-alloy

Figure 5 shows stress-strain plots of as-cast and multiaxial-forged alloy. The as-cast alloy had a low yield strength of 53 MPa and significant precipitation was observed along the grain boundaries which may introduce stress concentration, resulting in early fracture [3–5]. On annealing, the yield strength was decreased from 53 to 46 MPa because of grain growth. But, when the as-cast alloy was annealed and multiaxial forged, the yield strength was increased to 227 MPa, with tensile strength of 272 MPa, and high elongation of 30%. Direct multiaxial forging of as-cast alloy gave inferior properties, yield strength of 170 MPa, tensile strength of 220 MPa, and elongation of 16%. In summary, the MAF process involving forging along the three-dimensions and controlling strain/pass enabled texture-free and UFG structure with random orientation of grains to be obtained. The combination of UFG structure and mechanical properties was excellent for the ‘lean’ Mg-alloy after MAF [Table 1]. Texture-free and homogenous structure is difficult to obtain in other severe plastic deformation processes described in Section 2.

Figure 5. Stress-strain plots ofas-cast and as-cast + annealed + MAF Mg-2zn-2gd alloy [1,2]. .

Table 1. Average grain size and mechanical properties of CG (as-cast + annealed) and UFG (as-cast + annealed + MAF) (multiaxial forged, 2 passes) Mg-2zn-2gd alloy [1,2].

Property	Coarse-grained (CG) Alloy (As-cast + 500°C Annealed)	Ultrafine-grained (UFG) Alloy (As-cast + 500°C annealed + MAF process)
Grain Size	44±5 µm	416±140 nm
YS (MPa)	46±2	227±2
UTS (MPa)	70±2	272±1
% Elongation	7±1	30±2

6. Nanoscale deformation experiments (UFG versus CG alloy)

The following justifications led to the execution of nanoscale deformation experiments. In UFG structures where indentation is contained within ‘a single grain’, nanoindentation can first provide insights into tiny deformation mechanisms [24–26]. Second, the tested volume is scalable with respect to the microstructure, allowing us to examine the fundamental processes behind discrete deformation in a tiny amount of the material with a very low risk of encountering pre-existing dislocations prior to the start of plastic deformation. Third, single specimen geometry and loading configuration are used in nanoindentation. When paired with an applied force of the order of millinewtons (mN), a typical nanoindentation experiment’s contact area of nearly 100 nm² results in an applied stress in the GPa range [27–29].

The maximum load for the load-controlled mode of the nanoindentation tests was set at 25 mN, and the loading rate was 1 mN/s. The Berkovich three-sided pyramidal diamond indenter, with a nominal angle of 65.3° and an indenter tip diameter of 20 nm, constituted the nanoscale deformation experiment [3].

Based on TEM studies, the following slip systems in multi-axial forged (MAF) UFG Mg-alloy were found. Figures 6 and 7 show, respectively, post-mortem TEM analyses of the plastic zone surrounding the depression for UFG and CG Mg-alloy [1–3]. The TEM studies on the UFG alloy revealed substantial activity in around 60% of the grains and dislocation interaction. Complete Burgers vector analysis in more than 25 grains served as the foundation for the analysis. The UFG alloy indicated high density of dislocations characteristic of 1/3 < 11$\bar{2}$3> pyramidal slip systems, that were observed in the two-beam condition of g = [0002], while 1/3 < 11$\bar{2}$0> basal slip systems cannot be seen in this condition. The occurrence of <c+a> dislocation slip changes the active deformation mode on the pyramidal planes in the Mg alloy.

Figure 6. Representative bright field (a) and dark field transmission electron micrographs illustrating high density of dislocations in the plastic zone surrounding the indentation (top-right corner) in the ultrafine-grained (UFG: as-cast + annealed + multiaxial forged) Mg-2zn-2gd alloy. Zones 1 and 2 have high density of dislocations and zones 3A and 3B illustrate dislocations emitted by the grain boundary. (c) Selected area diffraction (SAED) pattern for grain in (a) and (b) showing [0002] operating reflection [3].

Figure 7. (a,b) Transmission electron micrograph of plastic deformation zone surrounding the indentation showing the occurrence of twinning in the CG Mg-2zn-2gd alloy and the diffraction pattern. (b) Another illustration of twinning in the CG alloy. Diffraction pattern was similar to a (ii). From the diffraction pattern, the twin {11 ̅01} <01 ̅12> belongs to the family of {101 ̅1} <101 ̅2> compression twin. a(i) and a(ii) show original and enlarged (and indexed) SAED pattern for (a) [3]. .

A typical illustration of the dislocation structure connected to nanoindentation is shown in Figure 6 [3]. Figure 6a shows a dark grain with extremely strong dislocation activity, which is also visible in Figure 6b’s dark-field image as a bright grain. It was analysed using the [0002] operating reflection in the SAED pattern of Figure 6c. In the bottom part of left grain in Figure 6a, dislocation profiles (zones 3A, 3B) are shown being emitted from the grain boundary, while the overall dislocation density in this grain is significantly lower than in the central (dark) grain in Figure 6a. Only two sources are activated by these profiles, which function as pile-ups against the opposing (lower) grain boundary and emit the dislocation profiles illustrated at 3A and 3B [3].

Table 2 highlights the frequency and type of dislocations discovered in the UFG alloy by post-mortem TEM investigation based on numerous TEM micrographs. Complete Burgers vector analysis of at least 25 randomly chosen grains from at least 5 separate samples, is the foundation of the analysis.

Table 2. Frequency of dislocations observed in the UFG Mg-alloy via post-mortem TEM analysis. The analysis is based on complete Burgers vector analysis in at least 25 random selected grains.

Dislocation type	Dislocation Burgers Vector	Amount (%)
Basal or Prismatic <a>	Perfect b = 1/3 [11$\bar{2}$0] and b = 1/3 [1$\bar{2}$10]	32
	Dissociated (SFI1+SFI2) b = 1/3 [1$\bar{2}$10]	2
Pyramidal <c+a>	Perfect b = 1/3 [11$\bar{2}$3] screw	32
	Perfect b = 1/3 [11$\bar{2}\bar{3}$] screw	2
	Perfect b = 1/3 [$\bar{2}11\bar{3}$] mixed	8
	Dissociated b = 1/3 [$\bar{2}$113]	18
	= 1/6 [$\bar{2}$023] + 1/6 [$\bar{2}$203]	6

According to the g.b. criterion, where b is Burgers vector and g is diffraction vector, dislocations with different Burgers vectors are visible when g.b#0 and invisible when g.b = 0. TEM analysis revealed that <c+a> dislocations of different vectors were present (Table 2), namely, b = 1/3[11$\bar{2}$3], b = 1/3[11$\bar{2}\bar{3}$] and b = 1/3[$\bar{2}$11$\bar{3}$]. Additionally, the collinear dissociation of b = 1/3[11$\bar{2}$3] dislocations according to 1/3[$\bar{2}11$3] → 1/6[$\bar{2}$023]+1/6[$\bar{2}$203] was observed. The observation of second order [$\bar{2}$113] pyramidal slip and <c+a> dislocations suggest that screw <c+a> dislocations are stable, while edge <c+a> dislocations decompose <a> and <c> dislocations [3,30].

In contrast to the high dislocation activity in the UFG structure, {10$\bar{1}$1} nanoscale twinning was an active deformation mechanism in the CG structure, however, the frequency of observation of twinning was ~15% considering all the grains. Typical twinning observations are shown in Figure 7. They exhibited similar orientation with the matrix. From the SAED pattern (Figure 7 a(i), (ii)), the type of the twin is {1$\bar{1}$01}[0$\bar{1}$12], which belongs to the family of {10$\bar{1}$1}<10$\bar{1}$2> compression twinning, consistent with the earlier studies on coarse-grained Mg alloys [30, 31,32]. The formation of twins early during the deformation process accommodates the applied strain [3,33].

The enhanced activity on non-basal planes in the UFG alloy is attributed to modified Peierls potential [3,34–39]. FE. Thus, the explanation of SFE varying with grain size is considered untenable.

As a result, under the same processing conditions, Figures 6 and 7 suggested a distinct distinction and fundamental transition in the deformation mechanism of CG and UFG alloy in a single material. The texture-free UFG Mg-RE alloy reduced twinning. Pyramidal non-basal <c+a> slip was activated in the UFG Mg alloy and twinned in the CG counterpart.

Load–displacement plots also showed the distinct plastic deformation of MAF UFG and CG Mg-alloys. Figure 8 [3] shows representative force-displacement plots for the alloys UFG and CG. Three distinct displacement bursts (pop-ins) were observed in the case of the UFG alloy, with the first one occurring at a displacement of about 70 nm and an applied load of about 0.12 mN. In contrast, in the CG alloy, there were no displacement bursts and deviation from Hertzian solution occurred at ~60 nm. The loading data prior to the first pop-in is described by a power law relationship (P~h^1.5), consistent with Hertzian elastic contact solution [40,41]. The broken-line in Figure 8 depicts the theoretical Hertzian elastic behaviour, which is consistent with the experimental plot. Moreover, it suggests that the region is elastic before the first discontinuity or pop-in and plastic thereafter.

Figure 8. Representative load-displacement plots for UFG and CG Mg-2zn-2gd alloy acquired via nanoindentation experiments in the load-controlled mode with a Berkovich indenter of radius 20 nm (maximum load of 0.5 mN and loading rate of 2 μN/s). (It may be noted that similar plots were observed in a large number of grains of each material, irrespective of their orientation of grains and hence they truly reflect the behaviour of the alloy) [3].

We are looking at a small portion of the material that is thought to have a low chance of coming across a pre-existing dislocation (essentially defect-free) [3], with the indenter tip’s diameter (20 nm) and the maximum displacement of about 200 nm (Figure 8). Thus, the pop-ins represent burst or nucleation of dislocations in the UFG alloy and are expected to occur when the maximum shear stress (τ_max) approaches the theoretical value, τ_theo. In nanoindentation, stresses can reach the theoretical strength of the material [41,42], as determined by Equations (1)(1) $P=\frac{4}{3}{E}^{\ast }\sqrt{R}{h}^{1.5}$(1) to (3(3) $p_0={\left(\frac{6P{E}^{\ast 2}}{{\pi }^3{R}^2}\right)}^{1/3}$(3) ):

(1) $P=\frac{4}{3}{E}^{\ast }\sqrt{R}{h}^{1.5}$(1)

where P is the applied load, E^* is the effective indentation modulus, R is the radius of the Berkovich indenter tip, and h is the corresponding indentation depth. The effective indentation modulus is related to the Young’s modulus, E, and Poisson’s ratios, υ, of the sample and the indenter via [41]:

(2) $E^{\ast }={\left(\frac{1-{\nu }_s^2}{E_s}+\frac{1-{\nu }_i^2}{E_i}\right)}^{-1}$(2)

The subscripts i and s refer to the indenter and sample, respectively.

Using the isotropic elastic constant for Mg-Gd alloy (E_s = 52.5 GPa and υ_s = 0.28 [43]) and E_i = 1141 GPa and υ_i = 0.07 for the diamond indenter tip [41], a value of effective indentation modulus of E^* = 54 GPa was obtained. Hertzian analysis of the data from the elastic portion of CG and UFG in Figure 8 gave similar E^* = 32 GPa and E^* = 41 GPa, respectively [3].

The maximum shear stress at the first pop-in was determined using τ_max = 0.31p₀, where [9],

(3) $p_0={\left(\frac{6P{E}^{\ast 2}}{{\pi }^3{R}^2}\right)}^{1/3}$(3)

based on Hertzian analysis. τ_max was determined to be ~3.6 GPa for UFG and ~2.7 GPa for CG using Equation (3)(3) $p_0={\left(\frac{6P{E}^{\ast 2}}{{\pi }^3{R}^2}\right)}^{1/3}$(3) . The measured τ_max is ~1/7th of theoretical shear modulus of Mg-Gd alloy at ~21 GPa [10] and is within the range for the theoretical strength of polycrystalline metals (~G/30 to ~G/5) [3,44].

6.1. Conclusions

To produce UFG Mg alloys with high strength-high ductility combination that are nearly texture-free, multiaxial forging of magnesium alloys is a practical manufacturing technique. The improvement in grain structure, which includes grain division and fragmentation as well as uniform distribution of nanoscale precipitates, is associated with the increase in strength. While the finding of a texture-free alloy is thought to be the result of MF and particle-stimulated grain nucleation working together over time. A new route for processing magnesium-based alloys for use in aerospace, automotive, and biomedical applications is made possible by using MAF to create UFG Mg alloys.

Acknowledgments

The author is grateful to the National Science Foundation for financial support through grant # CMMI 2130586.

Disclosure statement

No potential conflict of interest was reported by the author.

Additional information

Funding

The work was supported by the National Science Foundation.