Joining steel to aluminium and magnesium alloys.... a new way forward

TWI Bulletin, May - June 2007

Increasing demand for better fuel efficiency and reduced emissions has prompted the automotive industry to reduce vehicle weight using lightweight materials, particularly aluminium and magnesium alloys.

Steve Shi is a principal project leader in the Laser and Sheet Processes Group at TWI, which he joined in 2001 with a PhD in materials and experience in a range of metallic materials. He has managed projects concerned with laser materials processing and joining of high strength steels, with particular involvement in laser and laser-arc hybrid welding, in-process monitoring and adaptive control of laser welding, resistance and laser welding of high strength steels.

Steve Westgate is a consultant in resistance welding in the Laser and Sheet Processes Group at TWI. Steve joined TWI in 1972 with a degree in industrial metallurgy and since then has had involvement in, and managed, a wide range of research projects in all resistance welding processes. Close contact is maintained with industry through an active consultancy, trouble shooting and training role. His experience also includes mechanical fastening systems, adhesives and hybrid techniques, largely related to sheet metal joining.

Stephen Mulligan joined TWI in 1999 with a Masters degree in Materials Engineering from the University of Birmingham which included a placement with BAE Filton. He is an International/European Welding Engineer. Areas of expertise include pipeline welding, robotic MIG and TIG welding, AC pulsed MIG welding, tandem wire MIG welding and powder plasma arc welding of a wide range of materials.

Although thin-sheet steels are the dominant materials for most automotive applications, aluminium profiles and magnesium die-castings are being used increasingly in automotive structures. As Steve Westgate, Steve Shi and Stephen Mulligan report the co-existence of these materials in a structure, has led to a significant interest in the joining of steel to aluminium and magnesium alloys. However, because of the difficulties in fusion welding and indeed brazing these dissimilar materials, mechanical fastening, using rivets or bolts is still currently the preferred technique.

Steel, aluminium and magnesium have widely different thermal and mechanical properties ( Table 1). The main problem when fusion welding steel to aluminium and steel to magnesium is the limited solid solubility of Fe in Al and Mg, leading to the formation of brittle intermetallic compounds. In addition, it is necessary to remove the surface oxide layer (melting point 2040°C) from the aluminium alloy component before welding. Fluxing compounds available for aluminium tend to be based on chloride compounds and can be corrosive if not rinsed off adequately. Non-chloride based fluxes are also available. Fluxes also require high temperatures in order to work (typically ~575°C), so a localised heat source is desirable in order to minimise the extent of the heat-affected zone.

Table 1 Physical properties of C-Mn steel, and selected aluminium and magnesium alloys

A common industrialised solution for joining steel to aluminium is to use an explosively bonded or roll-bonded bimetallic insert. This technique is employed widely in the shipbuilding industry, and means that similar material joints can be made on either side of the bimetallic insert. However, if the interlayer is heated above about 300°C, the intermetallic compounds may grow and embrittle the joint. While this normally restricts their use in thin sections, the unaffected area of the transition joint can still provide sufficient strength in a spot welded joint. Such aluminium/steel transition joints have been applied for special cases in the automotive sector. The primary requirement of a fusion welding process for joining steel to aluminium is, therefore, to keep the heat input low in order to restrict the time above temperatures of about 300°C and thus minimise the formation of intermetallic compounds.

In the last few years, work has been conducted at TWI to develop techniques for joining dissimilar materials. This has included friction welding, to provide a joint with a partial mechanical interlock between aluminium and machined steel plates, and development of laser processes. Recent work at a machine manufacturer has also demonstrated spot welding of aluminium to steel by spot welding and cold metal transfer MIG welding of aluminium to steel with a thick zinc coating.

This work describes the result of preliminary studies at TWI on the joining of aluminium and magnesium to steel, using resistance spot welding and arc welding techniques. The experiences and results of using three approaches to this problem are reported, these being resistance spot welding using a triple layer (sandwich) lap joint configuration, resistance spot welding using a bimetallic insert, and TIG welding with trailing wire feed.

Materials

Materials used for the spot welding were as follows:

• AZ31B H24 magnesium alloy (Mg-Al-Zn) of 1.0mm thickness
• Galvanised low carbon steel of 0.8mm thickness, with a zinc coating of approximately 10mm on each side
• 5754 aluminium alloy of 1.5mm thickness
• Commercial roll bonded aluminium/low carbon steel transition material of 1.5mm thickness with equal thicknesses of each metal

The surfaces of the aluminium and magnesium were cleaned using abrasive pads to remove as-received surface oxide, and degreased with acetone immediately prior to welding. The zinc-coated steel and steel/Al transition piece were degreased solely with acetone.

TIG welding trials were performed using 3.2mm thick C-Mn steel sheet, and 2mm thick AA5083-O aluminium alloy. All sheets were 100mm wide and 350mm long.

Resistance spot welding trials for joining steel to aluminium and magnesium

In the first approach, aluminium and magnesium were resistance welded to steel in a triple layer lap joint configuration, with the steel both sides, as shown in Figure 1a. During welding, the current and electrode force in the spot welding operation were controlled so that the heat (generated mostly within the steel) was sufficient to melt the aluminium or magnesium sheet. It was expected that the molten metal, contained by the electrode force, would wet the steel surface, forming a bond at the interface between the two materials, as illustrated schematically in Figure 1b.

Fig.1. Joining aluminium and magnesium to steel using resistance heating

a) Joint configuration

b) Melting the Al or Mg to form a bond between the dissimilar materials

In a second set of experiments, aluminium was joined to steel with an aluminium/steel transition piece, again using resistance spot welding. In this case, a melted weld nugget would be formed on the steel side as in a conventional spot weld. The heat generated would also melt the aluminium from the interface of the transition piece into the outer aluminium sheet, as shown schematically in Figure 2.

Fig.2. Resistance spot welding aluminium to steel using a transition material:

a) Joint configuration

b) Formation of two weld nuggets at the interfaces between the similar materials

All of the resistance spot welding trials were carried out using a Martin scissors gun with a medium frequency DC power supply. Trials were carried out using Cu-Cr-Zr electrodes with 16mm diameter ISO 5182 caps and two tip shapes; the first a Type G electrode with a 6mm tip having a 40mm face radius, the second a Type A dome electrode, 16mm diameter with 40mm face radius. For each of the material combinations, trials were conducted by pre-setting the electrode force and weld time and then making welds at progressively increased current levels, to determine the maximum welding current that could be applied without splash.

Squeeze time was set at 50 cycles to ensure full achievement of the welding force. Hold time was set at ten cycles throughout. The applied forces were 1.3kN, 2kN and 4kN, depending on the material combination to be welded. Welds were peel tested to determine the fracture mode. Test welds were then made using conditions showing plug failure and no cracking in peel tests. Shear tests were conducted on these welds to measure the shear failure load.

TIG welding with trailing wire feed for joining steel to aluminium

In this approach, 1.2mm diameter aluminium alloy filler wires (AA4043A (Al-5Si) and AA4047 (Al-12Si)) were used to deposit 'buttering' weld beads onto C-Mn steel. Mechanised TIG welding was used with trailing filler wire feed. The filler wire was placed approximately 8mm behind the tungsten electrode impinging onto solidified steel material, using the extremity of the arc and the heat from the base material to melt the filler wire. These weld deposits were assessed visually for consistency. Adherence of the deposit was verified by a 90° bend test of the steel sheet with the 'buttered' surface in tension, and by chiselling with a screwdriver. Following the deposition of a satisfactory buttering layer, lap joints between the buttered C-Mn steel sheet and aluminium sheets were welded using 4043A filler wire and AC polarity.

Results and Discussion

Resistance spot welding trials for joining steel to aluminium and magnesium

The results of trials with different electrode forces and welding currents indicated that it was possible to produce bonding in the interface between the steel and magnesium alloy, using resistance heating. However, bonding could only be achieved in magnesium when the welding current was adjusted to melt only the central magnesium sheet, at a pre-set electrode force.

Initially, trials were carried out with low electrode force (1.3kN) and long weld time (30 cycles), in an attempt to achieve good wetting between the steel and the magnesium. A welded sample produced using such conditions is shown in Figure 3. In the peel test, a small diameter plug was pulled out from the magnesium sheet, indicating that a bond had been produced between steel and magnesium, as shown in Figure 3b.

Fig.3. Visual appearance and fracture mode of a spot welded joint between magnesium and steel:

a) A spot welded sample

b) Peel tested sample (magnesium piece in centre)

However, solidification cracking occurred in the melted magnesium ( Fig.4). Severe porosity and cracking were found in the centre of the nugget ( Fig.4a). No intermetallic compounds were observed at the interface between the steel and the magnesium under optical microscope examination ( Fig.4b). Although the welds exhibited weak, brittle interface failures in peel testing, three shear test samples at this condition gave an average shear failure load of 3.3kN.

Fig.4. Cross section of a resistance welded magnesium to steel sample. 1.3kN electrode force, 30 cycles weld time and 10kA welding current. Electrodes used were 16mm diameter, 40 mm radius dome Type A caps:

a) Macrosection

b) Detail of sound weld interface

c) Solidification crack at weld interface

A higher electrode force was found to be beneficial in preventing the shrinkage flaws in the magnesium. Figure 5b shows the cross section of a weld produced using the Type G, with a higher electrode force of 4kN. No cracks or porosity were found in this section, which had a minimum size melted zone in the magnesium of about 4mm diameter. Shear tests were not conducted at this condition but the welds still exhibited very low peel resistance.

Fig.5. Cross sections of resistance welds between magnesium and steel to examine the effect of electrode force, with welding current adjusted to suit. Electrodes used were 6mm tip diameter Type G caps:

a) 2kN electrode force, 8kA current and 15 cycles weld time

b) 4kN electrode force, 10kA current and 15 cycles weld time

The approach used for joining magnesium to steel was also used to join aluminium to steel. These welds were produced using 40mm radius dome electrodes, in an effort to produce a large weld area. The cross section of a joined sample is shown in Figure 6a. For the joining of aluminium to steel, the aluminium has been melted and a bond created at the interface, through wetting of the steel surface by molten aluminium. A continuous layer was formed in the interface between the steel and aluminium, but intermetallic compounds were formed, as can be seen in Figure 6b. The fused zone diameter was about 5mm. In this case, the preferential etching of the steel shows that the steel also melted within the sheet thickness and transferred heat to the aluminium sheet. The weld failed in a weak, brittle manner at the interface on peel testing.

Fig.6. Cross section of a resistance spot welded steel to aluminium joint. 1.3kN electrode force, 20 cycles weld time and 9.7kA welding current. Electrodes used were 40mm radius dome electrodes with ISO 16mm cap:

a) Cross section of the weld

b) Intermetallic compounds in the interface

Spot welding aluminium to steel using a transition material

Figure 7 shows the welded sample, peel test fracture mode and cross section of welds between steel and aluminium produced using the aluminium/steel transition piece. A nugget has formed in the steel and transferred heat to the aluminium side, which is melted from the transition material interface through into the aluminium sheet. The heat sink effect of the aluminium within the joint prevents the steel melting completely through to the contact with the aluminium. The section is shown in Fig.7c.

Fig.7. Fracture mode and cross section of a typical weld between steel and aluminium produced using a roll bonded steel/Al transition piece. 2kN electrode force, 10 cycles weld time and 10.65 kA welding current. Electrodes used were dome with ISO 16mm cap and 40mm radius:

a) Welded sample

b) Peel tested sample (steel on left, transition piece centre, aluminium on right)

c) Cross section of the weld

The shrinkage defects shown in the section are located against the steel interface as that part of the nugget was likely to be last to solidify. In addition, the interface within the transition piece exhibited the formation of intermetallic compounds. Consequently, the welds showed some brittleness and failed on peel testing by pulling a plug out of the thin aluminium layer of the transition material. Similar shear test samples gave a failure load up to 1.5kN with plug failure in the 1.5mm aluminium alloy strip. A minimum weld strength of about 2.5kN would normally be expected for this thickness of aluminium, although these welds were slightly undersized, having a plug diameter of about 4.5mm.

TIG welding with trailing wire feed for joining steel to aluminium

It was possible to deposit a weld bead of aluminium alloy onto steel using both 4043 and 4047 filler wires. Suitable parameters were 120A, 10V at 750mm/min travel speed. The wire was fed at 1m/min at 30° from the surface and aimed 7-8mm behind the tungsten electrode. Figure 8 shows a typical bead deposited using the trailing wire feed technique, and the consistency of the weld was found to be highly sensitive to the position of the wire. The deposit survived a 90° bend test without spalling off ( Fig.9), and therefore exhibited potential for use as a buttering layer.

Fig.8. Al-5Si weld on 3mm thick C-Mn steel sheet deposited using TIG welding with trailing wire feed

a) Appearance of overall weld

b) Close-up of weld bead

Fig.9. Al-5Si weld bead after 90° bend test

a) Testpiece after bending

b) Close-up of weld surface. Note the transverse cracking on the surface but with no spalling

Unfortunately, cracking occurred when the layer was subjected to a further heat cycle ( Fig.10). When welding directly onto the aluminium layer to create a dissimilar lap joint, cracking was audible during cooling and separation of the steel and aluminium occurred a few seconds after completing the weld. The parameters used for welding were 135A, 10.2V at 250mm/min travel speed, with the torch angled 5° toward the lap joint and aimed 1mm inside the edge of the upper aluminium plate. Wire feed speed was 1m/min. It was suspected that the driving force for cracking was differential thermal expansion.

Fig.10. Failure of aluminium weld deposit to aluminium sheet joint at interface with steel

Summary

This preliminary work has demonstrated the limited feasibility of joining steel to aluminium and magnesium alloys, as well as showing some of the difficulties encountered in making these dissimilar joints. Using resistance heating, it was possible to join steel to magnesium and steel to aluminium directly in a triple layer lap configuration. A bond could be formed at the interface between steel and magnesium or aluminium, by melting the magnesium or aluminium sheet in contact with the steel. However, the welds exhibited predictably weak brittle behaviour in peel testing despite showing moderate shear strength. No intermetallic compound was found in the interface between steel and magnesium under examination by optical microscopy.

A preferred method for welding steel to aluminium would be the use of a transition material as a means of minimising the effect of brittle intermetallic compounds. Reasonable joint strength can be produced but the disadvantage is the need to use a separate insert material.