Consideration of cold metal transfer welding of AA5754 to galvanized steel from mechanical and microstructural aspects

Nilay Çömez^a,*, Hülya Durmuş^a

^a Manisa Celal Bayar University, Engineering Faculty, Department of Metallurgical and Materials Engineering, Manisa, Turkey

*Corresponding author: nilay.comez@cbu.edu.tr

1. INTRODUCTIONTOP

The necessity for joints between dissimilar materials often arises in functionally complex industrial applications, economic and environmental concerns (Torkamany et al., 2010; Guo et al., 2014; Liu et al., 2015). For instance, the combination of steel with aluminum provides fuel-efficiency by reducing the weight of construction (Miller et al., 2000; Torkamany et al., 2010; Ahsan et al., 2016). One of the major challenges in transportation vehicle manufacturing is the joining of aluminum and steel by conventional fusion welding methods due to the large difference between their physical and chemical properties (Cao et al., 2014a ; Guo et al., 2014; Squires et al., 2015).

Al-Fe phase diagram reveals that the solid solubility of iron in aluminum is quite low (Torkamany et al., 2010; Su et al., 2014). Hence, the formation of brittle intermetallic phases such as Fe₃Al, FeAl, FeAl₂, Fe₂Al₃, Fe₂Al₅ and FeAl₃ are promoted especially at high temperature (Potesser et al., 2006; Torkamany et al., 2010; Cao et al., 2013; Zhou and Lin, 2014; Liu et al., 2015). Fe-Al intermetallic phases are required for an effective joint between aluminum and steel, nevertheless thickness of this IMC layer should be kept below 10 μm to supply technical demands (Jácome et al., 2009; Cao et al., 2014a). In order to control the thickness of Fe-Al intermetallic layer, weld heat input should be kept at a low rate.

Recently developed cold metal transfer (CMT) welding which is characterized with low heat input and spatter free weld, is recommended for joining dissimilar materials (Yang et al., 2013; Ahsan et al., 2016). In CMT welding, filler wire is withdrawn after short-circuiting and the metal transfer takes place by the detachment of the molten droplet with the aid of the wire movement (Yang et al., 2013; Zhang et al., 2013; Ahsan et al., 2016).

CMT welding of AA6061 to galvanized steel was studied by numerous researchers (Agudo et al., 2007; Cao et al., 2013; Yang et al., 2013; Zhang et al., 2013; Cao et al., 2014b). However, there is limited number of studies about CMT welding of AA5754 to steel in literature. AA5754 sheets are employed in inner panels of automobiles (Davies, 2003; Halim et al., 2007). As is known, numerous steel grades are the major materials of vehicle structures. These steels are mostly coated with zinc to provide galvanic protection (Ahsan et al., 2016). Zhou and Lin (2014) reported that zinc coating provides wettability of aluminum on the steel surface and also some heat can be removed from the system with the help of zinc vaporization.

AA5754 to steel joints were studied by several researchers. Figner et al. (2009) examined the friction stir spot welding of aluminum AA5754 and galvanized steel HX 340LAD. Yang et al. (2015) studied lap joining of AA5754 to DP 980 steel by diode laser welding and investigated the correlation between interfacial microstructure and mechanical properties. They obtained a hard and a brittle Fe₂(Al,Si)₅ (η phase) with micro cracks between steel and aluminum at high laser power. Watanabe et al. (2006) joined 2 mm thick SS400 carbon steel and AA5083 with friction stir welding. According to their study IMCs was found to be detrimental for joint strength. They also observed large pores at the Al/Fe interface.

Present study is aimed to research the weldability of AA5754 to galvanized steel which may provide benefit for automotive applications. For this purpose, effect of heat input on correlation between intermetallic formation and mechanical properties of joints was investigated. The originality of this study is the interpretation of tensile strength according to hardness and elastic modulus of intermetallic phases which were determined by nano indentation test.

2. MATERIALS AND METHODSTOP

2 mm thick AA5754 (Table 1) and galvanized steel sheets (EN 10143 – DX 51D+Z) (Table 2) were joined with CMT overlap welding, using ER4043 filler material (Table 3) having a diameter of 1.2 mm. In order to remove the oxide layer, welding groove of aluminum sheets were grinded and wiped with acetone. Weld pool was isolated by argon gas (15 l·min⁻¹) during the welding process. Welding parameters are given in Table 4. Torch angle and distance between the torch and the work piece were adjusted to 90° and 12 mm respectively.

2.1. Microstructural investigations: OM and SEM-EDXTOP

Cross-sections of welded specimens were grinded and polished. Keller and Nital reagents were used for the etching of aluminum weld metal and galvanized steel, respectively. AA5754 base metal was etched with a solution consists of 20 g NaOH and 100 ml distilled water. Nikon LV150 optical microscope and Clemex software were used for microstructural investigation of the weld zone and measuring the thickness of the intermetallic layers. Chemical composition of intermetallic phases was determined by SEM-EDX. XRD analysis was applied to Al/steel interface of 5G-2 which separated during the tensile test.

2.2. Tensile testTOP

Mechanical strength of the welded specimens was determined by Shimadzu Autograph tensile testing device with the rate of 1 mm·min⁻¹ at room temperature using extensometer. Tensile tests were conducted according to DIN EN 895 standard. Three specimens of each welding parameters were tested and the average results of tensile strength were calculated.

2.3. Micro hardness test and nano indentation testTOP

Micro hardness test was carried out in order to detect the heat affected zone. Vickers micro hardness values were measured in the order of aluminum-weld seam-steel using Future-Tech FM700, under 100 gf (gram-force) load applied for 10 seconds.

Berkovich indenter was utilized for the nano indentation test, which was carried out under 5 mN load in an attempt to determine the elastic modulus and the hardness of Al/Fe IMCs.

3. RESULTS AND DISCUSSIONTOP

3.1. Microstructural investigationsTOP

The cross-sectional microstructures of CMT welded AA5754 and galvanized steel are presented in Fig. 1. The grain size of steel base metal did not change by the heat input, while the aluminum weld metal grains became coarser with the increasing weld heat. Steel substrate and aluminum base metal generated a temperature gradient which caused the formation of columnar grain growth in the weld metal adjacent to steel and AA5754 (Fig. 1 and Fig. 2). Increasing weld heat input promoted the growth of Al/Fe IMC layer. Mean intermetallic thickness values of 5G-1 and 5G-2 were calculated as 2.103 μm and 4.655 μm respectively.

Torkamany et al. (2010) joined 0.8 mm thick low carbon steel (st14) and 2 mm thick AA5754 with Nd:YAG laser welding. They reported the formation of cracks, pores and cavitation in intermetallic layer between Al/Fe interfaces. They also mentioned that with the increasing laser peak power and the heat input, the percentage of IMC raised in weld metal. In present study, cracks or pores were observed in neither microstructure nor SEM investigations of IMC layers.

As is seen from Fig. 3, the intermetallic compound grew towards aluminum weld metal. This phenomenon can be explained with diffusion mechanisms and characteristics of Al and Fe elements. Shao et al. (2015) reported that iron atoms could diffuse more easily in liquid aluminum weld metal by virtue of the higher diffusion coefficient of iron in liquid aluminum in comparison with the diffusion coefficient of aluminum in solid iron.

In Fig. 4 SEM-EDX analysis of the weld toe of sample 5G-1 is given. It was observed that the IMC layer disappeared and Al weld metal exhibited cellular structure. Elemental mapping revealed that this region contains a large amount of Zn in Al weld metal (Fig. 4). When the molten aluminum droplet detaches from the wire, it replaces with zinc (melting point~420 °C) in liquid state. Although the density of liquid zinc (6.57 g·cm⁻³) is almost 3 times greater than that of aluminum, in present study it was observed that with the aid of both weld arc forces and mass of molten filler wire, liquid zinc was swept towards the toe of weld metal. It was thought that, this phenomenon was also supported by the tendency of aluminum to form Al/Fe IMCs with iron. Zhou and Lin (2014) stated that the replacement of aluminum and zinc occurs due to the higher tendency and reactivity of aluminum to iron than that of zinc to iron. They have proved their thesis with EDS analyses of interface and end point of weldment. They did not encounter any Zn trace at the interface of aluminum and steel. Besides, they observed zinc liquid film accumulated at the weld toe.

Another remarkable finding is the accumulation of zinc at the end of aluminum weld metal instead of steel base metal (Fig. 4). Zn has a higher affinity to Al than to Fe; hereby Al-Zn phases are formed at the weld toe (Agudo et al., 2007).

3.2. Chemical analysis of intermetallic phasesTOP

EDX analyses of intermetallic layers were carried out as shown in Fig. 5. EDX results are presented in Table 5. Silicon content was detected to be higher in the intermetallic layer of 5G-1 than that of 5G-2. Silicon avoids the growth of Al/Fe phases by occupying the vacancies of the Al/Fe intermetallic compounds and forms Fe_xAl_ySi_z phases which possess a slower growing rate than Fe_xAl_y phases (Eggeler et al., 1986; Jácome et al., 2009). As a result, intermetallic compound layer of 5G-2 exhibited finger-like growth towards aluminum while the intermetallic layer of 5G-1 grew smoothly due to the suppression of Si atoms (Fig. 5).

Chemical compositions of some Al/Fe and Al/Fe/Si IMCs are listed in Table 6 based on the literature findings. It was noticed that the chemical composition of 5G-1 was found to be similar to Al_7.2Fe_1.8Si phase as mentioned in (Song et al., 2009; Yang et al., 2015).

According to Al-Fe binary phase diagram, Fe₂Al₅ compound contains 53-57 wt.% aluminum (Kattner and Burton, 1992). Formation of brittle Fe₂Al₅ intermetallic phase was approved by EDX analysis of 5G-2 which can also be supported by the findings of (Maitra and Gupta, 2002; Agudo et al., 2007; Cao et al., 2014b) (Table 6).

Although Fe_xAl_y phases are required for an effectual joint between aluminum and steel, excessive formation of these phases leads to the tendency of crack initiation in intermetallic layer under tensile stress (Agudo et al., 2007; Su et al., 2014). As is known, the growth of brittle Al/Fe IMCs can be replaced with less detrimental Al/Fe/Si phases by Si addition (Song et al., 2009). Apart from the control of intermetallic layer thickness, mechanical properties of dissimilar Al-steel joints can also be adjusted with the aid of Si based filler wires. Although Al_7.2Fe_1.8Si phase hinders the growth of Fe_xAl_y phases (Song et al., 2009), increasing the heat input led to the formation of brittle Fe₂Al₅ compound in present study.

The result of XRD analysis taken from Al/steel interface of 5G-2 shows the peaks of Al, Fe and Zn elements as expected; in addition it indicates the probable existence of AlZn and Fe₂Al₅ intermetallics (Fig. 6).

3.3. Tensile testTOP

Hybrid structures such as dissimilar welding of aluminum and steel keep various phases with different physical and chemical properties together. For instance, Al/Fe IMCs exhibit superior mechanical properties in comparison to metallic materials. As a result, their respond to mechanical forces will not be similar to aluminum and steel. Therefore mechanical tests should be carried out to understand the fracture mechanism and ultimate tensile strength of hybrid joints in order to build safe constructions.

Stress-strain curves of CMT welded AA5754-galvanized steel joints are given in Fig. 7. CMT welded AA5754-galvanized steel samples exhibited different deformation and fracture characteristics. AA5754 aluminum alloy shows discontinuous deformation at room temperature which is called Portevin–Le Chatelier (PLC) effect. This effect is categorized according to serration types in the stress–strain curve (Halim et al., 2007). In this study, sample 5G-2 presents Type-B serrations which show hopping propagation (Halim et al., 2007) as it can be seen in Fig. 7. Although the PLC effect was observed clearly in plastic deformation zone of the sample 5G-2 (AA5754 base metal, Fig. 7), the fracture occurred at the Al/steel interface instead of the AA5754 base metal. It was thought that cracks initiated at the brittle Fe₂Al₅ intermetallic layer beyond the threshold load value and propagated rapidly causing an instantaneous rupture of the sample 5G-2. Similarly, the sample 5G-1 also fractured suddenly at the weld seam without plastic deformation (Fig. 7).

Figure 8 presents the fracture surface of 5G-1 which includes numerous micro-scale pores in weld seam. Oxygen and zinc rich zone was observed from the elemental mapping analysis right above the Al/steel interface where the aluminum content decreases.

Su et al. (2014) joined AA5052 and galvanized mild steel sheets by alternate-current double pulse gas metal arc welding in lapping configuration with Al–5Si filler wire and obtained a 134 MPa ultimate tensile strength. Song et al. (2009) focused on the effect of Si on the formation of Al/Fe intermetallic compounds which takes place in TIG welding of 5A06 aluminum alloy and AISI 321 stainless steel using 1100 pure Al, AlSi₅ and AlSi₁₂ filler metals. They found that intermetallic layer having 5 wt.% of Si additions showed the optimum mechanical properties with the tensile strength of 125.2 MPa. The tensile strength values of 5G-1 and 5G-2 welded with AlSi₅ filler wire were found as 105.29 MPa and 146.3 MPa, respectively. The higher joint strength of the sample 5G-2 is attributed to the formation of Fe₂Al₅ phase which has superior mechanical properties. However, Fe₂Al₅ induced the brittle fracture of intermetallic layer.

3.4. Micro hardness testTOP

Hardness distribution of Al/steel joints was almost homogenous as in Fig. 9. Heat affected zone (HAZ) was not observed both at the AA5754 and steel base metals due to the low heat input characteristics of CMT welding. Microstructural investigations also revealed that grain coarsening which is responsible for the hardness drop did not occur in base metals. Magda et al. (2013) reported that CMT did not lead to any changes in the HAZ of the base material. In the present study, a significant change in hardness was not observed, while increasing the heat input caused the grain coarsening of weld metal.

3.5. Nano indentation testTOP

Nano indentation test was applied to intermetallic layers from the cross section of weld metal. Load-displacement (P-h) curves of Al_7.2Fe_1.8Si and Fe₂Al₅ phases are given in Fig. 10. The effect of silicon content on mechanical properties can be seen clearly from Table 7. Al_7.2Fe_1.8Si which contains more Si has lower hardness and elastic modulus than that of Fe₂Al₅ with less Si content.

Although the increasing hardness and elastic modulus enhanced the tensile strength of dissimilar AA5754/steel joint, a fracture located at the brittle Fe₂Al₅ intermetallic layer due to the deterioration of plastic strain properties.

4. CONCLUSIONSTOP

AA5754 aluminum alloy and galvanized steel was joined successfully by CMT welding using AlSi₅ filler wire. The effect of heat input on the formation of intermetallic layer and mechanical properties was investigated. The findings can be summarized as follows:

ACKNOWLEDGMENTSTOP

The authors would like to thank to Scientific and Technological Research Council of Turkey (TÜBİTAK) for supporting this study financially with the project number of TÜBİTAK-215M623.

REFERENCESTOP