Microstructure and mechanical behaviour of CMT-welded Mg/Al dissimilar joint using Inconel 625 as filler metal

Abstract

Mg/Al dissimilar lap joint was produced by modified cold metal transfer process using Ni-based alloy wire Inconel 625 as filler metal. The characteristics of microstructure and mechanical behaviour of the joint were investigated. Microstructural analysis shows that the weld interface near Mg substrate consists of six types of areas: Mg₂Al₃-based area, Mg₁₇Al₁₂-based area, Al solid solution-based area, fusion zone-based area, Mg + Mg₁₇Al₁₂ eutectic area, and mixed area. With the characteristic of quasi-cleavage fracture, the tensile-shear load of the lap joint reached to 1.28 kN. The failure occurred primarily at the Mg₂Al₃-based area within the weld interface near Mg substrate. However, the harmful effect of the Mg₂Al₃ intermetallics on the strength of welded joint is limited.

Keywords

Introduction

With the continuously increasing requirement of lightweight constructions, there are a large number of Mg/Al hybrid structures required in the industrial field, especially in the automobile industry [1]. Schubert et al. [2] reported that a large potential Mg-to-Al assemblies required to be welded for mass saving of automotive applications in order to increase the fuel efficiency and reduce the greenhouse gases emission. The problem of Mg/Al dissimilar welding has to be solved. However, the application of Mg/Al dissimilar welded joint is still confined.

The challenge of Mg/Al dissimilar welding is the formation of hard-brittle intermetallic compounds (IMCs) which are detrimental to the joint strength, such as Mg₂Al₃ and Mg₁₇Al₁₂ [3]. To improve the joint strength, eliminating or reducing the Mg-Al IMCs reaction layer is essential. Numerous corresponding attempts using various welding processes were conducted with or without filler metal (or interlayer) [4 -13]. The results suggest that the formation of IMCs is inevitable in the Mg/Al dissimilar welded joints. Nonetheless, two feasible approaches can be concluded to improve the joint strength: reducing Mg-Al IMCs and improving their distribution state by controlling of the welding process, or replacing Mg-Al IMCs by the IMCs with relatively low brittleness using proper filler metal.

Cold metal transfer (CMT) welding, with the excellent characteristics in terms of low energy input, high deposition rate, no spattering, and extremely stable arc [14], has become a frontier, as well as a hotspot in the welding field. Since the formation of IMCs can be decreased greatly, CMT has been widely studied on dissimilar metals welding-brazing process [15 -19]. Currently, CMT is already used to join aluminium and steel for mass saving in automotive manufacturing. Owing to the low energy input and optional wire as filler metal, CMT is also a promising method to join Mg/Al dissimilar metals for vehicle body assembly. Some studies of Mg/Al dissimilar CMT welding were performed using Al-based or Mg-based wire as filler metal [20 -24], but the Mg-Al IMCs interface layers were all observed, resulting in relatively low joint strength. Thus, using proper wire as filler metal couple with the appropriate control of energy input process become the key to improve CMT-welded Mg/Al dissimilar joint.

It was noted that Ni interlayer has been used for Mg/Al dissimilar welding to improve the joint strength, such as diffusion welding [25], resistance spot welding [26], friction stir welding [27], and laser welding [28, 29]. Chang et al. [27] pointed out that using a Ni foil as filler material for friction stir welded Al6061-T6/AZ31 alloy showed significant improvement in the joint strength due to the presence of less brittle Ni-based IMCs instead of Mg₁₇Al₁₂. Yang et al. [28] claimed that no joints were achieved without Ni interlayer in Mg/Al dissimilar laser-welding/brazing process. By inserting a Ni interlayer, the joint strength increased from 0 to 410 N with the failure at Mg/Ni interface. Based on the results of these previous studies, it can be inferred that using Ni-based wire as filler metal could be a promising candidate to improve CMT-welded Mg/Al dissimilar joint. However, the characteristics of Mg/Al dissimilar CMT welding using Ni alloy wire as filler metal are still unknown.

In the present work, a modified CMT welding process was conducted to join Mg-AZ31B/Al-6061 dissimilar alloys, and the widely used Ni-based alloy wire Inconel 625 was selected as filler metal. The weldability and macrostructure features were studied. The microstructure of the fusion zone was investigated. Subsequently, the characteristics of the interface microstructures and the fracture behaviour were emphatically analysed and discussed in order to reveal the relationship between the IMCs characteristics and the joint strength.

Materials and experimental

In this study, 3-mm-thick Mg-AZ31B and Al-6061 sheets (150 mm × 75 mm) were used to produce CMT-welded Mg/Al dissimilar joint, and the welding wire of Ni-based alloy Inconel 625 with a diameter of 1.2 mm was selected as filler metal. Their chemical compositions are listed in Tables 1 and 2. Before welding, the surface of substrates was cleaned and degreased by stainless steel wire brush and acetone to remove the oxide films and oil.

Table 1

Chemical compositions of the Mg/Al dissimilar substrates (wt-%).

Materials	Mg	Al	Zn	Si	Mn	Fe	Cu	Ni	Ti	Cr
AZ31B	Bal.	2.9	1.03	0.019	0.42	0.0021	0.00078	0.00063	–	–
6061	0.8–1.2	Bal.	0.25	0.4–0.8	0.15	0.7	0.15–0.4	–	0.15	0.04–0.35

Table 2

Chemical composition of the Ni-based welding wire (wt-%).

Materials	Ni	Cr	Mo	Nb	Ti	Fe	Si	C	Mn	P	S
Inconel 625	64.389	22.7	8.7	3.56	0.21	0.21	0.04	0.012	<0.01	0.003	0.001

The welding configuration is demonstrated in Figure 1: the Al-6061 sheet was placed on the top of the Mg-AZ31B sheet with an overlap distance of 10 mm. A Fronius CMT Advanced 4000R arc welding system was employed with welding speed of 8 mm s⁻¹ and shielding gas (pure argon) flow rate of 20 L min⁻¹. To achieve effective Mg/Al dissimilar welding, based on our previous research [30], a modified Inconel 625 CMT process was conducted by control of the current and integrated wire motions using the characteristic parameters listed in Table 3. As a result, the wire feed speed was about 3.5 m min⁻¹.

Figure 1

Schematic of CMT welding configuration for Mg/Al dissimilar metals using wire Inconel 625.

Table 3

The characteristic parameters of modified Inconel 625 CMT welding process.

Filler metal	I_boost/A	t_I_boost/ms	I_sc_wait/A	I_sc2/A	vd_sc_ wait/m min⁻¹
Inconel 625	200	3	60	50	5

For metallography analysis, the cross-section specimen was prepared and polished. To reveal the microstructures of welded joint, the Al-6061 side and Mg-AZ31B side were etched in a solution of 5% hydrofluoric acid for 30 s and 4% nital for 180 s, respectively. Then it was observed and analysed by HITACHI SU8010 scanning electron microscope (SEM) coupled with a EDAX E1506-C2B energy dispersive spectrometer (EDS). It should be noticed that the values of EDS point analysis results are dependent on the area covered by the electron beam which is about 1 µm × 1 µm (or dependent on the area/particle size). According to AWS B4.0:2007, the tensile-shear test was conducted with a loading rate of 0.5 mm min⁻¹ at room temperature. The test specimens with the width (length of welded seam) of 20 mm were produced by wire cutting. To minimise the bending stresses, shims were used at each end of the specimens during the testing. Subsequently, the fracture surface was observed and analysed by SEM coupled with EDS and PANalytical Empyrean X-ray diffraction (XRD).

Results and discussion

Weldability and macrostructure

For Mg/Al dissimilar CMT welding process using Ni-based wire as fill metal, the major difficulty is the formation of magnesium vapour which is detrimental to the metal transfer process. As a result, the Mg/Al dissimilar butt joint was not obtained because the directly contact between the high-temperature Ni-based droplets and low-boiling Mg substrate caused a large number of magnesium vapours. In this study, to reduce the formation of magnesium vapour, lap welding was performed. On the one hand, as Mg substrate was placed under Al substrate, the directly contact between the droplet and Mg substrate was avoided. On the other hand, as characteristic parameters of the CMT welding process was modified, the temperatures of both droplet and substrate were reduced (according to our previous research [30], slow wire feed speed coupled with low burn phase current will cool the droplet and substrate during the burn phase of the CMT welding process).

As a result, Mg-AZ31B and Al-6061 sheets were effectively welded by the modified Inconel 625 CMT process. The macrostructures of welded joint are showed in Figure 2. On the top surface, some spatters were found due to the vapour pressure of magnesium hindered the droplets transfer during the welding process, as showed in Figure 2(a). On the cross-section, since the most of energy input was used to heat Al substrate which was on the top of Mg substrate, a lot of Al substrate was melted and flowed into the fusion zone (FZ) while a few Mg substrate was melted, as showed in Figure 2(b). Moreover, it can be observed that some porosities formed near Mg substrate.

Figure 2

Macrostructures of the CMT-welded Mg/Al dissimilar joint using wire Inconel 625: (a) top view, and (b) cross-section.

Microstructure of fusion zone

The microstructure of fusion zone was basically uniform due to the rapid flow and diffusion of Al melts (6061 substrate) in the high-temperature weld pool (mainly consist of molten wire Inconel 625). Figure 3 shows the microstructures in the centre of fusion zone, and the EDS analysis results are listed in Table 4. Since a lot of Al substrate was melted, the chemical composition of fusion zone with high Al content of about 28.5 at.-% (while Mg content is only 1.6 at.-%) clearly differs from that of Inconel 625 alloy. As a result, the fusion zone is mainly characterised with grey AlNi-based equiaxed grain [31], and there are lots of white Ni-rich and Cr-rich phases [32] in the grain boundaries. In addition, the microhardness of the fusion zone is 361.7 HV, which is much higher than that of Inconel 625 alloy (260 HV) [33].

Figure 3

Microstructures of the fusion zone: (a) centre of fusion zone, and (b) zone A shown in Figure 3(a).

Table 4

EDS analysis results of the various zones shown in Figure 3 (at.-%).

Position	Al	Mg	Ni	Cr	Mo	Nb	Possible phases
P1	41.3	–	49.3	9.4	–	–	AlNi + some Cr solid solution [31]
P2	10.2	2.5	48.0	25.8	5.1	8.4	Ni solid solution + some Cr₂Nb + few AlNi₃ [32]
P3	4.6	1.5	11.7	66.5	15.7	–	Cr solid solution + some Mo solid solution
Zone A	28.5	1.6	46.7	18.9	3.0	1.3	(P1 + P2 + P3)

Interface microstructure near Al substrate

Figure 4 shows the EDS map analysis results of the weld interface near Al substrate. Since the melted Al substrate flowed and diffused rapidly in the high-temperature weld pool during the welding process, there is a clearly boundary between the Al substrate and fusion zone. However, the high concentration gradients caused the Ni and Cr atoms (two major elements of wire Inconel 625) to diffuse from fusion zone into Al substrate at the weld boundary, and the Ni atoms would diffuse further than Cr atoms due to the high content of Ni elements of fusion zone. Subsequently, the Al-riched Al-Ni and Al-Cr phases, which has the relatively low melting point, can be formed within the Al substrate close to the weld boundary. According to the previous studies [32, 34] and EDS analysis results shown in Table 5, the microstructures of the weld interface near Al substrate shown in Figure 5 were analysed. Within the Al substrate near weld boundary, there are some reticular Al₃Ni intergranular precipitates in the grain boundaries of Al solid solution, as shown in Figure 5(b). Within the Al substrate next to weld boundary, there are lots of massive Al₃Ni phases and granular Al₇Cr (Al₁₃Cr₂) phases in the Al solid solution matrix, as shown in Figure 5(c).

Figure 4

EDS map analysis results of the weld interface near Al substrate.

Figure 5

Microstructures of the weld interface near Al substrate: (a) zone A shown in Figure 4, (b) zone B shown in (a), and (c) zone C shown in (a).

Table 5

EDS analysis results of the various zones shown in Figure 5 (at.-%).

Position	Al	Mg	Ni	Cr	Mo	Possible phases
P1	96.6	3.4	–	–	–	Al solid solution matrix
P2	92.3	3.0	4.7	–	–	Al₃Ni [32] + Al solid solution matrix
P3	85.2	3.7	11.1	–	–	Al₃Ni + Al solid solution matrix
P4	95.4	3.3	1.0	0.3	–	Al solid solution matrix
P5	94.7	4.4	0.9	–	–	Al solid solution matrix
P6	74.4	3.2	22.4	–	–	Al₃Ni + few Al solid solution matrix
P7	73.8	3.4	22.8	–	–	Al₃Ni + some Al solid solution matrix
P8	81.8	1.7	2.6	12.3	1.6	Al₇Cr (Al₁₃Cr₂) [34] + few Al solid solution matrix
P9	82.5	2.1	3.7	10.2	1.5	Al₇Cr (Al₁₃Cr₂) + few Al solid solution matrix
P10	78.2	1.2	12.6	7.3	0.7	Al₇Cr (Al₁₃Cr₂) + Al₃Ni
P11	79.1	1.5	19.4	–	–	Al₃Ni + some Al solid solution matrix
P12	49.3	1.6	43.2	5.9	–	AlNi + some Cr solid solution [31] (similar to P1 in Figure 3(b))

Interface microstructure near Mg substrate

Since some of the melted Al substrate flowed to the bottom of the fusion zone, the hard-brittle Mg-Al IMCs would form near Mg substrate. Figure 6 shows the EDS map analysis results of the weld interface near Mg substrate, which can confirm the enrichment of element Al at the weld interface. However, the distributions of the main elements within the weld interface are all uneven, indicating that there are several different types of areas with different microstructures formed within the weld interface near Mg substrate, as shown in Figure 6(b).

Figure 6

EDS map analysis results of the weld interface near Mg substrate: (a) micrograph, (b) phase map, and (c) distribution of elements.

Figure 7 shows the microstructures of the different areas within the weld interface near Mg substrate. For Mg-Al IMCs, Wang et al. [24] suggested that the relatively wide ranges of the composition of Mg₂Al₃, Mg₁₇Al₁₂, and Mg₁₇Al₁₂ + α-Mg solid solution eutectic were about 37–45, 45–60, and 60–70 at.-% Mg, respectively. According to the previous studies [24, 32, 34, 35] concerning the related binary/ternary phase diagram, EDS point analysis of the various zones were conducted to investigate the microstructure characteristics. The results are listed in Table 6. It can be concluded that the weld interface near Mg substrate mainly consist of six types of areas: Mg₂Al₃-based area mainly composed of Mg₂Al₃ matrix, Al₃Ni phase, Al₇Cr (Al₁₃Cr₂) phase, and Al–Fe-Ni phase, as shown in Figure 7(a,b); Mg₁₇Al₁₂-based area mainly composed of Mg₁₇Al₁₂ matrix and Al₃Ni phase, as shown in Figure 7(c,d); Al solid solution-based area near the fusion zone, which mainly composed of Al solid solution matrix, Al₃Ni phase, and Al₇Cr (Al₁₃Cr₂) phase, as shown in Figure 7(e); fusion zone-based area with the similar microstructure and composition compared to fusion zone, as shown in Figure 7(f); Mg + Mg₁₇Al₁₂ eutectic area near the fusion line, which composed of Mg + Mg₁₇Al₁₂ eutectic matrix and Mg solid solution, as shown in Figure 7(g); and mixed area mainly composed of Mg₂Al₃ matrix, Mg solid solution, AlNi phase, and Al₇Cr (Al₁₃Cr₂) phase, as shown in Figure 7(h). As shown in Figure 7(b), for Mg₂Al₃-based area, the stripy Al–Fe-Ni phases [35] denoted by point 1, Al₃Ni particles [32] (primary precipitated) denoted by point 2, and Al₇Cr (Al₁₃Cr₂) particles [34] denoted by point 3 are observed in the Mg₂Al₃ matrix. It should be noticed that some tiny reticular Al₃Ni phases [32] (secondary precipitated) denoted by point 5 were formed at the grain boundaries of Mg₂Al₃ matrix, which would reinforce the hard-brittle Mg₂Al₃-based area. Moreover, as shown in Figure 7(b), for Mg₁₇Al₁₂-based area, some tiny long strip Al₃Ni phases [32] are also observed at the grain boundaries of Mg₁₇Al₁₂ matrix.

Figure 7

Microstructures of the weld interface near Mg substrate: (a) area A shown in Figure 6(b), (b) zone F shown in (a), (c) area B shown in Figure 6(b), (d) zone G shown in (c), (e) area C shown in Figure 6(b), (f) area D shown in Figure 6(b), (g) area E shown in Figure 6(b), and (h) area F shown in Figure 6(b).

Table 6

EDS analysis results of the various zones shown in Figure 7 (at.-%).

Area	Position	Al	Mg	Ni	Cr	Mo	Nb	Fe	Possible phases
A	P1	71.1	19.6	7.6	–	–	–	1.7	Al–Fe–Ni phase [35] + Mg₂Al₃ matrix
A	P2	65.5	28.0	6.3	–	–	–	–	Al₃Ni [32] + Mg₂Al₃ matrix
A	P3	74.4	18.3	–	6.6	0.5	0.2	–	Al₇Cr (Al₁₃Cr₂) [34] + Mg₂Al₃ matrix
A	P4	59.8	40.2	–	–	–	–	–	Mg₂Al₃ matrix
A	P5	60.0	38.9	1.1	–	–	–	–	Al₃Ni (grain boundary) + Mg₂Al₃ matrix
B	P6	38.7	61.3	–	–	–	–	–	Mg₁₇Al₁₂
B	P7	11.4	88.6	–	–	–	–	–	Mg solid solution
B	Zone H	33.2	66.8	–	–	–	–	–	Mg solid solution + Mg₁₇Al₁₂ eutectic [24]
B	P8	40.3	59.7	–	–	–	–	–	Mg₁₇Al₁₂ matrix
B	P9	40.7	58.8	0.5	–	–	–	–	Al₃Ni + Mg₁₇Al₁₂ matrix
C	P10	87.7	9.1	–	3.2	–	–	–	Al₇Cr (Al₁₃Cr₂) + Al solid solution matrix
C	P11	81.9	10.9	6.8	–	–	–	0.4	Al₃Ni + Al solid solution matrix
C	P12	92.4	7.6	–	–	–	–	–	Al solid solution matrix
D	Zone I	40.6	9.1	32.2	14.8	2.3	1.0	–	similar to fusion zone (zone A in Figure 3(a))
D	P13	39.9	60.1	–	–	–	–	–	Mg₁₇Al₁₂ matrix
E	Zone J	44.7	55.3	–	–	–	–	–	Mg solid solution (etched) + Mg₁₇Al₁₂ eutectic
E	P14	38.1	61.9	–	–	–	–	–	Mg solid solution (etched)
F	Zone K	60.5	22.6	10.7	5.0	0.8	0.4	–	Mg₂Al₃ matrix + Mg solid solution (etched) + AlNi + Al₇Cr (Al₁₃Cr₂)
F	Zone L	57.4	8.9	22.5	9.1	1.5	0.6	–	similar to fusion zone (zone A in Figure 3(a))
F	P15	60.6	38.3	1.1	–	–	–	–	Mg₂Al₃ matrix + few AlNi
F	P16	62.4	19.1	15.3	2.4	0.8	–	–	AlNi + Mg₂Al₃ matrix + few Al₇Cr (Al₁₃Cr₂)

Fracture characteristic analysis

Figure 8 shows the tensile-shear test results of the welded joint. The average tensile-shear load of the CMT-welded Mg/Al dissimilar lap joint using wire Inconel 625 as filler metal was 1.28 kN (1.321, 1.252, and 1.266 kN). There was no obvious phenomenon of necking and plastic deformation in the fracture area. All specimens fractured at the weld interface near Mg substrate due to the formation of hard-brittle Mg-Al IMCs. However, the tensile-shear load is much higher than that of the CMT-welded Mg/Al dissimilar lap joint using Al-based wire ER4047 (0.38 and 0.93 kN) [22], indicating that using Ni-based wire is better than Al-based wire for Mg/Al dissimilar CMT welding process. Moreover, as using Ni interlayer for filler metal, the CMT-welded lap joint in this study is also much stronger than the laser-welded/brazed lap joint (0.41 ± 0.025 kN) [28].

Figure 8

Tensile-shear test results of the CMT-welded Mg/Al dissimilar joint: (a) fractured specimen, and (b) load-displacement curve.

As shown in Figure 9, the fracture surface of welded joint was characterised. The uneven river pattern with plenty of tearing ridges can be seen on the fracture surface, which confirm the quasi-cleavage fracture. Furthermore, there are also some particles (point 1 and 2) on fracture surface and tiny reticular phases (point 3) at the edges of tearing ridges, which would hinder the fracture propagation. As listed in Table 7, EDS analysis results suggest that the matrix on fracture surface is Mg₂Al₃, the particles on fracture surface are Al₃Ni [32] and Al₇Cr (Al₁₃Cr₂) [34], and the tiny reticular phases at the edges of tearing ridges also is Al₃Ni [32]. These results are consistent with the microstructure characteristics of the Mg₂Al₃-based area within the weld interface near Mg substrate, as shown in Figure 7(a,b). Figure 10 shows the XRD patterns of the constitutive phases on the fracture surfaces. As shown, peaks of Mg solid solution (dominant phase in the Mg substrate and Mg + Mg₁₇Al₁₂ eutectic areas near the fusion line), Mg₂Al₃ (dominant phase in the Mg₂Al₃-based areas), and Mg₁₇Al₁₂ (dominant phase in the Mg₁₇Al₁₂-based areas and Mg + Mg₁₇Al₁₂ eutectic areas near the fusion line) are presented at the Mg substrate side, and peaks of Al solid solution (dominant phase in the Al solid solution-based areas near the fusion zone), Mg₂Al₃ (dominant phase in the Mg₂Al₃-based areas), and AlNi (dominant phase in the fusion zone) are presented at the fusion zone side. Since the X-ray can penetrate the fracture surface (about 30 µm deep), referring to the interface microstructure near Mg substrate, the XRD analysis results suggest that the fracture mainly occurred from the location in the middle part of weld interface where the Mg₂Al₃-based areas are. Based on the above results, it can be concluded that the quasi-cleavage fracture of CMT-welded Mg/Al dissimilar lap joint using wire Inconel 625 occurred primarily at the Mg₂Al₃-based areas within the weld interface near Mg substrate.

Figure 9

Fracture surface of the CMT-welded Mg/Al dissimilar joint.

Figure 10

XRD patterns of the constitutive phases on the fracture surfaces: (a) Mg substrate side, and (b) fusion zone side.

Table 7

EDS analysis results of the various zones shown in Figure 9 (at.-%).

Position	Al	Mg	Ni	Cr	Possible phases
Zone A	65.4	33.6	1.0	–	Mg₂Al₃ matrix + some Al₃Ni [32]
Zone B	64.0	34.3	1.4	0.3	Mg₂Al₃ matrix + some Al₃Ni + few Al₇Cr (Al₁₃Cr₂)
P1	76.4	16.9	6.7	–	Al₇Cr (Al₁₃Cr₂) [34] + some Mg₂Al₃ matrix
P2	67.5	21.1	11.4	–	Al₃Ni + some Mg₂Al₃ matrix
P3	65.1	33.7	1.2	–	Al₃Ni (grain boundary) + Mg₂Al₃ matrix

Microstructure-strength relationship

The strength of CMT-welded Mg/Al dissimilar joint using wire Inconel 625 as filler metal was not as high as expected. This is because the formation of hard-brittle Mg-Al IMCs was not completely prevented. For weld interface near Al substrate, some Al₃Ni and Al₇Cr (Al₁₃Cr₂) phases formed in the Al solid solution matrix, which makes a slight effect on the interface strength. For weld interface near Mg substrate, the hard-brittle Mg-Al IMCs formed, which are detrimental to the interface strength, such as Mg₂Al₃-based areas, Mg₁₇Al₁₂-based areas, and Mg + Mg₁₇Al₁₂ eutectic areas. Moreover, the previous study under the compound casting process [36] shows that the three Mg-Al intermetallics with strength form high to low are Mg + Mg₁₇Al₁₂ eutectic, Mg₁₇Al₁₂, and Mg₂Al₃. Therefore, the CMT-welded Mg/Al dissimilar lap joint using wire Inconel 625 fractured primarily within the Mg₂Al₃-based area where is the weakest area within the weld interface near Mg substrate. However, the result of tensile-shear test shows that the harmful effect of the Mg₂Al₃ intermetallics on the strength of welded joint is limited. First, the distribution of Mg₂Al₃-based areas is uneven, avoiding the formation of continuous Mg₂Al₃ layer along the weld boundary [3]. Second, the precipitated particles such as Al–Fe-Ni phase, Al₃Ni, and Al₇Cr (Al₁₃Cr₂) formed in the Mg₂Al₃ matrix, reducing the brittleness of Mg₂Al₃ matrix. Thirdly, the tiny reticular Al₃Ni phases formed at the grain boundaries of Mg₂Al₃ matrix, reinforcing the strength of Mg₂Al₃-based area. Thus, the lap joint using Ni-based wire as filler metal in this study, with tensile-shear load of 1.28 kN, is much stronger than that using Al-based wire as filler metal [21, 22]. For further study, in order to improve the CMT-welded Mg/Al dissimilar joint using Ni-based wire, preventing the contact of substrates thereby eliminating the Mg-Al IMCs is essential.

Conclusions

Mg-AZ31B/Al-6061 dissimilar lap joint was effectively produced by modified CMT welding process using Ni-based alloy wire Inconel 625 as filler metal. To analyse the characteristics of interface microstructures and fracture behaviour, the following conclusions can be drawn.

There was a clearly boundary between Al substrate and fusion zone. Within the Al substrate next to weld boundary, lots of massive Al₃Ni phases and granular Al₇Cr (Al₁₃Cr₂) phases formed in the Al solid solution matrix. Within the Al substrate near weld boundary, some reticular Al₃Ni intergranular precipitates formed in the grain boundaries of Al solid solution.

Owing to the uneven distributions of the elements between Mg substrate and fusion zone, there were six types of areas with different microstructures formed within the weld interface near Mg substrate: Mg₂Al₃-based area, Mg₁₇Al₁₂-based area, Al solid solution-based area, fusion zone-based area, Mg + Mg₁₇Al₁₂ eutectic area, and mixed area.

With the characteristic of quasi-cleavage fracture, the tensile-shear load of lap joint reached to 1.28 kN. The failure occurred primarily at the Mg₂Al₃-based areas within the weld interface near Mg substrate. However, the harmful effect of the Mg₂Al₃ intermetallics on the strength of welded joint is limited due to the uneven distribution of Mg₂Al₃-based areas, the formation of Al-riched precipitated particles in Mg₂Al₃ matrix, and the formation of tiny reticular Al₃Ni phases formed at grain boundaries of Mg₂Al₃ matrix.

Using Ni-based wire is better than Al-based wire for Mg/Al dissimilar CMT welding process. To improve the joint, preventing the contact of substrates thereby eliminating the Mg-Al intermetallics is essential.

Footnotes

Disclosure statement

No potential conflict of interest was reported by the authors.

References

Cole

, Sherman

Light weight materials for automotive applications. Mater Charact. 1995; 35(1):3–9. doi: 10.1016/1044-5803(95)00063-1

Schubert

, Klassen

, Zerner

, Light-weight structures produced by laser beam joining for future applications in automobile and aerospace industry. J Mater Process Technol. 2001; 115(1):2–8. doi: 10.1016/S0924-0136(01)00756-7

Liu

, Ren

, Liu

A review of dissimilar welding techniques for magnesium alloys to aluminum alloys. Materials (Basel). 2014; 7(5):3735–3757. doi: 10.3390/ma7053735

Wang

, Wang

, Liu

Study on low temperature brazing of magnesium alloy to aluminum alloy using Sn–xZn solders. Mater Des. 2012; 39:14–19. doi: 10.1016/j.matdes.2012.02.021

, Li

, Xu

, Reduction of intermetallic compounds in ultrasonic-assisted semi-solid brazing of Al/Mg alloys. Sci Technol Weld Join. 2019; 24(2):163–170. doi: 10.1080/13621718.2018.1498657

Mahendran

, Balasubramanian

, Senthilvelan

Developing diffusion bonding windows for joining AZ31B magnesium–AA2024 aluminium alloys. Mater Des. 2009; 30(4):1240–1244. doi: 10.1016/j.matdes.2008.06.015

Dietrich

, Nickel

, Krause

, Formation of intermetallic phases in diffusion-welded joints of aluminium and magnesium alloys. J Mater Sci. 2011; 46(2):357–364. doi: 10.1007/s10853-010-4841-5

Chowdhury

, Chen

, Bhole

, Lap shear strength and fatigue behavior of friction stir spot welded dissimilar magnesium-to-aluminum joints with adhesive. Mater Sci Eng A. 2013; 562:53–60. doi: 10.1016/j.msea.2012.11.039

Yan

, Xu

, Li

, Microstructure characteristics and performance of dissimilar welds between magnesium alloy and aluminum formed by friction stirring. Scripta Mater. 2005; 53(5):585–589. doi: 10.1016/j.scriptamat.2005.04.022

10.

Patel

, Bhole

, Chen

Improving weld strength of magnesium to aluminium dissimilar joints via tin interlayer during ultrasonic spot welding. Sci Technol Weld Join. 2012; 17(5):342–347. doi: 10.1179/1362171812Y.0000000013

11.

Feng

, Li

, Zhao

, Mechanical properties and interface morphology of Mg/Al ultrasonic spot weld bonding welds. Sci Technol Weld Join. 2016; 21(8):688–699. doi: 10.1080/13621718.2016.1144262

12.

Borrisutthekul

, Miyashita

, Mutoh

Dissimilar material laser welding between magnesium alloy AZ31B and aluminum alloy A5052-O. Sci Technol Adv Mater. 2005; 6(2):199–204. doi: 10.1016/j.stam.2004.11.014

13.

Wang

, Liu

The characterization investigation of laser-arc-adhesive hybrid welding of Mg to Al joint using Ni interlayer. Mater Des. 2013; 50:463–466. doi: 10.1016/j.matdes.2013.02.085

14.

Schierl

The CMT process a revolution in welding technology. Weld World. 2005; 49(I):38.

15.

Mezrag

, Deschaux-Beaume

, Benachour

Control of mass and heat transfer for steel/aluminium joining using cold metal transfer process. Sci Technol Weld Join. 2015; 20(3):189–198. doi: 10.1179/1362171814Y.0000000271

16.

Madhavan

, Kamaraj

, Vijayaraghavan

Micro-structure and mechanical properties of cold metal transfer welded aluminium/dual phase steel. Sci Technol Weld Join. 2016; 21(3):194–200. doi: 10.1179/1362171815Y.0000000082

17.

Liu

, Sun

, Sang

, Microstructure and mechanical properties of cold metal transfer welded aluminium/nickel lap joints. Sci Technol Weld Join. 2015; 20(4):307–312. doi: 10.1179/1362171815Y.0000000011

18.

Cao

, Wang

, Cold metal transfer welding–brazing of pure titanium TA2 to magnesium alloy AZ31B. J Alloys Compd. 2014; 605:12–20. doi: 10.1016/j.jallcom.2014.03.051

19.

Cao

, Jing

, Feng

, Cold metal transfer welding–brazing of magnesium to pure copper. Sci Technol Weld Join. 2014; 19(6):451–460. doi: 10.1179/1362171814Y.0000000204

20.

Jing

, Kehong

, Qi

, Microstructure characteristics and properties of Mg/Al dissimilar metals Made by cold metal transfer welding with ER4043 filler metal. Rare Met Mater Eng. 2013; 42(7):1337–1341. doi: 10.1016/S1875-5372(13)60081-8

21.

Wang

, Feng

, Wang

Microstructure of Al–Mg dissimilar weld made by cold metal transfer MIG welding. Mater Sci Technol. 2008; 24(7):827–831. doi: 10.1179/174328408X278411

22.

Cao

, Wen

, Chen

, Cold metal transfer joining of magnesium AZ31B-to-aluminum A6061-T6. Mater Sci Eng A. 2013; 560:256–266. doi: 10.1016/j.msea.2012.09.065

23.

Madhavan

, Kamaraj

, Vijayaraghavan

Cold metal transfer welding of aluminium to magnesium: microstructure and mechanical properties. Sci Technol Weld Join. 2016; 21(4):310–316. doi: 10.1080/13621718.2015.1108070

24.

Wang

, Hu

, Shen

, Microstructure and mechanical behaviour of cold metal transfer welded Mg/Al dissimilar joint using wire AZ31 as filler metal. Sci Technol Weld Join. 2017; 22(4):353–361. doi: 10.1080/13621718.2016.1245240

25.

Zhang

, Luo

, Wang

, Effect of Al thin film and Ni foil interlayer on diffusion bonded Mg–Al dissimilar joints. J Alloys Compd. 2013; 556:139–142. doi: 10.1016/j.jallcom.2012.12.106

26.

Penner

, Liu

, Gerlich

, Feasibility study of resistance spot welding of dissimilar Al/Mg combinations with Ni based interlayers. Sci Technol Weld Join. 2013; 18(7):541–550. doi: 10.1179/1362171813Y.0000000129

27.

Chang

W-S

, Rajesh

, Chun

C-K

, Microstructure and mechanical properties of hybrid laser-friction stir welding between AA6061-T6 Al alloy and AZ31 Mg alloy. J Mater Sci Technol. 2011; 27(3):199–204. doi: 10.1016/S1005-0302(11)60049-2

28.

Yang

, Yu

, Li

, Laser welding/brazing of 5182 aluminium alloy to ZEK100 magnesium alloy using a nickel interlayer. Sci Technol Weld Join. 2018: 1–8. doi: 10.1080/13621718.2017.1314657

29.

Khodabakhshi

, Shah

, Gerlich

Dissimilar laser welding of an AA6022-AZ31 lap-joint by using Ni-interlayer: Novel beam-wobbling technique, processing parameters, and metallurgical characterization. Opt Laser Technol. 2019; 112:349–362. doi: 10.1016/j.optlastec.2018.11.034

30.

Wang

, Hu

, Shen

, Characterization the contribution and limitation of the characteristic processing parameters in cold metal transfer deposition of an Al alloy. J Mater Process Technol. 2017; 245:122–133. doi: 10.1016/j.jmatprotec.2017.02.019

31.

Akinlade

, Caley

, Richards

, Microstructural investigation of phase transformations and properties during thermal processing of a 6 w/oAl-modified Ni–Cr–Fe ternary alloy. Mater Sci Eng A. 2008; 488(1-2):221–234. doi: 10.1016/j.msea.2007.11.019

32.

Lippold

, Kiser

, DuPont

JN.

Welding metallurgy and weldability of nickel-base alloys. John Wiley & Sons; 2011.

33.

Badiger

, Narendranath

, Srinath

Joining of Inconel-625 alloy through microwave hybrid heating and its characterization. J Manuf Process. 2015; 18:117–123. doi: 10.1016/j.jmapro.2015.02.002

34.

Murray

The Al-Cr (aluminum-chromium) system. J Phase Equilib. 1998; 19(4):367. doi: 10.1007/BF02735058

35.

Chumak

, Richter

, Ipser

The Fe–Ni–Al phase diagram in the Al-rich (> 50 at. % Al) corner. Intermetallics. 2007; 15(11):1416–1424. doi: 10.1016/j.intermet.2007.04.012

36.

Hajjari

, Divandari

, Razavi

, Dissimilar joining of Al/Mg light metals by compound casting process. J Mater Sci. 2011; 46(20):6491–6499. doi: 10.1007/s10853-011-5595-4