Metallurgical studies of friction stir lap welding
Friction stir welding (FSW) is a solid-state joining technique which can be used for joining not only traditionally weldable aluminium alloys but also high strength aluminium and other metallic alloys that are hard to weld using conventional fusion welding processes. During FSW a rotating tool consisting of a shoulder and a pin is first plunged into the joint line of two plates to be welded. Intensive heat is generated as the tool rotates and travels against the plates, causing the latter to be softened without gross melting. The softened material is then stirred and deposited at the back of pin as the tool traverses the joint line, hence making a joint. The process of conducting FSW in lap joint configuration is called friction stir lap welding (FSLW).
Mechanical strength of FSL welds under static loading is commonly determined using tensile shear testing and fracture strength (σLap) corresponding to the maximum load in a test over the sample width is widely used the strength value. During FSLW of similar metallic alloys, a part of the original lapping interface between the plates is lifted up due to a specific material flow induced by the rotating threaded tool pin, taking a hook shape. Such a hook in a FSL welds often provides a favourable site for crack propagation under loading and thus adversely affects σLap. Furthermore, FS heat causes local softening but grain refinement by dynamic recrystallisation contributes to hardening in stir zone. Thus strength is location dependent. In FSLW of dissimilar metallic alloys with large differences in melting temperatures, a metallurgical bond is established through the formation of interfacial intermetallics. However, as these intermetallic compounds are generally believed to be brittle with limited ductility, they are commonly viewed to adversely affect σLap.
the aim of the present research is to provide thermomechanical explanation on how ω and ν affects hook formation during FSLW; to understand how hooking, FS softening, stress concentration and local deformation mechanisms (under loading) relate to σLap of Al and Mg FSL welds; and to reveal how the interface structure is affected by FS conditions and how the formation of interface structure affects σLap of Al/Steel and Al/Ti FSL welds.
For FSL welds of Al 6060-T5, it was found that low penetration required low v for a sufficient bonding. For sufficient penetration, (, v)-AB-SZ-h relationships were presented. Measured data have suggested that the increase of AB-SZ as increases is the result of increase in TSZ and thus the stir zone plasticity. A rapid increase in h to a maximum value (hMax) when AB-SZ increased from a minimum value was identified. Evidence of shoulder flow limiting the hMax will be shown and explained. When h tended to zero, despite of the existence of an un-welded lap and thus a high stress concentration, σLap (422 N/mm) was very close to BoP (fracture strength of butt joint geometry). This is due to local bending offered by the high ductility of the Al FCC structure thus reorientating to reduce considerably the stress concentration. It was found that when h < ~ 30% of the plate thickness (tPlate), σLap was not strongly affected by h. This surprising result has been explained by considering hook shape, hook discontinuity and FS softening that competes with hooking for the local deformation and fracture. When h > ~ 30% tPlate, it started to invoke a significant effect on reducing σLap, due to the increasingly larger reduction in load bearing area.
It will be demonstrated that FS flow and the subsequent tensile-shear mechanical behaviours of magnesium alloy AZ31B-H24 are significantly different from those of Al 6060-T5 alloy. For AZ31B-H24 alloy, FS zone was comparatively smaller and there was little discontinuity in each hook. It will be shown that due to lack of local sample rotation during tensile shear testing, as a result of low number of slip systems offered for plastic deformation by Mg alloy, high stress concentration of the lap joint geometry was maintained (during testing) causing failure with a significantly lower σLap value (290 N/mm), when h tended to zero. Also for the same reason, unlike Al 6060-T5 welds, FS softening could not compete with hooking for local deformation and fracture. Thus hooking location was always the location of fracture. Results of numerical modelling and artificial hook testing experiments showed that stress distribution at hook region (during tensile shear testing) was considerably affected by the orientation of hooking. This explains that fracturing proceeds away from the hook, for negative hooking samples. Significant stress concentration caused by tensile shear loading enhanced the localised operation of twinning at the hook region, facilitating a fracture in brittle manner.
Finally, detailed evidence has suggested that the major factor determining σLap of Al/Steel and Al/Ti FSL welds is the degree of contact between the bottom of the pin and surface of bottom plate (steel or to Ti6Al4V ) during welding. Insufficient contact resulted in discontinuous intermetallic layer in the form of outbursts. This type of interface will be shown to resist loading poorly and thus low σLap was obtained. When the pin was at a very short distance or just touching the bottom plate, a thin intermetallic layer formed at the interface. Further pin penetration into the bottom plate resulted in the formation of a mix interface region consisting of cut layers of the bottom plate material and irregular intermetallic layers. The mixed layer interface region corresponded to reasonable weld strength (σLap≈ 300 N/mm in Al/Steel FSLW and σLap≈ 340 N/mm in Al/Ti FSLW) consistent with data in literature. It will be demonstrated that a thin intermetallic layer formed under the condition of carefully positioning the pin very close or just touching the bottom plate (without severely penetrating) provided the most unfavourable crack propagation path. As a result, high strength joints with σLap=435 N/mm in Al/Steel FSLW and σLap=732 N/mm in Al/Ti FSLW was achieved.