Friction Stir Resistance Spot Welding of Aluminum Alloy to Advanced High Strength Steel

In recent years, a growing amount of attention has been paid to vehicle weight reduction in the automotive industry. Multimaterial vehicle structure is a promising solution, which increases the demand for developing reliable and economical dissimilar material joining techniques. Traditional fusion welding methods tend to generate an excessive amount of brittle intermetallic compounds (IMCs), which significantly deteriorate joint quality. In this regard, solid-state welding techniques are more favorable based on their relatively low heat input. One is the friction stir spot welding (FSSW) process, which can be applied for spot joints between dissimilar materials.

However, since materials are undergoing severe plastic deformation in the solid state, the FSSW process contains several inherent disadvantages, including a large axial plunge force, a relative long welding time, and insufficient material flow. To overcome these difficulties and enhance the weld performance, some hybrid improvements have been employed. Sun et al. [1] applied the laser as a preheating source to improve the friction stir welding (FSW) of S45C steel plates. In their study, different positions of the laser focal points were tested, and the welding speed could reach up to 800 mm/min when the laser beam was focused at the joint line. In addition, they indicated that the friction heat generated between the tool and the workpiece is reduced when preheating was applied on the advancing side. Chang et al. [2] studied the laser-assisted friction stir welding between aluminum alloy 6061-T6 to Mg alloy AZ31. To get a strong welding between the dissimilar materials, they inserted a Ni foil between the faying surfaces, and the resulted weld strength was 66% of the Mg base metal tensile strength. Merklein and Giera [3] applied laser-assisted friction stir welding to weld deep drawing steel DC04 and the aluminum alloy AA6016-T4. With the application of the laser, the welding feed speed increased up to 2000 mm/min and the welded tensile strength reached 80% of the aluminum base material. They also mentioned that there were no intermetallic phases generated even when applying the laser during the welding process. Besides the laser assisted FSW, ultrasonic assisted FSW was developed and studied to enlarge the welding process window and enhance the material flow. Park et al. [4] did ultrasonic assisted friction stir welding of aluminum alloys. The welding force was reduced and the mechanical properties of the welded part were improved, including both the elongation and yield strength. Besides, the chance of forming a welding defect decreased. Ahmadnia et al. [5] applied a vertical high-frequency vibration on the friction stir tool to weld aluminum 6061. Effects of four major process parameters on the welding quality were studied, including ultrasonic power, tool rotating speed, traverse speed, and axial force. The vibration power was showed to be the most dominant factor in determining the mechanical properties of the weld.

In addition to laser-assisted and ultrasonic-assisted process, researches have been performed to investigate effects of the electrical current on friction stir welding process. Luo et al. [6] did electrically assisted friction stir welding of similar materials, including Mg alloy AZ31B, Al alloy 7075, and steel Q235B. They found that the resistant heat produced by the electrical current refined the grain size and increased the hardness of the welding nugget. Liu et al. [7] performed electrically assisted friction stir welding of aluminum alloy 6061 to TRIP 780 steel. Two electrodes were placed at the sides and moved along with the welding tool during the welding process. From their study, the axial welding force was reduced.

Since a high-density electrical current is required to pass through the welding zone, the design of the electrically assisted FSW experimental system can be challenging considering the insulation and maximizing the current density. Luo et al. [6] developed an efficient way to maximize the current flow. Regarding the electrodes configurations, one electrode was connected to the welding tool through a brush and the other electrode was connected to the conductive brick attached to the backing plate. Mica sheets were used to insulate the workpiece from the backing plate. The designed current path allowed the current to flow through the workpiece in front of welding, which preheated the workpiece and reduced the welding force. The magnitude of electrical current varied from 0 A to 150 A, which was a relatively small value when compared to that in the resistance spot welding (RSW) process.

Santos et al. [8] designed a dedicated tool and fixture for the electrically assisted friction stir welding of aluminum alloy AA6082-T6. A schematic illustration of their experimental setup. A copper core was inserted in the center of the welding tool and connected to one electrode. The end of the other electrode was connected to copper, which was embedded in the backing plate beneath the workpiece. The applied current flowed along the copper in the welding tool and then passed through the workpiece to the backing plate. The designed current path maximized the current density and Joule heating effect on the welded material, and the corresponding maximum current density reached 800 A. However, one of the possible limitations of this experimental setup is that it could not support a large plunge force, since the hardness of the copper on the backing plate and tool center was insufficient.

In our previous studies [7], electrically assisted friction stir welding of dissimilar materials was performed and the experimental setup. The two electrodes were placed on the top surface of the welding materials and traveled together with the FSW tool during the welding process. In this configuration, the applied current flowed through the two base materials and the FSW tool was only passively involved in the circuit. This system works effectively for butt joint configuration of thin sheets. However, for a spot joint, the current density is insufficient as discussed in the following session.

In this research, a thoroughly improved electrically assisted experimental system is developed for spot joining configuration based on our previous studies. During conventional RSW process, the applied current is usually in the range of 10³ A [9]. As a comparison, the current applied in the hybrid friction stir resistance spot welding in this study is only in the order of 10² A. This relatively low electrical current input aims to avoid bulk melting and to suppress formation of large amount of IMCs.

The base materials applied in this experiment are Al alloy 6061-T6 and TRIP 780 steel, and the chemical compositions of the Al alloy 6061-T6 are listed in Table 1. The thickness of both materials is 1.5 mm, and the aluminum sheet is placed on top of the steel. During the experiment, the tool axis was aligned with the center of the overlapped area. To measure the axial plunge force, a Kistler dynamometer (type 9255ASP) was mounted below the workpiece. The measurement range is –20 kN to 20 kN in the x and y directions and –10 kN to 40 kN in the z direction. Obtained welds were sectioned in the center and then mounted, ground, and polished for microstructure analysis. Optimal microscope and scanning electron microscope (SEM) were applied to characterize the joint cross section. Energy dispersive X-ray spectroscopy (EDS) was employed to determine the elemental distributions at the Al/Fe interface. The welding tool used in this study is made of tungsten carbide with 10% cobalt with the hardness of 91.8HRA. The tool has the feature of a nonthreaded cylindrical pin and a flat shoulder. Other specific dimensions are shown in Fig. 1. An electrical power supply that can provide both direct current (DC) and pulsed current with different frequencies is employed in this experiment. The magnitude of the current can vary from 0 to 1000 A. The pulse on/off times can be set independently anywhere between 0.001 s and 6.5 s. This power supply can also be programmed to output customized waveform profiles. Different forms of current were applied in the experiment for understanding the possibly different effects of the DC and pulses on the joint performance. In this study, the plunge depth is defined as the distance between the end surface of the tool pin and top surface of the original aluminum sheet. Zero plunge depth is where the pin tip starts to touch the aluminum top surface, which is also illustrated in Fig. 1.

The initial investigation of electrically assisted FSSW is made with an experimental setup similar to that of Liu et al. [7]. The overall experimental configuration is shown in Fig. 2. Two electrodes are placed on the two sides of the welding tool, and the applied current flows through the stirring region. To insulate the workpiece from the backing plate, the mica film is inserted below welding materials. Two cases of electrode configurations were investigated and the top view is shown in Fig. 3. On the left side, both the electrodes are placed on the aluminum top surface. On the right side, one of the electrodes on aluminum is moved to the steel surface and a detailed description of the effects of the process parameters on the welding has been made from the previous work [10].

Based on the two configurations, axial welding force is measured and compared between conventional FSSW and electrically assisted conditions. When both the electrodes are placed on the top surface of the aluminum, the result shows negligible change in the peak force. In the other configuration where one electrode is placed on the steel surface, the maximum plunge force decreases from 9.49 ± 0.19 kN to 8.71 ± 0.44 kN, which is also an insignificant reduction.

The initial experimental design did not provide a significant evidence of decreasing the plunge force during the welding process. The possible reason behind is that the current density flowing through the steel is insufficient. To understand the current density distribution in different electrode configurations, a numerical analysis is performed with the comsolmultiphysics program. The model is simplified, which only includes the base materials and the electrodes. Besides, the simulation is based on the initial status of the workpiece and the actual material deformation during FSSW process is not taken into consideration. The contact properties between steel and aluminum, workpiece materials and the electrodes are considered as a perfect contact. One of the electrodes is assigned as a constant current source with current amplitude of 560 A, and the other electrode is treated as ground. The bottom surface of the workpiece, which is in contact with the mica film in the experimental setup, is assumed as electrically isolated. The calculated current density distributions for the initial experimental designs are shown in Fig. 4. It can be observed that a high current density concentrates near the electrode region. Moreover, in both conditions, the current density among the steel sheet is less than 10 A/mm². According to the works from Perkins et al. [11] in electrically assisted forging process of steel, a current density larger than 17.8 A/mm² can result in visible softening effects on the steel. Another research from Liu et al. [12] indicated that the electro-plastic effect cannot be initiated until the current density reaches a threshold value between 7.4 A/mm² and 11.4A/mm². Both of them indicate that the current density in the initial experimental design produces a trivial softening effect for steel during the welding process and therefore the reduction of axial force is insignificant.

To increase the current density in the steel, the experimental system is improved as shown in Fig. 5(a). In this setup, the electrical current is applied to an independent electrode with a ring geometry and the backing plate serves as the second electrode. The ring electrode is placed on the top surface of the aluminum sheet with the FSSW tool in the center. The inner diameter of this ring is slightly larger than the diameter of the welding tool, and the tool is only passively involved in the electrical circuit. During clamping, only the bottom side of the metal ring is in contact with the top surface of the aluminum, and it applies a compressive force for clamping the aluminum and steel sheet. On the top surface of the backing plate, a cylinder with the height of 0.1 mm, diameter of 3.7 mm protrudes out. The center of this cylinder is aligned with the tool axis. The rest region of the backing plate is covered with an insulated mica sheet, which is to guarantee that all the electrical current flows through this cylinder and therefore the current density can be maximized. Regarding the electrodes connection, one electrode is connected to the metal ring, and the other is attached to the backing plate. The current path can be described as following: the electrical current enters into the metal ring to the workpiece from one electrode. The current flows out from the workpiece to the cylinder protrusion on the backing plate, which is connected to the other electrode. The current flow path during the welding process is also illustrated in Fig. 5(b).

The current density distribution for the new fixture design is again analyzed using the comsolmultiphysics software, and the corresponding calculated result is shown in Fig. 6. Under a DC input of 560 A, the maximum current density reaches 73.5 A/mm² at the bottom surface of the steel. The calculated current density should therefore be capable of inducing the softening effect on the steel according to the Perkins et al. [11] and Liu et al. [12]. All the following results and discussions are based on this experimental configuration. The actual setup and welded sample are shown in Fig. 7. Geometry and dimensions of the welding tool are previously shown in Fig. 1.

To investigate the effects of the electrical pulses on the FSSW process, the magnitude of the pulse is selected at 900 A with a turn on time of 1 ms and duty cycle of 50%. Preliminary experiments were performed where the electrical current was applied under conditions of different tool rotational and plunge speeds. The reduction of the welding force was shown to be more significant under lower rotation speed and faster plunge speed. The final process parameters for a detailed comparison between electrically assisted and conventional FSSW process are listed in Table 2. The current is turned on immediately as the tool starts to touch the aluminum top surface, and it is stopped at the end of the plunge stage before the dwell stage.

where $I$ is the magnitude of the electrical pulses and $Idc$ is the magnitude of the equivalent DC. The total time duration $ttot$ is selected as the length of plunge stage. During experiments, the actual pulses are associated with a ramping up period and can hardly achieve an ideal rectangular shape. In order to more accurately calculate the amplitude of the equivalent DC current, the applied pulse signal is measured and the corresponding recorded current is shown in Fig. 9. Based on Eq. (2), the equivalent $Idc$ is determined to be 560 A.

The equivalent DC is then applied during the FSSW process, and the corresponding axial welding force is compared with that from the pulse condition in Fig. 10. It can be noticed that the difference between the two curves is trivial and the calculated mean absolute percentage difference is 7.42%. The maximum plunge force for 560A DC is 13.47 ± 0.05 kN while that for the pulse condition is 14.10 ± 0.01 kN. This indicates that with the equal amount of energy input, the force reduction is approximately the same regardless of the form of the electrical current. In other words, thermal effect dominates and there is no strong evidence to support the existence of the electro-plastic effect during the FSSW process. One of the possible explanations is that the material is deformed under a combined condition of high pressure, large strain rate, and high temperature, which impairs the electro-plastic effect on the steel.

An overview of a typical joint cross section is provided in Fig. 11(a). The marked area in rectangle is enlarged in Fig. 11(b), where a groove line can be observed in the thermo-mechanically affected zone (TMAZ). This line is located at a certain distance away from the tool shoulder. To identify the composition of this groove feature, an EDS analysis is performed along a straight line from point A to point B, as shown in Fig. 11(c). The analysis results are given in Fig. 12 and a Zn peak can be observed at the groove line. The Zn comes from the coating of the TRIP steel sheet. At the interface between the bottom surface of aluminum and the top surface of steel, the Zn coating is likely to be bonded to the bottom surface of aluminum under the combined state of high pressure and temperature during the plunge stage. The mixed Zn and aluminum then move together as the tool moves further down. In this perspective, Zn can be treated as a tracer for the material flow in the bottom region of the aluminum sheet.

Figure 13 further compares the region of Fig. 11(b) between traditional FSSW process and electrically assisted conditions. In Fig. 13(a), the Zn line starts from the bottom of the aluminum sheet and then extends upward. Regarding the aluminum flow during the welding process, the top region of aluminum is pushed outward as the welding tool plunges into the materials. When the flowing aluminum approaches the relatively cold heat affected zone, it moves upward due to the constraints of cold materials that have a higher deformation resistance. The material flow in the bottom region follows a similar pattern. For a quantitative description of the amount of material flow during the welding process, the distance L is measured from the outer boundary of the tool shoulder to the Zn line and it is also parallel to the original Al/Fe interface, as shown in Fig. 13(a). Comparing Fig. 13(a) with Fig. 13(b), a larger distance L is observed when there is a 560 A DC applied during the welding process, which has a value of 0.63 ± 0.04 mm and that without the current is 0.41 ± 0.03 mm. The larger value of L indicates a better material flow in the electrically assisted condition.

The joint shear strength of the weld samples obtained from conventional, direct current-assisted and pulse-assisted conditions are compared in Fig. 14. According to literature, for conventional RSW of steel to aluminum, the reported joint strength has an average value of 3 kN [14–16]. It can be observed that overall, the solid-state friction stir spot welding process can achieve a better joint quality. Application of the direct current increases the joint strength by around 43% while the pulses increase that by 28%. This improvement can be contributed from different interaction mechanisms between aluminum and steel during the conventional and electrically assisted welding process.

Figure 15 schematically illustrates the hook formation and material interactions at the boundary of the hook during the welding process. As the welding tool moves downward, the tool shoulder compresses the aluminum matrix, and the steel is extruded upward by the tool pin. As the tool reaches a deeper position, the aluminum is pushed toward the outside and more amount of steel is extruded to generate a larger hook. The new layer of extruded steel flows along the edge of the existing hook. In the meantime, a portion of the aluminum is embedded below this new steel layer. The embedded aluminum reacts with the steel under a certain temperature and pressure state, which results in the formation of IMC. As the tool plunge depth increases and more layers of steel are extruded, a periodic vortex morphology can be observed at the top side of the hook. Figure 16 shows the interaction layer between steel and aluminum of the conventional FSSW joints.

As the plunge depth further increases, a larger hook is formed and less amount of aluminum is allowed between tool shoulder and hook based on the geometrical constraints. It can be illustrated in Fig. 17(a), where the inside area is labeled as 1 and outside is labeled as 2. On the other hand, at this specific tool position, it is difficult for the aluminum surrounding the tool pin to flow outside. Accordingly, this amount of aluminum is compressed downward into the hook. At the same time, the extruded steel is pushed upward by the tool pin, which leads to a severe mixture between steel and aluminum at the inside region of the hook, as shown in Fig. 17(b). The mechanical mixing of the base materials also generates a large deformation resistance.

where the first term represents atomic flux induced by electrical current and the second term corresponds to that driven by the composition gradient. $Di$ is the diffusion coefficient, $Ni$ is the mole fraction of each element, T is the temperature, k is the Boltzmann constant, $zi*$ is the effective charge of atom, e is the charge per electron, $ρe$ represents material electrical resistivity, and $j$ is the current density. To compare the atom diffusion in these three welding conditions, outside boundary of the hook structure, marked as region 2 in Fig. 17(a), is further examined with a higher magnification using SEM. The results are shown in Fig. 18.

Under the nonelectrically assisted condition, a certain amount of aluminum penetrates into the steel matrix and forms an intercalation structure as shown in Fig. 18(a). This can be explained based on the mechanisms shown in Fig. 15 that the aluminum is embedded under a new layer of the extruded steel. After application of the direct current, the steel matrix is relatively intact. The penetration depth of aluminum into steel is much smaller, marked out as the region A in Fig. 18(b). However, the thickness of the IMC layer at the Al–Fe interface increases. These observations can be explained based on the hook formation mechanisms illustrated in Fig. 15. The additional Joule heat from electrical current enhances the flow of steel materials and softens the extruded steel, which ensures a smoother fusion between the newly extruded steel and the existing hook and restrains the penetration of aluminum. On the other hand, the associated higher temperature promotes reaction between aluminum and steel and accelerates growth of the IMC layer.

The Al–Fe interface under the pulse condition is shown in Fig. 18(c); the observations are similar to those obtained from the direct current condition. The penetration depth of aluminum into the steel matrix is smaller compared with the conventional FSSW process. On the other hand, the thickness of the IMC layer is even larger than the direct current condition. This indicates that the electromigration effect is more phenomenal from electrical pulses compared with that from the direct current with the same amount of total energy input. Diffusion rates of the atoms are increased due to the electromigration effect at the pulse condition, which forms a thicker IMC layer. In return, the thicker IMC layer accordingly results in a lower joint strength.

Where Z_m is the depth of the X-ray excited volume in um, E₀ is the acceleration voltage in keV, E_c is the critical excitation voltage in keV, and $ρ$ is the material density in kg/m³, Z is the atomic number, A is the atomic mass, D is the spatial resolution in um. The acceleration voltage used in the EDS analysis is 15 keV. Based on this calculation, the EDS spatial resolution for Fe is around 0.92 μm while that for Al is 2.78 μm. Accordingly, the IMC layer with thickness of less than 1 μm is hardly distinguishable as a horizontal line in the EDS composition analysis.

Similar patterns can be observed in Figs. 19(b) and 19(c) as well. After the IMC region B and C, the Fe composition increases a little. Both of which can be referred to the SEM images in Fig. 18, where the EDS analysis line passes though the isolated small piece of Fe. After that, the Al–Fe composition then gradually changes suggesting an inter diffusion transition layer with Al–Fe mixture. Still, this could also be induced by the insufficient resolution of EDS.

The compositions of the IMC layer are different under the three conditions. In the standard FSSW process in Fig. 19(a), the IMC layer has a thickness of around 2 μm and the corresponding weight percentage of aluminum is around 42%. After application of the direct current, the IMC thickness increases to 6 μm, and the weight percentage of aluminum is approximately 60%, as shown in Fig. 19(b). For the pulse condition, the IMC thickness is 15 μm and the percentage of the aluminum is 60%, which is approximately equal to that of the DC condition. This indicates that the compositions of the IMC layer from the direct and pulsed current assisted processes are the same. But the electromigration effect increases the diffusion rate and results in a thicker IMC layer.

At the inside of the hook, the aluminum is heavily mixed with steel under a large compression force from the tool shoulder in the conventional FSSW joint, as shown in Fig. 17. The corresponding position from the electrically assisted condition is shown in Fig. 20. It can be observed that the amount of aluminum mixed with steel is relatively smaller. This phenomenon can be explained similarly to the aluminum flow behavior at the TMAZ: The electrical current enhances the flow of aluminum materials that are located between the tool shoulder and hook. A larger fraction of aluminum is therefore capable of flowing outward, which results in less aluminum mixing with the inner side of the hook. The reduced mechanically mixing between aluminum and steel can also lead to a smaller plunge force.

This paper studies a hybrid friction stir resistance spot welding process for joining aluminum alloy 6061-T6 to TRIP 780 steel. An effective electrically assisted FSSW experimental system has been developed, which enables a high density electrical current in the weld zone with the tool passively involved in the circuit. Compared with the conventional RSW process, the applied current is lower and avoids bulk melting. Compared with conventional FSSW process, the axial plunge force reduces by 12.84% and the joint strength increases by 43% with application of the direct current. The electro-plastic effect, which is the direct material softening effect from electrical current, is studied by comparing direct and pulsed current conditions, where the total amount of energy input is equal. Regardless of the form of the electrical current, the force reduction is the same, which indicates the electro-plastic effect is insignificant and Joule heating is the dominant material softening mechanism during the FSSW process. Microstructure analysis on the joint cross section reveals that the electrical current enhances the material flow of aluminum and a more uniform hook is generated at the Fe/Al interface. The penetration of aluminum into steel is smaller and the intercalation structure is less obvious. However, the diffusion rate is increased with the electrical current due to the electromigration effect and an IMC layer with larger thickness is observed. The thickness of the IMC layer is larger in the pulsed condition compared with the direct current condition.

Special thanks to the United States Steel Corporation for providing the TRIP 780 steel as research propose.

• National Science Foundation (Grant No. 1537582, Joining of Dissimilar Materials through a Novel Hybrid Friction Stir Resistance Spot Welding Process).