Friction stir welding (FSW) is a solid-state joining technique which can be used for joining not only similar or dissimilar aluminium alloys but also between aluminium alloys and other high melting temperature alloys. Welding of an aluminium alloy to a high melting temperature alloy with traditional fusion welding processes is not suitable as it results in excessive growth of brittle intermetallics in the weld region. Recent studies by many researchers have provided well established data showing superior tensile and shear properties of Al to Ti FSL welds in comparison to other welds of aluminium alloys to other high melting temperature alloys made using FSW and tested in static loading. The process parameters have been often used to test the efficiency of FSL welds in order to improve their tensile and shear strength properties of Al to Ti FSL welds. However, thus far, there is a total lack of study on the fatigue properties of Al to Ti alloy welds made using FSW which has prevented Al to Ti FSW to be applied industrially. As a result, the study of fatigue behaviour of Al to Ti friction stir lap welding (FSLW) is important for establishing useful data for its various applications that engineers, designers and manufactures need. Lap weld (FSLW) is another major weld geometry widely used. Thus, the primary aim of this research was to determine the fatigue properties and investigate the fatigue behaviours of AA2024 to Ti6Al4V alloy welds made using FSLW at wide range of stress amplitudes. The alloys used were aerospace aluminium and titanium alloys and thus the study is particularly relevant for applications in aerospace industry.

Experimentally, fatigue tests have been conducted on welds of AA2024 (A14.5Cu1.4Mg0.5Mn) alloy to Ti6Al4V alloy welds made using FSLW, using a tool with a pin diameter of 6 mm. Pin bottom aimed for touching the Ti6Al4V plate but not penetrating (dP ≈ 0) although it was inevitable that pin could also readily penetrate (dP > 0) slightly with the FSLW being position controlled. FSLW experiments were also conducted by placing thermocouples at the weld interface regions so that temperature cycle data during FSLW could be obtained. Two other series of experiments have also been conducted, one refers to using a large pin and the other refers to using an interlayer so that dP = 0 is more certain. For the large pin experiment, a lower rotational speed was required to avoid insufficient stir flow and the top tunnel defects. After FSLW experiments and fatigue tests, microstructures in the weld interface region and on fracture surfaces have been studied in detail. This includes samples tested with crack growing during fatigue testing “frozen” so that the crack growth behaviours relating to the weld structures could be better observed and studied.

It has been found that, for both dP  0 and dP > 0, fatigue limits of the FSL AA2024/Ti6Al4V welds were slightly higher than the fatigue limits of the FSL Al to Al alloy welds reported in literature. Examination of cross sections of samples with crack growth interrupted thus “frozen” during testing and fracture surfaces of the tested samples have demonstrated diffusion welding forming the very thin interface intermetallic layer during FSLW. This thin layer being the major mechanism responsible for the good fatigue strength of the welds. For dP > 0, a significantly larger diffusion AA2024/Ti6Al4V weld distance outside the pin width than that for dP  0 welds have been observed. This was the reason for the fatigue limit of dP > 0 welds being comparable to the fatigue limit of dP  0 welds despite of the mix stir zone (MSZ) in dP > 0 welds being a brittle one due to excessive growth of intermetallics. The different thermomechanical conditions for the different diffusional weld widths will be illustrated. As the pin diameter was increased from 6 mm to 9 mm the total width of the AA2024/Ti6Al4V FSL weld was also increased for dP > 0. However, it has been found that there was no significant increase in the effective weld width outside the MSZ. The larger pin has also been found to readily penetrate more, having a detrimental effect that the deep MSZ could cause fracture in the Ti6Al4V plate. The effective weld width for dP ≈ 0 has also been found not to have increased, compared to that of the weld made using a normal size pin. This is likely due to the result of using a lower pin rotational speed, for avoiding top weld channel defects, so that effective diffusion welding is not enhanced. Thus, the fatigue performance of weld made using large size pin has been found to be comparable to the fatigue performance of the welds made using the normal size pin.

Using an Al interlayer, together with the standard size tool pin, has been found to have provided no benefit for increasing the fatigue strength, for both 8 mm and 10 mm wide interlayer. The reason for this has been found to be the similar diffusion weld width, as the stir flow width and possibly the temperature history are the same with or without the use of interlayer when the pin does not penetrate the bottom Ti6Al4V plate. Analysis of the fracture surfaces have confirmed that the thin interface intermetallic layer having been formed, thus still enabling comparable performance of AA2024 to Ti6Al4V FSL welds made with Al interlayer.