*3.2. Bonding Morphology of USW Joints*

The bonding morphology at the weld interface of the welds obtained by varying the welding energies was examined in the cross section along the center of the joints parallel to the vibration direction as presented in Figure 4.

**Figure 4.** (**a**,**b**) interface morphology of the welded joint obtained at 500 J at the center (**a**) and near the periphery (**b**) of the joint; (**c**,**d**) interface morphology of the welded joint obtained at 700 J at the center (**c**) and near the periphery (**d**) of the joint; (**e**,**f**): interface morphology of the welded joint obtained at 1000 J at the center (**e**) and near the periphery (**f**) of the joint; (**g**) microstructure interface of weld of 1000 J obtained by optical microscope. The white lines in Figure 4. (**a**,**c**,**e**) represent the locals where EDS line scans were performed.

Figure 4a,b, respectively, depict the interface morphology in the center and near the periphery of the weld obtained for a welding energy of 500 J. From Figure 4a it can be observed that the weld interface showed a wave-like pattern and the thickness of the Cu interlayer was uneven due to the effect of plastic deformation on the material surface, implying the occurrence of material flow at the interface of the deformed material. Under the position of indentation of the sonotrode tip in

the material, the thickness of Cu interlayer at the weld interface was thinner than that of the anvil. The weld interface was noticeably intact, resembling a friction-induced bonding characteristic [39]. Joining of NiTi with the Cu interlayer took place by the formation and growth of micro-welds at the weld interfaces owing to the close metal-to-metal contact, which produced mutual diffusion and metallurgical bonding along the interface [33,38,40]. Additionally, unbonded regions were observed near the periphery of the weld, indicating insufficient diffusion during the process. During the USW process, the highest temperature was located at the central area of the weld, thus the bonding between the NiTi and the Cu interlayer was significantly more effective close to this location that in the border of the weld [37]. In comparison, the welded joint obtained with a welding energy of 700 J, which is shown in Figure 4c,d, also presented unbonded gaps along the border of weld interface were also visible. However, these were less often observed due to increased diffusion occurring as a result of the increasing temperatures generated by the increase in the welding energy during the process [39,41].

Observations of the cross-section of the joint obtained at 1000 J showed that both at the center and periphery of the joint complete bonding was obtained, as depicted in Figure 4e,f. It can be observed that the joining interface in Figure 4f was better than that in Figure 4b,d, since no unbonded zones were observed. This can be explained by the greater plastic deformation and increase in temperature that occurred under the indent tips at both the sonotrode tip and anvil sides when 1000 J of energy were used. These results indicate that a good interfacial bonding was obtained during the NiTi with Cu foil USW process.

Figure 4g showed that NiTi and the Cu interlayer were complexly intertwined in the visible weld interface obtained for a welding energy of 1000 J, which contributes to a better bonding along the weld. During the USW process, under the clamping pressure and ultrasonic vibration, the NiTi BM adhered to the Cu foil and then shear deformation, as well as, mutual rubbing of the faying surfaces occurred. Some works [36] have shown that the strength of ultrasonic spot-welded joints is related to interfacial waves, mechanical interlocking and microbonds produced along the weld interface. From the results presented above, it can be concluded that with increased welding energy, more shear friction heat and plastic deformation energy were generated at the weld interface thus promoting joining between the NiTi BM and the Cu interlayer. Mutual extrusion and abrasion between the faying surface of NiTi and the Cu interlayer was a main source for friction heat and plastic deformation. The presence of unbonded zones in the weld interfaces obtained for lower welding energies, would promote premature fracture of joints during tensile testing, since when applying the load, the unbonded areas would become the starting points for crack initiation and propagation until failure of the joint occur.

The EDS line scan analysis at the weld interface center of different joints obtained by various welding energy conditions (the test positions for EDS analysis were depicted with white lines in Figure 4) were conducted to determine the chemical composition and to infer about the potential phases in the interface diffusion layer formed between NiTi and Cu. The results are shown in Figure 5.

It is apparent that the content distribution of Ti and Ni followed the same trend, which is opposite to the change for Cu: a decrease in either the Ni or Ti content would be compensated by an increase of Cu. The EDS line analysis results revealed a smooth and rapid change in composition of both NiTi and Cu across the weld interface border. As indicated by the ellipse in Figure 5a, changes in relative contents of NiTi and Cu elements suggests the onset of slight diffusion at the weld interface. However, the width of the diffusion zone at most weld interfaces was too small to put into evidence the formation of any IMCs layer. Therefore, the joining mechanisms of NiTi with Cu interlayer by USW can be concluded as the combination of solid-state shear plastic deformation, mechanical interlock, the formation and further expansion of micro welds, which was consistent with other previous studies [36,42,43]. Despite the element content fluctuation, it can be seen that the peaks and valleys in the element distribution profiles kept mostly stable in the rich Cu area, as shown in Figure 5a,c,e.

**Figure 5.** Chemical composition across the weld interface of NiTi joints with Cu interlayer: (**a**,**b**) 500 J; (**c**,**d**) 700 J; (**e**,**f**) 1000 J. The positions where the EDS line scans were performed are indicated by the white lines in Figure 4 (**a**,**c**,**e**).

The EDS analysis results also showed that the maximum content of Cu in the thinner Cu area (as depicted in Figure 4) gradually decreased with increasing welding energy, as presented in Figure 5b,d,f, with values of around 89.4%, 79.5% and 32.7%, respectively. It can be reasoned that more Cu foil was squeezed out at some positions of the weld interface with increasing welding energy. Since Cu is a soft metal with high conductivity and low yield strength, it is believed that the existence of Cu layer in the weld can compensate the thermal stress produced during the USW process [44].
