*3.3. Mechanical Performance and Failure Analysis*

To evaluate the mechanical performance of the ultrasonic spot welded NiTi joints using Cu interlayer, tensile shear tests were conducted. Figure 6 summarizes the load-displacement curves for all welding conditions. Additionally, the load-displacement curve for a NiTi/NiTi weld obtained without Cu interlayer for a 1000 J of energy is also added for comparison.

**Figure 6.** Load-displacement curves of NiTi joints with and without Cu interlayer at different welding energies.

The failure loads of all joints were defined as the peak tensile load on the load-displacement curve, with each joint representing its own weld strength under certain welding conditions based on the joint performance results. The failure load of the base metal with 0.15 mm thickness and 8 mm width, used in this present work is about 810 N. As shown in Figure 6, it is noticeable that a significantly higher failure load of 520 N was achieved at a welding energy of 1000 J. This higher failure strength is in accordance with the result of weld interface morphology of the 1000 J joint in Figure 4, where good bonding between NiTi and Cu was obtained due to sufficient mechanical interlocking and metallurgical adhesion as a result of a more concentrated vibratory energy. It further indicates that the weld joint with welding energy of 1000 J has the good load capacity, which can be used in distinct engineering fields. Furthermore, the use of the Cu interlayer can improve the mechanical properties of the welded joint: when no Cu interlayer is used the fracture load of the 1000 J weld is of approximately 200 N [45], which is significantly lower than the 520 N obtained when the Cu interlayer was used.

In this present work, the fracture location of all the joints occurred at the welded interface. Figure 7a,b show the schematic diagram of the fracture mode and the overall image of the fracture surface of the tensile test sample, respectively, which exhibits the interfacial fracture mode with some welded spots clearly observed. The dominant failure mode was characterized by a tearing behavior along the weld interface: it can be observed that the Cu foils present some tearing features on the fracture surface. In addition, to further understand the weld behavior and failure mode of ultrasonic spot-welded joints, the micro morphology of the corresponding fracture surface of 1000 J joint after tensile shear tests were observed by SEM. The EDS point scan analyses were performed on different areas of fracture surface, with these results presented in Figure 7e,f,h.

The judging criteria for a good weld quality could can be determined by the mechanical properties of the weld but also by the observation of developed weld spots on the fracture surfaces. Three distinct regions were observed on the fracture surface, as exhibited in Figure 7c: weld spots, scratched regions and tearing regions of Cu. The magnified images of the inserts indicated in Figure 7c suggested that on the ultrasonic spot welded NiTi joints with Cu interlayer, more significant deformation was observed under indentations of the sonotrode tip and anvil. As can be seen from Figure 7d, the fracture surface near the periphery of the weld spots was much smother and presented plastically deformed zones, while tear ridges and dimples can be observed at the center, as shown in Figure 7e,f, suggesting that the interface bonding strength near the periphery of weld spot was lower than that at the center of the weld interface.

**Figure 7.** (**a**) Schematic diagram of fracture mode; (**b**–**h**) fracture morphologies of the tensile failed sample made at a welding energy of 1000 J; (**b**) overall view of fracture surface; (**c**) overall view of fracture surface by SEM; (**d**) magnified image of a weld spot depicted in (**c**); (**e**,**f**) magnified image of box in (**d**); (**g**) overall view of Cu interlayer on the fracture surface; (**h**) magnified image of box in (**g**).

Brittle fracture characteristics can be observed in the weld spot zone as the high-magnification SEM image presented in Figure 7e reveals. Furthermore, the features in the fracture region of the NiTi weld spot exhibited smooth step patterns and river marks with cleavage-like characteristics. The EDS point analysis showed that this region had a composition of 50.1 at.% Ni and 49.9 at.% Ti, which is good agreement with the expected composition for the as-received NiTi BM. At the weld interface, some intergranular cracking can also be observed due to the strain incompatibility between different grains, which is similar to that observed in NiTi-Cu dissimilar laser welds [15]. In addition, fine dimples also existed in the ductile fracture zone of the weld spot, as shown in Figure 7f. The EDS result shows the composition of dimples in weld spot zone was 3.6 at.% Ni, 3.6 at.% Ti and 92.8 at.%

Cu, indicating that ductile failure occurred mainly in the position of softer Cu foil. This can justify the increasing load capacity of the joints obtained with the Cu interlayer: when Cu is used as an interlayer part of the deformation is accommodated by it which provides better mechanical properties than when no interlayer is used.

An overall view of the Cu interlayer on the fracture surface is presented in Figure 7g, and obvious tearing characteristics of Cu can be observed. The magnified image of the insert presented in Figure 7h consisted of both scratched regions and plastically deformed zones. Compared with the NiTi surface, compact dimples existed in the position of Cu interlayer, suggesting that shear fracture occurred through void nucleation, growth and coalescence. During the tensile shear process, a higher tensile load can be transmitted to the 1000 J joint due to a good interface combination, resulting in a higher ultimate tensile load, which is in accordance with the tensile results presented in Figure 6. Additionally, in this study, considering the mechanical performance of joints previously discussed, it is believed that the addition of Cu interlayer into the faying surfaces is critical to enhance the bonding strength due to the increase of friction coefficient [46,47].

The SEM images of the fracture surfaces also verified that the joint was composed of micro welds between NiTi and Cu interlayer. It is possible that the fracture originated from the weak regions in the weld spot border, since this is a suitable location for nucleation and propagation of a crack, and when the crack tip reaches at the boundaries of weld spot, it can easily propagate through it, resulting in the fluctuation of tensile loads due to the existence of the multiple spots connection mechanism of ultrasonic spot-welded NiTi with Cu interlayer, as shown in Figure 6. Meanwhile, the Cu foil between NiTi began to tear, contributing to the process of tensile test until the fracture failure, with the joint being able to deform even more.

X-ray diffraction analysis was carried out at room temperature to evaluate the influence of USW on the phases composition of NiTi weld. The diffraction patterns of both BM and of the fracture surface in the center of the 1000 J weld are depicted in Figure 8.

**Figure 8.** XRD patterns of NiTi BM and fracture surface of 1000 J weld.

The indexed pattern of the NiTi BM only consisted of B2 cubic austenite, without traces of B19' monoclinic martensitic phase, indicating the NiTi BM would be fully austenitic at room temperature. Comparing to the NiTi BM, the fracture surface of ultrasonic spot welded NiTi weld exhibited an additional pure Cu phase due to the addition of the Cu interlayer into the NiTi interface. It is noteworthy that no intermetallic phases, such as Ti2Ni or Ni3Ti which are usually formed during some fusion welding processes of NiTi [17–20], were detected in the weld region. Furthermore, the XRD results are consistent with the EDS line scan in Figure 5 where no intermetallic compound layer was found. It is believed that restricting the formation of the brittle phases at the weld interface is beneficial to reduce weld embrittlement, which can further prevent the formation of cold cracking [44]. In addition, the formation Cu-based intermetallics was not observed in this study either, although in fusion-based welding of NiTi to Cu they were already reported [13,15].
